Psychology Library Editions: Comparative Psychology, 16-Volume Set 9781138503298, 9781351128780, 9780815369363, 9781351252546

Citation preview

PSYCHOLOGY LIBRARY EDITIONS: COMPARATIVE PSYCHOLOGY

Volume 1

ANIMAL PSYCHOLOGY

ANIMAL PSYCHOLOGY Its Nature and its Problems

J.A. BIERENS DE HAAN

First published in 1948 by Hutchinson & Co. This edition first published in 2018 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 1948 J. A. Bierens de Haan All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: ISBN: ISBN: ISBN:

978-1-138-50329-8 978-1-351-12878-0 978-0-8153-6936-3 978-1-351-25254-6

(Set) (Set) (ebk) (Volume 1) (hbk) (Volume 1) (ebk)

Publisher’s Note The publisher has gone to great lengths to ensure the quality of this reprint but points out that some imperfections in the original copies may be apparent. Disclaimer The publisher has made every effort to trace copyright holders and would welcome correspondence from those they have been unable to trace.

ANIMAL PSYCHOLOGY Its Nature and its Problems by Dr. J. A. BIERENS DE HAAN, c.M.z.s.

SECRETARY OF THE DUTCH SOCIETY OF SCIENCES. LATE LECTURER IN EXPERIMENTAL ZOOLOGY IN THE UNIVERSITY OF AMSTERDAM.

HUTCHINSON'S UNIVERSITY LIBRARY 47 Princes Gate, London New York

Melbourne

Sydney

Cape Town

THIS

VOLUME

IS

NUMBER

15

IN

HtJTCHINSON 1 S UNIVERSITY LIBRARY

THIS 'SOOK IS PRODUCED IN COMPLETE CONFORMITY WITH THE AUTHORIZED ECONOMY STANDARDS.

Printed in Great Britain by Burrow's Press Ltd., Cheltenham and London.

H

200~

CONTENTS Page vi

Preface Chapter I

The Problem of Animal Psychology

7

The Problem of Animal Instinct

37

III

The Problem of Animal Learning

90

IV

The Problem of Animal Intelligence

94

II

v VI VII

The Problem of Animal Understanding

II2

The Problem of Animal Ideation

141

The Problem of the Animal's World

147

Bibliography

1SS

Index

IS9

PREFACE T H E reader will not expect to find in the pages of a book of such a modest size as this a full account of all facts known in the field of animal psychology and of all theories built on these facts. The author has rather preferred to follow, and to work out, one particular line of thought, to wit : the idea that the instincts; as they are defined on page 38 of this book, are the spring and basis of all animal behaviour (with the exception perhaps of play), and therewith the core of the animal's mind, and that individual experience, gathered by the animal in the course of its life, may influence and. reconstruct these instincts, so as to guide, in the form of intelligence and understanding, this behaviour along new (i.e., not innate) paths. Thus, instinct and experience become the pillars upon which animal behaviour is built up ; instinct, intelligence, and understanding form a triad round which the facts of the psychology of animals may be grouped. As a foundation of all this the author first tries to prove the good right of a real and genuine animal psychology, not hampered by objectivistic and behaviouristic scruples, while in a final chapter, by way of conclusion, he tries to give an image of how the world of the animal is built up. Along this road the reader will meet the principal facts of animal psychology and, for the rest, may add to them by studying some of the works mentioned at the close of this volume. The author expresses the hope that this way of treatment may serve its purpose and form an introduction to the fascinating science of the psychology of animals. Amsterdam, Christmas 1946.

CHAPTER

I

THE PROBLEM OF ANIMAL PSYCHOLOGY Tschuang-Tse and Hui-Tse were standing on the bridge across the Hao river. Tschuang-Tse said: "Look how the minnows are shooting to and fro ~ That is the joy of the fishes." "You are not a fish," said Hui-Tse, "how can you know in what the joy of the fishes consists ?" "You are not I," answered Tschvang-Tse, "how can you know I do not know in what the joy of the fishes consists ?" "I am not you," Hui-Tse conceded, "and I do not know you. All I know is that you are not a fish ; therefore you cannot know the fishes." Tschuang-Tse answered : "Let us return to your question. You ask me : 'How can you know in what the joy of the fishes consists ?' Essentially you knew that I know, and yet you asked me. No matter: I know it from my own joy of the water." The old Chinese Tschuang-Tse. (Quoted after Hempelmann.) WHOEVER for some moments has attentively watched the behaviour of an animal, be it that· of his dog who accompanied him on a walk through the fields, or of a blackbird singing in his garden, or even of an ant going along an antpath in the wood, will undoubtedly have felt the question arise in him : What are the springs of that animal activity ? Why does that dog suddenly start barking under that tree ; why does that bird suddenly fly away ; why does that ant all at once turn back, while there is no perceptible reason why it should not go straight on? And, further, what is the nature of these springs of activity ? Are the actions of animals purely mechanical by nature, and is the animal, therefore, wholly comparable to a finely constructed machine, differing from a motor-car or musical-box only in that it is born and

8

ANIMAL

PSYCHOLOGY

feeds itself and dies when its time has come ? Or are their actions built up from an integration of reflexes, of the same nature as the contraction of our pupil when light falls on the eye, or the secretion of the glands of our stomach when we see or taste food, vital phenomena indeed, but occurring beyond a consciousness that experiences them ? Or may we admit that the animal in his actions is guided by something comparable to the inner experiences that underlie our own deeds ; by feelings, perceptions, desires, perhaps even by understanding and judgment ? Mostly the spectator disposes of all these questions by saying that the animal is driven to its actions by its "instinct," not worrying very much about the import he has to give to this concept. But, as it is said that curiosity may be regarded as the mother of all science, so interest in the behaviour of animals may certainly be regarded as the origin of the science of animal psychology. Not everybody, however, who meditates on the nature of animal activity, has by this fact alone the right to term himself an animal psychologist. There are, as we saw just now, three main ways of explaining this activity. The first is, to regard the animal as a living machine and to explain its actions as the outcome of purely mechanical causation. The prototype of this view in modern times is found in the conceptions of Descartes. Descartes distinguished two fundamentally different realities or "substances": a material one, the "Substantia extensa," characterized by its extensiveness, and a spiritual one, the "Substantia cogitans," characterized by a self-conscious thinking. As a result of this distinction he admitted an extreme dualism of body and mind. The body, as a part of the Substantia extensa, to him was nothing but an ingeniously constructed machine, whereas the mind, because of its faculty of thinking, was part of the Substantia cogitans. Now the animals, as they do not speak and never make gestures as we do when we wish to express our thoughts, in his opinion do not think and therefore have no minds and are nothing but such machines-machines suited to the purpose for which they were constructed, but, in spite of their often seemingly intelligent behaviour, wholly corresponding to our watches and clocks, which are built up only of wheels

9 and springs and yet measure the time better than we do. It will be clear that in such conceptions there is no place for real animal psychology. There will probably not be many people who nowadays share Descartes' views, at least as far as the higher animals are concerned. But as regards the lower kinds of animal the case is somewhat different. It was only at the beginning of this century that Loeb, in his Theory of Tropisms, declared that the actions of animals were nothing but the direct effect of external forces, such as light, gravity, heat, and the like, on the bodies of these animals. These forces, being of a purely physico-chemical nature, worked in a way quite independent of the will of the animal, and were wholly comparable in their workings with that of a magnet on iron filings, or that of gravity on the movement of celestial bodies. Although Loeb's ideas have lost their influence through severe criticism from different sides, I am not sure that their after-effects do not still confuse the minds of some students. At any rate Loeb does not bring us nearer to animal psychology than Descartes did. The case is somewhat different when we come to a second way of explaining animal activity, to wit, the physiological one. The man who adopts this physiological way of explanation, that is, the man who tries to analyse the actions of animals into physiological stimulation of sense-organs and nerves, the contractions of muscles and secretion of glands, may have two reasons for doing so. First, he may be a professional physiologist, who need not concern himself about any other aspects of animal activity. In this he is certainly in his right, and we may respect his attitude as long as he is consistent and does not try to make us believe that what he gives us is a kind of psychology. Physiology and psychology, it may be said here by the way, are two wholly different branches of science, however many points of contact they may have. The former occupies itself with the material phenomena of sense stimulations, nerve impulses, secretions and kindred phenomena, the latter with psychical phenomena, such as sensations and perceptions, feelings and emotions, desires and memories, and the like. A confusion of the aims of two ANIMAL

PSYCHOLOGY

IO

ANIMAL

PSYCHOLOGY

so different branches of science has never led to a clearer understanding, and certainly this is not the case . with physiology and psychology. Besides, this physiological attitude in the explanation of animal activity is sometimes adopted not only by the professional physiologist but also by people who for some theoretical reason believe it to be the only possible or satisfactory one. We will come back later on to thia attitude. Anyhow, this physiological explanation does not bring us to any real psychology of the animals either. The third way of explaining animal behaviour, and the only one that leads to real psychology of animals is, as we have already indicated, that of explaining it in terms of psychic phenomena, of trying to find out the psychic phenomena that underlie their behaviour. What, then, are these psychic phenomena ? They are phenomena we experience in ourselves as immaterial and not occurring in space (as do the physiological processes in our body) phenomena we know more directly than any other phenomenon on earth, and know directly in ourselves alone. Our own visual or auditory sensations, our own perceptions of the world around us, our own feelings of joy or misery, our own desires and strivings, are phenomena we experience more directly than we do the trees in our garden or the flow of a river ; on the other hand we have to admit that the corresponding inner experiences of other men remain a closed book to us, as long as they do not reveal them to us by some way of communication. The most direct knowledge of our own, a fundamental inaccessibility of another man's psychic or subjective phenomena (as we may call them because of their being bound to a person or subject) are characteristic of these inner experiences. Only the man who tries to explain the actions of a fellow creature, be it a man or an animal, in terms of such subjective experience, may rightly call himself a psychologist. How far such an explanation of behaviour is possible, especially with the animals, is a question we shall soon have to consider. This statement that psychology is the science of subjective phenomena will perhaps not satisfy the man who regards psychology as the science of the soul. Modern psychology, however, is a branch of the natural sciences, and in natural

ANIMAL

PSYCHOLOGY

II

sciences there is no place for the concept of soul. Not that these sciences reject this concept, or doubt the reality of the soul, but they realize that it belongs to branches of knowledge, such as philosophy or metaphysics, that do not form part of natural science. The soul is an object of speculation, not of observation or experiment, and it is this difference in their method of approaching their object that separates natural sciences from the philosophic disciplines. Modern psychology, therefore, has rightly been called "a psychology without a soul." If animal psychology were to be the science of the animal soul, it would involve us in speculations about a survival of this soul after the death of the animal and the possibility of metempsychosis and so on. The scientific value of such a science would certainly be trifling. The object of animal psychology, then, is not the animal soul but the psychic phenomena in animals. Its ultimate aim is to arrive at so complete a knowledge of these phenomena that it becomes possible to understand the psychical structure, be it of an animal or of a group of them, i.e. the dependence of one such phenomenon on the other, the gearing of one into the other, the remodelling of one by the other, in a way similar to that in which the anatomist or the physiologist tries to acquire a knowledge of the anatomical or physiological structure of the animal he is studying. That animal psychology is yet far from having reached this ideal is a fact we need scarcely mention. But now the question arises whether an animal psychology conceived in this way is possible, or if it is an ideal that will never be realized, be it either because such phenomena do not occur in animals, or that they do occur but are not knowable to us. Many people, indeed, doubt the possibility of a science of psychic phenomena in animals. Human psychology, they say, is built up on two foundations : the introspection into one's own inner experiences and the communication in human language of that which another man experiences at a given moment or under given conditions. It will be clear that both methods fail us when we try to study the animal mind. Introspection is of course directly precluded. And as

I2

ANIMAL

PSYCHOLOGY

to communication of inner experiences, it is very doubtful if an animal could ever be really explicitly conscious of what it experiences at a given moment, that it could realize for instance: "Now I am afraid, now I desire to kill that prey," which, of course, is quite another thing from the being afraid or the feeling of the desire itself. And even if, perhaps, some higher animal should arrive at a more or less vague realization of its inner feelings or desires, it is doubtful if it will ever feel any need to communicate them to other creatures. The animal, as a rule, is an egocentric being. If in fear it utters a particular cry, it may be that other members of its species, or perhaps even members of other species, will know this cry and may be affected by it in such a way that the feeling of fear of the one individual spreads through the whole troop. Yet there is no reason to admit that the first animal uttered its cry in order to communicate its feelings to the others. The uttering of the cry was nothing more than the expression, or the effect, of the feeling itself. Perhaps where there exists a strong social bond between two animals of the same species, as between child and mother, or even perhaps in rare cases where there is a social bond between an animal and man, as in the attachment of the dog to his master, there may be something like a need for communication of inner experience. But even then it might be asked if such inner feelings are clearly and explicitly experienced by the animal, and if this seeming need for communication is anything more than a clinging to a fellow creature at a moment of strong inner feeling or desire. Anyhow, these cases are so exceptional, and the communication itself is so vague, that they cannot suffice as bases on which to build a science. For human psychology, to tell the truth, mere introspection into one's own inner experiences and the communication of such experiences by other people, are not sufficient as a basis for this science either. This is particularly the case when human psychology does not deal merely with the psychic life of the normal adult civilized man, as was too often the case formerly, but also with that of the young child, the primitive man, and the mentally diseased. Then the same difficulties arise as with the animals, though in a

ANIMAL

PSYCHOLOGY

lesser degree. These persons, also, are not able scientifically to gauge their own inner experiences, nor are they able, or even willing, to communicate them clearly to others. Yet these branches of psychology flourish, and no one would think of denying their right of existence. What, then, is the way these branches of human psychology overcome the difficulties arising from this lack or defectiveness of active assistance from the side of the objects of their study ? They do it by applying another method in which these objects play a more passive part, namely that of observation and interpretation of their behaviour. That is, we watch their actions while they believe themselves unobserved, we ask them questions of different kinds and note their answers, whether clear or vague, we bring them into unusual situations .and watch how they help themselves out of them, etc. We observe all their expressions of emotion and feeling, be they spontaneous expressions or reactions to the situation we brought them into. And from the store of observations we collect in this way we draw conclusions as to their inner experience, their feeling, their thinking, their desiring. Let us not overlook the fact that it is a great and relatively uncertain step we take when from objective observations of the behaviour of a man we draw conclusions as to the inner experiences that drive and guide him to his acts. For as we have seen, these inner experiences are known to us as existing in ourselves alone. That our fellow man has experiences of the same nature as we have, that he has kindred feelings and desires and sensations, is only a more or less probable supposition. A psychological solipsism, the attribution of such experiences to ourselves alone, is a logically defensible position, though doubtless nobody with a sane mind will adopt it. We are all sure that our fellow-man sees the light of the stars and feels the warmth of the sun just as we do, that our child is afraid in the dark or enjoys his play, that our adversary nurses a feeling of hatred or dislike to us and desires to thwart our plans. Nay, more, we know this to be sure knowledge. Why are we so sure about it ; how do we know it ? It has often been asserted that we know these inner experiences of our fellow men by way of analogy : if his movements

ANIMAL

PSYCHOLOGY

or gestures were the same as we make in particular circumstances, we might therefore conclude that he is stirred by similar feelings as we experience when making the same gestures. This assumption, however, only partly holds true. In many cases we know what stirs a man, even if he does not make any movement or strike any attitude ; in many cases also we feel that a man is feigning and with a kind smile covers a feeling of hatred towards us. The young baby, on the other hand, when it sees its mother smiling kindly at it, knows the gentle feeling of the mother even if it has never had the opportunity to see itself laughing in a mirror, and responds to her gentle feeling by smiling back at her. The truth is that this understanding of other men occurs in quite a different way, that is, by a sympathetic intuition, the nature and the working of which it is not our task to discuss here. We may, perhaps, provisionally define this intuition as the faculty for directly grasping the essence of a reality, without preceding intellectual analysis. That this way of understanding may later on be supplemented and facilitated by a knowledge of the meaning of certain facial and other expressions in our fellow men that we have learned to distinguish by the experience of our own life, does not detract at all from the fact that originally our knowledge of other people's inner experiences is acquired in a different way. If, now, we pass to the animals and the knowledge of their inner experiences, we are up against a similar, even though a greater, difficulty to that which we encounter when we study the psychology of the young child or primitive man. That the difficulty is greater nobody will deny, but fundamentally it is the same. It is greater because the mental difference between ourselves and the animals is far greater than that between ourselves and the child. But yet in principle it is the same : we make the great leap when we deduce from our inner experiences those of other people, and once granted the existence of psychic phenomena in other men and the possibility of our knowing them, the leap from them to the animals is not so formidable, especially if we believe that animal and man are cognate beings, originating from each other by way of evolution and forming a part of the same

ANIMAL

PSYCHOLOGY

stream of life. When we admit psychic phenomena in other men, and believe we are more or less able to understand them, there is in principle no reason to deny them to animals or to think that we shall never be able to know them here also. Since the difficulty in knowing the subjective experiences is the same in our fellow men and the animals, the method of overcoming this difficulty must also be similar in both cases. With the animals, therefore, our knowledge of their psychic life is also primarily based on sympathetic intuition. Let no one accuse us of unscientific phantasy or mysticism. Every one of us possesses this faculty and makes use of it in his everyday life. The man who whistles in vain for his dog and says: "He hears me very well, but he won't come, and tries to keep away from me, because he is afraid I will beat him," by way of sympathetic intuition has entered into the perceptions as well as into the feelings and the striving of his dog. And the man who warns a stranger to beware of a bull because he is vicious, shows the same kind of knowledge of the inner experiences· of the bull as the man who makes a similar remark about a tribe of head-hunters in the jungle. Why, then, should such a knowledge suddenly cease, or fade away, when we pass from the field or the street into the lecture room or the laboratory ? There is, however, a difference between an acknowledgment in principle and an acknowledgment in fact. The possibility that animals are mere automata without any inner life, as Descartes believed, or rather that the lower orders are such, as was contended by Loeb, cannot be ruled out by logical arguments alone, and intuition may deceive us. Empiricism must bring the solution, must support or invalidate our views. What, then, have the empirical facts to tell us about the reality of psychic phenomena in animals, quite apart from the question of their finer nature ? We have seen already that the method of studying the inner experiences of animals is to watch their behaviour. We are using here the word "behaviour" in a somewhat restricted sense, given to it by McDougall, who connotes by this word all human or animal activity in which mental processes find expression. Hence the question now is whether the animals,

r6

ANIMAL

PSYCHOLOGY

it may be only some of them, show behaviour in this sense of the word. How do we recognize behaviour, how do we know that the activity of animals is not merely the result of mechanical or physiological causation ? On this point William James has given us a useful hint. "The pursuance of future ends," he writes in the introduction of his Principles of Psychology "and the choice of means for their attainment, are the mark and criterion of the presence of mentality in a phenomenon." Where we find these facts we may conclude that inner experiences underlie the animals' activity, may infer the existence of real behaviour, therefore, in the McDougallian sense of the word. These remarks of James, however, are rather vague, and it will not always be easy to apply them to observed animal activity. McDougall, therefore, has rendered us a great service in further elaborating this idea of James and in analysing his criterion into seven more profound "marks of behaviour." By reason of their supreme importance to us it seems worth while shortly to record them here. The first of these marks that distinguish real behaviour, guided by inner experience, from the mechanical movements as we find in lifeless objects, is that of its spontaneity, i.e. the independence of the activity from external causation. The beginning of real behaviour is not determined by external forces working on the subject but by an inner state or inner experience in the subject itself. A stone remains lying still till some outer force sets it in motion ; when a man is sitting on a hill looking at the view before him, it is not external forces that drive him to move away but the consideration that it is time to go home or the fear of catching cold, or something similar. And when it is a threatening thunderstorm that impels him to move off and seek shelter he is not set in motion by the force of the wind or the darkness of the sky, but these perceptions inspire in him fear of getting wet, or confidence that the shower will pass over, and these inner feelings decide whether he goes or stays, and in the former case at what moment. Certainly, where there is such spontaneity in a phenomenon there is mentality in it.

ANIMAL

PSYCHOLOGY

The second mark of behaviour is that of the persistence of the action, independent of the continuation of what induced it. If the electric current that drives the train is interrupted, the train comes to a standstill. When the fear of the thunderstorm that induced the man to move has disappeared, he may continue his way home. And, then, there is a great difference between the motion of the stone rolling down hill when set in motion and that of the man running down hill in search of shelter. The former rolls down in an ever-increasing and computable velocity along a track that also may be indicated beforehand. The man may go more quickly or more slowly, as he likes, may make a detour around a block or rock or a pool of water, may choose this tree for a shelter or that, and may perhaps afterwards change his place for a better one. His action cannot be determined beforehand as was the movement of the stone. This brings us to a third mark of behaviour: that of the possible variability of its movements, that of the relative freedom of its action. A fourth mark of behaviour is its termination as soon as the goal is reached. When the man has found a hut to take shelter in he does not go farther, even if the path goes on and he himself is not tired. The stone, on the other hand, rolls down hill till the valley is reached, or some obstacle prevents it going further. And a fifth mark is that of the preparation for, or anticipation of, a future activity. Before the man reaches his house he will take the key out of his pocket to open the door. That the rolling stone does not show any such preparations need not be emphasized. We may summarize these five marks of behaviour as the mark of the independence of the action from outer forces, alike in its incipience, its progression and its completion. If any activity shows these five marks we do not doubt but that inner experience leads it, be it in a man or another creature. McDougall has given two other marks of real behaviour which, however, are not so generally applicable to all cases where activity is observed. The first of these is that of its increase in effectiveness if the activity be repeated a second or third time under similar circumstances. If brought back to its former place, the stone does not on a second occasion roll 2

18

ANIMAL

PSYCHOLOGY

down more quickly or better than before, but if the man a second time has to take refuge at the same point, led by the experience of the former occasion he will sooner and more easily find a good shelter than the first time. This profiting by former experience cannot of course always be demonstrated and certainly not in the case where an animal or man executes an act for the first time in his life. Therefore this mark is not of such primary importance as the first five are, although the profiting by former experience itself is prominent in the behaviour of animals and is, as we shall see in a later chapter, the basis of what may be regarded as intelligent actions. The seventh mark McDougall gave for the recognition of real behaviour is that of the totality of the action. Behaviour is a total reaction in which the whole individual is involved. This mark does not so much distinguish the behaviour of the man from the movement of the rolling stone, but does distinguish it from the working of a reflex. A few words on these reflexes, therefore, may not be superfluous, the more so as some confusion prevails on this point. What, then, is a reflex ? A reflex is a typical physiological phenomenon, wholly explicable by the laws of mechanical causation. When a physical stimulus strikes a sense-organ, or nerve-ends, this stimulus is propagated along a centripetal nerve-fibre to a centre, in which it is transmitted to a centrifugal fibre along which it is conducted to an organ, which hereby is roused to a characteristic activity, be it the contraction of a muscle or the secretion of a gland. Necessary for the functioning of a reflex, therefore, is a morphological structure, the so-called reflex-arc, consisting of at least five parts : the receptor in which the stimulus is received, the afferent nerve along which it is conducted to the centre, the centre itself as the place of transmission, the efferent nerve along which the stimulus is then carried to the effector, and the effector in which the stimulus provokes the movement. In many cases the reflexes are not so simple as described here : one reflex may induce a second and this a third, so that chains of reflexes may arise from the original stimulus ; also higher centres may interfere and influence the effect of

ANIMAL

PSYCHOLOGY

the phenomenon. This, however, does not essentially change the nature of the reflex-process. Characteristic of a reflex are thus the morphological structure that underlies the propagation of the stimulus and the mechanical course of the whole process. 1 Now it must he said here that many persons often use the word reflex in a much wider sense so that it completely loses its original meaning and means nothing more than a reaction to a stimulus, however complicated these two may be, whatever be their nature, and however many factors may be involved in them. If an animal sees and recognizes an enemy and flies in fear, then this is called a "flight-reflex." When one of Pavlov's dogs tried to liberate himself from an apparatus in which he had been shut up for an experiment, Pavlov attributed this action to a "liberation-reflex" of the dog, and when another dog showed interest in some changes in the experiment-room this was attributed to a "what-is-that-reflex"! It will be clear that such a watering down and so slovenly a use of scientific terms cannot be sufficiently condemned. The fault in doing so is that first one emphasizes a special characteristic of a term in the original restricted sense (e.g. its mechanical causation), then applies the term to a number of phenomena that were not originally comprised in the definition, and then endeavours to suggest that "therefore" all these phenomena show the characteristics that belong solely to the phenomena that answer to the original definition. To avoid such errors it is necessary to keep to the original meaning of the word reflex, and to speak of a reflex only when a real reflex-arc is, or at least can be theoretically, shown. If, then, we ask ourselves what is the principal difference between a reflex and a real action, we feel that this difference lies in the nature of the experience that attends both. When 1 Has the reflex reality ? Some physiologists (Bethe, Buytendyk) doubt it. In their opinion a real isolated reflex does not occur. The reflex, then, is only an abstraction made by the human mind, the reflex-arc only a hypothetical construction. The splitting up of a function into reflexes is artificial, as is the division of a body into organs, and of its activity into functions. In reality, they all form one cohering unity. It will be clear, however, that for us this question is of but secondary importance ; for the interpretation of animal activity the difference between a real or ideal reflex and an action remains the same.

20

ANIMAL

PSYCHOLOGY

executing a reflex movement we feel ourselves as passive ; when carrying out an action we feel ourselves as active. When by a blow on or below the knee-cap our shin and foot are suddenly thrown up in the well-known knee-jerk, we experience this movement as something that is done to us, not as something that we do ourselves. We have an entirely different experience if when sitting on a bench in the park we make a similar movement to chase a dog away. And when we cough because something tickles our throat we have a different experience than when we do it to draw the attention of a person who is walking before us in the street. Many reflexes, like the pupillary reflex, or the secretion of the g~stric glands, even occur below the threshold of our conscwusness. Now with animals this distinction mostly fails us. We do not as a rule know the inner experiences of a certain animal at a particular moment sufficiently to know for certain whether some movement it produces is undergone passively by it or is the result of an active striving. In such cases McDougall's seventh mark may be a great help. The reflex, then, is a partial reaction, a reaction of a limited part of the body to a simple stimulus ; in an act, in real behaviour, the whole animal is involved. Real behaviour, therefore, can be found only there where the unity of the organism has not been broken ; reflexes may be shown even by isolated parts of the body. If the head of an insect has been severed from the body, biting reflexes may be provoked, but only the intact animal really bites. A decapitated grasshopper may show the jumping reflex, but it does not spread out its wings in doing so, as does the normal animal when taking a flying leap. The decapitated copperhead-snake may strike when its tail is pinched and may even strike in the direction of the pinch, but according to Huxley there is a great difference between the blind and automatic striking of this part of the animal and the action of attacking, which is guided by perceptions and may be inhibited by fear or other vital interests of the animal as a whole. Where, therefore, we see an animal acting as an organic whole, we may regard this as a proof of real behaviour, and

ANIMAL

PSYCHOLOGY

21

need not consider this activity as the outcome of one or more reflexes. It will be clear that this knowledge gives us a weapon in our hands against the physiologist if he tries to divert us from the interpretation of the activity of some animal by proclaiming that this is nothing but the result of the working of some reflexes. But let us return to the seven marks of McDougall. Wherever we find in the activity of an animal these marks of spontaneity, of variation, of termination, of persistence, of preparation, of improvement, where we see the animal acting as a whole, we may be sure this activity is not governed by external mechanical causation, but is evoked by inner experiences, being an expression of perceptions and feelings, of desires and drives. Where we find them, we find real behaviour, and at the same time we find in the animal a serviceable object for our study of animal psychology. Let us now test some animals and see if they can be used to this end. For the higher animals this examination is not difficult, nor the issue of it uncertain. Let us take the case of a dog lying quietly in the room while his master is reading. Suddenly, and without any perceptible reason, the dog rises and goes to his master. His course through the room is unpredictable : he may go round the right side of the table or round the left side, he may go slowly, or quickly, and certainly his going will differ one evening from that of the day before. When he reaches his master he sits down and looks at him in expectation. If the master gives some sign, if, for instance, he rises from his chair, the dog runs to the door and stands there waiting till the master opens the door to let him out. If this happens several times the dog will show an improvement in his behaviour, in so far as he does not wait till his master has risen, but runs to the door as soon as his master looks inquiringly at him, as he has learnt by experience that this look is a preliminary to his master's rising from his chair. Doubtless the behaviour of the dog is accompanied and guided by inner experiences, as weariness, impatience, desire, expectance, joy, in some cases disappointment and sadness, which, in the case of the dog, may easily be read from his

22

ANIMAL

PSYCHOLOGY

attitude and sounds, the movements of his body and tail, the look in his eyes, etc. Instead of a dog we might have chosen a monkey, a bird, a snake, a fish. Everywhere in these groups we should have found unmistakable proofs of behaviour, guided by inner experiences. It is not necessary to quote examples. One has merely to watch the behaviour of some voracious fish, say a pike or a perch in a stream or aquarium, to find here the marks of spontaneity, of variability, of persistence and so on. That fishes may learn, that is, may profit by good or bad experience, is a fact every angler knows and that is surely known also to the student of sense physiology, who realizes that the experiments on colour vision, form discrimination, hearing and so on in fishes, are all founded on the method of training, i.e. of the modification of innate reactions in the light of past experience. Let us now pass to another class of animals, the insects : ants, bees, wasps, and the like. At the very beginning of this chapter our curiosity was aroused by the spontaneity of the ant returning on her path without any perceptible external reason, and the same characteristics may be observed in the bee flying round a heather bush in search of food. Solitary wasps dig a hole in the ground, and there endeavour to collect a store of caterpillars or spiders as food for their future offspring. The Peckhams, who closely studied the behaviour of these animals, lay stress on the great variability in their behaviour in their striving to attain this end, and even go so far as to term this variability the "one prominent, unmistakable and ever-present fact" in this behaviour. This holds true for the conduct of different individuals of the same species as well as for the successive actions of one individual. The making of the hole itself may be regarded as a preparation for the future storing of provision. In her search for a caterpillar, or a spider, the hunting wasp persistently goes on exploring the territory round her nest till a prey has been discovered ; then she stops flying about and proceeds with a different action. That insects may learn by experience to change or improve upon innate actions is again shown by the ease with which bees can be trained to react

ANIMAL

PSYCHOLOGY

23

to special colours or forms, if these prove to indicate a source of food, and from the fact that they easily and readily come back to a place where once food has been found, be it a field of flowering heather in the case of the bee, or, in the case of the wasp, the table of a family taking tea in the open air. · It will be clear that the same must be true of the larvae of insects. From a psychological point of view there is of course no reason to make a distinction between the larva and the perfect insect. The insect is a living, perceiving and striving being before as well as after metamorphosis, however great the morphological differences between the two stages of life may be. Caterpillars, caddis flies, larvae of beetles and suchlike show the same characteristics in their behaviour that we found sufficient in other animals for the assumption of psychic phenomena. Let us show this by one example, which some time ago acquired some notoriety as an instance of the contrary. The ant-lion, larva of the Neuropteron Myrmeleo, passes its larval life at the bottom of a funnel-like pit, built by itself in the sand. If ants, or other small insects, happen to get into this pit they slide down, together with the sand of the walls of the pit, and fall into the open mandibles of the larva, which seizes them and pulls them under the sand, where it kills them and sucks them out. If by chance an ant succeeds in freeing itself from the ant-lion's mandibles and tries to climb the walls of the pit, the ant-lion throws up the sand that the ant in its flight sends down. In most cases this hits the fugitive ant and causes it to slip down again. Now, some .thirty years ago Doflein alleged that the whole behaviour of the ant-lion could be explained by assuming three simple mechanical reflexes : a digging-reflex, causing the larva to disappear under the sand, a throwing-reflex, which makes it throw up the sand in digging its pit and in pursuit of the flying ant, and a snapping-reflex, causing the mandibles to close whenever an insect is between them. He herewith declared the ant-lion to be a reflex-automaton, a machine without any trace of psychical life, thereby returning to the view which Descartes had defended long before him

ANIMAL

PSYCHOLOGY

regarding the whole animal kingdom. That the term reflex was not very felicitous will be clear : when the animal disappears under the ground there is more involved in this action than a local reaction of some part of the body, and clearly the animal as a whole is working to some end. But the essential question is : Is the entire activity of the animal mechanically determined by some few external stimuli ? Later observers of the ant-lion's behaviour came to quite a different result. They found spontaneity in the behaviour of the animal, especially when, without any outer stimulus, it suddenly began to dig a new canal in the sand, probably stirred by some feeling of uneasiness as a result of nonoptimal conditions of the surroundings. A similar spontaneity might sometimes be observed when, after the prey was taken away from him by the experimenter, the animal suddenly started to throw up sand, although in this case it had not been struck by sand thrown down by the escaping prey. It was even observed that the ant-lion in the pursuit of a fugitive ant went so far as to turn round in its pit, and even now and then to leave it. In the act of catching its prey a great variability could be observed. Thus a wildly struggling ant is sometimes caught and stunned by being beaten against the walls of the pit ; an ant that is not rightly caught is thrown up and caught better when it comes down again. In the position the ant-lion takes at the bottom of its pit, with the mandibles open to catch any falling victim, we may see the preparation for a coming act. Even an ant-lion showed itself capable of learning by experience : it could be trained to accept and eat dead flies instead of living ants. And when the author put an ant-lion and an ant together in a glass tube and let them fight together there, he observed in the larva as well as in the ant a purposive activity with sudden outbursts of spontaneity, a persistence in the action when stimulation from the side of the enemy was wanting, and an effective variation in the action itself, such as a machine or an automaton would never show. It will be clear that Doflein's judgment of the character of the ant-lion was only the result of a too superficial and biassed observation. The source of his error may be partially

ANIMAL

PSYCHOLOGY

25

found in the fact that the ant-lion is a highly specialized creature, living in simple natural surroundings with simple and almost unvarying characteristics. In such surroundings the actions of an animal are often inclined to become simplified to a few uniform and seemingly automatic movements, and thus, on a superficial observer, may make the impression of being no more than reflexes. Some simple experiments, however, suffice to demonstrate their real nature. If the antlion is taken from its pit and brought into another environment, for instance into a small box covered with a glass plate, its behaviour at once loses much of the stereotypy it evinced in the sand, and certainly no longer makes the impression of being of a reflex character. It is not necessary to pass through all classes of the animal kingdom in search of marks of real behaviour, of proofs of psychic phenomena. Let us rather pass on directly to the Protozoa. Is real behaviour, bearing the marks of psychic life, observable in these creatures ? The importance of this question will be clear. If real psychic life be found here, where we stand at the roots of animal life, it will certainly be difficult to deny it in other animals. If, then, we ask what observers of Protozoan life have found with regard to McDougall's criteria in the activity of these animals, thereby entitling us to assume psychic life and inner experiences in them, we must first of all rid ourselves of the prejudice that the small size of these animals would make it improbable, or even impossible, that their activity be governed by psychic experiences. It will be clear that there is no reason at all to believe that these immaterial phenomena are limited to a certain minimum size of the individual exhibiting them or that psychic life in this world correlates with the visual acuity of the human eye. Only a close observation of their activity and a critical interpretation of it can answer this question. Now, first of all, many students of Protozoa have been struck by the spontaneity in their behavour. Even so relatively early a writer as Verworn has pointed out the frequent occurrence in them of spontaneous movements, as contrasted with movements caused by external stimuli. As instances of these

z6

ANIMAL

PSYCHOLOGY

spontaneous movements he mentioned the swimming movements and especially the abrupt changes in the direction and velocity of the swimming in Bacteria and Flagellata, the seeking movements of the flagellum of Euglena and kindred forms, the movements of the pseudopodia of Amoeba, the sudden contractions of stalked Infusoria, like Stentor and Vorticella, for which he could never find any external stimulus, etc. Especially striking are the spontaneous movements of Infusoria which, like Halteria, possess spring-cilia and first swim forwards with the help of their adoral cilia, then suddenly, without any perceivable reason, turn back and swim on in another direction. In Amoeba, Jennings described a remarkable spontaneous action : if Amoeba has been suspended for some time in the water without contact with any fixed object, it stretches out long pseudopodia to all sides and searches the surroundings till one of the pseudopodia comes into contact with a solid object. Then the tip of that pseudopodium attaches itself to the object, and the animal, while drawing in the other pseudopodia, gets on to it. Here we have an instance, not only of spontaneity of behaviour on the part of the Protozoan, but also one of persistence in the action till the end, viz. contact with a fixed body, has been attained, and of the termination of the act of seeking as soon as this is accomplished. Certainly we cannot find the motive for the seeking action in any external stimulus; it is, on the contrary, the lack of all external stimulation that induces the activity of Amoeba. But here perhaps the physiologist who has so far silently, although more or less reluctantly, followed the line of our argumentation, will stop and interrupt us. As long as it was only a question of higher animals, say of dogs, or fishes, or even insects, he could acquiesce in the idea that inner experiences played a part in their behaviour. But with unicellular being, "a speck of protoplasm with a nucleus," as he has learned to call them, this goes too far for him. He is willing to grant that no external forces led Amoeba to its search. But does this prove that psychic processes underlie this activity ? Is it not possible to ascribe the searching for a fixed object by Amoeba to some physiological state in the animal,

ANIMAL

PSYCHOLOGY

say some tiredness, that impels it to seek a resting place ? Or, to put the question on a higher plane, is the mark of spontaneity, and are the marks of McDougall in general, really sufficient to prove psychic phenomena in animals ? Is there not another possibility for the explanation of their behaviour when external mechanical causation is ruled out, viz., the working of purely physiological inner processes ? In answer to this we may say we are willing to assume some form of "tiredness" as the origin of the action described of Amoeba. But, then, is such a "tiredness" of purely physiological, i.e. physico-chemical, nature ? Let us guard against the old confusion between physiology and psychology, between physiological and psychological phenomena I A purely physico-chemical state of tiredness will never lead to anything else than inactivity in an animal, and only if it be accompanied by some feeling of uneasiness, discomfort, or such like, however vague and diffuse these feelings may be, will this feeling of tiredness induce a spontaneous and, in its details unpredictable, behaviour like that of Amoeba, when seeking for some fixed point with outstretched pseudopodia. More generally speaking : a purely physiological, i.e. physico-chemical state or a physico-chemical phenomenon in a living being, not accompanied by any psychical concomitant, may have as a result some physico-chemical process in the body of the animal, and may even perhaps in Amoeba be the cause of an expulsion of protoplasma in the form of pseudopodia. It is unconceivable, however, that such a physico-chemical state, or an integration of physicochemical processes alone, may induce an animal to purposive directive striving, exhibiting characteristics belonging to an essentially different domain from that of physico-chemical reactions and physico-chemical causality. Wherever we find a spontaneous, persistent, variable, directive behaviour of an animal as a whole, be it in a dog or in an Amoeba, we come no further with an explanation by inner physiological processes than with an explanation by outer mechanical causation. It will further be clear that an appeal to the concept of reflex offers no help to the physiologist in this dilemma.

28

ANIMAL

PSYCHOLOGY

In the first place, as far as the Protozoa are concerned, it is questionable if real reflexes occur in them. Jennings, at least, doubts this. But even if we confine ourselves to the higher animals, where reflexes certainly are to be found, they cannot explain real behaviour. We have already seen that the reflex, if this concept is to have any meaning at all, comprises partial mechanical reactions to physico-chemical stimulations. Such reflexes, then, may be integrated to complex wholes, but do not thereby change their real nature. Reflexes, or integrations of reflexes, may show a passive mechanical purposefulness, as the purposefulness of a typewriter or an atomic bomb ; they never show the directiveness, the active purposiveness, the "pursuance of future ends" of behaviour. The purposive can never be explained by the mechanical, nor can real behaviour ever be wholly analysed into reflexes. But let us return to the Protozoa to see if they show more marks of real behaviour. The best occasion to observe such behaviour with all its characteristic signs is to be found when an animal shows the highest tension of its vitality, that is, when retreating from an advancing enemy, or when itself hunting a prey. This hunting by the Protozoa has several times been observed and reported on by various authors. Engelmann described such chasing by a Voticella, that followed a smaller one for some time but was unable to catch it, and finally lost it when the smaller one made a sudden turn. Binet described the hunting by Ciliates as a real chase, in which perception and localization of the prey and purposive activity, played a role. The most elaborate description of the chasing by Amoeba we find in Jennings. He describes the pursuit by an Amoeba of a round cyst of Euglena, and, more interesting yet, the chase by an Amoeba of a smaller one, that not merely rolled passively away as did the cyst, but actually tried to escape from the pursuer. It would take too long to quote here Jennings' description in full; for the details we must refer to Jennings' own account. It may suffice here to state that, according to this writer, Amoeba behaves in a way similar to that which would be adopted by some higher animal in the same case, the entire hunting being one coherent process in which the

ANIMAL

PSYCHOLOGY

29

whole animal is involved. A dissolution into a number of simple reactions to stimuli is, according to Jennings, difficult, or even impossible ; a variation of the means employed, a persistence till the end is reached and a termination of the action when this has taken place, are the most striking characteristics in this behaviour. The imbibing of food by Amoeba is also a more complex activity than was formerly believed. That it cannot simply be explained by changes in the tension of the surface layer alone, as has sometimes been asserted, that on the contrary it requires actual activity on the part of the Amoeba, has been shown by Mast and Root, and by Beers. More important, however, is the great variability in the behaviour while capturing the prey, as has been shown by Kepner and his fellow workers. An immobile prey is closely embraced by the protoplasma of the Amoeba ; to a mobile prey the flight is first blocked in a wider embrace, then the prey is embraced more closely in different ways depending on the behaviour of the prey itself. Amoeba may form one or more secondary pseudopods of different sizes, it may withdraw its pseudopods at unsuitable places while it may form new ones more suitable, etc. According to Kepner, all these reactions are entirely independent of the stimulation of Amoeba by its prey ; they may be stopped or reversed, should that be required by the aim Amoeba is striving for by ever-varying means. We therefore find in Protozoa the same marks of behaviour that we know in higher organisms: spontaneity, variation in the means employed, persistence of action and cessation of it when the end has been reached, independently of the momentary working of the stimulus. That Amoeba reacts as an organized whole has been especially stressed by Mast. And even some years ago it was shown by Bramstedt that Protozoa may learn to profit by experience. This was demonstrated in the following way. Bramstedt put some Paramaecia into a drop of water in which a temperature gradient of 15° was produced. The warm half of the drop was then illuminated, while the cold one was kept dark. Now, if the animals swimming in the drop happened to get into the warm half of it, they showed their well-known

30

ANIMAL

PSYCHOLOGY

evading reaction, and returned to the cold part. Then after 1-1! hours both parts of the drop were equally heated. Although there was now no external reason why the animals should avoid the lighter half of the drop, they yet remained in the dark part and exhibited evasive reactions when crossing the boundary between the dark and light, which reactions were not shown by animals that had not been trained before. They had therefore learned by experience to associate light with heat. This experience was only retained for some 15 minutes, after which the effects of training were lost, probably by the new experience that light now no more meant heat. With a creeping Infusorium, Bramstedt succeeded in a similar way in forming an association between light and a rough surface, between darkness and a smooth one. Bramstedt's results have later been confirmed by Soest, and extended by the latter to other Protozoa. From all these facts we may draw the conclusion that the behaviour of Protozoa shows the marks we had previously recognized as proofs of psychic phenomena underlying the activity of a living being. We may therefore credit the Protozoa with experiences such as sensations and feelings, desires and drives, of the same nature, though undoubtedly vaguer and more diffuse, than those we know in ourselves. It will be clear that this conclusion is highly important. For if we admit an inner psychic life in Protozoa, we cannot deny it to other animals. It is very improbable, to say the least of it, that psychic phenomena occurring in Protozoa are lac king in higher animals that have descended from similar b eings in the course of the evolution of animal life. The difference between a perceiving, feeling and striving being and a finely constructed machine is too fundamental to make such a supposition admissible. This of course does not imply that everywhere in the animal world psychic life can be demonstrated with such relative ease as in Protozoa (especially in Amoeba and similar active species). In sessile animals, like the Coelenterata, the Ascidia and the Bryozoa, it will be difficult to prove real behaviour. Their sessile life give them little opportunity to show a purposive striving. In them we observe little of any spontaneity of behaviour, of a variation

31 in the means, of a persistence and a preparation, and of a profiting by experience. In their way of living they show great resemblance to plants, regarding which Aristotle said that they were "animals fallen asleep." Yet most of them, during their larval life, pass through a stadium of free movement, and in this stage show the same kind of behaviour as do free-living plankton organisms. Larvae of Ascidia at the beginning of their life are negatively geotactic and strive to reach the surface of the water ; later they become positively geotactic and make for the bottom of the sea, where they attach themselves in preparation to their metamorphosis. It is of course incredible that all psychic life should wholly disappear with their transition to their final phase. And even such sessile animals from time to time show marks of real behaviour. Spontaneous activity may be observed in Hydra which, if it has been left undisturbed for a long time, may suddenly leave its place and move about until finally it fixes itself at another place, probably having been stirred by some feeling of uneasiness or discomfort. Loeb found that if the Actinia Cerianthus is put upside down into a test tube, the animal in directive action wriggles to get back into its normal position. In the origin of the symbiosis between an anemone and a crab the active part is not always played by the crab ; in many cases the initiative emanates from the Actinia that strives to get on the back of a crab and in this sometimes shows a distinct preparation for the act of climbing on to the crab's back even before a crab is anywhere in the neighbourhood. These few examples may suffice to show that in sessile animals psychic life is not always so dormant that it could not be recognized in their activity. This justifies the recognition of psychic phenomena in all animals. Of course such an admission does not say anything further about the nature of these phenomena ; in particular it says nothing about the degree of consciousness that accompanies and illuminates it. In ourselves we know all gradations of clearness of consciousness of our inner experiences, from those which occur wholly, or nearly, subconsciously to those that take place in the clearest light of our apperceptive attention. We will do well ANIMAL

PSYCHOLOGY

ANIMAL

PSYCHOLOGY

to admit that, especially with the lower animals, psychic phenomena will for the most part take place in the twilight of a half-conscious state, illuminated only from time to time by flashes of a clearer consciousness. We now know that we may admit psychic life in all animals, that the whole animal kingdom belongs to the field of study of the student of animal psychology. We know, too, that in principle these psychic phenomena in animals are knowable to us, in the same way as those of our fellow men. We might expect, therefore, that all students of animal behaviour would follow us and try to deduce from the observed behaviour the psychic phenomena of which this behaviour is the expression. But it is just here that we meet with unwillingness on the part of the objectivists. The objectivist is the man who (in most cases) does not deny psychic life in animals as Descartes did, but does not take pains to know it, pretending either that he has no interest in it, or that he believes the knowledge he can obtain of it to be too uncertain. He, therefore, voluntarily renounces a knowledge he might be able to obtain. He studies animal behaviour for its own sake, and is not willing to go farther than the objectively knowable facts, and therefore remains quite at the fringe of this behaviour, describing the acts of the animals, asking what stimuli evoke these acts, but omitting to consider what links an external stimulus to an external reaction. In Germany there first arose such an objectivistic movement when, at the end of the last century, three biologists, Beer, Bethe and von Uexkiill, in a comprehensible reaction to the rather uncritical animal psychology of those days, published a manifesto in which they tried to clear the sense physiology of men and animals from all expressions that were contaminated with a subjective meaning, since, in their opinion, subjective phenomena were only knowable in ourselves, whereas in our fellow men, not to speak of animals, they were knowable only by analogy. They therefore proposed to speak of "receptors" instead of sense organs, of "recep..: tions" instead of sensations, of "icono-receptions" instead of perceptions, and the like. Characteristic of their attitude was

ANIMAL

PSYCHOLOGY

33 von Uexkull's pronouncement: "We renounce deducing from observable action of the animals psychical phenomena in them." These phenomena, therefore, are not denied, but only ignored. From this standpoint von Uexkull later built up his "Umweltlehre," an interesting objective doctrine of animal behaviour, but one the further consideration of which would here be out of place. The influence of this German objectivistic trend has never been very great. Most of the adherents turned to pure physiology. More influential was an American tendency to objectivism, the "Behaviourism," first expounded by Watson in a manifesto of 1913. There are among these Behaviourists many different tendencies and gradations, from the mechanistic and somewhat simplistic behaviourism of Watson himself to the "purposive Behaviourism" of Tolman, who oscillates between the Behaviourism of Watson on the one end and the Purposivism of McDougall on the other, and the "hormic Behaviourism" of Russell, who, although openly recognizing the animal as a feeling, perceiving, and striving subject, yet focusses attention not so much on these subjective phenomena themselves, but on the biological significance of the behaviour and its relation to other bodily activities. But they all agree on the one fact of acknowledging as the object of science only what is objectively knowable. In his later books Watson goes farther, and not only ignores subjective phenomena but even more or less denies them. Such concepts as sensation, sentiment, drive, memory, and the like, according to him, are scientifically worthless, and must be left to philosophy. It cannot be denied that from a theoretical point of view such an objectivism is quite in its right. A man of science need not interest himself in every subject on earth and has the right to restrict the scope of his work : a student of history need not learn the languages of all the peoples whose past he studies, nor is it absolutely necessary for the student of comparative anatomy to study the distribution on the earth of the animals he uses in his comparisons. On the other hand it must be stated that in this case the objectivist remains much in default. If one is interested in describing and explaining 3

ANIMAL PSYCHOLOGY 34 the behaviour of animals, why stop at the description of objectively observable phenomena and not at least try to penetrate into the inner experiences of an animal that undoubtedly influence its behaviour ? Why speak only of a stimulus when we know that the animal reacts to what for it is more than a stimulus, namely : a perception ? Why ignore the feelings and strivings of the animal or at the best summarize them as "unknown internal factors"? In the case of man surely nobody can be content with such a defective explanation. If an observer were to explain the behaviour of a man who tries to reach the shore after having fallen into the water simply as a complex movement of arms and legs, released by the stimulus of the humidity of the water, we should call this a very unsatisfactory interpretation of the facts observed. We should feel that in this description highly important and fundamental elements had been overlooked, such as the image in the man's mind of the expanse of deep water he is floating in, his fear of being drowned, his looking out for a place of safety and his actual striving to reach the shore, and we should know that all these inner experiences have influenced and governed his actions much more than the mere stimulation by contact with the water would do. In the same way if an objectivist (Tinbergen) believes to have sufficiently analysed the behaviour of a herring-gull settling on her eggs by saying that the visual stimulus of the eggs releases the innate movement of sitting down, we know that he overlooks the fact that in reality the chain has more links : the gull perceives the eggs, recognizes them as objects to be hatched, experiences some tender feeling towards those round objects, desires and strives to sit on them and, as a result of all this, executes the act of settling. If, by some cerebral lesion or mental disease, the gull should be struck by "psychic blindness" and should no longer recognize those round white objects that she sees as eggs. she probably would not settle on them, although the visual stimulus would remain the same. The whole chain, therefore, is built up in this way : visual stimulus-perception-feeling-drive-action, from which the three middle, psychic, factors of course cannot be omitted if a complete analysis is desired.

ANIMAL

PSYCHOLOGY

35

For the rest the attitude of many of the objectivists is curiously inconsistent. They claim to restrict themselves to the objective facts alone, but as they often feel the impossibility of really describing in this way what happens, they often supplement their description by using terms which have a psychol9gical meaning, like "drive" or "mood," or the like, excusing themselves by saying that they use these terms "in an objective or physiological sense only." What a word like "mood," which describes a certain complex of feelings or emotional states, means in a physiological sense is not quite clear ; for the rest, the insertion of such subjective terms into an objective description is of course apt to cause confusion. This declaration of objectivism when using subjective terms is evidently done to quiet the objectivistic conscience of the writers concerned. Sometimes also they honestly own up to the insufficiency of their theoretical attitude when they say, like Nissen and Crawford in connection with the use of words like "unhappiness," "friendship," "sympathy" and such-like in their interesting description of the behaviour of chimpanzees : "We shall conserve time and space by employing anthropomorphic terms (without quotation marks) when these permit briefer or more adequate description than would available objective terminology." We may doubt if ever any really objective terminology will permit of an adequate description in this case. Among these objectivists we may again distinguish two main trends. Some are content with the tracing of the stimulation that provokes a reaction, and the description of this reaction to the stimulus. Others go farther, and regard as their ultimate aim the analysis of the behaviour of the animals into the muscle contractions that are involved in it, or even, like Fraenkel and Gunn, into fundamental terms of physics and chemistry. Practically speaking there is no great difference between the latter and the physiologists we mentioned before, and with them we may speak of a "flight into physiology" for fear of acknowledging psychical phenomena in animals. It is clear that where the behaviour of an animal is thus split up into these ultimate elements, the animal as a whole is disregarded. For this reason we would give the

ANIMAL

PSYCHOLOGY

preference to the former trend, if we had to choose between them. But it is not our intention to depreciate the objectivists. They have done, and are still doing, excellent work in .their careful descriptions of animal behaviour. Our task is to draw conclusions as to the inner factors in behaviour. Objectivistic study of the behaviour is the first step to be taken. But it is not necessary to limit ourselves to this first step. We may and must go farther. Let us end this chapter now by summarizing our conclusions as to these fundamental problems of animal psychology. The object of animal psychology, then, is not the animal soul, a conception we readily leave to other sciences to speculate on ; it is to be found, rather, in the subjective or psychical phenomena of animals. Such phenomena we know directly in ourselves as our inner experiences of perceiving and feeling, of desiring and striving, of remembering and understanding and the like. In other creatures, in our fellow-men as well as in the animals, we know them only in an indirect way, namely, either with the aid of that other creature himself, who communicates to us what he experiences at a given moment, or without his aid, by studying and interpreting his behaviour. In animals we are confined to the latter way of study. The interpretation of their behaviour then is mainly based on a sympathetic intuition anent the animal's inner experiences, a faculty every man possesses to a certain degree and of which he makes use in his everyday life. Errors, of course, may be made, as regards the feelings and drives of the animal as well as the feelings and desires of the man, but in principle those inner experiences of animals are certainly knowable to us. And in applying to the animals the seven marks McDougall has brought to the fore for the recognition of real behaviour, i.e. activity provoked by inner experiences, we found such behaviour everywhere in the animal world, from the higher orders down to the Protozoa, so that we are entitled to include all animals in our psychological studies. For these two reasons there is no ground for us to share the objectivist's scepticism : we may

ANIMAL

PSYCHOLOGY

37

commit ourselves to a genuine and frank animal psychology. It will now be our task in the following pages to summarize in a few chapters the results of many years of study of this animal psychology.

CHAPTER II

THE PROBLEM OF ANIMAL INSTINCT For, where there is sensation, there is also pleasure and pain and, where these, necessarily also desire. Aristotle, De Anima. (Translation J. A. Smith.)

EVERY science which is more than a simple agglomeration of assumed facts, but which tries to build up a logical construction of theories and hypotheses that bind such facts together, has generally one or more central problems to which all these theories and hypotheses are related. For animal psychology this central problem is that of animal instinct. What, then, is an instinct ? In the philosophy of the man in the street the view may often be heard that the animal is impelled to its acts by "instinct," man on the contrary by "intelligence." According to this way of thinking, instinct comprises all the psychical faculties of the animal ; it is, perhaps, even the only psychical quality with which the animal is credited. It is supposed to be some general, never-failing, mysterious, innate knowledge, which tells the animal in every circumstance of its life what it has to do and what it must avoid. Intelligence, on the other hand, a prerogative of man, is not such a "savoir faire inne" (Spaier), but a mental faculty which enables man consciously to choose the right means to attain some end, and to act after ample deliberation about the pros and cons of the performance of some act in a special circumstance. This view, however, is certainly no more than a very crude approximation to the truth. That in the behaviour of

ANIMAL

PSYCHOLOGY

animals instinct stands out more clearly than in that of man, that on the other hand the acts of man in many cases are executed after deliberation and are guided by the result of reflection, cannot be denied. But we know that in man also instinct, though it may be unnoticed by himself, often determines his actions. And that in the behaviour of animals intelligence may also play a part will be made clear in the course of this book. This old antithesis between instinct, as being alone responsible for the acts of animals, and intelligence, as being responsible for those of man, certainly cannot be maintained. This antithesis, it must be admitted, is a very old one. Already the ancient Greeks, especially the Stoic philosophers, assumed in the animals as the springs of their actions a faculty of mind they called the "Horme" (i.e. that which impels), and Scholasticism translated this word by "instinctus" from the Latin "instinguere," which also means instigating, or impelling. And although both philosophies were more or less subject to the same fault we criticized just now in the popular psychology of the man in the street, the derivation of both terms from verbs that mean something like "incitation" or "impelling" shows that they consciously gave to the word instinct a psychological meaning. We will hold to this psychological meaning of the word. It cannot be our intention to give a review of the great number of descriptions and definitions given in the course of time of this concept of instinct. Many of these satisfy us no longer, either because they give to the notion of instinct too wide a scope, whereby it comprises many phenomena that are better not included in it, or, on the other hand, they restrict it so much that even its psychological core sometimes gets lost. One of the best provisional definitions probably is that given by Romanes in his article on instinct in the ninth edition of the Encyclopaedia Britannica which runs as follows: "Instinct is a generic term comprising all those faculties of mind which lead to the (conscious) performance of actions that are adaptive in character but pursued without necessary knowledge of the relation between the means employed and the ends attained." We have put the word "conscious"

39 between brackets because, as we saw in the first chapter, the degree of consciousness with which the animal performs an action is something we do not know about. For the rest, the word is not of primary importance in this definition. Instinct, then, according to this definition is a psychological factor, comprising a number of "faculties of mind," i.e. of psychical phenomena, of inner experiences. What is the nature of these psychical phenomena, what is the relation between them, are questions we shall have to consider in this chapter. Instinct, or rather, instincts (for we have learned to give up here the idea of one simple general psychological faculty, and have come to distinguish between several instincts in the plural)-instincts, then, lead the animal to perform adaptive actions. Let us call these actions "instinctive actions," and distinguish them clearly from the instinct itself that calls them up, a distinction that is not always kept in view. These actions, then, are performed without the animal understanding their meaning and purpose : it does not know their "why" and "for what." The animal, acting on the urge of an instinct, mostly acts as if driven by a blind impulse. We may add yet two other characteristics of instincts, not mentioned by Romanes in his definition : they are innate and not acquired during the life of the animal, and they are characteristic for the whole species to which the animal belongs, not only for the individual animal. Let us illustrate all this by some examples from animal life. ANIMAL

PSYCHOLOGY

Instincts and instinctive actions are characteristic of the species to which the animal belongs. All members of a species, or of a wider systematic group of animals, in similar circumstances behave in a similar instinctive way. At the end of the summer all our storks migrate to the South, and it does not depend on the individual decision of any one among them if he will depart or stay in the country. On the other hand, no house-sparrow undertakes such a journey but all remain in our country in winter. All spiders of the family of Agelenidae spin an irregular web of fine texture, and no one of them ever comes so far as to spin a regular orb-web which we all know as a characteristic of the family of Epeiridae. Among

ANIMAL

PSYCHOLOGY

the birds the family of the Columbidae, according to Lorenz, is distinguished from other families of birds by no common morphological characteristic, but they distinguish themselves from other birds by their drinking instinct, the act of which consists in a rhythmical sucking up of the water with submerged beak, while the other birds drink by putting their bill into the water and then raising their head. Spaier even goes so far as to distinguish animals in general from plants as organisms subject to the working of instincts. Instincts and instinctive actions in this way become characteristics of a species, just like the morphological ones on which as a rule the distinction between the species of animals is founded. The same holds true for the results of instinctive actions. By the nest of a bird we may recognize the species of the animal that built it, just as we may recognize this by the colour of its feathers. The insects stored in the nest of a digger-wasp tell us the species of the wasp that buried them, just like the colour of the abdomen or the form of the antennae of the wasps themselves. In some cases the instinctive actions of animals may give even better and clearer marks of distinction than do morphological ones. We have seen this already for the family of pigeons. According to Wachs the courtship of sea-birds shows characteristic differences that are much more important than the morphological differences generally used to characterize the different species. The common and the Arctic tern (Sterna hirundo and paradisea) show but slight external differences. Van Oordt, however, discovered a typical difference in their brooding instinct: the former begins to sit directly after laying her first egg, while the latter does not do so before the whole clutch is complete. This has the result that the young chicken of the latter species are all of the same size, while in the former there are differences in size. It was only at the beginning of this century that De Winton's yellow-necked mouse (Apo...; demus flavicollis Wintoni) was distinguished from the common field mouse (Apodemus sylvaticus sylvaticus) as a separate variety. As Frances Pitt remarks, a clear difference between the two, more marked than the morphological ones, is found in the fact that the latter never, or hardly ever, invades

ANIMAL

PSYCHOLOGY

41

dwellings while the former is very fond of doing so. Thus many examples might be quoted in which instincts yield us a better criterion of a species than morphological characteristics. More important, however, for an understanding of instinct is the general fact that instincts are characteristic of the species, not of the individual. This resemblance between the morphological or physiological structure and the instinctive outfit of an animal, the fact that both are specific and not individual in character, brings us to the question whether there is a mutual connection between this morphological structure and these instincts, and if so, which of them then is primary and which is dependent upon the other. That in many cases such a connection exists, nobody will deny. Without spinnerets no spider would be able to spin a web and without well-developed wings no bird could migrate to the South. The influence of hormones, especially of the sex hormones, on the instinctive behaviour of adult animals is beyond all doubt. But does this mean that instincts are the outcome of the animal's morphological structure or physiological functions, as has been asserted in times when psychical phenomena were not held in high esteem? Is it true, as has been asserted by Miiller-Erlangen, that an instinct is nothing more than the need to use an organ that the animal possesses ? Already the philosopher von Hartmann has combatted this view. All spiders possess the same spinning-glands ; yet some of them make a regular orb-web and others an irregular web, while yet others live in holes they only line with their spinnings. Further, this admission at best would explain why an animal uses an organ, but not the way in which it uses it ; that for instance the spider empties her spinning-glands, not that she weaves a web with their secretion. As Morgan remarked, nobody from the bodily structure of the eel would deduce the remarkable migration of these animals, nor from the anatomy of the ant-lion conclude that it makes a pit and there lies in wait for ants to fall down. On the other hand, similar instincts are shown by animals with quite a different bodily structure. Ants and termites, it is known, although belonging to quite different orders of insects, show a remarkable correspondence in their social

ANIMAL

PSYCHOLOGY

instincts. Yet more striking perhaps is the resemblance, to which Wheeler has drawn our attention, between the instincts of the ant-lion and the worm-lion. The latter, the larva of the Dipteran Vermileo, is an animal which, with an entirely different bodily structure from that of the ant-lion, also makes a pit in the sand by throwing the sand up, lies there in wait for small insects which it pulls under the sand if they fall down into the pit, and also throws sand up if the prey tries to escape. That there exists a connection between instinct and bodily structure nobody will deny ; that on the other hand the instincts are determined by the bodily structure cannot be maintained. We must consider them both as two different aspects of one phenomenon that lies at the background of both, and is realized in the structure as well as in the instinct. This basic phenomenon is empirically unknowable to us. Speculations as to its nature would lead us into the realm of metaphysics and cannot therefore find a place here. Another characteristic of instinct and instinctive actions is the fact that they are innate and the actions not acquired during lifetime, be it by imitation or trial or by instruction from older members of the species. It has often been proved that animals reared in isolation from their congeners showed normal instincts and normal instinctive activity when the time for these had arrived. Young birds of prey, reared in isolation and always fed by hand, showed on the first occasion the same way of attacking and killing a prey as their congeners grown up in normal circumstances. Birds, bred in nests of other species, built a nest typical of their own species, so showing that they were not influenced by any remembrance of the nest in which they had spent the first days of their life. A young moorhen, reared in isolation by Morgan, was swimming in a pool when it was frightened by a young yelping dog. It then at once dived, and swam away under water as if it had learnt this way of flight from older moorhens. Typical .instinctive actions, further, do not change in principle with repetition : the first web of the young spider is built after the same pattern as that of the old one, although of smaller dimensions. And all those instinctive actions of

ANIMAL

PSYCHOLOGY

43 insects and other animals often so complicated and performed but once in their life, such as those connected with the propagation of the species, are carried out in their characteristic way by all members of the species without any tuition from their parents, which, as a rule, are already dead when the young ones arrive at the age to perform these acts. Characteristic of the instincts is therefore that they are innate and not acquired. This, however, does not mean that all instincts are manifested directly after the birth of the animal. This will at once be clear of instincts that require a certain bodily development of the animal before coming into action. Birds that are not yet able to fly cannot fly away if danger is menacing, but will press themselves down and keep as motionless as possible, while their mother flies away. Instinctive actions connected with the propagation of the species will not show themselves before the animal has reached sexual maturity. Instinctive actions of animals that have to pass through a metamorphosis cannot be performed by the larva. All such instincts that show themselves only after a certain development of the animal were called by Morgan "deferred instincts." The most interesting cases of these deferred instincts are those in which the delay of execution is not caused by a bodily development but by psychical maturation. A good example of this is provided by the division of labour in the bee-hive. It was formerly believed that the different activities of the bees in the hive, such as the care for the brood, the building of the comb, the collecting of food and so on, were carried out by different groups of bees, so that the whole population was composed of different castes, each with its own task. Careful observations of Roesch, however, have shown that this is not the case, but that, on the contrary, every bee in the hive has during its lifetime to pass through all the different jobs. First, in the first ten days of their life, the young bees have the task to attend to the brood, to clean the cells in which the queen will deposit her eggs, to warm the developing brood by sitting on it, to feed the older larvre with honey and pollen from the store-cells, and, when

44

ANIMAL

PSYCHOLOGY

eventually their own food-glands have developed, to feed the young larvre with the secretion of those glands. Then, in the second period, running from their tenth to twentieth day, they have other occupations, such as to take the food from the returning bees and bring it to the store-cells, to clean the hive and, further, to build the comb when their wax-glands have fully developed, which happens between the tenth and the seventeenth day. Finally, in the third period, which runs from the twentieth day up to their death, which in summer occurs between the thirtieth and thirty-fifth day, the bee works as a field-bee, collecting food and bringing it to the hive. It is clear that we have here a number of examples of deferred instincts : although they are all innate, many of the instinctive actions of the bee are executed only at a certain age. It is also clear that in some of these cases there is a clear connection between the execution of the instinctive actions and bodily development. The feeding of the young larvre with the bee's own feeding-juice has to wait for the full functioning of the feeding-glands ; the building of the comb for that of the wax-glands. Yet it would be wrong to believe that the appearance of all these instincts is determined by bodily development alone. This was shown in a very interesting experiment by Roesch, in which he succeeded in dividing a population of bees into two groups of different ages : one younger and one older than eighteen days old. The result of this division was that in the younger population the menace of famine arose, as the foraging bees were wanting. Then, after some days, young bees of 7-15 days old flew out in search of food. At first they were still in possession of food-glands, and so flew out "physiologically precociously." In adaptation to the abnormal circumstances the population was thus split up into two groups of the same age with different occupations. On the other hand, in the group of older bees there arose a menace to the care of the brood, till some older bees took this task upon themselves. Also bees older than the .normal builders began to build on the comb, and to that end even developed new wax-glands. Here, therefore, we have a case

ANIMAL

PSYCHOLOGY

45

where the instinctive activity is certainly not dependent upon bodily development, but, on the contrary, bodily development is adapted to the psychological wants of the individual. Roesch further made the important observation that bees which had been isolated during the first ten days of their life, when brought back to the hive first went through the activities they had missed during their absence and so had to work off the arrears. This again shows that the instincts are relatively independent of bodily development and that the influence of mental maturity is an overriding one. Now it is a curious fact that instincts which for their full realization require a special kind of bodily or psychical development, may, before this maturity has been reached, sometimes show themselves in a more or less schematic form and towards an object other than the adequate one. Young animals of prey, still nourished by their mother, practise their hunting instinct on a living or lifeless prey ; young animals not yet sexually mature treat their brothers and sisters in a schematical way as sex partners. Herring-gulls sometimes show a sham nest-building, which precedes the real building of the nest ; the birds then pick up all sorts of material for nest-building in their beak, only to lay it down at another place without further attention to it. Howard observed with the whitethroat that the female one day takes some blades of grass in her beak but directly lets them fall again. The nex~ day she carries the blades somewhat longer with her ; the following day she deposits a number of them in the fork of a branch ; some days later again she builds the beginning of a nest, till, finally, after having built such unfinished nests at different places, the final nest is built. Here we see before our eyes the maturation of an instinct and its development from a first tentative utterance up to its full expression. Much of what has been described as play in animals finds its explanation as a premature execution of instinctive activity. Groos, in his classic book on animal play, has stressed the value of such premature instinctive activities as a practice for later serious occupations. It is by no means certain, however, that the animal really requires such practice ; certainly the normal maturation of its instincts is by

ANIMAL

PISYCHOLOGY

itself sufficient to enable the animal to execute the moveM ments concerned with the necessary skill. Anyhow, these premature instinctive activities again show the relative independence of the instincts from bodily development. We have already seen that Romanes in his definition of instinct pointed to two characteristics of instinctive actions to wit, that they are adaptive in character, i.e. generally purposeful, and are executed without the animal knowing the relation between the means employed and the end attained ; in other words that they are executed under the urge of a blind drive. This purposiveness presupposes a definite end towards which instinctive actions are directed. This, now, is an essential characteristic of instincts ; their activity is all directed towards one special vitally important end, be it the propagation of the species or the preservation of life, or something else. By this characteristic instinctive activity is distinguished from other forms of activity that, although innate, are not directed towards such an end. Activities like swimming, flying, creeping, diving, burrowing, and the like, are innate and in the· special way in which they are performed often typical of the species of larger groups of animals. Yet as such, they are not directed to one special end. A waterMbird may dive in order to escape or to seek for food ; a bird may fly for migrating, or in search of material to build its nest. On the other hand the escape of a water-bird may take place with the help of flying, of swimming or of diving. For this reason it would be erroneous to regard these general motor mechanisms as instinctive actions, as is sometimes done, only because they are innate and typical of the species. They are the means with which instincts are executed, not instinctive actions themselves. The directiveness of an instinct to one special end is an important feature of it. This fact, namely, that instincts are all directed to a special end, puts the means into our hands to build up a system of instincts. It has sometimes been asked how many instincts some animal, or group of animals, possesses. This question cannot be answered so long as it has not been determined what is to be regarded as a unit of instinct. May we speak of one brood-caring instinct of the fossorial

ANIMAL

PSYCHOLOGY

47

wasp, or must we distinguish in this animal a digginginstinct, a hunting-instinct, a paralyzing-instinct, a burying instinct, or perhaps even more subordinate instincts, which together form the care of the brood in the wasp ? According to Weyrauch no less than 57 different instincts are involved in the construction of the envelope of the nest in social wasps, every not farther analysable connection between a perception and a reaction being regarded as an independent instinct. In this way the animal instincts run into thousands I It will be clear that instead of trying to enumerate in such a way the instincts of animals, it is much more convenient to construct a system of them, consisting of larger groups embracing smaller ones. By so doing the survey of animal instincts is facilitated. If, therefore, in order to construct such a system, we ask what is the end towards which instincts are directed, we may say that. in general they are all directed towards the one great end of self-maintenance, be it that of indi·vidual self-maintenance, the maintenance of the individual life, or that of maintenance of the species, the maintenance of the life of the species. This brings us to a first classification of instincts into two principal groups. Both of them in their turn may be divided into three sub-groups. The instincts of personal self-maintenance may be divided into those in service of bodily development, of self-sustenance and self-defence. The first sub-group embraces all instincts that lead the larvre to activities necessary for their metamorphosis, such as the seeking for a proper place for pupation, the weaving of a cocoon, the digging into the soil by larvre of Polychaetre ; further, the opening of the egg-shell by the young bird at the moment of hatching, etc. The second subgroup comprises a great number of instincts that are all related to the satisfying of bodily wants, such as the seeking for water and hunting for food, the attacking of the prey and the killing of it, the seeking for a place for sleep, etc. Even so simple an action as the pecking up of a grain by a bird must be regarded as an instinctive activity belonging to this class. Many, more complicated, actions of other animals too belong to this sub-group : the construction of

ANIMAL

PSYCHOLOGY

the pit by the ant-lion, and of the web by the spider as means to catch a prey. To the_ third group, that directed to the defence of life, belong all instincts of self-defence against an enemy, of flight in danger, the so-called instinct of "deathfeigning," the instincts of migration or of hiding when the bad season approaches, the making of holes, or shelters, for protection, and several more. To this sub-group must also be reckoned the . instinct of care for the body that many higher animals show, the aim of which is to keep the body in a good state of health (the so-called "comforting instinct"), like the cleaning of the body, the scratching induced by itching, etc., and further, the instinct of curiosity, an instinct that drives many animals (monkeys, rats, but also many birds) to examine new surroundings or an unknown object in them, the biological significance of which is to give the animal timely warning in case of eventual danger. The second group, that of maintenance of the species, may also be divided into three sub-groups. These are that of propagation of the species, that of attending to the progeny and that of associating with congeners. To the first sub-group belong all instincts directed towards the winning of a sexual partner, as, for instance, the seeking and following of such a partner, the courtship, and so on, and further, all that belongs to the mating itself. The second sub-group comprises a great number of instinctive activities : that of making preparations for the laying of eggs or the delivery of their young, the building of nests and holes for the protection of the young ones, the breeding and defending of the eggs, the nourishing,. caring and defending of the young, and so on. Finally, to the third sub-group belong all instinctive actions that are sometimes taken together under the head of one "social instinct." These are the instincts of keeping together with other animals of the same species, of communal attacking, hunting or defending, and so on, in short, all instincts that make social life in animals, be it in a herd of deer or a breeding colony of birds or a nest of ants or termites, possible and advantageous for the partners of such a union. In this way we arrive, not at an infinite catalogue of instincts, but at a system, that is practically useful for the

ANIMAL

PSYCHOLOGY

49 ordering of instincts, and at the same time giving us a survey on the manifold instinctive activities of animals. It is of course possible, and perhaps even necessary, to sub-divide again the instincts of the last sub-groups into smaller and more restricted sub-instincts and these again into groups that embrace yet simpler activities. We have not done this here because it would only be detrimental to the general surveyability. For the rest it will be clear that we shall not find all instincts, or even instincts of all these six classes, in every group of animals. It will, on the contrary, be the task of the student of animal behaviour to state what instincts each species of animal shows during the different phases of its life. The characteristic of instinets, however, which has always struck most strongly the imagination of all observers, is not the fact that instincts are innate, or that they are typical for the species, but the fact that instinctive activity in its whole complexity generally takes place so purposefully, although the ani1'11cal executing these acts does not know, and often cannot even know, the end towards which its own actions are directed. The latter is certainly the case with many of the instinctive activities that are related to the well-being of a progeny the animal will never see, and of the future existence of which it even has no idea. Let us imagine what passes in the feeble mind of a fossorial wasp, say a Sphex or an Ammophila, when setting out to procure provision for her future offspring. The only feeling she experiences at that moment is one of something in her body she will have to get rid of. That this something is an egg, i.e. a tiny little product of her genital glands, that will grow and develop into a living being, wanting for this development fo;>d and shelter, is of course something she does not know even on a second or third occasion, since she scarcely ever sees this egg and even then does not understand anything about the future history of such an object. Yet she flies out and behaves as if by careful study she had learnt that the egg would develop into a larva, and that this wanted food in the form of a paralysed grasshopper or caterpillar, and would now do her best to satisfy this want. The same holds true for simpler instinctive activities as well: the sitting of a bird on her eggs and even the sucking of the 4

so

ANIMAL

PSYCHOLOGY

hungry newly-born mammal, that sucks at the breast of its mother, without possessing any knowledge of the function of lacteal glands and the physiology of nutrition. This fact has led some philosophers to admit that instinct might include some, if not reflective then intuitive, knowledge of the world, and by this knowledge guide the animal to its goal. This view, however, is certainly untenable, if by intuitive knowledge we understand a real and explicit knowledge of things, not metaphysical wisdom. First of all, from a standpoint of science it is unthinkable how such an explicit intuitive knowledge could have been acquired by the animal in the course of evolution. But, above all, the animal often makes too many errors in executing instinctive actions to allow us to believe that it really possesses some knowledge of the end it is unconsciously pursuing. Spaier indeed admits that the instinct is not a knowledge but an innate knowinghow-to-act in a restricted number of cases, beyond which it may be stupid and awkward. Even this, however, if taken in the direct sense of the words, must either be rejected for the same reason we rejected the idea of intuition, or does not mean anything more than the acknowledgment that instinctive actions are purposeful to the end to which they are directed, and does not help us any farther. Instincts, anyhow, may be said to compel animals to perform actions that are purposeful to the end towards which they are directed. This thesis scarcely needs any evidence. That the sucking of the young mammal is purposeful to the appeasing of its hunger, that the migration of the bird is purposeful to the preservation of its life during the bad season, that the weaving of the web by the spider is purposeful to the catching of insects, and the collecting and storing of caterpillars by the wasp purposeful to the nourishing of its future larva are facts that are so self-evident that it is not necessary further to elucidate them. More interesting and puzzling are the cases in which instinctive actions lose their normal purposiveness and become futile, or even harmful to the animal. What may be the reason that in some circumstances instinctive actions lose their purposeful11ess ? The most

ANIMAL

PSYCHOLOGY

fundamental cause again is the fact that instinctive actions are performed by the animal without knowledge of the end towards which they are directed. If the animal does not know to what end it performs actions, how can it leave off doing so in a case where they become inappropriate ? A striking example is reported by Hingston. Messor barbarus is an ant which collects seeds of grasses and carries them into the nest, where they are peeled and stored, after which the husks are carried out of the nest to a special refuse heap, about eight inches away from the nest. Hingston once found a nest that by way of exception was built in a vertical wall. Now the ants did not take advantage of the convenient position of the nest opening to throw the husks out, but continued carrying them eight inches away and laying them carefully against the wall, as if to make a heap there. Naturally the husks always fell down directly ; yet for months the ants continued doing this useless work. The deposition of the husks at a distance from the nest had become a no longer understood rite. Hereby comes the fact that instinctive actions, if repeated for some time, show a tendency to rigidity, to become routine work ; the action then is executed without any discrimination of changes in the outer conditions. It is a well-known fact that if in the absence of the bees the hive be displaced a few metres, the old bees on their return assemble at the place where, before they left, the entrance of the hive had been, and only gradually find the new place of entrance. Their return to the hive then is controlled not by perception of its place, but by a routine act that now fails them. The most important reason, however, of such unpurposefulness of instinctive action lies in the fact that the perception that incites the instinctive action in many cases is too indefinite to give to the animal a discrimination between appropriate and inappropriate execution of this action. Instincts, it must be remembered, do not unexpectedly and without any inducement come into full action, as Athena suddenly appeared in full armour from the head of Zeus. An instinct, as we shall have to illustrate presently, is evoked by a cognition, be it a simple sensation or a more complex

52

ANIMAL

PSYCHOLOGY

perception. And it is in this initiating perception that the cause of an inappropriate functioning of an instinct is often to be found. Observations and experiments in recent years have shown that in many cases the initiating perception is much simpler and much poorer in content, than one would suppose, and that for this reason it becomes a key that may open the door to a wrong road. Young chickens, for instance, as we know from Morgan's classic observations, in the first hours of their life do not pick only at grains of corn and other edible objects, but at any small object within their reach, so that they also pick at little pebbles, pieces of paper, the toes of their brothers, little spots of light on the ground and other not edible things. That experience of the good and bad effects of their picking speedily intervenes and leads the chicken to pick at edible objects is a question that does not concern us here. In this case the picking instinct, therefore, is released, not by the perception of small pieces of food, but by the much less definite, and much simpler perception of "a small object." New-born mammals do not only suck the teats of their mother, but anything warm and soft, so that young dogs may be observed sucking the tail or the ears of their mother, or the snout of one of their brothers. Hungry larvre of dragonflies not only throw out their "mask" (their prehensile organ) towards moving insects, but to any not too large object moving in their neighbourhood, such as pieces of paper or even moving spots of light. The cause of all these errors lies in the imperfection of the releasing perception. Other instincts than the feeding instinct are also sometimes misled by the too simple character of the releasing perception. The black-headed gull, according to Kirkman, sits not only on its own eggs and on those of ducks, but also on wooden eggs and little tins and balls, and even on pieces of coal-apparently on any object of small size. For the laying of their eggs insects are attracted by the smell of appropriate food for the larvre. But often they are misled by similar odours given forth by wrong objects. Insects, for instance, that lay their eggs in putrefying flesh may also lay

ANIMAL

PSYCHOLOGY

53

them in flowers with a similar scent, and the butterfly Papilio demolus that lays its eggs on orange trees does this, according to Vosseler, also on stones to the leeward of those trees that are impregnated by their odour. Lissmann observed that the first phase of the battle of the fighting fish Betta splendens can be released by the simple perception of a moving object of its own size, even by that of a cross with arms of I em. large and 4-5 cm.long. Of a couple of partridges reared from the eggs by Heinroth, the female, when arrived at sexual maturity, treated her nurse as a sexual partner, while the cock showed a hostile disposition towards her, only because she wore a brown apron of the same colour as the breast fleck of the cock. Apparently the brown patch on his breast, not the cock as such, stirred the sexual instincts of both animals. It would not be difficult to quote many more cases in which the all-too-simple character of the releasing perception, of the "releaser," as it is often called, is the cause of instincts working in an unpurposeful, or even detrimental way. It must be acknowledged, however, that in most of these cases this unpurposiveness is the fault of man, that great disturber of the harmony of nature, be it that by interfering with nature he creates conditions in which an instinct is apt to function in an inappropriate way, be it that he confuses the animal by teasing it with what he himself calls his "experiments." Hingston records that the Indian ant Camponotus compressus builds her nest at the foot of high trees. This is a suitable place for it, as the ant feeds herself on the juices of insects that live on the leaves of trees, so that she merely has to climb the trunk to arrive at the food. Since, however, man builds walls, nests also are found at the bottom of such walls, along which the ants go out in vain for food. The instinct to build a nest at the foot of a high object was purposeful only as long as such objects were trees. The new-born lamb of the sheep of the Pampas, according to Hudson, has the instinct to run away from every approaching object, and to follow every object that retreats. In the natural life of the herd this has the result that the young animal stays with the mother and so keeps in contact with the herd. In herds that

ANIMAL PSYCHOLOGY 54 are looked after by man, on the contrary, this has often the effect that the animal joins a departing shepherd, or horse, and so gets lost from the herd. But it is especially by human experiments that animal instincts are often turned into caricatures. The caterpillars of the processionary moth Cnethocampa processionea ·owe their name to the fact that in the evening they leave the common nest in great numbers in search of food. Then they march in single file one after another, the head of each animal touching the back-end of the animal before, at the same time spinning a thread which, united to those of the others, forms a silken band. This band keeps the animals together and facilitates the way back to the nest. In this procession there is no special leader : any one that accidentally proceeds at the head of the column is slavishly followed by the others. Fabre now succeeded in letting the one in front fall in behind the one in the rear, whereupon it at once gave up its leading function and followed the animal before it. The result of this was that a ring was formed in which the animals kept marching on round and round for more than seven days. It was only on the eighth day that finally the ring was broken by some smaller groups. Fabre computed that they were in motion for at least 84 hours, and covered the circle about 355 times. A similar result was later obtained by Wheeler with the wandering ant Eciton schmitti. It is therefore clear that for the efficient working of instinct it is necessary for the releasing perception to be so unequivocally determined that any erroneous reaction towards an inadequate stimulus is precluded. The examples just quoted show that this is not always the case. The indefiniteness of the perception, then, is the cause of the errors that so mar instincts. But it is better in such a case not to speak of errors ; as Russell rightly remarked, the animal does not make errors, but is misled. One striking case in which in nature a beautiful and complicated instinct is led astray by the all-too-simple character of the releasing perception may be quoted from Fabre. The larvre of the beetle Sitaris pass their development in the cells of the solitary bee Anthophora. To this end the Sitaris

ANIMAL

PSYCHOLOGY

55

mother lays her eggs at the entrance of the nest of this bee. The young larvre hibernate in these galleries, till in spring, when the larvre are already seven months old, the young bees leave their nest. At that moment the Sitaris-larvre attach themselves to the hairy bodies of the bees. Now these young bees are mostly males, as the males come out earlier than the females, and it therefore is necessary for the larvre to go over on to the females, which can only take place at the moment of copulation. If this succeeds the larvre attach themselves to the thorax of the female bee and try to pass over to an egg of Anthophora at the moment this is laid. If this also succeeds the larva can feed herself first on the egg and then on the contents of the cell of the Anthophora. Now the value of this interesting instinct of the Sitarislarva is, however, diminished by the fact that the perception which has to release the transition to the female Anthophora is too indefinite. Not only do the larvre fail to distinguish sufficiently between a male and a female Anthophora, whereby many of them remain attached to bees of the wrong sex, but they even do not distinguish between an Anthophora and other insects, and attach themselves to wrong species, like flies or honey-bees. Fabre even observed that they may cling to inanimate hairy objects. In such cases they naturally perish, as for their further development they are strictly adapted to the life of Anthophora. Here again the releasing perception is not differentiated enough : it is not the perception of a female Anthophora, but of any hairy object moving along, which releases the act of attaching. That the number of eggs laid by the beetle is so great that it compensates this imperfection of the instinct, does not alter the fact that by this too-schematic character of the releasing perception the purposeful working of the instinct is often disturbed. These cases, now, bring us to another point of importance. Is the animal, when performing an instinctive action, so strictly bound, so little capable of adaptive changes of its action, that it cannot by any means leave the road prescribed by its instincts so as to make its behaviour more appropriate to circumstances differing slightly from the normal ones ?

ANIMAL

PSYCHOLOGY

This question brings us to the much-discussed point of the rigidity or suppleness of the instincts. Are instincts so rigid as not to allow any freedom of action to the animal if altered circumstances need some change in the normal instinctive behaviour ? Fabre indeed maintained this, not only for theoretical reasons, but also from the result of his experiments. Let us, therefore, describe some of these experiments and their results. One experiment, often quoted, was the following. The fossorial wasp Sphex catches and paralyses grasshoppers, drags them to her nest and there buries them after having laid an egg on them. Before drawing the grasshopper into the nest she first lays it down at the edge of the nest and enters without it for a last inspection. If the nest is found to be in a good state, not occupied by another wasp or damaged in any way, she seizes her prey and draws it in. Fabre now, during such an inspection, removed the grasshopper some inches away from the nest. When the wasp after finishing her inspection came out and discovered the grasshopper at some distance from the nest, she drew it back to the edge of the nest again, but did not pull it in, and anew entered her nest for an inspection. Fabre took this opportunity to remove the grasshopper once more to some distance ; the wasp on coming out of the nest once more drew it to the edge of the nest and then again entered for an inspection. As Fabre then again laid the grasshopper aside, the wasp acted again as before. This little comedy was repeated some forty times ; then Fabre abandoned his experiment. The wasp, therefore, proved to be unable to bring a grasshopper directly into her nest if it lay at some distance from it, but had first to inspect the nest. In other words : the perception of a grasshopper lying at some distance from the nest impelled the wasp first to draw it to the edge of it, then to inspect the nest and only then to pull it in, even if she had inspected the nest just before. Or more generally : a special perception always released a special action, even if this action had become unnecessary and senseless. In some other cases, too, in which, as often happens in complex instinctive activity, one part of the complex is

ANIMAL

PSYCHOLOGY

57

executed only after a preceding part of it has been finished, it appeared that the connection between the two parts was so firm that the second part could only be performed in direct connection with the first. The mason-bee Chalicodoma muraria builds a cell from a mixture of earth and the secretion of her own salivary glands ; this cell is then filled with honey and pollen on which she lays her egg. When provisioning the cell the bee first goes forward into the cell to empty her crop filled with honey ; then she goes backward into it to wipe the pollen from her abdomen. Fabre now prevented a Chalicodoma from executing this second act by pushing her away with a straw at the moment she tried to enter the cell backwards. Then the bee again started to enter the cell in a forward direction, although her crop was now empty, and only after this tried to enter it backwards to get rid of the pollen. As this once more was prevented by Fabre, she again went head first into the cell, and so on. Ultimately the bee only partially entered the cell or merely put her head into it ; yet she was not able wholly to omit the first action before performing the second one. In this case the link between the two parts of an instinct-complex proved to be very rigid. In many cases it was shown that if animals are performing one part of such an instinct-complex, they are not able to return to an earlier part of it. When a Chalicodoma was building her cell, Fabre made a hole in the bottom of it. The bee discovered and repaired the damage. When, however, the hole was made when the bee was collecting provision for the cell, she did not repair it, although she discovered the damage and touched the edges of the hole, but went on provisioning the damaged cell with honey and pollen, which were immediately lost through the hole. Finally she laid her egg in the empty cell. During and after the provisioning the bee, therefore, was not able to return to an earlier part of the instinctive complex, namely, building the cell. It has not been Fabre only who observed such cases of rigidity of instincts. In a species of the wasp Ammophila, that brings more than one caterpillar in her nest, Ferton met one specimen that was just closing her nest after having

.ss

ANIMAL

PSYCHOLOGY

brought in two caterpillars and having laid an egg on them. He then laid a caterpillar, paralysed by another wasp, close to the nest. When the wasp found this caterpillar she inspected it, laid it down at the edge of her nest and re-opened it. When she then found the nest already filled she closed it again, but then she found the caterpillar again, which induced her to open the nest anew. When for the second time she found the nest already filled she closed it definitely and flew away, anxiously avoiding the place where the paralysed caterpillar was lying. Although the rigidity of the instinct in this case was not so absolute as in the former cases, yet the perception of the paralysed caterpillar twice forced her again to open the nest she had herself just filled and closed. An interesting case has been reported by Hingston. The solitary wasp, Eumenes conica, builds a dome-shaped cell and after finishing it fixes an egg in the top of the dome and provisions the cell with a number of small caterpillars. Hingston now cut away the top of the cell before an Eumenes had laid her egg in it. The wasp at once noticed the lack of the normal place for the egg and, although within the dome there was plenty of room to fix her egg, she could not bring herself to lay the egg on any other than the usual place, so that after some vain efforts she laid it there, where the top of the cell should be, with the result that the egg was shot into the air and lost. . Most descriptions of such a rigidity of instinct are concerned with insects. But higher animals may also sometimes provide examples of such an incapacity for changing innate behaviour in cases that deviate but slightly from the normal. When the viper Vipera aspis finds a mouse, she bites her, lets her go and after some time follows her trail till she finds the dead mouse and swallows her. Baumann in an experiment changed the bitten mouse for a fresh one. When the viper found her, the fresh mouse was not bitten but treated as if this had happened already. The mouse then defended herself against the viper and in many cases succeeded in escaping. Only if the mouse in defending herself jumped on the viper was she bitten. There was nothing to prevent the viper from biting again, but the rigidity of its hunting instinct was such

ANIMAL

PSYCHOLOGY

59

that it prevented it biting a mouse more than once, even if this was necessary. Even birds sometimes give proof of an amazing rigidity in their instincts. With breeding pigeons the parent birds relieve each other at a fixed hour of the day : the female sits from the later afternoon till the next morning ; the male from the morning till the afternoon. Lorenz, now, once watched a pair of pigeons of which the female was killed by a cat. The ·male then did not brood longer than normally, went out for food at the time the female had to relieve him and did not sit on the nest the night, but beside it. As the night was cold the young died. Nevertheless, next morning the male sat on the dead bodies of the young ones till the afternoon. For two days this senseless behaviour continued. Here, again, the animal showed an incapacity to adapt its behaviour to changed conditions and was not able to desist in an action that had lost its purpose. It would be wrong, however, to believe that with these examples of rigidity of an instinct the last word on this question has been said. On the contrary, as many examples of the opposite kind might be quoted. First of all, other observers, when repeating Fabre's experiments, did not always find the same proofs of a rigidity of instinct as Fabre had found. When the Peckhams, with another Sphex, repeated Fabre's experiment of removing the grasshopper, the wasp after four times dragging the grasshopper to the edge of the nest, the fifth time drew it straight into the nest. To tell the truth, even Fabre, when repeating his own experiment another year with another Sphex, found that after some two or three times this wasp also drew the grasshopper straight into the nest. The same was the case with a Pompilus, with whom Ferton tried Fabre's experiment. And when Ferton repeated the experiment he had before done with Ammophila, and laid a paralysed spider beside the closed nest of a Pompilus, a species of fossorial wasp that collects spiders in her nest, the wasp, after finding the new spider, did not open her nest again, but dug a new hole and buried the spider in it, after having laid its egg on it. Also Hingston, who reports so many instances of rigidity

6o

ANIMAL

PSYCHOLOGY

of instincts, found cases of suppleness in the working of instincts where Fabre had found none. We have already mentioned that Fabre's Chalicodoma did not repair a hole made in her cell while she was filling it with honey, but went on filling it as if nothing had happened. When Hingston did the same with an Eumenes that had flown out in search of caterpillars, the first caterpillar brought back by the wasp fell down through the hole. A second one was caught just hanging partly in and partly outside of the cell. At first the wasp did not pay attention to the damage, but when she had brought in enough caterpillars she inspected the hole, pushed the hanging caterpillar with great trouble into the cell and flew away to fetch sand to close the hole. A rigidity as asserted by Fabre, therefore, is certainly not the rule in instinctive behaviour. On the contrary, a closer examination of the instinctive behaviour of insects, spiders and other animals has shown that there are three groups of facts that stand out in contrast to this idea of their absolute rigidity. First, there is the great variability in instincts and instinctive actions. Then, instincts in many cases show adaptation to small changes in the external conditions. And, finally, they often show a regulation of behaviour that goes farther than simple adaptation. Let us now consider some instances of these three aspects of instinctive life. First a few words concerning this variability of instincts. Nearly all investigators who after Fabre studied the instinctive life of animals, have emphasized the fact that although instincts in general follow a fixed schema, yet in details instinctive actions show a variability that is incompatible with the idea of their absolute fixedness. Especially the Peckhams, in contrast to Fabre, have always laid stress on this variability even in the behaviour of fossorial wasps, of which Fabre quoted so many instances of rigid behaviour. Never, according to them, are two nests of Ammophila entirely similar ; never do we observe in them exactly the same way of digging a hole, of paralysing a caterpillar, of dragging the prey to the nest. One time the caterpillars in the nest will be dead, another time they are alive and well or only slightly paralysed. According to Hingston, Sphex

ANIMAL

PSYCHOLOGY

6r

lobatus on one occasion stings her prey only once, on another two to five times. With Eumenes we shall sometimes find four paralysed caterpillars in the nest ; at others eight to ten. As a rule the wasp lays one egg in each cell; sometimes, however, we find nests without any egg or with three. And when we come to the instinctive behaviour of higher animals, like birds and mammals, the variability is so great that if often seems as if their actions are not ruled by any leading instinct at all. It will be clear that, although all such variations themselves are of minor importance and follow the general rule of variability of all vital phenomena, yet they are contrary to the idea of an absolute rigidity of innate instincts. More important as an evidence against their absolute rigidity are the adaptations instincts often show if required by special circumstances. If the normal material for a nest, or the normal prey, is not available, the animal often knows how to manage with other objects. The fossorial wasp Pelopaeus hunts spiders of the genus Epeira, but if these are not to be found she may catch other ones. The parasitic bee Osmia lines her cell with the petals of flowers ; Osmia papaveris does this with the red petals of the poppy. Ferton now found that in Corsica, where this flower is lacking, the wasp uses the yellow petals of Glaucium luteum, and in the Pyrenees the blue petals of Malva. The same Osmia builds her nest in holes in wood or in the spaces between stones, by preference, however, in the empty shells of snails. Then the form of the nest is adapted to the available space : in long stalks the cells are placed in one long row, in openings between stones in an irregular heap, in snail-shells cells are built in a simpler row in the narrower convolutions of the shell, while nearer to the opening of the shell the cells are laid side by side. Even the very stereotyped form of the web of the spider may be more or less modified if the available space requires an adaptation· of it. Now even these adaptations are mostly not very radical deviations from the normal instinctive course, although it would be difficult to reconcile them with the idea of strict rigidity. More decisive, therefore, are the cases in which animals, in executing instinctive actions, prove to be able to

ANIMAL

PSYCHOLOGY

change their behaviour if important changes in the conditions demand this. The clearest instances of such regulations we find where an animal is under the necessity of repairing damaged constructions. We have already mentioned two cases with contradictory results. Fabre's Chalicodoma did not repair its damaged cell while at work provisioning it; Hingston's Eumenes did, after first proceeding with its provisioning. The former wasp, therefore, was not able to break off her normal behaviour and return to a preceding link in the chain of instinctive actions ; the latter showed a regulative change in it. With another Eumenes that already had filled her cells and had laid her eggs on the collected caterpillars, and that was now beginning to build a cover of mud around it, Hingston opened one cell and took the caterpillars away. At first the wasp went on working on the cover; after a dozen journeys, however, she began to repair the damage. When this was done she came back with a caterpillar, started again with provisioning and finally finished her normal task. More important and complete in this respect were the results Verlaine obtained with a Pelopreus, a fossorial wasp of the Congo. In imitation of Fabre, he partially destroyed the cells of these wasps at different stages, took the collected spiders away, and so forth, to see if the wasps would be able to repair the damage. It was only seldom that the wasps went on as if nothing had happened; in 86 out of 91 cases they did mend the damage. A hole in the side of the cell was repaired when this had been made during the building of the cell, or when the cell was ready, or when the egg was laid, or even when the wasp discovered the damage when she was already working at the provisioning of a new cell. If Verlaine gave her a cell filled with spiders, the wasp at once commenced closing it. If her egg was taken away, the wasp laid a new one ; if the spider with the egg was removed, the wasp brought in a new spider and laid an egg on it. In all this behaviour the normal sequence of actions, namely, the building of the cell, the bringing in of a spider, the laying of an egg, the provisioning of the cell with more spiders and the final closing of the cell, was broken off by the wasp ;

ANIMAL

PSYCHOLOGY

she was able to return from a later to an earlier action and also to omit an action which was part of the regular sequence. Of a fixed rigidity of the instinctive behaviour here no trace could be observed. A construction that often served as an object for experiments on the rigidity, or suppleness, of instinct is the web of the spider. Does the spider repair her web after damage, or not? To understand these experiments better we must first say a few words on the way a cobweb is constructed. In the common garden spider Epeira this in brief takes place in the following way. The spider begins by constructing a framework of threads, first of triangular form but afterwards made irregularly quadrilateral by spinning a horizontal thread within the triangle. From the centre of this framework she then spins a number of radii, or spokes, alternately on opposite sides of tl;le centre, and making almost equal angles with each other. The number of these spokes varies with the species, each species of spider having its own fixed number. Then, in the centre, these radii are connected together by other threads, forming an irregular texture: the hub. When this is ready the spider constructs a provisional spiral of ordinary thread, spun from the hub outwards to the circumference and passing from one spoke to the next one. The function of this provisional spiral is to serve as a scaffolding for the spider when she makes her final, or viscid, spiral that serves to hold the insects that are unhappy enough to fly into the web ; this final spiral, in contrast to the provisional one, is spun from the periphery to the centre of the web. In spinning this final spiral the spider uses the provisional one to cross over from one radius to the next one. The viscid spiral differs from the provisional one by being provided with viscid globules by which the insects are held. When the viscid spiral is ready, the spider destroys the provisional one. Finally round the centre a notched zone is constructed in which she sits awaiting the prey that will fly into the web. In some species we find slight deviations from this general schema, which it is not necessary to enter into here. Fabre cut the web of an Epeira into two. The spider,

ANIMAL PSYCHOLOGY 64 when returning to her web, discovered the damage, since she no longer found support for her legs on one side. She then spun two· threads from the hub to the other side of the breach, but did not repair the web and kept sitting on the remaining half. In another experiment he cut away the whole viscid spiral of a web. Then, too, the spider did no repairs, but kept sitting on a web consisting only of the remaining radii. Hingston came to similar results with the spider Araneus nauticus. When the spider was weaving her viscid spiral he, in one sector of the web, cut away one or two parts of the provisional spiral: Although the spider seemed to notice the absence of the threads she did not replace them, but went on to an inner winding of the spiral to pass from there to the next radius. When Hingston destroyed all the parts of the provisional spiral in one sector of the web, the spider did not lay new bridges between the radii, but every time she came to that sector she went to the hub to reach the next radius. And when he destroyed all connecting threads between the radii, the spider went every time from the periphery to the centre, and from there back to the periphery, rather than build a new provisional spiral. The result of this behaviour was an entirely irregular web with threads partly sticking together. The conclusion to be drawn from these experiments would seem to be that the spider does not repair her web if it is damaged after being once finished, and even does not repair damage made while she is engaged in another part of her web-building activity. Once again, however, other experiments had different results. In particular it appeared that the radii of the web are easily repaired, probably because the perception of the loss of the tension of the web induces the spider to rebuild them. Hingston found that if during the weaving of the web a new radius was removed, the spider replaced the removed radius by a new one as often as 25 times. Wiehle and Peters found also that, if during the weaving of the web the tension was much disturbed, for instance by cutting through the framework or taking away one of the spanning threads, the spider repaired the damage by making new threads. If a large part

ANIMAL

PSYCHOLOGY

of the web was taken away, Peters observed that not only radii and threads of the framework, but even the viscid spiral might be repaired. Lesser damages of the web, according to Peters, are repaired if made by the spider herself, for example when she loosens the prey from the place where she has spun it in, in order to suck it out in the centre of the web. The results Verlaine obtained with young spiders of the genus Epeira, only a few millimetres in length, go yet further. If in this case the provisional spiral was destroyed, this was wholly or partially repaired, even when the spider was already working at her viscid spiral. If in twa sectors of the web all provisional threads were taken away, they were replaced by some irregular threads and then the viscid spiral completed. If the whole provisional spiral, and the first viscid thread, were removed, the spider built a new web on the old radii. If a great number of radii were removed, the spider went so far as to demolish the already completed part of the viscid spiral, the provisional spiral and sometimes even the remaining radii, to build a new web in the old framework. With such young spiders Verlaine therefore found a much greater power of regulation of the instinctive behaviour than had been found by other authors with older ones. That instinct is not so rigid as has been supposed, has thus been made clear for the web-building of the spider. But now the question arises how it can be that different experimenters obtain such different and contradictory answers with regard to the question of the rigidity or suppleness of instinct, and that, e.g., Hingston might be quoted for, as well as against, both views. Of course it must be taken into account that they often worked with different species and not always under quite the same circumstances. But there must be yet another reason for their differing results. It seems probable that there are periods in animal life when the animal is more susceptible to outer impressions, and therefore more supple in its behaviour than during other periods. Especially the youth of an animal, or rather the period in which an instinctive action is performed for the first time or one of the first times, seems to be a period in 5

66

ANIMAL

PSYCHOLOGY

which deviations from the innate succession of actions are easier than at a later time. It is significant in this respect that Verlaine, who obtained the strongest evidence for the suppleness of instinct, worked with quite young animals, while others, like Fabre, probably worked with older ones. In later life the activity has lost its youthfulness, has become more or less a routine-action, from which it is more difficult for the animal to deviate. But besides this, it is probable that during the unwinding of the chain of instinctive actions there are moments of greater and lesser suppleness and rigidity coinciding with fluctuations in the attention of the animal which is sometimes more intensely directed to the work it is executing (to the exclusion of what is happening around it) than at others. We have already seen that Hingston's Eumenes did not at first repair the damage done to its cell, but when she had finished with her provisioning she paid attention to it and began to close the hole ; in another experiment by the same author she first finished making the cover of her cells and then started anew to repair the damaged cell and fill it with caterpillars. It may be possible, therefore, that adaptations to changes in the situation are lacking, simply because these changes did not attract the animal's attention, were not observed by it. An interesting case of this variation in the attention given to changes in the situation was reported by Baerends. He worked with Ammophila campestris, a species that differs from other Ammophila-species in that it makes two nests at the same time, and after having brought one caterpillar into each of them and having laid an egg on it, further provisions the nests according to needs. For this she opens her nest for an inspection, at first without bringing a caterpillar with her. If it then appears that the developing larva wants more food, she brings in more caterpillars. Baerends now found that at this inspection the wasp takes notice of any striking changes in her nest and adaptively reacts to them. If, for instance, the larva has been taken out of the nest, or a living larva replaced by a dead one, or if the larva has been replaced by an undeveloped egg, the wasp stops further provisioning. If, on the other hand, the egg has been replaced by a larva, the

ANIMAL

PSYCHOLOGY

wasp at once starts provisioning. If the experimenter augmented the number of caterpillars in the nest, the wasp herself brought in only a few ; if, instead of a growing larva a spinning larva or a cocoon was brought into the nest, the wasp closed it without provisioning it further. In all these cases the behaviour of the wasp showed adaptive regulation to the situation noted : the opening of a cell was not blindly followed by provisioning, whether necessary or not. Quite different, however, was the case when the wasp had already commenced upon the provisioning itself. If then the larva was taken out of the nest or replaced by an egg, the wasp continued bringing in new caterpillars ; if every time the new caterpillar was taken away, the number of caterpillars brought in was no larger than usual, and then the nest was closed. During this provisioning all regulation was absent, and the rigidity of the action once started stood out clearly. But it is time now to come to a conclusion about the rigidity, or suppleness, of instinct. This conclusion must be that this rigidity is certainly not so fundamental a character of instinctive behaviour as was formerly believed. By nature an instinct is more or less supple, and it is only when an instinctive action has been repeated several times, that the original suppleness may give place to a greater or less rigidity. Rigidity, then, must be regarded as a sign that the instinct has aged, or as the result of flagging interest for, and attention to, changes in the environment. It goes without saying that this suppleness, on the other hand, must not be overrated : the instinctive behaviour follows the schema of the innate instinct, and we may as little expect an ant-lion to come out of its pit to chase insects as a tiger to get aversion to killing living animals, and be converted to the eating of fruits and grass. One more question about this suppleness must, however, be answered. How must we understand these regulative adaptations of animals to changes in the situation ? Must we regard them as a proof of intelligence and ascribe them to an intelligent insight on the part of the animal ? Some students of animal instinct have indeed done so.

68

ANIMAL

PSYCHOLOGY

This, however, cannot be right. First of all, as we will see in a later chapter, insight is a faculty of mind only to be found in higher animals, not with the animals in which we here found these examples of suppleness in instinct : the insects and spiders. Further, such an insight presupposes a conscious knowledge of the end to be attained and the means necessary to obtain it, and we saw that at the level of instinctive behaviour such a knowledge is and often must be wanting. If we admit insight in the case of adaptive regulation of behaviour, the errors of instinct in other similar cases remain inexplicable. If there is insight in instinct, this insight is of a supraindividual, not to say a metaphysical, nature. But we prefer to reserve the term insight to individual performances, and not to class all cases of teleological behaviour under the concept of insight. For the explanation of such regulative adaptations therefore we shall have to look into the concept of instinct itself. In doing so, we shall have to admit that within the range of instinctive activity there are more possibilities than can be discovered by simple observation alone. The perception of the cell-in-the-making evokes in Eumenes the action of fetching sand and building further on it ; the perception of a hole in the finished cell evokes the action of fetching sand and repairing the hole. The perception of a nest wanting some caterpillars evokes in Ammophila campestris the action of hunting for caterpillars ; once her attention is focussed on the fact that there is not a larva but a cocoon in the nest, this perception induces her to close the nest, although otherwise she would have gone on provisioning it. And, likewise, once the attention of Verlaine's young spiders is drawn to the fact that the provisional spiral is missing, this discovery brings them to stop spinning the viscid spiral and to recommence building the provisional one. The laying of another egg if the first one has been taken away is induced in Pelopreus by the perception that the egg is missing on the spiders ; egg laying, therefore, is not only evoked by the perception that she has filled her cell with spiders. The same holds true for the little adaptations cited above : the adaptation of the form of the nest of Osmia to the room available, the accepting of an

ANIMAL

PSYCHOLOGY

abnormal prey if the normal victim is wanting, and so forth. On the other hand, just as different perceptions may give rise to similar actions, similar perceptions may evoke different actions. We have already spoken of the variability observed by several students in the instinctive activity of fossorial wasps. The perception of the cell to be filled sometimes evokes in Eumenes the collecting of four caterpillars, another time of eight or ten. The perception of the nest to be closed, according to the Peckhams, may once incite Ammophila to lay a clump of earth on it, another time to put a little stone into it and to fill it up with sand ; again another time to lay two clods of sand just in the opening and to cover them with smaller clods and sand. It therefore becomes clear that at both ends of the instinctive chain, on the side of the perception as well as o·n the side of the action, there may be some margin, some breadth, that allows of variations in the normal course, adaptations to small changes in the surroundings, and a regulative change where the conditions require it. We have already pointed out, however, that one must not overestimate the possibilities of this instinct-breadth. We must also keep in mind that there are differences in this breadth of instinct, differences between the different instincts of one animal as well as differences in the same instinct of different animals. The herring-gull, which is omnivorous and feeds itself with nearly everything it finds on the beach, with the eggs and young of other birds, and the waste from the tables of man, has a much broader feeding instinct than the tern, which catches its food only by diving into the water. In song-birds the nest-building instinct shows greater breadth than the feeding-instinct : the individual nests of such birds, therefore, show greater difference than the food they eat. In general we may say that the higher the animal, the wider its instincts. So the instincts of birds and mammals generally are wider and less definite than those of insects and spiders. We may say that in insects the instincts determine not only the ends but also the means ; in the higher animals little more than the ends are determined. This is the reason why in higher animals instincts are more easily overlooked than in the lower ones, and their role in

ANIMAL

PSYCHOLOGY

the behaviour of the higher animals is often underestimated. As we shall see in a later chapter, this breadth of instinct is of great importance in the acquisition of experience and the working of intelligence. Before ending these general remarks on instincts and instinctive actions we must say a few words regarding a curious feature of this activity, recently discovered by Kortlandt and by Tinbergen. If, for some reason the normal issue of an instinctive activity is interrupted, the energy, set free by the drive, may "spark over" to another movement mechanism and so seemingly evoke another instinct, for which there is no motive at that moment. For instance : two male birds meet at the boundary of their breeding territory and there make threatening movements at each other as a preliminary to a fight ; yet they do not proceed to actual fighting, but begin to pick up food or nest-material, in which, however, in reality no food is swallowed and the bill of the bird often does n~t even reach the ground. The reason why the fighting instinct was suppressed is to be found in the antagonistic instinct of flight, released by the perception of the opponent on the boundary of his own territory and ready to defend it. Another reason of such a sparking over of energy may be found in the fact that the end of an instinctive action has been reached too soon, so that there remains a residue of unused psychical energy. This may happen if the opponent evades the fight by flight. Then the victor often pretends to pick up food, bathe, clean his feathers, and the like. Hitherto most of the examples of such "sparking-over movements," or "displacement reactions," as they are also called, are found in birds, but there are indications that they may also be found in higher animals, perhaps even in man. The origin of this sparking-over of energy, the function of which must be to provide a means of discharge for a not wholly realized drive, remains obscure. In the foregoing pages we have described the general characteristics of instincts : their specificity, their innateness, their direC'tiveness, their purposefulness. We further discussed their rigidity and suppleness. We will now endeavour shortly

ANIMAL

PSYCHOLOGY

7I

to pass in review the principal forms of instinct found in the animal world. We have already pointed to the fact that an instinct does not come into action without some preceding cause, but is evoked by a cognition of some kind. What is the character of this cognition, and are there differences in the nature and degree of complexity of instinctive actions, correlated with the nature and complexity of their releasing cognition ? Here we must first make a distinction between a cognition of inner, and one of outer, origin. Instinctive actions are sometimes evoked by inner cognitions, sometimes by outer ones, often by both. In these inner cognitions we may group all bodily sensations, like hunger, thirst, fatigue, sexual stimulation, etc., evoked by special changes in the body. Here we must sharply distinguish, however, between the initial sensation and the agreeable, or disagreeable, feeling that follows it : the sensation of an empty stomach or the sensation of weakness caused by the fact that no food has been eaten for some time, is quite different from the feeling of discomfort or pain that sooner or later follows this sensation. These inner sensations now evoke a striving, directed to the appeasement of these bodily wants. If this is achieved, if an object is perceived suited to satisfy the want, a second phase is initiated which for the moment does not interest us here. The actions evoked by this striving to appease bear the character of spatially undirected searching activities. The hungry animal, experiencing the bodily sensation of hunger and the feeling of discomfort caused by it, goes in search of food till an adequate object is found with which to satisfy its hunger. The same is the case with the animal which experiences thirst, or is sexually stimulated, and the same with the insect that feels in itself the urge to lay eggs (i.e., experiences the sensation of having eggs to get rid of), and again with the caterpillar that experiences the sensation of being ripe for pupation, and so on. Perhaps only the tired animal, experiencing the sensation of bodily exhaustion, is usually not obliged to seek for a suitable place to rest but can lie down where it is. Yet for their night-rest many animals have to seek for a suitable place.

72

ANIMAL

PSYCHOLOGY

It is not necessary to quote instances of such a seekingbehaviour, evoked by inner sensations, in the higher animals. As to the lower ones, we have already seen in the first chapter in this book that if Amoeba has been suspended in the water for some time without contact with a fixed object, it stretches out its pseudopodia to all sides in search for such an object to attach itself to. If this object is found, the second phase of the behaviour is initiated, and the animal goes over to it. It is probable the feeling of hunger that incites Hydra, after it has remained quietly in one place for a long time, to leave this and crawl round till it fixes itself at another place in the aquarium. And even in the case of the ant-lion, that, as we saw before, was regarded by Doflein solely as a reflex machine, Doflein could not help stating that when the animal has been waiting for some time in vain for a prey, it becomes restless and moves away till at another place it starts again to build a new pit. In the figures Doflein gives of the path described by the animals in these locomotor activities, in which circles and spirals are drawn by the animals, the lack of spatial directiveness of these seeking-movements is clearly expressed. Now it is an interesting fact that, if such inner drives are very strong, if the animal is very hungry, or very strongly sexually stimulated, or very much driven by a need to breed, and the appropriate object to appease this drive is lacking, the perception which is able to release the second phase of the instinctive activity becomes more simple, less differentiated, so that other objects than the usual ones are able to r:elease the action which, in cases of a less strong drive, would be released ·by the normal object alone. The actions released by such inadequate objects are sometimes called "substitutive" or "symbolic" actions. Birds may then be observed to brood on stones instead of their eggs, hungry animals to eat things that are not suited for their food. Many cases quoted before as examples of an inappropriate functioning of instinct through errors of perception may be regarded as examples of such substitutive actions as well. A strong inner drive, then, requires so little outer stimulus for releasing an instinctive activity that this activity in many

ANIMAL

PSYCHOLOGY

73 cases becomes sterile. Some investigators even go so far as to believe that under the urge of a very strong drive instinctive actions may be executed without any outer perception at all ("Leerlaufreaktionen" of Lorenz). The examples quoted to show this, however, are not wholly convincing : it is not always easy to prove that an animal really does not perceive anything at all at such a moment. Anyhow it is certain that under a strong inner drive perceptions but little differentiated may be able to release the whole chain of instinctive activity. A greater number of instincts, however, are released by cognitions of outer origin, be these simple sensations or more or less complex perceptions. But in many of these cases inner cognitions also play a role. It is only in a state of hunger, when experiencing the bodily sensation of an empty stomach and the discomfort caused by it, that the beast of prey attacks its victim ; it is only when it is in heat that the male behaves in a special way at the perception of a congener of the other or the same sex. In such cases, therefore, an instinct may be said to be of mixed origin. But in such cases the role of outer cognitions is so much more decisive for the execution of the action than that of the inner sensations, that we may regard the latter simply as the conditions necessary for releasing and full functioning of the instinct on outer perceptions, and therefore for the present we may pass over them here. Now, these outer cognitions may be simple sensations, or more or less complex perceptions. Let us begin with the instincts evoked by simple sensations, such as those of light, gravity, heat, moisture, etc. These instincts may be divided into two classes, according to whether the instinctive action is, or is not, directed with regard to the location of the source of stimulation. In the former case we speak of orientatinginstincts, in the latter of alarming-instincts. Let us first say some words about these alarming-instincts. Many animals show the peculiarity that sudden sensations evoke in them a typical self-protecting behaviour. Animals that live in tubes or other protecting coverings contract, when stimulated by light, or by shadow, or by both of these. So the tube-worm Serpula uncinata contracts in its tube when a shadow falls on it, while it does not react to a ray of light ;

ANIMAL PSYCHOLOGY 74 snails, like Helix or Limnaea, in this case retract their antennre, Pecten doses its shell, etc. Other animals only react to light, as for instance the Ascidian Ciona intestinalis. Some mussels, such as Mya arenaria, react to light as well as to shadow ; in the former case by contraction of the sipho, in the latter by contraction of its tentacles. Even the tortoise Testudo ibera draws its head in under its shell if a shadow strikes it, but does this also at the perception of a moving object in its neighbourhood, even if this does not throw a shadow on it. Here the same reaction may be evoked by a simple sensation as well as by more complex perception. 1 Other sensations, too, may evoke similar actions. Sensations of touch have the same effect on sea-anemones and tube-worms : they retire under the sand or contract their tube. The tube-worm Spirographis spallanzanii according to Winterstein retires into its tube if, at a short distance from it, an organ-pipe is blown in the water. Hydra, according to Mast, contracts if a capillary tube with hot or cold water is placed near to its neighbourhood. All these actions are to be understood as actions of flight or self-protection at the perception of menacing danger. In other cases such alarming sensations may have a stimulating effect on the animal. This is especially the case with sensations of light. If the little Medusa Gonionemus is struck by a ray of light, according to Yerkes, it begins to move and swim round till by chance it comes into a shaded part of the aquarium, where it comes to rest. The result of this is that after some time all the animals find themselves in the darker part of the aquarium. This flocking together in the dark must be carefully distinguished from the directed movements towards the dark which we shall soon have to speak of. A similar alarm by light may be observed in many other animals. That this is more than a simple physiological stimulation by the light is shown by the fact that Asterias forreri, according to Jennings, if struck by a strong light, 1 Physiologists are inclined to call these reactions light- or shadow"reflexes.' For reasons stated in the former chapter we cannot adopt this terminology, but regard them as innate actions of the animal as a whole, directed to personal self-protection.

75 becomes restless and creeps about until it comes to a place where the intensity of the light is less. Should the animal be · engaged in eating at the moment of the illumination, it leaves its food and creeps away. The sensation of strong light, therefore, apparently evokes a feeling of discomfort or pain in the animal ; otherwise it would certainly not have stopped eating and, if it had been a question of physiological stimulation alone, probably would have started to eat faster than before. Other sensations also may induce flight-reactions of this kind. We have already seen that Hydra contracts when a tube with hot water at about 6o 0 is brought near to it. With water of a still higher temperature the animals loosen themselves from the soil and swim away, but the direction of this swimming, according to Mast, shows no relation to that of the stimulation. The woodlouse Porcellio scaber, according to Gunn, becomes active when the humidity of the air diminishes ; with the grasshopper Locusta migratoria, according to Kennedy, the contrary is the case, the animal becoming active when the humidity of the air increases, while it comes to rest when the air grows drier. For those who know the biological needs of these animals the purposefulness of these instinctive flight actions is not difficult to understand. Physiologists are inclined to regard all these activities of animals under the influence of light sensations as the effect of purely physiological stimulation. But the case of the starfish leaving its food when struck by light has already led us to understand that there is more in these reactions than a simple physiological phenomenon. This is shown still more clearly in a case described by Hovey with the marine flat-worm Leptoplana. This animal shows a strong reaction to light, and starts moving at a sudden light-sensation, caused when the stone under which it is hiding is turned over. On the other hand it shows a negative reaction and withdraws for a short distance when its head is touched. Hovey now succeeded in partly breaking the animals of their light reaction by touching their head every time they moved as the result of a sudden illumination. In control experiments it was shown that neither adaptation to light, nor tiredness, or any injury ANIMAL

PSYCHOJ.,OGY

ANIMAL

PSYCHOLOGY

to the head of the worms, was the cause of the change of their behaviour. The experience acquired by the animal, namely, that illumination was a signal for a subsequent touch on its head, induced it more or less to change its innate reaction to the light. The psychological element of experience, to which we shall return in a following chapter, therefore interferes in the innate reaction of flight from strong light. In this respect these alarming-instincts do not differ from the higher instincts we shall presently have to describe. More important, however, than these instincts in which animals are merely alarmed by special sensations are those in which such sensations induce animals to move in a direction related to that of the source of stimulation. These orientatinginstincts are known as " tropisms," " taxes " or " tactic movements.'' First a few words as to these terms. The term "tropism" was introduced by the old botanists for the description of the curving of growing plants under the influence of light. When later the movements of free-moving organisms, or plant-spores, towards, or away from, the light were discovered, these at first were also called tropisms. Later on a distinction was made between "tropisms," such as the curving of fixed plants, and "taxes" or "tactic movements" such as the movements of free-moving organisms towards, or away from, a source of stimulation. It is advisable to adopt this distinction. There are only a few cases of such tropisms in the animal world; thus, growing polyps of Hydrozoa, according to Loeb, grow out in the direction of the light. This directed growth, however, is probably caused simply by an unequal growth at the illuminated, and the non-illuminated side of the polyp, and there would seem to be no reason to admit that any psychical element is involved in it. So we may leave these phenomena and the study of them to the physiologist. There remain, then, the taxes and tactic movements. It seems profitable here to make a distinction similar to that we made before between instincts and instinctive actions. Taxes, then, are the inner factors, responsible for the directed

AN I MAL

P SY CH 0 L 0 GY

77

movements towards, or from, a source of stimulation ; these movements themselves we call tactic movements. Among these tactic movements we may again distinguish two main classes. The first one bears the name of "phobotaxis," the second one that of "topotaxis." These names were invented by the botanist Pfeffer, and afterwards adopted by Kiihn in his pioneering work on the orientation of animals in space. The term "phobotaxis" has undergone some criticism of later years ; it indeed suggests something not quite defensible, namely, the fact that something like fear (the Greek "phobos") lies at the bottom of these reactions. We, of course, do not know very much about the feelings of animals, especially of the lower ones, but it does not seem very probabie that anything like a real feeling of fear "'ould induce a Protozoan to react in a special way to light stimulation. In the term "topotaxis," on the other hand, nothing is suggested about a concomitant feeling. But as it is not wise to change technical terms that have become current, we will continue to speak of phobotaxis as well as of topotaxis. It is not necessary in this work to go very deeply into the difference between phobotaxes and topotaxes. There is a difference on the side of cognition that must be ascribed to a difference in the structure of the respective sense organs of the animals : phobotaxes are evoked by a sensation of difference in stimulation between two succeeding stimuli ; in topotaxis on the other hand there is a sensation of difference in stimulation between two simultaneous stimuli. Connected with these differences in the cognition there are differences in the movements that are caused by them : in phobotaxes the movements are not directly orientated with regard to the direction of the stimulus, while in topotaxes the movements are so orientated. But the final result is the same : either directly or indirectly the animal moves towards, or away from, the source of stimulation, in phobotaxis as well as in topotaxis. For the rest, phobic and topic orientation may alternate in the same animal ; many animals react in a phobic way at some distance from the source of stimulation, while nearer to it they react in a topic way, and in many cases the reaction is such that it cannot definitely be said if we have to do with

ANIMAL

PSYCHOLOGY

a phobo- or a topotaxis. From the psychological point of view the differences between the two groups are not very essential. We may therefore here disregard these differences and consider what the two classes of taxes have in common. This common character is that their movements are directed towards, or away from, a source of stimulation. This stimu.,; lation may be caused by a ray of light, by the pressure of gravity, by the heat from a warm spot. If the movement is directed towards the stimulus we speak of a positive taxis ; if directed away from it of a negative one. Sometimes also the animal orientates itself transversely, or obliquely, with respect to the stimulus. We may, however, disregard this form here. The best-known cases of such taxes are found where animals react to a stimulation by light. These phenomena are called "phototaxes." Most of the lower animals that have organs capable of perceiving either the intensity or the direction of a ray of light, show a positive or negative phototaxis. If a pseudopodium of a forward-moving Amoeba is struck by a ray of light, the forward movement is stopped, the pseudopodium retracted and a new one formed at another place. If this new pseudopodium again reaches the light spot, it is again retracted and a new one toi:'med, and so on, till finally a pseudopodium is formed that is no longer touched by the light. Then the whole body of the Amoeba follows this new pseudopodium, and the animal creeps away from the light. Here the sensation of a difference in successive light stimula~ions is followed by a movement that indirectly leads the animal away from the light, a good example therefore of phobotaxis. Similar phobic reactions towards, or from, the light may be observed in many swimming Protozoa, although executed in a somewhat different way. Another case is found with animals that possess eyes capable of the perception of the direction of light. If the young larvre of the Polychrete worm Arenicola cristata is affected on one side by the light, the animal, according to Mast, at once turns sharply towards it. The contrary is shown by animals like the flat-worm Planaria maculata, which

ANIMAL

PSYCHOLOGY

79 immediately turn away if a ray of light falls on them. Here we have examples of topotactic reactions, the former positive, the latter negative. It is especially in insects and their larvre that many examples of such light reactions are to be found. It is not possible to give a definite rule about the sign of the taxis ; this often changes with the age or the state of the animal, while hunger and sexual maturation may also influence the direction of the movement. Further, the sign of the taxis often changes with the intensity of the stimulus : animals that are positive to weak light react negatively in strong light . . It ia not difficult to find a parallel to these changes in the instinctive life of higher animals : if a small enemy is met with, the instinct of attack prevails, while if the enemy is big and strong the instinct of self-protection may prevail, and the animal seeks safety in flight. For the rest, in the higher animals movements released by such simple light sensations are rare ; their place is taken by actions released by more complex perceptions, although phototaxis may be found in the larvre of fishes and in young mammals before the eyes are open. Similar taxtic movements are to be found as reactions to other sensations as well. "Geotaxes" are movements upwards or downwards, against, or in the direction of, gravity, according as the geotaxis is negative or positive. If Medusre are disturbed, they swim to the surface of the water in a negative geotaxis ; on a strong illumination, on the contrary, they swim, in positive geotaxis, downwards to the bottom of the aquarium. In "hygrotaxis" the animal strives to go to, or away from, the water or a moist place. When Savory studied the behaviour of two species of spiders, Zilla-X-notata and Zilla atrica in a basin with water on one side and dry on the other side, he found that while the former species always went to the dry side, the latter looked for a place near the water. In "thermotaxis" we find a flight from extreme temperatures and an endeavour to find an optimal temperature, different for each species. "Rheotaxis" brings Planaria or fishes to react to the direction of the stream of water, whether by these movements it is prevented from being carried away

8o

ANIMAL

PSYCHOLOGY

by the stream, or whether by creeping against the stream it reaches places where food is more abundant. In the same way we speak of "anemotaxis" when animals sit with their heads directed towards, or away from, the wind. The sensation of touch may evoke positive or negative "thigmotactic" reactions : sometimes animals in a positive thigmotaxis try to keep in contact with a fixed body, like a stone, or stay in a fissure or hole ; in other cases they contract, or turn aside, in negative thigmotaxis if suddenly touched by a movirig object. Finally, we speak of "vibrotaxis" if animals react to the sensation of vibrations in their neighbourhood. The bestknown instance of such a positive vibrotaxis is the reaction of the spider, if some insect has flown into its web. Whether we may rightly speak of "chemotaxis" in the strict sense of the word is doubtful. Of course there are many positive or negative reactions to chemical stimulations in animals, but as a rule these stimuli are too complex to be regarded as arousing simple sensations. It is not thesensation of smell in general, but a special smell-perception that causes the animal to follow its prey or to seek for its sex-partner, the propinquity of which it has discovered by smell. These chemical reactions form a transition between taxes in the strict sense of the word and the more complex forms of instincts, which we shall soon have to consider. Now, what all these taxes have in common is that a sensation of some kind has the effect of making the animal move in a direction that stands in a certain relation to that of the source of the stimulation. For one who knows the vital wants of an animal at a special period of its life, it is not difficult to understand the purposefulness of these movements. If water is saturated with carbon dioxide, swimming towards the light in normal circumstances brings the waterflea Daphnia into regions where there is more oxygen, and a similar effect is reached by the negative geotaxis, shown by Paramrecium, if numbers of them are together in a small tube or if the carbon dioxide tension of the water is heightened. While tadpoles and young fishes in general are indifferent to light, they may be made positively phototactic if brought together in greater numbers in a small glass.

ANIMAL

PSYCHOLOGY

8r

Thus, Franz found that if r 5 tadpoles were put together into a glass of 4·7 em. diameter with 5 em. height of water, the animals swam round in all directions. If, on the contrary, he put them into a similar glass with 1.5 em. water, they suddenly showed a marked positive phototaxis and swam to the light. Phototaxis in this case must be regarded as an instinctive flight reaction, evoked by the constant contact with the other tadpoles. Koehler, at least, found that a similar phototaxis may be aroused in tadpoles swimming in small numbers in fresh water, if they are gently touched with a brush. It is a well-known fact that flies in a room, when suddenly roused, fly to the windows in positive phototaxis in an effort to escape. Positive thermotaxis brings parasites of warm-blooded animals to their hosts; positive rheotaxis, as we saw already, has the effect that fish in a stream remain at about the same place, thus preventing their being washed down by the current, although in many cases this effect is reached more by visual perception of the apparent movement of the banks than by the sensation of the flow of the water. So all tactic movements, aroused in the laboratory by various stimulations, may be regarded as instinctive actions, being under natural conditions directed towards the preservation of life or the flight from danger. Now these simple instincts or taxes may, like the more complex instincts, in some cases be changed by the influence of acquired experience. We have already seen that Hovev partly succeeded in breaking Leptoplana of its photokinesis. In a similar way Blees succeeded in temporarily breaking tht~ positive phototaxis in Daphnia. To this end he placed positively phototactic Daphnias in a bent tube, one arm of which stood vertically in the water, while the other arm, ro em. long, was kept horizontal. The animals were inserted with a pipette to the point where the horizontal arm began. At first this horizontal arm was directed towards the light ; the animals then immediately swam through it to the light. Then the horizontal arm was turned somewhat way from the light, first 45° 1 then 90°. The animals had now to learn that they had. first to swim obliquely or transversely to the light through this horizontal arm of the tube, and only after having 6

ANIMAL

PSYCHOLOGY

left it could they swim straight in the direction of the light. This learning was measured by the number of times they struck against the wall at the light side of the tube and by the time they needed to pass through the horizontal arm. After they had learned to do so, the horizontal arm was turned 135° and finally even 180° away from the light. The animals were now able, first to swim away from the light in the horizontal arm and then to move towards the light as soon as they found the way free. In learning this, the number of times they touched the wall of the tube and the time they required for the passage through it lessened. Thus, when the horizontal arm was turned away 135° from the light, the averages for three successive experiments with four animals on one day went down from 87 to 78 and 45 touches and from 425 to 325 and 235 seconds respectively. \Vhen the arm was turned straight away from the lig4t, the averages of three experiments with two animals on· one day were 39, 19, and 13 touches and 110, 48, and 30 seconds respectively. Untrained animals were not able to do this: four of them remained for I 5 minutes at the bend of the tube and could not be brought to swim against the light. It is clear that this change in the behaviour of the trained animals must be ascribed to experience. This experience was that to give up the tendency to swim to the light for some moments would bring them to a point from where swimming towards the light would again be possible. True, this experience is assuredly not so explicit as we here describe it, a point to which we will come back in a following chapter. But of whatever nature it may be, this influence of acquired experience on innate behaviour points to a psychical element underlying this behaviour itself. Taxes are not simple physiological phenomena, as some physiologists assert. They are the expressions of psychical phenomena in animals, just as the higher instincts are. For the rest, there are several intermediate cases between taxes and these higher instincts, between instincts evoked by simple sensations and those evoked by real perceptions. We have already seen that it is doubtful whether we may speak of real chemotaxis. That is, the positive- or negative reactions

ANIMAL

PSYCHOLOGY

to chemical stimuli are mostly evoked not by a simple chemical sensation, by the sensation of some chemical stimulus as such, but by a definite chemical perception, such as the smell of food or the scent of an enemy. In the visual sphere, too, simple perceptions may often evoke instinctive actions. As a first instance of these we may quote the cases of so-called "scototaxes." We have seen that many animals are negatively phototactic, i.e., they go away from the light. On the other hand, if black screens are placed in the field, some animals react positively to these screens and try to reach them. This was formerly regarded as a case of negative phototaxis, but later it has been shown that it is distinct from this. It is a positive reaction to the perception of a dark screen (and for this reason the term "taxis" here ia not altogether correct), whereas, in the former case, there was a negative reaction to a simple sensation. We saw further that the sensation of shadow or light induces tube-worms to contract. The same effect may be reached with Testudo, if it perceives a moving object in its neighbourhood. The mussel Pecten is alarmed by the perception of the movement of its enemy, the starfish Asterias, but also, as von Uexkiill observed, by the slow moving of a hand with open fingers along the wall of the aquarium. Another case of such instincts evoked by simple perception has been described by Mast of the firefly Photinus pyralis. The males of this beetle fly round in the dark, and send out a flash of light about every five seconds. The females sit in the grass and react to the perception of these flashes with their own flashes of light, of lesser intensity and of longer duration than those of the males. The males then turn towards the females until, after repeated flashes, the animals find each other and copulate. The males know how to distinguish between the flashes of their own sex and those of the females, and still turn towards the latter after their flashing is finished. The points of agreement and difference between these phenomena and positive phototaxes will be clear: it is a perception of light, not, however, of light in general, but of a special form of light emission, that evokes the approaching of the males. We have already seen that the perceptions that release

ANIMAL

PSYCHOLOGY

instinctive actions are in many cases much more simple than was formerly admitted and that just this simplicity is the cause why so often animals are led astray and commit errors, sometimes fatal to themselves. It is necessary, therefore, to distinguish between the apparent and the· real releasing perception in instincts. A closer analysis of this releasing perception will certainly show us many more examples of their simplicity. Of recent years much good analytical work in this line has been done. All this again proves that there are only gradual transitions between taxes and the higher instincts, and that it could be wrong to regard taxes as phenomena of a nature essentially different from that of instincts. The reader will not expect us here to give a survey of all these higher instincts in animals, higher in so far as they are evoked by more or less complex perceptions and not by simple sensations, higher also in so far as their actions are not simple movements, but form a more or less complicated behaviour. This field is so immense and the instinctive actions of animals are so innumerable that it would fill many larger volumes than this to enumerate them. For the rest we have already described a number of them in our general considerations about the nature of instinct. Let us conclude this section by describing one more example of such higher instincts, a case that may be regarded as· a standard example, in so far as several characters of instinct, e.g., its innateness, its purposefulness, the fact that the action is performed without being understood by the animal, the correlation between instinct and morphological structure, etc., stand out clearly. This, then, is the instinct of the care for the brood in the Yucca-moth, as it was first described by Riley some sixty years ago. The Yucca-moth, Pronuba yuccasella, is a small moth, the imagines of which come out of the pupa at the same time as the flowers of the Yucca-plant (Yucca filamentosa) open for a few nights. Males and females of the moth find each other and copulate ; then the female flies to an open flower of the Yucca, takes some pollen from the stamina of it and. kneads it into a ball with the help of her great sickle-shaped maxillary tentacles. She then carries the collected pollen

ANIMAL

PSYCHOLOGY

ss

away to another Yucca-flower. There, with the aid of her sharp ovipositor, she opens the ovary of the flower and deposits her eggs between the ovules of the plant ; then she climbs upwards along the style to the stigma, and presses the pollen she brought with her into the opening of it. Herewith she achieves the fertilisation of the plant, which without her help probably would not have occurred, as this moth is one of the few insects, or the only one, that pollinates Yuccas. Through this complicated behaviour, however, the moth achieves the propagation of her own species as well as that of the plant, as her growing larvre feed only on the developing ovules of the Yucca and would therefore perish if these ovules were not brought to development. And as, on the other hand, so large a number of ovules of the plant are fertilised that only a part of them will serve as food. to the larvre, ·both the plant and the moth have to thank this instinctive behaviour of the moth for their propagation. These actions of the moth indeed bear all the marks of instinctive behaviour. They are innate : no other moth has served as an example to the animal when performing its actions. The moth does not know, cannot know, the end towards which her behaviour is directed. That she has no idea about the necessity or desirability of the continuance of the life of her species, not to speak of that of the plant, is obvious. The behaviour is characteristic of the species ; all Yucca-moths behave in the same way. It is purposive, in this case not only for the species of the moth but, curiously enough, also for that of another living being, the plant. Interesting, further, are the morphological adaptations of the moth to her work; she is the only moth with an ovipositor, and the only one that possesses the great sickle-shaped maxillary tentacles which enable her to knead and to carry away the pollen. What outer perceptions exactly release the different parts of her behaviour has not yet been determined. Whether, for instance, she is attracted by the smell or the sight of the white flowers is not known, nor if some negative geotaxis plays a role in her climbing the style after having laid her eggs. As to the inner sensation, we may safely assume that the sensation of the ripe and fertilised eggs in her body

86

ANIMAL

PSYCHOLOGY

provide the inner condition for the seeking of a flower. With this example of a highly specialised and complex instinct of an insect we may conclude the description of animal instincts. We did not end with this one for the reason that animal instincts are not found above the insect level. As we saw already, instincts and instinctive tendencies are to be found in the higher animals as well. With them, however, they lose their narrower specialised character, so that even careful students of the life of higher animals are sometimes apt to overlook them. But it would be wrong to suppose that the instincts are not there. The whole behaviour of animals is based on the foundation of their instincts. As we will see in a later chapter, such actions of theirs as deserve to be called intelligent issue from their instinctive activity ; their intelligence is built up on their instinct. Instincts are the kernel of all animal behaviour, be it simply innate, or guided by intelligence. We quoted at the beginning of this chapter as a provisional definition of instinct, that of Romanes, pronounced some sixty years ago. Since then our knowledge concerning the essence of instinct has been considerably increased. Are we now, at the end of this chapter, in a position to give a more definite, and perhaps more analytical, definition ? With this in view we must bear in mind that an instinct is released by a cognition, the nature of which we have discussed above. We also know that an instinct ends in a striving, a conation, that impels to some activity, till an end is reached, the perception of which makes the animal cease its action. But is this all that happens within the animal, or is there one more link in the chain uniting the conation to the cognition? Let us consider the case of some higher animal, for instance, of the hare that sees or hears the hounds, and flies for its life. Or, rather, let us first consider that of a man in similar circumstances, a man who has committed some crime and now becomes aware that the police are at his heels. In such a case the criminal behaves like a startled animal, and

ANIMAL

PSYCHOLOGY

flies for his life, till some safe place is reached. But we know that in this case there is one more inner experience between the perception and the striving, namely, the feeling of fear, evoked by the perception of the approaching police. No one will deny that a similar feeling of fear must be experienced by the hare when pursued by the hounds : the scudding animal clearly bears all the signs of it. The truth is, therefore, that the perception of the pursuing enemy evokes the feeling of fear, and the feeling of fear in its turn evokes the desire to escape and the striving to get away from the danger, which causes the animal to fly, till a place is reached which evokes a feeling of security that brings the flight to a stop. In other words, between the cognition and the conation there is some experience that links them together,. namely, an affection of some kind. In a similar way another affection, say, a feeling of satisfaction at the perception of the goal reached, brings the conation to an end. The same is the case with other instincts. The perception of the grain would not induce the chicken to pick it up were it not moved by some feeling of appetence ; the female could not induce the male in heat to try to copulate with her without some feeling of lust evoked in him by the perception of the female. The same must be the case with lower animals, although the feelings of lower animals are certainly so very different from ours that it is difficult, if not impossible, to describe them adequately in terms of human experiences. But some feeling must lead the wasp to paralyse the caterpillar it has discovered and the termites in the nest to attend to the queen. We have already seen that with bodily sensations, like those of hunger or thirst, it is a feeling of pain or discomfort, caused by these sensations, that leads the animal to go round in search of the food or water. Some feeling of discomfort makes the Daphnia, in unclean water, fly towards the light. Even in Amoeba, floating freely in the water, as we saw in our first chapter, some feeling we shall not try to describe brings it to seek for a fixed object. Our inability to make any very definite statements about these feelings should not be a reason for us to ignore, or even to deny, them. \Vhen sailing on the sea we are not able to say very much that is definite about the animal

88

ANIMAL

PSYCHOLOGY

life below our vessel ; yet it would be foolish and erroneous if for that reason we should wholly deny its existence. These considerations bring us to a more profound and analytical definition of instinctthan that of Romanes. Instinct, then, is a linking, and indeed an innate and specific linking, between a cogn£tion, an affection, and a conation. But this linking does not confine itself to one direction only. Just as the perception instinctively calls forth the feeling, and the feeling the striving, so also inversely the striving influences the perception and the feeling of the animal. The perceptions of the animals are as dependent on their strivjng as their striving is on their perceptions ; the animal striving to have its hunger appeased perceives the prey that will serve as food, but probably scarcely perceives objects not suitable for this purpose. The male hare in heat will probably not notice the female mouse ; the hunting wolf will not perceive the branches· of the tree so important for the bird seeking a place to build its nest. By its striving the objects in its world acquire for the animal a special meaning, a special "valence," as it has been called. The rabbit has no valence for the calf, nor the clover for the weasel. And in the same way the feelings of the animals are dependent on their striving : in a male in heat a male of his own species evokes the feeling of hostility, whereas it is an indifferent being to him if he is not stirred by his sexual instinct. To a dog gnawing a bone his playmate of a quarter of an hour ago becomes a rival and an enemy. We may therefore conclude with the following dej£nit£on of an instinct. Under instinct in animals we understand that characteristic innate psychic disposition, through which special sensations or perceptions, evoked by special stimuli, evoke themselves special feelings and emotions, and these in their turn evoke special drives and strivings, which express themselves in special actions, directed to a special goal; while, on the other hand also, specialperceptions and feelings are evoked and influenced by special strivings. Or, to put it more briefly : Instinct is the innate psychological structure which couples a special affection to a special cognition and a special conation to a special affection, and on the other hand

0

-.1

c.::

s 0

u.J

>

i=

Q .....J ~

0

$

LJ..I

>

--, c:o

i= u U-1

VI

0

1...) U-.1

::::>

@

.st... ,...~ ~

e= ...

....= '()I)

~

ANIMAL

PSYCHOLOGY

couples a special cogmtion and an affection to a special conation. We have endeavoured to express this in the diagram on the previous page (p. 89). It will be remarked that this definition of instinct is a purely psychological one, i.e. it contains only psychological concepts, in contrast to many older definitions which contained biological concepts like "purposefulness," "vital needs," and the like. Instinct is a psychological phenomenon, and must therefore be defined in psychological terms ; the instinctive action is a biological phenomenon and may be defined in terms of biology. But as we are writing on the psychology of animals, the latter aspect holds no interest for us for the moment.

CHAPTER III

THE PROBLEM OF ANIMAL LEARNING I N the preceding chapter we have seen that instinctive actions are innate and specific, i.e. that they have not to be learnt by the animal, and are characteristic of the species to which the animal belongs and not of the individual animal itself. But it is the individual, and not the species, which is the real living being, a being which grows and develops, which perceives and feels, remembers and gathers experience. And as a result of all this, the innate behaviour of an animal may change during the course of its life. All these changes in the innate behaviour we summarize under the concept of "learning." Learning, then, taken in the broadest sense of the word, may be defined as that inner process by which changes are introduced into the original behaviour of an individual. But we must at once add that not all these changes have a psychological origin. Therefore we must distinguish between learning based on physiological and that based on psychological phenomena, and, although strictly speaking the first form of learning falls beyond the

ANIMAL

PSYCHOLOGY

scope of this book, it would seem necessary, to avoid confusion, to say a few words on this physiological mode of learning too. First, then, there is learning based on physiological maturation. The body of the young animal changes, it grows and develops. Its muscles become stronger, nerve connections grow and the innervation of special muscles becomes better than it was before. The result of this is that the execution of special movements improves. A good instance of this form of learning is that of the young bird learning to fly. That this learning in many cases is not the result of experience has been stated by Lorenz. Young wood-pigeons (Columba palumbus) leave the nest very early, before the flight-feathers are fully grown, whereas with the rock-dove (Columba Iivia) the young ones fly out later, when the wings are quite developed. Although the young rock-doves have no experience in flying and have had no room to practise, according to Lorenz they fly from the very start just as well as the young wr d.-pigeons which have been flying for some days. This is su1ely due to bodily maturation. That this is so has been proved experimentally by Grohmann. He kept young pigeons in small boxes in which they could not turn round nor by any means open their wings. One of them was set free after it had been confined in such a box from the 12th to the 37th day of its life. The bird at once began to flutter, twelve minutes later flew from the floor on to a 2s-cm. high box, after one hour flew up vertically to a height of half a metre and the next day flew as high and as well as a pigeon that had not been confined at all. Another bird, set free on the 55th day, at once flew away 40 metres on to a fence I .So m. in height. It was clear that the flying had not had to be learnt by experience or practice, but that the skill was the result of a physiological maturation. Similar results had indeed been obtained by Spalding some seventy years ago, with swallows, titmice, and wrens. Yet, as we soon shall see, these results must not be generalized. . Another change in behaviour, not attributable to a psychological factor, is that which is caused by physiological

92

ANIMAL

PSYCHOLOGY

fatigue or habituation of the sense organs. If not too strong a stimulus be applied for some time to an animal, the animal gets habituated to the stimulus, and after some time changes its· behaviour in so far as to react no longer to the stimulus in the same way as it originally did. There are a great number of cases known in the animal world of such a change in behaviour due to habituation. Thus Jennings found that Stentor and other Infusoria soon get habituated to a strec.m of water directed against them, and no longer contract. If Jennings let a drop of water fall from a height of 30 em. on to the disc of the Anemone Aiptasia annulata, the animal at once contracted completely ; if, after the animal had again expanded, a second drop fell, the animal in many cases hardly reacted at all and the reaction wholly ceased after some further drops. Arbacia, according to Holmes, reacts on a shadowing by erecting its spines, but if this shadowing is repeated three or four times the animal stops doing it. This disappearance of the reaction sometimes stands in relation to the interval between the stimulations. Thus Hargitt observed that if a colony of the tube-worm Hydroides dianth].ls was shadowed by means of a swinging pendulum, so as to give a rhythmical shadowing at intervals of one second, all animals contracted at each shadowing, while at an interval of half a second a number of them reacted no longer after the first shadowings, and with intervals of a quarter of a second practically all the animals became indifferent to the shadowing after the first time. It will be clear that in all these cases it is not necessary to admit psychological experience as. the cause of the change ; the phenomenon is based on a physiological habituation of the sense organs. The reverse of this habituation to repeated stimulation may perhaps be found in cases in which the sensibility to a stimulus is raised by another previous stimulation. Sgonina found in Paramrecium that electrical stimulation lowers the threshold for optical stimuli and so causes a negative phototaxis in animals that before were indifferent to the light. If the conclusions he draws from these experiments are correct, we must admit that here again there is a change of behaviour on a physiological basis.

ANIMAL

PSYCHOLOGY

93 Less easy to analys~, and probably based on physiological as well as on psychological phenomena, is the change of behaviour through practice. When a billiard-player improves his play by practice, the change from his original more or less clumsy performance on the billiard-table is certainly for a great part based on the experience of former occasions when trying to hit the balls, not to speak of an element of reasoning and computation at the moment of striking. But if a child learns to walk, the improvement in its walking is not so much due to former experience, as to a better command over its movements caused by a better physiological functioning of necessary reflexes. In animals there are several instances of improvement in behaviour by practice, in which it is not easy to say whether it is effected· in the way of the billiardplayer or of the child. Kortlandt reports that young cormorants at first fly very awkwardly, but that after twenty hours of flying there is no longer any difference between their flying and that of adult birds. The fact that it takes rather a long time to reach this stage seems to prove that in this case physiological maturation alone is not responsible for this improvement, but that it must also be partly ascribed to practice. The same was found by Dennis with young buzzards. Unfledged birds, kept in boxes for ten weeks, after being set free, flew much worse than equally old birds not encaged. The cause of this difference can only be lack of practice. If this be so, the learning to fly must with these birds be of the same nature as that of the child learning to walk. If, on the other hand, a bird of prey by practice obtains skill in seizing its prey, there probably is also an element of experience in this improvement. But more interesting to us are the cases in which learning is based on psychological factors alone. Here again we must distinguish between two possibilities : the learning may either be the result of a psychical maturation, or it is the result of experience. When a child is learning to speak, its learning for a large part is the result of psychical development : a backward child learns to speak only with great difficulty, no matter how much trouble its parents may take to hasten the process. The instances of deferred instincts, mentioned in our last

ANIMAL PSYCHOLOGY 94 chapter, may be regarded as cases of "learning" by psychological maturation, the word learning being taken in the broad sense as defined before. As an example of this we may refer to the development of the activity of the bees in the hive, as amply discussed in the preceding chapter. The most important to us, however, are the cases in which the change of behaviour, in other words, the learning, is due to acquired experience. For here the behaviour of an animal reaches a higher level. Where former experience plays a role in the behaviour of an animal, we may say that this behaviour is no longer guided solely by instinct, but a new element has appeared on the scene, namely, intelligence. This intelligence in its different forms will be the subject of the discussions in the following chapters. It was only to preclude confusion that we thought it advisable, before proceeding, to say a few words on the different ways in which learning may be effected in the animal world, and to point to the difference between a non-intelligent and an intelligent way of learning.

CHAPTER

IV

THE PROBLEM OF ANIMAL INTELLIGENCE IF, now, we are to discuss in this chapter the problem of animal intelligence, it will be well to ask ourselves first, what this idea of intelligence may mean in the field of animal psychology. The answer to this question is not at once clear. The concept "intelligence," it will be realized, is borrowed from human psychology. Now, in man, intelligence is not a simple function of the mind ; the concept of intelligence, therefore, is for the human psychologist not so simple and clear as is the concept of "heart" for the anatomist, or that of "bloodcirculation" for the physiologist. "Intelligence" embraces in man a number of collaborating and interlocking psychical functions, like those of attention and memory, of imagination and ideation, of reason and judgment and the like, which,

ANIMAL

PSYCHOLOGY

95

all together, influence and improve his original innate behaviour. In human psychology, therefore, the word intelligence is often used in rather different senses, according as the accent is laid on one or another of these functions, so much so that Spearman even goes so far as to doubt if the word has yet any scientific meaning at all. As a rule it is restricted to denoting a function of practical life, while the word "reasoning" or "intellect" is reserved for the capacity for conceptual thinking. Intelligence, then, is the capacity to deal adequately with new and unexpected circumstances, the capacity in such circumstances to grasp the essential in an object or situation, and to use this insight to reach a desired end. But, as we said, there is no general agreement in this matter. With the animals the case is somewhat different and up to a certain point, more simple. Some of the functions mentioned above are not to be found in animals, or are to be found only in such an embryonic degree of development that we may ignore them. Many animals, and certainly the higher orders, exhibit the phenomenon of attention, and even among the lower orders many animals show memory. But do animals show imagination, and do they have free ideas ? One may doubt it. Anyhow, the faculty of abstraction certainly fails them. So with the animals the concept of intelligence may have a more restricted meaning than in man. With them we may relinquish its abstract aspects and use the word in its practical sense alone. But, on the other hand, if with the animals we try to apply the above-mentioned definition of "grasping the essential in an object or situation," and the "using this to reach a desired end," that we found useful in the case of man, we encounter the facts of instinct and instinctive activity. Let us consider once more the case of the Yucca-moth, fertilising the Yucca-plant, or that of a fossorial wasp, paralysing her prey and dragging it into her hole as a food for her future offspring, or even the simple case of the young mammal, sucking at his mother's teats to appease his hunger. We certainly encounter in these animals a grasping of the essentials of the parts of the plant or of the body of the

AN I M AL

P'S Y C H 0 L 0 G Y

caterpillar or the breast of the mother, and the using of them as means to reach their particular biological end. Yet in these cases there is no reason to speak of intelligence. If, as we saw to be the case in instinctive activity, the end is not explicitly known, there can be no question of intelligence in the ·sense defined above of apprehending and using the essentials of a situation to attain a psychologically desired end. At best we might speak of an "implicit intelligencl-," if this expression has any meaning at all. The above-mentioned definition of intelligence, therefore, fails us in the case of animals. With animals, therefore, we must approach the subject from another side. We have seen that all their activity is stirred by instincts, instincts that are inborn and not acquired, and are typical for the species and not for the individual, so that originally all animals of the same species may be expected to act in a similar way under similar circumstances (provided for a moment we pass over such slight differences as are caused by the normal variation in instincts). If, then, in animals we wish to distinguish intelligent actions, actions that bear the mark of intelligence, from these instinctive ones, we must lay stress on such characteristics of animal activity as deviate from those of their instinctive behaviour. We then observe that much of animal behaviour is acquired and not innate, that therefore it is characteristic of the individual and not of the species. This behaviour, then, we will call "intelligent." If one agrees with this or, perhaps better, if this distinction between intelligent and instinctive activity in animals be accepted as being the only way tp obtain a clear distinction between the two and to outline them sharply, one will agree with the view which defines intelligence as "the faculty to improve upon inborn instinctive actions in the light of past experience," or, better perhaps, "to rebuild innate instincts in the light of such experience." It will be clear that this definition is wider than that given above for human intelligence, and is based on another aspect of intelligence qua general phenomenon. But it will also be clear that this is the most helpful definition if, in animals, we wish to distinguish between instinctive and intelligent activity.

ANIMAL

PSYCHOLOGY

97 It will be understood, now, that this profiting by past experience so as to improve upon innate behaviour presupposes two other faculties of the animal mind. The first of these is that the animal must be able consciously to notice and to experience its own actions and their consequences to itself ; the second one is that it must be able to retain these experiences for a longer or shorter time. The first fact is connected with the psychological level on which the animal lives. An animal like a dog or a bird experiences much more than does a worm or a snail ; it has therefore more opportunity to let this experience influence its later actions. The retaining of this experience, then, is the function of its memory. Many researches have been made on memory in animals, about the duration of their retention, its different modalities, and so on. We cannot enter into all these questions, which stand only in indirect relation to the question which concerns us here. But we must not fail to take into account that it is not necessary to admit that what is experienced is also explicitly kept in the animal's mind, that, in other words, it is not necessary to believe that the animal later explicitly remembers : "When I was in these circumstances before, I did so or so, with such or such consequences for me." Doubtless the process as a rule occurs on the lower level of implicit remembrance. That is to say, the action or the result of an action evokes in the animal a certain feeling, be it one of pleasure in the case of an agreeable issue, or one of pain or discomfort if the action had a disagreeable effect. Now this positive or negative feeling overflows on to, and effectively colours, the action which evoked the feeling, so that the action itself becomes attractive or unattractive to the animal. This feeling may even colour the object or perception which released the original action, so that this object is invested with a special agreeable or disagreeable feeling. The champagne we have learned to drink at feasts afterwards "looks festive" ; the draught we tasted "looks nauseous." The perception, then, is infiltrated with the pleasant or unpleasant experience of the action released by it; or, to use an expression of Hobhouse, the feeling aroused by the releasing perception has assimilated the character of 7

ANIMAL

PSYCHOLOGY

another feeling, which had followed it previously. If this process is repeated several times and the experience gained is strengthened by repetition, the original action in the one case will be executed directly and with pleasure, in the other case he§itatingly or not at all. Then the animal has changed its inborn behaviour in the light of past experience ; it has shown intelligence. We may express this also in another way. In our second chapter we stated that if the perception of some object or situation instinctively evokes some action from the side of the animal, this obj~ct may be said to possess some innate meaning, some innate "valence" for the animal. The cat, whether perceived by hearing or by smell, has for the mouse the innate meaning of danger, and releases the affect of fear and the instinct of flight. Now ~n object or situation which originally did evoke a certain instinct, in other words, which had a special innate valence for an animal, by experience may acquire another valence and release another instinct, or perhaps may lose its original valence and no more evoke any instinct at all. In the same way an originally valence-less object may by experience acquire one, and so release some instinctive activity. We may consider these to be cases of "re-valentiating" and "de-valentiating," and in the last mentioned case speak of "valentiating" objects by experience. All this new valentiating, all this change of valence, is the work of intelligence. A simple example may make all this clear. A young dog meets for the first time in his life some animal unknown to him. Then this animal releases his instinct of curiosity, better perhaps, his instinct of examination. Suppose now this animal to be a cat, which herself reacts in an instinctive way towards the approaching dog. If the dog is big and the cat young and small, the perception of the dog will instinctively evoke fear in the cat and release her instinct of flight. This reaction of the cat stirs the hunting instinct of the dog and imparts to the animal "cat" the new valence of a prey to be hunted. But if the cat is strong and valorous she will not turn to flee at the sight of the dog but will defend herself, in a way perhaps which induces fear in the dog, causes him to stop

ANIMAL

PSYCHOLOGY

99

in his examination, and releases his instinct of flight. The animal which originally had the valence of something to be examined may thus by experience acquire the valence of something to be hunted or to be feared. An object without any valence may in the same way acquire a valence by experience. We have already seen an example of this in the behaviour of Paramrecium in Bramstedt's experiments. Paramrecium, which originally is indifferent to light, for which originally light has no valence, may learn to evade the lighter side of a drop of water if it has learnt by experience that this light involves a high temperature. The originally valenceless light herewith acquires the valence of something to be avoided. It will be clear now that the breadth of instincts, of which we spoke in a former chapter, is of great importance for the acquiring of such experience, leading to the acquiring of a ne}V valence, and therefore for the entrance of intelligence into the activity of an animal. If the instincts were absolutely rigid, if they had no breadth at all, so that only one particular perception is able to evoke one particular action, the possibility of acquiring experience would be very much limited. In the best case the animal would only learn to desist from such an innate action, if this action had unpleasant effects for it, supposing the drive to be not so strong that the animal cannot suppress it. If, on the other hand, an animal instinctively reacts to a number of kindred perceptions in a number of different ways, there is more occasion for it to experience that one reaction will have a more agreeable result for itself than another. The tern, which we saw has a narrow feedinginstinct and is specialized for catching its food by diving into the water, has much less opportunity to obtain experience of different kinds of food and different ways of obtaining it than the gull, which is an all-eater and catches its food in different ways. If in the case of the dog the only possible reaction of his examination-instinct was that of running at high speed straight towards every unknown object, there would be little occasion for him to gain experience as to the way to treat such strange things. Actually he may now run fast, now walk slowly, or even creep towards it, may walk

IOO

ANIMAL

PSYCHOLOGY

prudently around it, may bark at it, may lie down and wait till the object moves, etc., according to whether the object is big or small, moves towards him or lies still. This variety of action gives him great opportunity to acquire experience, a very important thing, especially for an animal living in nature with all its threatening dangers. Now, as a rule such experience is gathered from the casual happenings of daily life. But these happenings may be influenced in two ways. First man may occupy himself with an animal and thrust some special experience upon it. This process is called the "training" of an animal. By agreeable and disagreeable experience (food or caresses in one case, punishment or scolding on the other) the master trains his dog to do what he wants him to do : to follow him at his heels in the street, or to stay in his basket during dinner. In the same way, in the laboratory, the monkeys or fish, whose sense perceptions are being studied, are brought to learn by experience that if they choose a box or an opening with a special mark this means food, whereas choosing another one means an electric shock or some other kind of punishment. More important to the animal, however, than this human training is the influence of the conduct of any other animals with whom he is living in a social bond. By imitating the acts of its parents or other older congeners (which occurs especially where an individual with less initiative lives in a social bond with a more enterprising congener, as is the case with a young animal living together with older ones), by following their example and doing what it sees them doing and omitting what it sees them avoiding, the animal obtains a store of useful experience more rapidly than if it had been living alone. Morgan has reported several observations on the effect of this imitation. If the hen begins to pick, the chickens follow her example ; they pick at what they see other chickens picking at. If one chicken begins to drink from a dish, the others soon follow. According to Lorenz, the flight-reaction of young jackdaws is not at first released by the perception of a dangerous animal as such, but by the sight of the frightened parents. If this has happened several times the younger ones are frightened by the

ANIMAL

PSYCHOLOGY

IOI

dangerous animal itself. Jackdaws that are reared by man therefore do not know danger, and the young of tame parents remain tame. Imitation therefore accelerates the process of obtaining experience. Before ending these general remarks on animal intelligence we must point to three further aspects of it that are of importance to the animal. The first one is, that by experience an originally blind drive may be changed into a more or less conscious striving. The young animal which originally, without knowing why, makes sucking movements on the teats of its mother, learns by experience that these movements serve to appease its hunger. If, then, later it is hungry, it will probably seek the teats with the explicit knowledge of the end it wishes to reach. The yourtg bird may the first time sit upon its eggs without any clear knowledge why it does so ; if it has several times seen young ones hatching from its eggs it is possible that a memory remains of this fact, so that then it will sit on the eggs with some vague idea of the young ones that will appear, without, of course, understanding anything about the physiological effect of its brooding. The effect of a former action, then, is no longer merely implicitly preserved in the strengthening or weakening of an innate instinctive activity, but experience may influence the animal's acts in a more explicit way, so that we may then have a gre::1ter right to speak of a striving, explicitly directed towards a goal. The originally purely instinctive action may thus reach the higher level of a willed action, of a volition. The second point that· we ·must stress is that, even if intelligence influences and rebuilds instincts and plays a part in the instinctive behaviour of an animal, it would nevertheless be erroneous to believe that its actions are now prompted by intelligence, an error often made especially in the case of man. It is always the instinct which provides the drive and the energy necessary to perform an action, intelligence only plays the secondary role of directing instinctively released actions in a special way. Intelligence-actions are a special case of instinctive actions ; behind the intelligence there is always the instinct which compels the animal to its actions. And then, finally, the most important change which

IOZ

ANIMAL

PSYCHOLOGY

intelligence imparts to an animal is that it places the individual on a higher level. An animal acting only under the drive of its instincts is no more than one of his species. By the working of its intelligence, by the after-effects of its own particular experience, a personal note comes to the fore. Intelligence transforms an animal to a personality among its congeners. After these general considerations on the intelligence of animals we shall try to give a survey of the different forms of it found in the animal world. Although we begin with the simpler forms, we shall summarize them without pretending to give them in any hierarchical order or to construct a system in which each further type means a higher form of intelligence. Needless to say that here, as always where distinctions are made in matters of living nature, sharp distinctions cannot be made, and some cases might be classed equally well with one type as with another. The first form we may distinguish, then, is that of psychological habituation. In the preceding chapter we saw that there is a habituation on a physiological basis, a simple physiological adaptation of the sense organs. There exists, besides ·this, a habituation based on acquired experience, although perhaps it will not always be possible to distinguish sharply between the two. The most marked, be it perhaps only a relative external difference between them, is that while physiological habituation is but temporary, psychological habituation may have a more permanent effect. The clearest example of such a habituation is provided by a wild animal living in a game reserve park, or captured by man and forced to live in his proximity. At first such an animal will show fear of men, and the sight of an approaching man will release its instinct of flight. When after some time it notices that no danger threatens from the side of man, and that its keeper even brings food and tends it, by this experience man will get a new valence for it and the sight of man will no more release the instinct of flight, but perhaps will even release the action of approaching the man in the expectation of getting food. ·Such a habituation may even occur in animals as low in the animal scale as larvre of insects. Thus Sondheim

ANIMAL

PSYCHOLOGY

IOJ

brought larvre of the dragon-fly Aeschna grandis so far as to accept flies from his hand, and even to come to the front of the aquarium as soon as he approached it. Animals in this way may become habituated to all kinds of originally threatening perceptions. The Peckhams found in the case of females of the spider Epeira labyrinthica that they let themselves drop from their web if the experimenter approached with a vibrating tuning-fork. After some repetitions, however, this reaction disappeared. This habituation developed gradually ; an animal which the first time had let itself drop 15-18 inches and there waited for several minutes before venturing to climb up again into the web, after the seventh experiment evinced much less alarm and after some twenty experiments let itself drop only an inch or two and then came back at once. After the twenty-second experiment it merely lifted its forelegs if the vibrating fork was brought near to it. It is clear that this change cannot be regarded as a case of simple physiological adaptation of the sense organs ; here the animal must have learnt by experience that the vibration of the tuning fork did not imply impending danger from which it had to flee. Miss Fielde reports how her ants in an artificial nest gradually became tamer, did not allow themselves to be disturbed by the cleaning of the nest and after some time no longer stung her if they were touched. Brun succeeded in forming alliance-colonies between different species of ants (e.g., Formica rufa, pratensis and sanguinea) in which the animals almost completely ceased their original mutual hostilities. This was not due to the forming of a complex odour out of the odour of the component colonies, for the animals still distinguished their congeners from those of the other species, and members of their own nest were treated as friends, while those of the other part were attacked although only mildly. This habituation occurred especially in difficult or unusual situations, as, for instance, in the face of a common enemy. An interesting case of such a psychological habituation was observed by tenCate-Kazejewa. The hermit-crab, Pagurus arrosor, lives in the shell of a snail. If the abdomen of the animal be touched through an opening in

104

A N I MAL

P SYCH 0 L 0 GY

the shell it at first contracts within its shell, but after a stronger stimulation leaves the shell and crawls round till it finds shelter in another shell, sometimes its own. If, now this stimulation be repeated several times, the animal at the first slight touch on its abdomen immediately escapes out of the shell, but then remains in the neighbourhood of it and returns to it as soon as the stimulation has ceased. By experience the simple reaction of contraction on being touched is replaced by one of a real flight, but on the other hand the animal also has learnt by experience that the danger is soon over, and remains in the neighbourhood. By this it proves to be more or less habituated to the originally frightening sensation of being touched. Habituation may, of course, also come about in the reverse direction. Animals in regions where no human beings occur first show no fear of them, as Darwin had already observed on his visit to the Galapagos Islands. Bitter experience, however, soon suffices to make them afraid of man. Although the term "habituation" does not seem to be very appropriate in this case the working of the intelligence in altering the animal's actions is in effect the same in both cases. Habituation may be regarded as a negative effect of intelligence, that is, the animal learns to desist from a certain action. More positive effect is shown in a second type of intelligence, namely, that in which one perception comes by experience to function as a signal for another. We already found an example of this in the behaviour of Paramrecium, which learned by experience that light was a signal for an uncomfortable temperature of the water. Instances of this signal-forming are very numerous among the animals. The worm Nereis virens lives in a tube in the sand at the bottom of the sea. If, now, in the aquarium some meat be placed before the tube the feeding instinct of the worm is released, and it comes partially out of the tube in search of the food. If, on the other hand, light or shadow is cast upon the tube, the worm retires deeper into it by way of flight. Copeland, now, in a dark room threw a light on the animal, and I 5-20 seconds later laid a piece of mussel

ANIMAL

PSYCHOLOGY

IOS

before the tube. After a number of repetitions of this experiment the worrn not only came out of the tube on the perception of the slowly-diffusing meat-juice, but reacted in a positive way as soon as it was stimulated by the light. By experience it had learnt to regard the light as a signal for the arrival of food coming, and the light released the feedinginstinct instead of the flight-instinct. A curious instance of such a signal-forming in lower animals was observed by Ada Yerkes. The tube-worm Hydroides dianthus contracts in its tube if a shadow is cast upon it. After some more shadowing this reaction ceases as the result of a physiological or psychological habituation. More regular and lasting is the reaction if no shadow is cast upon the animal but it is touched gently with a little rod. If Mrs. Yerkes let the shadowing be followed by a gentle touch, the reaction to the shadow did not stop, but the animal kept on contracting at the shadow. Here experience had taught that the shadow was a signal for the subsequent touching, so that the normal habituation was suspended. As we stated above, examples of such signal-forming are numerous among the lower animals ; signal-forming is indeed the most common form of intelligence in them. But among the higher animals also signal-forming is a very common phenomenon. The carp in the fish-pond which approach the land when a man comes to the edge of the pond, having learnt by experience that the man brings food to them, the dog who has learnt that if his master is dressed in a particular way this is a signal for a walk for them both in the fields, and many other cases might be quoted as instances of such a signal-forming. The whole life of intelligent animals is full of such signals, perceptions standing for something else. In this connection we must make a brief mention of the so-called "conditioned reflexes." That the term reflex, used by Pavlov and his followers, is in such cases mostly an inadequate indication of what really happens with the animal is a point we have already treated in a former chapter. It is certainly better to call these processes "conditioned responses," if it is preferred to keep to the term "conditioning" instead of using the more simple term of "learning." But apart from

106

ANIMAL

PSYCHOLOGY

this question of terminology, we may say that these conditioned responses, too, are good examples of an intelligent signal-forming in animals, based on acquired experience. If one of Pavlov's dogs secretes saliva on hearing a bell ringing, it is because this ringing has been followed several times by the giving of food. The sound thus has become a signal for the forthcoming meal, and the secretion of saliva is part of the action of his feeding instinct, an anticipation of the following actions of seizing upon food and chewing it. The so-called conditioned reflexes of dogs and other animals involve intelligence on the part of these animals, and Pavlov's term "psychic salivation" expresses this feature of the phenomenon much better than that generally in use. A third type of intelligence is found where experience creates a distinction between different perceptions which originally were not kept apart. We have already seen that the young mammal at first does not know where to suck, but sucks all that is warm and soft on his mother's body, her hair as well as her teats. But experience soon teaches him that sucking his mother's hair does not bring the desired appeasement, and that for this end the teats are the parts of her body to suck. This brings with it a distinction among the warm and soft parts ; some are likely to satisfy his hunger, others are not. Herewith the perception of the animal becomes more differentiated. A classical example of this learning to distinguish is presented by Morgan's experiments with chicks. Young chicks when newly hatched from the egg begin to pick at all small objects within their reach, whether they be grains of corn, or pebbles, or pieces of paper, or spots of light on the ground, or even their own toes. Experience soon teaches them, however, what can be picked up and what not, what is edible and what is not, what has a good and what has a bad taste. Intelligence thus brings more distinction into their perceptive world. Young animals, which at first do not generally know how to distinguish between innocent and dangerous creatures, learn this by experience with them or, as we saw, from the examples given by their parents. Especially the first hours of their life are for many animals often very important in this respect. Lorenz has shown that birds during the first hours

ANIMAL

PSYCHOLOGY

107

of their life gather much experience especially with regard to their congeners, an experience which is preserved for their whole life and in his opinion can never be reversed. It may, however, well be doubted whether this view concerning the irreversibility of first experience is right and whether there is an essential difference between this early experience and that which is acquired later in life. Nevertheless, it is certain that birds during their first hours pass through a specially sensible period of their life, in which their intelligence is very active in the assimilation of lasting experience. A further type of intelligence is that of the formation of simple motor habits. Jennings laid individuals of the starfish Asterias forreri on their back and then forced them to use a particular pair of arms in turning over again, by holding the other arms fast. One individual learnt in t8o experiments to use these arms only in three out of ten cases after all the arms were set free. Moore was able to train an Asterias to begin its reversal-movements with one particular arm by touching the other arms every time the animal tried to use them. A well-known example of such motor habits was given by Yerkes in his experiment with the earthworm Allolobophora foetida. He put this animal into a T -shaped tube of glass, the right arm of which ended in a dark wooden box, while the left arm was furnished with a piece of blotting paper soaked in a strong salt solution, or, in later experiments, with electrically charged copper wires. So by reward and punishment the animal had to learn to choose the right arm and to form the simple motor habit of turning to the right, every time it came to the point of bifurcation. This was attained rather quickly although the learning was not absolute and good days alternated with bad ones. These results of Yerkes were later corroborated bj other experimenters, while similar motor habits, such as the choosing of the right or left way out of an apparatus, were formed also by other animals, like snails and crabs. From this simple choosing between a right and a left path there is a gradual transition to the learning by higher animals to follow intricate paths in a maze. The learning of these complicated paths, however, shows some aspects that are lacking in the learning of these

108

ANIMAL

PSYCHOLOGY

simple paths. We shall come back to this maze problem in our next chapter. As an example of the forming of such a simple motor habit in a higher animal we may mention that of a guinea-pig, trained by Grindley to turn his head to the right as soon as a buzzer was switched on, which the animal learned by being rewarded with a piece of carrot as soon as it did so. These motor habits bring us to another type of the modification of instinctive action by intelligence, viz., that of the acquirement of the necessary skill in the performance of such an action. In the preceding chapter we saw that the improvement in instinctive action by practice may be based on physiological as well as on psychological phenomena, and that it is not always easy to say if only a physiological maturation is responsible for it, or if the improvement is the result of experience concerning the effect of certain movements. In that chapter we mentioned as such a doubtful case the learning to fly by young cormorants. Young chickens that from the earliest moments of life. peck at small objects, do not at first always get them. Their pecking, however, is soon improved, and here again it is not easy to decide from the partly contradictory results of different students how much of this improvement must be ascribed to a physiological maturation only and how much to the good or bad experience gained during the pecking movements. That a bird of prey profits by experience in the catching of his prey, that the same is the case with the squirrel improving his leaping from one branch of a tree to another, seems to leave no doubt. One of the clearest examples of this improvement in an instinctive action by experience is given by Lorenz with the red-backed shrike (Lanius collurio). It is part of the instinctive equipment of this bird to impale insects and other small animals on the thorns of branches. This behaviour is, however, not wholly innate ; innate is the drive to do so, but the bird has to learn that this impaling must be done on thorns. Young shrikes at first make the required movements with the food in their bill, but at random anywhere in their cage, and without paying attention to suitable places, such as nails or thorns. But if the bird by accident has experienced that food

ANIMAL

PSYCHOLOGY

remains hanging on nails, these subjects will then be used in the performance of its innate instinctive activity. A more' complex phenomenon is that of "homing" in animals. Under a "home" we may understand all places which for some reason the animal desires to reach again after having left it, be it because it finds a shelter there, or food, or because its progeny has been left behind. The finding of its home, if this is not directly perceptible by vision or scent or some other sense perception, requires the help of former experience, in other words of intelligence. The lowest animal of which it is known that it possesses such a home and is able to find it again after leaving it is the limpet (Patella), a snail which lives on the rocks in the tidal zone of the coast and in rest occupies there a slight hollow in the stone corresponding exactly to the outline of its own shell. For long it has been known that each Patella has its own place on the rocks, towards which it returns after having left it to feed on the algre of the rock. The problem now is, how the animal is able to find its own place again. This has proved to be rather a complex case. Patella does not simply retrace the tracks of its outward path. The direction of the light seems to play a certain, although not a preponderating part in the finding of the home again, in so far as the animal in returning may try to keep the light on the side other than that on which it was when it left its place. It seems that in the homing of Patella several kinds of experience are utilized : a general remembrance of the direction of the home with regard to the direction of the light ; a remembrance of the distance covered since the home was left; but, above all; a general knowledge of the structure of the surroundings nearest its home, based on tactile perceptions of the surface of the rock. Out of all this a "place-memory" is formed, based on the experience obtained on former occasions about the meaning of special characteristics of the surroundings in finding its place again. This knowledge of the surroundings is of course limited ; it therefore need cause no wonder that if a Patella be taken away from her place and put down on the rocks at some distance from it, the greater this distance is, the fewer will be the number of snails which succeed in finding their way back.

IIO

ANIMAL

PSYCHOLOGY

This acquiring of a place-memory by perceiving and retaining impressions regarding the characteristics of the home surroundings is very marked in insects which possess such a home, be it bees, ants or wasps, living in communal nests, or fossorial wasps, for which the home is the place where they dug their nest for storing provisions for their future offspring. Before leaving such a place such animals are seen first to make orientation-flights around the nest, during which they get acquainted with the surroundings and learn to know plants and stones and other objects as landmarks to guide them on their way back. Displacement of such landmarks by the interference of man causes the animals to become confused and they are no longer able to find the nest again. It is not possible to review here the extensive literature on this homing in animals, but it will be clear that this homing must be regarded as another and somewhat more complicated type of animal intelligence. As a last type of the lower forms of animal intelligence to be treated in this chapter, we must mention the so-called "learning by trial and error," a term introduced by Morgan and generally adapted by students of animal psychology, although it would certainly have been better to name it a "learning by success and failure," as these two words express much better what really is the foundation of the learning which results from it. This learning by trial and error, then, takes place in the following way. An animal finds itself confronted with a special task which it wishes to accomplish, e.g., to reach some place, mostly because food is to be found there, or to free itself from some confinement, or to do some other thing which an instinct impels it to do. This it tries to attain by making some movements or performing some actions which, although they are directed towards this end, are mostly at first merely accidental. It may then happen that most of these movements or actions are unsuccessful and do not further its aim, but one, or some, of them bring it to the end desired. The unsuccessful movements then get clothed with the negative feeling of being ineffectual ; those which were successful with the positive feeling of bringing the animal nearer to its goal. In the way described above the latter movements are

ANIMAL

PSYCHOLOGY

III

stamped in, and the former will be omitted, if another time the animal finds itself in the same situation. A classical example of this learning by trial and error is provided by Thorndike's experiments with cats and dogs in his problem-boxes. In these experiments the animals were confronted with a problem which originally lay beyond their capacities to solve. They had to open a box or case closed by a complicated mechanism, in which a handle had to be turned, or a latch had to be lifted, or a loop had to be pulled down ; or sometimes a combination of these manipulations had to be performed, in order to free themselves from a cage in which they were confined, or to open a box in order to reach food. At first therefore their movements were quite haphazard, but by selecting the successful ones out of all those that were made the animals as a rule sooner or later got so far as to be able to open the box at once by making the required movements. From the results obtained Thorndike drew some conclusions regarding the intelligence of his animals which were not very flattering to them, but were partly unjust and certainly were over-generalized. They were subsequently much criticized, and it is therefore needless to dwell upon them here. Much of the behaviour of higher animals, seemingly based on understanding, is undoubtedly no more than the result of such learning by trial and error. A good example of this was provided by one of Morgan's dogs. This animal, if let free in the garden, was wont to stand on a parapet wall and look through the railing to see what happened in the street. Once when, accidentally, he looked out under the latch of the gate, he happened to raise his head, so that the latch was lifted and the gate swung open. At first the dog did not notice what he had done, but as he saw the gate opened he went out of the garden into the street. According to Morgan's description, after some ten or twelve of these casual occurrences, in which the dog's exit from the garden was gradually effected with greater speed and with less gazing out at wrong places, he learnt by experience to go straight to the right place and there make directly the right movement. To a passer-by, then, it might have seemed as if the

Il2

ANIMAL

PSYCHOLOGY

dog had at once performed the right action with a clear understanding of the effects of his raising the latch, but to one who, like Morgan, had followed the whole process of learning, it is clear that this performance developed gradually by a strengthening of accidentally successful movements. This learning by trial and error leads us to a higher type of animal intelligence than any hitherto described. If a dog, like that of Morgan's, learns only by experience to 'execute one particular movement at one particular place to attain one particular end, this process cannot be estimated much higher than the learning of a crab to move to the right side at a bifurcation of the path, or even that of a worm coming out of its tube in order to get food if a shadow is cast on it. But are performances like these the best an animal can show ? Might not the animal be able in some cases to understand more explicitly the results of its own actions during those trials, and to apply them to other cases more or less dissimilar from those in which this experience was originally gathered ? If this were the case, we should have to admit that there might be more in such a performance than the plain execution of an action that has been stamped in by experience of its results ; some real understanding might in fact be involved in its actions. It will be our task in the next chapter to discuss whether anything like such understanding is actually exhibited in animal behaviour, whether an animal ever carries out with understanding a purposive action which is not innate, whether in other words we can find in the animals something more akin to what at the beginning of this chapter we described as intelligence in man.

CHAPTER V

THE PROBLEM OF ANIMAL UNDERSTANDING W E ended the preceding chapter by describing how, by a process known as "trial and error," an llnimal may come to

ANIMAL

PSYCHOLOGY

113

the proper performance of some particular, not innate, actions, necessary to reach an end towards which it is driven by an instinct, as for instance the opening of a box in order to get food or the unlocking of a garden-gate in order to get out. We then asked if this was the highest pitch that animal intelligence could reach. We may now answer this question by saying that it is not, but that, be it by this "trying and erring" or in another way, some of the more gifted among the animals may attain to a form of intelligence that we must consider as a higher type than that of only giving new or other valences to their actions ; to put it more accurately, we can state that some animals can in fact arrive at a real understanding of the means necessary to reach a desired end. Of what does this understanding consist, and how is it attained ? The answer must be that this understanding is based on the noticing of the essential elements in a situation, and the grasping of their mutual relations with regard to the end desired, followed by the ability to apply this knowledge in order to reach the goal. But here again we must realize that it is not necessary to assume that in such cases the animal forms explicit ideas of ·these elements and their relations, that it explicitly forms connections between them in its mind, that it judges and thinks. Here again the whole process takes place on a lower, more immediate, level ; it is rather a direct sensorial insight than the result of prior deliberation, more a "seeing" than a "thinking." Further, the understanding takes place on the level of the concrete, i.e. of the spatial and the temporal, and the abstract is not involved. For this reason we will term this form of animal intelligence "concrete understanding." It will be clear that, in order to obtain such an understanding, an animal requires a certain endowment, be it a specific or an individual one. Some species are better endowed, are cleverer, than others, but also within one species different degrees of endowment may be found. Even animals of the same litter may differ in this regard. Age and psychical maturation also play a role in this endowment: older animals, with much experience, are often cleverer than younger ones. The cunning of an old fox is proverbial. But even without 8

ANIMAL

PSYCHOLOGY

particular experience the ability to grasp such relations may increase with age. Lorenz reports that young Jackdaws which had at first been ·unable immediately to make a detour through a door, although they several times found their way through it by trial and error, after being caged up for four weeks at opce took the right way. As practice in this case was precluded, the increase of understanding can only be ;1scribed to psychical growth. As in other similar cases, it is not possible to draw a line anvwhere in the animal realm and assert that below this line only the lower forms of intelligence are found, whereas above the line a real understanding may also manifest itself. Higher and lower grades of intelligence cannot be sharply distinguished, higher and lower forms of intelligence may work together in the performance of intelligent actions by the same animal. If we let animals run through a simple or a more complicated maze, where does the frontier lie between the simple learning of a motor habit, as was the case with Yerkes' earthworm, and the real understanding of the path it has to follow, as we shall presently describe it in higher animals like rats ? Insects, in particular, with their instincts, on the one hand so complicated and on the other so often subject to regulation, frequently show behaviour which a casual observer would perhaps be inclined to ascribe to understanding, but which we shall probably do better not to value so highly. The Peckhams once observed a Pompilus enlarging the opening of her nest as it appeared that the spider she was bringing with her was too big to be got into it. Must we ascribe this to insight on the part of the wasp that the dislodging of some sand would remove the hindrance to bringing the spider into the nest ? Or must we believe that to enlarge the opening belongs to the instinctive equipment of the species, so that every Pompilus instinctively adapts the dimensions of the nest-opening to that of her prey ? It is difficult to decide this. Some authors, like Hingston, have been rather generous in ascribing all such behaviour to insight. It would seem better to remain somewhat more critical and, for the rest, leave the answering of this question open till more knowledge respecting the instinctive endowment of the species is available.

ANIMAL

PSYCHOLOGY

us

Now, the basis of all such understanding is the experience gathered by the animal in the course of its life, as we have described this in our preceding chapter. From its first day onwards the animal has to deal with all kinds of objects and situations. In this way it acquires experience about their properties, about the distances and directions of particular places, about the hardness and impenetrability of certain objects, about the results of its own intervention in situations. The young chicken learns by experience that if it turns round that wall it will reach the garden ; the young monkey that if he draws a branch towards him, the leaves and fruit on that branch will follow the- movement. According to its intellectual endowment, it learns bv this more or less to understand the relation between different events. Then the circumstances, or the caprice of the experimenter, place the animal in a situation not encountered before. None of the higher animals of a certain age comes quite unprepared into such a situation ; each brings with it some experience which may now be a help to it. If, as we supposed, this new situation is not quite identical with those which the animal has met with before, there will yet be some similarity between the new situation and the former ones. What will happen now depends on the ability of the animal to see the agreement between the essential elements of both situations and to apply its former experience of those elements to the new situation. The outcome of this, then, may be different. First it may be that the animal directly sees through the new situation, and with the help of its former experience immediately finds the right solution of the problem which confronts it. The chicken sees the new opening in the wall and goes through it into the garden; the monkey draws in the rope on which the experi.:. menter has fixed a piece of food. We then say that the animal has shown a "primary solution" of the problem, although in the strict sense of the word the term "primary" is not quite adequate, as the solution in such cases is the result of experience gathered in former more or less similar situations. If there is no such primary solution, two things may happen. First the animal may wholly fail to find the solution and after some futile attempts give up the problem before which

II6

ANIMAL

PSYCHOLOGY

it is placed. But it is also possible that the animal begins to try different means and so by purposive trials sooner or later comes to the desired solution, as we have already described in our former chapter. This, then, may be called a "secondary solution." It will be clear that the difference between a primary and a secondary solution is only relative, and not so essential as has been believed by authors who, like Kohler, reserve the term "insight" to those primary solutions, especially if these appear suddenly and unexpectedly. There may be several reasons for such a sudden solution, reasons that have nothing to do with intelligence as such. The animal, for instance, may suddenly and accidentally perceive something it did not perceive before, or the animal may suddenly for some reason be brought to hasten its action. But even if the understanding does suddenly arise, there is no reason to attach to this understanding an essentially different character. Primary and secondary solutions show only quantitative differences ; the now primary solution owes. its primarity to previous secondary ones. There may be a primary and a secondary insight, but both are insight. 1 Primary and secondary solutions are, indeed, not always clearly distinguishable ; before an animal shows a so-called primary solution it may have performed small actions that smooth the path for the final solution, and the decision, whether something is done primarily or secondarily may sometimes be difficult. But the result is the same : whether by a primary or by a secondary solution, the animal executes the actions necessary to reach its end. And the shorter its trials on the next occasion when it is placed in a similar situation, the better was its understanding in the first case. The question may now be asked, what kind of mutually related elements of a situation may be grasped in this way by the animals, and what forms of concrete understanding do we therefore meet in them. The answer must be that these elements are of a threefold nature. First, the animal may grasp the relation between spatial elements in its world, 1 Some authors speak of "insight-learning" if this embraces more than a mechanical stamping in of some movements. This, however, is wrong : insight may be the result of learning, but it is not a mode of learning itself.

ANIMAL

PSYCHOLOGY

117

the relation' of a "next-to-each-other" or "before-and-behindeach-other" of elements. Then, the animal may grasp the relation between temporal elements, the relation of a "beforeand-after-each-other" of elements. And, thirdly, it may grasp causal relations, the relation of "cause-and-effect" between phenomena. We will now demonstrate these three types of concrete understanding by some examples. Let us begin with the spatial relations. Here we may distinguish between two different types of situation, according as such relations are either existing independently of, or arise only during, the movements of an animal and in relation to these. For convenience sake we may distinguish them as "static" and "kinetic" relations of space. As an example of the first we may quote the experiments made by Yerkes in his so-called "multiple-choice-apparatus." This apparatus consists of a number of similar boxes, placed next to each other in a row or a sector of a circle. First Yerkes used twelve such boxes, later nine,· but this number, of course, is immaterial. In each experiment a number of the boxes were closed and thus put out of use, while the remaining ones were left open. The number of the open boxes changed with each experiment, and so did their place in the whole apparatus ; thus one time a number of them at the right side of the row were left open, another time a different number at the left side, a third time again another in the middle, and so on. The task for the animal, now, was to choose one box among those left open, which was characterized by its relative position among the open boxes, as for instance the first one on the left, or the second on the right, or the middle one in an uneven number, etc. As the number and the place of the open boxes were changed after each experiment it was not possible for the animal to direct itself towards some more simple absolute characteristics, as for instance for distance of the box from the edge of the apparatus, or the place of the box in the room, or such like. In order to make a comparative test between a number of different animals with this apparatus, Yerkes had devised ten different tasks with an increasing degree of difficulty.

II8

ANIMAL

PSYCHOLOGY

The easiest task was, of course, to choose the first box on the right or the left. A more difficult task was to choose the second box in the row, or the middle one. Still more difficult for the animal was to choose alternatively the first box on the right and the second on the left, as herein also another factor, viz. that of time, was involved. It soon became evident that all these more complicated tasks were too difficult for the animals ; only with the most simple ones were good results obtained. Curiously enough, relatively the best results were obtained, not with monkeys or even anthropoids, but with birds. A Macacus cynomolgus of Yerkes learned to choose the first on the left in 150 experiments, and to choose the second on the right in 1 ,o8o experiments, but as he still made errors when new combinations of boxes were opened, it appeared that he had learned rather to choose the correct box in a special number of combinations than to choose according to the principle required. A Rhesus monkey learned somewhat more quickly, and, in 480 experiments learned to choose alternatively the first on the left and the second on the right. Trying to ·teach him to choose the middle box in a varying uneven number of boxes met with no success in 320 experiments. With anthropoids the results were no better. An Orang did not get farther than choosing the first on the left, while of four chimpanzees only one was able to learn to choose alternatively the outermost boxes, and none to choose the middle box. Somewhat better were the results with lower animals and birds. Two swine of Yerkes and Coburn learnt to choose the first box on the right and the second on the left, and even alternatively the first on the right and the first on the left, but to choose the middle box was again too difficult. Two crows learned no more than to choose the first on the right ; but better results were obtained by Sadovinkova with songbirds, which rather quickly learnt the easier problems, and of which a siskin even learnt to choose the middle box in different combinations. Later the author tried whether better results would be obtained with monkeys if the boxes were arranged, not in a horizontal but in a vertical row, but his results were not much better than those of Yerkes.

ANIMAL

PSYCHOLOGY

119

It appears thus that learning such static relations of space is rather difficult for the animal. This is not surprising. In nature an animal is never placed before such a task, so that the endowment and probably even the attention to the solution of such problems is lacking in them. Better results, however, are obtained if spatial relations are to be grasped while the animal is in movement, a task more in harmony with its natural needs. Grasping the relation between spatial elements while in movements is necessary when an animal is forced to make a detour, a round-about way, if the direct road to its goal is blocked, or has occasion to take a short cut, if the possibility arises to shorten the usual way by the absence of some customary obstacle. It will be clear that these are problems with which an animal is often confronted in its natural life, where the direct way to the nest, or to some prey perceived, is not always possible, and the animal must try to reach it by a longer indirect way. It may be thought that, perhaps, to make such a detour offers scarcely any difficulty to an animal, and that any animal which is able to obtain a total impression of a certain area is also able to make the necessary detours in this area. That such is not the case was shown by the author's experiments with Octopus. When the animals were lying in a corner of the aquarium he put a short distance away a small wirenetting, and behind this a live crab, the favourite food of the octopus. Although it appeared that the animals clearly distinguished the edges of the network and so quite easily could have reached the crab by creeping round it or stretching out one of their long arms, none of them did so, but all tried in vain to take the direct way to the prey through the wire netting. Nor were they able to catch a live crab out of a tumbler set in the aquarium, but continued to bump against the glass instead of putting an arm into it. It must be mentioned that Buytendyk, when repeating these experiments with a number of octopus, found that one of them, after many failures to reach the crab in the direct way, suddenly crept round the wire netting and seized the crab. This shows how the endowment for the solution of this problem may vary

120

ANIMAL

PSYCHOLOGY

in animals of the same species ; probably in this case the task lies at the limit of their ability. But higher animals as a rule are able to make detours, be it directly, in a primary solution of the problem, or secondarily, after a longer or shorter process of trying. When surveying the literature on this problem we see that there are four different types of detours with increasing degree of difficulty. To the first type belong such detours in which both the goal and the way to it can be seen from the starting point, and in which there is then no need for the animal to turn its back to the goal during its march towards it and ·so lose it from sight for some moments. The second type is that in which at the starting-point the goal is visible and the road to it can be seen, but in which during part of the march the animal is forced to turn its back to the goal. The third type is such that the goal is visible from the starting-point, but the path cannot be seen and must be guessed more or less from former experience. And a fourth degree of difficulty is given, if the path not only cannot be seen from the starting-point, but even the goal is not visible from there, so that the animal must know from former experience that the goal (food or such like) is to be found within the apparatus. We willshow by some examples how animals deal with the four types of detours. A detour of the first type we have found already in the path the octopus has to take in order to reach the crab that it saw before it, and which could be kept in view while going to it. Other examples can easily be imagined: the dog that sees his master standing outside the garden runs first to the open gate and from there to his master. To quote another example with invertebrates : Drzewina made use of the positive phototaxis of a crab to make the animals move from the darker part of an aquarium into a part illuminated by a lighted candle, in doing which they had to pass through an opening at one side of the glass partition which divided ,the aquarium into two. The first day four animals found the right way in I, 5, I 5 and 35 minutes respectively after several bumps against the glass-plate. By the third day, however, they went directly through the opening, in from ten seconds

ANIMAL

PSYCHOLOGY

I 2I

to two minutes. In this case the burning candle or the illuminated part of the aquarium could be kept constantly in view. With various fishes similar detour experiments have been made with like results. That the making of a detour is made difficult to some animals if the goal is lost sight of for a time was clearly demonstrated by Fischel in experiments with lizards. If, when in making their detour, these animals had to pass round a piece of cardboard, so that the food they were trying to reach was invisible for some moments, most of them stopped there, or forgot the food, and only went to it if later they saw the food again by accident. How variously different animals may behave in such circumstances was shown by Kohler in his experiments with dogs and chickens. These animals were brought into a cul-de-sac 2m. long and I m. broad, and closed with wire netting, behind which food was laid. A dog showed a primary solution of this problem, turned round, and went without stopping out of the cui and round a fenced-in piece of ground straight to the food. Chickens, on the other hand, were not able to do this, even if a shorter detour was required than in the case of the dog, but kept running to and fro before the network. Only gradually did some of them succeed in finding the way to the food. If Kohler had carried on his experiments with his chicks there would certainly have come a moment in which these too would have been able to make the proper detour, after a longer or shorter period of trying. This shows that a problem, which can be solved primarily by one animal, with another requires a longer or shorter time of trying, and is solved only secondarily. That even higher animals may find it difficult to make a detour if having first to move away from the goal, was shown by the author with monkeys and lemurs. These animals had to make two detours. The first one was, when sitting on the top of a large bird-cage, to climb down and enter the cage in order to reach the food the experimenter was offering them there. The other one was, when sitting on a branch in that cage, to get down from it to the floor of the cage and to go out of it and to climb up outside it in order to reach the food

122

ANIMAL

PSYCHOLOGY

the experimenter was offering them above the cage. Monkeys found no difficulty in solving these two problems. Lemurs, on the other hand, had no difficulty with the first problem but did with the second. They always tried to snatch the food through the bars of the cage and only reached it if, after having given up trying and having left the cage, they happened to see the food from outside the cage and then went to it. After repeating this experiment several times they finally learned to make the right detour directly, which, in this case, was learned as a secondary fusion of two originally separate actions. The cause of this unequal difficulty in the two problems was obvious : in the second problem the animals had to turn their backs to the food while descending from their perch. Other animals, however, showed better performances when they had to make even more difficult detours. The cats of Adams, the rats of Maier, the dogs of Hobhouse made complicated detours to a point, where they saw food lying or their master standing, along paths which were not visible from the starting point and while during their walk to the goal they had repeatedly to lose sight of this goal. Sarris got good results with dogs when he ordered them to go to a point known to them but not visible from the starting place, as for instance their basket, even if they had to cover a complicated path in order to reach it. That tree-anim~ls, like monkeys and beech martens, find no difficulty in making perpendicular detours, too difficult for ground-animals like dogs and polecats, might be expected and has been confirmed by experiments of Muller and others. In close affinity to the making of detours is that of taking short cuts. Here, too, the relation of spatial elements must be grasped. Animals which have a nest, or a hole, generally know the spatial relations in the neighbourhood of this place and know how to return to it by the shortest way from every point in the surroundings, without being obliged to follow the frequently complicated path they took when leaving the nest. This is even the case with animals like ants and wasps. Hunting animals, like dogs and wolves, are often seen to cut off the path to the prey and so catch it at a point not yet

ANIMAL

PSYCHOLOGY

123

reached by the prey at the moment when the short cut was made. In experiments in which rats were trained to follow a complicated path in an apparatus, they often showed themselves able to take a short cut if, after having leanied the way, occasion was given to them to do so. We mentioned just now experiments made by Sarris with dogs that had to find their way to a point which they could not see from the starting-point, but which they knew from former occasions. This fourth degree of difficulty in making detours is now being specially studied in a type of apparatus called a maze. These mazes are too well known to require detailed description ; and the results of maze experiments are too numerous to make it possible to give a full account of them here. We may briefly define a maze as an apparatus consisting of a complicated system of alleys, some coming to a dead end and some leading to a goal in the centre of the apparatus, where some reward, mostly in the form of food, is to be found. The T-tube used by Yerkes with the earthworm may be regarded as the simplest type of such a maze. To follow the right path in a maze it is thus necessary to make a complicated detour, a detour so complicated that it is impossible for the animal to arrive at a primary solution-the more so as it has first to learn that there is in the apparatus a goal worth reaching (because food is to be found there), while, further, the path to this goal can never be seen in its entirety. White rats in particular have given evidence of being able to learn very intricate types of mazes. What is of more interest to us here is that the animal in learning the path through the maze has often proved that this involves more than a mere blind and mechanical following of a particular route. Often rats have shown themselves to possess some image of the road to be followed, and of the relations of a certain point in the maze to the position of the goal. For instance, they often seemed to have a feeling for the direction of the goal and a desire to move in that direction : they thus made more errors by entering blind alleys pointing in the direction of the goal than by entering those pointing in the opposite direction. If at some place they had to turn to the right or the left,

124

ANIMAL

PSYCH01.0GY

they frequently made errors by anticipating these turnings and thus entering a preceding alley. Occasionally they also showed some understanding of the distance they were from the goal at some point of the maze. As we mentioned above, they sometimes manifested the capacity to take short cuts if they got the occasion to do so. At times they also appeared to have grasped the relation between parts of the apparatus, as was the case in an experiment of Hsiao, in which rats understood that if one door to the goal was closed, all alleys leading to this door had consequently to be avoided, and that they would have to make a detour to another door that would lead to the goal. All this demonstrates once again the capacity of certain animals to grasp spatial relations, even in such unfavourable conditions as are involved when they are forced to run through a maze. Experiments on the capacity of animals to grasp temporal relations have revealed a much poorer ability than in the case of spatial relations. This need not excite wonder. Time plays a much less important part in animal life, and for this reason their attention is directed much less towards time and temporal elements. Here again we may distinguish between a "static" and a "kinetic" way of grasping such relations. Like Yerkes, who, with his multiple-choice-apparatus, studied the capacity of the animals to choose a box which was distinguished by its relative place in a row, Atkins and Dashiell investigated the question whether rats were able to learn to choose a door which was distinguished by its place in a temporal sequence. This was done in the following way. Three out of four doors in an apparatus were illuminated every time for a moment in varying order ; in one experiment the doors Nos. 4, 2, and 3 ; in the next the doors Nos. 2, 3, and I, and so on. Then three out of twelve rats had to learn · to enter the door that was illuminated first, six others to enter the door illuminated second, and three again to enter the door illuminated last. The result of these experiments was quite negative. In about 400 experiments no rat had learned this task, but they all formed position habits, i.e. they always chose the same door, or chose the doors in a

ANIMAL

PSYCHOLOGY

I25

fixed order. The results, therefore, showed an ability inferior to those revealed in Yerkes' experiments. Somewhat better were the results if animals had to do something in relation to a sequence of time. That is to say, if they had to do first one thing, then another thing, and then again some other thing. The simplest example of this is alternation, when an animal has to do something, then another thing, then the first thing again, and so on. Thus Carr placed rats in an apparatus which they had to leave alternately through a door on the right and one on the left side, in order to reach a place where food was to be found. His eight animals learned this with great difficulty ; it was only after 400 experiments that the choice made was right in 85% cases. When, afterwards, Carr lengthened the intervals between two experiments from I6 seconds to one minute, the results became much poorer still. Another type of alternation was studied by Hunter. In an apparatus of T -form rats had to learn to choose alternately the right and the left exit in order to reach food. This was learned rather quickly, but when the animals had to learn the double alternation (l-l-r-r-l-l-r-r-l-1) in soo-6oo experiments none of the rats could master this. Even the simple combination 1-1-r-r could not be mastered by any of them in zoo experiments. Such a double alternation was apparently too difficult for them. This was also shown by Hunter in his so-called "temporal maze." This apparatus consists of a central alley, at the end of which the animals can turn either to the right or to the left and so again come back through a side-alley to the starting point, from where they can again follow the central alley, and so on. Hunter, in this maze, tried to teach rats to run the double alternation 1-1-r-r-1-l-r-r-l-l. This again was too difficult for them. Of as many as seven other rats which had to learn the simple alternation l-r-l-r-l-r-1-r-l-r, only one succeeded to a certain degree in 59 experiments. It must be admitted, however, that other animals proved to be able to do more in such alternation experiments. Four raccoons of Hunter could not learn a fourfold double alternation (r-r-1-1-r-:-r-l-l), but were able more or less to learn the

126

AN I MAL

P SY CH 0 L 0 GY

two-fold double alternation r-r-l-l. Two young Rhesusmonkeys of Gellerman learned the double alternation r-r-l-l in 8o and 315 experiments respectively, and similar results were obtained by Karn with cats. And when Karn and Patton taught this double alternation to cats, and then brought them to a similar temporal maze of different dimensions, the cats were able to perform their task in this new maze as well, which shows that their performance was based on an understanding of temporal relations, and not on some simple physiological phenomenon, as perhaps might be supposed. In another way the author trained animals to perform similar rhythmic actions. When an inverted tin was placed on a table beside his cage, a monkey was trained to overturn this the first two times the tin was placed there, and to omit this the third time. (The first two times food was put under the tin and the third time not.) The animal thus was trained on a "yes-yes-no"-rhythm. This was learned in about 650 experiments, but when the interval between the experin'.ents was lengthened from 10 to zo seconds, the animal made 3 errors in 10 experiments, and when the interval was extended to 30 seconds it made 9 errors in 10 experiments. Doubling the interval, therefore, partly destroyed the acquired rhythm of action ; trebling it destroyed the rhythm almost entirely. If, on the other hand, an animal was trained on the more simple rhythm "yes-no," a doubling of the interval had no influence, while with an interval of 30 seconds 6 errors were made in a series of 10 experiments, and with an interval of 45 seconds, 10 errors. A simpler rhythm apparently can admit of a greater change in. the interval than a more complicated one. Another monkey could even be trained to a yet more complicated rhythm, viz., that of "yes-yes-yes-no." Such rhythms of action were also learned in other ways. Rats of Ellis in a choice-apparatus learned to jump down three times to the right side and then once to the left. Koehler and his collaborators trained pigeons to pick up only a certain number out of a greater number of grains and to leave the others alone. First the animals learned to pick up 2 grains out of rows of from 3 to 6 ; finally, one bird came so far as to be able to pick up 6 grains out of a row of 7 to 13 and·

ANIMAL

PSYCHOLOGY

127

let the others alone. Although not all Koehler's results can perhaps be explained by admitting such rhythms of action, yet most of them do. Anyhow, it is certain that they cannot be ascribed to any counting on the part of the animals, as has been supposed by some authors. For real counting it would be necessary to possess the concepts of number, and no animal is able to form such concepts. Yet all the results mentioned in this section point to a more or less successful grasping of temporal relations. For, if an animal knows the right path to follow in a temporal maze, or turns a tin in a required sequence, or picks up some grains and leaves the others alone, it must have understood the "after-each-other" of certain actions it has to perform, i.e., it must have understood that the action a must be followed by the action b, and that after the action p all action must stop. This understanding of temporal relations may not go very far in animals, yet it is certain that in some of them it exists to a certain degree. More interesting, however, are the results obtained with animals regarding their understanding of causal relations. The first systematic experiments 1n this grasping of causal relations by animals were made by Thorndike with his problem-boxes, already mentioned in the preceding chapter. His dogs and cats had to open boxes with an intricate opening mechanism, too complicated to be understood by the animals at first sight. That the results were rather poor, and that the conclusions drawn by Thorndike from his results as to the understanding of the animals were not very favourable for the latter was, therefore, rather the fault of Thorndike than of his animals, the more so indeed inasmuch as Thorndike was somewhat prejudiced against them and even tried to explain away physiologically facts which actually did speak for some understanding on the part of his animals. Even in monkeys Thorndike could find no trace of real understanding. Later, however, other experimenters, working with other animals, obtained better results than Thorndike, even with similar problem-boxes. Raccoons of the McDougalls very soon learned to open a box closed with a vertical latch,

128

ANIMAL

PSYCHOLOGY

either by pushing the latch away with a paw or with their nose. Then a second latch was.fixed which held the first one, so that to open the box first the second latch and then the first latch had to be pushed aside. After this was learned a third and a fourth latch were fitted, and so on, till finally the box was closed by a system of 24 horizontal and vertical interlocking latches, distributed over the different walls of the box. One of the raccoons learned very quickly every time to open the new latch, and to push away all the latches in the proper order. With every new latch that was fitted, he understood almost at once the additional hindrance it formed to the opening of the other latches. He was also able to discover what was wrong if the latch was not sufficiently moved. His dexterity in opening the latches gradually improved, showing that during the work his understanding was increased by new experience. In a similar way rats of the McDougalls learned to open boxes with 14 latches, also to be undone in a certain order. To study the understanding by animals of causal relations there are, however, better methods than the use of problemboxes with their intricate closing-mechanisms, namely, the study of their handling of different objects, as this was first started by Hobhouse. We may divide these experiments into three groups : those of the clearing away of obstacles, those of the utilising the possibilities of moving objects, and those of the use of tools. Of each of these groups we will now give some examples. With regard to the clearing away of obstacles, we may distinguish between obstacles which hamper the animals themselves in their movements, and those which stand in their way when trying to reach some goal. In both cases it is necessary that the animal should first recognize the ob!ltacle as such, and then understand what to do in order to overcome it. On the first type some early experiments were carried out by Hobhouse. When the chain of a Rhesus monkey was laid round a box so that the monkey could not reach food laid on the floor, the animal immediately pushed the box away and so showed he understood that it was the box which hampered his movements, and that he had to push it away in order to

ANIMAL

PSYCHOLOGY

I29

be able to reach the food. Another time Hobhouse laid the chain round a heavy table. Then, after an interval, the animal managed to help himself by going around the other side of the table, in other words, by making a detour. In similar circumstances a pig-tailed macaque (Nemestrinus nemestrinus) of McDougall knew how to help himself in the same way, and if his chain was looped round a stake, he tried, sometimes with success, to lift the chain over the stake. A raccoon of McDougall in such a case even drew the stake out of the ground and then ran for the food. A good solution was presented by a rat of Schaff and Sgonina, which, when a rope held him from his goal, took the rope between his paws and gnawed it through. All these results show an understanding of how to deal with obstacles. The same was often the case with objects which stood in the way of the animals when trying to reach a goal. Here again the first experiments were made by Hobhouse. A stool was placed before the door of a dog's kennel. Mter a slight hesitation the dog clawed the stool away and entered the kennel. Kohler repeated the same experiment with chimpanzees and put a box inside the cage of the animals just in front of the place where they had to reach for a banana laid outside the cage. His results, curiously enough, were rather disappointing. Not one of his six animals directly manifested the proper insight, and it was only after some futile endeavours to reach the fruit in another way that they gradually found the right solution. This· is the more remarkable, inasmuch as the present author, in comparative experiments confronting different animals with a similar situation, obtained much better results from many of them. A pig-tailed macaque, a Cebus monkey, a young coati and a raccoon, all directly took the box and pushed or threw it away ; other monkeys and Lemurs did the same sooner or later after some had first tried to reach to food in another way, and only a squirrel did not get further than making a direct attack on the bars of the cage. This proves that it would be wrong to believe that the intelligence of animals correlates highly with their place in the animal system, and it warns us especially against the belief that in all such experiments chimpanzees must 9

ANIMAL

PSYCHOLOGY

give the best results, just because morphologically they most resemble man and are therefore placed at the top of the system. We may also mention here that a raven of Hertz drew a stone out of an opening if food was laid behind it. That many animals, even birds, are able to turn over a box or tin if they have seen food placed under it, may also be mentioned as a case of clearing away obstacles in order to reach a goal. A great number of experiments with many animals have been performed to see whether they were able to use the possibility of some object being moved in order to reach a goal, in most cases, again, food. In this problem, too, experience in their daily life may help them. Every animal which has the power to set some object moving has the opportunity to observe that everything fixed on this object moves in the same direction. As we remarked already, a monkey which, in play, draws the branch of a tree towards him may notice that the leaves of the branch follow the movement, and the same is the case with the dog that pushes away his foodtrough and may notice that the food it contains is pushed away with it. The question, whether an animal understands that in order to get at some object it must first move another object to which the former one is fixed, is therefore not an unreasonable one. Some authors have regarded such actions as examples of the using of tools. 'Ve believe it better to reserve this name for cases in which an animal itself actively brings an object into contact with the desired food and thereupon uses it for obtaining the food. Cases such as we will now describe might, then, be called examples of a "spurious use of tools." We find them to be of three types : that of drawing in some object, that of pushing some object away and that of rotating some object, all in order to reach the otherwise unattainable food. In the first case, then, the animal has to draw an object towards it in order to obtain something attached to it. To study this problem food may be fastened to a cord, the one end of which lies within the reach of the animal, whereas the food itself lies beyond its reach. The cord with the food may also be hung on a branch on which the animal is sitting, so

ANIMAL

PSYCHOLOGY

131

that it must pull up the cord in order to get the food. Instead of a cord a thin lath may be used, with food lying on one end. The task is somewhat more difficult for the animal if a little rake be used, which is laid behind the food, so that the animal must draw in the rake and so rake up the food. Apes, as well as monkeys and lemurs, generally find no difficulty in solving this problem and immediately draw in the cord or the lath. With other animals, however, the results of such experiments were rather different, and as a rule less good than with monkeys. It must be kept in mind, however, that the lower animals are less well equipped for such work than monkeys with their supple hands, and have to do it with their feet or teeth. Yet most of these animals, like the raccoons and coatis, or the rats and guinea-pigs of different experimenters, understood fairly well how to help themselves. The same was found in experiments with a fox, while with dogs and cats varying results were obtained. Some of these animals failed entirely, while others learned to do it after various trials, or even pulled in the cord at once. This drawing in of a cord is rendered more difficult if the cord be laid round a pole so that, when the animal draws the cord in, the food first moves away, or if the cord be laid on the floor in loops, so that drawing in the cord at first does not make the food approach. Yet this did not discourage the cats of Adams and the monkeys of Guillaume and Meyerson from continuing to pull. Drawing in a cord with food has even been performed by birds (crows, parrots, different singing-birds) and may be observed in the garden when titmice sit on a tree and draw up a string with a peanut some bird-friend has hung there for their food. The raking in of food, if the rake is laid behind it, is not generally a very difficult task for an animal able to draw in anything at all. Even squirrels can do so. This, however, lasts only as long as the food moves with the movement of the rake. If it happens that the food slips from the rake, or if the rake is laid, not behind, but beside the food, no animal, as a rule, with the exception of anthropoid apes, is able, on his own initiative, to get the rake behind the food again. At best we see only some vague movements with the rake in the

132

ANIMAL

PSYCHOLOGY

direction of the food. In such cases more is required from: the animal, viz. the capacity to use an object as a real tool. We shall return to this question presently. Much more difficult is it for an animal to push something away, in order to get it after making a short detour. The difficulty here is, of course, that in pushing, the food moves away in a direction opposite to that wanted by the animal. To study this, food is laid inside a tube with a stick partly projecting from it, between the food and the animal, so that the animal must push the food through the tube in order to get it. Even for monkeys this problem is mostly too difficult ; only some chimpanzees learned to do it, and some of them were even able to push the food out of the tube with a stick laid beside it, which implied the understanding of the use of a tool. , A third type utilizing the possibilities of moving objects lies in rotation. The method employed to study whether animals understand how to make use of rotation is to fix a lath on a hinge, so that it can be rotated in a vertical or horizontal direction, and to lay food on its free end beyond the animal's reach. A rotating plate also may be used, on the farther end of which food is laid. The latter task is somewhat more difficult than the former, as the animal must learn not to try to draw the plate to it, as is its natural tendency, but to make it turn and to continue the movement till the food is within its reach. Monkeys and lemurs generally solve this problem primarily, as soon as the possibility of rotating the plate is noticed by them. So do raccoons, and coatis, while with cats the results once again varied greatly, and a squirrel belonging to the author did not get farther than biting the plate. The problem can be rendered still more difficult if a lath has to be rotated around a vertical axis, fixed in the middle of the lath. Then the food moves away in a direction opposite to that of the end of the lath which the animal has set in motion. Even the chimpanzees of Guillaume and Meyerson had great difficulty with this problem. Another type of the rotation problem was devised by Kohler. A stretched cord was fixed to a point outside of,

ANIMAL

PSYCHOLOGY

133

and in front of, the cage. Its free end was lying inside the cage in such a way that the cord subtended an acute angle with the bars of the cage. On the middle of the cord, beyond the reach of the animals, a piece of food was fastened. To obtain the food the animals had to rotate the cord round the fixed outer point by passing the free end of it through the bars of the cage till in so doing the food came to a point from where it could be reached from within the cage. Curiously enough, here again lower animals often showed a better understanding than Kohler's chimpanzees. Two of these latter got no farther than a futile pulling at the rope, while three others solved the problem only secondarily, and only one found the right solution almost immediately. On the other hand, two monkeys of Trendelenburg, a pig-tailed macaque and a cebus of the author, as also two macaques of Verlaine and Gallis, found the right solution almost at once, and the same was the case with two raccoons and a coati of the author. With his lemurs the solution was less certain. Two gibbons of Guillaume and Meyerson were unable to solve the problem at all, and an orang not every time, whereas some lower monkeys and four chimpanzees found it directly. All this again proves that it is erroneous to suppose that in such problems insight correlates with the animal's place in the system. Lower animals often arrive at better solutions than higher ones, and the understanding is rather a matter of individual, than of specific, endowment. The most striking cases of the understanding of causal relations by an animal are found, where it proves able to use a tool. What is a tool ? By a tool we understand an animate, or inanimate, object, not forming part of the body of its user, which temporarily is brought by an animal into its actions, in order to make possible, or facilitate, the attaining of some goal towards which it is striving. But we must at once add that the using of such a tool must not always be regarded as a proof of an explicit understanding of causal relations, as, curiously

ANIMAL PSYCHOLOGY 134 enough, the using of such tools sometimes occurs on a level on which there is not yet any question of such an understanding, viz., on the level of instinctive activity. The best-known example of such instinctive using of tools in animals is found with the weaver-ant, Oecophylla smaragdina. These ants build nests of leaves of trees by sticking them together. For a long time it was not known how they managed it, as they do not produce any viscous substance. About half a century ago, however, it was discovered that they perform this with the help of their larvre, which possess glands that produce a spinning substance able to be used for sticking the leaves together. In building or repairing their nests, the ants take the larvre in their mandibles and move them to and fro from one leaf to the other ; while doing so, the larva disgorges a sticky secretion, which soon dries and thus holds the leaves together. The same behaviour, by the way, was also observed later in other species of ants. These weaver-ants, thus, use their larvre in two ways as tools : first as a tool to produce the sticking material, and then as a tool to bring it to the leaves. And even a third way of using a tool may be observed in them. If the leaves to be united are too far apart to permit an ant standing on one leaf to draw the other leaf towards it, a second ant comes and takes the first one in her mandibles and uses her mate as a tool to draw in the leaf. If necessary, chains of five or six ants may be formed in this way, the first ant standing on the leaf, while the last draws the other leaf in. That we have here a case of the using of tools will not be denied. It was, however, not always clear how this behaviour had to be interpreted. Some observers were inclined to see in it an understanding of the result of their own actions by the ants, in other words, an intelligent using of tools. We cannot follow them in this interpretation. First of all, this behaviour is found with animals which show a highlydeveloped instinctive life but very little proof of real understanding. Again, the behaviour is not executed only by some very clever individuals but by all the members of the species, and is certainly innate, as is shown by the fact that even ants, removed from the nest when they came out of the pupa,

ANIMAL

PSYCHOLOGY

135

showed this behaviour just like those living in normal conditions. It seems that the larvre even assist in the weaving, in so far as they move their heads to and fro between the leaves. Still again, the larvre themselves show a morphological adaptation to the behaviour of the perfect ants in their possession of spinning-glands, which are more developed in them than in the larvre of related species, although they themselves do not spin a cocoon. Here once more we have an adaptation of the body of an animal to instinctive activity, of which we spoke in our second chapter, the more remarkable now, as here the instinctive activity of a perfect insect finds a correlate in a morphological structure of the larva. That a larva should have a morphological structure adapted to an intelligent action of an individual congener is, of course, unacceptable : that in these ants the using of their larvre as a tool happens on the level of instinct, is therefore beyond doubt. There are some more examples of such an instinctive using of tools by animals, although perhaps not so striking as that of the weaver-ants. It has been observed that some Ammophilas of different species used a small pebble to stamp the sand down the opening of their nest. The different observers of this fact disagree as to whether this was a performance of one special animal or was done by all animals of its species. If the latter be the case we may rightly ascribe the action to instinct. If only one, or several animals are found to do so, it would seem to be an example of individual intelligence. But then, perhaps, it may be ascribed to an old instinct of the species, yet living in other species, and in this species partly extinct but suddenly revived in one, or some, individuals. Other cases of an instinctive using of tools are the throwing up of sand by the ant-lion, if the prey tries to escape out of the pit, and the blowing up of drops of water by the fish Toxotes jaculator, which captures insects flying near the surface by shooting drops of water at them. And an interesting case was discovered a short time ago by Lack with the finch Camarhynchus pallidus of the Galapagos Islands. This bird searches for insects living in holes in the stems of trees, and to this end pokes into the hole with a

ANIMAL

PSYCHOLOGY

little twig or a spine of a cactus and then throws it away as soon as the insect comes out. Since, according to Lack, all individuals of this species do so, there is again no reason to estimate thi~ behaviour higher than an inborn instinctive activity. But what interests us here are not the instances of an instinctive using of tools, but such in which the using of tools is built up upon an understanding of causal relations. This intelligent using of tools may take place in two forms. Either the animal may use a stick, or something equivalent, as a means to lengthen its arms, or it may use a box or some similar object as a means to lengthen its legs, both in order to reach food laid down or suspended beyond its reach.1 Let us begin with the stick. The most common way of employing a stick is that of using it as a rake to rake food lying beyond the reach of the animal. In such a case the animal must understand that when it draws towards itself a stick it has laid behind a piece of food, it causes the food to follow the movement of the stick. We have already discussed as a precursor of this kind of free using of a tool the cases in which the animal draws in a stick the experimenter has laid behind the food. That much more is required from the animal, if it has itself to lay the stick behind the food, is proved by the fact that but few animals are able to do so. Whereas it does not generally occasion much difficulty to anthropoid apes to use a stick or some object with the same functional valence, e.g. a cloth, a bundle of straw, a water-bowl, a piece of iron-wire, and so on, as a rake, lower monkeys are usually not able to do so of themselves, even if they are able to draw in a stick laid by the experimenter behind the food. Yet there are exceptions to this rule : sometimes anthropoid apes fail, while by way of 1

A few instances are known of the using of a tool by an animal in play.

A chimpanzee belonging to Mobius took his water-bowl and struck back

a nail driven into the wall of his cage, and a monkey belonging to Yerkes drove a nail into a board with a hmer. Chimpanzees belonging to Kohler used a stick in their playful fighting (but never in real fight I) and in play pricked chickens with sticks after first having lured them towards their cage. These, however, are exceptions ; generally the using of a stick as a tool is prompted by the feeding-instinct.

ANIMAL

PSYCHOLOGY

1 37

exception lower monkeys show the necessary insight. Here again the individual endowment of the animal is worth more than its place in the animal system. While gorillas and gibbons were nearly always lacking in the necessary understanding, the lower monkeys, especially some Cebus monkeys, were able to do so. A Cebus fatuellus of Kluver even used a living animal for this purpose. When a living rat on a cord was put into his cage, he threw the rat out of the cage in the direction of the desired food and drew the animal with the food in by the cord, sometimes waiting till the rat while walking round found itself behind the food I In one exceptional case also it was reported that an elephant took a branch with his trunk and used it as a rake to rake in bread lying beyond his reach. There are still other ways in which a stick may be used. Monkeys were sometimes seen to use it as a striking-tool, as a means to knock down food hanging above their reach. Sometimes the stick is thrown up towards such hanging food, although in this case it is not quite clear if this is done simply to hit the food or rather as an expression of the animal's desire for the hanging food. That sometimes the stick may be used to push food out of a tube has been mentioned above. Occasionally, apes used a little stick or twig as a spoon which they put into the water and then licked off . . Kohler's chimpanzees used the stick to dig with, or as a lever to turn over stones. Chimpanzees and orangs, and even some lower monkeys, are reported to have used sticks as climbing-poles, which they placed under a hanging piece of food and along which they then quickly climbed up and grasped the food just as they fell down, stick and all. The stick may sometimes be used also as an indirect tool. That is to say, if a piece of food lies outside the cage, and beside it, also beyond their reach, lies a stick long enough to rake the food in with, while a shorter stick lies within their reach, chimpanzees may use the shorter stick to rake the long one in, and then use the longer one to rake in the food. A Cebus of Cope drew in a strap he had lost with the help of a poker, and then used the strap to rake food in. That this roundabout way: "Short stick-long stick-food," is a

ANIMAL

PSYCHOLOGY

method which the animals find difficult, is proved by the fact that even chimpanzees, who know how to use a single stick, are often unable to accomplish it. The using of a box as a means to lengthen the body is more interesting than using a stick, in so far as in this case animals often do not stop after having drawn one box under the suspended food, but are able if necessary to draw a second or even a third box towards the same place and to put them one upon the other so as to construct a kind of tower. Placing a box under suspended food and mounting upon it in order to be able to grasp it, has been observed of many anthropoid apes. Instead of a box, a chair or a stone, or a roll or iron~ wire, or a block of wood may be used ; chimpanzees of Kohler and Yerkes even tried to draw the experimenter under the food and to use him as a climbing-tool. Orangs and chimpanzees often showed themselves capable of this piling of boxes, although sometimes only after a period of trial and error. Some chimpanzees of Kohler and Bingham even got so far as to pile up four boxes. This has been the best performance known of animals. For technical reasons, but no less as a matter of insight, placing one box on another is far more difficult than simply drawing a single box under suspended food. The floor on which the animals walk, or lie, is the normal base on which they carry out their actions : the top of a box is a much less familiar plane, and it requires more insight to understand that this surface, too, may serve as a starting-point for placing a box than it is to understand that a box can be drawn or placed somewhere on the floor in order to be mounted. Added to this the building of a tower of two or more boxes is also technically rather difficult for the animals. Those built by Kohler's chimpanzees ·were often so shaky that the animals together with the boxes fell down when they tried to climb upon them. It was only owing to their finely developed sense of balance that they often succeeded in grasping the food before they tumbled down, together with the boxes. In lower monkeys the understanding of the use of a box in order to obtain suspended food is not generally very well developed. As an exceptional case among lower monkeys a

ANIMAL

PSYCHOLOGY

139

Cebus of the author may be mentioned. This animal, after the experimenter had set it gradually more difficult tasks, was finally able to build a tower of three boxes. Among the lower mammals a dog of Sarris may be mentioned, who, when given a box mounted on rollers and provided with cords, pulled the cords so as to bring the box under a suspended piece of meat, while a coati of the author learned to do the same after a long period of trial and error. These, however, are exceptions. In general we may say that this task is too difficult for mammals lower than monkeys. We may add that sometimes animals have been found capable of combining the use of a stick and a box, and so to integrate the two ways of using a tool into one action. Chimpanzees of Kohler sometimes piled up some boxes, placed a pole on them, and then climbed up the pole, or, standing on a box they had drawn under the food, used a stick to strike the food down. The same was done by the author's Cebus and a Cebus of Kluver. When Kohler once hung a stick high on the wall and placed a box in the neighbourhood, one of his chimpanzees drew the box under the stick mounted it, took the stick down from the wall and used it to rake in food. One animal, on the other hand, used a stick to draw in a box in order to mount it. In this piling up of boxes a higher degree of intelligence is shown than in simply using an object as a tool. Here, indeed, the animal not onlv uses a tool, but also constructs one. In other cases animals were known to improve their tools, if these were inadequate for their purpose. Thus an oblong box may be turned up if lying on its side and not high enough for the animal. When the author's Cebus built a tower of two little boxes and a tin, he once missed the cover of the tin, so that it was difficult for him to stand on the tin. He thereupon went into the sleeping cage, came back with the lid he had left there, placed it on the tin, and then mounted it. Some animals which wanted a stick but could not find one, were able to make a rake for themselves. A chimpanzee of Kohler to this end broke an iron bar out of a boot-scraper and used it as a rake, or broke off a branch from a tree for this purpose. The same animal unrolled a_sroll of iron-wire

ANIMAL

PSYCHOLOGY

and used it as a rake; others broke boards out of boxes, etc. A macaque of Verlaine broke a twig off a plant, or pulled a bristle out of a brush ; a Cebus of Kluver rolled pages of a newspaper between his hands and the floor, and so made a roll he could use. Objects used as rakes may also be improved on if they do not satisfy. When a bundle of straw was too soft to be used as a rake, one of Kohler's chimpanzees folded the bundle into two and so obtained a shorter but firmer rake. On the other hand, a chimpanzee of Guillaume and Meyerson showed that he knew how to lengthen his rake when he unfolded a foot-rule. And one of Kohler's chimpanzees, when two bamboo-rods were given to him both of which were too short to reach the food, understood how to put the smaller rod into the larger one and so to construct a rake of greater length. Now, it cannot be denied that all these performances require a certain imaginative faculty, a certain phantasy on the part of the animal. The animal must "see" the rake in the bars of the boot-scraper, or in the branches of the tree, while they still form part of that object ; it must "see" the higher box in the oblong box on the ground, and the shorter but more solid rake in the bundle of straw before folding it. Some animals also showed that they were able to see in their imagination objects, or parts of objects, wanted for the construction of their tools. One of Kohler's chimpanzees, when building a pile of boxes, went into the corridor to seek for a box that he knew was standing there and which he wanted for completing his tower, and the Cebus of the author went for the same purpose into the sleeping-cage and fetched the tin he had left there behind in play. Some imagination is often shown even in the mistakes the animals make in their performances. For instance, monkeys were often seen to lift a box up in the direction of the food, or to press it against a wall of the cage, as if expressing their understanding that if the box could only be made to remain there, they would be able to climb upon it and get the food. Even if the animal simply picks up a stick in order to use it as a rake, or goes to a box in order to place it under the food, some imagination must be involved in this action.

ANIMAL

PSYCHOLOGY

Similar imagination, it may be said by the way, is shown by an animal in its play. If a dog treats as a prey a ball or a piece of wood thrown away before his eyes by his master, if he runs after it and catches it, shaking and biting it, just as he would do with a living animal, although he undoubtedly knows very well that this object is not a real prey at all, and, again, if in his playful hunting he alternately runs after his mate and lets his mate run after him, if in a sham fight he pretends to bite his playmate but does not really do him any harm, there must certainly be imagination underlying this behaviour, by which he harmonizes, but distinguishes, the playful case and the serious one. And must there not be some imagination in the dog or the wolf when he cuts off the path to the prey he is pursuing, a "seeing" of the prey at a piace where it is not yet, but where it will be after some moments? These faint rays of imagination, shining on some of these higher intelligent performances of animals and their plays, lead us to ask the question, whether it is really true that the concrete understanding described in this chapter is the highest form of animal intelligence. May there not be something in the animal more like our own power of abstraction, like our thinking, like our more theoretical understanding ? Is there nothing in the behaviour of an animal to be regarded as a proof of ideation ? It will be our task in our next chapter to test the evidence of this as it is presented by the animals themselves.

CHAPTER

VI

THE PROBLEM OF ANIMAL IDEATION WE saw in our last chapter that the summit of animal performances in the field of practical understanding is the using, the improving and the constructing of tools. The question now, finally, to be considered, is whether there is or

ANIMAL

PSYCHOLOGY

is not in the animal something of a more abstract nature than this understanding of concrete causal relations, something more akin to the higher mental faculties of man. We have already credited the animal with some fantasy, which manifests itself in its play and sometimes enters into its intelligent performances. But does the animal possess free images, images free of what it perceives at some particular moment ? The dog may in his mind link the threatening whip with pain, and the particular spot in the field with the rabbit started there the day before. But can he, lying in his basket, ponder over the punishment he received from his master an hour ago, and remember the pain and the feelings he experienced during the punishment, or recall the walk with his master that morning and remember the pleasures he enjoyed then ? We may doubt it. If the monkey goes into his sleepingcage in order to fetch a box he wants for building his tower, the image of the box to be found there is associated with the perception of the unfinished building. If he "sees" the rake in a branch of a tree, he really does see the branch at that moment. There seems no reason to credit him with more. Some will perhaps allege that the animal ha:s free images during its dreams. That at least the higher animals dream seems very probable, although it will never be possible to prove this. The movements they make in their sleep, and the sounds they utter, resemble too much some characteristic movements and sounds of their waking life to be ascribed to accident or to casual reflexes of their body. Yet what we know about the dream-life of animals is so vague that it cannot give us any positive information respecting the images that pass through their mind at such moments and cannot even be regarded as a proof that any experiences really approximating to our dream-images actually occur in such supposed dreams. As proof of the existence of such imagery in the animals their dreams are of little value. That animals do not form concepts, even such concrete and vital concepts as "food" or "water," as "mate" or "enemy," may be regarded as indubitable. If they did so,if they possessed such concepts, they would certainly have a word to denote it, a sound specifically meaning that object. The animals, it is

143 known, have no words. All endeavours to train higher animals to use special sounds to denote objects in their world have failed. Animal language, if we may use the word "language" for the sounds an animal utters, is mostly no more than an expression of the animal's emotions and desires, uttered by the animal for itself alone, independent of the question whether another being hears them or not. Only rarely are these sounds intended to be heard by another living being as may be the case with animals living in a herd, or between mother and child, or exceptionally, as in the case of an animal like the dog, between animal and man. But even then the sounds uttered are inarticulate and innate, not articulated and learned in the course of life as is the case with the wcrds of man. And even then, their sounds do not denote objects, but affections and desires ; they are not indicative, but expressive. The animal has no words, and therefore has no concepts, and as it has neither words nor concepts it is not able to do anything with them, to combine them, to manipulate them ; in other words, to think. Even the cleverest ape does not think while building his tower ; he does not say to himself: "Let me put one more box on the others and then I shall be able to reach the food." He simply sees what he has to do and does it. Such seeing, however, is not thinking. Animals, thus, of themselves do not form concepts. Some experimenters have studied whether it might be possible to train animals to performances, for which some simple conceptions are required. Thus Hamilton placed his subjects (some normal and defective men and children, some monkeys, cats and a horse) in a space with four similar doors, one of which could be opened by pressing against it, whereas the others were closed. The position of the unlocked door varied irregularly from one trial to the other, with the proviso that the door which could be opened was never the same as the one that had been unlocked in the preceding trial. All that was expected of the subjects was, thus, that they should never try to open the door that had given exit the time before. Only the human subjects arrived at the understanding of this principle ; none of his animals (and a child of two years old) could discover it. They all tried the doors in an ANIMAL

PSYCHOLOGY

1 44

ANIMAL

PSYCHOLOGY

irregular sequence or, at the best, tried to open them in a succession one after the other. Revesz placed four boxes before some lower monkeys and put the food successively in the first, the second, the third, and the fourth box. None of his animals was able to discover this simple principle of changing the place of the reward. Robinson, Kluver, and others, studied whether monkeys were able to form the conception "other than." In some of Kluver's experiments, for instance, four boxes were used, one of which was marked with a yellow circle, while the other three were marked with a blue one. Food was always put in the box which was different from the others, in this case, therefore, the box with the yellow circle. After the animal had learned to open this box, four boxes were presented, one of which was marked with an orange circle, the three others with a violet one. The monkey, then, in 97·5% of the cases opened the box with the orange circle. In other experiments his monkey had learned to choose a box with a black square in contrast to three boxes with a white square. Then, when four boxes were presented, three of which were placed at a distance of Ioo em. and one at IIS em., this last one was chosen in 94% of the cases. But does all this prove that the monkey had formed the conception "other than," or "the only one of its kind" ? We may doubt it. An object which in one conspicuous character differs from others always strikes us ; so all that had been gained during the training was probably no more than to direct yet more the monkey's attention to the different element and to train him to choose this. That explicit conceptions like "other than" are required for this performance is not necessary ; the whole process again may take place on the perceptual level. This is confirmed by the fact that the monkey in Robinson's experiment only gradually learned to choose the different box. Had he discovered the principle, had he grasped that he had to open the box "other than" the others, from that moment he should have made no more errors. But now the different only gradually obtained the valence of being the indicator of the place of the food. In all such explanations we must keep to the principle formulated half a century ago by Morgan, namely, that we must not

ANIMAL

PSYCHOLOGY

1 45

ascribe an animal's action to some higher psychical process, as long as it is possible to describe it as the result of a process which stands lower in the scale of psychical development. If we relinquish this sane standpoint, all our explanations lose their power of conviction. The animal, as we saw, is not able to form abstractions. But on the other hand it must be admitted that sometimes animals have proved to be able to perform something which may be regarded as a precursor to such abstractions, be it again on a lower level, viz., the forming of perceptual generalizations. Animals that were trained to choose a box which was marked with a triangle, and to disregard another one marked with a circle, afterwards also chose this box if it was marked, not with the triangle used up till then, but with some other kind of triangle, say a right-angled triangle instead of an equilateral, or a triangle with the apex downwards or to the left or right side, instead of upwards, or a triangle of different size or painted in another colour than that which had been used during the training. In such a case we must admit that the animal has grasped the sensorial generalizatiqn of "something with three sides and three angles."1 To facilitate the forming of such generalizations it is advisable during the training period to change the triangle used from one trial to the other, as was done by Fields with rats. In one of his experiments, instead of the equilateral triangle, used up till then, a right-angle triangle was presented in eight different positions. His rats then chose it in 96% of the cases. Similar results were obtained by Buytendyk with a dog, and by Gellerman with chimpanzees. But in other circumstances, too, animals may give evidence of being able to form perceptual generalizations. If an ambulance-dog is trained to indicate "wounded" persons (i.e., lying) and to disregard "unhurt" persons (i.e., standing or walking), the dog must 1 One danger, however, must always be kept in mind in such experiments. Animals sometimes train themselves, not to choose the box with the positive mark, but to avoid that with the negative mark, it being to this, therefore, that their attention is directed. If, then, some change is made in the positive mark, the animals may fail to notice it as it does not enter their consciousness with sufficient clarity. Before drawing conclusions, control experiments should be made to escape this trap I

Io

q6

ANIMAL

PSYCHOLOGY

have been able to form the perceptual generalization "man in a lying position," quite apart from the other characteristics of all these men and the way they are lying and in contrast to all other forms of human attitude and movements. But in these cases, too, the generalization does not overstep the perceptual level. The animal discovers the point of similarity in a number of partly-different perceptions, and learns to react in a particular way to this common element. But this does not pass beyond the level of the perceptual and we have no right to credit the animal with real concepts here either. Does an animal recognize living beings or objects, known to it, on pictures ? The result obtained by the author with two monkeys were not very encouraging, although other experimenters seemed to obtain somewhat better results. But even then, if an animal reacts to a picture as before it had reacted to the object itself, must we admit that it really recognizes that object, i.e., that it understands that the picture stands in the place of the object represented ? Here, again, some doubt is permitted. If, for instance, a dog reacts in a hostile way to a picture of a cat or barks if he sees dogs running on a film, this, in our opinion, does not imply that he recognizes a cat, or his mates, in the pictures of them, but, on the contrary, believes he has to deal with real animals. If he had understood that it was only a picture, he probably would have kept quiet. The birds of Zeuxis in Pliny's story would never have pecked at the grapes in the painter's picture if they had understood that they were not real. The story may demonstrate that Zeuxis was a great realistic painter, but not that his birds recognized the grapes as paintings. I believe all such results testify rather to the poor vision of some animals, or at least to their inadequate use of their powers of vision than to their recognizing something as representing something else. It is not necessary to continue these . questions much farther. There is, we must conclude, a broad rift between animals and man. On the one side stands the animal, a creature living in the realm of the natural, the perceptual, the concrete only; on the other side stands man, living also in the world of the spiritual, the conceptual, the abstract.

ANIMAL

PSYCHOLOGY

From the standpoint of evolution we must, of course, admit that man and the human mind have evolved from an ancestry mentally akin to animals such as are living now on earth. In the animal world, therefore, we are justified in searching for some indication of the higher faculties of the human mind. We may find perhaps a few faint traces, pointing to higher possibilities, abilities from which higher mental forms may develop. But we shall not find very much of this nature. Notwithstanding all similarity between animal and man in their instincts and feelings, there will in the mental sphere always remain a deep cleft between the two. For even at the apex of animal intelligence, this intelligence will never be more than the intelligence of an animal.

CHAPTER VII

THE PROBLEM OF THE ANIMAL'S WORLD WHEN reading the concluding words of the preceding chapter, the reader will perhaps feel inclined to ponder what, then, is the world in which the animal lives ? Is it possible to get some idea of it ? Let us, in closing, try to form for ourselves an image of it. When doing so, it must from the start be emphasized that the animal lives in quite another world than man, even if, as is for instance the case with our domestic animals, he lives in the same surroundings. First of all, as we have already seen in our two preceding chapters, there exists scarcely any spiritual world for him. The animal has no religious experiences, does not venerate a higher being, and even if he may feel himself inferior and submissive towards the leader of his herd or troop, or towards some human being, such a submission shows only a natural and not a religious character. He has no aesthetic feelings and does not admire the beautiful nor despise the ugly, even though he himself may, to our eyes, be beautifully coloured,

ANIMAL

P8YC}IOLOGY

or make constructions which we admire as beautiful. He does not conform to moral considerations and does not distinguish between good and evil, and even when he aids other beings, as for instance his young, he does not know that he is performing a good deed. He may be honest and may not be able to feign or lie, but is not aware of it himself. He has no sense of humour, and does not laugh, even if a dog or a monkey may sometimes grin as an expression of well-being. He has few free images, if any, and has only a few faint glimmerings of phantasy. The spiritual world of the animal is very poor indeed, so far as it exists at all. His intellectual world, too, as we saw in the preceding chapters, is much more simply constructed than that of man. He does not possess concepts, and therefore cannot work with them, cannot manipulate them, cannot think, although he may find practical solutions for situations requiring understanding and insight. In this respect also he stands far beneath man. On the other hand the animal certainly experiences the urge of his instincts much more than man does. His instinctive world preponderates over other inner experiences. The most important and ever alert of these instincts is that of flight. His whole life long the animal, at least the animal living in a natural state, is threatened by dangers and must be on his guard against enemies. Some few animals may enjoy a paradise on earth where they are free from all danger (for instance, the dwellings of men for his domestic animals and the Zoological Gardens for the few wild animals that have found a refuge there), but such paradises are exceptions. The animal in nature lives in constant danger, and this danger will impart a speci;ll colour to his world, unknown to most of us. Next to this instinct of flight the feeding instinct is the most important, while, at particular moments of his life, the sexual and parental instincts prevail. All such urges are undoubtedly experienced much more strongly by animals than by us, partly because they have so few other experiences to counterbalance them. The feelings aroused by these instincts, too, are for the same reason probably stronger than in man, at least in the higher animals, although on the

ANIMAL

PSYCHOLOGY

1 49

other hand they are doubtless more elementary and much less blended with other emotions than with us. Moreover, the perceptual world of the animal also differs greatly from that of man. Man is a visual being, that is to say, sight is his most important sense, the sense he depends on. His perceptual world is chiefly a visual one. As long as he is in a waking state his sense of sight is in action, whether he walks or works, writes or reads. Visual perceptions direct his actions. He acts much less on auditory impressions though these of course are operative, as when he moves aside for a motor car he hears behind him, or answers the questions of a fellow-man. The sense of smell plays only a subordinate role in his life. His perceptual world is in substance a visual one. The perceptual worlds of animals not only differ from that of man, but are also different from each other. To start with, the order of the senses is different in different animals. In some of them the sense of sight plays the leading part, as in man, but in others the principal role is played by smell. The spider lives chiefly by feeling in a predominantly tactual world. Moreover, with regard to these perceptual worlds, we have to acknowledge that man is far from enjoying in all respects an unquestionable superiority. Man is a visual being, indeed, but the visual world of many animals is richer and more differentiated than ours. Many of them, and not only the nocturnal animals but also an animal like the dog, can see better in the dark than we do. The power of vision of the bird of prey, which descries from a great height the prey moving on earth below, is greatly superior to that of man. Man's world is a coloured one, indeed, and in this respect is richer than that of many animals which are colour-blind, and live in a world of lighter and darker hues, like man's world in the uncoloured films of the cinema. This is, for instance, the case with most of the mammals : the dog and the cat, the rabbit, the cow. Of the mammals, only monkeys and apes have a colour-sense comparable with that of man. On the other hand, man does not see ultra-violet light as coloured, as is the case with the bee, which, however, by way of compensation, does not see

ANIMAL

PSYCHOLOGY

red light as a colour, so that its visual world is shifted towards the side of the shorter wave-lengths; This does not render the colour-world of the bee richer, only different from ours. But it has been shown that some fishes distinguish no less than three different colours in the ultra-violet, while their red-vision is undiminished, so that their world is much richer in colour than ours. The sense of smell is not very well developed in man, so that many animals live in a much richer osmatic world than we. This fact is well known in the case of mammals, especially the beasts of prey. Our dog, for instance, lives in a world of smells of which it is hard for us to form any adequate idea. Sniffing, the dog follows the track a rabbit has left in the wood; his mates, and human friends, are recognized by him chiefly by their smell. It has been demonstrated that he is able to distinguish special smells when several are mingled, a task man cannot perform, as to us in a smell-mixture either one smell predominates, or a new smell is formed in which the components can no more be analysed. The dog, moreover, is able to recognize smells in a very weak solution ; for instance, iodoform in a dilute of I : 4,ooo,ooo, and sulphuric acid even in one of one to ten million. He is able to smell sodium chloride and quinine, substances odourless to man, and these even in weak dilutions. His retentiveness of scents is certainly much more developed than with us and he must have "smell-recollections" of persons and objects, just as we have visual recollections of them. It is not impossible, too, that his dreams are "smell-dreams," in which "smell-images" pass through his mind, just as our dreams are mostly built up of visual images. It also seems probable that to him human beings, and other dogs, have a "smell-physiognomy," i.e., that they may have for him a "malicious' or "good-natured" smell, just as for us persons may have a trustworthy, or unreliable, face, and that by emitting special scents that we ourselves cannot perceive, they convey their moods to him, so that the master "smells joyful" or "anxious" to his dog. The osmatic world of the dog is indeed of quite another nature than that of man 1 Birds, on the other hand. do not seem to smell at all.

ANIMAL

PSYCHOLOGY

The acoustic world of the dog is also much richer than ours. His hearing is very keen. In experiments of Engelmann a dog could hear a faint noise (the falling of a small steel ball of 3 rom. in diameter on a steel plate from a height of 3 em.) at four times as far away as did a man of acute auditory capacity. His hearing was thus 16 times that of man I He also perceives, and can be trained to react to so-called supersonic sounds, sounds so high in pitch that the human ear is unable to perceive them. Many other mammals, too, have better audition than man, and can perceive higher tones. Recent investigations have shown that bats, when flying, evade collisions with objects by emitting similar supersonic sounds and using them for the purpose of echo-location. Some lower animals, too, sometimes give proof of such acuteness of sense organs as we can scarcely imagine. When Fabre once let a female of the Emperor Moth Saturnia pavonia break out of the pupa in his room, in the evening about twenty males of that species collected round the wire cage in which the female had been put, and in a week's time he had in this way about I so males in his house, although the moths were rather rare in his country and must have come from distances of several miles. And when Forel reared some females of Saturnia carpini in a room in his house at Lausanne, the number of males besieging his window was so great as to attract a crowd of boys in the street below. That it was the sense of smell which brought the males to the female was shown by Fabre with another species of butterflies (Bombyx quercus). When a female was put under a glass-bell in the room, the arriving males paid no attention to her, fluttering under the bell, but went directly to the place where the day before she had been sitting on the sand in a dish. Such attainments by the sense of smell are inconceivable to us. Forel has pointed to the fact that in insects like ants and wasps the organs of touch and smell are situated together on the free movable and mobile antennre. He therefore suggests that such animals may perhaps integrate their tactile and smell-perceptions into one complex perception such a~

I

52

ANIMAL

PSYCHOLOGY

to distinguish "round" from "square smells," etc., in the same way as man in his form-perception may integrate optic and tactile perceptions. This idea is perhaps somewhat fantastic, but, be that as it may, it must be admitted that such a capacity of building "smell-forms" would impart to the~e animals' sensory world constituents wholly differing from those of our own smell-world. Have animals sense perceptions at their disposal of a kind wholly unknown to us ? This cannot be regarded as impossible, although it is naturally very difficult, if not impossible, for us to imagine such perceptions. It is known that some animals achieve very remarkable performances in finding their way back to their homes. Although many of the stories about dogs and cats finding their way home after displacement must certainly be treated with some scepticism, yet there have been some experiments, carried out with sufficient exactness as to be taken seriously, which have yielded surprising results. Bastian Schmid took a dog in a closed basket in a closed motor-lorry 20 km. along a road to a place 6 km. away, where the dog had never been before, and separated from the house by hills and woods. On being released, the dog, after a short period of hesitation, started in a direction almost exactly corresponding to that of his house and arrived there, although not in quite the shortest way, about one hour and a half later, having covered a distance of about I I km. On a repetition of the experiment some weeks later he started after a few minutes in the right direction and arrived home in about three-quarters of an hour. What sense perception may have revealed to him the location of his house ? That migratory birds in spring ·often come back to the place of their former nesting after their sojourn in the South is a fact known to every bird-friend, although we do not understand how they manage to find this place again. Still more remarkable are the results obtained by Riippell with birds taken by him from their nests and carried away great distances. To quote only the most astonishing among many other results, out of twenty swallows taken from their nests near Berlin and sent by airplane to Madrid and Athens, a distance of about I ,8oo km., in seven days no

ANIMAL

PSYCHOLOGY

1

53

less than four came back to their nests, two from each of the towns. How they were able to perform this is incomprehensible to us. The idea of some sense unknown to us intrudes itself upon our mind, although, it must immediately be added, the admission of such an unknown sense is a somewhat easy and cheap solution of a problem, and may therefore be used only if all other possibilities have been rejected with certainty. It is clear, therefore, that the animals not only live in a world which differs vastly from that of man, but also that among the animals themselves the worlds they ·live in are quite different. Furthermore, we have to admit that there exist as many animal worlds as there are species of animals. The world of one animal will, intellectually and perceptually, be much richer and much more finely constructed than that of another. If it could observe and could appreciate such differences, the worm would look up as much to the ant as the chick would to the monkey. Generally speaking, the further one descends along the scale of animal life, the more the world of the animal shrivels. We saw this already in the preceding chapters with regard to the intellectual worlds. But the same holds true for the perceptual worlds. Where the structure of the eye becomes more primitive, the animal no longer perceives distinct images of its surroundings nor sees the forms and contours of things, but receives merely more or less vague impressions of them. At a low stage of development of the eye only the direction of the light is perceived ; at a still lower stage only differences in the intensity of the light, so that the animal merely distinguishes between light and dark, or perceives the changes caused by the movement of light and shadow. The visual world of such animals is of course much poorer than that of better equipped ones. In eyeless animals the whole light perception, if any, is based on the photic sensibility of the epidermis, as a matter of course much weaker than that of the eyes. That fishes do hear, a faculty much doubted, has been sufficiently proved of late years by von Frisch and his fellow workers. But among the invertebrates only a few insects really hear, so that for almost all lower animals the whole world of

ANIMAL

PSYCHOLOGY

sounds is closed. Most animals are less sensitive to pain than we are, as is shown by the fact that foxes and rats sometimes gnaw off their legs when caught in a trap. Birds do not show any sign of pain when their wings are clipped and thereby their "hand" is cut off. Among the invertebrates pain does not seem to exist at all. An ant cut in two by Forel went on sucking honey as if nothing had happened. The chemical sense, too, disappears more or less in the lower orders. Probably only the tactile sense, the sense of being touched, is found throughout the whole animal realm. According to von Uexkiill the whole world of the Medusa Rhizostoma, when floating on the blue waves of the Mediterranean under the southern sky, consists only in the perception of the rhythmical contractions of its own muscular apparatus. Whether this is quite true may perhaps be doubted, but it may serve as a good example of the poverty of the world of some lower animals. Thus every animal lives in his own world, differing from that of other animals, differing especially from that of man. The knowledge of all these different animal worlds, embracing a knowledge of each animal's sensations and perceptions, of his feelings and drives, of his intellectual faculties and spiritual experiences, if any, must be regarded as the ultimate aim of the study of animal psychology.

ANIMAL

PSYCHOLOGY

155

BIBLIOGRAPHY

IN a volume of this size there is, of course, no place for a complete list of all the publications mentioned in the text. The author, therefore, believes he may restrict himself to naming some books of a more general character in the field of animal psychology, and for the rest refer to his "Tierischen lnstinkte," in which a list of about 900 titles of articles and works on the subject of this book may be found.

J. A. Bierens de Haan. Animal Psychology for Biologists. Three

Lectures. London, 1929. - - - Die tierpsychologische Forschung. Ihre Ziele und Wege. Leipzig, 1935. - - - Labyrinth und Umweg. Ein Kapitel aus der Tierpsychologie. Leiden, 1937. - - - Die tierischen Instinkte und ihr Umbau durch Erfahrung. Eine Einfuhrung in die allgemeine Tierpsychologie. Leiden, 1940. J. H. Fabre. Souvenirs entomologiques. Etudes sur !'instinct et les moeurs des insectes. Serie I-X. Paris, 1879-1910. P. Guillaume. La Psychologie Animale. Paris, 1940. F. Hempelmann. Tierpsychologie vom Standpunkte des Biologen. Leipzig, 1926. R. W. G. Hingston. Problems of Instinct and Intelligence. London, 1928. L. T. Hob house. Mind in Evolution. 2nd Edition. London, 1915. H. S. Jennings. Behaviour of Lower Organisms. New York, 1906. D. Katz. Animals and Men. Studies in Comparative Psychology. London, 1937. H. Kluver. Behavior Mechanisms in Monkeys. Chicago, 1933. W. Kohler. The Mentality of Apes. New York, 1925. A. Kuhn. Die Orientierung der Tiere im Raume. Jena, 1919. J. Loeb. Forced Movements, Tropisms, and Animal Conduct. Philadelphia, 1918. N. R. F. Maier and T. C. Schneirla. Principles of Animal Psychology. New York, 1935. W. McDougall. Psychology. The Study of Behaviour. London, 1912.

ANIMAL

PSYCHOLOGY

W. McDougall. An Outline of Psychology. London, 1923. G. de Montpellier. Conduites Intelligentes et Psychisme chez !'Animal et chez !'Homme. Etude de Psychologie comparee. Louvain, 1946. C. L. Morgan. Habit and Instinct. London, 1896. - - - Animal Behaviour. 2nd Edition. London, 1920. G. W. and E. G. Peckham. Wasps Social and Solitary. Westminster, 1905. E. S. Russell. The Behaviour of Animals. An Introduction to Its Study. 2nd Edition. London, 1938. E. C. Tolman. Purposive Behavior in Animals and Men. New York, 1932. J. B. Watson. Behaviorism. New York, 1924. R. M. Yerkes. The Mental Life of Monkeys and Apes. A Study of Ideational Behavior. Behavior Monographs, No. 12. New York, 1916. G. Zunini. Animali e Uomo. Visti da uno Psicologo. Milano, 1947.

INDEX

INDEX So Animal activity, explanation of, S Animal psychology, methods of, I3 object of, II possibility of, I I Attention in instinctive activity, 66 ANEMOTAXIS,

I5 marks of, r6 Behaviourism, 33 Box, using of, 13S BEHAVIOUR,

CHEMOTAXIS, Concepts, I4~

Instincts, adaptations of, 6I alarming, 73 deferred, 43 orientating, 73 regulations of, 6~ system of, 46 unpurposefulness of, 50 variability of, 6o Intelligence, 67, 94 definition of, 96 Intuition, sympathetic, 14 LEARNING,

So

definition of, 90

physiological, 91 psychical, 93 Maze, 1~3 Motor habits, I07 Moving objects, utilising the possibilities of, I 30 MATURATION,

DETOUR, II9

Displacement reactions, 70 Distinction, Io6

EXPERIENCE,

94

GENERALIZATIONS,

Geotaxis, 79

HABITUATION,

psychological, Homing, I09 Hygrotaxis, 79

perceptual, 14.5

physiological,

9~

IO~

I4I Images, free, 14:a Imagination, I40 Imitation, roo Insight, u6 Instinct and instinctive action, 39 breadth of, 69, 99 characteristic of the species, 39 and cognition, 7I definition of, 38, 88 diagram of, Sg innateness of, 4~ maturation of, 43 and morphological structure, 41 purposiveness of, 46 rigidity and suppleness of, 56, 67 IDEATION,

OBJECTIVISM,

3~

Obstacles, clearing away, 12S

and Topotaxis, 77 Phototaxis, 7S Physiology and Psychology, 9 Pictures, recognition of, 146 Place-memory, I09 Practice, 93 Protozoan behaviour, 25 Psychic phenomena, xo PHOBOTAXIS

rS Reflexes, conditioned, 105 Relations, grasping causal, 127 , spatial, II7 , temporal, 124 Releaser, 53 Repairing damaged constructions, 62 Rheotaxis, 79

REFLEX,

S3 Short cuts, 12:a

SCOTOTAXIS,

159

x6o

INDEX

Signal function, 104 Skill, acquirement of, ro8 Solution, primary and secondary,

us, u6

Soul, I I Stick, using of, 136 Substitutive actions, 72. and tactic movements, 76 Thermotaxis, 79 Thigmotaxis, So Tool, definition of, 133 Tools, construction and improvement of, 139 instinctive use of, I 34 spurious use of, 130 intelligent use of, I 36

Training, roo Trial and Error, no Tropisms, 76 theory of, 9 UNDERSTANDING, Il2

concrete,

I I

3

TAXES

88, 98 Vibrotaxis, So Volition, ror VALENCE,

the animal's, 147 Words, 143 WORLD,

PSYCHOLOGY LIBRARY EDITIONS: COMPARATIVE PSYCHOLOGY

Volume 2

THE BEHAVIORAL SIGNIFICANCE OF COLOR

THE BEHAVIORAL SIGNIFICANCE OF COLOR

Edited by EDWARD H. BURTT, JR.

First published in 1979 by Garland Publishing Inc. This edition first published in 2018 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 711 Third Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 1979 Garland Publishing Inc. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: ISBN: ISBN: ISBN:

978-1-138-50329-8 978-1-351-12878-0 978-1-138-57496-0 978-1-351-27044-1

(Set) (Set) (ebk) (Volume 2) (hbk) (Volume 2) (ebk)

Publisher’s Note The publisher has gone to great lengths to ensure the quality of this reprint but points out that some imperfections in the original copies may be apparent. Disclaimer The publisher has made every effort to trace copyright holders and would welcome correspondence from those they have been unable to trace.

l BeliThe. avtora Significance of Color

Edward H. Burtt, Jr. Department of Zoology Ohio Wesleyan University Delaware, Ohio

Garland STPM Press New York & London

Copyright

© 1979 by Garland Publishing,

Inc.

All rights reserved. No part of this work covered by the copyright hereon may be reproduced or used in any form or by any means -graphic, electronic,. or mechanical, including photocopying, recording, taping, or information storage and retrieval systemswithout permission of the publisher.

15

14

13

12

11

10

9

8

7

6

5

4

3 2

Library of Congress Cataloging in Publication Data Main entry under title: The Behavioral significance of color. (Garland series in ethology) "Based on a symposium held as part of the Animal Behavior Society's meeting in June 1977." Includes bibliographies and index. 1. Color of animals. 2. Animals, Habits and behavior of. I. Burtt, Edward H., 1948II. Animal Behavior Society. III. Series. QL767.B413 591.5 77-14618 ISBN 0-8240-7016-X

Printed in the United States of America

1

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . ix . . xi

Contributing Authors Introduction Edward H. Burtt, Jr. PART I:

-.

.xiii

PHYSICAL PRINCIPLES

1. Physics of Light: An Introduction for Light-Minded Ethologists . . . . . . . . . . . . . . . . . . '. 3 B. Dennis Sustare Emissivity, A Little-Explored Variable: Discussion . . . . . . . . . . . . . . . . . . 28 C. Richard Tracy PART II:

PHYSIOLOGICAL FUNCTIONS OF ANIMAL COLORATION

2. The Influence of Color on Behavioral Thermoregulation and Hydroregulation . 35 David M. Hoppe Further Thoughts on Anuran Thermoregulation: . . • . 63 Discussion . C. Richard Tracy Maximization of Reproduction: Discussion . . . . . . William J. Hamilton III

. . . . . . . . . . . 69

Audience Questions: Discussion . . . . . . . . . . . 70

vi

Contents

3. Tips on Wings and Other Things . . . . . . . . . . . 75 Edward H. Burtt, Jr.

The Evolutiono-Engineering Approach: Discussion . . 111 C. Richard Tracy

Where is the Evidence for Ultra-violet Damage: Discussion . . . . . . . 114 William J. Hamilton

Audience Questions:

III

Discussion . . . . . . . . . . . 120

PART III: PHOTORECEPTION 4. Extraretinal Photoreception . . . . . . . . . . . . 127 Herbert Underwood

Extraretinal Photoreception: Words of Caution: Discussion . . . . 179 C. Richard Tracy

5. Mechanisms of Color Vision: An Ethologist's Primer . . . . . .

. . . . 183

Samuel H. Gruber

6. Visual Discriminations Encountered in Food Foraging by a Neotropical Primate: Implications for the Evolution of Color Vision . . . . . . . . . 237 D. Max Snodderly

Comments on Coevolution: Discussion . . . . . . . . 280 C. Richard Tracy

Are Selection Pressures Different? William J. Hamilton

Audience Questions:

Discussion . . . 282

III

Discussion . . . . . . . . . . . 284

PART IV: COLORATION FOR COMMUNICATION 7. Environmental Light and Conspicuous Colors . . . . . 289 Jack P. Hailman

Audience Questions:

Discussion . . . . . . . . . . . 355

vii

Contents

8.

Optical Signals and Interspecific Communication . . 359 Jeffrey R. Baylis

9.

The Use of Color in Intraspecific Communication . . 379 William J. Rowland Visual Functions of Color: Approach: Discussion C. Richard Tracy Audience Questions:

The Predictive

Discussion.

• • . . 422 424

10. Conclusion . • . . . Edward H. Burtt, Jr.

427

Index • . .

433

Preface This volume grew out of a symposium presented at the meeting of the Animal Behavior Society in June 1977. The chapters are based on the papers read before the society, but the authors have included new material not presented and frequently unknown at that time. Each chapter is followed by a discussion that includes comments from C. Richard Tracy or William J. Hamilton, III or both, questions from the symposium's audience, and answers prepared by the author of the chapter. I hope that such a format captures the enthusiasm and spontaneity so evident at the symposium. Animal coloration has received too little study. If this book excites further thought and research into the world of colors and patterns of color, then it will have fulfilled its purpose. Many are those who have contributed to this book. Edward 0. Price, Program Chairman for the Animal Behavior Society, was an invaluable source of encouragement during the planning and organizing of the symposium. The authors have been exceptionally pleasant and cooperative and have made outstanding contributions. Although the questioners are infrequently identified, their queries and the ensuing discussion add immeasurably to the book. The symposium was organized while I was a member of the Department of Psychology at the University of Tennessee; I am grateful to the department for its support of the entire project, especially to Becky George and Lorraine Simmons, who typed many letters and relayed many telephone messages. The Department of Zoology at Ohio Wesleyan University provided facilities for the preparation of the manuscript; to the members of the department, especially A. John Gatz -- whose advice was freely

X

Preface

given -- and to Ohio Wesleyan University, I offer heartfelt thanks. Gladys Hummel, who typed the entire, edited manuscript, has earned my eternal thanks as have Brian Fortini and Donna Wierbowicz who helped with the proofreading. And for George Narita, our long-suffering but ever helpful publisher, I offer the hope that the book is worthy of his expectations. Now I leave the reading to you and return to the family too often neglected during the past months of writing and editing. Edward H. Burtt, Jr. 5 March 1978

Contributing Authors Jeffrey R. Baylis Department of Zoology University of Wisconsin Madison, Wisconsin Edward H. Burtt, Jr. Department of Zoology Ohio Wesleyan University Delaware, Ohio

David M. Hoppe Department of Zoology & Entomology University of Minnesota Morris, Minnesota William J. Rowland Department of Biology Indiana University Bloomington, Indiana

Samuel H. Gruber Division of Biology & Living Resources University of Miami Miami, Florida

D. Max Snodderly Eye Research Institute of Retina Foundation Boston, Massachusetts

Jack P. Hailman Department of Zoology University of Wisconsin Madison, Wisconsin

B. Dennis Sustare Department of Biology Clarkson College Potsdam, New York

W. J. Hamilton III Division of Environmental Studies University of California Davis, California

C. Richard Tracy Department of Zoology & Entomology Colorado State University Fort Collins, Colorado

Herbert Underwood Department of Zoology North Carolina State University Raleigh, North Carolina

Introduction Coloration and the pattern of coloration of animals play a central role in behavior -- even among species in which vision is not the dominant sense. For example, an animal's color or pattern is often related to its display movements (Lorenz 1941, Tinbergen 1952, Blest 1957, Cullen 1957, Hailman 1977a, 1977b); cryptic coloration often goes hand in hand with cryptic behavior (Ruiter 1952, 1956, Cott 1957, Hailman 1977b); and coloration that absorbs solar energy may show a close relation with thermoregulatory behavior (Ohmart and Lasiewski 1970, Burtt 1973, Storer 1974). Despite these pervasive correlations between behavior and coloration, we understand few of the principles that predict the specific pattern or color that best serves a particular function in conjunction with behavior. This volume brings together diverse experts and disparate facts as a step toward understanding the behavioral significance of animal coloration. Hypotheses that account for the coloration of animals fall into three major categories: (1) physiological functions of pigmentation; (2) coloration that affects the animal's visibility to other animals, including conspecifics; and (3) coloration that affects the animal's own VlSlon. Predicative hypotheses in all three categories are outlined and data relevant to these hypotheses are presented in the chapters, discussions, questions, and answers that follow. REFERENCES Blest, A. D. 1957. The function of eyespot patterns in the Lepidoptera. Behavior 11: 209-256.

xiv

Introduction

Burtt, E. H., Jr. 1973. Warbler leg coloration and migratory behavior. Amer. Zoologist 13: 1263. Cott, H. B. 1957. Adaptive Coloration in Animals. Methuen, London. Cullen, E. 1957. Adaptations in the kittiwake to cliffnesting. Ibis 99: 275-302. Hailman, J. P. 1977a. Communication by reflected light. In How Animals Communicate. T. A. Sebeok (ed). Ind. Univ. Press, Bloomington. Hailman, J. P. 1977b. Optical Signals: Animal Communication and Light. Ind. Univ. Press, Bloomington and London. Lorenz, K. Z. 1941. Vergleichende Bewegungsstudien on Anatinen. J. f. Ornithol. 89: 194-294. Ohmart, R. D., and R. C. Lasiewski. 1970. Roadrunners: energy conservation by hypothermia and absorption of sunlight. Science 172: 67-69. Ruiter, L. de. 1952. Some experiments on the camouflage of stick caterpillars. Behavior 4: 222-232. Ruiter, L. de. 1956. Countershading in caterpillars. An analysis of its adaptive significance. Arch. Neerl. Zool. 11: 285-342. Storer, R. W. 1974. Sunbathing in grebes. Abstr. 16th Int. Orni thol . Congr. Tinbergen, N. 1952. "Derived" activities: their causation, biological significance, origin, and emancipation during evolution. Rev. Biol. 27: 1-32.

Partl

Physical Principles

Chapter 1

Physics of Light:

An Introduction for Light-Minded Ethologists B. Dennis Sustare Introduction Electromagnetic Radiation Waves or Particles? Mathematical Description What is Color? Geometrical Optics Reflection Refraction Focal Points Diffraction Sources of Electromagnetic Radiation Properties of Black Bodies Emissivity The Measurement of Radiation Inverse-Square Law Radiometry Photometry Sunlight Solar Spectra Sunlight's Earthly Fates Polarization Interference Water and Light Zoochromes and Phytochromes Bioluminescence Recommended Reading

5

Physics of Light

INTRODUCTION Energy is absorbed, reflected, and radiated by all plants and animals and most animals use some of the information carried by radiant energy to modify their subsequent behavior. Regardless of the behavior examined, the unceasing flow of radiant energy affects the behavioral system. Hence ethologists must renew or make their acquaintance with the physical properties of radiation. For the sake of those whose memories are rusty, I briefly survey some features of the physics of electromagnetic radiation. ELECTROMAGNETIC RADIATION Waves or Particles? Electromagnetic radiation can be thought of as little packages of energy, each package being a discrete unit. The amount of energy in a package characterizes the type of radiation. Along the bottom scale in Figure 1, energy is plotted in attojoules. As a reminder, a joule is the energy required to apply a force of one newton over a distance of one meter, a sizable amount of energy. Because the metric prefix atto- corresponds to a factor of 10- 1 8, an attojoule is a very tiny unit of energy. You can see in Figure 1 that there is little energy in the packages of a radio wave; infrared, visible light (what we can see), and ultraviolet show increasing amounts of energy/package respectively. If the figure continued to the left, more highly energetic radiation would be displayed, for example, X rays. Energy/package is only one way to characterize electromagnetic radiation. Under some experimental conditions, electromagnetic radiation acts like a stream of particles; under other conditions radiation acts like waves. The more energetic each package of radiation, the more it acts like a particle. High-energy radiation can be more accurately located in time and space and tends to penetrate substances better than low-energy radiation. Low-energy radiation is very wavelike; it bends around objects easily, and is hard to locate in time and space. However, regardless of the energy content, any radiation may demonstrate wave properties, such as the ability to interfere with other radiation in the same way that water waves or sound waves interfere,

6

B. Dennis Sustare

• PARTICLE-

WAVE-·

LIKE

LIKE

I Visible I Frequency in Hz

1200

I

250

1000

I

300

800 I

I

.7

600

I

200

400

I

I

I

I

.6

.5

.4

.3

.2

I

5000

Radio

Infrared

Visible

0

I

400 500 600 750 1000 Wavelength in nm (=10-• m)

Ultraviolet

.8

(=10' 2 Hz)

.1

0

Energy in aJ (= 10-" J)

Fig. 1. Lower scale: Energy of photons, measured in attojoules. Spectral equivalences are shown above the scale (only shown to 0.8 aJ). Upper scales: Correspondence between frequency and wavelength of photons that have the amount of energy shown on the bottom scale. Only that portion of the spectrum up to 1200 THz is shown.

adding in some places, canceling in other places. Radiation tends to seem more wavelike when it is being transmitted through space, and more particle-like when it interacts with materials. High-energy radiation acts as though its waves were very close together, with a high repetition rate, or frequency, and a very short distance between adjacent wave crests. On the upper scales of Figure 1 are the frequencies and wavelengths corresponding to the energies plotted on the lower scale. Frequency is measured in hertz, one hertz being one wave crest/second; a terahertz is 101 2 hertz. Wavelength, the distance between adjacent wave crests, is measured in nanometers in Figure 1, one nanometer equaling 1Q-9 meter. Interestingly, visible light displays a good mix of wave and particle properties, given the scale and sensitivity of our usual measuring devices. Note that the perceived color of a beam of light is a function of the beam's energy. Radiation at the low-energy

7

Physics of Light

end of the visible light segment appears red. As we progress toward higher energies, the light appears to pass through the colors of the rainbow: red, orange, yellow, green, blue, and violet, the last being at the high-energy end of the visible range. Mathematical Description There is a direct relation between the frequency of the radiation and the energy in one radiation package, called a quantum or photon. Note that frequency and energy are plotted arithmetically in Figure 1; they differ only by a constant factor:

u =

hv

(1.1)

where u represents the energy of a single quantum and v is the frequency of that quantum. If energy is measured in joules, and frequency in hertz, the proportionality constant ish, Planck's constant (6.6256 x 1o-34 J•s). There is also a regular relationship between the wavelength and the frequency of radiation. The speed of an object is the distance it travels per unit time. If you multiply the wavelength, a distance, by the frequency, the reciprocal of time, the result is the speed of light (or any other electromagnetic radiation): c

= AV

(1.2)

where A is the symbol for wavelength and c for speed of light. One useful fact about electromagnetic radiation is that all quanta travel at the same speed in a total vacuum (approximated by outer space), no matter what their energy content; speed is independent of the wavelength or frequency of the radiation. This speed is very close to 3 x 10 8 m·s-1. You may find the wavelength (A) by dividing the speed (c) by the frequency (v); similarly, divide the speed (c) by the wavelength (A) to find frequency (v); but be careful that your units are consistent. WHAT IS COLOR? What happens when light strikes something? (The following discussion deals with the behavior of a beam of many quanta or packages, not with the behavior of a single quantum.) In general, three things may happen to the light.

B. Dennis Sustare

8

It may bounce off the surface (reflection), it may be absorbed and transfer its energy to the absorbing material, or it may pass entirely through the substance (transmission). Often all three processes occur, with some quanta meeting each fate according to the relationship: (1. 3)

where I is the number of incident quanta, R the number of reflected quanta, A the number of absorbed quanta, and T the number of transmitted quanta. The perceived color of an object depends on the frequencies of visible light remaining after absorption. If the object reflects frequencies that we would call green, then we call the object green. Similarly, if we look at an object by transmitted light, we would call it green if it transmits green light, i.e., light in the frequency range of about 540-600 THz (about 0.36-0.4 aJ or 500-555 nm). The color of an object is determined by the light that is not absorbed by the object. If you shine a green light on an object that absorbs green light, it appears black. The process of absorption is complicated, and is not discussed in detail; but, briefly, absorption of light by the molecules of the substance causes those molecules to become excited, i.e., to have more energy. The molecules may give up this energy by reradiating it (but at a lower frequency, and hence with less energy, than the radiation absorbed--a phenomenon called fluorescence), or by bumping into a nearby molecule and giving up some of the energy to the bumped molecule, or by flying off from the surface or otherwise converting the energy into work performed. GEOMETRICAL OPTICS A quantum (package of radiation) is in motion during its entire existence and tends to maintain its direction, all other things being equal, in a manner similar to the inertia of objects in motion. Those "other things" add many complications that are of little or no importance to ethologists, so I will not discuss such other things as strong gravitational fields, or what "maintain its direction" really means. The movement of radiation in straight lines allows some useful approximations, called geometrical optics, to be developed.

9

Physics of Light

Reflection Fig. 2. Reflection: The angle of incidence equals the angle of reflection; here, it is measured relative to the surface plane rather than to the perpendicular with the surface.

Reflection If light strikes a mirror and bounces off again, the angle between the reflected beam and the mirror is the same as that between the impinging beam and the mirror (Fig. 2). These angles are customarily measured with respect to the perpendicular to the mirror surface rather than as in the figure, but you will recall from your high school geometry that this makes no difference in the equality of angles entering and leaving. Refraction When light passes from one medium into another, it changes direction, bending closer to the perpendicular when it enters a denser medium (Fig. 3). This process is called refraction. The relation between the angles is not as simple as with reflection. sin 6

=n

sin

~

(1. 4)

Equation (1.4) states that the sine of one angle is a constant (n) times the sine of the other angle. What is this constant of proportionality? You may have noticed the earlier qualification, that the speed of light is constant in a vacuum. The qualification is necessary because radiation moves more slowly when passing through a medium. The

B. Dennis Sustare

10

Refraction

Fig. 3. Refraction: the light bends as its speed undergoes a transition between two media differing in refractive index.

amount of slowing depends on the medium: slower in water than in air; slightly slower in air than in a vacuum. The frequency of the radiation does not change as it enters the new medium, so that with a new speed the distance between wave crests (wavelength) must change. Wave crests for the waves on both sides of the interface between two media must have the same spacing along the interface, so that there must be a bending for rays that strike the interface at an angle other than the perpendicular. Now, for the proportionality constant n: it is the ratio of the apparent speeds of the radiation on the two sides of the interface, speed after interface speed before interface and is called the index of refraction. Equation (1.4) is sometimes called Snell's Law. I stated that the amount of bending depends on the density of the medium. More precisely, the index of refraction depends on the number of charges (i.e., electrons) per unit volume in the material. The index is also a function of the frequency of the radiation, being somewhat higher (more bending) for blue light

Physics of Light

11

than for red light. This is why a prism separates the frequencies of light in a beam of white light, bending the blue components more than the red. What appears white to us is of course a combination of frequencies, not light of a single frequency. One interesting effect of refraction is a change in the apparent depth of an object located in a different medium (Fig. 4). In Figure 4, consider a source of radiation located at point o'. When the rays reach the surface of the glass and enter the air, they bend outward, according to Snell's Law. If one is in the air region and looks toward the source, the rays appear to be corning from a point source located at 0. Therefore the source is actually deeper in the glass (or water) than it appears to be.

Bringing many electromagnetic rays together at the same point is called focusing (Fig. 5). The bending of

o'

Fig. 4. Change in apparent depth due to refraction. Due to the refractive bending of light, a light source in the glass at point Q will appear to be at point O'to an observer in the air.

12

B. Dennis Sustare

Focal Points Bringing many electromagnetic rays together at the same point is called focusing (Fig. 5). The bending of radiation by glass or other materials allows the construction of focusing devices such as a lens. In the upper part

LENS

PINHOLE

Fig. 5. Top: Focusing by means of a lens. The dotted lines show the focusing of parallel rays (from an infinitely distant source) at the focal point of the lens. The solid lines show that rays may follow different paths to reach the same point in the image. Bottom: Focusing by means of a pinhole. Note that the image is in focus at all distances on the other side of the pinhole, but that there is only one path for light originating at any one point at the source.

Physics of Light

13

of Figure 5, the dotted lines show parallel rays from an infinitely distant source reaching the lens and bending to focus at a point. The distance from the lens to this point of focus (ignoring the size of the lens itself) is the focal length of the lens. Placing a source at the focal point produces a collimated beam of radiation, i.e., a beam in which all the rays are parallel. If an object, such as the arrow in Figure 5, is located farther away than the focal point (but not at infinity), visible light from each point on the object passes through the lens and comes to a focus at one point on the other side of the lens, beyond the focal point. If the lens is properly shaped, light along many paths will come to the same focus, forming an image of the object. The proper shape can be difficult. Figure 5 represents a lens with a spherical surface on either side, which gives a fairly good approximation to a perfectly focusing lens. Nevertheless, rays passing near the center of the lens will not focus at the same point as rays passing near the edge, causing what is called spherical aberration. If you recall that different frequencies are bent by different amounts, you might anticipate that another source of focusing error is due to the failure to focus all frequencies at the same point, called chromatic aberration when referring to visible light. Having a properly designed lens to form an image is one way to resolve the ambiguity caused by radiation traveling outward in all directions from each point of an object. This principle is used to form an image on film in a lenstype camera. The lower figure shows another way in which an unambiguous image may be formed on film in a camera: using a pinhole instead of a lens. Note that there is only one route for visible light (or any other type of radiation) leaving a point on the object. Unlike the lens, which is in focus at only one distance for an object at a fixed distance from the lens, the pinhole forms an image that is in focus at all distances. Conversely, objects at all distances from the pinhole are in focus at any given plane on the other side of the pinhole. Thus a pinhole camera has an infinite depth of field; the more a conventional camera is stopped down by closing up the hole in the shutter through which the light passes, the greater its depth of field. One problem with the pinhole, of course, is that it lets very little light pass. The lens is able to collect light that was going in different directions and bring it together at a point; the pinhole selects only one ray of light from each direction.

14

B. Dennis Sustare

Diffraction Another phenomenon that appears with the pinhole is diffraction (not to be confused with refraction). Diffraction is another source of bending, and occurs when radiation passes through a small opening or past a sharp edge. Diffraction is tied to the interference of waves, and is linked to the uncertainty principle. In any case, all that need be remembered is that diffraction tends to fuzz up images and reduce resolution. Diffraction is a drawback to trying to form an image through a very small pupil of an eye. SOURCES OF ELECTROMAGNETIC RADIATION Passing from optics to the properties of radiation sources, let us consider what happens when a source radiates. One easy way to get an object to radiate is to heat it. You need not heat the object much; all objects above absolute zero constantly radiate. What's more, every macroscopic object radiates at all frequencies; you are, in theory, radiating visible light, radio waves, and X rays as you read this chapter, simply because your temperature is greater than -273°C (0°K). Why aren't we all dazzled by the glare? Of course, it is because the probability of radiating is not the same at all frequencies. Properties of Black Bodies An object that absorbs all radiation that falls on it, no matter what the frequency, would be an ideal black body. No radiation is reflected, so that the object must appear black. However, blackbody is a misnomer, since any blackbody above absolute zero (as it must be as soon as it absorbs energy) is radiating energy. Nonetheless, this theoretical model is convenient for discussing radiation, since the complications of differential absorption and emission may be ignored. So what kind of radiation does a heated blackbody give off? Figure 6 shows the radiation given off by a series of blackbodies, at temperatures ranging from 200°C to 6000°K; the latter is the temperature of a blackbody whose radiation approximates that of the sun. The radiation spectrum is plotted as wavelength, with short wavelengths (high frequencies) to the left. The line marked VIS indicates the visible spectrum. Notice that increasing the

Physics of Light

Q)

15

104

(.)

c:

co

:0102

co

a:

1.0

Fig. 6. Radiation from blackbodies. Each curve represents the radiance from a blackbody at a given temperature (in Kelvins). Wavelength is plotted on a log scale, in micrometers, with the shortest wavelengths (highest frequencies) to the left. The line labeled VIS indicates the visible spectrum. Radiance is measured in W·m- 2 -sr-l•Jlm-1. The diagonal line intersects the curves at their maxima (after Barnes 1968). temperature increases the energy radiated at all wavelengths, though the maxima of .the curves shift to the left with higher temperatures. This result should not be surprising; heating up the black body results in photons of higher energy (thus, higher frequency) being radiated. Note how the energy from the 6000°K source is at a maximum in the visible range. Notice also that radiation from the earth at a temperature of 273°K, the freezing point of water, is concentrated in wavelengths longer than the visible, in the infrared region, with a peak between 10 and 20 m. The shift in maxima with temperature is known as Wien's Displacement Law. Emissivity The radiant emittance from a hot surface is a function of the temperature of the surface, even in the case of real

16

B. Dennis Sustare

objects that deviate from the behavior of ideal blackbodies. In fact, the radiant emittance varies as the fourth power of the absolute temperature; doubling the temperature increases the radiant emittance by a factor of 2 4 , or 16. If you consider two objects that are near one another, they are of course radiating to each other. The net flow of energy depends on their relative temperatures: (1.5) where QR is the net radiative transfer (in watts) and T is the absolute temperature, measured in degrees Kelvin. The term A is the view factor, the area through which the radiation passes between the two sources. The emissivity, E, tells how well the source approximates a blackbody; a perfect black body would have an emissivity of 1.0 at all frequencies. The Stefan-Boltzmann constant, cr (5.6697 x 10- 8 W·m-2•°K-4) is a proportionality derived from a combination of K (Boltzmann's constant), the speed of light in a vacuum, and Planck's constant. Equation (1.5) is at the heart of any model dealing with the exchange of thermal radiation between objects. The emissivity of biological tissue is approximately 1.0 in the middle infrared portion of the spectrum, due largely to strong absorption by water in these frequencies. The high emissivity is independent of pigmentation in the visible region. Plants and animals act very nearly as black bodies in the middle infrared. They absorb and emit these radiations very effectively. These are the frequencies of maximal radiation for objects at normal living temperatures. Therefore, coloration has essentially no effect on the gain and loss of infrared radiation by organisms. Emissivity in the visible portion of the spectrum varies with coloration. Hence, organisms with dark skins, feathers, or fur absorb more visible radiation than those with light surfaces; recall that the peak of solar output is in the visible range. THE MEASUREMENT OF RADIATION Inverse-Square Law Energy from a point source is radiated in all directions. In a vacuum, with nothing to intercept the rays, they spread out as they travel away from the source. The amount of energy passing a cross section of a radiating

Physics of Light

17

cone in a unit of time is the same for any section of the same cone. This means that the intensity of radiation per unit area in the cross section decreases as the area increases. Doubling the distance from the point source causes a fourfold increase in the area of the cross section of the cone. Hence, the intensity per unit area decreases with the square of the distance from the source: the Inverse-Square Law of radiation from a point source. Radiometry The measurement of radiation can be very confusing, since there are many ways to take a measurement. Many persons have fallen prey to this confusion, and are not aided by the similarity of some of the terms used in the measurement of radiation, known as radiometry. Consider radiation emitted from a source, passing through an intervening medium, and striking a target. Measurements may be relative to the source of the target; total energy or energy per unit time may be recorded; radiation may be measured in all directions or only in certain directions. A rather cumbersome diagram (Fig. 7) outlines some of these concepts. Electromagnetic energy represents the total output of the source, over an infinite range of frequencies. Any real sampling device can detect only a portion of that range, of course, though radiant energy refers to an attempt

RADIOMETRIC CONCEPTS

{onto a surface)

(from a surface)

(onto a surface)

(from a surface)

(onto a surface)

(from a surface)

Fig. 7. Logical relationships of radiometric concepts; further explanation in the text (after Preisendorfer 1976).

B. Dennis Sustare

18

to measure the energy at all frequencies without respect to the sensitivity of our eyes to perceive the radiation. Radiant energy is measured in joules, and is an indication of the total ability to perform work. The key concept in this diagram is that of radiant flux, defined as the rate of flow of radiant energy across a theoretical surface. This is a measure of power, or energy per unit time, and thus is in units of watts (joules/second). Since flux is measured across the spectrum, there is the problem of breaking up a continuous distribution into small bands of equal frequencies, unless you have a source that radiates at only a single frequency. This is a problem of integrating measurements across the distribution; it is not discussed further. The next step is to consider the geometry of the . measurement. If a parallel flow of quanta through an area is considered, the area density of radiant flux is measured (in watts/m 2 ), sometimes called the flux density. If a measure is made of the solid angle density of radiant flux from a point, this is recorded in units of watts/steradian (the steradian being a measure of the solid angle). The solid angle density of radiant flux is sometimes given the shorthand name of (radiant) intensity. A combination of the two modes of measurement considers the little cones of radiation from each point in an area, called the phase density, measured in watts/m 2 ·sr. Unfortunately, there is one rema1n1ng complication iri this diagram. A distinction is made between the flow of radiant energy onto a surface and from a surface. When this is done, each of the three densities may have either of the two interpretations. The names and customary symbols for these measures are shown in the right-hand column of boxes. Each·pair of boxes is measured in the same units as given for the generating density; i.e., irradiance and radiant emittance are both examples of area density measures of the radiant flux. Photometry Now that there is no confusion as to how one measures radiation, it is time for another monkey wrench (Fig. 8). Because most scientists are human beings, and vision is an important part of our sensory world, a complete alternative scheme of measurement has been devised for visible light, based upon the sensitivity of the human eye. In order to measure visible light in units that correspond to our perception of the brightness of light, the photometric system was developed. Photometric units are based on the selective

Physics of Light PHOTOMETRIC CONCEPTS

19 (onto a surface)

(from a surface)

(onto a surface)

(from a surface)

(onto a surface)

(from a surface)

Fig. 8. Logical relationships of photometric concepts (after Preisendorfer 1976).

sensitivity of the light-adapted (as opposed to darkadapted) human eye to the various frequencies of visible light. Figure 8 is the photometric counterpart of the radiometric diagram (Fig. 7). The key concept is luminous flux, measured in lumens. Notice the modification of certain terms: illuminance replaces irradiance, luminous emittance instead of radiant emittance, and so forth. Properly speaking, all of these measures should be made with the light-adapted human eye. Use of other types of light sensors will introduce inaccuracies in proportion to their departure from the sensitivity curve of the human eye. An additional unit, the candela, is equal to one lumen/steradian, a measure of luminous intensity. Figure 9 compares radiant and luminous flux. The ordinate measures absolute luminosity, expressing the ratio of luminous flux to radiant flux. Hence, the ratio is a conversion factor with the units of lumens/watt, and is plotted on a log scale in Figure 9. The right-hand curve (labeled Photopic) shows the luminosity for the light-adapted human eye, and is the basis for the relation between the photometric and radiometric measures. Notice that since the eye is not equally sensitive to all frequencies of light, a conversion between the two measurement systems must take into account the frequency of visible light (the curves here are plotted on a wavelength scale). The general procedure is to break up the spectral curve of a radiant source into numerous

B. Dennis Sustare

20

IOOO

Scotopic

100

>. (/)

0

c

E

IO

:J ....J

...

Wavelength (nm)

Fig. 9. Absolute luminosity curves as functions of wavelength (response of the human eye to radiation of a given wavelength). Absolute luminosity is measured in lumens/ watt. Photometric units are based on the photopic response of the light-adapted human eye (after Barnes 1968). small intervals and apply the appropriate conversion factor at each of these intervals. The left-hand curve shows the luminosity for the darkadapted human eye under low light conditions; this scotopic response is dominated by the rod photoreceptors in our eyes. Note that the peak luminosity for the scotopic curve is located at a shorter wavelength (i.e., shifted toward the blue) than is the peak for the photopic curve. The photopic peak luminosity is at a wavelength of about 556 nm, whereas the scotopic peak is at about 511 nm. The peak luminosity for scotopic vision is greater than that for photopic vision

21

Physics of Light

because the standard light source used to define the lumen (the standard candela) radiates less energy in the bluegreen, where scotopic vision peaks, than in the yellowgreen, where photopic vision peaks. The difference in peak luminosity is not due to differences in sensitivity of the two types of vision, although the scotopic system is more sensitive than the photopic system. SUNLIGHT Solar Spectra The primary radiant source for all organisms is the sun. Figure 10 shows the spectral distribution of sunlight reaching the surface of the earth. The dotted line is a curve of black body radiation at a temperature of 5900°K, and is a very good approximation to the solar irradiance curve outside the earth's atmosphere. The solid line shows the irradiance of the sun as measured at sea level, after the atmosphere has absorbed certain portions of the spectrum .

,,""\

•20

, I

.15 (]) (.)

c ctl ·-

-o ctl ".....

.10

• 05

~ 00

.2

.4

------

.6

3.0

Wavelength (l!m)

Fig. 10. Spectral distribution of sunlight reaching the surface of the earth. Spectral irradiance measured in W·m-2·A-l. The dotted line represents the theoretical curve for a black body at 5900 K. Position of the visible spectrum is indicated (VIS) (after Barnes 1968).

B. Dennis Sustare

22

There are three molecules responsible for most of this absorption. Ozone absorbs strongly in the ultraviolet and part of the visible range. Absorption through the infrared is due mainly to water vapor and carbon .dioxide. Notice the concentration of solar irradiance in the visible spectrum. Sunlight's Earthly Fates In addition to absorption, particles in the atmosphere also scatter sunlight. Scattering is a complex phenomenon that depends on the refractive index and size of a particle, the frequency of light striking the particle, and the angle at which the particle is viewed. You may even have scattering from a nonabsorbing material when that material is in bulk form. The atmosphere scatters blue frequencies more strongly than red frequencies, which is why the sky appears blue. On a clear day at sea level about one-fifth of the total illuminance (remember, illuminance refers to the lightadapted human eye) of the earth's surface is from the sky rather than directly from the sun. When you look at the setting sun it appears red because the blue light has been scattered during the long passage through the atmosphere at a narrow angle above the horizon. Some animal colors are produced physically through scattering by small particles in feathers or other integumentary structures. Scattering produces the blue color of kingfishers (Alciedinidae) and blue tits (Parus caeruleus) (Auber 1957) as well as the blue in certain insect wings and in the cuticles of some other invertebrates (Gardiner 1972). Colors produced by scattering from integumentary particles do not change their characteristic frequency (and wavelength) with slight changes in the viewing angle. When the sun shines on water, the amount of visible light penetrating the water depends on the angle of incidence. A flat air-water interface reflects about 2% of the visible light that strikes perpendicular to the interface, with the percentage reflected increasing as the angle from the perpendicular increases. If the interface is ruffled by wind or waves, slightly more light is transmitted at every angle. Polarization Some other interesting phenomena may occur at the airwater interface, including polarization and interference.

Physics of Light

23

Electromagnetic radiation is transmitted as transverse waves; i.e., the displacement vector of the wave is normal to the axis of propagation (like. water waves, but unlike sound waves, which are longitudinal waves). If the displacement vector of the wave is always in the same plane, the wave is plane-polarized. Normally, radiation from a source such as a heated object has unpolarized waves, oriented in random directions normal to the axis of propagation. When a beam strikes a nonmetallic surface, the tendency for the radiation to be reflected or to penetrate the surface depends on the polarization of the radiation, resulting in a partial polarization of the reflected beam, and dependent on the angle of incidence of the beam. Waves whose displacement vector is parallel to the surface tend to be reflected; waves whose displacement vector is oriented otherwise tend to be absorbed. Certain materials (such as Polaroid sheet) are dichroic, that is, they selectively absorb visible light that is not in a "preferred" plane of polarization relative to the orientation of the dichroic material. These substances can be used to produce polarized, visible light. Numerous arthropods can see polarized light, and can use the directional information provided by the polarization produced by scattering and reflection of light in their environment. Interference When two radiation waves are superimposed slightly out of phase, they tend to cancel each other through the process of interference. Interference may also produce an adding effect, causing a local increase in light intensity when two light waves are in phase with one another. A thin film may cause interference by the action of light reflected from the first surface interfering with light reflected from the second surface. The thickness of the film must be compar~ able to the wavelength of the light for this to occur. Such interference produces the iridescence seen at the surface of soap bubbles. Interference between a transmitted ray and a doubly reflected ray at a thin film may produce similar results. One characteristic of thin film interference is that the frequency of interference depends on the angle of incidence of light (and thus the path length traveled through the film). This is why an oil film on a water surface changes color as you change your angle of view. Interference at the surface of the integument (or at the surface of pigment granules within the integument) is another source of animal colors, such as the iridescent or

24

B. Dennis Sustare

metallic colors of pigeons (Columbiformes), hummingbirds (Trochilidae), peacocks (Pavo eristatus), mallard ducks (Anas platyrhynchos) (Fox and Vevers 1960), and many insects (Gardiner 1972). Most of the blues and greens (the latter in combination with pigments) of bird feathers are produced by structural colors involving scattering or interference, rather than selective absorption by pigments. Water and Light Once the radiation passes the surface of the water, refraction has caused the rays to bend so that an underwater observer is effectively looking out of a manhole, with an angular radius of about 48°. Beyond this angle there is total internal reflection from the surface, according to Snell's Law. As the radiation descends through the water column, the infrared is absorbed very rapidly. Infrared comprises about half the total irradiance at sea level on a clear sunny day; this is essentially all absorbed by the first meter of water, falling off in an exponential fashion. Other frequencies of light show this exponential decay as well, though the decay rate is frequency-dependent. For natural waters the irradiance transmittance is usually maximal in the 460-500 nm range; i.e., blue light tends to penetrate the deepest in the water column. Transmittance falls off very rapidly at wavelengths above 580 nm, so that red light is filtered out very quickly. ZOOCHROMES AND PHYTOCHROMES Probably the key source of coloration for most plants and animals is pigmentation, the presence of materials that absorb in the visible range. Figure 11 shows absorption curves for two of the earth's key pigments, chlorophyll a and b. Each pigment shows two major peaks, one in the spectral region we call blue and one in the red. Failure to absorb in the green provides chlorophyll with its green appearance. The chlorophylls are related to several other classes of pigments that contain four pyrrole units, including porphyrins, heme pigments, cytochromes, and bile pigments (Wolken 1975). The location of major peaks for some other plant pigments is shown in Figure 11: B-carotene is a common carotenoid pigment found in many species (the carotenoids include the carotenes and the xanthophylls); phycoerythrin and phycocyanin are two algal pigments. There are numerous other pigments in the biological world,

Physics of Light

25 p- Carotene

z

0

1-

a..

a:

0

(/)

CD

0.7 kw·m- 2 ), a 1% change in a is usually equivalent to a several-percent change in E. On the other hand, when there is little sunlight (e.g., when T- Te = 20°K), a 1% change in E is equivalent to a several-percent change in a. However, this latter set of circumstances is usually unlikely or even impossible. That is to say, an ectothermic animal is not likely to achieve a very elevated body temperature without a large amount of absorbed radiant energy. Therefore, it would appear that under most natural circumstances a small change in solar absorptivity would be equivalent to a larger change in thermal emissivity, and thus natural selection should usually favor morphological adaptations related to a more strongly than similar adaptations related to s. It seems, therefore, that animals such as Liolaemus represent a very interesting exception

Emissivity, a Little Explored Variable

31

::..::: 0

~--~

I 1-

ABSORBED DIRECT SOLAR RADIATION

kwatts·m2

Fig. 12. The values of the partial derivative 3€/3a (see equation (1.8)): the steady-state energy-balance equation for a hypothetical ectothermic animal as a function of the direct solar radiation absorbed by the animal and the temperature difference between the animal and its environment. The combinations of environmental variables which give 3€/3a> 1 represent situations in which differences in solar absorptivity, ~. are more influential on the steady-state body temperature of the hypothetical animal than identical differences in the thermal emissivity, s. Combinations of environmental variables resulting in 3s/3a < 1 are situations in which differences in s are more influential on the energy balance than like differences in a.

to a logical rule, and must be exposed normally to unusual thermal environments.

32

C. Richard Tracy

ACKNOWLEDGMENTS I would especially like to thank 0. P. Pearson, who shared his ideas and a manuscript that was the focus of my discussion.

REFERENCES Gates, D. M. 1962. Energy Exchange in the Biosphere. New York: Harper. Norris, K. S. 1967. Color adaptation in desert reptiles and its thermal relationship. In Lizard Ecology: A Symposium, W. H. Milstead (ed.). Columbia, Mo.: Univ. of Mo. Press. Pearson, 0. P. 1977. The effect of substrate and skin color on thermoregulation of a lizard. Camp. Biochem. and Physiol. 58:353-358. Porter, W. P., and Gates, D. M. 1969. Thermodynamic equilibria of animals with environment. Ecol. Monogr. 39:245270. Porter, W. P.; Mitchell, J. W.; Beckman, W. A.; and DeWitt, C. B. 1973. Behavioral implications of mechanistic ecology. Oecologia 13:1-54. Tracy, C. R. 1972. Newton's Law: Its applicability for expressing heat losses from homeotherms. BioScience 22: 656-660 (erratum, 23:296).

Part2

Physiological Functions

of Animal Coloration

Chapter 2

The Influence of Color on Behavioral Thermoregulation and H ydroregulation David M. Hoppe Introduction Thermoregulatory and Hydroregulatory Behavior Reptiles Amphibians The Effect of Color on Absorption and Reflection Study Systems for Examining the Role of Color Metachrosis in Lizards Metachrosis in Frogs Amphibian Color Polymorphism Desiccation Experiments with Chorus Frogs Reflectance Spectra of Chorus Frogs Distributional Aspects of Color Polymorphism Adaptive Compromise

Thermoregulation and Hydroregulation

37

INTRODUCTION Animal coloration represents many evolutionary adaptations. The first consideration by those seeking to identify coloration's adaptive value has often been the visual appearance to other animals of pigmented skin, scales, feathers, or hair. Investigators have considered and extensively studied such phenomena as cryptic coloration, aposematic coloration, parasematic coloration, and other aspects of inter- and intraspecific communication (refer to Chapters 8 and 9) as possible forces affecting the evolution of animal color. But color has nonvisual effects, through physical and biochemical changes that result from absorption and reflection of direct solar radiation and indirect or secondary radiations. Pigmentation of the skin and its derivatives influences the quality and quantity of radiation absorbed, and may filter out harmful radiation (see Chapter 3), or affect body temperature or the rate of evaporative water loss. Where radiation, temperature, or moisture is a critical factor limiting animal populations, color could have considerable adaptive significance through these nonvisual phenomena. In this chapter, I review some of the evidence indicating that amphibians and reptiles behave so as to regulate body temperature or water content, or at least confine fluctuations within narrower limits than environmental conditions would otherwise dictate. I discuss the effect of color on some physical responses of organisms to solar radiation, and the behavioral responses that are affected as a consequence. I limit the discussion to responses of amphibians and reptiles, although some conclusions from my experiments with these ectotherms may apply to endothermal vertebrates as well. I consider the differential absorption of solar radiation by differently colored skin or scales, and I conclude by speculating on the adaptive significance of color in amphibians and reptiles from the standpoint of behavioral thermoregulation or hydroregulation, referring to studies of color change and color polymorphism in these vertebrate groups.

David M. Hoppe

38

THERMOREGULATORY AND HYDROREGULATORY BEHAVIOR Reptiles Thermoregulatory behavior is well documented for many species of reptiles, and one can sort through literally hundreds of papers in reviewing this phenomenon. Perhaps the most widespread thermoregulatory behavior is basking; organisms gain heat by prolonged exposure to solar radiation (heliothermic basking) or to warm substrates (thigmothermic basking). Heliothermic basking is of more interest here, since color affects the absorption of solar radiation. Basking enables some species to attain body temperatures well above ambient air temperatures, a phenomenon most pronounced in high-altitude populations such as an Andean smooththroated iguanid lizard (Liolaemus multiformis) (Pearson 1954, Pearson.and Bradford 1976) that attained body temperatures as high as 30°C above that of the ambient air. Vitt (1974) reported that many species of high-latitude lizards and snakes achieve body temperatures higher than that of the ambient air or substrate by basking. Norris (1967) demonstrated similar thermal phenomena with three species of high-altitude American lizards of the genus Sceloporus. The efficiency of basking can be increased by changes in the animal's posture or orientation with respect to incident radiation. For example, Brattstrom (1971) described several basking postures used by the bearded dragon (Amphibolurus barbatus), showed that these lizards orient themselves to maximize or minimize exposure to the sun or a heat lamp under experimental conditions, and demonstrated that such behavior is influenced by body temperature. Similarly, Cogger (1974) described orientation and posturing behavior involved in the basking of the mallee dragon (Amphibolurus fordi) and DeWitt (1967) showed that the desert iguana (Dipsosaurus dorsalis) adjusts its position with respect to incident radiation under laboratory conditions, thereby achieving quite precise regulation of its body temperature. Further modification of basking behavior was shown by Case (1976), whose studies indicate that endogenous triggers to basking have become integrated with seasonal behavior of the chuckawalla (Sauromalus obesus). He showed that springcollected animals readily basked, whereas fall-collected animals were hesitant to bask under the same conditions of temperature and light. This result suggests that endogenous controls allow for basking when increasing body temperature and metabolic rate is advantageous, but inhibit such behavior when the animal should be entering a winter torpidity.

Thermoregulation and Hydroregulation

39

Whereas basking and its modifications help to increase the body temperature, other action patterns are necessary to keep this temperature from reaching lethally high values under some circumstances. Mosauer (1936) reported that reptiles die from thermal stress due to insolation, not from ultraviolet radiation or desiccation. He also stated that a diurnal desert lizard cannot tolerate any more heat than a nocturnal snake, and therefore must adjust behaviorally to keep its body temperature down. Cole (1943) has confirmed Mosauer's finding that desert reptiles die from thermal stress when overexposed to sunlight. Some disagreement exists in the literature regarding thermal tolerance; however, as Cowles and Bogert (1944) state: "Contrary to previous reports, nocturnal reptiles not only tolerate but prefer temperatures lower than diurnal reptiles." Regardless of any comparison of diurnal and nocturnal forms in this respect, diurnal lizards (particularly desert species) can become lethally overheated in sunlight, and must behave so as to avoid this occurrence. One tactic that minimizes heat gain from the surroundings is to reduce the surface in contact with warm substrates. For example, the lizard Phrynocephalus mystaceus adjusts its posture at midday so that its legs hold its body well off the substrate (Kashkarov and Kurbatov 1930). The shovel-snouted lizard (Aporosaura anchietae) arches its body, which limits the substrate contact to a small portion of its ventrum (Brain 1962). Brattstrom (1971) and Cogger (1974) also describe postures which serve to minimize substrate contact. Avoidance behavior is another adaptation to unfavorable thermal environments. Shade-seeking, shelter-seeking, or burrowing, as in the reports of Cowles and Bogert (1944), Norris (1953), Brattstrom (1971), Cogger (1974), and Aleksiuk (1976), are behavioral patterns aimed at temperature regulation by retreating from unfavorable environmental conditions. By splitting their time between sunlight and shade, being in burrows or on the surface, and other such behavior, reptiles have been able to achieve rather precise control over their body temperatures. Thermoregulation can be assisted by heat-dissipating behavior, such as respiratory movements or panting (Cowles and Bogert 1944, Brattstrom 1971, Cogger 1974), or licking the lips to allow for evaporative cooling (Brain 1962). Further refinement of thermoregulatory behavior is seen in species such as the taipan (oxyuranus scutellatus), which appears to behave so as to achieve an even more precise regulation of head temperature than body temperature (Johnson 1975). Johnson described

40

David M. Hoppe

specific reptilian action patterns (e.g., gaping, elevating the head above the substrate, tucking the head under or inside a body coil, or burying the head in the substrate) that may precisely regulate the temperature of the head. Specific regulation of head temperature may be an important factor when considering differences between coloration of the head and body of some species. The significance of behavioral thermoregulation is further elaborated by studies of natural populations. Hamilton (1973) has reported that a large vegetarian lizard (Angolosaurus skoogi) spends most of its above-surface time on dunes, behaving so as to first increase body temperature, then stabilize it near 39°C, with only brief visits to feeding sites. These data suggest that the significance of thermoregulation may be in allowing for more efficient metabolic food processing rather than increasing activity capabilities of the whole organism for food-seeking and subsequent higher food intake. Cost/benefit ratios may be important factors to consider in studying the evolution of thermoregulatory behavior. For example, Huey (1974) reported that a Puerto Rican anolid lizard appears to behaviorally thermoregulate in open habitat, but tolerates decreased and more variable temperatures in forest habitat where fewer and more widely separated basking sites are present. Another tropical forest species, Anolis marmoratus, is also more eurythermal than most lizards, showing little thermoregulation (Huey and Webster 1975). Thus, the energy expenditure or increased exposure to predation in seeking the less abundant basking sites may make behavioral thermoregulation disadaptive. Amphibians Some amphibians are thought to use behavioral thermoregulation. That thermal preferences are present in this group of vertebrates can be seen even in the larval stages. Brattstrom (1963) has reported that tadpole aggregations are sometimes thermotaxic, and that larval salamanders may select specific, favorable water temperatures within ponds. Licht and Brown (1967) found that both larval and adult redbellied newts (Taricha rivularis) exhibit distinct thermal preferences. One report of thermoregulatory behavior among frogs is that of Lillywhite (1970), who studied the behavior of bullfrogs (Rana catesbeiana). He found that certain action patterns were correlated with air temperature, with

Thermoregulation and Hydroregulation

41

stronger correlations seen for juveniles than adults. Such behavior included shuttling in and out of shade, shuttling between water and terrestrial surfaces, basking in exposed areas, and changing between prostrate and sitting postures. Valdivieso and Tamsitt (1974) have also reported basking behavior in semi-arborial neotropical tree frogs (Hyla labialis) which resulted in body temperatures higher than ambient air temperatures. Western toads (Bufo boreus) exhibit basking as a means of increasing body temperature (Lillywhite, Licht, and Chelgren 1973). Such behavior is more pronounced among juveniles than adults, and may serve to maximize growth rate. Some Peruvian toads were also shown to increase body temperatures somewhat by basking, but not nearly as well as lizards of the same region (Pearson and Bradford 1976). The water-permeable skin of amphibians results in a problem not faced by the scaled reptiles, that of rapid desiccation. Thus there exists for amphibians a close relationship between thermal changes and body water balance. The resulting evaporative cooling may reduce or even eliminate the possibilities for amphibian thermoregulation. These relationships are pointed out in observations that "thermoregulating" toads gained heat inefficiently because they returned to shaded regions of lower temperature often, perhaps to rehydrate, but could be active during unfavorably hot, dry hours of the day, presumably due to evaporative cooling (Pearson and Bradford 1976). Similarly, an African toad (Bufo mauritanicus) may have adapted to hot, dry climates by losing water more rapidly and thus maintaining a lower body temperature through evaporative cooling (CloudsleyThompson 1969). The relationship of heat and water budgets is emphasized by the observation that heliothermic anurans are forms that remain close to permanent water (Brattstrom 1963). Tracy (1976) showed that there may be little thermoregulatory ability among leopard frogs. His computer simulations imply that under most natural conditions, these frogs would have a core temperature very close to the ambient air temperature, due to evaporative cooling, so thermoregulatory behavior, if present, would be very inefficient. My own work with chorus frogs, to be discussed later, confirms this observation. Since amphibians lose water freely, many of their action patterns may be aimed at conserving body water, which I consider to be hydroregulation. Bentley (1966) described hydroregulatory behavior of amphibians occupying arid Australian regions, including burrowing when conditions are dry

42

David M. Hoppe

and emerging in response to rain, and becoming cryptozoic by seeking refuge under litter and in soil cracks. Such amphibians may seek shade or moist soil, or may position themselves so as to contact more moist substrate or to contact it with the more permeable skin of the groin region. Leopard frogs subjected to desiccation experiments (Gillis and Hoppe 1975) initially sat erect and attempted to escape, and later in the trials assumed a tucked posture that reduced their surface area and put more ventral skin in contact with the substrate. Salamanders may seek refuge under leaf litter, and coil to reduce surface area. The bullfrogs in Lillywhite's (1970) trials, in splitting time between water and terrestrial surfaces, may have been regulating body water content. THE EFFECT OF COLOR ON ABSORPTION AND REFLECTION Different pigments or densities of pigment are known to absorb solar radiation at different rates and in different qualities or quantities. In this way, the color of an animal can affect its temperature and possibly its water dynamics, thus influencing behavioral thermoregulation or hydroregulation. Cole (1943) noted that a dark-colored animal is subject to more rapid heating than a lighter animal if exposed to the same intense radiation, and dark lizards overheated more rapidly than lizards of lighter color. He and other early investigators erroneously concluded, however, that the dark-colored animal also lost heat more rapidly than the light-colored animal by reradiating it, when sheltered from incident radiation. Norris (1967) detailed the reasons why a dark lizard is not necessarily a better emitter of radiation than a light one. This is a crucial point, since any advantage a dark animal holds in absorbing radiant energy would be lessened if it also emitted that energy more rapidly in shade or early evening hours, which would be the case if Cole's (1943) assumption were true. The thermal relationships of reptiles with respect to color have been well discussed by Norris (1967), and I refer the reader to that publication rather than reviewing the material here. Bartlett and Gates (1967) reported that the total energy gain of a lizard would be altered by about 4% with a 10% change in the lizard's absorptivity. While 4% may seem a small amount, they concluded, "It would be worthwhile to examine the spectral variations of both individuals and

Thermoregulation and Hydroregulation

43

ecotypes in these terms." I report the spectral properties of different color phenotypes of chorus frogs in a later section of this chapter. That reptiles have adapted evolutionarily to meet thermal problems by altering the spectral properties of their surfaces is indicated by some observations of Hutchison and Larimer (1960). They pointed out that heat gain of some lizards via absorption of solar radiation is correlated to habitat, desert species absorbing least and tropical forest species the most. Lower heat gain was seen with increasing aridity and temperature. They also found that the light ventral surfaces of desert species reflect more of the radiation being emitted and reflected by the substrate, which may be another color-related thermal adaptation. STUDY SYSTEMS FOR EXAMINING THE ROLE OF COLOR Two types of color variation among reptiles and amphibians provide systems for studying the influence of color on the behavioral regulation of temperature or water content. One is the capacity for color change by individuals, a phenomenon which has been called metachrosis by some authors. This can be a distinct change between two visible hues, such as from brown to green, or a change in the saturation of a color, such as from light brown to a darker brown. The other is color polymorphism, where individuals of different genetically determined colors coexist in the same population. Metachrosis in Lizards Pronounced color changes have been studied in species of both reptiles and amphibians, most of which are affected by environmental conditions in a predictable fashion. Some of these changes have been shown to affect the thermal biology and consequently the behavior of the organisms involved. The pioneering studies of Atsatt (1939) on sixteen species of desert lizards revealed that both geckos and iguanid lizards become lighter colored in response to high temperatures and darker when temperatures are decreased, which is adaptive for a temperature-regulating organism. Cogger (1974) reported that the melanophores of the mallee dragon (Amphibolurus fordi) expand to yield a darker color

44

David M. Hoppe

soon after emergence in the morning, increasing the efficiency of its basking. Norris (1967) also described cases in which lizards, after emerging and increasing their body temperatures by basking, become paler, some of them becoming "superlight," thereby greatly reducing their absorption of solar energy. These daily cycles of metachrosis appear to be adaptive strategies of thermoregulation among diurnal lizards, animals becoming darker in color when it is advantageous to absorb more light, and lighter when body temperature must be stabilized or decreased by reflecting more light. One might get into a "cause-vs.-effect" argument at this point. That is, the animal turns darker as an "effect" of lower temperatures (and perhaps changing intensities of illumination), so we should be careful of speaking teleologically in saying that the animal turns darker to "cause" greater absorption of radiant energy. Regardless of the viewpoint taken, such metachrosis seems to represent an evolutionary adaptation to modify the thermal changes occurring as a result of insolation. Ontogenetic color change could represent another level of thermoregulatory adaptation. Fitch (1955) described a color pattern metamorphosis in the Great Plains skink (Eumeces obsoletus). In this species the hatchling is usually jet black dorsally, changing to a golden brown after the first hibernation period, and eventually to a dull, speckled adult pattern varying from "grayish brown or yellowish brown to olive, with irregular black markings." The darker pigmentation of juveniles, combined with their higher ratio of surface area to volume, allows them more capacity for gaining heat through thermoregulatory behavior, and results in an increased rate of growth. Metachrosis in Frogs Considerable research has been carried out which examines the phenomenon of color change in amphibians, the earlier work being summarized by Parker (1948). From this summary, Edgren (1954) generalized that "frogs respond to cool dark environments by release of melanophorotropic hormone and darkening of integument, whereas light and warmth result in melanophore contraction and light colors . . . . " Anyone who collects or maintains frogs for research purposes has probably observed the darkening of the skin that occurs when the organisms are refrigerated. Both leopard frogs and chorus frogs exhibited this type of metachrosis in my own research activities. In some species, most

Thermoregulation and Hydroregulation

45

notably the tree frogs, environmental factors cause color changes between distinctly different hues. Edgren (1954), for example, noted that gray tree frogs (Hyla versicolor) remain brown under constant light conditions but turn green in constant dark, and that cold temperatures induce darkening of color. In fact, frogs with parts of their bodies exposed to different environmental conditions may show color differences between these body regions. For example, Porter (1972) described green frogs (Rana clamitans) which were bright green on portions of their bodies that were in direct sunlight, and a bronze or olive-green color where they were in the shade or under water. These amphibian color changes could play a thermoregulatory role in addition to contributing to the cryptic coloration of the animals involved. Experimental data quantifying the thermal effects of amphibian color change are not presently available, however. Amphibian Color Polymorphism Some populations of a variety of amphibian species exhibit color polymorphism. By examining such phenomena as seasonal changes in the frequencies of various color morphs, environmental correlates with color morph frequencies, or other physiological traits that may accompany color phenotypes, the adaptive significance of color might be further elucidated. One well-studied example of color polymorphism is the situation existing in Rocky Mountain populations of the boreal chorus frog (Pseudacris triseriata) (Matthews 1971, Hoppe 1975, Tordoff, Pettus, and Matthews 1976, Tordoff and Pettus 1977). Eight color phenotypes may occur in these populations; the dorsum can be brown, red, green, or redand-green, and the accompanying spots or stripes can be brown or green. Chorus frogs do not change from one color to another, although they become darker or lighter in response to environmental factors as do other amphibians and reptiles previously discussed. The relative advantages or disadvantages of different colors appear related to sunlight or visual phenomena, since these frogs feed diurnally during most of July and August, whereas nearby prairie populations are largely nocturnal and contain only the brown dorsal color phenotype (Spencer 1964, Hess 1969).

46

David M. Hoppe

Two types of experimental evidence, plus some observations regarding the dynamics of these mountain populations, contribute to my speculation regarding the nonvisual significance of color in chorus frogs. One experimental finding is that a green frog loses water more rapidly than a brown frog when exposed to direct sunlight. The other experimental finding is that green chorus frogs absorb and reflect radiant energy differently from brown frogs, in a manner that might account for their different desiccation rates. Desiccation Experiments with Chorus Frogs Two sets of desiccation experiments, designated dark and sunlight trials, were carried out in order to examine the role of color in the thermal and water dynamics of these frogs. The dark trials were performed under laboratory conditions similar to the method of Claussen (1969). The sunlight trials involved recording similar data, but the frogs were exposed to direct solar radiation during the experiments. Dark desiccation trials were carried out by placing frogs in wire screen cages in a 0.05 m3 oven at room temperature (21-23°C). Air was first pumped through a liter jar filled with Drierite (CaS0 4 ), then through the oven chamber from bottom to top at a rate of 2500 cm 3 /min. An enamel pan containing another 800 cm3 of Drierite was placed in the bottom of the oven, resulting in 10-12% relative humidity in the desiccation chamber. Frogs were acclimated for 24 hours at 18°C in petri dishes that contained filter paper saturated with distilled water to the point where a layer of standing water just covered the paper. Each frog was weighed before the trial, at half-hour intervals during the trial, and at the frog's critical activity point (CAP). Critical activity points were determined by using the combined criteria of Ray (1958) and Farrell and MacMahon (1969). These criteria are (1) loss of righting ability, (2) cessation of buccal movements, (3) dried appearance of the skin, and (4) ultimate survival of the frog upon immediate rehydration. Temperature and desiccation responses to sunlight were determined by modifying the above procedure somewhat. Adults were acclimated as described above, then fastened to a sheet of white styrofoam by means of strings tied around the knees. They were placed in direct sunlight on a flat

Thermoregulation and Hydroregulation

47

surface. Weight, dorsal surface temperature, and cloacal temperature were recorded initially, after 30 minutes, and at CAP. Dorsal surface and cloacal temperatures were measured electronically with flat surface and 20-gauge blunt needle thermistor probes, respectively. Air temperature, relative humidity, and wind velocity (which was negligible on the days selected) were also recorded. All trials were carried out between 1230 and 1400 Mountain Daylight Time (MDT) to standardize the angle of exposure to the sunlight. Desiccation data were pooled from different days during which the weather conditions were similar enough that overall desiccation means were not significantly different. Least-squares regression coefficients of desiccation rate on initial weight were calculated with phenotype as a covariant. Analyses of covariance were then performed to examine the role of color phenotype in determining desiccation rate. Changes in dorsal surface and cloacal temperatures of the frogs during the sunlight trials are shown in Table 1. Table 1 Temperature Changes of Adult Frogs in Sunlight Trials, in °C

Brown

Green Area Measured

N

X

30 minutes

14

CAP

+ S.E.

+

S.E.

t

N

X

-2.0 + 1.3

20

-0.9 + 1.0

0.696

13

+6.2 + 0.8

18

+6.4 + 0.6

0.159

30 minutes

14

-1.1 + 0.6

20

-0.7 + 0.6

0.471

CAP

13

+1.3 + 1.2

18

+1.9 + 0.5

0.472

I':. Dorsal

D. Cloacal

48

David M. Hoppe

Whereas the relatively large error factor in these data discourages rigorous analysis or interpretation, the data indicate that temperature response is probably not a major limiting factor for these frogs. The average environmental conditions to which they were exposed were temperatures of 26-29°C, relative humidities of 23-42%, and a dry substrate at a temperature of 29-33°C. But both surface and cloacal temperatures (initially 20-24°C) dropped by 1-2° during half an hour under these circumstances. Only when the animals reached CAP (which took 30-45 minutes under these circumstances) did these temperatures rise appreciably, and even at CAP the animals were usually cooler than the air and the substrate. In these trials, evaporative cooling was effective enough that frogs maintained temperatures below ambient, and would die of desiccation before encountering seriously high body temperatures. In fact, frogs which were not placed in water for rehydration immediately after reaching CAP did not recover. The final cloacal temperatures of these frogs ranged from 25-29°C. Since the critical thermal maximum (CTM) temperature of chorus frogs from these same populations, acclimated under similar circumstances, is reported as 37-38°C (Miller and Packard 1974), death must have resulted from desiccation in the sunlight trials discussed above, not from high body temperature. Water losses of the frogs during these sunlight trials are summarized in Table 2, expressed as percentages of initial weight lost. Tolerance to desiccation, as indicated by percentage of body weight lost at CAP, did not differ between green and brown frogs, in either adults or juveniles. However, green adults lost water more rapidly than browns, losing 29.3% of their body weight in 30 minutes compared with 26.5% in browns (p < 0.05). Similar differences were not detected in juvenile frogs. I suggest that the high surface area/volume ratio of juveniles allows for such a high relative rate of loss that any differences due to other factors are less likely to be expressed. Also, the pigmentation of juveniles may not be completely developed, a suggestion supported by the observation that colors become more distinct during the first few weeks of growth of these frogs. Table 3 contains similar data from adults and juveniles subjected to the dark, laboratory-chamber desiccation trials. Again, neither adults nor juveniles show different CAP values between the phenotypes. In these dark trials, differences in rates of water loss were not seen between green and brown frogs, juveniles or adults.

Table 2 Weight Loss Due to Desiccation of Frogs in Sunlight Trials, as Percentages of Initial Weight Green Category

Brown

N

X + S.E.

N

X+ S.E.

30 minutes

21

29.3 + 0.8

21

26.5 + 0.8

2.462*

CAP

13

36.1 + 0.6

18

35.4 + 0.6

0.768

20 minutes

12

21.9 + 0.4

12

21.1+0.4

1.536

CAP

12

38.3 + 0.9

12

38.9 + 0.9

0.479

t

Adults

Juveniles

*.E.< 0.05.

Table 3 Weight Loss Due to Desiccation of Frogs in Dark Trials, as Percentages of Initial Weight Brown

Green Category

N

X+ S.E.

N

X+ S.E.

t

Adults 1 hour

12

11.6 + 0.6

12

11.5 + 0.4

0.153

CAP

12

42.6 + 0.5

12

42.1 + 0.4

o. 774

1 hour

19

23.3 + 1.2

24

24.9 + 0.8

1.211

CAP

15

40.5 + 0.9

23

38.6 + 0.8

1.583

Juveniles

50

David M. Hoppe

;:,

;:,

;:,

;:,

c

;:,

s

;:,

0

~

"' 00

5l

.;:;

.~ ill

~

" ill

u

~

v

;:,

P.;

,6green

• brown Br

20.~~~--~----~--~--~--~----~--~--~--~1.23

3.27

Weight (g)

Fig. 13. Regression of adult weight loss on initial frog weight, sunlight trials.

Anurans can lose water at different rates due to their body size, as a consequence of surface area/volume relationships. Therefore, the desiccation data reported here have been plotted as functions of the frogs' weights (Figs. 13-16), in the format of Farrell and MacMahon (1969). The resulting linear relationships are similar to those of other hylids, as reported in the Farrell and MacMahon paper. F-statistics from analyses of covariance on each of these pairs of regression lines are listed in Table 4. (The roman numerals subscript to F in the two

Thermoregulation and Hydroregulation

51

Table 4 Analysis of Covariancea of Weight Loss Due to Desiccation of Frogs (Fig. 13-16)

F b

Trial

-I

Sunlight Adults

0.212

7.834**

Juveniles

0.060

3.956

Adults

0.087

0.089

Juveniles

1. 725

0.169

Dark

aTesting regressions of rate of weight loss of initial frog weight, with color phenotype as a covariant. equal.

bTesting the hypothesis that the two slopes are

cTesting the hypothesis that the two Y-intercepts are equal, given that the slopes are equal. **.E. < 0.01. right-hand columns refer to hypothesis r and hypothesis rr~ respectively.) The only significant difference indicated by these tests is between green and brown adults in the sunlight trials (F = 7.834, p < 0.01). The interpretation, as shown in Figure 13, is that a green frog of a given weight loses water more rapidly in sunlight than a brown frog of the same weight. These desiccation findings suggest a terrestrial disadvantage of green frogs compared with brown frogs. Since these montane frogs feed diurnally, exposure to sunlight results in relatively more water loss in green adults, which

52

David M. Hoppe 24.0

d

·s

6

IS.

0

~

6

m m

6

.8

:.::

Gr

.~

v

~

0 v

u

6

"ID

11.

Br

6. green • brown

l8.1L---L---~--~--~----~--~--~--~----~~

0.43

0.23

Weight (g)

Fig. 14. Regression of juvenile weight loss on initial frog weight, sunlight trials. may limit the distance traveled or time spent in feeding, or both. It may be argued that the frogs limit their activity to moist meadows, hence overcoming evaporative loss through continual water uptake. Two considerations tend to counter this argument. One is that the water balance model of Tracy (1976) suggests that under the desiccating circumstances of montane air temperatures, insolation, and wind, chorus frogs could not replenish their body water, even on water-saturated substrates. Secondly, as reported by Spencer (1964), these frogs migrate considerable distances across relatively dry terrain. Unless such

Thermoregulation and Hydroregulation

53

15 . 4

•

~green

Br

• brown

8.1L---~--_J----~--~----~--J---~~--L---~--_J 1.0

2.5

Weight

(g)

Fig. 15. Regression of adult weight on initial frog weight, dark trials.

migrations were completed during periods of rainfall or darkness, brown frogs would have a slight advantage in losing water more slowly while exposed to sunlight. Reflectance Spectra of Chorus Frogs Preliminary spectrophotometric data relate to the different desiccation rates of brown and green chorus frogs. Spectral reflectances of the dorsal surfaces of

54

David M. Hoppe 3 0. 7

6.

0.27 Weight (g)

Fig. 16. Regression of juvenile weight loss on initial frog weight, dark trials.

these color morphs were determined with a Beckman DK-2A spectroreflectometer for the wavelength interval of 350 to 2500 nm. The response curves of two brown and two green frogs are shown in Figure 17. As illustrated, the percentages of light reflected by the frogs' dorsal surfaces were generally lower for green frogs than for brown frogs, particularly in the spectral interval of 500 to 1000 nm. In other words, green frogs absorb more of the incident radiation than do brown frogs, green acting as a "darker" color in this respect. It must be pointed out that these

55

Thermoregulation and Hydroregulation 40 --brown ---green

-

......

/

..........

///-.__ ........ // ............ ......

// \ ' // \,_-;/

/

__

......... ..... ..... ........

..............

.... II '/'-

(lines converge)___..

500

1000

1500

Wavelength (nm)

Fig. 17. Reflectance spectra of brown and green chorus frogs.

animals' surfaces are nearly perfect absorbers and emitters of long-wave or thermal radiation, regardless of color (see Chapter 1). However, the wavelength interval over which these color morphs differ in their absorptive properties contains significant energy which is converted to heat on absorption, and could therefore affect the thermal and water budgets of the animals. Distributional Aspects of Color Polymorphism Some distributional data regarding amphibian color polymorphism relate to the role of color in the thermal or water dynamics of the species involved. Data from at least two montane populations of chorus frogs indicate that the green phenotype may be favored in the aquatic phase of the life cycle, but at a disadvantage terrestrially (Tordoff 1971, Hoppe unpublished data). The frequencies of green frogs in these populations appear to decrease throughout the summer, during the period of time when the frogs are feeding diurnally. Similar phenomena were reported in two other species of hylid frogs studied. In populations of Pacific tree frogs (Hyla regilla), green individuals are

David M. Hoppe

56

favored in the spring, whereas non-greens are favored by late summer (Jameson and Pequegnat 1971). Similarly, in cricket frog populations (Acris crepitans and A. gryllus), green frogs appear to be favored in spring and early summer, but grays are favored by fall (Pyburn 1961, Neva 1973). In Pacific tree frog populations again, the frequency of green frogs in West Coast populations is inversely proportional to aridity, relatively fewer green frogs being found in populations located in more arid regions (Resnick and Jameson 1963). All these distributional observations suggest that the increased solar absorptivity of small diurnal or semidiurnal frogs with a green dorsum may be placing them at an adaptive disadvantage in desiccating circumstances. Conversely, as long as they have adequate substrate moisture or standing water to replace evaporated water, these green frogs could possibly gain a thermal boost through more efficient basking. For at least two reasons, I hesitate to apply my speculations to green versus brown frogs generally. One is that surface/volume relationships may mask any color effect. Sunlight desiccation trials similar to my experiments with chorus frogs were run with leopard frogs, and no differences in desiccation rate were seen between brown and green frogs (Gillis and Hoppe 1975). With the lower surface/volume ratio of the larger leopard frogs, the differences in absorptivity of the dorsal surfaces of the color morphs may not alter the thermal or water balance significantly. The other reason is that a green frog is not necessarily a "green" frog. All frogs that the human eye perceives as green do not necessarily have the same combination of pigments or the same spectral properties. Schwalm, Starrett, and McDiarmid (1977), for example, reported that some frogs that appear green reflect nearinfrared light, whereas others do not, so the "green" color of these frogs may differ in its effect on their thermoregulatory or hydroregulatory abilities. ADAPTIVE COMPROMISE In oversimplifying some ectothermal color relationships, I have of necessity ignored a considerable volume of literature dealing with other effects of color, both from a visual and a nonvisual standpoint. The evolution of animal color in relation to thermal or water dynamics is interrelated with selection for cryptic coloration, aposematic coloration, parasematic coloration, protection

Thermoregulation and Hydroregulation

57

from harmful radiation, and other selection pressures dealt with in other chapters. The relative importance of these factors varies with the species or the particular environmental circumstances. ' The most obvious role of color and color change in several of the examples I have described is color-matching between the animal and its surroundings. For example, a number of species of desert lizards change color to match their surroundings. But among these, representatives of at least two genera tend to overshoot the ideal color-match by becoming lighter than their surroundings at very high temperatures (Norris 1953, Norris and Lowe 1964). Cowles and Bogert (1944) have pointed out the evolutionary compromise between cryptic color change and thermoregulation, noting that color change in diurnal desert lizards can extend their survival time when thermal conditions are marginal. Norris and Lowe (1964) have extended the consideration of adaptive compromise to discussing seven different phenomena which influence the degree of background colormatching in amphibians and reptiles. They have summarized that "background color-matching of color-labile reptiles in nature is most effective when the animal is within its activity temperature range, whereas the same animal is, to some extent, maladapted to concealing coloration during the warming period. Even during the warming period some radiation absorbing efficiency may be sacrificed to the maintenance of a degree of concealing coloration." The ideas of adaptive compromise and sacrifice lead to another important consideration when studying animal behavior. The interacting effects of color on the various problems faced by an animal may necessitate more behavioral adaptation. For example, if the animal must sacrifice cryptic color advantage to gain more heat, then behavioral predator avoidance becomes important. If selection for cryptic coloration results in unfavorable spectral properties with regard to thermal or water dynamics, the species may have simultaneously been selected for more efficient thermoregulatory or hydroregulatory behavior. ACKNOWLEDGMENTS I am grateful to David Pettus for his advice and encouragement related to the chorus frog experimentation reported here, and for the use of his laboratory facilities.

David M. Hoppe

58

Thanks also go to C. Richard Tracy for determining reflectance spectra for me, and to Barbara Hoppe for her many hours of assistance in the field and laboratory. The chorus frog desiccation data are part of a Ph.D. dissertation submitted to Colorado State University. REFERENCES Aleksiuk, M. 1976. Metabolic and behavioral adjustments to temperature change in the red-sided garter snake (Thamnophis sirtalis parietalis): an integrated approach. J. Thermal Biol. 1:153-156. Atsatt, S. R. 1939. Color changes as controlled by temperature and light in the lizards of the desert regions of southern California. Publ. Univ. Calif. Los Angeles Biol. Sci. 1:237-276. Bartlett, P. N., and Gates, D. M. 1967. The energy budget of a lizard on a tree trunk. Ecology 48:315-322. Bentley, P. J. 1966. Adaptations of amphibia to arid environments. Science 152:619-623. Brain, C. K. 1962. Observations on the temperature tolerance of lizards in the Central Namib Desert, South West Africa. Cimbebasia 4:1-5. Brattstrom, B. H. 1963. A preliminary review of the thermal requirements of amphibians. Ecology 44:238-255. Brattstrom, B. H. 1971. Social and thermoregulatory behavior of the bearded dragon, Amphibolurus barbaratus. Copeia 1971:484-497. Case, T. J. 1976. Seasonal aspects of thermoregulatory behavior in the chuckawalla, Sauromalus obesus (Reptilia, Lacertila, Iguanidae). J. Herpetol. 10:85-95. Claussen, D. L. 1969. Studies on water loss and rehydration in anurans. Physiol. Zool. 42:1-14. Cloudsley-Thompson, J. L. 1969. Water relations of the African toad, Bufo mauritanicus. Br. J. Herpetol". 5:425426.

Thermoregulation and Hydroregulation

59

Cogger, H. G. 1974. Thermal relations of the mallee dragon, Amphibolurus fordi (lacertilia: Agamidae). Aust. J. Zool. 22:319-339. Cole, L. C. 1943. Experiments on toleration of high ternperature in lizards with reference to adaptive coloration. Ecology 24:94-108. Cowles, R. B., and Bogert, C. M. 1944. A preliminary study of the thermal requirements of desert reptiles. Bull. Amer. MUS• Nat. Hist. 83:265-296. DeWitt, C. B. 1967. Precision of thermoregulation and its relation to environmental factors in the desert iguana, Dipsosaurus dorsalis. Physiol. Zool. 40:49-66. Edgren, R. A. 1954. Factors controlling color change in the tree frog, Hyla versicolor Wied. Proc. Soc. Exp. Biol. Med. 87:20-23. Farrell, M.P., and MacMahon, J. A. 1969. An eco-physiological study of water economy in eight species of tree frogs (Hylidae). Herpetologica 25:279-294. Fitch, H. S. 1955. Habits and adaptations of the Great Plains skink (Eumeces obsoletus). Ecol. Monogr. 25:5983. Gillis, J. E., and Hoppe, D. M. 1975. Variation and possible adaptive significance of shape in Rana pipiens and R. blairi. Bull. N. Mex. Acad. Sci. 15:51. Hamilton, W. J., III. 1973. Life's Color Code. New York: McGraw-Hill. Hess, J. B. 1969. Changes in frequency of the green-spot phenotype in piedmont populations of the chorus frog. Ph.D. dissertation, zoology, Colo. State Univ. Hoppe, D. M. 1975. Selection pressures affecting color polymorphism in montane populations of the boreal chorus frog, Pseudacris triseriata. Ph.D. dissertation, zoology and entomology, Colo. State Univ. Huey, R. B. 1974. Behavioral thermoregulation in lizards: importance of associated costs. Science 184:1001-1003.

60

David M. Hoppe

Huey, R. B., and Webster, T. P. 1975. Thermal biology of a solitary lizard: Anolis marmoratus of Guadeloupe, Lesser Antilles. Ecology 56:445-452. Hutchison, V. H., and Larimer, J. L. 1960. Reflectivity of the integuments of some lizards from different habitats. Ecology 41:199-209. Jameson, D. L., and Pequegnat, S. 1971. Estimation of relative viability and fecundity of color polymorphisms in anurans. Evolution 25:180-194. Johnson, C. R. 1975. Head-body thermal control, thermal preferenda, and voluntary maxima in the taipan, Oxyuranus scutellatus (Serpentes: Elapidae). Zool. J. Linn. Soc- 56:1-12. Kashkarov, D., and Kurbatov, V. 1930. Preliminary ecological survey of the vertebrate fauna of the Central KaraKum Desert in West Turkestan. Ecology 11:35-60. Licht, P., and Brown, A. G. 1967. Behavioral thermoregulation and its role in the ecology of the red-bellied newt, Taricha rivularis. Ecology 48:598-611. Lillywhite, H. B. 1970. Behavioral temperature regulation in the bullfrog, Rana catesbeiana. Copeia 1970:158-168. Lillywhite, H. B.; Licht, P; and Chelgren, P. 1973. The role of behavioral thermoregulation in the growth energetics of the toad, Bufo boreas. Ecology 54:375-383. Matthews, T. C. 1971. Genetic changes in a population of boreal chorus frogs (Pseudacris triseriata) polymorphic for color. Amer. Midland Naturalist 85:208-221. Miller, K., and Packard, G. C. 1974. Critical thermal maximum: ecotypic variation between montane and piedmont chorus frogs (Pseudacris triseriata, Hylidae). Experientia 30:355-356. Mosauer, W. 1936. The toleration of solar heat in desert reptiles. Ecology 17:55-56. Nevo, E. 1973. Adaptive color polymorphism in cricket frogs. Evolution 27:363-367.

Thermoregulation and Hydroregulation

61

Norris, K. S. 1953. The ecology of the desert iguana, Dipsosaurus dorsalis. Ecology 34:265-287. Norris, K. S. 1967. Color adaptation in desert reptiles and its thermal relationships. In Lizard Ecology: A Symposium, W. H. Milstead (ed.). Columbia, Mo.: Univ. Mo. Press. Norris, K. S., and Lowe, C. H. 1964. An analysis of background color matching in amphibians and reptiles. Ecology 45:565-580. Parker, G. H. 1948. Animal Color Changes and Their Neurohumors. Cambridge: Cambridge Univ. Press. Pearson, 0. P. 1954. Habits of the lizard Liolaemus multiformis multiformis at high altitudes in southern Peru. Copeia 2:111-116. Pearson, 0. P., and Bradford, D. F. 1976. Thermoregulations of lizards and toads at high altitudes in Peru. Copeia 1976:155-170. Porter, K. R. 1972. Herpetology. Philadelphia: Saunders. Pyburn, W. F. 1961. The inheritance and distribution of vertebral stripe color in the cricket frog. In Vertebrate Speciation, W. F. Blair (ed.). Austin: Univ. Texas Press. Ray, C. 1958. Vital limits and rates of desiccation in salamanders. Ecology 39:75-83. Resnick, L. E., and Jameson, D. L. 1963. Color polymorphism in Pacific tree frogs. Science 142:1081-1083. Schwalm, P. A.; Starrett, P. H.; and McDiarmid, R. W. 1977. Infrared reflectance in leaf-sitting neotropical frogs. Science 196:1225-1226. Spencer, A. W. 1964. The relationship of dispersal and migration to gene flow in the boreal chorus frog. Ph.D. dissertation, zoology, Colo. State Univ.

62

David M. Hoppe

Tordoff, W. 1971. Environmental factors affecting gene frequencies in montane populations of the chorus frog, Pseudacris triseriata. Ph.D. dissertation, zoology, Colo. State Univ. Tordoff, W., and Pettus, D. 1977. Temporal stability of phenotypic frequencies in Pseudacris triseriata (Amphibia, Anura, Hylidae). J. Herpetol. 11:161-168. Tordoff, W.; Pettus, D.; and Matthews, T. C. 1976. Microgeographic variation in gene frequencies in Pseudacris triseriata(Hylidae). J. Herpetol. 10:37-42. Tracy, C. R. 1976. A model of the dynamic exchanges of water and energy between a terrestrial amphibian and its environment. Ecol. Monogr. 46:293-326. Valdivieso, D., and Tamsitt, J. R. 1974. Thermal relations of the neotropical frog Hyla labialis (Anura: Hylidae). Roy. Ont. Mus. Life Sci. Contrib. 26:1-10. Vitt, L. J. 1974. Body temperatures of high latitude reptiles. Copeia 1974:255-256.

Further Thoughts on Anuran Thermoregulation: Discussion

C. Richard Tracy

Dave Hoppe's discussion of the role of color in regulatory processes (i.e., thermoregulation and/or hydroregulation) is cryptically complex. As Dave has suggested, it is extremely important to view any one regulatory process in terms of competition with other processes. Consider the concept of thermoregulation in amphibians. Figure 18 illustrates a simulation of the body temperatures of a frog, Rana pipiens (using the model of Tracy 1976), and a lizard, Sceloporus undulatus (using the model of Porter et al. 1973), as a function of time during an idealized day at 40° N latitude at an altitude of 250 m where the maximum air temperature during the day is 26°C (see Porter et al. 1973 and Tracy 1975 for examples of this sort of simulation). The frog and lizard are each simulated as having two solar absorptivities, 0.65 and 0.85. Notice that the body temperatures of the lizard are remarkably influenced by a 20% difference in solar absorptivity, whereas the body temperatures of the frog differ by less than one degree Celsius. The reason for these differences in response to the 20% increase in absorbed radiation lies in the fact that increased radiation absorbed by the lizard is available to become stored energy, which changes the animal's body temperature (Porter et al. 1973), whereas increased radiation absorbed by the frog is used in vaporizing water from the frog's mucus-covered skin (Fig. 19), and thus, the frog's body temperature is very little influenced

64

C. Richard Tracy

70 60

u

0

50

~

~

0:::

w a..

40

~

~ 30

w 0:::

8 20 10 0

0

3

6

12

15

18

21

24

HOUR OF DAY

Fig. 18. Model simulation of body temperatures of a lizard, Sceloporus undulatus, and frog, Rana pipiens, as a function of time of day for an idealized clear and warm day. The animals were each modeled as having solar absorptivities of 0.85 and 0.65. The frog model is described in Tracy (1976), and the lizard model is described in Porter et al. (1973). by increased radiant input (Tracy 1976). It is also important to notice that the 20% increase in solar absorptivity elicited an increase in evaporation rate of more than 1.2 g·hr- 1 , which is enough to dissipate energy at a rate of more than 0.81 watt. The difference between lizards and frogs supports the hypothesis that selection for an ability to precisely control body temperature is strong in lizards and weak in frogs (Tracy 1975, 1976). The hypothesis predicts that lizards should generally regulate body temperature precisely and function best within a narrow range of temperatures. On the other hand, frogs should not regulate body temperature precisely and should function best over a wide range of body temperatures. These predictions have been tested by prodding s. undulatus and R. pipiens until they

Further Thoughts on Thermoregulation

65

.15

.Ec .....

01

v-i Vl

9

.10

0:::

w ~ ~

w > .05

~ 0::: 0

!1.

g;

w 00

3

6

9

18 15 12 HOUR OF DAY

21

24

Fig. 19. Model simulations of the evaporative losses for the frog in Fig. 18.

"escaped" by running (lizard) or hopping (frog) away. s. undulatus was able to sprint fastest (Fig. 20) over a very narrow range of body temperatures, whereas R. pipiens was able to hop farthest (Fig. 21) over an extremely broad range of body temperatures. These data suggest that, in lizards, natural selection has produced physiological mechanisms for efficient bodily operation over a narrow range of body temperatures within which the lizard presumably maintains itself by behavioral and physiological thermoregulation. The frog, on the other hand, seems equipped to function equally well over a broad range of body temperatures which could be maintained by avoiding extreme environments. The conclusion is supported by the fact that body temperatures of frogs in the field (where the data were taken over a period of several days) span a wide range of temperatures (Brattstrom 1963). Hence the concept of thermoregulation in frogs may be one that needs reevaluation in light of the other regulatory processes that may operate simultaneously in this animal group (Tracy 1975, 1976).

66

C. Richard Tracy

150 (10)

aJ ~ 100

(14)

u

·a.c

50

If)

Body Temperature,

Fig. 20.

•c

The rate at which Sceloporus sprinted as a function of body temperature. Seven lizards were chased no more than two times each across a gridded "runway" one meter long. The sprints were recorded on super-eight motion picture film, and later transduced into a sprint speed by noting the number of picture frames that were necessary to photograph the lizard transversing the runway. The camera speed was calibrated to a digital stopwatch readable to the nearest 0.01 s. undulatus

67

Further Thoughts on Thermoregulation

30

(36)

(36)

(33)

(33)

T 25 E 20 u

g:

I c

(12)

15

m

2 10

5

Body Temperature,

oc

Fig. 21.

The mean distance hopped by Rana as a function of body temperature. Each frog was induced (by tapping the hind foot) to jump three hops at each body temperature.

pipiens

ACKNOWLEDGMENTS I woul~ like to thank Keith Cristian, Steve Waldschmidt, and Bobbi Tracy, who helped collect the empirical data presented in my figures.

C. Richard Tracy

68

REFERENCES Brattstrom, B. H. 1963. A preliminary review of the thermal requirements of amphibians. Ecology 44:238-255. Porter, W. P.; Mitchell, J. W.; Beckman, W. A.; and DeWitt, C. B. 1973. Behavioral implications of mechanistic ecology. Oecologia 13:1-54. Tracy, C. R. 1975. Water and energy relations of terrestrial amphibians: insights from mechanistic models. In Perspectives in Biophysical Ecology, D. M. Gates and R. Schmerl (eds.). New York: Springer-Verlag, pp. 325346. Tracy, C. R. 1976. A model of the dynamic exchanges of water and energy between a terrestrial amphibian and its environment. Ecol. Monogr. 46:293-336.

Maximization of Reproduction: Discussion

William]. Hamilton I II

My view differs from Dick's (above) because my basic perception is that of an ecologist and field biologist. I see a thread of continuity that runs throughout lizards, across insects that I have studied, and across certain amphibians. The behavior of some insects is understandable only in terms of thermoregulation. I refer to species that maintain the body's core temperature between 38°C and 42°C, for example, many diurnal Namib Desert tenebrionid beetles (Hamilton 1971, 1973). This may not apply to the montane chorus frogs discussed by David Hoppe, but we must address a basic question: Why did the frog come out of the water in the first place? Why is the frog sitting on the edge basking when it could be in the water? I think maximization of throughput and reproductive capacity through behavioral thermoregulation offers a viable explanation. REFERENCES Hamilton, W. J., III. 1973. Life's Color Code. New York: McGraw-Hill. Hamilton, W. J., III. 1971. Competition and thermoregulation behavior of the Namib Desert tenebrionid beetle genus Cardiosis. Ecology 52:810-822.

Audience Questions: Discussion

Question:

chorus frogs?

What are the origins of polymorphism in

Hoppe: There is little evidence that allows speculation as to the origins of polymorphism in chorus frogs. The literature, personal communications, and my own observations indicate that color-polymorphic populations of chorus frogs are not unique to the Colorado Rockies, but have been reported from Wyoming, North Dakota, Minnesota, Wisconsin, and Michigan, to cite a few populations. The widespread distribution suggests that whatever environmental factors have exerted selection on color phenotypes in the montane populations that I have discussed may have also played a role in the phylogenetic history of other populations. However, I think that speculation on the origins of color polymorphism must await an understanding of the maintenance of balanced polymorphism under present environmental conditions. Question: Are color differences in chorus frogs correlated with sex or age? Hoppe: Color differences among chorus frogs are not correlated with sex. The published genetic model. (Matthews and Pettus 1966) indicates that dorsal coloration is controlled by three pairs of alleles at three autosomal loci; hence sex-linked inheritance is ruled out. In addition,

Audience Questions

71

the frequencies of differently colored phenotypes among males and females suggest no influence of sex on the expression of color genotypes. Age-related differences in the frequencies of differently colored phenotypes may be found as a consequence of the seasonal selection to which I have alluded in Chapter 2. For example, in several populations a higher frequency of greens was found among juveniles than among adults. I have suggested (Chapter 2) that this may be a consequence of terrestrial selection against greens under some circumstances, since they desiccate more rapidly in sunlight than do brown frogs. There is no ontogenetic color change. The individual's coloration develops during the first few days after metamorphosis and remains unchanged throughout adulthood. Question: Do similarly colored frogs pair more frequently than differently colored frogs? Hoppe: No. There appears to be no assortative mating based on color phenotypes. This has been verified by analyses of the color combinations of breeding pairs of frogs captured in amplexus (Matthews 1971). Furthermore, these frogs breed nocturnally, so color perception is unlikely. Question: What is the activity cycle of these frogs? Are they nocturnal or diurnal? Hoppe: These montane frogs breed nocturnally during late May and most of June, then migrate to feeding meadows, where they are active diurnally during July and August. As I mentioned in Chapter 2, the nearby piedmont populations (in which green- and red-color phenotypes are not found) feed and breed primarily at night. This is one reason for my suggestion that the selection factors that maintain color polymorphism in the montane population relate to daylight conditions. Question: Considering that animals absorb and emit long-wave thermal radiation at pretty much the same high rates (with emissivities and absorptivities of ess~ntially unity), do you think that the reflectance differences you have shown between color phenotypes, which are pronounced only in the visual and near-infrared spectrum, significantly affect the thermodynamics of these frogs?

72

Audience Questions

Hoppe: At the moment I cannot explain why the differently colored phenotypes desiccate at different rates in sunlight but at the same rates in the dark, except as a result of differential absorption of solar radiation. Light with wavelengths between 500 and 1000 nm represents energy that is converted to heat on absorption (Chapter 1) and that heat appears to increase the rate of evaporation from the frog's skin.

Consider also that chorus frogs in their natural surroundings are not exposed primarily to direct sunlight, but more importantly to light reflected and transmitted by the vegetation around them. In such light, visual wavelengths represent a much higher proportion of the energy than in direct sunlight, since much of the thermal radiation is absorbed by the surrounding vegetation. Question: The fact that the absorption curves for green and brown chorus frogs diverge in the visible range and converge in the near and middle infrared region of the solar spectrum suggests to me that the adaptive significance lies in the color perception of predators, not in thermal adaptations. Otherwise one might expect that the infrared and unseen part of the spectrum would remain divergent. These animals have the potential to adapt to differences in the infrared, but have failed to do so. Hoppe: The hypothesis of differential predation based on color-matching of frog and substrate has been tested by Tordoff (1971). He considered robins and gray jays to be the major predators, and tested their behavior when frogs were made available on substrates of different colors. The jays hunted by perching and scanning the substrate, and did prey more heavily on phenotypes that contrasted with the background color. The robins, however, hopped around on the substrate flushing the frogs, finding frogs when they moved rather than by their contrasting color.

The problem with selective predation as an explanation of color polymorphism is the lack of correlation between phenotypic frequencies and habitat heterogeneity. For example, if one population had twice as many green frogs as another population, the first population should have correspondingly more green background in its habitat, but such correlations have not been found. In fact, of the breeding populations studied by Tordoff, the pond with the highest frequency of greens (about 20%) is separated by only 490 meters from the one with the lowest frequency of greens

Audience Questions

73

(less than 1%), and the two populations share much of the same feeding habitat. As to your statement that these frogs "have the potential to adapt to differences in the infrared," I assume this is based on reports that a few species of tropical tree frogs show differential reflectance in the infrared region of the spectrum. Richmond (1960) has pointed out that major climatic changes have occurred within the last 6000 years on the east slope of the Colorado Rockies. The montane populations of chorus frogs that I have studied have probably existed as semi-isolated entities under conditions approximating the present climate only 4500-6000 years. Conversely, the tropical tree frogs mentioned above may have had hundreds of thousands of years of relatively uniform conditions in which to evolve. Given this time factor, I consider it remarkable that the reflectance curves of different montane, color phenotypes differ as much as they do into the near-infrared spectrum, and see no "failure" to adapt in such a direction. I do not deny that crypticity may be part of the adaptive significance of color polymorphism in chorus frogs. However, I think that no single selection pressure (e.g., color perception by predators, desiccating environmental conditions, or solar radiation for absorptive heat gain) can be paired with habitat heterogeneity to maintain this color polymorphism. Instead there are numerous selection pressures that change in intensity and relative importance both seasonally and sporadically, with the vagaries of montane weather and climate. I have emphasized the factors related to the absorption of solar radiation and deemphasized an array of other considerations in an attempt to stick to the topic of thermoregulatory and hydroregulatory behavior. Question: Schwalm et al. (1977) demonstrate that the green color of some green frogs differs in the infrared region from the green coloration of the habitat. Thus colors that appear the same may not have the same reflectance spectra. Have you a comment? Hoppe: I have alluded to those data in the previous answer and have discussed them in Chapter 2. The paper you have cited makes several points that arise in this symposium also. For one thing, we are throwing the term "color" around rather loosely. Are we speaking of a particular wavelength of light, a human perception, a biochemical

74

Audience Questions

pigment, or the printed appearance of exposed film of some type? Dr. Gruber (Chapter 5) will discuss the perception of color in more detail and Dr. Sustare (Chapter 1) has discussed the relationship between visible and other wavelengths of electromagnetic radiation. Schwalm's paper emphasizes that we cannot and must not generalize about "green vs. brown," or even "green frog vs. brown frog," based on our own perception of color. One should test the absorptive and reflective properties of the skin of some perceived "color" before speculating too freely on the adaptive significance of that color. REFERENCES Matthews, T. C. 1971. Genetic changes in a population of boreal chorus frogs (Pseudacris triseriata) polymorphic for color. Amer. Midland Naturalist 85:208-221. Matthews, T. C., and Pettus, D. 1966. Color inheritance in Pseudacris triseriata. Herpetologica 22:269-275. Richmond, G. M. 1960. Glaciation of the east slope of Rocky Mountain National Park, Colorado. Geol. Soc. Amer. Bull. 71:1371-1382. Schwalm, P. A.; Starrett, P. H.; and McDiarmid, R. W. 1977. Infrared reflectance in leaf-sitting neotropical frogs. Science 196:1225-1226. Tordoff, W. 1971. Environmental factors affecting gene frequencies in montane populations of the chorus frog, Pseudacris triseriata. Ph.D. dissertation, zoology, Colo. State Univ.

Chapter 3

Tips on Wings and Other Things Edward H. Burtt,]r. Introduction The Coloration of Warblers The Topography of a Warbler The Munsell Color System Measurement of Reflection and Transmission Spectra Abrasion Resistance Statement of the Hypothesis Past Evidence of Differential Wear Tactics for Evaluation Experiments on the Abrasion of Warbler Feathers Methods Results Discussion The Topography of Abrasion-Resistant Coloration Abrasion by Airborne Particles Abrasion by Airborne Particles in Flying Birds The Dorsum The Tail The Remiges The Observed Distribution of Abrasion-Resistant Colors on Warblers Conclusions Regarding Abrasion Resistance Protection from Ultraviolet Radiation Statement of the Hypothesis The Molecular Basis for Damage from Ultraviolet Radiation The Potential for Ultraviolet Damage to Animals Protection from Ultraviolet Radiation Tactics for Evaluation

Reflection and Transmission of Ultraviolet Radiation Methods Results Discussion Topography of Ultraviolet-Resistant Coloration Predicted Patterns of Coloration and Behavior Observed Patterns of Coloration and Behavior Coloration Behavior Methods Results External Coloration as a Defense Against Ultraviolet Radiation: Evaluation of the Hypothesis

Tips on Wings and Other Things

77

INTRODUCTION Look at a monarch butterfly or a scarlet tanager and at once you are struck by their brilliant colors and bold patterns. Perhaps you have wondered why these and many other animals are brilliantly and boldly colored? Perhaps you have wondered what information is broadcast by these colors and patterns of color, or when and under what conditions the optical signal is broadcast, or for whom the signal is intended; but have you ever wondered about the noncommunicative functions of pigments, functions that are independent of the pigment's absorption, reflection, or transmission of visible light? In this chapter I evaluate the hypotheses that differently pigmented feathers resist abrasion differently, that the abrasion-resistance of differently colored feathers accounts for the general pattern of color of wood-warblers (Parulidae), and that external coloration protects underlying tissue from the potential damage of ultraviolet radiation. The hypothesized functions of color predict the behavior of differently colored woodwarblers. The predictions are tested by comparative study of the behavior and coloration of wood-warblers. THE COLORATION OF WARBLERS The Topography of a Warbler I divided the warbler's body into twenty-two regions (Fig. 22). The eyebrow-stripe, eye ring, eyeline, and whisker are referred to collectively as the face. The nape, collars, and throat comprise the neck. The dorsum includes the back, rump, and upper tail coverts and the venter includes the breast, belly, and under tail coverts. The Munsell Color System I determined the coloration of male and female woodwarblers in nuptial plumage by using the Munsell color system in direct sunlight. The Munsell system is a spherical array of colored paper samples. The three dimensions are hue, value, and chroma. Ten hues are arranged in spectral order around the equator of the sphere. Value as measured along a central axis grades from black at one pole

Edward H. Burtt, Jr.

78 Eye ring Eyebrow stripe Eye I ine

Mondibles:--~~~~:t'" upper lower

---.~=-Toil spots

Tail

Fig. 22. Topography of warbler. Burtt, in press.)

(From

to white at the other pole. The proportion of white, called the chroma, is measured along the radii of the equatorial plane from a maximum at the central achromatic axis to a minimum at the perimeter. Chroma approximates saturation. Every color can be specified by a numerical designation representing hue, value, and chroma. I used the Munsell color system to determine the coloration of the 22 body regions of 115 species of Parulidae. The Munsell system recognizes ten hues: red, orange, yellow, yellow-green, green, blue-green, blue, purple-blue, purple, and red-purple. I followed the Munsell system except that I divided red into chestnut, red, and brown. If the color's chroma was less than two, then the color was close to the neutral (white to black) axis; and I called it white if its value was nine or above, black if its value was two or below, and gray if its value was between two and nine. I determined the hue of each of the nineteen feathered regions of the warbler. The chroma of the upper and lower mandibles and the legs was usually two or less. I categorized these three unfeathered regions by color value exclusively.

Tips on Wings and Other Things

79

Measurement of Reflection and Transmission Spectra The Beckman DK-2A spectroreflectometer determines reflectance or transmittance of a sample by comparing the reflection from a sample surface to the reflection from a reference surface. Reflectance is measured by comparing the reflectance spectrum of a white (BaS04) surface with the reflectance from a sample, such as a feather. To measure transmittance both the sample and the reference beams are reflected from BaS0 4 blanks. The specimen is mounted in the path of the sample beam and reflection from the blanks is measured and compared. For a more detailed explanation, see Burtt (in press). ABRASION RESISTANCE Statement of the Hypothesis Knowledge of the structure of pigments and the effect of pigments on the structure of feathers makes it possible to predict the relative extent of damage to differently colored feathers. Melanin is a granular pigment whose granules are deposited in dense layers between layers of keratin. Furthermore, the deposition of melanin in the barbs and barbules induces increased keratin formation (Voitkevich 1966). Carotenes and xanthophylls are diffuse pigments and in moderate concentrations, such as found in warbler feathers, carotenes and xanthophylls have little effect on the structure of barbs and barbules (Brush and Siefried 1968). White in warblers is due to the reflection and scattering of all wavelengths from closely packed, unpigmented fibers in barbs and barbules. Only in feathers that contain melanin is there reason to expect increased resistance to abrasion. Therefore, if abrasion resistance is different in differently colored feathers, Prediction 1: melanin-impregnated feathers will be the most abrasion resistant. Past Evidence of Differential Wear Dwight (1900) appears·to have been the first to associate differential wear with differently colored barbs, although much earlier Bachman (1839) concluded that birds change color without molting, simply by wearing away the

80

Edward H. Burtt, Jr.

differently colored edges of the feathers. For example, the black bib of the male house sparrow (Passer domesticus) is revealed in the spring when the buff edges of its throat feathers wear away. Averill (1923) observed that the white barbs of the recently shed primaries of gulls were worn away whereas the black barbs, although worn, remained intact. The observation was replicated by Test (1940) in the flight feathers of common flickers (Colaptes auratus) and by Bowers (1959) in the wrentit (Chamnea fasciata). Reinroth and Heinroth (1958) picture the molted primary of a peregrine falcon (Falco peregrinus) showing the worn appearance of the white barbs as contrasted to the unworn appearance of the black barbs. These examples show that black feathers resist abrasion better than white feathers. Because melanin is the black pigment, the implication is that feathers containing melanin are more resistant to abrasion than feathers that lack melanin. However, the examples fail to mention colors other than black and white, and they fail to assess the quantitative difference between black and white. Tactics for Evaluation The hypothesis that differently colored feathers resist abrasion differently was tested by subjecting feathers of different colors to a measured amount of abrasion. The effect of such abrasion was quantified and the results cornpared for feathers of different colors. How are the feathers of living birds abraded? I discuss abrasion that results from airborne particles colliding with the feathers during flight. From this discussion emerge predictions of the location of the most intense abrasion and the areas most in need of abrasion-resistant coloration. Experiments on the Abrasion of Warbler Feathers Methods The outermost left tail feather was plucked from warblers captured in mist-nets at Itasca, Minnesota, during the summers of 1973 and 1974 and from warblers killed at television transmitting towers in Madison, Wisconsin, on 26-28 September 1973. This feather was chosen because it was readily identifiable, easily removed, frequently

Tips on Wings and Other Things

81

contained a contrastingly-colored patch, and occurred in six different colors. sure Only were each

Prior to abrasion, all feathers were examined to enthat there were no broken barbs or missing barbules. feathers in perfect condition were used. The feathers individually tagged and traced so that the area of feather or colored patch could be measured.

The feathers were abraded in the laboratory of K. Westphal with the help of B. Morgan, to both of whom I am greatly indebted for the use of their time and equipment. Each feather was exposed for one minute to a stream of powdered silicon from an air pressure gun. The feathers were held vertically with the ventral surface flat against a metal plate and the dorsal surface about 10 em from the nozzle of the air gun. During the one-minute exposure, 0.4 g of powdered glass hit the feather. Following abrasion I counted the number of broken and unbroken barbs. The percentage of broken barbs (P) for each feather was calculated from the ratio of the number of broken barbs (b), to the total number of barbs (B): P

=

100b/B

(3 .1)

The mean percentage of broken barbs for each color was compared using an analysis of variance, and the individual means were compared using the Scheffe test (Roscoe 1975). Following abrasion, each feather was again traced. I used a polar planimeter to find the area from the tracings of each feather before and after abrasion. The change in area was found by subtracting the area after abrasion from the area before abrasion. Size varied among the differently colored feathers. Therefore I calculated the percentage of area destroyed by abrasion (L), according to the equation:

L

=

lOO(A-a)/A

(3. 2)

where a is the area of the feather after abrasion and A is the area of the feather before abrasion. The mean percentage of area destroyed by abrasion was calculated for each color and an overall comparison made with the analysis of variance. Individual means were compared using the Scheffe test.

Edward H. Burtt, Jr.

82

Results A quantitative measure of the effect of natural abrasion cannot be made from Dwight's (1900) photographs of naturally abraded feathers, but the qualitative effects of natural and artificial abrasion are similar. In both cases the barbules have been stripped from large sections of the barbs and numerous barbs have been broken. Therefore artificial abrasion seems to be a reasonable substitute for a natural process that takes place throughout an entire year. The mean percentages of broken barbs in differently colored feathers are shown in Figure 23. The percentages differ significantly from random based on an analysis of variance with 5 and 95 degrees of freedom (F = 9.32, p < 0.001). The results of comparing the mean percentages of 60

en

...

£J

0

10

c 40

""'::"'

10

-30 0

"'c

0'

~ 20 ~

a.."'

N=9

Orange

N=6

Ye flow

Yellowgreen

Black

Brown

Fig. 23. The mean percentage of broken barbs in feathers of different colors exposed to equal amounts of abrasion. The number of feathers abraded (~) is indicated in the appropriate bar. The horizontal lines group feather colors in which the mean percentage of broken barbs is not significantly different. (From Burtt, in press.)

Tips on Wings and Other Things

83

broken barbs for feathers of different colors are shown by horizontal lines (Fig. 23) that connect mean percentages that are not significantly different. Black, brown, and yellow-green feathers, the only feathers that contain melanin, have a significantly lower percentage of broken barbs than white, which contains no pigment. Black and brown feathers also have a significantly lower percentage of broken barbs than orange feathers, which contain a carotenoid pigment but lack melanin. The percentage of broken barbs is not significantly different among black, brown, and yellow-green feathers, all of which contain melanin, nor is there a significant difference among yellow, orange, and white feathers, none of which contains melanin. The mean percentage of area of differently colored feathers destroyed by abrasion is shown in Figure 24. The overall difference is significant (F = 37.65; df = 5, 92; 50

c: 0

"'.._

0

.0

Cl

40

0'

c:

.._

-"' ::1

Cl

30

0

_J

0

~

Cl

0 GJ

0'

~ c:

GJ

~

N=6

N=l9

N=9

GJ

Cl.

Yellow

White

Orange

Black

Yellowgreen

Brown

Fig. 24. The percentage of area lost by differently colored feathers exposed to equal amounts of abrasion. The number of abraded feathers (N) is indicated in the appropriate bar. The horizontal lines group feather colors in which the mean percentage of area lost is not significantly different. (From Burtt, in press.)

84

Edward H. Burtt, Jr.

p < 0.001).

The horizontal lines in Figure 24 connect means that are not significantly different. Brown and yellowgreen feathers lost a significantly smaller percentage of area than yellow or white feathers. The percentage of area lost by brown, black, and yellow-green feathers is not significantly different, nor is there a significant difference in the percentage of area lost by orange, yellow, and white feathers. Discussion The hypothesis that differently colored feathers exposed to a constant amount of abrasion show different amounts of damage is strongly supported by the data, Feathers whose color is wholly or partially the result of melanin impregnation are more abrasion-resistant than those feathers lacking melanin (Prediction 1). The addition of moderate amounts of carotenoid pigment, such as found in yellowgreen feathers, does not affect abrasion-resistance. The data suggest no difference in the abilities of carotenoidpigmented and white feathers to withstand abrasion. THE TOPOGRAPHY OF ABRASION-RESISTANT COLORATION There are at least two sources of abrasion: that caused by the collision of barbs and barbules with airborne particles and that caused by feathers rubbing against one another or against a substrate. Only abrasion caused by airborne particles is discussed (Burtt [in press] discusses other types of abra-sion). Abrasion by Airborne Particles Damage (D) from kinetic energy (K.E.) lision (Probstein and Maki 1974, Schmel and

windborne particles depends on the of the particles at the time of colFasso 1970, Waldman and Reinecke 1971, Sutler 1974, Smith 1976): D = f(K.E.)

(3. 3)

Collisions of barbs and barbules with airborne particles are almost certainly nonelastic. The energy transferred from the particle to the barb or barbule struck is probably dissipated as heat or by the destruction of the keratin molecule.

Tips on Wings and Other Things

85

If the extent of damage that results from the collision depends on the kinetic energy of the particle, then the damage will be greatest where the kinetic energy of the particles is greatest. Kinetic energy is energy due to motion. It is defined:

(3.4) where m is the mass of the particle and v its velocity in any direction. If we assume all particles to be some average size that is characteristic of the air of some habitat --for example, the air of boreal forests, which might contain conifer pollen as its primary particulate--then the kinetic energy of the particles depends only on the velocity of the particles at the point of collision with the barbs and barbules. Abrasion by Airborne Particles in Flying Birds The mass of an average particle traveling toward a collision with a barb or barbule is small compared with the mass of the feather. Therefore, take the feather about to be hit as the reference point: all changes in velocity of either the feather or the particle appear as changes in the velocity of the particle only. An airborne particle appears to approach along the bird's line of flight at a velocity equal and opposite to the bird's velocity, assuming still air. If the velocity of the bird were the only factor that determined the velocity of the particles at the point of collision, then the head, neck, breast, shoulders, and leading edge of the wings would receive the most abrasion. However, the airfoil shape of the wings and possibly of the body as well means that air moving across the dorsum has a higher velocity than air moving across the venter, turbulence around the wings and tail means that these are areas of rapid air movement, and the wings move and therefore contribute their own kinetic energy to the collision. Because the velocity of the air moving past a flying bird varies according to a predictable pattern, the kinetic energy of airborne particles colliding with the barbs and barbuies varies predictably, as does the potential damage caused by abrasion. The following predictions are based on the airflow pattern around a flying bird.

86

Edward H. Burtt, Jr.

The Dorsum If the bodily shape of a bird in flight deviates from teardrop shape, then air flow is not equal over all body surfaces. Figure 43 in Storer (1948, p. 30) shows a flight profile of an anhinga (Anhinga anhinga) that suggests an airfoil. When the body profile approximates an airfoil, the velocity of airborne particles moving across the dorsum is faster than the velocity of those moving across the venter and their kinetic energy on impact is greater. The wing must be an airfoil in birds that fly. Hence airborne particles moving across the dorsal surface of the wing have a higher kinetic energy than particles moving across the ventral surface. The wing anterior to the wing bars is included in the dorsum. Prediction 2: The dorsum is more likely to be melanin-impregnated than the venter. The Tail The tail is used to supplement the control movements of the wings, especially in the rapid aerial maneuvers of warblers. During such maneuvers the tail is spread to form an auxiliary surface behind and below the wings. Such a surface draws air over the main surface of the wings and so keeps the flow attached to the wings at high angles of attack, thereby increasing the maximum lift coefficient of the wing and lowering the stalling speed (Pennycuick 1972, 1975). Such use of the tail subjects the entire tail to rapidly moving air similar to the rapid flow of air across the dorsum, but also subjects the lateral and trailing edges to turbulence. The velocity of particles in areas of turbulence is greater than that of particles in the air moving across the dorsum or the dorsal surface of the tail because the air in the turbulence is moving relative to the nonturbulent flow of air. Where the velocity is greater, the kinetic energy of the particles is greater and the damage from abrasion is greater. The tail is subjected to abrasion from two sources: rapidly moving air and turbulence. Prediction 3: The tail and particularly the lat~ eral and trailing edges of the tail are more likely to be melanin-impregnated than the dorsum. When the tail is expanded, the more medial feathers overlie the adjacent, more lateral feathers with the lateral barbs of each feather above the medial barbs of the adjacent feather. Therefore, rapidly moving particles in the air from the dorsum and the wings abrade the medial

Tips on Wings and Other Things

87

tail feathers more than the lateral feathers, and the lateral barbs of each feather more than the medial barbs. When the tail is furled in rapid flight, the medial feathers are uppermost with the most lateral feathers on the bottom. During such flight, only the medial feathers of the tail and the lateral edges of the underlying feathers are exposed to abrasion from airborne particles moving around the tail. Prediction 4: The medial feathers and lateral edges of all tail feathers are more likely to contain melanin than other parts of the tail. The Remiges The remiges are subject to abrasion from the rapid flow of air that is characteristic of an airfoil. Such abrasion is similar to that occurring on the dorsum and tail. The remiges are also subject to abrasion from turbulence, as is the tail, but the remiges are subject to turbulence both at their trailing edges and at the wing tip. Unlike the other two regions, the remiges, particularly the distal remiges or primaries, are moving. From our reference point on the feather such movement appears as changes in the velocity of the particle. Since the wing moves at an angle with respect to the line of flight, the apparent change in the velocity of the particle is equal and opposite to a vector of the velocity of the remiges that is parallel to the line of flight. So the velocity of a particle approaching the remiges has increased velocity due to movement over an airfoil, increased velocity due to turbulence, and increased velocity due to flapping of the remiges. Prediction 5: The remiges are more likely to be melanin-impregnated than either the dorsum or the tail. THE OBSERVED DISTRIBUTION OF ABRASIONRESISTANT COLORS ON WARBLERS The occurrence on the dorsum of male and female warblers of colors whose abrasion resistance has been measured is shown in Figure 25. As predicted (Prediction 2), the most abrasion-resistant colors (brown, yellow-green, and black) predominate on the dorsum whereas white and yellow predominate on the venter (Fig. 26), but rarely occur on the dorsum. The tail shows a much greater tendency toward abrasion-resistant coloration than the dorsum (Prediction 3).

' Dorsal Colors •

Brown

•

Yellow-green

A

Block

0

Orange

D White 6. Yellow

Q.>

c> 0

c

Q.>

u

....

Q.>

Q.

N=9

White

N=6

Orange Yellow

Brown

Block

Yellowgreen

Fig. 30. The mean percentage of ultraviolet light (290-400 nm) transmitted by feathers of six different colors. The number of feathers measured (li) is indicated in the appropriate bar. The horizontal lines group feather colors in which the mean percentage of ultraviolet light transmitted is not significantly different. (From Burtt, in press.)

Edward H. Burtt, Jr.

98

than 60% of the incident radiation. Orange and yellow feathers, both of which contain carotenoids, transmit significantly less ultraviolet radiation than white feathers and significantly more than feathers that contain melanin. Brown, black, and yellow-green feathers, all of which contain melanin, transmit the least ultraviolet radiation: less than 20% of the incident radiation. Legs with a Munsell color value of two (dark) transmit no ultraviolet energy, whereas legs with a Munsell color value of six (light) transmit 0.6% of the incident ultraviolet radiation. Measurements of transmission through the legs may be underestimated because the light passed through the entire leg not just a layer of scales. However, light legs have a larger diameter than dark legs. Therefore, the ultraviolet beam passes through more tissue in the light legs. With the extraneous tissue removed the difference in transmission between dark and light legs may increase. Discussion The percentage of incident ultraviolet light transmitted to the skin through layers of differently colored feathers is unknown, but is undoubtedly less than the percentage shown in Figure 30. The feat.hers of birds overlap one another. Hence ultraviolet light must penetrate several feathers before reaching the skin. The intensity of the radiation is reduced by each feather the light passes through. Warblers with white, yellow, or orange feathers could reduce ultraviolet irradiance on the skin by increasing the number of feathers covering the skin. However, there is no evidence that such a strategy has been adopted (Wetmore 1936). Thus the best protection offered by colored feathers is provided by feathers that contain melanin. However, feathers that contain carotenoids transmit less ultraviolet radiation than unpigmented feathers and therefore offer some protection from exposure to excessive ultraviolet radiation. Ultraviolet flux increases with altitude. Likewise, the concentration of carotenoids in feathers of Ramphocelus tanagers increases with altitude (Brush 1970), although Brush does not attribute the correlation to ultraviolet-shielding. Melanin-impregnated scales offer complete protection from ultraviolet light, whereas unpigmented scales transmit small amounts of radiation.

Tips on Wings and Other Things

99

TOPOGRAPHY OF ULTRAVIOLET-RESISTANT COLORATION Predicted Patterns of Coloration and Behavior Short-wave radiation is scattered by the earth's atmosphere. Therefore, the sky acts as a hemispherical source of ultraviolet radiation. As the angle between the sun and the zenith (zenith angle) decreases, atmospheric absorption and back-scattering of ultraviolet radiation decrease. Ultraviolet irradiance is most intense and most harmful when the sun is at its zenith (Shettle et al. 1975). With the sun at its zenith, about 50% of the ultraviolet radiation incident at sea level is direct radiation from the sun; the other 50% is diffuse radiation from the sky (Shettle and Green 1974, Allen et al. 1975). The sun approximates a point source. Therefore, when ultraviolet irradiance is most intense about half the radiation comes from a point source and half from a hemispherical source. Under these conditions a bird that is upright exposes the upper mandible, crown, nape, and dorsum to the most intense ultraviolet radiation. Less intense radiation falls on the face, collars, flank, wing patches, and remiges. If coloration of the warbler's external surface has evolved to shield the warbler from excessive ultraviolet radiation, then Prediction 1: The upper mandible, crown, nape, and dorsum will be darker than the face collars, flank, wing bars, and remiges, which will be darker than the lower mandible, throat venter, and legs. If tolerance of ultraviolet radiation (e.g., molecular repair mechanisms) and internal coloration are similar in all warblers, then Prediction 2: Species whose upper mandible, crown, nape, or dorsum is lightly colored will spend less time in direct sunlight than species whose upper mandible, crown, nape, or dorsum are dark. Observed Patterns of Coloration and Behavior Coloration The upper mandible is significantly darker than the lower mandible (Prediction 1; Figs. 31 and 32). The percentage of regions that contain dark, melanin-impregnated feathers is highest in male warblers on the dorsal surface, less on the lateral surface, and least on the ventral surface (Prediction 1; Fig. 33). The pattern in females is almost identical. The legs are the darkest of the ventrally located structures. Almost 50% have Munsell color value two (dark) .

100

Edward H. Burtt, Jr.

QJ

::::J

0

> ~

0 0

u

~ ~

~

•

Upper Mend i ble

0

Lower Mondib··le

I

I

p

0

I

20

I

40

60

80

Number of Species of Worblers

Fig. 31. The Munsell color values of the upper and lower mandibles of males of 115 species of wood warblers (Parulidae). (From Burtt, in press.) Behavior Variation exists in the coloration of the upper mandible, crown, and rump on the dorsal surface, and the legs are unusually dark for ventrally located structures. Therefore, the percentage of time spent in sunlight is compared in species whose upper mandible, crown, rump, and leg color differ. Methods I measured the time spent in sunlight for warblers breeding at Itasca, Minnesota, in 1974 and Chapel Hill, North Carolina, in 1975, and for warblers migrating through Madison, Wisconsin, in 1974 and 1975. I identified three light zones: sunlight, when the body was entirely illuminated; shade, when the body was entirely in shadow; and dappled, when the body was simultaneously in shadow and sunlight. I identified a warbler, waited ten seconds, and

101

Tips on Wings and Other Things

a

::::J

0

> 0

u

"'c ::::J

:E

~ ~

b

•

Upper Mandible

0

Lower Mandible

I

b

PI I

0

20

40

60

80

Number of Species of Warblers

Fig. 32. The Munsell color values of the upper and lower mandibles of females of 106 species of wood warblers (Parulidae). (From Burtt, in press.) then began recording on a tape recorder each change of light zones. I followed an individual for as long as possible or ten minutes, whichever came first. The time each species spent in sunlight is expressed as a percentage of the total time I observed the species under sunlit conditions. Percentages are used because the total time each species was observed varies greatly. Dark legs (Munsell color value two) transmit no ultraviolet radiation whereas light legs (Munsell color value six) transmit some ultraviolet radiation (see above). Therefore, as Munsell color value increases (the color becomes lighter), more ultraviolet light is transmitted. Kendall's tau (Roscoe 1975) was used to evaluate the correlation between the Munsell color value of the upper mandible and the percentage of time spent in sunlight. The Pearson product-moment correlation coefficient (Roscoe 1975) was calculated to evaluate the correlation between the percentage of time spent in sunlight and the percentage

102

Edward H. Burtt, Jr. 50

Crown, Nope, Dorsum

40

•

30 "0

.,

20

.,u

10

...

•

Q. !/)

... 0

8

00

.,

50

-

40

~

30

.1:.

0

-

~

10

20

....

10

ct

30

70

50

•

0., 50 40

40

Face, Collars, Flank, Wing bar, Remiges

20

~.,"'

30

20

0

30

40

50

60

70

Throat, Venter

0

u

20 10 40 60 50 30 Percent of Ultraviolet Light Transmitted

e

Brawn

& Black

0 White

•

Yellow-green

0

1:::,. Yellow

Orange

70

Fig. 33. Comparison of the percentage of differently colored regions on the dorsal, lateral, and ventral surfaces of male warblers. The percentages do not add up to 100: some of the colors that occur on warblers were not measured because suitable feathers could not be obtained. (From Burtt, in press.) of ultraviolet light transmitted by differently colored feathers of the crown and rump.

Tips on Wings and Other Things

103

The light falling on the legs was recorded separately from that on the body. I identified the species of warbler, waited ten seconds, and just before starting the taped record, I noted the light zone of the body and of the legs independently of the body. From these data I calculated the probability of exposing the legs to sunlight. A chisquare test was used to compare the probability of exposing the legs to sunlight in dark-legged (Munsell color value two) and light-legged (Munsell color values four, six, or eight) warblers. Results There is a significant negative correlation between the Munsell color value of the upper mandible and the percentage of time spent in sunlight (Fig. 34). Species whose upper mandible is dark (Munsell color values one or two) occasionally spent time in sunlight whereas species whose 60 ~

~~ c

:I

II)

c

..

·- 40

c

a.

..

II)

E

30

I-

._

.. "' .. 0

~ c

.. ~

20

10

a..

0

• • •

' i,, •

...................

'

... • I t DARK

.......

............

'

.......

.......

,

'm

-------61 4 LIGHT

Munsell Color Value of the Upper Mandible

Fig. 34. Species of warblers plotted by the percentage of time spent in sunlight and the Munsell color value of the upper mandible. The dotted line connects median values. (From Burtt, in press.)

Edward H. Burtt, Jr.

104

upper mandible is lighter (Munsell color values three or four) spent little time in sunlight. Among species I observed, only the bay-breasted warbler (Dendroica Castanea) is sexually dimorphic in the coloration of its upper mandible. The upper mandible of the male is darker (Munsell color value one) than the upper mandible of the female (Munsell color value two), and the male spent 7% of his time in sunlight as compared with 3% for the female. There is no correlation between the coloration of the crown and the percentage of time spent in sunlight by species with differently colored crowns (r = 0.12, r 2 = 0.01) nor is there a correlation between coloration of the rump and the percentage of time spent in sunlight by species with differently colored rumps (r = 0.12, r 2 = 0.01). The probability of exposing dark legs (Munsell color value two) to sunlight is not significantly different (X 2 = 3.00; df = 1; 0.10 > p > 0.05) from the probability of exposing light legs (Munsell color value four, six, eight) to sunlight (Table 5). Table 5 Number of Observations in which Dark- and Light-Legged Warblers Held Their Legs in Sunlight or Shade When the Body was in Sunlight, Dappled Light, or Shade (Species Combined)

Color of Legs

Sun

Shade

Dark (Munsell color value 2)

44

469

25

169

Light (Munsell color value 4, 6, 8)

H.: of legs.

Exposure to sunlight independent of color value

3.00; df

1; E

=

0.10.

0.10

> ~ >

0.05.

Tips on Wings and Other Things

105

EXTERNAL COLORATION AS A DEFENSE AGAINST ULTRAVIOLET RADIATION:

EVALUATION OF THE HYPOTHESIS

Transmission of ultraviolet light is correlated with pigmentation of the feathers. Melanin-impregnated feathers transmit significantly less ultraviolet radiation than feathers that contain carotenoids. Feathers that contain carotenoid pigments transmit significantly less ultraviolet radiation than unpigmented feathers. The pattern of coloration of wood warblers is consistent with the hypothesis that external coloration evolved to shield the animal from excessive ultraviolet radiation. Feathers that contain melanin occur most commonly on the dorsal and lateral surfaces of warblers where ultraviolet radiation is most intense (Prediction 1). Only coloration of the crown and rump depart significantly from the predicted pattern. The behavior of species with differently colored plumage is not consistent with the hypothesis that coloration of the plumage evolved to shield the animal from ultraviolet radiation. Species whose crown and rump feathers transmit large amounts of ultraviolet light spend just as much time in direct sunlight as congeners whose crown or rump feathers transmit little ultraviolet light (contra-Prediction 2). Therefore, the risk of damage incurred by exposure to ultraviolet radiation is outweighed by the advantages of bright, ultraviolet-transparent colors which serve communicative or other functions. The density of melanin granules in keratin determines how dark the structure appears (Bowers 1959). Therefore, the varied shades of the mandibles and legs are due to different concentrations of melanin. Where the melanin concentration is low, in the light legs of the ovenbird (Munsell color value six), for example, ultraviolet radiation penetrates the keratin and reaches living tissue. The upper mandible is darker than the lower mandible as predicted (Prediction 1) and species with lightly colored upper mandibles (Munsell color values three and four) avoid sunlight as predicted (Prediction 2). Therefore, both the coloration of the mandibles and the behavior of species with differently colored mandibles are consistent with the hypothesis that external coloration is an adaptation to protect tissue from ultraviolet radiation. However,

Edward H. Burtt, Jr.

106

see Burtt (in press) for other interpretations of mandibular coloration. The legs of warblers are frequently as dark or darker than the dorsum (contra Prediction 1). Species with lightly colored legs show no tendency to shield the legs from ultraviolet radiation (contra Prediction 2). Therefore coloration of the legs of warblers is not explained by the hypothesis that external coloration has evolved to shield the animal from excessive ultraviolet radiation. ACKNOWLEDGMENTS I examined specimens of warblers through the courtesy of Melvin R. Traylor and the Field Museum in Chicago, Robert W. Storer and the University of Michigan, Frank B. Gill and the Philadelphia Academy of Sciences, and Raymond A. Paynter and the Museum of Comparative Zoology at Harvard. I thank Jack P. Hailman, Timothy C. Moermond, and Warren P. Porter for their criticism of my ideas and manuscript. My field work was supported by a Josephine Herz Fellowship from the University of Minnesota in 1973 and by the Frank M. Chapman Fund of the American Museum of Natural History in 1974. This chapter is based on a doctoral dissertation submitted to the Department of Zoology, University of Wisconsin, Madison, Wisconsin. REFERENCES Allen, L. H., Jr.; Gausman, H. W.; and Allen, W. A. 1975. Penetration of solar ultraviolet radiation into terrestrial plant communities. In CIAP Monograph 5: Part ]-Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 2-78--2-108. Averill, C. K. 1923. Black wing tips. Condor 25:57-59. Bachman, J. 1839. Observations on the changes of colour in birds and quadrupeds. Trans. Amer. Philos. Soc. (Phila.) new series 6:197-239. Blake, C. H. 1966. Warbler tail spots. EBBA News 29:54-55. Blum, H. F. 1975. Ultraviolet radiation from the sun and skin cancer in human populations. In CIAP Monograph 5:

Tips on Wings and Other Things

107

Part ]--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 7-87-7-103.

Bollum, F. J., and Setlow, R. B. 1963. The action spectra for ultraviolet-light inactivation of systems containing 5-bromouracil-substituted deoxyribonucleic acid. Biochem. Biophys. Acta 68:446-454. Bowers, D. E. 1959. A study of variation in feather pigments of the wrentit. Condor 61:38-45. Brush, A. H. 1970. Pigments in hybrid, variant and melanic tanagers (birds). Camp. Biochem. Physiol. 36:785-793. Brush, A. H., and Siefried, H. 1968. Pigmentation and feather structure in genetic variants of the Gouldian Finch, Poephila gouldiae. Auk 85:416-430. Burtt, E. H., Jr. The coloration of wood warblers (Parulidae). Nuttall Ornithol. Monogr.: in press. Caldwell, M. M., and Nachtwey, D. S. 1975. Introduction and overview. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 1-3--1-29. Calkins, J., and Nachtwey, D. S. 1975. UV effects on bacteria, algae, proLozoa, and aquatic invertebrates. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects,

D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 5-3--5-9. Cole, L. C. 1943. Experiments on toleration of high temperature in lizards with reference to adaptive coloration. Ecology 24:94-108. Collette, B. 1961. Correlations between ecology and morphology in anoline lizards from Havana, Cuba, and southern Florida. Bull. Mus. Compar. Zool. 125:137-162. Dwight, J., Jr. 1900. Sequence and plumage of moults of the Passerine birds of New York. Ann. N.Y. Acad. Sci. 13:73-360.

108

Edward H. Burtt, Jr.

Hanawalt, P. C. 1966. The u.v. sensitivity of bacteria; its relation to the DNA replication cycle. Photochem. Photobioi. 5:1-12. Harm, W. 1969. Biological determination of the germicidal activity of sunlight. Radiat. Res. 40:63. Heinroth, 0., and Heinroth, K. 1958. The Birds. Ann Arbor: Univ. Mich. Press. Klauber, L. M. 1939. Studies of reptile life in the arid southwest. Bull. Zool. Soc. San Diego 14:1-100. Kleczkowski, A. 1971. Photobiology of plant viruses. In Photophysiology, Vol. 6, A. C. Giese (ed.). New York: Academic Press, pp. 179-208. Kubitschek, H. E. 1967. Mutagenesis by near-visible light. Science 155:1545-1546. McLaren, A. D., and Shugar, D. 1964. Photochemistry of Proteins and Nucleic Acids, Vol. 22. New York: Macmillan. Maki, T. 1974. Prevention of wind frictional scratch on citrus fruit and the suppression of transpiration from eggplant by a surface coating agent. J. Agric. Meteorol. 30:39-44. Murphy, T. M. 1973. Inactivation of TMV-RNA by ultraviolet radiation in the sunlight. Int. J. Radiat. Biol. 23: 519-526. Pennycuick, C. J. 1972. Animal Flight. London: William Clowes. Pennycuick, C. J. 1975. Mechanics of flight. Avian Biol. 5:1-75. Porter, W. P. 1967. Solar radiation through the living body wall of vertebrates with emphasis on desert reptiles. Ecol. Monogr. 37:273-296. Porter, W. P. 1975. Ultraviolet transmission properties of vertebrate tissues. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 6-3--6-15.

Tips on Wings and Other Things

109

Porter, W. P., and Norris, K. S. 1969. Lizard reflectivity change and its effect on light transmission through body wall. Science 163:482-484. Probstein, R. F., and Fasso, F. 1970. Dusty hypersonic flows. AIAA Journal 8:772-779. Resnick, M. A. 1969. A photoreactivationless mutant of Saccharomyces cerevisiae. Photochem. Photobiol. 9:307312. Resnick, M. A. 1970. Sunlight-induced killing in Saccharomyces cerevisiae. Nature 226:377-378. Roscoe, J. T. 1975. Fundamental Research Statistics for the Behavioral Sciences. New York: Holt, Rinehart. Rupert, C. S. 1964. Photoreactivation of ultraviolet damage. In Photophysiology, Vol. 2, A. C. Giese (ed.). New York: Academic Press, pp. 283-327. Schmel, G. A., and Sutter, S. L. 1974. Particle deposition rates on a water surface as a function of particle diameter and air velocity. J. Recherches Atmospheriques 8:911-920. Shettle, E. P., and Green, A. E. S. 1974. Multiple scattering calculation o£ the middle ultraviolet reaching the ground. Appl. Opt. 13:1567-1581. Shettle, E. P.; Nack, M. L.; and Green, A. E. S. 1975. Multiple scattering and the influence of clouds, haze, and smog on the middle UV reaching the ground. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects,

D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 2-38--2-49. Smith, K. C. 1971. The roles of genetic recombination and DNA polymerase in the repair of damaged DNA. In Photophysiology, Vol. 6, A. C. Giese (ed.). New York: Academic Press, pp. 209-278. S1nith, D. H. 1976. Debris shielding in regions of high edge velocity. AIAA Journal 14:94-96. Storer, J. H. 1948. The Flight of Birds Analyzed Through Slow-Motion Photography. Bloomfield Hills, Mich: Granbrook Press.

110

Edward H. Burtt, Jr.

Test, F. H. 1940. Effects of natural abrasion and oxidation on the coloration of flickers. Condor 42:76-80. Voitkevich, A. A. 1966. The Feathers and Plumage of Birds. New York: October House. Waldman, G. D., and Reinecke, W. G. 1971. Particle trajectories, heating, and break-up in hypersonic shock layers. AIAA Journal 9:1040-1048. Watkins-Pitchford, W. 1909. The Etiology of Cancer. London: William Clowes. Wetmore, A. 1936. The number of contour feathers in Passeriform and related birds. Auk 53:159-169. Witkin, E. M. 1966. Radiation-induced mutations and their repair. Science 152:1345-1353.

The EvolutionoEngineering Approach: Discussion

C. Richard Tracy

Jed Burtt's approach to understanding the color patterning of warblers is what I called the "evolutiono-engineering approach." In this approach, you begin with basic physical and chemical principles and attempt to answer the question, "How can I best design this animal?" If the animals studied have the same design as that reasoned from the principles employed, you assume that the principles and processes originally assumed to be important in the evolution of particular adaptations are indeed important and that you have deduced some knowledge of the process and/or importance of particular adaptations. This approach differs from the more inductive process of surveying the adaptations of a large sample of animals, and then inferring knowledge about the process and/or importance of the adaptation studied. I have two comments on the use of the evolutionoengineering approach. First, this deductive analysis necessarily carries assumptions which should be tested individually. For example, Jed implicitly assumed that a one-minute blast of powdered glass on feathers was significantly related to the normal wear that feathers would get in nature. This assumption must be tested in some way, for if the normal wear on feathers is not great enough to cause reduced fitness in birds having feathers without the "appropriate" amounts of melanin, then the supporting predictions from the "wear hypothesis" must be reevaluated in

112

C. Richard Tracy

terms of multicolinear support for alternative hypotheses. For example, Jed's Prediction 2 (above) states that from the hypothesis of protection from wear, the dorsum of warblers is more likely to be melanized than the venter. If the prediction is supported by evidence, we also should ask how many alternative hypotheses does it likewise support. We know from the material given by Jed that a melanized dorsum also supports the hypothesis of protection from ultraviolet radiation. Thus, predictions in the evolutionoengineering approach must be based on reasonable and/or well tested assumptions, and confirmed predictions must be evaluated in terms of alternative hypotheses or the entire approach could mislead us r~ther badly. One way to strengthen the evolutiono-engineering approach is to form hypotheses in terms of optimality principles (Rosen 1967). Such principles do not generate hypotheses concerning evolutionary adaptations purely from the benefit derived from the adaptation, but rather from some benefit-cost relationship. That is to say, for every adaptation, one must evaluate and analyze the cost of that adaptation concomitant with the benefit derived from it. Indeed, if one were to make predictions entirely on the basis of benefit, and work strictly from the principle that melanin provides a benefit in terms of protection from abrasive wear, one would have to predict that warblers should be encased in feathers containing melanin. Since none of the warblers studied were completely armored with melanin, the prediction seems inadequate without also considering the costs of encasing the animal in melanin. The evolu~ tiona-engineering approach using optimality principles of benefits and costs has yielded amazing insights into the selective advantages of complex biological adaptations (Parkhurst and Loucks 1972), and its use in ethological analyses promises very inventive breakthroughs. The hypothesis that coloration can be involved in adaptations for protection from ultraviolet (Porter 1967) is very interesting and somewhat controversial (see Hamilton 1973). The initial assumption from which Jed has made predictions is that incident ultraviolet radiation exists in doses which are damaging to warblers. Thus, all of the predictions stemming from the ultraviolet hypothesis seemingly would have to be confirmed or else the initial assumption would not be true. In other words, any ''damage," however slight, decreases fitness, so no adaptation implies no (or at least weak) potential for ultraviolet damage. Thus, it seems that each of Jed's predictions

The Evolutiono-Engineering Approach

113

approximates a good null hypothesis, and rejection of any one should allow rejection of the overall hypothesis (that color patterning is a response to pressures from ultraviolet damage). REFERENCES Hamilton, W. J., III. 1973. Life's Color Code. New York: McGraw-Hill. Parkhurst, D. F., and Loucks, 0. L. 1972. Optimal leaf size in relation to environment. J. Ecol. 60:505-537. Porter, W. P. 1967. Solar radiation through the living body walls of vertebrates with emphasis on desert reptiles. Ecol. Monogr. 37:273-296. Rosen, R. 1967. Optimality Principles in Biology. New York: Plenum Press.

Where is the Evidence for Ultraviolet Damage?: Discussion

William]. Hamilton III

Jed has shown that a darkly colored dorsum may shield an animal from harmful ultraviolet radiation or increase the abrasion resistance of darkly colored feathers, fur, or scales; such hypotheses have received scant attention in the past. In addition, two classic hypotheses, abs.orption of solar radiation (Porter et al. 1973) and countershading (Ruiter 1956, Cott 1957), remain to be integrated into an overall hypothesis that can predict the coloration and pattern of color on the dorsum. The questions I pose, but cannot answer, are: How can we separate these four hypotheses? What is the relative importance of each selection pressure? Are the four hypotheses an inseparable, adaptive mix? I suggest that, in the case of protection from ultraviolet radiation, despite all the effort to establish a quantitative argument, Jed and others who have concentrated on comparative studies provide the only evidence to date that melanic coloration shields the animal from ultraviolet radiation. Jed mentioned that the mutation rate of E. coli increases when it is exposed to the radiation levels that would reach the lizard's gut (where E. coli reside) if the lizard's black peritoneum were removed. However, E. coli is an internal organism that lives in a sheltered environment. Isn't it possible that, if E. coli lived in another environment, it would have a different tolerance to ultraviolet radiation?

Evidence for Ultraviolet Damage?

115

What is the critical experiment, what are the critical data that will give us a quantitative measure of the relevance of the ultraviolet protection hypothesis? I fail to see such evidence. I see supportive evidence in Porter's (1967) comparison of the pigmentation of the peritoneum in lizards, in Collette's (1961) comparison of anole species, and in Jed's data from warblers. These studies provide supportive, circumstantial evidence that ultraviolet radiation is an evolutionary problem to free-ranging animals in natural environments. Can anyone show where quantitative data fit into the argument? Burtt: Bill raises two important points: (1) the evolution of ultraviolet tolerance and (2) the lack of quantitative evidence that ultraviolet radiation is a potential hazard to lizards and warblers. Calkins and Nachtwey (1975) found that present flux rates are lethal to many classes of organisms even after tolerance mechanisms are accounted for. Hence these organisms must shield themselves. E. coli has a low tolerance for ultraviolet radiation and lives in an environment that shields it from the potentially damaging radiation. When E. coli's shield, the lizard's black peritoneum, is removed, the bacterium's mutation rate rises. Hence the melanin-impregnated peritoneum is an effective shield beneath which E. coli is protected and without which E. coli is irreparably damaged.

As to quantitative evidence, there is none to show that ultraviolet radiation is a real hazard to warblers or lizards. Humans who are exposed to excessive ultraviolet radiation develop severe sunburn, freckling, and skin cancer (Blum 1975). I expect that the same conditions could develop in other vertebrates whose skin is exposed to excessive ultraviolet radiation, but a single lethal dose is unnecessary. Radiation damage is cumulative. Constant, low-intensity irradiation can cause damage just as surely as a single intense dose. Hamilton: The human evidence is irrelevant because the human animal has been introduced recently into a habitat in which it did not evolve. An animal driving a tractor all day with its neck exposed is living outside its natural haunts. I would like to see evidence from the nonhuman world. Question: Would you rephrase your question regarding quantitative data?

116

William J. Hamilton

III

Hamilton: Are animals, living under natural conditions, ever exposed to excessive ultraviolet radiation? Is melanin an adaptation for shielding animals from excessive doses of ultraviolet radiation? Question: Are you suggesting a more manipulative approach to the question of ultraviolet protection? Hamilton:

Yes.

Question: Relative to critical tests, what criteria would you use to assess radiation damage? Hamilton: The animal dies after exposure to ultraviolet radiation, is denied profitable access to space or resources because of ultraviolet radiation, or reproduces or grows less rapidly when it is not exposed to ambient levels of ultraviolet radiation. Question: Isn't the fact implicit in your question, that in order to define a critical test, you must know what you are being protected against? No one has referred to data that tell you what sort of damage occurs or where it occurs. Burtt: On a macroscopic level the effects of excessive exposure to ultraviolet radiation are severe sunburn, freckling, and skin cancer (Blum 1975). These symptoms could serve as criteria, although their occurrence in nonhuman vertebrates is poorly documented.

Microscopically, absorption of ultraviolet occurs when the frequency of the incident radiation coincides with the resonant frequency of electrons in the outer shell of an atom or molecule. Absorption raises the electrons in the outer shell to a higher energy level. If the electrons are part of a covalent bond, that bond may rupture. If DNA absorbs an ultraviolet photon into a thymine molecule, the hydrogen bond between adenine and thymine ruptures and two adjacent thymines on the same backbone unite to form a dimer linkage (Beukers and Berends 1960) that is more stable than the former hydrogen linkage. The result is a DNA molecule that fails to replicate properly (Setlow, Swenson and Carrier 1964, Hanawalt and Hayes 1967). Question: Is there evidence that ultraviolet radiation affects metabolism?

Evidence for Ultraviolet Damage?

117

Hamilton: Yes, but that is not my question. My question is: Where is the evidence that intensities of ultraviolet radiation sufficient to cause metabolic or genetic damage are relevant to animals living in their natural habitats? to animals living in a natural environment? Question: Have the necessary experiments not been done in the natural environment? Hamilton: How would you design such an experiment? Is the fact demonstrable only on a comparative basis? Question:

experiment?

Is an albinistic animal a natural

Hamilton: No, an albinistic animal is a poorly controlled experiment. In many species albinos certainly have radiation problems. Such freaks also lack the repertoire of color adaptations to their particular color environment. The ultraviolet problems such individuals encounter are probably a consequence of a shift to an ultraviolet environment not encountered by the organism in its recent evolutionary history. Hence such problems cannot be evidence for or against the supposed evolutionary basis of pigments. Tracy: Bill makes the important point that we have a problem with multicolinear effects. How can you separate one hypothesis fyom all other hypotheses, when all hypotheses make similar predictions? One test cannot single out one hypothesis; the results of that test can only support the possibility of several hypotheses. Bill's point is excellent and can be leveled at Jed's work. From the data presented we cannot conclude that the ultraviolet hypothesis dictates selection. At the same time Jed's thrust was not to conclude that potentially harmful ultraviolet radiation dictates coloration, his thrust was to say that there are many predictive hypotheses worthy of careful study. That is the real defense of Jed's work. [Hamilton: Subsequent reflection following this conference has suggested to me one line of quantitative evidence which would demonstrate that ambient levels of ultraviolet radiation may impose limits upon the fitness of animals living in undisturbed natural environments. If decrements in ambient levels of ultraviolet radiation decrease mortality, increase reproductive or growth rate, or

118

William J. Hamilton

III

otherwise enhance fitness, the conclusion that ambient levels of radiation are hazardous and limiting would seem to be justified. It is experimentally much easier to increase radiation above ambient levels and to observe the result. Since such experiments are taking place under conditions not experienced in nature there is no selection pressure for adaptation to such conditions. It is thus not surprising that the general result is a demonstration of deleterious consequences. For marine algae, Lorenzen (1975) concludes that natural levels of ultraviolet radiation suppress photosynthesis in naturally occurring marine phytoplankton populations. To the extent that this is true, the case for ultraviolet as a barrier to further adaptation is apparently demonstrated. However, in interpreting such experiments it is critical to apply ultraviolet dosages available in the space and time occupied by the population under investigation.] REFERENCES Beukers, R., and Berends, W. 1960. Isolation and identification of the irradiation products of thymine. Biochem. Biophys. Acta 41:550-551. Blum, H. F. 1975. Ultraviolet radiation from the sun and skin cancer in human populations. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 7-87--7-103. Calkins, J., and Nachtwey, D. S. 1975. UV effects on bacteria, algae, protozoa, and aquatic invertebrates. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. S-3--5-9. Collette, B. 1961. Correlations between ecology and morphology in anoline lizards from Havana, Cuba, and southern Florida. Bull. Mus. Compar. Zool. 125:137-162. Cott, H. B. 1957. Adaptive Coloration in Animals. London: Methuen. Hanawalt, P. D., and Haynes, R. H. 1967. The repair of DNA. Sci. Amer. 216:36-43.

Evidence for Ultraviolet Damage?

119

Lorenzen, C. 1975. Phytoplankton responses to UV radiation and ecological implications of elevated UV irradiance. In CIAP Monograph 5: Part 1--Ultraviolet Radiation Effects, D. S. Nachtwey, M. M. Caldwell, and R. H. Biggs (eds.). Springfield, Va.: National Technical Information Service, pp. 5-83--5-91. Porter, W. P. 1967. Solar radiation through the living body wall of vertebrates with emphasis on desert reptiles. Ecol. Monogr. 37:273-296. Porter, W. P.; Mitchell, J. W.; Beckman, W. A.; and DeWitt, C. B. 1973. Behavioral implications of mechanistic ecology; thermal and behavioral modeling of desert ectotherms and their microenvironment. Ecologia 13:1-54. Ruiter, L. de. 1956. Countershading in caterpillars. An analysis of its adaptive significance. Arch. Neerl. Zool. 11:285-342. Setlow, R. B.; Swenson, P. A.; and Carrier, W. L. 1964. Thymine dimers and inhibition of DNA synthesis (in bacteria) by ultraviolet irradiation of cells. Science 142:1464-1466.

Audience Questions: Discussion

Question: Unpigmented species of rodents frequently live in competition with melanistic species of rodents on white sands and adjacent black lava beds (Benson 1933, Hooper 1941). Could such a parametric situation be exploited to tease apart the hypotheses of abrasion resistance, ultraviolet protection, thermoregulation, and conspicuousness? Hamilton: Let me speak to that, because I spent three months last year (1976) in the Namib Desert working with two species of tenebrionid beetle, Onymacius. One has a white dorsum, the other a black dorsum. They inhabit approximately the same environment, but have different activity rhythms.

The separation of hypotheses is enormously difficult. You must follow individuals in order to know their radiation dosage. You must know the longevity of individuals who have received different cumulative doses (see Jed's discussion, above) and you must know what constitutes radiation damage (see Jed's discussion, above). What we find is that inside the abdomen ultraviolet radiation is about equally intense in black and white animals, because the externally white beetle has a dark, melanized layer beneath its carapace. Predation pressures upon the two species are different and the white species loses water at almost twice the rate of the black one (Hamilton 1973), so

Audience Questions

121

the question arises: Is water a limiting factor in the desert environment? Hence, before you get an answer the whole web of adaptations comes into play. The approach you suggest is a valid one, but not an easy one. Question: Have you thought about how the different texture, number arrangement, or pigment concentration of different plumages (e.g., natal down, juvenile plumage, first winter plumage, first nuptial plumage, and so forth) affects ultraviolet protection or abrasion resistance? Burtt: I have thought about the variables you mention, but in an attempt to use the comparative method rigorously, I studied only adult wood warblers (Parulidae), a group of physiologically and morphologically similar birds. Hence I eliminated the variables you mention. Morse: Many warblers are dark dorsally and light ventrally, but many of the light feathers are light distally and gray to black proximately. I hope you realize that. Burtt: Yes, you raise an important point also illustrated by Bill's study of black and white Onymacius beetles (above, Hamilton 1973). The ultraviolet shield need not be on the animal's outermost surface. The feathers of birds and the cuticle of beetles are nonliving structures that are not seriously damaged by ultraviolet radiation. Living cells that are easily damaged by ultraviolet radiation (e.g., nerve cells, gametes) must be shielded. The melanin shield may enclose only easily damaged cells in the nervous and reproductive systems (e.g., Dipsosaurus dorsalis, Porter 1967), may lie just beneath the nonliving cuticle, feathers, fur, skin, or scales (e.g., the white Onymacuis beetle), or the shield may comprise external coloration of the body. The location of the ultraviolet shield is less important than the fact that the shield exists in all animals that are exposed to ultraviolet light. When the melanin shield is internal, external coloration is free to conform to other selection pressures. Indeed, the hypothesis of ultraviolet protection may be a poor explanation of external coloration because there is no necessity to absorb ultraviolet radiation at the outermost surface of the body. Question: If melanin strengthens the cuticle, could one predict that raptors would have darker claws than birds that do not "abuse" their claws?

122

Audience Questions

Hamilton: I do not know about that, but the most perfectly camouflaged crickets, green and beautifully matching their substrate, have black, melanic tips on the jaws and tarsal tips. A structural function whose benefits exceed the cost of more perfect camouflage is strongly implied. Burtt: Are the claws of raptors subject to more abrasion than the claws of a sparrow or warbler that scratches on the ground for its food? Question: Your prediction of abrasion resistance was based on evidence that airborne particles abrade aircraft. Have you any evidence that airborne particles abrade bird feathers? Could the wear be the result of repeated bending? Burtt: I have only circumstantial evidence that airborne particles are the source of abrasion to bird feathers. The worn feathers pictured by Dwight (1900), the worn feathers of warblers that I have captured in mist nets, and the feathers I abraded with airborne particles all show a similar pattern of broken and pitted barbs. The pitted barbs and the irregularity of the breaks in the vane of the feather (e.g., one barb broken near its joint with the shaft of the feather whereas the adjacent barbs are unbroken) suggest collision with some small, hard object such as an airborne particle. Because the vane o£ the feather bends as a unit, breaks due to bending should extend uniformly across many barbs. Breaks from contact with twigs, branches, or the ground sh,ould involve many barbs and not infrequently the shaft. I commonly find such breaks, but they cannot account for the pitted barbs or the isolated, broken barbs. I recognize that the evidence is circumstantial, and I am working on wind tunnel measurements that may provide direct evidence for the particulate nature of feather abrasion.

REFERENCES Benson, S. B. 1933. Concealing coloration among some desert rodents of the southwestern United States. Univ. Calif. Publ. Zool. 40:1-70. Dwight, J., Jr. 1900. Sequence and plumage of moults of the passerine birds of New York. Ann. N.Y. Acad. Sci. 13:73-360.

Audience Questions

123

Hamilton, W. J., III. 1973. Life's Color Code. New York: McGraw-Hill. Hooper, E. T. 1941. Mammals of the lava fields and adjoining areas in Valencia County, New Mexico. Misc. Publ. Mus. Zool., Univ. Mich. 51:1-47. Porter, W. P. 1967. Solar radiation through the living body walls of vertebrates with emphasis on desert reptiles. Ecol. Monogr. 37:273-296.

Part3

Photoreception

Chapter 4

Extraretinal Photoreception Herbert U nderu;ood Introduction Insects and Extratetinal Photoreception The Role of ERRs in Entrainment of Insect Clocks The Role of ERRs in Insect Photoperiodism Vertebrates and Extratetinal Photoreception Pineal System: Photosensitivity Entrainment of the Biological Clock Photoperiodic Photoreception Pineal Biochemical Rhythms Physiological Color Change Phototaxis and Photokinesis Orientation Photoreception in Adult Mammals Conclusion

Extraretinal Photoreception

129

INTRODUCTION For over a century we have known that vertebrates can respond to light by using photoreceptors other than the lateral eyes. Much of the earlier work concerned the phototaxic and photokinetic behavior of blir,ded animals and many early researchers assumed that lighl was perceived by photoreceptors in the skin--a so-called "dermal light sensitivity" (Steven 1963). Interest in extra1:etinal photoreception revived following the demonstration that important physiological and behavioral responses such as photoperiodism, changes in external coloration, or entrainment of the biological clock can be influenced by light after removal of the eyes. Recent studies show that extraretinal photoreception is a consistent and important aspect of the sensory repertoire of all vertebrates, with the possible exception of adult mammals. However, extraretinal photoreception is not confined to vertebrates. Many invertebrates (e.g., insects, molluscs, crustaceans) also employ extraretinal photoreceptors. This chapter discusses the role of extraretinal photoreceptors among the invertebrates only in insects since they have received more intensive study than other invertebrate groups. The reader is referred to Lickey et al. (1976) and Page and Larimer (1976) for recent reviews on extraretinal photoreception in other invertebrate groups. The daily light-dark cycles associated with the earth's rotation and the annual changes in the length of the photoperiod are used by many organisms to time important physiological and behavioral events. By using alternating light-dark cycles to synchronize daily (circadian) rhythms animals can gain a degree of temporal coordination that would be impossible by utilizing more labile stimuli such as temperature or humidity. All eukaryotic organisms examined to date exhibit entrainment (synchronization) of circadian rhythms to 24-hour light-dark cycles. In addition, the annual change in day length offers a noise-free cue that relatively long-lived organisms can use to time annual cycles in such physiological and behavioral processes as diapause, fattening, migration, and reproduction. The adaptive significance of such photoperiodic responses is obvious--animals can anticipate and prepare for adverse conditions (e.g., diapause, migration) and they can confine reproduction to the time of year that is most conducive to

130

Herbert Underwood

the survival of the organism and its offspring. Most of the research into the role of extraretinal photoreceptors (ERRs) in insects as well as vertebrates concerns entrainment of the biological clock and photoperiodic photoreception. INSECTS AND EXTRARETINAL PHOTORECEPTION The Role of ERRs in Entrainment of Insect Clocks In only two insects, cockroaches (Blattidae) and crickets (Gryllidae), is there evidence that ERRs (extraretinal photoreceptors) are not involved in entrainment of circadian rhythms. In both insects the compound eyes are the only routes by which daily light cycles can entrain the clock (Roberts 1965; Nishiitsutsuji-Uwo and Pittendrigh 1968a, 1968b; Loher 1972; Sokolove 1975; Sokolove and Loher 1975). Transection of the optic nerve, for example, produces a free-running (that is, the animal expresses its endogenous circadian rhythm) locomotor rhythm regardless of lighting conditions. The optic lobes of the brain appear to be the sites of the driving oscillators (or biological clocks) responsible for maintaining rhythmicity, since sectioning the neural pathways between the optic lobes and the rest of the brain or removing the optic lobes causes arrhythrnicity in locomotor (cockroaches and crickets) or stridulatory (crickets) activity. In all other insects examined, entrainment persists after removal of the compound eyes (Truman 1976). In the "long-horn" grasshopper (Ephippiger sp.), Dumortier (1972) showed that entrainment of the circadian rhythm of stridulatory activity persisted after removal of compound eyes or ocelli. However, localized illumination of the head area indicated that the compound eyes and/or ocelli also had an input into the grasshopper's biological clock. In most insects the compound eyes are apparently not involved in the perception of entraining light cycles. Several investigators have used the daily rhythm of eclosion in order to reveal the nature and location of photoreceptors that have inputs into the biological clock of insects. Many insects exhibit a daily rhythm of eclosion; in a population of insects, individuals emerge from their pupal cases only during a restricted period of the day.

Extraretinal Photoreception

131

In Drosophila, for example, individual flies under natural conditions emerge near dawn, the coldest and wettest part of the day. Emerging flies lose water far more rapidly than mature adults and the wings may fail to expand properly if the humidity is too low. Although an individual emerges only once, the event is timed by the insect's biological clock. The existence of ERRs in Drosophila was established by Engelmann and Honegger (1966) who found that the circadian eclosion rhythm of Drosophila melanogaster mutants, which lacked compound eyes and ocelli, entrained normally to an LD 12:12 cycle. Zimmerman and Ives (1971) determined the spectral sensitivities for both the compound eyes and the circadian eclosion rhythm of Drosophila pseudoobscura (Fig. 35). In this experiment, the spectral sensitivity of the compound eye photoreceptors was determined by an electrode placed on the corneas of immobilized flies. Action spectra for the eclosion rhythm were determined by exposing populations (white-eyed and wild-type) of

B >-

t-

>

-2

2

t-

"'zw "'

"'

1\

\ \

-4

.

.~

400

500 WAVELENGTH (nm)

600

400

500 WAVELENGTH (nm)

~~

600

Fig. 35. Spectral sensitivity of the compound eyes of the adult (A) and the circadian rhythm (B) in white-eyed (open circles, dotted arrows) and wild-type (closed circles, solid arrows) D. pseudoobscura (Zimmerman and Ives 1971. Reproduced with the permission of the National Academy of Sciences).

132

Herbert Underwood

Drosophila, which had been released from LD 12:12 into constant darkness (DD), to 15-minute light signals 5 hours after the light-to-dark transition. The populations were exposed to either monochromatic light, white fluorescent light (1100 lux), or no light (free-run controls). Action spectra were obtained by determining the relative number of quanta at wavelengths between 354 nm and 800 nm which generated a phase shift in the eclosion rhythm equal to about 50% of the saturating phase shift generated by the white light signal.

A comparison of the action spectra for phase shifting of the eclosion rhythm and the electrical response of the photoreceptor cells of the eye indicates that the two responses are mediated by different photopigments (Fig. 35). The eclosion rhythm is insensitive to light of wavelengths greater than 570 nm, but the eye is sensitive to light up to 666 nm. The spectral sensitivities of the circadian rhythms of the few insects examined to date are similar to those described for D. pseudoobscura. The circadian rhythms of hatching, oviposition, and eclosion in the moth Pectinophora gossypiella are sensitive to blue but not to red light (Pittendrigh et al. 1970). A more complete action spectrum for the initiation of the larval-hatching rhythm of P. gossypiella is very similar to the Drosophila spectrum; the most effective wavelengths were between 390'nm and 480 nm and wavelengths above 520 nm were ineffective (Bruce and Minis 1969). Another technique for determining the nature of insect photopigments takes advantage of the fact that in all known visual pigments the chromophore of the photopigment is the carotenoid derivative retinaldehyde. Carotenoids are synthesized by plants and can only be obtained by insects from their diet. Zimmerman and Goldsmith (1971) raised Drosophila melanogaster on diets without S-carotene and subsequently assayed both the photosensitivity of the circadian eclosion rhythm of the pupae to 15-minute monochromatic light signals and the photosensitivity of the compound eyes (Fig. 36). The photosensitivity of the visual receptors in carotenoid-depleted flies was about three log units lower than that of the carotenoid-supplemented flies. Most significantly, there was no difference between the deprived and supplemented flies with respect to the photosensitivity of their circadian rhythms. These results indicate that a carotenoid-deprived chromophore is not

Extraretinal Photoreception

:::

;

~ 3 51 c

...c

0.

.....

0

:> 0

::5 51 c: 0

5!-

>.

...c

"'

0

t ~ ~

~~

2

~ E ...c

133

~ t+ + -1

0

f

+

2

~

0

~ ~ 20

51

1:

0

5!-

~

+

+

30

y 3

t

t

~

..

.2 10

0. u

"~

0

...c 0..

4

cm2 )

Energy flux (log erg.seC:1

Fig. 36. Response-energy curves for phase shifts of the circadian rhythm (squares) and for the retinal action potential (circles) of D. melanogaster grown on aseptic diets with s-carotene (closed symbols) and without scarotene (open symbols). The energy scale gives the light flux used in both experiments, but the exposures were 900 times longer for resetting the rhythm. The responses of the visual receptors shown to the right of the break in the abscissa were obtained with a bright white light (Zimmerman and Goldsmith 1971. Science 171:1167-1168. Copyright 1971 by The American Association for the Advancement of Science). involved in mediating circadian rhythms. However, Zimmerman and Goldsmith (1971) point out that an alternative but less likely explanation of the results involves transmission of some carotenoids through the egg and preferential utilization of these carotenoids by the circadian system. Drosophila ERRs are probably located in the brain. Opaquing (painting) the anterior ends of the pupae prevented phase shifting by dim monochromatic light whereas painting the posterior half did not (Zimmerman and Ives 1971).

134

Herbert Underwood

The compound eyes have no role in the entrainment of the silkmoth flight activity rhythms (Truman 1974). A dim (one lux) light cycle was capable of entraining the flight activity rhythm of normal Hyalophora cecropia and Samia cynthia. However, silkmoths which had their heads, except the compound eyes, covered by black wax showed free-running activity rhythms even though the eyes were exposed to the light cycle. These results also indicate that the ERRs are located in the brain. The relatively large size of silkmoths allowed various surgical manipulations which indicated that the brain was the site of both the circadian oscillators controlling circadian rhythms and the site of the ERRs (Truman and Riddiford 1970, Truman 1972). For example, if the brains of silkmoths are removed early in adult development, the eclosion rhythm of the brainless animals is abolished-they emerge at random times with respect to the light-dark cycles. Implantation of the brain into the abdomen of brainless animals restores rhythmicity, and the site of photosensitivity shifts to the abdomen. Implanting various parts of the brain into brainless hosts showed that the optic lobes are not required for entrainment of the silkmoth eclosion rhythm and the clock and photoreceptors are probably located in the cerebral region of the brain (Truman 1972). The location of the clock in these insects, therefore, is different from its location in cockroaches and crickets, where intact optic lobes are necessary for the persistence of rhythmicity. The Role of ERRs in Insect Photoperiodism Many insects have evolved time-measuring mechanisms whereby they can discriminate between day lengths and so initiate seasonally appropriate events such as induction or termination of diapause. The reduced metabolic rate which occurs during diapause enables insects to overwinter or to withstand a dry season. Another adaptive photoperiodic response which is common in insects is the control of seasonal morphs. For example, the long days of summer induce the production of viviparous parthenogenetic (virginoparae) aphids which can reproduce rapidly in the presence of abundant food supplies, whereas short days induce the production of egg-laying (oviparae) offspring (Saunders 1976).

Extraretinal Photoreception

135

Insects appear to measure photoperiod in two different ways. Measurement in some species appears to use an endogenous daily (circadian) rhythm of responsiveness to light (Pittendrigh and Minis 1964, Pittendrigh 1972). The second kind of time-measuring system envisages an "hourglass" or interval timer (Saunders 1976). This hypothesis assumes the accumulation of a reaction product during the dark (or light) which is inactivated during the other phase of the light-dark cycle. The duration of dark (or light) is "measured" by the quantity of substance that has accumulated; if enough substance accumulates, a photoperiodic response begins. This process begins anew with each cycle of the lighting regimen and lacks endogenous periodicity. The existence of two different measuring systems for photoperiodic time, one involving a circadian clock, increases the likelihood that different photoreceptors are involved in the different measurement systems. The experiments of Pittendrigh et al. (1970) and Pittendrigh and Minis (1971) on P. gossypiella support this view. As discussed previously, Pittendrigh et al. (1970) demonstrated that several circadian rhythms in P. gossypiella are redinsensitive. However, the photoperiodic control of diapause in this insect is fully red-sensitive, suggesting the existence of separate photosystems for entrainment and photoperiodism. The results suggest that photoperiodic time measurement in P. gossypiella is accomplished by an hourglass mechanism. However, an equally plausible explanation is that the photoperiodic clock is a separate circadian oscillator coupled to light by a red-absorbing pigment. Among insects, photoperiodic responses appear to be solely mediated by ERRs located in the brain. Williams and Adkisson (1964) showed that the site of photoperiodic photosensitivity in the oak silkmoth (Antheraea pernyi) could be transferred from the head to the abdomen by transferring the brain to the abdomen. The photoperiodic response of A. pernyi pupae persists after removal of the optic lobes (Williams 1969). Similarly, Claret (1966a, 1966b) transferred photoperiodic photosensitivity along with the brain in the cabbage butterfly (Pieris brassicae). These studies suggest that the ERRs for photoperiod·ic photoreception and entrainment of the biological clock are both located in the cerebral lobe of the brain, but whether these receptors are identical or not remains unresolved.

Extraretinal Photoreception

137

A technically different approach was utilized by Lees (1960, 1964) to localize the photoperiodic photoreceptors in the aphid Megoura viciae. Lees maintained parent aphids on a short day length (LD 14:10) but gave two hours of additional illumination to discrete areas of the aphid by means of fine light guides. If the additional illumination was positioned over the brain, the aphids reacted to the long day length (LD 16:8) and produced virginoparous daughters, whereas illumination of other areas of the body, including the compound eyes, was less effective and the aphids reacted to the short day by producing oviparous daughters. Workers have investigated the photoperiodic spectral sensitivities of insects in an attempt to determine the nature of the photopigments. In general, most species are maximally sensitive to light in the blue-green region of the spectrum and largely insensitive to red, but exceptions are known (e.g., P. gossypiella). Saunders (1976) lists eleven red-insensitive species and five red-sensitive species. Assuming that these reflect the sensitivities of photopigments located in the brain, the screening effects of tissue overlying the brain may be a factor in modifying the intensities and spectral qualities of light. This screening will, of course, be of greater magnitude in larger insects. One of the most complete action spectra studies for photoperiodism in insects was conducted by Lees (1971) with Megoura (Fig. 37). Parent aphids were exposed to a one-hour pulse of monochromatic light beginning 1~ hours into the dark portion of an LD 13.5:10.5 light cycle or to a ~-hour pulse placed 7~ hours after the inception of the dark period. These positions corresponded to an early or late night break, both of which induced the production of the long-day (virginoparous) offspring. The early night

Fig. 37. Action spectrum for the maternal control of virginoparaproduction in the aphid Megoura viciae. In (A) near-monochromatic light was applied in the early night, 1.5 h after the beginning of a 10.5-h dark phase. The curve is drawn for incident energies at which approximately 50% of the parent aphids become virginopara-producers. (B) Action spectrum showing the effect of 0.5 h of nearmonochromatic light applied in the late night, 7.5 h after the beginning of darkness (Lees 1971. Reproduced with the permission of The National Academy of Sciences).

138

Herbert Underwood

interruption showed a maximum sensitivity in the blue (450 nm-470 nm) with a threshold at that wavelength of approximately 0.2 W•cm- 2 • The maximum sensitivity of the late night interruption was also in the blue but sensitivity extended to 600 nm. The differences in action spectra indicated that different events, possibly involving different photopigments, occurred during these two times of night. Other studies on spectral sensitivity indicate that the intensity threshold for photoperiodic responses may be quite low. The approximate intensity thresholds for photoperiodic responses of a total of eleven different species are listed in Saunders (1976) and Truman (1976). Most of these insects show thresholds between 0.1 and 10 lux. Truman (1976) suggested that more advanced groups of insects show an increased reliance on extraretinal photoreception. In primitive insects such as cockroaches or crickets, circadian rhythms are exclusively entrained via retinal pathways, whereas in insects which undergo a complete metamorphosis, such as flies and moths, photoreception may be mediated entirely by extraretinal receptors located in the brain. Truman (1976) speculates that brain receptors may be of value during metamorphosis because they are unaffected by the extensive reorganization occurring in external structures (such as eyes). VERTEBRATES AND EXTRARETINAL PHOTORECEPTION Extraretinal photoreception is widespread among the vertebrates. Extraretinal photoreceptors participate in the responses of vertebrates to light in at least six general areas: (1) entrainment of the biological clock; (2) photoperiodic photoreception; (3) control of certain biochemical rhythms in the pineal; (4) control of physiological color changes; (5) phototactic and photokinetic responses; and (6) orientation. However, among vertebrates the only extraretinal photoreceptive system which has been definitely localized is the pineal system of fish, amphibians, and reptiles. The pineal system has attracted a host of investigators because of the grossly "eye-like" morphology of some components of the system and because of the ease with which the members of this system can be removed and probed cytologically, biochemically, or electrophysiologically. A comprehensive review of the pineal system is beyond the scope of the present discussion, but a number of excellent reviews can be consulted for additional

Extraretinal Photoreception

139

information (Ariens-Kappers 1965, Wurtman et al. 1968, Wolstenholme and Knight 1971, Quay 1974, Relkin 1976), Pineal System:

Photosensitivity

In many fish, frogs, and lizards the pineal system is composed of two elements, both of which are derived embryologically as evaginations of the roof of the diencephalon. In these cases one component, the pineal organ proper, remains attached to the roof of the diencephalon. The other component, the parapineal organ, seems to originate either as an outpouching from the pineal organ (as in frogs) or as a separate diverticulum from the diencephalon (as in some lizards) (Kelly 1962). The parapineal component is most often located just beneath the skin of the head and in many lizards is highly differentiated into an eyelike organ (the parietal eye) complete with cornea, lens, and retina. A less specialized version is common in anuran amphibians and is termed the frontal organ. Although some lower vertebrates lack the superficial component (e.g., urodele amphibians, snakes), practically all vertebrates retain a pineal organ. Electron microscopy has revealed that both components of the pineal system in lower vertebrates possess a number of cells that resemble the photoreceptive cones of the lateral eyes (Wurtman et al. 1968, Wolstenholme and Knight 1971). However, many of these cells are more degenerate in appearance than the cones of the eyes and have been termed "rudimentary photoreceptors." Electrical responses to illumination have been recorded from the pineal organ of fish, the pineal organ and frontal organ of anuran amphibians, and from the pineal organ and parietal eye of lizards (Dodt and Heerd 1962, Dodt 1963, Dodt and Jacobson 1963, Dodt and Scherer 1968, Hamasaki and Dodt 1969, Hamasaki 1969, Hamasaki and Streck 1971). In general, the pineal organs give achromatic responses whereas the parietal eye and frontal organs show chromatic responses. The achromatic response typically involves inhibition of ongoing (spike) activity by all wavelengths of light. The chromatic response usually shows inhibition of electrical activity by shorter wavelengths and excitation by longer wavelengths of visible light. Some representative examples of the spectral sensitivity peaks of the (achromatic) pineal organs are: fish (Salmo irideus, 505 nm; Scyliorhinus caniculus, 500 nm; Pterophyllum scalare, 525 nm); amphibians (Rana temporaria, 560 nm); and lizards (Iguana iguana, Lacerta

140

Herbert Underwood

sicula and Acanthodactylus erythrurus, 570 nm) (Dodt 1963, Hamasaki and Streck 1971, Morita and Bergmann 1971, Dodt and Jacobson 1963, Hamasaki 1969). Examples of the peak chromatic responses from amphibian frontal organs and lizard parietal eyes are: amphibian frontal organs (R. temporaria and R. esculenta, 355 nm and 515 nm), and lizard parietal eyes (I. iguana and L. sicula, 460 nm and 520 nm) (Dodt and Heerd 1962, Dodt and Scherer 1968, Hamasaki 1969).

The achromatic responses of fish and amphibian pineal organs are typically very sensitive to light and threshold sensitivities of dark adapted (exposed) pineals range from 10-2 to 10-6 lux. The liaard's pineal is less sensitive and shows a threshold for exposed pineals of 4 lux (Hamasaki and Dodt 1969). In intact animals the skin and skull overlying the pineal reduce the amount of light reaching the pineal by factors of 100 to 1000. In some cases both graded (slow) potentials and action potentials (spike activity) are observed. The graded potentials are most noticeable in frontal organs of frogs and lizard parietal eyes. In the lizard parietal eye, for example, there is a graded potential similar to the electroretinogram (ERG) of the lateral eyes (Dodt and Scherer 1968, Hamasaki 1969). With light stimuli in the blue range this ERG shows an "on" response that consists of a relatively rapidly rising positive wave. Stimuli of longer wavelengths elicit a rapidly falling negative wave. The positive component is associated with inhibition of spike activity whereas the negative component is associated with excitation of spike activity. The photoreceptor cells in pineal and parapineal organs synapse with a single kind of neuron which sends its axons to the brain. A recent study by Engbretson and Lent (1976) shows the possibility of mutual interaction between the lizard's pineal and parietal eye. In the lizard Crotaphytus collaris, afferent fibers leaving the parietal eye pass over and possibly innervate the pineal on their way to the rest of the brain. Efferent fibers originating in the pineal organ can modify the parietal eye's response to light. The efferent nerves are not photosensitive but they are chemosensitive; norepinephrine and serotonin when applied to the pineal organ will initiate activity in these nerves. The dual pineal system present in many lower vertebrates is absent in birds and mammals. The pineal organs

Extraretinal Photoreception

141

of adult birds and mammals are apparently not directly photosensitive (Wurtman et al. 1968, Wolstenholme and Knight 1971, Quay 1974). The main cell type of bird and mammalian pineals is the pinealocyte, which is believed to belong to the same cell line as the photoreceptor cells of the lower vertebrates. The pinealocyte is an active secretory cell and is involved in the synthesis of polypeptides and biogenic amines such as serotonin and melatonin (Fig. 38). In this regard, the photoreceptive cells in the lower vertebrates, in spite of their obvious photoreceptive ability, are also active sites of chemical synthesis of such products as the biogenic amines. The only known innervation 0

0

" C.OH

N

g HO(JrJ'XC'OH

~II I 1) ::,... N

2

I Tryptophan I

NHz

15J/ /

N

c~

Ju

-

f'''· I

13 12

zw

I

1-

X X

2.1 5.2 3 •9 3.1

hardwood leaf radiance

vertical irradiance, 0731 EDT (Fig. 74d)

sweet gum trunk radiance, 0741 EDT

broadleaved litter radiance

2.0

holly leaf radiance

aSee footnotes to Table 15.

9.6 x 10

sweet gum radiance (1 ) X

10

102

1.1

sweet gum leaf radiance (20') X

1.7

10

vertical irradiance, OS57 EDT (Fig. 74d)

X

51S

527

529

614

51S

62S

527c 0.42 0.54

529c

--

626

633

--

531c

629

--

0.3S

o.os

0.37

0.31

520

10

0.11

617

10 2

0.50

0.30

0.11

531

519

615

10

10

102

X 10 3

SOd--Cedarville

X

3.1

sweet gum trunk radiance, 0625 EDT

X

X

2.5

vertical irradiance, 0631 EDT (Fig. 74d)

SOc--Cedarville

t-t

0 :::1

1.0

......, ......,

ill

0 t-o 0 "!

(")

s:::: ill

0

s::::

'-'· Q

'a

ill

(")

:::1

~

...

rt

'§.

'-'·

t-o

~

rt

:::1

~

g"!

~

-

10::: :::J o_

z

0

0.4

1-

GRASS •

J ::0

SWEET GUM LEAVES

•

::::>

1-

u

::0

X

w 0.3

_j

•

HOLLY LEAVES (UNDERSIDES)

OLDFIELD•

c w

; they appear to learn that it is associated with nonreinforcement. This is demonstrated by the fact that stimuli that resemble the SL> stimulus evoke more responding than the SL> stimulus itself. What this implies is not exactly clear. It may imply that an SL> stimulus acts at a cognitive level to inhibit the reinforced instrumental response (e.g. Spence, 1937; Mackintosh, 1983), or it may imply, as other theorists have suggested, that the sf:, stimulus elicits behaviors which compete with the reinforced response. Staddon (1983) has suggested that SL> stimuli come to elicit behaviors which are unrelated to either the reinforcer or the reinforced response. We know, for instance, that on a food-reinforcement schedule, periods when food is unavailable are characterized by a variety of stereotyped behaviors which are unrelated to feeding. These have variously been called adjunctive behaviors (Falk, 1971), interim responses (Staddon and Simmelhag, 1971) or schedule-induced responses (see pp. 83-91, this chapter), and appear to be manifestations of behaviors which are representative of motivational systems other than feeding (cf. Staddon, 1977; but see Timberlake and Lucas, 1985). The fact that these adjunctive behaviors are elicited by stimuli which signal nonreinforcement may be important in this context, since the inhibitory generalization gradients obtained from SL>s may reflect the suppression of instrumental responding by competing adjunctive behaviors elicited by the SL> as a signal for nonreinforcement. Studies of discrimination and generalization tell us quite a number of things about learning. First, they suggest that external stimuli can come to control the emission of instrumental responses. In such circumstances the animal may learn to associate the S 0 with reinforcement (e.g. Trapold and Overmier, 1972) or to associate the SD with the fact that responses emitted 67

Instrumental learning

Excitatory generalization

w

I

1-

u z ::::)

-, 0

.

01

'Q)

c

...,

Q)

references during this sensitive period appears to be relatively permanent (Hess, 1973). Both of these characteristics suggest that this kind of food preference learning resembles other kinds of imprinting processes. Imprinting in mammals So far, we have discussed imprinting primarily in relation to precocial and altricial birds. What evidence is there that learning processes like imprinting exist in other groups of animals? A process which closely resembles imprinting can be identified in the formation of the mother-young bond in sheep and goats. Collias (1953) found that if the lamb is removed from the ewe at birth and returned 2-3 hours later, the ewe will refuse to accept the lamb. However, if the lamb is removed 1 hour after birth and returned 2-3 hours later, it will subsequently be accepted by the ewe. Thus, the sensitive period for the formation of this mother-young bond appears to be 1-2 hours after birth- a finding which has been replicated in goats (Klopfer, Adams, and Klopfer, 1964). This kind offilial imprinting differs from that in precocial birds in that,_ the learning appears to be done by the parent and not the offspring. However, since sheep and goats give birth on a number of 260

Phase-specific learning

100

90

Qlll f-::J ...J

w ::J lll~

6;::

(L ll1 .

80 70

60

lllZ

ww

w wo::

50

'j

40

0::

t?? ~

Z(9

Wz

lEWet::~

30

(Lf-

20

Operant level ot naive chicks

10

2

3

4

5

6

7

8

9

10

DAYS OF AGE

Figure 9.2 Preference for pecking at a triangle-green stimulus only had a

lasting effect when the chick subjects were reinforced for this behavior between the ages of 3-5 days. (After Hess, 1973)

occasions during their lifetime, this learning can recur more than once. It thus differs from the relatively stable and durable filial and sexual imprinting that occurs in birds. 4 Learning mechanisms in imprinting It is important to be clear about what is special about imprinting as a learning process. It is that preference learning occurs only during a sensitive period, and that the preference that is formed is fairly durable and resistant to change. It is also important to be clear about what is learned. A "preference" is not a behavior as such, and we infer preference either from the fact that an animal is willing to approach and follow a stimulus or from the fact that it 261

Phase-specific learning

chooses that stimulus in preference to others in a choice test. These are all points we must be clear about before we can begin to discuss what learning processes might be involved in imprinting. Traditionally, theorists have been concerned about discovering the role that both Pavlovian and instrumental conditioning might play in imprinting, or, if they are not involved, what specialized learning processes can be devised to account for imprinting. The characteristics that make imprinting special as a . learning phenomenon by no means exclude the involvement of Pavlovian or instrumental learning, as we shall see below. It is the relative resistance ofthe imprinting behavior to subsequent change that makes it unlikely, however, that an explanation of imprinting could be couched solely in terms of conditioning principles. While simple associative factors might contribute to the acquisition of the preference, some other factors need to be invoked to explain its subsequent durability. Imprinting as associative learning Moltz (1960, 1963) originally claimed that filial imprinting could be thought of simply as a sequence of Pavlovian followed by instrumental conditioning. He suggests that on emerging from the egg, hatchlings have little fear of the world around them and orient toward any relatively large, attentionprovoking object. By a process of Pavlovian association the object which is I oriented toward (the target or imprinted stimulus) becomes associated with these low anxiety levels such that, when the hatchling does eventually begin to acquire fear to unfamiliar objects, the imprinted stimulus becomes a source of anxiety reduction which instrumentally reinforces the following response. Certainly, there is some evidence that imprinted stimuli do have fear or arousal reducing effects (Bateson, 1969; Hoffman, 1968), and indeed, animals do appear to approach stimuli which have been paired with fear reduction (Leclerc, 1985; seep. 27). Imprinted objects also acquire reinforcing properties and act in much the same way as do conditioned reinforcers (see pp. 69-70). For instance, young ducklings or chicks can easily be trained to press a pedal to switch on a motor bringing the imprinted object into motion, and will also perform an instrumental response in order to present themselves with different angled views ofthe imprinted object (Bateson and Reese, 1969). However, there are a number of problems with a simple associative model of this kind. First, while imprinting can in some instances be reversed by certain kinds of procedures (e.g. Salzen and Meyer, 1967, 1968), the following response appears, by and large, to be insensitive to punishment during the sensitive period -something not predicted if conditioning processes are acting alone. Subjecting young

262

Phase-specific learning

chicks to electric shocks contingent upon approaching the imprinted object during the sensitive period actually fails to prevent approach behavior (Kovach and Hess, 1963). Second, imprinted responses do not appear to show a readiness to extinguish when reinforcements are subsequently withheld. For instance, altering the pecking preference of a 3-4-day-old chick with food reinforcement appears to have a permanent effect, even when food reinforcement is subsequently discontinued (Hess, 1962, 1964). Third, if the following response were a simple conditioned behavior then we might expect that any target stimulus would be equally effective in establishing approach and follow responses. However, it is quite clear that some stimuli- especially stimuli possessing characteristics that resemble the hatchling's own species -are more effective as filial and sexual imprinting stimuli. While we now know that Pavlov's principle of equipotentiality of stimuli is probably incorrect (see pp. 125~ 7), it is still difficult to understand why conspecific features are so salient unless there is some genetically predetermined tendency to prefer those features (cf. Hess, 1973; Immelmann, 1985: 283~5). Nevertheless, none of the foregoing precludes the possibility that associative mechanisms are at work somewhere in imprinting; but if they are, they still need to be clearly identified, and traditional conditioning principles would not appear to provide a complete picture of the imprinting process. Imprinting as a special learning mechanism Lorenz (1935) first pioneered the view that imprinting was a special learning mechanism dedicated to a particular function such as kin recognition. Many theorists were subsequently skeptical of this claim - if only because it appeared to violate scientific principles of parsimony and run contrary to generalized laws of learning that were popular during the 1960s and 1970s (e.g. Bateson, 1966; Hoffman and Ratner, 1973). Nevertheless, even some years ago theorists did admit to the possibility of genetically-determined biases in imprinting. The tendency of precocial hatchlings to imprint more readily to stimuli which possess conspecific features is important in this respect, and led Hess (1973) to suggest that we must consider that young chicks innately possess a schema of the natural imprinting object, so that the more a social object fits this schema, the stronger the imprinting that occurs to the object. This innate disposition with regard to the type of object learned indicates that social imprinting is not just simply an extremely powerful effect of the environment upon the behavior of an animal. Rather, there has been an evolutionary pressure for the young bird to learn the right thing - the 263

Phase-specific learning

natural parent - at the right time - the first day of life - the time of the sensitive period that has been genetically provided for. (Hess, 1973: 380) There are two issues here that should be clarified: one suggests that there is a good deal of genetic predisposition in imprinting, the second suggests that imprinting is a specialized learning process which differs from other more conventional learning processes. The two are quite independent issues. For instance, we have already seen in earlier chapters that some animals might bring a genetically-determined predisposition to a learning situation, but the actual learning appears to be mediated by conventional learning mechanisms (see, for instance, taste aversion learning, Chapter 6, pp. 183-90). In such cases the genetic or species predisposition determines what kinds of stimuli will be attended to, and as a consequence, what other stimuli they will become associated with. The fact that ducklings imprint more readily to conspecifics than other more neutral stimuli may be another example of this geneticallydetermimid stimulus selectivity (see also pp. 125-7). However, there is some evidence on the neurophysiology of imprinting which throws some light on these issues. One concerns the brain centers involved in imprinting, the other concerns the states of neuronal plasticity that exist during sensitive periods. Bateson and his colleagues have demonstrated that filial imprinting in domestic chicks enhances biochemical activity in the roof of the forebrain. For instance, when chicks were allowed to imprint using only one eye (the other being covered by a patch), the biochemical activity was 15 per cent higher on the trained side of the forebrain roof than on the untrained side (Horn, Rose, and Bateson, 1973). Similarly, there is a correlation between the strength of the imprinting response and biochemical activity in the roof of the anterior forebrain (Bateson, Horn, and Rose, 1975). The critical brain center which appeared to be the focus of these effects was the intermediate region of the medial hyperstriatum ventrale (IMHV) (cf. Horn, McCabe, and Bateson, 1979), and this center had connections to visual projection areas (Bradley and Horn, 1978). In subsequent studies involving ablation of IMHV, this area was found to be crucial to the recognition of the imprinted object. Subjects which had IMHV lesions showed no subsequent preference for the object with which they had earlier received imprinting training (McCabe, Horn, and Bateson, 1981). Furthermore, while IMHV lesioned chicks failed to learn the characteristics of a moving imprinted object, they could still successfully learn instrumental responses involving a visual discrimination for heat reinforcement (cf. Bateson, 1984). This suggests that IMHV is not a generalized center

264

Phase-specific learning

mediating visual recognition, but is specialized in mediating certain kinds of recognition- especially those concerned in imprinting, and possibly with the storage of information acquired during imprinting. In other neurophysiological studies the readiness to learn early in life can often be found to be correlated with the formation of new synapses and dendritic branching and to terminate with the disconnection of neurons and their physiological death (e.g. Wolff, 1981; Bischof and Herrmann, 1984). Such periods of neural plasticity often coincide with sensitive periods for imprinting. The evidence is beginning to suggest that this may not only account for phase-specific learning but that it might also account for the relative durability of the learning that occurs during a sensitive period. For instance, Immelmann suggests that In view of the possible correlations between sensitive phases and neuronanatomical development, it seems likely that the critical mechanism underlying the temporal limits of plasticity might consist of genetically determined programs for the time course of neuronal plasticity. These programs could determine at what age and for how long morphological changes are pssible and thus put constraints on the age period during which information can be stored with a high degree of permanance. (lmmelmann, 1985: 283) The fact that at the end of a sensitive period particular brain centers may become effectively neuronally inert may indeed account for the permanence of information learned during the sensitive period. If information acquired during imprinting is stored in a specific center which is characterized by this brief period of neuronal plasticity followed by neuronal inactivity thereafter, then this could go some way to explaining the relative permanence of preferences acquired during sensitive periods of imprinting. 5 Summary

What the above evidence is beginning to suggest is that (1) animals may indeed have genetically predetermined tendencies to imprint to some objects rather than others, but this does not necessarily imply that imprinting is a specialized learning system but may simply reflect stimulus selectivity predispositions found in learning generally; and (2) special areas of the brain appear to be important in the storage of information learned during imprinting and these areas may be characterized by the phasic nature of their neuronal plasticity. This results in information acquired during imprinting being relatively resistant to change later in life.

265

Phase~specific

learning

This still leaves open the question of what mechanisms are implicated in the actual learning about imprinted objects. These may indeed be basic associative mechanisms, but we must await further evidence on this point. Song learning in birds There are just under 9 000 extant species of bird in the world, of which around 4 000 are perching songbirds (the passerines), that are differentiated from other species by their striking and often complex vocalizations. While most passerine species possess their own distinctive song, individual members of a species may elaborate on and augment the basic species theme. Similarly, while some aspects of a species song appear to be innately determined, individual birds clearly have to learn important features of their species song. This learning is interesting in that, while it employs many distinct characteristics, it is in particular phase~specific and durable -characteristics which set it aside from simple forms of adaptive learning and which make comparisons with imprinting possible. An example of the characteristics of this song learning is provided in a study on white-crowned sparrows (Zonotrichia leucophrys) by Marler (1970). The song of the white-crowned sparrow consists of an initial whistle followed by a more individual complex trill. The whistle is highly stereotyped, but the trill represents a locality dialect which is learned by the bird. Within the first 150 days post-hatching, premature singing begins with both the whistle and the trill developing through a series of stages which gradually begin to approximate the adult song. At about 8 months of age, when the bird reaches sexual maturity, the song crystallizes and no further learning takes place, either in that year or in subsequent years. Thus, as the bird reaches sexual maturity it is as though the learning mechanism is switched off, and what has been learned up to that point remains durable and resistant to change. First, let us look at the characteristics of the learning processes that contribute to song learning. I Characteristics of the learning process We know that learning is an essential aspect of the development of a mature bird 's song because birds that are reared in isolation will develop abnormal songs (Kroodsma, 1982). Nevertheless, as we shall see below, although learning is essential, there are also some predispositional features of this learning.

266

Phase-specific learning

Acoustical feedback While some aspects of song structure develop even in birds that have been deafened, severing the auditory feedback loop will result in birds learning abnormal songs (Konishi, 1965). This feedback appears to be important only in songbirds, however, since the calls of many nonsinging birds such as doves and chickens are unaffected by early deafening (Konishi and Nottebohm, 1969). Thus, songbirds must hear themselves sing in order to learn the species song correctly. However, even in deafened birds, there are dear species differences in the songs that emerge, suggesting at least some genetic predetermination (e.g. Marler and Sherman, 1982). Imitation and improvisation Imitation appears to have an important, but complex, effect on song learning. For instance, birds that are raised with exposure to their conspecific song usually develop songs which are more clearly normal than birds reared in isolation. Some species, such as the northern mockingbird, may mimic many species, whereas others, such as North American sparrows, will never mimic heterospecific songs (Baylis, 1982; Dowsett-Lemaire, 1979; Marler and Peters, 1980). However, many species of bird are as likely to improvise as to imitate directly, although the imitation process appears to facilitate this improvisation. They tend to utilize the conspecific song as a kind of template around which they can improvise. Juncos, for instance, appear to utilize a model as a template for developing their own individualized and highly stereotyped song. As a consequence, juncos share very few song syllables in common (Marler, 1967; Marler, Kreith, and Tamura, 1962). Hearing a conspecific model performing the species-specific song can also act as a trigger for the song-learning process. For instance, some birds can readily learn heterospecific sounds, but only if they are intermingled with conspecific sounds (cf. Marler and Peters, 1980). The presence of the conspecific sounds appears to act as a trigger which permits the subject to imitate what follows. Nevertheless, some other species of bird rely almost entirely on improvisation rather than imitation. Blackbirds, for instance, can recombine syllables to form completely new songs even when reared in acoustical and social isolation (Hall-Craggs, 1962; Thielcke-Poltz and Thielcke, 1960). Sensitive periods In passerine birds, song learning appears only during what seems to be a sensitive period during early development. There appear to be two stages to

267

Phase-specific learning

this developmental process: first, the sensitive period itself when learning occurs, and second, the process of crystallization when no further learning takes place (Kroodsma, 1981; Marler, 1970). The sensitive period itself varies from species to species. In species such as the white-crowned sparrow, the song sparrow, and the zebra finch, the sensitive period occurs very early in the first year (Immelmann, 1969; Kroodsma, 1977, 1978). In other species, the sensitive period may last for the whole of the first year and crystallization will be less well defined. Such species include the European and cardinal chaffinch (Lemon, 1975; Nottebohm, 1970). However, a few other species, although learning the rudiments of their song early in the first year, do not have a well-defined crystallization stage and may continue learning in subsequent years. This produces a large song repertoire which may have important biological functions for those species concerned (for example, female canaries exposed to the songs of males with large repertoires lay more eggs than those exposed to males with small song repertoires: Kroodsma, 1976). Nevertheless, well-defined sensitive periods and crystallization stages are more commonly the rule than the exception. Memorial abilities During song learning there is a sensory phase, when the bird is exposed to the model song, followed by a sensorimotor phase, where the bird begins to produce an amorphous imitation of the model, through to the development of the full song and eventual crystallization. What is interesting is that the sensory and sensorimotor stages may be separated by several months during which the bird neither hears the model nor rehearses the song (Marler and Peters, 1982). Similarly, when the song has crystallized it remains relatively unchanged, even though the bird does not rehearse the song during the nonbreeding season (Konishi and Nottebohm, 1969). Genetic predispositions Even when birds have been experimentally deafened early in life, Marler and Sherman (1982) found significant species differences in the frequency characteristics and the duration, structure, and repertoire size of the songs that these birds eventually learned, suggesting that at least some aspects of the song are preprogramed. Similarly, in species such as song sparrows and canaries, certain aspects of the song temporal pattern develop quite normally in deafened individuals even though the refinements of the song may be abnormal (Guettinger, 1981; Marler and Sherman, 1982). Also, the fact that around 85 per cent of song birds learn to mimic only their own species songs and reject heterospecific songs does possibly suggest a genetic predisposition 268

Phase-specific learning

(although in some cases it may be the social conditions during rearing which determine this preference; cf. Immelmann, 1969). 2 Is song learning a specialized learning process?

Any theoretical account of the learning process that underlies song acquisition must take into account a number of salient facts. First, the learning tends to occur during a brief sensitive period early in life and subsequently is relatively resistant either to change or forgetting. Second, there are certain genetic predispositions and species-specific factors which need attention: certain components of the song appear to be learned regardless of acoustical and social isolation, and a large number of species prefer their own species song and reject others. Third, song learning appears to possess remarkable memorial capacities: birds retain sensory information for months on end before performing the song, and also retain the song even after non-use during the months of the non breeding season. Associative learning There appears to be little evidence that song learning in birds is shaped either by instrumental or Pavlovian contingencies (cf. Marler, 1984), but that the presence of certain triggering conditions is sufficient for the learning to take place. However, two points should be raised here. First, there is some evidence that contingent social interaction may be necessary and sufficient for song learning to proceed in some species (Payne, 1981); and if there is any shaping from instrumental contingencies social interaction may be the most viable reinforcer. Second, hearing the conspecific song can act as a very powerful reinforcer itself to some species of bird. For instance, Stevenson (1969) found that presentations of the conspecific song to chaffinches contingent upon a perching response acted to increase the frequency of that response. Specialized brain centers A number of studies have shown that certain forebrain nuclei (for example, nucleus hyperstriatum ventrale, pars caudale (HVc) and the nucleus robustus archistriatalis (RA) occur only in birds capable of song learning and are absent in suboscine or nonpasserine birds which lack this capacity (Nottebohm, 1980; Paton, Manogue, and Nottebohm, 1981; Konishi, 1985). Similarly, variations in the size of the HVc and RA nuclei can be found within species: for instance, canaries with larger numbers of syllables to their song tend to have larger HVcs and RAs than those with a smaller number of 269

Phase-specific learning

syllables (Nottebohm, Kasparian, and Pandazis, 1981), and the size of HVc and RA nuclei fluctuate seasonally, becoming larger in the spring season when the canary is most vocal (Nottebohm, 1981). These nuclei appear to be involved in both the sensory and sensorimotor aspects of song production since they contain both motor and auditory neurones (e.g. McCasland and Konishi, 1981; Konishi, 1985). Single-unit recordings from the HV c suggest that both motor and auditory nuclei are involved in song learning, because certain neurones of the HV c have been shown to fire only just prior to and during the delivery of a specific learned piece of the bird's song (McCasland, 1983; Margoliash, 1983). Lesion studies of the HVc nucleus also support its involvement in song learning. HVc-lesioned canaries, for instance, show fewer changes in their songs than birds lesioned in the hypoglossus region (Nottebohm, 1980; Nottebohm, Manning, and Nottebohm, 1976; Nottebohm, Stokes, and Leonard, 1976). Furthermore, there even appears to be a lateralization of function in the HV c nuclei; lesions of the left HV c cause greater loss or song deterioration than a lesion of the right HVc (Nottebohm, Manning, and Nottebohm, 1976). It is not entirely clear yet what the existence of these specialized brain nuclei means for theories of song learning. The discovery of a specific brain center related to song learning does appear to correspond with the finding of similar specialized centers for imprinting, and may obviously reflect the evolution of special learning requirements. It is possible that these song control nuclei may contain the auditory templates that appear to exist within different species (e.g. Marler, 1984), or they may possess the neural characteristics necessary to delineate sensitive periods for song learning (see pp. 264-5). We must await further neurophysiological studies for clarification of these issues. 3 Summary

We can say tentatively that song learning does appear to be a specialized learning process. Songs are not obviously learned by processes of instrumental or Pavlovian conditioning, and song learning appears to possess certain predispositions which indicate the possible existence of auditory templates within species. In many species the conspecific song is learned rapidly during a short period of early development, and is durable over many seasons without being lost during periods of disuse. The existence of brain nuclei specializing in song learning and production suggests that song learning may be a highly specialized process, but how these nuclei mediate any specialization that may exist has yet to be determined.

270

Phase-specific learning

The biological function of phase-specific learning The general features which characterize phase-specific learning are (1) the existence of sensitive periods when the learning occurs, and (2) the high degree of stability and durability that this learning bestows on the animal. Both imprinting and avian song learning possess these characteristics in varying degrees. The next question to ask is why phase-specific learning should have evolved as a specialized learning adaptation possessing these characteristics. It can be seen from the contents of this chapter that phase-specific learning has been discovered and studied almost exclusively in birds ( cf. Immelmann and Suomi, 1981), and the existence of such learning may well be related to the rapid ontogenetic development in these animals. Birds reach their adult size and weight within 1 per cent of their normal life expectancy compared with a figure of over 30 per cent for most mammals. This rapid ontogenetic development appears to have the benefit of enabling birds to reach the ideal ratio between body-weight and a constant wing surface as quickly as possible (cf. Immelmann, 1985; Mason, 1979). Phase-specific learning may represent the learning and memorial equivalent of the bird's rapid bodily development, with conventional learning processes perhaps being inadequate to cope with the kinds of information that need to be acquired in the brief period between hatching and adulthood. Furthermore, while phase-specific learning does have distinct characteristics it is adaptable in that sensitive periods do appear to adapt to specific ecological conditions. In precocial birds, for instance, the sensitive period begins as early as a few hours after hatching when the hatchling needs to recognize and follow the parents. With altricial species these periods are prolonged by the fact that the young bird has more time available in which to learn the characteristics of the parents. In many species the termination of the sensitive period is geared to the age at which the young bird leaves the parents, and this process appears to protect what is already learned and prevent "misimprinting" on either wrong species or wrong habitats. Very brief sensitive periods have, for instance, been found in species which disperse and form flocks very early and where the dangers of"misimprinting" are high (cf. Immelmann, 1985). All of these ecological conditions favor the evolution of a durable early-learning mechanism which is particularly sensitive to certain environmental stimuli.

271

Phase-specific learning

Chapter summary 1 Certain types of learning are characterized by the fact that they occur only during a specific, brief period of the animal's lifetime. This kind of learning is called phase specific and occurs during a sensitive period of the animal's development. 2 There are two main types of phase::.specific learning that have been studied. These are imprinting and avian song learning. 3 The young of certain precocial birds (such as chicks and ducklings) would, on emerging from the egg, learn to approach and follow the nearest moving object. This attachment is known as imprinting, is subsequently difficult to break, and occurs during a specific period of the bird's early development. 4 Hatchlings will imprint to many different stimuli but appear to have a preference for their own species. 5 The actual duration of the sensitive period for imprinting does not appear to be rigorously fixed but is determined by hormonal and neuronal changes. 6 As well as filial imprinting (attachment to conspecifics) there appear to be other kinds of preference formation which share similar characteristics to imprinting. These include locality imprinting, food preferences, and motheryoung attachments in some mammals. 7 We are still unclear as to the actual mechanisms which underlie imprinting. However, (1) animals do appear to have genetically predetermined tendencies to imprint to some objects rather than others, and (2) special areas of the brain appear to be important in the storage of information learned during imprinting. These latter areas are characterized by the phasic nature of their neuronal plasticity which gives imprinting its characteristic of being resilient to change in later life. 8 Many species of bird have their own distinctive song. The acquisition of such songs appears to reflect a combination of genetically-predetermined factors and individual learning. The learning component is normally characterized by its phase-specific nature and its durability. 9 Birds reared in complete isolation often develop abnormal songs, suggesting that learning is a vital component of song acquisition. Important factors include acoustical feedback, imitation and exposure to relevant information during a sensitive period of song development. 10 Song learning appears to be a specialized learning process because (1) songs are not obviously learned by processes of conditioning, and (2) song learning possesses certain predispositions which indicate the existence of auditory templates within species.

272

Phase-specific learning

11 There appears to be a specific brain center related to song learning, which corresponds to the fact that similar specialized centers eJ~1llfNT"'5 OF T'Hti' T'Otf- .5E IVS tJA Y ('OINT J( CAVDAL ME.DV~~~~ liND CtJtfD

SPt~lf£

8£TWGEH 10 AND .2/.0AYS

FrG. 3. At left, tracings of frontal sections of the mid-medulla in a series of rats from 7 to 25 days old to show the growth of the corticospinal tract which forms one of the pyramids (black) in the medulla at this level. At right, the quadrants of the motor-sensory cortex are diagrammed on an outline of a mature rat's brain. The arrows indicate some major areas of projection of the corticospinal tract to the caudal medulla and (cervical) spinal cord.

tract and differentiation of the isocortex, they suggested hypothetical mechanisms for functional plasticity. The first was the response of the developing corticospinal tract to unilateral ablation of the motor-sensory cortex or to hemispherectomy in the newborn. In the normal rat, the corticospinal tract in the caudal medulla is completely crossed. After the ablation in the newborn, a considerable number of axones did not cross but formed a miniature uncrossed corticospinal tract whose fibers terminated, as far as our methods can reveal, in the same nuclei in the caudal medulla and spinal cord as normal crossed fibers would have. This aberrant response could be stimulated by ablation in rats five days old, but by nine days it was scant, and by two weeks we could not longer detect it (Leong & Lund, 1973). Older animals did not form the uncrossed tract (Hicks & D' Amato, 1970). The time sequence of the development of the aberrant, uncrossed tract paralleled that of the normal as far as we have studied it, progressing through the medulla in the third week. Whether the little tract functioned usefully was not established. About the only difference we could find between rats hemispherectomized at birth or maturity was the length and character of stride during running on a measured track as shown in slow motion movies (Hicks & D' Amato, 1970). A normal part of the rat's stride was a forward thrust of the forefoot, fanning of the digits, and then precise downward placing of the foot, digits first, on the ground (Fig. 2). Animals operated on as infants retained these characteristics, whereas those hemispherectomized at maturity lost some of the thrust and had a shorter than normal stride. We

ADAPTATION AFTER EARLY BRAIN INJURY

43

suggested that the uncrossed tract in the animals operated on as infants, by providing bilateral corticospinal tract representation, might have spared the stride, but this is wholly speculative at present. The other alteration following ablation of motor-sensory cortex in the newborn was a transformation in the differentiation of the isocortex adjacent to the site of the ablation. It was seen regularly in the medial dorsal isocortex in all animals in a series with subtotal motor-sensory cortex ablations: what would normally have differentiated into medial isocortex instead became an extension of the cingular cortex. Cingular cortex, forming the medial cortex of the hemisphere, has a peculiarly distinctive cytoarchitecture, including among other features, a condensation of small neurons in layer 2. The result of the ablations was that the area of this kind of cortex was considerably increased, rising lateralward toward the vertex. The experiments emphasized how immature and modifiable the infant cortex was, and recalled that many young neurons continued to migrate into the cortex during the early postnatal period (Hicks & D' Amato, 1968), and that many afferent and efferent fiber connections as well as local synapses were yet to be foimed. The cingular cortex transformation led one to think that regions of cortex normally destined for one functional role might be altered to take on another one. Two rats of a dozen that had had both motor-sensory cortices removed at birth retained their tactile placing reflexes as expected, but after several months these reflexes began to fade out. In one animal they were lost unilaterally; in the other, they were lost bilaterally. The possibility is being investigated that posterior cortex might have been transformed into aberrant motor-sensory cortex whose slowly forming "corticospinal" axones had turned off the responses. Ablations of this supposedly transformed cortex with Fink-Heimer-Nauta techniques are the first step toward examining the hypothesis. Morphologic and functional aspects of different parts of the motorsensory cortex. Much of what we believe we know about the function of the nervous system is inferred from the effects of surgical ablations of parts of it or the consequences of disease in animals and humans. Mechanisms for some functions are certainly concentrated in particular parts, but the nervous system is not an aggregate of independent units that do this or that, but of closely interlocked systems. Ablation of a part may tell us by the resulting functional deficits what machinery resided there, but the deficits may also reflect the expression of mechanisms situated elsewhere that are no longer held in check by what was ablated. To find out more about how the two motor-sensory cortices and their derivatives, the corticospinal tracts, governed the placing and positioning of limbs in locomotion, we have begun to search for relations between the anatomic projec. tions of the corticospinal tract from different parts of the motor-sensory cortex and disturbances of motor functions resulting from ablation of those parts. We have divided the cortex arbitrarily into quadrants (Fig. 3 ), and we are studying the

44

SAMUEL P. HICKS AND CONSTANCE J. D'AMATO

anatomy and function from birth to three weeks, and to about seven weeks. We have preliminary results from these experiments that can be summarized as follows. The more caudal projections, with which we will be concerned here, from the anterior lateral quadrant went largely to the spinal trigeminal and lateral reticular nuclear regions of the caudal medulla. There was very little projection to posterior column (gracile and cuneate) nuclei or into the spinal cord. This quadrant overlapped the cortical regions mapped by Woolsey (1952) and his colleagues, in which motor and sensory functions involving head, face, mouth, and probably forefoot were partly represented. Woolsey's cortical regions were mapped by recording potentials evoked in the cortex by stimuli over the body surface, or by stimulating the cortex directly and recording movements of various parts of the body. When we ablated this anterior lateral quadrant on one side in the newborn, or in three- or seven-week-old rats, tactile placing or locomotion on a flat track were not impaired, and direct observation did not reveal any handicap in traversing a narrow path, but movies of this activity remain to be analyzed. The anterior medial quadrant projected predominantly to the dorsal column nuclei. Tactile placing laterally with the forefoot was sluggish after ablation of it and sometimes could not be elicited. The hind foot opposite the lesion was retrieved somewhat awkwardly when it slipped off the narrow pathway (normal animals slipped, too, but retrieved instantly). The quadrant overlapped the head, mouth, and forelimb motor region of Woolsey. The projections from the posterior medial quadrant, which included some of both motor and sensory representations of the posterior half of the body in Woolsey's topography, went to the posterior column nuclei and the spinal cord. Of all quadrants, ablation of this one, on one side, affected tactile placing responses and locomotion on irregular terrain the most seriously. If the animal was carefully placed on the edge of a flat track, it would let the limbs opposite the lesion dangle over the side until it initiated walking again, resembling the animal at left in Fig. 2. On a narrow path these limbs, contralateral to the lesion, slipped off easily, grasped awkwardly, and were retrieved clumsily. On a fait surface, walking and running appeared normal. Anatomic data are not yet complete on the posterior lateral quadrant, but ablation of it, on one side, impaired placing and locomotion to about the same degree the anterior medial quadrant ablation did, and so far as our experiments have progressed, affected reflexes only on the side contralateral to the lesion. The region corresponded partly to the sensory cortex, called somatic sensory 2, that had some bilateral representation of the body surface in Woolsey's topography. We mentioned reliance earlier on testing lateral tactile placing and observation of placing and positioning of limbs in natural locomotion. We have used or explored a variety of examinations, including other tests of tactile placing, visual placing, hopping, stepping, running on horizontal ladders and on narrow longitudinal strips, climbing, and such elements of a neurologic examination that can be applied to a rat (Hicks & D' Amato, 1970; Hickset al., 1969). One cannot be sure

ADAPTATION AFTER EARLY BRAIN INJURY

45

in advance that one has selected proper tests for an experiment, but in the present studies we have continued the tactile placing tests with fore and hind limbs, observations of locomotion on a board track, and walking and running along elevated narrow pathways (Fig. 2). The latter were a length of wood strip a half-inch wide and two approximated parallel quarter-inch dowels. Traversing these latter was just difficult enough for normal rats to slip a foot or a limb off occasionally and to magnify deficits in animals with ablations. Whether the ablations were performed at one day, three weeks, or about seven weeks, the subjects were members of a litter that were introduced to the experimental apparatus at about 13 days of age and allowed to explore it in daily sessions (but less often after they were about five weeks old) until termination of the experiment. Two to four members of a litter were operated on and two or more normal littermates went through the same performances as controls. The animals ran across or otherwise negotiated (Fig. 2) the track and narrow pathways enthusiastically (or at least persistently) by the middle of the third week. Even quite handicapped animals persisted at this activity. Besides contributing to the preliminary results of the quadrant ablations just outlined, this approach suggested that there were differences in effects of similar ablations at the three different ages, and differences in effects of bilateral ablations depending on what quadrants were removed. Seven-weeks-old rats seemed to show more flagrant dangling of the limbs contralateral to certain unilateral lesions than animals operated on at the earlier ages (Fig. 2). Our first experience with ablation of the posterior medial and the posterior lateral quadrants, sparing the anterior quadrants, on both sides showed that the dysfunction following this partial ablation differed from that following removal of all four quadrants. With removal of the whole motor-sensory cortex, the animal though awkward in initiating locomotion, as noted earlier, tactually placed with both forelimbs, and also visually placed. The animal with the posterior quadrants bilaterally removed, but the anterior ones spared, not only retained tactile placing in all four limbs, but the responses were much exaggerated and the limbs over-shot the mark considerably. It did not place visually, ignoring a ledge on either side until limbs or whiskers touched it. Its locomotion was on a wide base and awkward, and it had great difficulty negotiating the narrow pathways, being almost unable to get across. When the animal was several weeks old, it could visually place, but the character of its locomotion and exaggerated placing remained about the same. Remarks on the growing motor-sensory cortex-corticospinal system. The experiments illustrated how much nervous system development went on after birth, both structurally and functionally, and how potentially plastic some of this postnatal growth was. Major development of the corticospinal system and its functions occurred in the third week, coinciding with a rapid transition from infantile to mature patterns of locomotion. That a developmental defect in infancy might not show itself functionally until some time later was illustrated by the loss of placing reflexes 17 days after

46

SAMUEL P. HICKS AND CONSTANCE J. D'AMATO

motor-sensory cortex ablation at birth. The difficulty of determining "localization of function'' in the brain by ablations, especially in early life, was also illustrated in this latter experiment. The loss of function involved mechanisms, not at all understood presently, in both the growing cortex that had been removed and the growing cortex that remained. The problems of localization of function and predicting what kind of adaptation may follow an injury were further illustrated by the bilateral ablations in the newborn. The disturbances of function that might be attributed to ablation of one part of the cortex depended very much on whether other parts were present or absent.

REFERENCES Barron, D. H. The results of unilateral pyramidal section in the rat. Jour/Ull of Comparative Neurology, 1934, 60, 45-55. Dennenberg, V. H. Experimental programming of life histories in the rat. In A. Ambrose (Ed.), Stimulation in early infancy, New York: Academic Press, 1969. Falk, J. L. & D' Amato, C. J. Automation of pattern discrimination in the rat. Psychological Reports, 1962, 10, 24. Fowler, H., Hicks, S. P., D'Amato, C. J., & Beach, F. A. Effects of fetal irradiation on behavior in the albino rat. Journal of Comparative and Physiological Psychology, 1962, 55, 309-314. Furchgott, E. Behavioral effects of ionizing radiation. Psychological Bulletin, 1963, 60, 157-199. Galambos, R., Norton, T. T., & Fromme!, G. P. Optic tract lesions sparing pattern vision in cats. Experimental Nuerology, 1967, 18, 8-25. Ginsburg, B. E. Developmental behavioral genetics. InN. B. Talbot, J. Kagan, & L. Eisenberg (Eds.), Behavioral science in pediatric medicine. Philadelphia: Saunders, 1971. Hicks, S. P., & D'Amato, C. J. How to design and build abnormal brains using radiation during development. In W. S. Fields & M. M. Desmond (Eds.), Disorders of the developing nervous system. Springfield: Thomas, 1961. Hicks, S. P ., & D' Amato, C. J. Malformation and regeneration of the mammalian retina following experimental radiation. In L. Michaux & M. Field (Eds.), Les phakomatoses cerebrales, deuxieme colloque international, malformations congenitales de 1' encephale. Paris: S .P .E.I., 1963. Hicks, S. P ., & D' Amato, C. J. Effects of ionizing radiations on mammalian development. In D. H. M. Woolman (Ed.), Advances in teratology. London: Logos Press, 1966. Hicks, S.P., & D' Amato, C.J. Cell migrations to the isocortex in the rat.AIUltomical Record, 1968, 160, 619-634. Hicks, S. P., & D'Amato, C. J. Visual and motor function after hemispherectomy in newborn and mature ;ats. Experimental Neurology, 1970, 29, 416-438. Hicks, S. P., & D'Amato, C. J. Visual function of rat retinas malformed by irradiation at birth (Abstract and exhibit), Society for Neuroscience, Proceedings of lst Annual Meeting, Washington, D.C., October, 1971. Hicks, S. P ., & D' Amato, C. J. Effects of ablation of motor-sensory cortex are different in newborn and mature rats. (Abstract) Society for Neuroscience, Proceedings of 3rd Annual Meeting, San Diego, November, 1973. (a) Hicks, S. P., & D'Amato, C. J. Normal and aberrant development of corticospinal tract in rat. (Abstract and exhibit) Michigan Chapter: Society for Neuroscience, Proceedings of the 4th Annual Spring Meeting, University of Michigan, Ann Arbor, May, 1973. (b) Hicks, S. P., D'Amato, C. J., & Falk, J. L. Effects of radiation on structural and behavioral development. International Journal of Neurology, 1962, 3, 535-548.

ADAPTATION AFTER EARLY BRAIN INJURY

47

Hicks, S. P., D'Amato, C. J., French, B. C., Klein, S. J., & Austin, L. L. Effects of regional irradiation or ablation of the infant rat cerebellum on motor development. In M. R. Sikov & D. D. Malum (Eds.),Radiobiology of the fetal am/juvenile mammal: Ninth annual Hanford biology symposium. Oak Ridge, Tenn: USAEC Div. Tech, Inform. Extension, 1969. Hicks; S. P., D'Amato, C. J ., & Lowe, M. J. The development of the mammalian nervous system. I. Malformations of the brain, especially the cerebral cortex, induced in rats by radiation. II. Some mechanisms of the malformations of the cortex.Jouranl of Comparative Neurology, 1959, 113, 435-469. Jacobson, M., & Hunt, R. K. The origins of nerve cell specificity. Scientific American, 1973, 228, 26-35. Joffe, J. M. Prenatal determinants of behavior. Oxford: Pergamon, 1969. Lashley, K. The mechanism of vision. I. A method for rapid analysis of pattern-vision in the rat. Journal of Genetic Psychology, 1930, 37, 453-460. Lashley, K. The mechanism of vision. XVI. The functioning of small remnants of the visual cortex. Journal of Comparative Neurology, 1939, 70, 45-67. Leong, S. K., & Lund, R. D. Anomalous bilateral corticofugal pathways in albino rats after neonatal lesions. Brain Research, 1973, 62, 218-221. Lewellyn, D., Lowes, G., & Isaacson, R. L. Visually mediated behaviors following neocortical destruction in the rat.Journal of Comparative and Physiological Psychology, 1969, 69, 25-52. Liddel, E. G. T., & Phillips, C. G. Pyramidal section in the cat. Brain, 1944, 67, 1-9. Lund, R. D., & Lund, J. S. Development of synaptic patterns in the superior colliculus of the rat. Brain Research 1972, 42, 1-20. Money, J. & Ehrhardt, A. A. Man & woman, boy & girl: The differentiation and dimorphism of gender identity from conception to maturity. Baltimore: Johns Hopkins University Press, 1972. Woolsey, C. N. Patterns of localization in sensory and motor areas of the cerebral cortex. In The biology of mental health and disease; The 27th annual conference of the Milbank memorial fund. New York: Roeber, 1952.

3 SOCIAL RESPONSES TO BLIND INFANT MONKEYS! Gershon Berkson Illinois State Pediatric Institute

In The Descent ofMan (p. 475), Darwin stated that herds of animals sometimes gore or worry to death a disabled member of the group so that his presence will not attract a predator. Social selection for defects has since been more or less assumed to be an obvious aspect of natural selection in animals. However, instances of one animal of a group killing a defective individual who is not already moribund are difficult to find in the literature. More commonly, it is believed (Wynne-Edwards, 1962, p. 547) that density-dependent social mechanisms, such as competition for territory, force disabled individuals into situations where they are more vulnerable to predation or starvation (Errington, 1967). Intraspecific contributions to mortality are therefore generally indirect. In the same passage, Darwin described social compensation for defects. He cited instances in which birds kept a blind member of a flock alive by feeding him. Disabled animals apparently do sometimes survive in natural habitats (Berkson, 1974; Schultz, 1956), and it is likely that they benefit from the same social forces that protect normal individuals. Survival of disabled animals seems likely to the degree that protective social behaviors are characteristic of the group in which the animal lives. More specifically, it seems that animals who live in permanent social groups that have low intrinsic fertility rates and that raise dependent young for long periods are likely to tolerate and compensate for defects in group members (Berkson, 1973a). Most species of higher primates fulfill these criteria, and it appears that they do compensate for defects in infants and tolerate them in older animals. Rumbaugh (1965) and more recently Rosenblum and Youngstein (1974) have shown that monkey mothers compensate for temporary experimentally produced disabilities 1 I am especially indebted to Lilian Tosic, Linda Massen, and Russell Puetz who did the observations and analyses.

49

50

GERSHON BERKSON

in their infants, and Lindburg (1969) and Berkson (1973b) indicate that this is also true of permanent defects. Studies preliminary to the one I am going to describe have shown that infant monkeys with a visual acuity deficit survived until seven months of age in a natural habitat in Thailand. While they were alive, their group adapted its behavior to the infant's deficit (Berkson, 1970). A repetition of this experiment (Berkson, 1973b) in a free-ranging monkey colony where food is freely available and predators absent has thus far shown that partially blind animals have survived until almost three years of age. For practical reasons, it has not been possible to examine intensively social responses to these monkeys. However, I have been able to do a longitudinal study of totally and partially blind individuals in four laboratory groups of monkeys. We have recently completed the first year of observations which show relatively normal social development in the home cage. METHOD

The animals were part of four crab-eating monkey troops (M acaca fascicularis) housed in 4.8-m X 3.9-m x 1.7-m cyclone fence cages, each troop in a separate room. Food and water were freely available. The troops were initially composed of an adult male, four adult females, and two 1.9-2.6-kg male juveniles. Each adult female was pregnant when the troop was formed, and their offspring were the experimental and control groups. Within three weeks after birth, six infants were made partially blind (Berkson & Karrer, 1968). In addition five infants were made totally blind and eleven infants underwent a control operation. The experimental and control groups were randomly distributed to the four troops (Table 1). The troop was observed on three days every two weeks, with eight momentary observations of 37 categories of behavior taken on all animals on each observation day. Percentage of agreement on occurrences averaged 95 for the various categories. In addition to the routine observations which continued throughout the year, each group was exposed to a series of "stress tests" every three months to determine the response of the stressed troops to the experimental and control animals. Detailed quantitative data were collected and log books were kept of both routine and stress observations. Only the data on the routine quantitative observations will be reported here. RESULTS

In our previous study, partially blind animals survived in a free-ranging environment where food was plentiful and there were no predators. No mortality was therefore expected in the first year in this laboratory study, even though the visual deficit was more·severe and the density of the animals very high. This expectation

SOCIAL RESPONSES TO BLIND INFANT MONKEYS

51

TABLE 1 Composition and Characteristics of the Groups on February 15, 1974

No. Group

Sex

Birth

Mother

No. Group

Sex

Room I 01 Adult 02 Adult 03 Adult 04 Adult 05 Adult 06 Juv 07 Juv 08 Partial 09 Control 10 Partial 11 Control 12 Control 13 Control 14 Adult 15 Total

F

04 02 03 05 03 05

01 Adult 02 Adult 03 Adult 04 Adult 05 Adult 06 Juv 07 Juv 08 Control 09 Partial 10 Total ll Partial 12 Total 13 Control 14 Control

M F F F F M M F F M M M F F

M F F F F M M F F M M F M F

02 03 04 05 02 05 03

Room IV

(Died 2/12/74)

1/12/72 1/24/72 5/19/72 2/17/72 10/31/72 6/23/73 7/20/73

9/12/71 10/31/71 8/ 6/72 12/ 2/71 9/18/72 10/17/72 4/ 6/73

02

Room III 01 Adult 02 Adult 03 Adult 04 Adult 05 Adult 06 Juv 07 Juv 08 Control 09 Partial 10 Control 11 Total 12 Control 13 Control 14 Control

Mother

Room II

M F F (Died 4/25/72) F M M M 1/ 3/70 M 1/29/70 4/17/70 M M 7/28/70 M 8/19/71 M 5/10/72 Replaced 04 F M 12/10/72

Birth

04 05 03 02 05 02 04

01 Adult 02 Adult 03 Adult 04 Adult 05 Adult 06 Juv 07 Juv 08 Total 09 Partial 10 Control 11 Control 12 Control

M

F F F F M M M F M F F

6/10/72 8/ 2/72 11/25/72 11/30/72 8/27/73

04 03 05 02 04

was confirmed: No animal died in its first year. We therefore have no evidence of social selection for visual deficit in macaque monkeys in the first year of life. Despite the fact that they were blind, all experimental animals learned to feed themselves by eating the easily available monkey biscuits and drinking at one of the six watering faucets in the cage. Even the totally blind infants knew their way around the cage. Although they were obviously awkward in traveling, they were able to get to the vicinity of the faucet, feel their way to its exact location, and drink normally. Initially we thought that locomotion paths were more stereotyped in

52

GERSHON BERKSON 100

50

__.

Total

~

Partial

o---o

Control

Mother 0 ~~1--~2~~3---4~--s---6~ rn

c

........0

..>...

10

10

Adult Females

Adult Male

1

2

3

4

5

6

Ill

.g IJl

...,

c Ill u

... Ill

--Total

-

'"c

...."'

'0 Ill

::~~ ~~~~~ ~ p

y

0

1

I

IIIJ

2 .l 4 5 6

~8

20 MINUTE PUIOOS FIG.

5.

•

9 10

I

IIIJ

-

~

..

2

~

;

~ ~

;

~ ~

10

20 IIIIIUTE PERIODS

Effect of 1.0-mg/kg d-amphetamine on time spent in contact with rotating anesthetized

adult.

labile emotionality, and so forth. With continuted cerebral maturation, the directing and organizing effects of amphetamine may diminish and the generalized arousal-inducing properties emerge and become dominant. The tendency for amphetamine to direct behavior is probably not restricted to amphetamine and its derivitives. On the basis of tentative pilot data it appears that many other stimulant conditions that produce generalized, non-directed arousal in the adult, including food deprivation and electric shock, increase approach and contact in the neonatal rat. This may also be true of the human infant. Bridger (1962) reports, for example, that immersing the child? s foot in cold water increases nursing. In both rat and man these sequences of development are closely correlated with, and presumably controlled by, central nervous system development. The brains of both develop in a caudal to rostral direction, the brain stem being far more mature at birth than the forebrain. In man, many of the reflexes of infancy which are organized at the brainstem level (e.g., rooting, suckling, Babinski) are known to be actively inhibited by maturation of specific regions of the cortex and to reappear following injury to those areas (Paulson & Gottlieb, 1968). For the rat it is similarly assumed that the reflexive behaviors of infancy which disappear during ontogenesis are inhibited by analagous regions of the brain, although direct experimental evidence on this point is lacking. Given this interrelation between maturation of the central nervous system and the sequential emergence and disappearance of neonatal reflexive behaviors, it is tempting to assume that a comparable process mediates the transition from canalized arousal to generalized arousal. In the rat, forebrain areas such as the cortex and hippocampus are known to exert a high degree of inhibitory influence

AMPHETAMINE AND AROUSAL IN RATS

111

on behavioral excitability in the adult. In particular the hippocampus, a structure that at birth contains a large number of mitotic cells (Altman & Das, 1965), is neurochemically incomplete (Mathews, Nadler, Lynch, and Cotman, 1974 ), and is particularly sparce in synaptic clefts (Crain, Cotman, Taylor, and Lynch, 1973), has been implicated repeatedly in the control of arousal. Lesions of this area typically produce an animal that is chronically hyperactive, as well as hyper-responsive to phasic factors (e.g., estrous, food deprivation) normally resulting in locomotor activity increments (Altman, Bruner, & Bayer, 1973; Douglas, 1967). Frontal cortex, another area slow to mature, is believed to be involved in the suppression of reactive increases in arousal. Ontogenetically, lesions of these structures have no effect on activity prior to the period of decreasing normal activity subsequent to 15 days of age (Moorcroft, 1971). During the subsequent 5 days very rapid growth and development occurs, both anatomically and electrophysiologically, so that by 25 days the hippocampus has essentially adult characteristics (Crain et al., 1973; Myslivecek, 1970). Furthermore, pharmacological agents producing an effect on activity through cholinergic or serotonergic action do not become effective until ages at which hippocampal and frontal lesions produce adult-like effects (Fibiger, Lytle, & Campbell, 1970; Campbell & Mabry, 1973). In this connection there is good evidence that at least some components of the hyperactivity seen in hippocampal animals is a result of the disruption of a serotonergic projection from the raphe nucleus to the hippocampus (Trimbach, 1972). The limbic system, and particularly the hippocampal complex, has been implicated in the control of emotional arousal. Furthermore, in humans it seems as well to be involved in attentional and short-term memory processes. In view of this and its extensive postnatal development as demonstrated by both anatomical and lesion techniques, the limbic system seems a good candidate for the neurological substrate of hyperkinesis. Late maturing structures, such as the hippocampus, may be particularly susceptible to neurotoxins of many sorts, for example, lead, which may play a role in the etiology of some forms of hyperkinesia. It is well-known that stresses of all sorts have a greater effect on rapidly developing organs than on mature structures (see Dobbing, 1970). Lastly, limbic system involvement seems particularly appealing since a general intellectual deficit is not a critical characteristic of the syndrome. If there were a general neurodevelopmentallag in the CNS, including cortex, one would also expect slower intellectual development. Instead, the hyperkinetic child seems plagued by impulsiveness, hyperactivity, anq short attention span with minimal impairment of intellectual function. Two recent studies in infrahuman primates tend to further support this analysis of juvenile behavior and its dependence on neurological development. First, infant chimpanzies have been reported to cease exploratory or play behavior and to cling to the mother following amphetamine administration (Mason, 1971). Second, lesions of the cortex are said to severely delay the normal transition from neonatal to juvenile behavior in the rhesus monkey. In particular, clinging and contact with the mother was prolonged in the lesioned animals.

112

BYRON A. CAMPBELL AND PATRICK K. RANDALL

These two observations offer further support for the view that amphetamine does indeed produce a different if not "paradoxical" effect on behavior in the neonatal animal and that the disappearance of this paradoxical effect is dependent upon further maturation of the central nervous system.

REFERENCES Altman, J., Bruner, R. L., & Bayer, S. A. The hippocampus and behavioral maturation. Behavioral Biology, 1973, 8, 557-596. Altman, J., & Das, G. D. Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. Journal of Comparative Neurology, 1965, 124, 319-336. Bridger, W. H. Ethological concepts and human development. Recent Advances in Biological Psychiatry, 1962, 4, 95-107. Campbell, B. A., Lytle, L. D., & Fibiger, H. C. Ontogeny of arousal and cholinergic inhibitory mechanisms in the rat. Science, 1969, 166, 637-638. Campbell, B. A., & Mabry, P. D. The role of catecholamines in behavioral arousal during ontogenesis. Psychopharmacologia, 1973, 31, 253-264. Crain, B., Cotman, C., Taylor, D., & Lynch, G. A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat. Brain Research, 1973, 63, 195-204. Dobbing, J. Undernutrition and the developing brain. In W. A Himwich (Ed.), Developmental neurobiology. Springfield, Ill.: Thomas, 1970. Pp. 241-261. Douglas, R. J. The hippocampus and behavior. Psychological Bulletin, 1967, 67, (6), 416-442. Fibiger, H. C., Lytle, L. D., & Campbell, B. A. Cholinergic modulation of adrenergic arousal in the developing rat. Journal of Comparative and Physiological Psychology, 1970, 72, 384--389. Gesell, A., & Ilg, F. Infant and child in the culture of today. New York: Harper and Row, 1943. La!, S., & Sourkes, T. L. Ontogeny of stereotyped behavior induced by apomorphine and amphetamine in the rat. Archives Internationales de Pharmacodynamie et de Therapie, 1973, 202, 171-182. Mason, N. A. Motivational factors in psychosocial development. In W. J. Arnold & M. M. Page (Eds.), Nebraska symposium on motivation. Lincoln: University of Nebraska Press, 1971. Mathews, D. A., Nadler, J. V., Lynch, G. S., & Cotman, C. W. Development of cholinergic innervation in the hippocampal formation of the rat. I. Histochemical demonstration of acetylcholinesterase activity. Developmental Biology, 1974, 96, 130--141. Moorcroft, W. H. Ontogeny of behavioral inhibition by forebrain structures in the rat. Brain Research, 1971, 35, 513-525. Moorcroft, W. H., Lytle, L. D., & Campbell, B. A. Ontogeny of starvation-induced behavioral arousal in the rat. Journal of Comparative and Physiological Psychology, 1971, 75, 59-67. Myslivecek, J. Electrophysiology of the developing brain-Central and Eastern European contributions. In W. A. Himwich (Ed.), Developmental neurobiology. Springfield, Ill.: Thomas, 1970. Paulson, G., & Gottlieb, G. Development reflexes: The reappearance of foetal and neonatal reflexes in aged patients. Brain, 1965, 91, 37. Routh, D. K., Schroeder, C. S., & O'Tauma, L. A. Development of activity level in children. Developmental Psychology, 1974, 10 (2), 163-168. Trimbach, C. Hippocampal modulation of behavioral arousal: Mediation by serotonin. Unpublished doctoral dissertation, Princeton University, 1972. Wender, P. H. Minimal brain dysfunction in children. New York: Wiley, 1971.

8 THE CONCEPT OF A CUMULATIVE RISK SCORE FOR INFANTS Arthur H. Parmelee, M.D. Marian Sigman, Ph.D. Claire B. Kopp, Ph.D. Audrey Haber, Ph.D. University of California at Los Angeles

INTRODUCTION For this discussion, the term risk is used to imply an increased probability of handicap in childhood. At present, we generally identify infants at biological risk for later sensory, motor, or mental handicaps on the basis of pregnancy, perinatal, and postnatal factors related to infant mortality. Justification for this procedure derives from the concept that a continuum of casualty exists which has both lethal and sublethal manifestations. The lethal components consist of abortions, still births, and neonatal deaths, while the sublethal manifestations include sensory, motor, and mental disabilities (Lilienfeld & Parkhurst, 1951; Parmelee & Haber, 1973; Sameroff & Chandler, in press). This concept is helpful in identifying potentially important variables, but it does not aid us in determining the predictive power of these factors. Such information is not available from present studies. Correlations between single perinatal or postnatal events and later disabling sequelae have been very low in several large prospective studies (Buck, Gregg, Stavraky, et al. 1969; Niswander, Friedman, Hoover, eta!., 1966; Parmelee & Haber, 1973; Sameroff & Chandler, in press). Similarly, the English risk registers, which attempted to classify infants with items selected on the basis of clinical impressions, have failed because too many unimportant isolated events were included (Rogers, 1968). Studies that have focused on more comprehensive ''risk'' events such as prematurity and neonatal asphyxia or anoxia have demonstrated greater incidence of disabling sequelae among infants who have suffered such trauma than among control infants. However, even these results have varied between studies because of the heterogeneity of the risk groups studied. In all follow-up studies of risk 113

114

A. H. PARMELEE, M. SIGMAN,

C. B. KOPP, A. HABER

factors, a broad spectrum of outcomes has been obtained rather than a bimodal distribution of normal and abnormal outcomes between groups. While this is consistent with the concept of a continuum of casualty, such results do not aid in the identification of the strength of relevant variables (Braine, Heimer, Wortis, & Freedman, 1966; Douglas, 1960; Drage & Berendes, 1966; Drage, Berendes, & Fisher, 1967; Drillien, 1964; Graham, Ernhard, Thurston, & Craft, 1962; Heimer, Cutler, & Freedman, 1964; Keith & Gage, 1960; Lubchenco, Delivoria-Popadopoulos, & Searles, 1972; Parmelee & Haber, 1973; Sameroff & Chandler, in press; Schacter & Apgar, 1959; Wiener, Rider, Oppel, & Harper, 1968). One important recurring observation is that outcome measures are strongly influenced by the socio-economic circumstances of the children's environments, and this influence is often stronger than that of earlier biological events. However, there is also evidence that early biological problems lead children to be more vulnerable to adverse environments. Since health problems during pregnancy and early infancy are related to socio-economic status, the two variables must be considered inextricably interwoven (Braine et al., 1966; Douglas, 1960; Drage et al. 1969; Drillien, 1964; Heimer et al., 1964; Knobloch & Pasamanick, 1960; Parmelee & Haber, 1973; Sameroff & Chandler, in press; Werner, Simonian, Bierman, & French, 1968; Weiner et al., 1968). Thus, with our present information, we can discuss which groups of infants are at risk of later disabilities on the basis of socio-economic and/or biological indicators, but we cannot specify the degree of risk or identify the individual infant who will suffer a disability in childhood. A CUMULATIVE RISK CONCEPT The majority of the infants in any ''risk'' group so far identified do sufficiently well on all outcome measures later in childhood that they can not be considered truely handicapped. As a result, the manpower required for intervention programs with the truely handicapped infants is critically diluted by our inability to precisely identify these children. We would like, therefore, to devise more accurate predictive measures of the degree of risk for individual infants. Several studies have demonstrated that multiple factors may be considered as additive in determining degree of risk. Some have cumulated pregnancy, perinatal, and neonatal events and others have included socioeconomic factors (Braine et al. 1966; Drage & Berendes, 1966; Drage et al., 1969; Drillien, 1964; Heimer et al., 1964; Werner et al., 1968; Wieneret al., 1968). A recent study demonstrated high prediction of behavioral achievement at 7 years of age using a cumulative score of biological factors during pregnancy, birth events, socio-economic factors, and performance items during the first year of life (Smith, Flick, Ferriss, & Sellman, 1972). On this basis, we decided to design a method of identifying infants at risk that used multiple assessments at different ages and measured a wide range of variables

A CUMULATIVE RISK SCORE FOR INFANTS

115

(Parmelee, Sigman, Kopp, & Haber, 1974). In devising our risk scoring system we considered the following clinical observations and deductions: 1. Many perinatal problems cause only transient insult, rather than permanent brain injury. Thus, in the newborn period, babies may appear equally ill upon examination but some will recover completely. 2. Some pregnancy and perinatal problems cause brain injury that is not manifest in obvious ways in the neonatal period but the deviance becomes more evident as complex behaviors unfold dufing infancy. 3. Some parents appear intuitively able to provide an optimal environment for an infant with mild neurological deviances, allowing him to compensate. With these points in mind a useful risk scoring system might be one that (a) scores pregnancy, perinatal, and neonatal biological events and behavioral performances in an additive fashion; (b) reassesses the infant in the first months of life to sort out those infants with transient brain insult from those with brain injury who remain deviant; (c) reassesses the infant again primarily on a behavioral basis later in the first year of life, providing time for environments to have an effect on developmental progress.

Cumulative Risk Score Design Our cumulative risk score system is composed of five items in the neonatal period, four at three and four months, and five at eight to nine months, as follows: Neonatal risk score items: 1. obstetric complications 2. postnatal events 3. newborn neurological examination 4. visual attention 5. sleep polygraph Three- and four-month risk score items: 1. pediatric events and examination 2. Gesell test 3. visual attention 4. sleep polygraph Eight- and nine-months risk score items: 1. pediatric events and examination 2. Gesell test 3. cognitive test 4. hand precision/sensory-motor schemes 5. exploratory behavior

116

A. H. PARMELEE, M. SIGMAN, C. B. KOPP, A. HABER

Pilot studies were conducted on these measures to determine the range and distribution of scores. A range of performance scores from normal to abnormal was established for each test, and the raw scores were converted to standardized scores with means of 100 and standard deviations of 20. In this way, the scores could be treated as equivalent and all tests summed and averaged to obtain a cumulative risk score at nine months. We arbitrarily determined that infants having an average cumulative risk score of 96 or less at nine months would be designated as high-risk, and those with scores greater than 96 as low risk. The significance of this cumulative risk score system can only be established through validation studies which relate scores at nine months to later performance. In other words, the cumulative risk score is an independent variable which must be compared with subsequent dependent measures. The dependent variables in the present study will consist of assessments made at two years of age. These outcome measures will include Bayley and Gesell developmental examinations, a cognitive test, measures of expressive and receptive language, and an assessment of exploratory behavior. In addition, we hope to continue collecting information on outcome at older ages so the predictive value of the cumulative risk score can be established. The research flow plan is illustrated in Fig. 1. Subjects

In order to test the validity at two years of age of our risk score system, a sample of infants is being followed from birth. The longitudinal sample consists of premature infants, of 37 weeks gestation or less with birth weights at 2500 gm or less, and a control group of full term infants, of 39-41 weeks gestation and a birth weight greater than 2500 gm. All infants are tested at equivalent conceptional ages, defined as gestational age plus age from birth. Gestational age is calculated from the onset of the mother's last menses. Thus, the newborn tests are administered at 40 weeks conceptional age, which is the expected date of delivery for the premature infants and the gestational age of the full-term infants. Date of testing for the later measures is calculated from the expected date of delivery for the prematures rather than the actual date of birth (Parmelee & Schulte, 1970). Subjects will include infants from all socio-economic groups and attempts will be made to equalize the representation of different socio-economic groups within the risk categories. At present in our study socio-economic status signifies the level of education completed by the infant's mother. Several previous studies have demonstrated a correlation between the level of mother's education and the intellectual development of the child (Drage et al., 1969; Werner et al., 1968). Data is also being collected on other aspects of family background so alternative systems of classifying socioeconomic status can be used. Every infant and family who participates in the longitudinal study receive medical, nursing, and social work help in an effort to provide support services

A CUMULATIVE RISK SCORE FOR INFANTS

117

RESEARCH FLOW CHART

DIAGNOSTIC STUDIES

Cumulative Risk Scare 9 Months

Term Date

Interim Evaluative Measures 18 Months

Outcome Measures 2 Years

/

LOW-RISK I·J$.JFANTS Infants Enter >-----1-----~------------~ 1 No Educational Intervention Project

\

HIGH-RISK INFANTS No Educational Intervention

HIGH-RISK INFANTS Educational Intervention 10 Months INTERVENTION STUDIES FIG.

I . Research flow chart.

regardless of risk category. The family is referred to special community resources whenever these are needed for the infant. A subgroup of "high-risk" infants is participating in a specialized educational intervention program from 10 to 24 months to test the value of such treatment (Kass, Sigman, Bromwich, & Parmelee, 1974). Preliminary Results

These data are preliminary since we have tested about half the 200 cases to be studied before the final analyses. As of March 1 , 1974, a total of 76 premature infants were assigned term risk scores. Of these, 65 babies had completed all tests through four months and thus, had four month risk scores as well. Only 39 babies had nine month risk scores. The means and standard deviations of the three separate risk scores for the 39 babies who have completed all tests throughout the first nine months are reported in Table 1.

118

A. H. PARMELEE, M. SIGMAN, C. B. KOPP, A. HABER

TABLE 1 Mean Risk Scores at Term, Four Months, and Nine Months of Infants Completing All Measures

Risk scores

N

Mean

Term Four months Nine months

39 39 39

96.15 100.13 99.76

SD

10.91 9.46 9.11

%Infants classified as high risk 56 33

28

The results of an analyses of variance indicate that the Risk Scores are significantly lower (indication of higher risk) at tenn than at four and nine months [ F(2, 76) = 10.01,p < .01 In addition, among these 39 infants, more babies scored 96 or below at tenn and four months than at nine months. Thus, on the basis of tenn risk score, 22 infants would have been classified as high risk, and on the basis of four month risk score, 13 would have been classified as high risk. In comparison, 11 infants have been classified as high risk on the basis of the 9-months risk score. In order to get a preliminary idea of which individual diagnostic measures most closely mirrored the aggregate risk scores, the correlation coeficients of each diagnostic measure with the risk scores at tenn, 4 months, and 9 months were examined (Table 2). Most of the early measures are significantly related to risk status at term and four months. Correlations may be somewhat inflated since the score on each measure is included in the total risk score. However, even at 9 months, where any one measure is a very small part of the cumulative risk score because of the greater number of diagnostic measures included, there are significant relationships between many measures and risk score.

J.

DISCUSSION

The data presented in this paper are truly preliminary in that many of the correlations will probably change when all the data are analyzed. Furthermore, the really critical questions can not be addressed until information from the outcome measures is available. While it is interesting to determine the interrelationship between the measures comprising the high risk scale, the more significant problem concerns the relationship between the various measures in the high-risk score and outcome variables. The cumulative risk system was designed to feature the use of multiple measures. In fact, we expect that the most valid predictions will be made using clusters of these measures. The strength of this approach is that it will make possible the identification of the contributions in predictive outcome made by the various measures independently and in combination. With this information, it may be

A CUMULATIVE RISK SCORE

FOR INFANTS

119

possible to design a more effective system either by eliminating certain measures or utilizing a weighting system. In addition, the strength of the various components of each measure can be evaluated in relation to risk score and later performance so the individual measures can be improved. As a separate part of our project, but not included in our present risk score, mother-infant interaction is observed in the home at one and eight months (Beckwith, 1974). A retrospective analysis will attempt to determine those qualities of mother-infant interaction which had an ameliorating effect on infants who were at risk early in infancy and improved by nine months or two years. In TABLE 2 Correlations between Individual Diagnostic Measures and Cumulative Risk Scores Cumulative risk scores Diagno~tic

measures

Obstetrical complications scale **N Newborn neurological Term visual attention Term sleep polygraph Postnatal factors 3-month sleep polygraph 4-month visual attention 4-month Gesell 4-month pediatric examination 8-month precision and sensory-motor schemas 8-month exploratory behavior 9-month cognitive (Piaget) 9-month Gesell 9,month pediatric examination

=

Term

4 Months

9 Months

0.60* 76 0.33* 76 0.471* 74 0.32* 76 0.65* 76

0.55* 65 0.32* 65 0.23 63 0.37* 65 0.57* 65 0.22 65 0.48* 61 0.50* 65 0.53* 65

0.62* 39 0.01 39 0.39* 37 0.37* 39 0.46* 39 0.39* 39 0.43* 38 0.50* 39 0.38* 39 0.49* 38 0.49* 38 0.54* 34 0.43* 38 0.62* 39

*p < 0.05. **N = Number of subjects in each correlational analysis.

120

A. H. PARMELEE, M. SIGMAN, C. B. KOPP, A. HABER

addition, we will attempt to isolate those factors in mother-infant interaction which correlated with a decline in the infant's development. This information may enable us to identify environmental factors that should be included in the risk score. It will also be helpful in the selection of intervention procedures. The subjects in the longitudinal sample include infants from all socioeconomic groups, so the relationship between socioeconomic status, biological risk, and mother-infant interaction can be examined. Our current hypothesis is that even though mother-infant interaction is related to social class, mother-infant interaction will have powerful effects on infant development irrespective of social class. Analyses will be done which will relate mother-child interaction and social class to each other, biological risk factors, and later development. The need for follow-up studies is essential to any research of this nature. However, the general failure of long-term studies to demonstrate a direct relation between early medical events and later outcome suggests that another strategy is needed. Long-term studies fail to take into account the ongoing changes that affect outcome measures. In the past, most studies have attempted to predict development in early childhood from the neonatal period. An alternative strategy is that of repeated predictions over short periods of time. Such a strategy allows for the probable occurence of changes resulting from transactional processes between environment and individual and provides the opportunity to identify the nature of these processes. Our cumulative risk score was designed with this in mind, and our follow-up will continue this strategy. ACKNOWLEDGMENTS This research was supported by NIH Contract N01-HD-3-2776 "Diagnostic and Intervention Studies of High-Risk Infants"; and NICHD Grant No. HD-04612, Mental Retardation Research Center, UCLA. Computing assistance was obtained from the Health Sciences Computing Facility, University of California at Los Angeles, sponsored by National Institutes of Health Special Research Resources Grant RR-3.

REFERENCES Beckwith, L. Care giver-infant interaction and the development of the risk infant. Conference report. Early Intervention for High Risk Infants and Young Children. Chapel Hill, N.C., 1974. Braine, M. D. S., Heimer, C. B., Wortis, H., & Freedman, A. M. Factors associated with impairment of the early development of prematures. Monographs of the Society for Research in Child Development, 1966, 31 (Whole No. 106). Buck, C., Gregg, R., Stavraky, K., et al. The effect of single prenatal and natal complications upon the development of children of mature birthweight. Pediatrics, 1969, 43, 942-955. Douglas, J. W. B. "Premature" children at primary schools. British Medical Journal, 1960, 1, 1008-1013. Drage, J. S ., & Berendes, H. W. Apgar scores and outcome of the newborn. Pediatric Clinics ofNorth America, 1966, 13, 635-643.

A CUMULATIVE RISK SCORE FOR INFANTS

121

Drage, J. S., Berendes, H. W., & Fisher, P. D. The Apgar scores and four-year psychological examination performance. In Perinatal factors affecting human development. Pan American Health Organization Scientific Publication No. 185, 1969. Drillien, C. M. The growth and development of the prematurely born infant. Baltimore: Williams and Wilkins, 1964. Graham, F. K., Ernhard, C. B., Thurston, D., & Craft, M. Development three years after perinatal anoxia and other potentially damaging newborn experiences. Psychological Monographs, 1962, 76 (Whole No. 522). Heimer, C. B., Cutler, R., & Freedman, A.M. Neurological sequelae of premature birth.American Journal of Diseases of Children, 1964, 108, 122-133. Kass, E. R., Sigman, M., Bromwich, R., & Parmelee, A. H. Educational intervention with high risk infants. Conference report. Early Intervention for High Risk Infants and Young Children. Chapel Hill, N. C., 1974. Keith, H. M., & Gage, R. P. Neurologic lesions in relation to asphyxia of the newborn and factors of pregnancy: Long-term follow-up. Pediatrics, 1960, 26, 616-622. Knobloch, H., & Pasamanick, B. Environmental factors affecting human development before and after birth. Pediatrics, 1960, 26, 210-218. Lilienfeld, A.M., & Parkhurst, E. A study of the association of factors of pregnancy and parturition with the development of cerebral palsy. A preliminary report. American Journal of Hygiene, 1951, 53, 262-282. Lubchenco, L. 0., Delivoria-Papadopoulos, M., & Searls, D. Long-term follow-up studies of prematurely born infants. II Influence ofbirthweight and gestational age on sequelae.Journal of Pediatrics, 1972, 80, 509-512. Niswander, K. R., Friedman, E. A., Hoover, D. B., et al. Fetal morbidity following potentially anoxigenic obstetric conditions. I. Abruptio placentae. II Placenta previa. III Prolapse of the umbilical cord. American Journal of Obstetrics and Gynecology, 1966, 95, 838-845. Parmelee, A. H., & Schulte, F. J. Developmental testing of preterm and small-for-dates infants. Pediatrics, 1970, 45, 21-28. Parmelee, A. H., & Haber, A. Who is the "Risk Infant"?ClinicalObstetricsand Gynecology, 1973, 16, 376-387. Parmelee, A. H., Sigman, M., Kopp, C. B., & Haber, A. Diagnosis of the infant at high risk for mental, motor or sensory handicap. Conference report. Early Intervention for High Risk Infants and Young Children. Chapel Hill, N.C., 1974. Rogers, M. G. H., 1968. Risk Registers and early detection of handicaps. Developmental Medicine & Child Neurology, 10, 651-661. Sameroff, A. J ., & Chandler, M. J. Reproductive risk and the continuum of caretaking casualty. In F. D. Horowitz, M. Hetherington, S. Scarr-Salapetek, & G. Siegel (Eds.), Review of child development research. Chicago, Ill.: University of Chicago Press, in press. Schachter, F. F., & Apgar V. Perinatal asphyxia and psychologic signs of brain damage in childhood. Pediatrics, 1959, 24, 1016-1025. Smith, A. C., Flick, G. L., Ferriss, G. S., & Sellman, A. H. Prediction of developmental outcome at seven years from prenatal, perinatal and postnatal events. Child Development, 1972, 43, 495-498. Werner, E., Simonian, K., Bierman, J. M., & French, F. E. Cumulative effect of perinatal complications and deprived environment on physical, intellectual, and social development of preschool children. Pediatrics, 1968, 39, 490-505. Wiener, G., Rider, R. V., Oppel, W. C., & Harper, P. A. Correlatesoflow birthweight. Psychological status at 8 to 10 years of age. Pediatric Research, 196&, 2, 110-118.

9 EARLY DEVELOPMENT OF SLEEPING BEHAVIORS IN INFANTS Evelyn B. Thoman University of Connecticut

An infant's behavioral state is his most continuous characteristic. He expresses his state by sleeping, wakefulness, or even by giving mixed signals of sleeping and wakefulness simultaneously. Along with the behaviors that indicate state, an infant may display additional spontaneo·us behaviors such as startles, jerks, rhythmic mouthing, sucking, smiles, frowns, etc. as well as crying which primarily occurs when the infant is awake. The purposes of our research are: (a) to describe the characteristics of behavioral states of newborn infants, (b) to identify changes that occur in state organization throughout the first weeks of life, and (c) to identify measures of behavioral states that characterize individual infants during these early weeks. Researchers have been interested in infants' states for as long as they have been interested in studying infants. Crying or sleeping are obvious behaviors to record, and these general state categories are very useful in describing infants. Definition of "state'' as a concept and as a research variable is still evolving, and a variety of sleeping and waking states are identified by various researchers (Ashton, 1973; Komer, 1972). Interest in state behaviors has converged from two primary directions: (1) studies of the function and physiology of sleep, and (2) behavioral studies in which the infant's state is expected to account for variability in behavior. Sleep researchers have produced an enormous literature on neural, autonomic, and endocrine changes associated with sleep states. Aserinsky and Kleitman (1955) and others have demonstrated the cyclical nature of sleep states. The ontogeny of sleep states has been studied by a number of investigators (e.g., Dittrichova, 1969; Dreyfus-Brisac, 1967; Parmalee, 1961; Parmalee & Stem, 1972; Roffwarg, Muzio, & Dement, 1966). From these studies two major categories of sleep have been identified: (1) an active sleep state, generally characterized by irregular respiration, low muscle tone, and the occurrence of 123

124

EVELYN B. THOMAN

rapid eye movements (REMs); and (2) a quiet sleep state, characterized by a lack of motor movement, absence of eye movements, and regular respiration. Depending on the investigator, different labels have been applied with very little difference in referents. This situation was somewhat remedied by a conference (Anders, Emde, & Parmalee, 1971), at which nomenclature and criteria, behavioral and physiological, were agreed on for scoring of states of sleep and wakefulness in newborn infants. This was a major step in operationally defining the state concept, using a combination of behavioral and physiological parameters. The guiding principle for physiological definitions of states is that they are defined by the concordance of measures of physiological parameters observed, including EEG, EOG, EMG, heart rate, and respiration (Prechtl, Weinmann, & Akiyama, 1969). Researchers interested in organization of the infant's behavior and behavioral development have applied the same notion, that is, concordance of behaviors, to define states. The classic example, and the most extensive system for defining behavioral states, was provided by Wolff (1966). For researchers who have a primary interest in behavior, physiological measures are not only irrevelant, but obtaining such measures by means of electrodes and other attachments are considered an intrusion and a possible interferences with the infant's most naturalistic behaviors. This is the position we maintain, as we are interested in the developing behaviors of the infant, how these behaviors may constitute stimuli for the mother, and modifications of behavior as a function of the infant's interaction with the mother. If the concept of behavioral state is to be a useful one in research, a great deal more work is yet to be done in determining the most appropriate clusters of behaviors for defining state categories. This point was emphasized by Ashton ( 1973) in a recent review of the state literature. Our research is aimed at clarifying some of the issues with respect to the state concept. Justification for the definitions we use comes from our findings that infants show individual differences over the first weeks of life on the state measures employed for our systematic observations. Even more, our data are beginning to provide a basis for predicting some aspects of both behavioral and physiological development in infants. Consistency and predictive power are the basic requirements of any meaningful descriptive system. The following sections will describe both the consistencies and the consequences of behavioral states as we have observed them. METHODS

Subjects The subjects for these studies were selected as full-term, normal infants, without prenatal, perinatal, or postnatal complications. They include males and females, with mothers of varying parity. All are Caucasian. Our studies have been made at three hospitals located in different parts of the country. Although all

SLEEPING BEHAVIORS IN INFANTS

125

infants were initially "normal" by medical standards, not all of these infants have remained in this category throughout their first months oflife. One objective of our research is to identify the neonates that may have developmental difficulties which are not detected by the usual examinations given newborn infants. Procedure

The categories of behavioral states and the criteria we use for their definition are as follows:

Quiet sleep A. The infant's eyes are firmly closed and still. There is little orno motor activity, with the exception of occasional startles or rhythmic mouthing. Respiration is abdominal and relatively slow (average around 36 per minute), deep and regular. Quiet sleep B. All of the characteristics of Quiet Sleep A apply for this category except for respiration, which deviates somewhat from the slow regularity seen inA. In this state, the respiration check may be relatively fast, above 46, and show some irregularities, or the respiration may be slower but show some irregularities. Respiration is primarily abdominal in this state. Active sleep without REM. The infant's eyes are closed, but slow, rolling movements may be apparent. Bodily activity can range from minor twitches to writhing and stretching. Respiration is irregular, costal in nature, and generally faster than that seen in quiet sleep (average of 46 per minute). Facial movements may include frowns, grimaces, smiles, twitches, mouth movements, and sucking (actually face movements are not very often seen in this category of active sleep). Active sleep with REM. The infant's eyes are closed, and REM occur during the 10-sec epoch; other respiration and movement characteristics are the same as those just described for active sleep without REMs, except that the facial activity is highly likely to accompany REMs or to be interspersed between groups of REMs. Active sleep with dense REM. Characteristics of this category are the same as those of the other two categories of active sleep; however, this one is distinctive for the continuous occurrence of REMs throughout the 10-sec observation epoch. REMs in this category are often accompanied by raising of the eyebrows and by eyeopening. Drowsy state. The infant's eyes may either open and close or they may be partially or fully open, but very still and dazed in appearance. There may be some generalized motor activity, and respiration is fairly regular, but faster and more shallow than that observed in Quiet sleep. Alert inactivity. The infant's body and face are relatively quiet and inactive, and the eyes are "bright and shining" in appearance (Wolff, 1966). Waking activity. The infant's eyes are generally open, but may be closed. There is generalized motor activity, accompanied by grimacing, grunting, or brief vocalization.

126

EVELYN B. THOMAN

Fussing. The characteristics of this state are the same as those for waking activity but mild, agitated vocalization is continuous; or one cry burst may occur. Crying. The characteristics of this state are the same as waking activity, but generalized motor activity is more intense, and cry bursts are continuous. Indefinite state. The infant's eyes may be closed, or opening and closing. There is generalized motor activity, but there are no sufficient criteria by which the infant's state can be classified as waking or sleeping. There is much overlap between the behavioral state categories as we define them and the behaviors included in the physiological states as defined by Anders et al. (1971) for the newborn infant. Primarily, our changes consist of subdividing the sleep categories. Justification for these separate categories is derived from our studies which have shown individual consistency in the newborn period and individual differences over the early weeks of life in these behavioral states (Thoman, in press). Respiration is recorded by means of a small, portable analog chart recorder which derives a signal from a sensor placed under the infant's mattress pad. In addition to the behavioral states listed above, we also record startles, jerks, rhythmic mouthing, sucking, smiles, frowns, grimaces, small body movements, large body movements, grunts, sneezing, hiccups, the occurrence of a bowel movement, gagging or spitting up, REM, eyes open during sleep states, vocalization during sleep states, and noncrying sounds made by the baby. Data for the studies to be reported have been obtained from observations of newborn infants in hospitals and also from observations that have been made in the infant's homes. Home observations consist of a 7- hr day that is spent observing the mother and her infant once a week during the first four weeks after birth. These observations are part of a major project in which we are interested in the ongoing state behaviors of the infant, both when he is alone and in the crib and when the mother and infant are interacting. When the mother is holding the infant or is nearby, we record both mother and infant behaviors. Whenever the infant is out of the crib, sleep is recorded as "eyes closed," as it is impossible to note specific sleep states reliably under these circumstances. When the infant is in the crib, all of the sleeping and waking states are recorded along with respiration. Thus, on each day of observation, we obtain a total of7 hrofthe infant's state of which about 2.5 breach time is sleep observed when the infant is in the crib. The sleep data from the crib observations will be the basis for the present report. When the infant is observed in the hospital, he is taken to a separate research room, placed in a warmer, wearing only a diaper, and is maintained at neutral temperature. Under these conditions, the infant is observed until two hours of sleep has occurred or until time for the next feeding. States and state-related behaviors are recorded every 10 sec. This short time period is used in all of our research because (a) we have found that state durations in highly volatile infants may be extremely short, and (b) we have found this form of variability in behavior

--?? .._.....,.....4

60

Crying Waking Active Alert I noctive Drowsy D-Rem

Rem's

40

Active Sleep

en w .....

No Rem's

:::::>

z

::E

20 Quiet Sleep

..!!.

Quiet Sleep

A

1. Mena number of minutes per hour spent in each state. FIG.

"• OF TOTAL TIME

.82

QUIET

s. A

.69

.... 55

>

QUIET S. .§...

.05

INDEFINITE DROWSY

.27J 13

CRY FIG.

.83

,

.......--=--.37

.30J

W. ACTIVE

·~ ~ .65

AL. INACTIVE

2. Transition probabilities among eight states.

127

128

EVELYN B. THOMAN

to be a significant measure of individual differences among infants. These findings will be described. When the infants are observed in the home, they are placed in the crib dressed, wrapped and positioned as the mother chooses, as we do not intervene in any way with the interaction between the mother and the infant. Results of Studies of Sleeping and Waking Behaviors in Neonates Figure 1 presents the data on the study of 41 babies and indicates the distribution of time that infants spend in each sleep or wake state. About 40% of sleep time is spent in the two categories of Quiet sleep. Relatively small periods of time are spent in the waking states. The sequence of states that an infant may have is not a random affair. Figure 2 presents a flow diagram, showing the highest probabilities of transitions among states. This diagram was derived from a sequence analysis of the data from the 14 newborn infants referred to above. Each subject was observed for 2 hr. The percent figures in the diagram indicate the probability of a change in state in the direction of the arrow. The highest probabilities are among the sleep states and among the wake states. Given that the baby makes a transition from sleep to wakefulness or vice versa, such changes are made primarily via either the Indefinite State or the Drowsy State. However, this transition is unlikely, and once the infant is asleep or awake, he is likely to stay in that category of states. As indicated by this diagram as well as the previous figure, 82% of the total observation time was spent in some form of sleep. Figure 2 depicts the transition probabilities among states for full-term normal infants. An infant that generally deviates from such a pattern may be considered as nonnormal in some respects. The significance of such deviations are, of course, not yet known. In describing Fig. 1 we noted that approximately 40% of sleep time was spent in Quiet Sleep. Variability in infant populations in amount of Quiet Sleep is indicated in Fig. 3. Three groups of infants obtained in three different hospital showed varying amounts of quiet sleep. These differences may reflect a variety of factors. Infants from Hospital I were most stringently screened for weight, amount of maternal medication given during labor, length of labor, and other prenatal and perinatal factors. In addition, the infants at this hospital are born to parents from a higher socioeconomic level than those in the other two groups. Since Quiet sleep may require more homeostatic control than Active sleep, and since it is a state which increases relative to active sleep with maturation, these data suggest that the advantages associated with the population of infants born at the hospital I may have influenced the state organization of these newborn infants. Much more study would be required to identify the specific factors that account for these differences as well as for the actual implications of such differences. At this time the data are simply descriptive. Since babies are highly variable, a major question that arises is whether any

SLEEPING BEHAVIORS IN INFANTS

129

100

80

60 ~

z

LlJ (.)

lr LlJ

Q.

40

20

I

n

m

HOSPITAL FIG.

3.

Amount of Quiet sleep as a percent of total slee_p, subjects from three hospital populations.

particular observation is typical for a baby or simply a random sample from the total variable repertoire of his behaviors. Table 1 presents data on this issue. Twoand three-day-old infants were observed in the hospital mid-feeding in the morning and mid-feeding in the afternoon for one hour each time. Onlyinfants that slept both hours are included in the data reported here. Correlations for the several sleep states that we have defined are indicated on this table. It is apparent that infants are highly consistent with respect to the amount of Quiet and Active Sleep they have at this age. There is also significant consistency in each of the three categories of Active Sleep. However, it should be noted that there is no consistency whatsoever (r = .13) in the amount of Active Sleep taken as a single state. This is an important finding, because it points up the value of subdividing Active Sleep into these three specific components. A state category can be of little use if it occurs without any consistency. The appropriateness of our subcategories of both Quiet sleep and Active sleep are also demonstrated in Table 2, which shows the rate of occurrence (per hour) of the various state-related behaviors recorded for each of the various sleep states. Differences in these rates among the several sleep categories were found to be significant. It is clear from the data presented that rates during quiet and active sleep are very different. They also differ for each of the three active sleep

130

EVELYN B. THOMAN

TABLE 1 Mean Time (minutes) Spent in Each State During a.m. and p.m. Observations, Intraclass Correlations for a.m. and p.m. Durations (N = 24 Babies)

Quiet sleep A Quiet sleep B Active sleep without REM REM dense REM

a.m.

p.m.

r

15.8 8.7 35.2 19.5 14.4

15.3 6.1 38.4 20.6 15.8 2.0

.72 .71 .13 .63 .48 .87

1.3

TABLE 2 Hourly Rates of State-Related Behaviors During Sleep (N = 41 Babies)

Smile .Frown Quiet sleep A Quiet sleep B Non-REM Active REM Active Dense REM Active

Grimace

Mouthing or sucking

Sigh sob

Startle

Jerk

Small movement

Large movement

0.0 0.0

0.4 0.7

1.4 2.5

9.7 8.3

3.4 3.4

13.3 4.3

7.6 5.4

7.2 13.7

1.8 5.7

2.8 3.6

7.4 11.6

25.0 8.8

31.6" 37.6

5.6 4.7

0.7 0.0

8.8 3.2

45.1 28.4

42.3 13.0

3.6

5.0

4.0

25.9

2.8

0.0

8.6

11.5

0.7

categories. For example, smiles, frowns, and grimaces rarely occur in Quiet sleep; mouthing or sucking, and body movements occur at much lower frequencies; and startles occur almost exclusively in Quiet sleep. In Quiet sleep B, the infant is less likely to startle and more likely to make small or large body movements than in Quiet sleep A. Among the Active sleep categories, there are more grimaces and more large and small body movements in Quiet sleep without REM and there are more frowns in Active sleep with REM. The distinct rate of occurrence of the various behaviors in the subcategories of Quiet and Active Sleep states strongly suggests the appropriateness of using these separated state categories to describe infants. Characteristics of Infants' States over the First Five Weeks of Life Distribution of sleep state time. Figure 4 presents the mean percent of total sleep that is Quiet Sleep for each of 5 weekly observations. There are ten infants represented at each time point except for the 3-month observation, at which time

100

eo

60 t--

... .....

z

LLJ

u

0:

w

40

...

...... ... ........

..,.-"'

ll.

20

2 Days

3

2

4

5

12

WEEKS

Fig. 4. Amount of Quiet sleep as a percent of total sleep, at 2 days and 2, 3, 4, 5, and 12 weeks of age.

100 D

eo

£

60 t--

z

w

u

0:

w

ll.

40

!!.

20

2 Days

2

3 WEEKS

FJG. 5. Amount oftime spent in each of the four categories of quiet sleep (expressed as a percent of total Quiet sleep), at 2 days and 2, 3, 4, 5, and 12 weeks of age.

131

132

EVELYN B. THOMAN

8

6

a:

:::J

0

:r: 0::

~

4

(/)

LU Cl

z

.05). The interaction of these effects was not statistically significant. During open-field testing two abnormalities were observed in a number of experimental subjects. These animals lacked normal neuromuscular control of the hind limbs, causing the feet to splay outward when rearing, resulting in an

164

R. E. BUTCHER, K. HAWVER, T. BURBACHER, W. SCOTT 200 180

Backward -Forward

[=::J

160 ~

e 140 w C> N

120

0

::::;: 100 0

~

80

c ~

60

::::;:

40 20

0 ......_ _ Control

HU-375

HU-500

FIG. 2. Mean total errors made during maze testing by 50-day-old rats exposed prenatally to hydroxyurea (HU groups) and their controls.

unsteady locomotor pattern. Also, kinked tail, which had previously been observed in only 3% of the 20-day fetuses receiving 500 mg/kg HUon day 12 of gestation, was found in a substantial portion (47%) of the older experimental subjects examined in this study. All subjects were weighed just before being maze tested. An analysis of variance performed on these data indicated significantly reduced weights in the HU-500 subjects (p > .01) and in female subjects (p > .01). All groups displayed a similar sex difference in weight, which was reflected in the analysis by a nonsignificant interaction term. No differences in weight between cross-fostered and non-cross-fostered subjects could be detected by directed t test for either sex (tmale = 1.92, d{ = 49, ]J > .05; [female = 0.09, dj = 73, jJ > .05). No significant differences attributable to any of the experimental variables were indicated in an analysis of variance performed on the data from the swimming speed trials administered prior to maze testing. An analysis of variance performed on mean total errors for the 4 days of maze testing (Fig. 2) examined two principal effects (treatments and phase of testing) and their interaction. Administration of HU had a significant effect (p = < .01). Significantly more errors were made by all groups in learning the backward path through the maze than in forward-path testing (p < .01). A nonsignificant interaction term indicated that in all experimental groups the difference between the number of errors made during the forward and backward maze paths was similar. A posteriori (Scheffe, 1953) analysis indicated that the combined performance of experimental offspring was significantly poorer (p < .01) than that of the controls. Although the number of errors made by the HU-375 group was greater than those of the controls a direct comparison of these two groups approached but did not achieve statistical signifi-

ANTENATAL EXPOSURE TO TERATOGENS

165

cance. The direct (Scheffe, 1953) comparison of the HU-375 and HU-500 groups, however, reached statistical significance (p < .01). Results consistent with those from the error data were obtained in an analysis of elapsed time (treatment, p < .01; testing phase, p < .01). To insure that the maze error results were not influenced by other experimental variables, two additional tests were performed which demonstrated that neither cross-fosterin~ (t = .751, df = 120) nor sex (t = .116, df = 120) had significant effects upon maze errors. Separate t tests were also made to determine whether the physical impairments present in the HU groups led to poorer maze performance. These tests revealed that the number of errors of subjects with kinked-tail or locomotor impairments or both were not significantly (p > .05) greater than those of animals without these abnormalities in either the HU-375 or HU-500 groups. EXPERIMENT II A second series of experiments has examined the postnatal behavioral effects of prenatal Diamox administration. This substance rather uniformly produces deformity of the forepaw in rats (usually ectrodactyly and ulnar hemimelia of the right forepaw). Diamox is thought to affect the central nervous system only rarely (Wilson, Maren, Takano, & Ellison, 1968). Method We administered 500 mg/kg of acetazolamide subcutaneously to four pregnant Sprague Dawley rats on the afternoon of the tenth day and the morning of the 11th day of their pregnancy. This dose of Diamox was expected to produce forepaw malformation rates between 35--45%. Four additional control females received distilled water adjusted to a pH of 9.2 with NaOH. To provide positive control subjects, three additional groups of four females were prepared. These subjects received 500 or 625 mg/kg hydroxyurea or distilled water (negative control) i.p. on the twelfth day of pregnancy. Our previous investigation indicated that offspring of females treated with hydroxyurea in this way would show impaired performance on our learning tasks. Diamox, hydroxyurea, and control subjects were prepared and subsequently tested concurrently. All litters were allowed to deliver normally, were reduced to number 10, and were examined externally for malformations at 5 days of age. Of the Diamox subjects 40% were observed to have a forepaw abnormality at this time. Litters were weaned and housed in individual cages at 21 days of age. Results At 30---35 days of age all subjects were examined in the open field test described previously. No difference in the amount of exploration was found. The malformed Diamox animals were observed to compensate for the affected limb and showed a "rocking" limp. Locomotor abnormalities, which included rear legs splaying

166

R. E. BUTCHER, K. HAWVER, T. BURBACHER, W. SCOTT

CJ BACKWARD 120

-FORWARD

-

~100 ~ Ill

:5a:

a: w w

80

;so N

.....

c ..... 0

..... 40

z

c

w

:IE 20

0...__ _ pH9.2

DIAMOX HU-500 HU-625 500

FIG. 3. Mean total errors made during maze testing by 50-day-old rats exposed prenatally to Diamox, hydroxyurea, and control solutions (H2 0 and pH 9.2 groups).

outward and a hopping gait, were present in 40% of the hydroxyurea 500 and 66% of the 625 mg/kg hydroxyurea animals. At 50-55 days of age all subjects were tested in the water filled six unit t-maze. No differences in preliminary swimming speed tests were noted. The results of this testing in terms of errors are displayed in Fig. 3. Approximately 30 subjects are represented in all groups except the hydroxyurea 625, which had 16. The performance of the Diamox animals was as good as either of the control groups prepared, while the hydroxyurea subjects made substantially more errors. An analysis of variance performed on these data indicated a statistically significant effect among the treatment groups (p < .01) and the comparison of individual groups (Scheffe, 1953) indicated that the Diamox subjects did not differ from the controls. The Diamox subjects, however, did perform significantly better than either of the hydroxyurea groups (p < .01 for all comparisons). The performance of malformed and normal Diamox was also compared by direct t test. No significant differences could be detected. Similar analyses performed on the elapsed time resulted in almost identical results.

ANTENATAL EXPOSURE TO TERATOGENS

167

Discussion

We have observed learning impairments without locomotor defects (sodium salicylate, hypervitaminosis-A) and with locomotor defects (hydroxyurea). Using Diamox we observe locomotor defects, but no learning impairments. These results suggest that the behavioral impairments we have observed previously are not effects of muscular or skeletal defects alone. It appears that agents which cause malformation of or damage to the fetal nervous system will produce behavioral impairments in offspring when given in moderate amounts. It may also be that substances which do not have a major effect upon fetal central nervous system will not produce behavioral impairment even when administered in quantities that induce frank malformation.

ACKNOWLEDGMENTS Supported in part by NIH grants HD00324, HD02792, and HD05221, and Contract No. 74-6 from the U.S. Food and Drug Administration.

REFERENCES Butcher, R. E., Brunner, R. L., Roth, T., & Kimmel, C. A. A learning impairment associated with maternal hypervitaminosis-A in rats. Life Sciences, 1970, Part 1, 11, 141-145. Butcher, R. E., Scott, W. J., Kazmaier, K., & Ritter, E. J., Postnatal effects in rats of prenatal treatment with hydroxyurea. Teratology, 1973, 7, 161-165. Butcher, R. E., Voorhees, C. V. & Kimmel, C. A. Learning Impairment from maternal salicylate treatment in rats. Nature New Biology, 1972. 236, 211-212. Scott, W. J., Ritter, E. J. & Wilson, J. G. DNA synthesis inhibition and cell death associated with hydroxyurea teratogenesis in rat embryos. Developmental Biology, 1971, 26, 306-315. Wilson, J. G., Maren, T. H., Takano, K., & Ellison, A. Teratogenic action of carbonic anhydrase inhibitors in the rat. Teratology, 1968, l(No. 1), 51-61. Scheffe, H. A. A method for judging all possible contrasts in the analysis of variance. Biometrika, 1953, 40, 87-104.

13 MORPHOLOGICAL AND BEHAVIORAL CONSEQUENCES OF CHEMICALLY INDUCED LESIONS OF THE CNS 1

Patricia M. Rodier William Webster Jan Langman University of Virginia

With the development of autoradiography, it has become possible to date the time of origin of cells in developing organisms. Through the work of many investigators, a chronology of the spinal cord and many brain regions has evolved. One effect of this work has been to push our estimates of the time when cell formation ceases past the time of birth. In the rat, for example, some neurons are formed as late as the onset of puberty, while some glia are still produced in mature adults (Altman, 1966). In man, as in rodents, many cerebellar neurons are produced after birth (Raaf & Kernohan, 1944; Rakic & Sidman, 1970). Teratologists have tended to focus on the period of organogenesis, when insults lead to gross malformations. The same insults late in pregnancy or early in postnatal life do not result in obvious abnormalities. Yet these periods include the birthdays of unique cell types in the CNS, and loss of even a few neurons could have serious consequences for an otherwise healthy animal. We have used several drugs believed to interfere exclusively witlt cell proliferation to eliminate specific cells in the developing CNS (Langman & Shimada, 1971; Andreoli, Rodier, & Langman, 1973). The treatments are systemic, but since they damage only those cells that are synthesizing DNA, they are extremely selective. Since the neurons destined for various adult structures are produced in a sequential pattern, the region of cell loss should depend on the time at which cell proliferation is interrupted. This is true not only of drugs that interfere with cell proliferation, but of x-ray, and, most important to those interested in congenital brain damage, of DNA viruses, such as measles. Our purpose, then, was to compare treated animals to controls, both morphologically and behaviorally, and 'Supported by NIH grant NS06188.

169

170

P.M. RODIER, W. WEBSTER, J. LANGMAN

a

POSTNATAL DAY 3

FIG.

1. (a) Adult position of heav-

ily labeled cells in animals injected

b

GESTATION DAY 19

10

with tritiated thymidine on postnatal day 3 (2 inj. x 6 {tC/g, 6 hr apart). (b) on gestation day 19 (2 inj. x 6 {tC/g, 6 hr apartf (c) on gestation day 15 (2 inj. X 6 {tC/g, 6 hr apart).

14

7

c

GESTATION DAY 15

further, to compare animals treated at different stages of development to one another. The immediate effects of treatment with 5-azacytidine have been described by Langman and Shimada (1971). About 2 hr after a pregnant mouse was injected with azacytidine, a few abnormal mitotic figures appeared at the lumen of the neuroepithelium in the fetuses. Over the following hours the affected cells became more and more pycnotic and their number increased. Surprisingly, they migrated away from the lumen in an apparently normal fashion, but 24 hr after treatment many were disintegrating, and by 48 hr most had disappeared. When animals treated with azacytidine are compared to those injected with tritiated thymidine, the regions of cell death match the regions of cell proliferation. When we have treated animals with azacytidine and labeled simultaneously it is clear that only labeled cells become pycnotic. Thus, the structures damaged by azacytidine should be the same as those labeled at the same stage of development. Figure 1 shows the adult position of cells heavily labeled on postnatal day 3, embryonic day 19, and embryonic day 15. According to the labeling experiments, the cells undergoing their final division on PN 3 are: (1) glomerular layer cells of the olfactory bulb; (2) internal granular layer cells of the olfactory bulb; (3) subependymal cells; (4) granule cells of the dentate gyrus; (5) internal granular layer cells of the cerebellum and the cells surrounding the Purkinje cells.

c

CONSEQUENCES OF CHEMICALLY INDUCED CNS LESIONS

171

The same areas were labeled by injection onE 19 as shown in Fig. lb. More internal granular layer cells and fewer glomerular layer cells were involved by the earlier labeling, and the distribution of labeled cells in the cerebellum seemed more restricted to the vermis. Scattered labeled cells were observed in the corpus striatum (6). OnE 15 neurons for many structures were undergoing their final divisions (See Fig. lc). Many cells in the corpus -striatum were heavily labeled in adults after thymidine injections onE 15. Labeling in the olfactory bulb was almost exclusively related to mitral cells (7). Several layers of cerebral cortex (8) showed heavy labeling, along with many pyramidal cells of the hippocampus (10). The label in the olfactory tubercle was striking (11). Regions where cell proliferation must have been nearing completion were the septal nuclei (9), the inferior colliculus (14), the mammillary nuclei (12), and some pontine nuclei (13). The amygdala, especially the medial cortical nucleus, included many heavily labeled cells. It is too lateral to appear on the diagram. The description of missing cells is considerably more difficult than the description of labeled cells. Some of the damage produced by azacytidine on E 15 is obvious. Figure 2a and b shows the thin, small cerebral cortex of a treated mouse, and in plastic sections, Fig: 2c and d, the gaps in the pyramidal cell layer of the hippocampus. The corpus striatum is not obviously affected, but its width in histological sections is consistantly reduced. Quantifying deficits in the scattered small cells involved by later treatments is extremely difficult. One cell type is amenable to counting, the large light-staining cells in the Purkinje cell layer, which may be Golgi epithelial cells. The number of these does appear to be reduced by azacytidine treatment on PN 3 (Figs. 2e and f). Whatever their other characteristics, the azacytidine-treated animals are true small-for-date babies. Table 1 shows data on the physical development of the three experimental groups and controls. Weight differences between treated animals and controls were significant for all groups in infancy. The trend for body weights to decrease from controls to PN 3' s to E 19' s to E 15' s persists into maturity, but only E 15' s differ significantly from controls as adults. The same pattern of results occurs in measures of brain weight. The attainment of developmental landmarks was retarded in all groups, and the earlier the treatment the greater the delay. No gross malformations were observed in any of the animals studied-a total of several hundred. All the animals used for behavioral testing were conceived on the same day. When any group received an azacytidine treatment (8 mg/kg in 2 injections, 6 hr apart) all other groups received saline injections. Prenatal treatments were administered intraperitoneally and postnatal injections were given subcutaneously. In contrast to the general retardation observed in physical characteristics, behavioral observation revealed different deficits in the different treatment groups. The pattern of behavioral abnormalities suggests that the functional changes observed are related to specific sites of cell loss, rather than the extent of cell loss.

FIG. 2. (a) Paraffin section (4 µ,) of adult mouse injected with saline on gestation day 15 (hematoxylin-eosin). (b) Paraffin section (4 µ,) of adult mouse injected with 5-azacytidine (2 inj. X 4 mg/kg, 6 hr apart) on gestation day 15 (hematoxylin-eosin). (c) Methacrylate section (I µ,) of adult hippocampus after injection with saline on gestation day 15 (azure II). (d) Methacrylate section (Iµ,) of adult hippocampus after injection with 5-azacytidine (2 inj. X 4 mg/kg, 6 hr apart) on gestation day 15 (azure II). (e) Paraffin section (4 µ,)of adult cerebellum after injection with tritiated thymidine (2 inj. X 6 µ,C/g , 6 hr apart) on postnatal day 3 (hematoxylin-eosin). (f) Paraffin section (4 µ,) of adult cerebellum after injection with tritiated thymidine (2 inj. x 6 µ,C/ g , 6 hr apart) and 5-azacytidine (2 inj. x 4 mg/kg, 6 hr apart) on postnatal day 3 (hematoxylin-eosin).

172

CONSEQUENCES OF CHEMICALLY INDUCED CNS LESIONS

173

Table 2 gives some examples of functional anomalies observed after azacytidine treatment. On postnatal day 6, all animals were tested for righting responses. Time-toright was greatly increased in the PN 3 and E 19 groups, but not in theE 15 group. Thus, the animals believed to have cerebellar damage had difficulty in righting, while theE 15's righted normally despite brain damage in many other areas. By the 9th day after birth all groups could right themselves almost immediately. Pivoting is an infantile locomotor behavior in which the animal rests on a collapsed hind leg and moves the front legs with the result of spinning in place. This behavior persisted in PN 3's and especially in E 15's. Evidently damage in several motor areas can lead to this abnormality, for no damage to motor systems is common to the two affected groups. It is interesting that E 19-treated mice showed no locomotor impairment, and were even precocious in some ways. PN 3's showed a gait characterized by poor control of the hind limbs. Besides pivoting, they held the hind legs abducted and tended to walk on the medial side of the foot rather than on the sole. The forward excursion of the hind limbs was reduced, so that the feet were often posterior to the body during all phases of the gait. Early in development, E 15's had an efficient gait, but in the second week they began to show locomotor difficulties. Like the PN 3's, they pivoted and held the hindlimbs in abduction, but unlike the later treatment group, they walked on the soles of their feet and had an exagerated forward excursion of the limb. Some crossing of the front limbs was noted-a form of incoordination rarely seen in other groups. Coarse intentional tremors were seen in all treatment groups. They appeared during the first week in E 19' s and PN 3' s and later in the E 15 animals. Although these axial tremors are a classic "cerebellar sign" they have also been associated with lesions of the basal ganglia (Martin, 1967). General activity as measured by an induction coil activity meter (LKB Farad) demonstrated a clear dependence on time of treatment. BothE 15's and E 19's were significantly different from controls, but E 15's were hyperactive, while E 19's were hypoactive. Thus, the same treatment delivered on different days of development had opposite effects. No physiological psychologist can resist the temptation to relate these behavioral anomalies to the brain structure damaged--one is reassured to see hippocampal damage associated with hyperactivity or cerebellar lesions associated with locomotor difficulties. Yet, with these combination lesions of several structures, a firm relation of structure to functional deficit is presumptuous. What is certain is that different patterns of cell loss do result in different syndromes of behavioral abnormalities. The description of such syndromes is directly relevant to human brain damage, for many insults interfere with cell proliferation and congenital brain damage must frequently involve the kinds of subtle, widespread lesions produced in these experiments. If one considers the many experiments involving agents known to produce cell loss in proliferative populations, there is a wealth of evidence that suggests our

TABLE I Physical Development

Mean body weight at 5 days (gm) Mean body weight at weaning (gm) Mean body weight at 8 weeks (gm)a Mean brain weight at 8 weeks (gm)a Order of eye.opening Order of ear opening Order of fur growth

Control

PN 3

(N = 11)

(N = 10)

3.7

3.5*

E 19 (N

=

11)

2.9**

E 15

= 22)

(N

2.1 **

21.4

22.2

21.9

15.8**

38.4

35.8

33.6

24.3**

.555

.547 2 2 2

.534 3 3 3

.413** 4 4 4

*p < .05; **p < .01. aThese data were taken from groups prepared for antomical study only. Other figures are from animals prepared for behavioral study.

TABLE 2 Behavioral Development

Controls (N

Proportion above median righting timeat 6 days Mean seconds/minute pivoting at 8 days Proportion showing tremor at 7 days Proportion showing tremor at 17 days Mean activity/2 minutes in adults **p

174




E

z

6.0

...J ILJ

>

ILJ ...J

z 0

1-

0

a:

en I

ILJ

a: a..

-

!i" f\

-

-

:r

ILJ ...J

w a.. 2.0 ::E z

0

-

1

ILJ 0

0

~0

RELATION OF s 1 to S 2

o-o SAME e-e DIFFERENT

r~ 1~\ 0

lJ..

0

{' to\ 52 ~

~

:r a..

2.0

z

1-

(/)

4.0

~s, ~

D ~

~0

52

•

SUCCESSIVE 5-SECOND INTERVALS

FIG. 9. Mean average evoked responses to successive auditory stimuli, S 1 and S2 , when the two were the same or different frequencies. Plotted separately are the responses to those pairs involving no intervening stimulation (top panel) and those pairs separated by a distractor stimulus, D, as indicated (bottom panel). (Reprinted from Whitlow, 1975. Copyright 1975 by the American Psychological Association. Reprinted by permission.)

the Terry (1976) observation. The experiment was closely patterned after the previous study with the exceptions that (a) the S 1 -S 2 interval was consistently 60 sec, and (b) on half of each of the occasions in which S2 was the same tone as sl or was different from sl, the interval included a dis tractor stimulus 20 sec after S1 and thus 40 sec prior to S2 . From the data presented in Fig. 8, it could be expected that in the absence of the distractor stimulus there would be ample stimulus-specific response decrement to S2 at this S1 -S 2 interval. The question was whether or not the distractor (which was a sequential compound of a 1-sec flashing light and a 1-sec electro tactile stimulation of the cheek) would remove the decrement, as extending the duration of the S1 -S 2 interval to 150 sec had been found to do.

196

ALLAN R. WAGNER

Figure 9 summarizes the relevant data in the same manner as Fig. 8. When there was no intervening stimulation (see top panel) the response to S2 was reduced on occasions in which S 1 was the same stimulus relative to occasions on which S 1 was a different stimulus. When the distractor was presented between sl and s2 (see bottom panel), it produced an evoked response itself, but more importantly, it removed any differential effect upon S2 responding of the same vs. different S1 events. These findings are what one would expect if the stimulusspecific response decrement were dependent upon a perseverating representation of the test stimulus in STM, and the distractor acted to remove such representation. There are a variety of reasons why the unconditioned response to a stimulus might be reduced on the second of two temporally adjacent presentations. Sensory adaptation or response system fatigue are two obvious possibilities and each would lead one to expect a recovery function. Whitlow's (1975) studies are particularly informative in indicating that such accounts would not suffice. The stimulus-specificity of the decrement argues against simple response system fatigue, while the restorative effect of the distractor is outside of our notions of what might remove sensory adaptation. The distractor, furthermore, did not appear to act as a general sensitizer. It appears preferable to suppose that the refractory-like response decrement as observed by Whitlow (1975) is due to the persistence of S1 representation in STM. We can then assume that S2 , if it is the same as S1 , will be less likely than it otherwise would be to occasion full processing (rehearsal) in STM. V.

EXTRAPOLATION

There are a considerable number of implications of the priming notion for phenomena outside of the data domain which initially provoked the formulation. In the space available I will mention just two that are particularly relevant to the overall theme of the paper. Even thus restricted, the discussion will necessarily be rather superficial. But it should suffice to indicate certain potentialities for theoretical interpretation and some promising areas for further research. A. Distribution of Trials

First, consider the implications of the model for distribution -of-trials phenomena. What is performed and what is learned on trial n of a sequence of training trials should depend importantly on what remains in STM from preceding trials. In some situations the dependencies may be relatively complex (see, e.g., Pfautz & Wagner, 1976), but arathersimplesetofexpectationscanbeseen to be confirmed by a dissertation conducted in our laboratory by Michael Davis (1970). The study was provoked by the authoritative assertion of Thompson and Spencer (1966) that one of the nine defining characteristics of "habituation" is that it proceeds faster the shorter the intertrial interval. This is an adequate character-

7.

EXPECTANCIES AND THE PRIMING OF STM

197

ization of the response decrement seen with repeated stimulation at different fixed intervals, but seemed to us likely to be an oversimplification if there is any short-term, refractory-like decrement in stimulus processing as we have now seen in the Whitlow (1975) studies. Then one might well expect to see, throughout an habituation sequence, less unconditioned responding to a stimulus the shorter the time since its last presentation, i.e., the more likely it was still represented in STM. But if an habituation decrement also follows from alterations in Long-term Memory, such alterations might be less likely to be witnessed the shorter the intertrial interval. Davis employed a stabilimeter device to measure the startle response of the rat to 50-msec, 4,000-Hz tones that increased the sound pressure level from 80 to 120 db. In an initial prehabituation test and an identical posthabituation test, all subjects were exposed to a series of 300 tones during which the interstimulus interval was varied among 2, 4, 8, and 16 sec in an order that equated the first-, second-, and third-order sequential probabilities. Between these two tests all subjects received an habituation series of 1,000 tone stimulations, half of the subjects with a constant 2-sec interstimulus interval, the remainder with a constant 16-sec interval. 100 . .-.--~---,----,

z

0

a_

HABITUATION

PRE-

w

80

w a:: 1--

a::

60

(.)

z

!c....

.20

...-- ,/'--

o neg • pos

i= .60 !"'----Jl

0

z

0

D.

(/)

UJ

a:

IL

0

UJ

~ a:

140

~117

~ .,.

20

UJ

(/)

~

~

120

X:'•

100 80

60 40

~

20 0

1231234567123 SESSION

BLOCKS

FIG. 11. Latency ratios and response rates during the poststimulus for three pigeons. During the first and the last three blocks of five sessions, vertical and horizontal lines were presented during the poststimulus periods. During the middle seven blocks of five sessions, only an orange field was presented.

2. Results. The latency ratios are presented in Fig. 11. The most notable aspect of the data is the very small disruption of appropriate control key responding when the vertical and horizontal lines were removed. S18 at first suffers a considerable decline in accuracy, but then recovers to a level of performance quite comparable to the baseline conditions. The performance of S115 was practically unaffected. S117 shows rather little effect for 15 sessions, then some decline in the latency ratios of positive trials, then a recovery, and then a decline. All birds do very well during the return to the previous training condition with lines alone as poststimuli. This result was unexpected. Why would performance on the memory task not collapse in the absence of feedback from the poststimuli? One possibility

8.

WORKING MEMORY IN THE PIGEON

239

was that the birds "knew" that the trials ended in the reinforcement or extinction conditions even in the absence of differential cues on the key. We determined this by looking at the response rates during "positive" terminal periods following correct behavior toward the control key, and during "negative" terminal periods following an error. These rates are also plotted in Fig. 11. They differ markedly during the baseline and the postexperimental condition, when vertical and horizontal lines were displayed. The difference vanishes when these stimuli are removed, and rates are maintained at a high level. If anything, they are slightly higher for two birds during the extinction condition, presumably because no time is spent consuming the reinforcer.

3. Discussion. The criterion performance in our working-memory paradigms was established with differential reinforcement of the control key response. It is hard to imagine how it would be achieved otherwise. Why this differential reinforcement was only marginally necessary to maintain the behavior in very welltrained subjects is certainly not obvious. One possible explanation is this: Once the control-key paradigm is mastered, every trial can end with a positive outcome, and a bird that performs well has contact with only one post stimulus (S+) over the course of many training sessions. Thus, the presentation of different stimuli may not be necessary to maintain the behavior. The data do provide some modest support for this notion. S18 made the most errors during baseline training (8.67 per session, on the average), and was most disrupted by the removal of the lines from the poststimuli. SllS made only about four errors per session, and S117 about two, and these subjects were less affected. Nonetheless, SIS improved his performance during the experimental sessions even without differential feedback, and S117 did likewise for a few sessions after a deterioration of his performance. One empirical approach to testing this interpretation would be to run subjects in a parallel experiment with the advance key rather than the control key procedure. In the former, half the trials end with S+ and half end with S- whether the subject makes the advance response or not. Thus, the subject will continue to have contact with both poststimuli. Another possible interpretation of the present results is simply that our highly trained subjects failed to process the poststimuli in relation to their own performance. If the stimuli did not functionally serve as "feedback," then the lack of a differential outcome would not be expected to affect the behavior. This notion is compatible with a theoretical account of working memory that follows in the next section. VIII. GENERAL DISCUSSION

In the current set of experiments, we have established a phenomenon, namely a stable, relatively long-term working memory in the pigeon. The most obvious and important single feature of our data is the relative insensitivity of this

240

WERNER K. HONIG

memory to the passage of time, at least within the limits of 30 sec. This insensitivity is manifested by the "best data" plots derived from various training conditions, and also by the lack of an interaction between delay intervals and conditions which reduce performance, as observed in the study on the role played by reinforcement delivery and pecking at the food key. This does not, of course, imply that our pigeons never forgot during the delay interval, nor that such forgetting is independent of the passage of time. But forgetting does not seem to be inexorably related to the passage of time, and this affects theoretical interpretations of the data. I shall discuss some interpretations of the data that would be suggested by the theoretical approaches that were reviewed briefly in the introductory section of this chapter. A. Trace Theory Perhaps the most traditional interpretation of our results would be that the prestimuli establish particularly strong traces which are well maintained over extensive memory intervals. Extensive data summarized by Roberts and Grant (1976) show that those treatments which are expected to produce "strong" traces also retard the rather rapid forgetting that typically occurs with DMTS in the pigeon. For example, increasing the duration of the sample improves performance. It could be argued that our S+ prestimuli generated particularly strong traces because of their duration, the correlated occurrence of reinforcement, and the high rate of pecking at the food key. The S- prestimuli may have generated weaker traces because some of these cues were not available. A reasonable explanation of our results could be this: The pigeon simply learns not to peck the advance key in the presence of the strong trace generated in S+ trials. Otherwise (in the absence of a trace, or in the presence of a trace of S-) he pecks the key. When components of the positive trace are removed, then the trace is weakened, and performance following S+ will be reduced, as it was in our experiment relevant to this question. In my view, such an explanation is not adequate to deal with the data presented here. By its very nature, a trace should be sensitive to the passage of time. Forgetting should be orderly. Furthermore, any treatment that strengthens or weakens a trace should enhance or reduce the performance based on memory, but always within the context of an orderly decay function. Data fitting such a model are shown in Fig. 1, where the effects of sample duration upon DMTS are shown. We have no evidence in our results of a clear limitation upon working memory, nor of an orderly decay function, even when the trace. of the positive stimulus is presumably weakened by removing food reward and the opportunities to peck at the response key. A second problem with the trace theory is that it cannot account for our transfer data. A trace is presumably a trace of a physical stimulus. Presumably the trace acts as a cue for the correct response following the interstimulus. In

8. WORKING MEMORY IN THE PIGEON

241

that case, our subjects must have transferred control of responding from traces of line orientations to traces of colors, which had never before been presented within the context of the memory paradigm. Our interpretation of these results was based on the equivalence of the associative values of the stimuli, but in trace theory, to this point, it has not been suggested that associative values generate particular traces. B. Activation Theory

Lett (in press) has offered a theory of associative memory, according to which the memory of an initial event become "inactive" following the event, and is then reactivated when the subsequent event occurs which is to be associated with the first. Such reactivation normally results from presenting a "retrieval cue," and Lett explains her own long-delay learning in rats on this basis, namely that the memory of the rat's response is "protected" in the home cage while it is inactive, and is then activated when the rat is fed in the training apparatus. It may be possible to extend such a theory to cover working memory as well, and, indeed, where the associative and working memories must be acquired conjointly, as in alternation studies in the runway with long intertrial intervals, this seems like a reasonable approach. In our case, one would have to assume that the memory of the prestimulus becomes inactive when that stimulus is terminated, or shortly thereafter, and is then reactivated when the advance or control key is illuminated. Such a theory could account for the slow decline of memory as a function of the delay interval. It should have no difficulty with our transfer data on the assumption that the inactive memory need not be restricted to the purely physical aspects of the prestimulus. It is not quite clear, however, why the illumination of the advance key should serve as a retrieval cue. This stimulus is not present during the prestimulus, as retrieval cues usually are. Furthermore, there is no clear mechanism for the termination of the working memory. The theory was designed to account for the establishment of stable associations. If inactive memories are activated through the retrieval cue, such as the illumination of the advance key, why would there not be a lot of interference between the memory laid down on previous trials and the memory from the current trial? C. Forgetting as Discrimination Failure

Little of our work is directly relevant to the views of D' Amato (1973) and of Worsham (1975) in the sense of providing explicit tests of their hypothesis. But a pragmatic analysis of our findings would not seem to support them. Relevant data for their view are largely provided by proactive interference. In our situation, the same stimuli are generally used as pre- and poststimuli, and appear on the food key. It is thus reasonable to suppose that proactive interference on a

242

WERNER K. HONIG

given trial would primarily arise from the rather lengthy poststimulus from the previous trial. Now consider two consecutive trials with the advance-key procedure. If these trials begin with the same prestimulus, then the poststimulus of the first must always be opposite in sign from the prestimulus of the second, and. this should create interference. Conversely, if they begin with different prestimuli, then the adjacent post- and prestimuli are of the same sign, which should reduce interference. Unfortunately, we have not generally obtained trialby-trial records, so the appropriate analyses cannot be carried out at this time. But other considerations lend little support to this view. A well-trained bird will make rather few errors, and these tend to be concentrated in the first few trials of a session. We have no evidence that trials within a session "cluster" into good and bad sets, depending on the outcome of the previous trial. In the control-key procedure, a subject that makes few errors can generally obtain S+ as the poststimulus on most trials. This suggests that performance should be poorer on trials starting with S- rather than S+. There is no evidence for this. Furthermore, one would predict that if the pre- and poststimuli are taken from different dimensions (with colors as pre stimuli, and lines as poststimuli, for example), performance should be better than if they are from the same dimension. There is no evidence for this either. Admittedly, these arguments are weak, since the evidence is indirect. Trial-bytrial analyses are indeed crucial, particularly during the earlier training phases, when the subject may not have had the opportunity to learn a discrimination between the prestimulus which follows the intertrial interval, and the previous poststimulus, which precedes it. And yet we should note that the theory of D'Amato and Worsham is a theory of forgetting, unlike trace theory and reactivation theory, which are theories of remembering. Since most of our evidence concerns remembering rather than forgetting, I am really more interested in an adequate theory that deals with the former. D. Memory as an "Instruction"

I want to offer a fourth alternative. This is that the prestimulus establishes an "instruction," which is a state or process that determines the criterion response following the memory interval-whether the pigeon will peck or not peck at the control or advance key. This instruction has the following characteristics: (1) Unless it is forgotten in the course of the memory interval, it is maintained at full strength, possibly with the assistance of "rehearsal" or some form of differential behavior. (2) The instruction is a discrete state; the initial stimulus establishes it at full strength, if at all, and if it is forgotten, it is forgotten completely. (3) Once the criterion response is made, the instruction is terminated, and has no effect upon subsequent behavior. (4) The instruction is established by the initial stimulus, but it is not necessary for the subject to remember the stimulus itself. This hypothesis has certain clear implications for characteristics of working memory. We can review relevant aspects of our findings, which are, not surpris-

8. WORKING MEMORY IN THE PIGEON

243

ingly, quite compatible with these implications. Furthermore, we can make predictions about the effects of treatments that have not yet been carried out. First, we should note that forgetting is not an inherent characteristic of the instruction, unlike the fading of a trace. The probability of forgetting increases with duration of the memory interval, but the causes of forgetting are probably interference, or a failure to maintain the instruction through rehearsal or some other means (inattentiveness?). Thus, the notion of an instruction is compatible with the relatively long delays that we have observed with little decrement in performance. Second, the notion is compatible with our transfer data. An instruction could be cued by any of a number of functionally equivalent stimuli, separately or together. If these cues are reduced in number, or weak, then the instruction is less likely to be correctly established, and this is quite compatible with our experiment in which the role of various aspects of the initial period was investigated. In that experiment, there was no interaction between the characteristics of the initial period and the memory interval. It would seem that the probability of establishing the instruction, rather than the forgetting of the stimulus, was affected by the experimental treatments. Finally, the notion of an instruction may also explain the last experiment described in this chapter. If the pigeon learns to terminate the instruction upon making the criterion response, then the poststimulus as a reinforcer may not enter into the evaluation by the subject ofits preceding behavior. 1. Prospective tests of the instructional hypothesis. Certain suggestions for future research emerge from this hypothesis, and fairly clear predictions can be made. Since the instruction is established at full strength, if at all, working memory ought to be relatively insensitive to physical parameters of the initial stimulus. For example, duration should not markedly affect performance, although there would be some limit below which a duration would be ineffective. We actually carried out some work on this question with our very first pilot subjects run on the advance key procedure. Shortening initial periods from 20 to 10 and even to 5 sec, and increasing them to 30 sec, had rather little effect. Latency ratios on S+ trials declined somewhat with the shorter durations, but they were unaffected on S- trials. This change may have been due to the reduced number of initial S+ periods of short duration in which a reinforcer was presented. At the time, we did not realize that reinforcers could provide at least part of the basis of the working memory. The matter of stimulus duration needs to be studied again more systematically, with numbers of reinforcers equated for initial periods of different durations, and with the use of several memory intervals. Another prediction based on the instructional interpretation is this: Once the criterion response is made, the instruction is terminated; therefore, the subject should respond at a chance level if the memory interval were to be continued. This could be tested with two opportunities for a criterion response within a trial. Pigeons could be trained with memory intervals of, say, 20 and 30 sec on

244

WERNER K. HONIG

different trials in each session. Then the advance or control key could be illuminated for 5 sec within selected 30-sec trials, but after a 20-sec delay. A response to it would be ineffective, aside from turning off the key. At the end of the 30-sec delay, the key would go on again, the the pigeon could make another criterion response. If the instruction is terminated after the first opportunity, performance on the second opportunity should be poor. Further, it would be good to determine whether there is any behavior that could be interpreted as accompanying, sustaining, or directing an instruction during the memory interval. This need not be mediating behavior in the usual sense of providing a stimulus that elicits a criterion response, but merely a behavior that facilitates the maintenance of a "set." We have videotaped some of our subjects during the interim periods. While they respond to the food key fairly steadily, they frequently make head movements in the direction of the control key if a peck at it will be the correct response, and they make other head movements ~ up and down, or in the opposite direction - if such a response should be withheld. However, interruptions of these patterns, occasioned, for example, by delivery of food during the interim period, do not necessarily disrupt the correct response. Nonetheless, a correlation between the occurrence of such behaviors and the accuracy of responding to the control key is worth further study. If differential behavior patterns are seen during the memory intervals of S+ and S- trials, the experimenter could deliberately illuminate the advance or control key either while the pigeon is performing the "appropriate" pattern, or when he switches to the "inappropriate" pattern. This may result in very different probabilities of making the correct response to the control key. 2. Other research that is relevant to the instructional hypothesis. An informal way of describing the instruction is that the pigeon remembers "what to do" rather than "what it saw." This suggests that the instructional hypothesis is not applicable to those paradigms where the memory of particular physical characteristics of a stimulus is required for the criterion response. In DMTS, it is reasonable to suppose that the choice between comparisons is directed by the relationship between the memory of a particular initial stimulus and the comparison stimuli. In pigeons, performance on this task is characterized by an orderly decline as a function of the duration of the memory interval, even with intervals of 20 to 60 sec (Grant, 1976). It is also sensitive to the parameters of the initial stimulus, particularly duration (Grant, 1976; Roberts & Grant, 1974). It appears to be subject to proactive interference, which we would not expect to find from a "terminated" instruction (Grant, 1975). In our situation, the pigeon had to remember what to do in the sense of making a particularly well-trained response. An instruction may not be able to encompass a discrimination between two choice stimuli to be presented later in a trial. Thus, if we were to present two different control keys, and the pigeon would have to choose the correct one on the basis of an initial stimulus, we

8. WORKING MEMORY IN THE PIGEON

245

might well obtain results that are close to those typical for DMTS. But possibly other, more advanced species are not limited in this way. Monkeys and dolphins, for example, generally have to choose one of two very familiar stimuli at the end of a trial in a working memory paradigm. Now this choice may be under the control of an instruction established earlier in the trial, e.g., "choose the red panel," or "swim to the high tone." In other words, the subject's capacity may permit him to include a characterization of the correct stimulus in the instruction. The choice would then not be a discriminated response made to the fading trace of a sample stimulus presented earlier. However, we should note that the instructional hypothesis is "stretched" by the study by Herman and Gordon {1974), who used a set of 17 different auditory stimuli to compose their DMTS problems, each of which was unique. We would have to postulate that the instruction could incorporate a representation of a relatively unfamiliar stimulus in such an intelligent species. Some of the data obtained with monkeys and dolphins conforms more closely to an instructional than to a decay hypothesis of working memory. For example, D'Amato and Worsham (1972) showed that the duration of the sample stimulus had little effect upon DMTS in the monkey, even though durations were reduced to less than .1 sec. In a later paper (D'Amato & Worsham, 1974), they also show that with sufficient preliminary training, monkeys can do as well on conditional delayed matching as they do on DMTS. These findings are not easily handled by a trace theory, nor are they accounted for by the notion that in matching problems the subject d;epends upon retrieval cues provided by the comparison stimuli. They are more readily explained by the hypothesis that the sample establishes an instruction to choose a particular stimulus which it need not resemble. Furthermore, in much of the work reported by D'Amato and his colleagues, there is little decline in performance over substantial delay intervals. Typically, intervals up to two minutes are used. At the longer intervals within this range, there is considerable forgetting in many cases, but this is not a universal finding during "baseline" training. It is most often seen under experimental conditions designed to induce forgetting. The "decay" functions are high and flat at lower intervals, and then drop at longer ones; this is in contrast to most data obtained from pigeons, where the memory loss occurs mostly at the shortest intervals. Herman and Gordon (1974) similarly report excellent retention of sounds in a DMTS paradigm by a bottlenosed dolphin for intervals up to two minutes. A loss of memory seems, in general, to be due to retroactive interference, rather than to the passage of time alone. This has been demonstrated systematically by D'Amato and O'Neill (1971) and by Worsham and D'Amato (1973); such a phenomenon is, of course, compatible with the instructional hypothesis. Herman (197 5) studied the effects of interpolating sounds during the delay interval in his auditory DMTS task with the dolphin, and obtained interference effects if the sounds filled a large portion of the interval.

246

WERNER K. HONIG

If the instruction is canceled when the criterion response is made, then there should be no proactive interference in subsequent trials. But such interference has been observed, and it causes problems for the instructional hypothesis. It has been documented by Worsham (1975) in monkeys, and Herman (1975) also reports that if only two stimuli are used in a delayed matching task, the dolphin is more likely to make an error on the second of two consecutive trials if the sample stimuli are different than if they are the same. Clearly, if the performance of the criterion response cancels the entire memory of the events on that trial, the instructional hypothesis cannot meet the test of such data. The possibility remains that the animal may cancel the instruction as such, but that some memory remains which could influence the establishment of instructions on subsequent trials. This could well be the case when the discrimination requires that the instruction contains particular information on the characteristics of a stimulus that is to be chosen. If the animal has to remember what to do in relation to what it saw or heard, the memory of the latter may persist, although not in the form of an instruction. The instructional hypothesis does account for much of the data obtained by Olton and Samuelson (1976) and Olton (this volume, Chapter 12) on spatial memory in rats. It will be recalled from the introduction that these subjects could remember within the course of a trial which of eight possible arms they had entered. Olton suggests that the rat stores information on where he has been, and implies that this information is used to instruct itself to avoid returning to the same places throughout the duration of the trial. Furthermore, Olton explicitly incorporates a "reset" mechanism in his model, because the memory of earlier trials does not seem to interfere with performance on later trials. Error analysis based on the sequence of choices within trials does not suggest that earlier choices are forgotten sooner than later ones in the course of a trial, which argues strongly against intratrial interference and against a fading of traces. Instead, it supports my earlier suggestion that some animals, at least, can incorporate representation of specific stimuli - in this case the alleys that the subject has entered within the same trial - into a working memory that directs subsequent behavior. Whether the pigeon has any such capacity we do not know. Many results, reviewed briefly in this last section, converge upon a hypothesis that suggests a memorial process in which information is held with rather little time-related loss, and which directs subsequent behavior within particular trials. The information may incorporate the representation of a future stimulus that is to be chosen, or the representation of stimuli that have already been sampled. The memorial process is likely to vary in complexity and in the details of its mechanism from species to species and possibly from procedure to procedure. It leads to the conclusion that the control of behavior is not limited to the fading traces of particular physical stimuli, but that behavior can be governed by information which is representative of particular stimuli, and which is stored, used, and forgotten as the occasion requires.

8.

WORKING MEMORY IN THE PIGEON

247

REFERENCES Blough, D. S. Delayed matching in the pigeon. Journal of the Experimental Analysis of Behavior, 1959,2, 151-160. Capaldi, E. J. Partial reinforcement: A hypothesis of sequential effects. Psychological Review, 1966, 73,459-477. Capaldi, E. J. A sequential hypothesis of instrumental learning. InK. W. Spence and J. T. Spence (Eds.), The psychology oflearningand motivation (Vol. 1). New York: Academic Press, 1967, pp. 67-157. Capaldi, E. J. Memory and learning: A sequential viewpoint. In W. K. Honig and P. H. R. James (Eds.), Animal memory. New York: Academic Press, 1971, pp. 111-154. Capaldi, E. J., & Stanley, L. R. Temporal properties of reinforcement aftereffects. Journal of Experimental Psychology, 1963, 65, 169-175. Cohen, L. R., Looney, T. A., Brady, J. H., & Au cella, A. F. Differential sample response schedules in the acquisition of conditional discriminations by pigeons. Journal of the Experimental Analysis of Behavior, 1976,26, 301-314. D'Amato, M. R. Delayed matching and short-term memory in monkeys. In G. H. Bower (Ed.), The psychology and learning and motivation: Advances in research and theory (Vol. 7). New York: Academic Press, 1973, pp. 227-269. D'Amato, M. R., & O'Neill, W. Effect of delay-interval illumination on matching behavior in the capuchin monkey. Journal of the Experimental Analysis of Behavior, 1971, 15, 327-333. D'Amato, M. R., & Worsham, R. W. Delayed matching in the capuchin monkey with brief sample durations. Learning and Motivation, 1972,3, 304-312. D'Amato, M. R., & Worsham, R. W. Retrieval cues and short-term memory in capuchin monkeys. Journal of Comparative and Physiological Psychology, 1974, 86, 274-282. Gleitman, H. Forgetting of long-term memories in animals. In W. K. Honig and P. H. R. James (Eds.), Animal memory. New York: Academic Press, 1971, pp. 1-44. Grant, D. S. Proactive inhibition in pigeon short-term memory. Journal of Experimental Psychology: Animal Behavior Processes, 1975,104, 207-220. Grant, D. S. Effect of sample presentation time on long-delay matching in the pigeon. Learning and Motivation, 1976, I, 580-590. Grice, G. R. The relation of secondary reinforcement to delayed reward in visual discrimination learning. Journal of Experimental Psychology, 1948,38, 1-18. Herman, L. M. Interference and auditory short-term memory in the bottlenose dolphin. Animal Learning and Behavior, 1975, 3, 43-48. Herman, L. M., & Gordon, J. A. Auditory delayed matching in the bottlenose dolphin. Journal of the Experimental Analysis of Behavior, 1974,21, 19-26. Hoffman, H. S., Fleshier, M., & Jensen, P. Stimulus aspects of aversive controls: The retention of conditioned suppression. Journal of the Experimental Analysis of Behavior, 1963, 6, 575-583. Honig, W. K., & Beale, I. L. Stimulus duration as a measure of stimulus generalization. Journal of the Experimental Analysis of Behavior, 1976,25, 209-217. Honig, W. K., Beale, I. L., Seraganian, P., Lander, D. G., & Muir, D. Stimulus and response reduction: Two aspects of inhibitory control in learning. In R. A. Boakes and M. S. Halliday (Eds.), Inhibition and learning. New York: Academic Press, 1972, pp. 41-71. Honig, W. K., & Lindsay, H. Transfer of the control of stimulus duration across discrimination problems. Learning and Motivation, 1975, 6, 157-178. Jarrard, L. E., & Moise, S. L. Short-term memory in the monkey. In L. E. Jarrard (Ed.), Cognitive processes of nonhuman primates. New York: Academic Press, 1971, pp. 3-24. Kamin, L. J. Temporal and intensity characteristics of the conditioned stimulus. In W. F. Prokasy (Ed.), Classical conditioning: A symposium. New York: Appleton-CenturyCroft~ 1965,pp. 118-147.

248

WERNER K. HONIG

Lavin, M. S. The establishment of flavor-flavor associations using a sensory preconditioning training procedure. Learning and Motivation, 1976, 7, 173-183. Lett, B. T. Delayed reward learning: Disproof of the traditional theory. Learning and Moti· vation, 1973,4, 237-246. Lett, B. T. Long delay learning in the T-maze. Learning and Motivation, 1975, 6, 80-90. Lett, B. T. Long delay learning: Implications for learning and memory theory. In N. S. Sutherland (Ed.), Tutorial essays in experiment!ll psychology (Vol. 2). Hillsdale, New Jersey: Lawrence Erlbaum Associates, in press. Leyland, C. M., & Honig, W. K. Maintenance of behavior controlling the duration of discriminative stimuli. Journal of the Experimental Analysis of Behavior, 1975,24, 207-214. Olton, D. S., & Samuelson, R. J. Remembrance of places passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes, 1976, 2, 97-116. Revusky, S. The role of interference in association over a delay. In W. K. Honig and P. H. R. James (Eds.),Animal memory. New York: Academic Press, 1971, pp. 155-213. Revusky, S. H., & Garcia, J. Learned associations over long delays. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 4). New York: Academic Press, 1970, pp. 1-84. Roberts, W. A. Short-term memory int he pigeon: Effects of repetition and spacing. Journal of Experimental Psychology, 1972,94, 74-83. Roberts, W. A. Spaced repetition facilitates short-term retention in the rat. Journal of Comparative and Physiological Psychology, 1974,86, 164-171. · Roberts, W. A., & Grant, D. S. Some studies of short-term memory in the pigeon with presentation time precisely controlled. Learning and Motivation, 1974, 5, 393-408. Roberts, W. A., & Grant, D. S. Studies of short-term memory in the pigeon using the delayed matching-to-sample procedure. In D. L. Medin, W. A. Roberts, and R. T. Davis (Eds.), Processes of animal memory. Hillsdale, New Jersey: Lawrence Erlbaum Associates, 1976, pp. 79-112. Spear, N. E. Forgetting as a retrieval failure. In W. K. Honig and P. H. R. James (Eds.), Animal memory. New York: Academic Press, 1971, pp. 45-109. Worsham, R. W. Temporal discrimination factors in the delayed matching-to-sample task in monkeys. Animal Learning and Behavior, 1975, 3, 93-97. Worsham, R. W., & D'Amato, M. R. Ambient light, white noise, and monkey vocalization as sources of interference in visual short-term memory of monkeys. Journal of Experimental Psychology, 1973,99, 99-105.

Selective Attention and Related Cognitive Processes in Pigeons

Donald A. Riley H. L. Roitblat University of California, Berkeley

I. SELECTIVE ATTENTION AND RELATED COGNITIVE PROCESSES IN PIGEONS

It is a common observation that when we notice, think about, or observe some

stimulus, it may be difficult to notice, think about, or observe other stimuli. In other words, perception and other cognitive processes are often selective. One explanation for these observations is that in order to notice some stimulus, we must attend to it and we can only attend to a limited number of stimuli at a time. Often, other explanations are available. For example, it may be difficult to notice two things because they are spatially separated, and we can only look in one direction at a time. Such peripheral explanations are usually suggested as alternatives to an attentional explanation. Attention can, then, be viewed as a central adaptation to information overload. There is more information present in the stimulus array than is possible to process so it is necessary to selectively attend to some stimuli at the expense of others. Selective attention has also been proposed as an explanation for the performance of animals in discrimination problems (e.g., Sutherland and Mackintosh, 1971; Riley and Leith, 1976). Every time a selective attention interpretation has been proposed by some investigator, however, other experimenters have been quick to demonstrate other more peripheral and perhaps simpler mechanisms to account for the apparent attentional effects. After almost 50 years it is still a matter of controversy whether an attentional hypothesis is at all necessary to explain animal discrimination performance (Mackintosh, 1975). This lack of resolution may be true, in part, because experimenters attempting to investigate attention in animals have usually failed to structure their experi· 249

250

DONALD A. RILEY AND H. L. ROITBLAT

ments to provide any overload of the animal's ability to take in all the available information within the allowed time. It is our view that when this condition is met, that is, when the animal's information processing ability or capacity is taxed, the likelihood of demonstrating selective attention will increase. The modern study of attention in humans, beginning with Broadbent's book Perception and Communication (1958) has been sensitive to this position. Broadbent (1961) also suggested that such a strategy would be useful in studying animal attention. This chapter presents research that has used two different methods to vary the load placed on the animal's information processing system: Variations in the time available to the animal in which to process the information and variations in the number of stimuli which the animal must process. By manipulating the available time and the number of messages, we can investigate the processes controlling the animal's ability to attend to more than one input at the same time. We have tried to ensure that we are actually taxing the animal's ability by shortening the available time or increasing the number of messages until errors begin to increase. In the following pages we examine the selective-attention hypothesis and several competing interpretations in the context of the matching-to-sample paradigm in which the number of messages presented and the sample duration are manipulated. An experiment conducted by Maki and Leith (1973) demonstrates the effect of both variations. The procedure is illustrated in Fig. 1.

Warning Signal

-I

~ m

Sample

~ Test

IE] RED •

BLUE

FIG.1. Two typical matching-to-sample sequences as used by Maki and Leith (1973). An element trial is depicted on the left; a compound trial, on the right. Warning signals and samples appear on the center key; the two test stimuli appear on the side keys.

9. ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

251

First, a white warning signal was projected onto the center key of a threekey pigeon-intelligence panel. In some experiments this warning signal was terminated by a single peck, in others it remained on for a fixed 5 sec. In either case, a single peck to the warning-signal key was necessary for the presentation of the rest of the trial. Failure to peck resulted in the initiation of a dark intertrial interval and the re-presentation of the entire trial beginning with the warning signal. If the warning-signal key was pecked, it was followed, after a 100-msec delay, by a sample stimulus whose duration was manipulated as a parameter of interest. Sample duration was independent of the number of pecks to the key on which the sample was projected, and none was required. The sample was then followed by the presentation on the side keys of two test stimuli, one of which matched the sample and was designated correct. A single peck to one of the side keys within 5 sec was necessary to indicate the subject's choice. A random half of the trials on which the subject made a correct response were reinforced with access to a lighted food hopper. Incorrect choices caused the test lights to remain on for 5 sec during which pecks had no effect. Following the termination of the reinforcer presentation or the test stimuli on those trials on which an incorrect choice or no choice was made, a 5-sec dark intertrial interval occurred. Failure to make a choice resulted in the presentation of an intertrial interval and the re-presentation of the same trial. As Fig. 1 indicates, the animal was confronted with an element sample on half the trials and a compound sample on the other half. In these experiments, an element was either a solid colored red or blue disk, or three white lines, either horizontal or vertical on a black surround. A compound sample consisted of the pairing of a color element with a line element, and appeared as three white lines on a colored surround. The test stimuli were always two value~ from the same dimension: The animal was never asked to choose between a color and a line, nor were compound tests presented in this first experiment. On those trials during which an element sample was presented, the animal had only the information from that single element to process and store for use. On compound trials, however, it had to process the elements from both dimensions because it could be tested on either; line and color tests occurred equally often and in random order. Because of this procedure, the information load on compound trials was greater than on element trials. A summary of Maki and Leith's results appears in the upper two curves of Fig. 2. Several effects are worth noting. First, element performance was consistently superior to compound performance at all sample durations. This was characteristic of both birds and both dimensions. The second important feature is that the birds appear to process information about both dimensions on every trial, even with very short sample durations. The lowest curve in Fig. 2 is the performance that would occur if, on compound trials, the bird attended to only one dimension per trial, choosing each dimension equally often on a random basis. This simple model assumes that the bird is a perfect selective attender and performs as well on the attended dimension as it does on element

252

DONALD A. RILEY AND H. L. ROITBLAT

100 1-

ELEMENT _.__...COMPOUND

90

u

~ 80

a:

8 12

70

LIJ

u a: LIJ

0..

60

50 -.;t

0

0

OON-.Q-.;fN 0--NM

00000

0 0 Ll")

SAMPLE DURATION (SEC) FIG. 2. Matching-to-sample performance for two birds. The open and closed circles represent respectively performance on element and on compound trials. The diamonds represent performance that would result from a hypothetical bird attending to only one element per compound trial. The resulting curve is a combined average of element performance (perfectly attending to the to-be-tested dimension) and chance (perfectly ignoring the to-be-tested dimension), (Data taken from Maki & Leith, 1973.)

trials. Because it attends to one dimension and ignores the other, if the neglected dimension is tested, performance is dictated by chance. Thus, according to this model, average performance following compound trials will be halfway between chance and element performance. This is true because on half the trials the animal will select the correct dimension, be tested for it, and show element trial performance; while on the other half of the trials it will select the dimension that is not tested and show chance performance. As can be seen in the figure, the model does not come close to mirroring performance following compound samples. We conclude, therefore, that the experimental birds do not behave as predicted by the model and instead appear to attend to both dimensions on most or all trials, sharing their attention between the two dimensions. The third important feature of Maki and Leith's results is that matching-tosample performance improves as a function of sample duration over the range of sample durations indicated here and, indeed, may continue to improve even with sample durations longer than 5 sec. These facts are not peculiar to research from this laboratory. Roberts and Grant (1974) have found such improvements following an 8-sec sample as opposed to a 4-sec sample, and Farthing and Opuda (1973) have found improvements as a function of the number of times the bird

9.

ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

253

pecks at the sample. As Roberts and Grant (1974) observed, these facts differ . from similar investigations with monkeys. D' Amato and Worsham (1972) found virtually perfect matching-to-sample in monkeys with 40-msec exposure time, and we have made similar observations with humans using our pigeon apparatus. In this chapter, we shall inquire into the reasons for both the element-compound difference and the change in performance with sample duration.

II. THE ANALYSIS OF ELEMENT -COMPOUND DIFFERENCES

The most obvious interpretation of the superiority of element over compound performance is that the bird divides attention and thus capacity between the two elements of the compound during sample processing. This divided-attention hypothesis states that when capacity is not adequate to process all information in the brief time available, and when this information is more or less equally divided between two or more incoming messages, each message will suffer relative to performance with a single message. Such a general hypothesis does not state where in the processing system the loss occurs, but it does distinguish between an information processing loss and losses due to other factors. Maki and Leith (1973) and Maki, Riley, and Leith (1976) have considered two additional classes of explanations to account for element-compound differences that are alternatives to the divided-attention hypothesis. One of these, a generalization decrement hypothesis, states that performance following a compound sample is inferior to performance following an element sample because compound samples are always followed by element tests. Consequently, no correct stimulus precisely matches the compound sample. Following an element sample, on the other hand, one of the test stimuli is an identical match. The other interpretation attributes element-compound differences to stimulus degradation during sample presentation. White lines on a colored field may desaturate the color, making color differences less distinct while the bright colors surrounding the white lines make the lines harder to detect than do black surrounds. This interpretation has also been suggested by Zuckerman (1973) to account for analogous effects in a maintained discrimination. A. Generalization Decrement

Maki et al. (1976) conducted three tests of the generalization decrement hypothesis and found no support for it. In the first two experiments, element and compound samples were both followed by compound test stimuli that differed from each other on the relevant dimension but were the same on the other dimension. In Experiment 1, the irrelevant dimension was the same as one dimension of the compound sample. For example, if the sample was blue and horizontal, one test stimulus might have been blue and horizontal and be the

254

DONALD A. RILEY AND H. L. ROITBLAT

correct choice while the other was blue and vertical. In Experiment 2, the irrelevant dimension in the two test stimuli had the value that did not appear in the sample. Hence, the blue horizontal sample mentioned above might be followed by a red horizontal and a red vertical test with the former being the correct choice. In both these experiments, the superiority of element relative to compound performance remained undisturbed. In Experiment 1, the correct test choice exactly matched the compound, but never matched the element sample. Thus the experimental arrangements were reversed from those used in Maki and Leith's (1973) experiment. Consequently, according to the generalization decrement hypothesis, matching of the compound sample in this new experiment should have been superior to matching of the element. Nevertheless, no change in the superiority of the element matching over compound matching occurred. In Experiment 2, neither test stimulus exactly matched either the element or the compound sample. Again, however, the superiority of element over compound remained the same. These facts argue strongly against a generalization decrement hypothesis but are consistent with an information overload hypothesis. It is possible to argue, however, that in both experiments the irrelevant dimension interfered with matching performance. In Experiment 1, both tests were at least partially correct in that both keys matched the sample on the irrelevant dimension. In Experiment 2, neither test was completely correct in that both keys mismatched the sample on the irrelevant dimension. Both of these factors may have affected the subject's performance following compou~q samples. B. Response Competition and Redundant Relevant Cues

For this reason, Experiment 3 was conducted. In the condition of major interest, all compound sample presentations were followed by test presentations in which one key matched the sample on both dimensions and the other test key differed from the sample on both dimensions. In this way, both dimensions were redundant and relevant in that they both signaled the correct choice. While the results bear on the generalization decrement hypothesis, the issues raised by them are more far reaching and create new questions for the analysis of element-compound differences. Consequently, we will discuss this experiment in some detail. The experiment compared processing of element and compound samples under two sets of test conditions. For one set of conditions element test stimuli always followed both element (EE) and compound (CE) samples (the first letter refers to the sample, the second to the test). These conditions are identical to the earlier conditions (Maki & Leith, 1973) in which EE trials produced consistently higher performance than CE trials (Fig. 2). Days with such element tests were alternated with days on which compound tests followed element (EC) and compound (CC) samples. In the compound-test conditions, when a sample was an element, one of the compound-test stimuli was the same as the sample on one dimension and irrelevant on the other; the

9. ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

255

other compound-test stimulus differed from the first test stimulus on both dimensions. For example, if the sample was a red key, the positive test stimulus might be red and vertical, the negative blue and horizontal. Finally, when a compound sample was presented, the compound-test condition (CC) carried information on both dimensions and so contained one test stimulus that was identical to the sample stimulus and another test stimulus that differed from the sample on both dimensions. For example, a red and vertical sample would be followed by a choice between a red and vertical correct stimulus and a blue and horizontal incorrect stimulus. Thus, with such a sample-test pair, the generalization decrement hypothesis would predict that performance with compound samples and compound tests would be identical to performance obtained with element samples and element tests. In both cases, there is a test stimulus identical to the sample stimulus and a test stimulus completely different from the sample stimulus. Since the values on both dimensions of the compound sample agree with the values on both dimensions of the correct compound test, the animal could base his decision on either or both elements and be correct. Consequently, there should be no generalization decrement and no possible response competition as in Experiments 1 and 2. One would expect, according to either of these alterna-

100 1u w

a: a:

90

0

u

80

1-

z

w

u

70

a: w

CL

SAMPLE

DURATION

(MSEC)

FIG. 3. Matching-to-sample performance for two birds under four different combinations of sample-test conditions. Conditions EE and CE refer to element and compound samples with an element test condition. Conditions EC and CC employed the same sample conditions but the tests were always compound tests. Following both element and compound samples, the test stimuli differed from one another on both dimensions. Thus, following an element sample, one of the test keys had a stimulus value which matched the sample on only one dimension. Following a compound sample, one of the test keys matched the sample· on both dimensions, the other on neither dimension. (Data taken from Maki et al., 1976.)

256

DONALD A. RILEY AND H. L. ROITBLAT

tives to the information overload hypothesis, that performance following compound samples and compound tests would be identical to performance following element samples and element tests. The results of this experiment, shown in Fig. 3, do not support these predictions. The presence of redundant relevant test pairs did not facilitate performance. In fact, CC trials produced the same level of performance as that produced by CE trials which were both lower than the performance obtained on EE trials. Performance following a compound sample was the same whether the test stimulus was an element or a compound. Thus, neither response competition nor generalization decrement are necessary to account for the obtained difference between performance with element and with compound samples. Response competition does, however, seem to figure into the performance of the other novel condition in this experiment: The EC condition in which compound tests followed element tests. Apparently, in this condition, the birds' performance was partially controlled by the irrelevant dimension. The type of competition found in this condition, however, could not account for the element-compound difference of interest. With compound samples and element tests, there is only the single relevant dimension present on the test keys and thus no competition could occur. Recall that our impetus for investigating these phenomena was an attempt to understand selective attention in animals in situations in which we were reasonably satisfied that the animals were operating under conditions of information overload. Do the data of the experiments by Maki et al. (1976) come from situations in which the animal is overloaded? Performance in the compound sample conditions of Experiment 3 in which there was no response competition was superior to performance in the compound conditions of Experiment 1 or 2, in which there was both information overload and response competition. Despite the elimination of response competition, however, performance on the compound sample condition CE and CC remained below the performance of the element sample condition in which there was no competition. We feel confident, therefore, in concluding that the superiority of matching performance with an element sample over compound sample is due to the effect of information overload. As persuasive as we find this argument, it does raise the problem of why the birds in Maki et al.'s (1976) CC condition did not utilize the redundancy of compound samples and tests to reduce the information load. C. Redundant Relevant Cue Facilitation

Because the same values of both dimensions were present in the sample and in the correct test choice, but not in the incorrect test choice, the animal could behave perfectly accurately by attending to only one of the available dimensions and not spending time partially encoding the second. If the animal were to

9.

ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

257

behave in this manner both during sample presentation and during the test, CC performance should equal EE performance. Because the animals in this experiment performed equally in the CC and CE conditions, it follows that there was no reduction in information load. This is a curious fmding. Other experiments have not only found that discriminative performance is as good with compound stimuli as with element stimuli, but actually found that performance improves with compound stimuli above that obtained with element. That is, redundant relevant cues can facilitate discriminative performance. Zuckerman (1973) found such facilitation when he compared discriminative performance with stimulus elements (hues and line tilts) with performance with redundant relevant compounds of the same stimuli. In the first phase of the experiment, Zuckerman trained the birds to make one response if any of four shorter wavelength hues (shorter than 580 nm) or any of the four smaller angled line tilts (less than 22.5" counterclockwise rotation from vertical) was presented, and to make another response if any of four longer wavelength hues (greater than 580 nm) or any of four larger angled (greater than 22.5") line tilt stimuli appeared. By testing each dimension separately, he obtained psychometric functions for element discriminative performance. In a second phase, Zuckerman then presented eight unique pairs of stimuli in which the shortest wavelength hue was always presented in compound with the smallest angle line tilt, the second hue with the second line tilt, etc. Training with these compound stimuli resulted in an improvement in the discriminability of the redundant relevant compounds over the discriminative performance obtained with each dimension individually. In a third phase, he then presented all possible pairs of hue and line tilt. This means that in addition to the eight pairs of stimuli presented during the second phase, there were also 56 other possible pairs. The eight pairs presented in both Phase II and Phase III were equally redundant, equally relevant, and equally often reinforced in both phases, but the facilitative effect of pairing them disappeared in Phase Ill. It is not immediately obvious why redundant relevant cue facilitation should depend on the context in which the pairs are presented, but it seems to depend on just that. When the stimuli were consistently paired, facilitation occurred; when the same pairs were presented among all possible pairs, facilitation disappeared. In the experiment presented by Maki et al. (1976) all possible pairs of stimuli were also presented, and in that experiment, too, no facilitative effect was found. Animals appear to respond differently depending on how the cue pairs are formed. This notion is reminiscent of the work on configura! conditioning in which prolonged training with compound stimuli can result in a difference between the signal strength of the compound relative to the signal strengths of its components (Thomas, Berman, Serednesky, and Lyons, 1968; Baker, 1968; Booth and Hammond, 1971; Rescorla, 1972). Perhaps under the consistent pair-

258

DONALD A. RILEY AND H. L. ROITBLAT

ings situation used by Zuckerman (1973), and perhaps under the conditions used to produce configuring (these may or may not reflect the same phenomena), the animal comes to treat the compound as a new stimulus which is other than a simple combination of its components. Ill. STIMULUS COMPOUNDS AND INFORMATION PROCESSING

Another impetus for the view that different kinds of stimulus compounds are processed in different ways comes primarily from the work of Garner (1970, 1974) using huma~ subjects. He and his associates have presented evidence that the nature of the stimulus compound can affect one's ability to process information about it. In particular they have drawn two sets of distinctions that may be of value in the analysis of these two experiments and animal information processing in general. Garner (1970) distinguishes between two types of compounds, separable and integral, and between two sources of difficulty in a discrimination, state limitations and process limitations. Integral stimulus compounds are those in which the elimination of one dimension in a bidimensional compound necessarily results in the elimination of the other dimension. An archetypical example is a set of cues to be discriminated on the basis of hue and saturation. Elimination of hue necessarily eliminates saturation and vice versa. Separable compounds involve independent dimensions in which the elimination of one dimension does not need to effect the other dimension. An example is a discrimination involving a colored light and a tone. Discriminability can be limited in either of two ways. A process limitation refers to a discrimination being difficult because the stimuli to be discriminated are so close together that the subject cannot tell, them apart very well. A state limitation refers to a discrimination in which the difficulty is caused by the stimuli each being hard to detect alone (e.g., because the illumination is too low or the presentation too brief). If energy were to be added to them, however, the differences between them would be sufficiently large to result in rather good performance. Garner and his associates (e.g., Garner, 1974) have presented evidence regarding the different ways in which these two distinctions interact. Separable stimulus compounds facilitate performance when discriminability is limited by a state limitation but not when discriminability is limtied by a process limitation. Integral compounds, on the other hand, facilitate process-limited discriminations but not state-limited discriminations. Garner and his associates have also discovered something interesting about the way in which humans rate the similarity and difference between separable and integral stimuli. Scaled differences between integral stimulus compounds obey a Euclidean metric in which the differences along each dimension form the legs of a right triangle and the judged difference corresponds to the hypotenuse of the triangle. Scaled differences between separable compounds, on the other hand,

9.

ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

259

obey a city block metric in which the differences on the two dimensions are added. A. Garner's Analysis Extended to Animal Research

Garner's analysis seems directly applicable to the understanding of the issues raised in the preceding section. Recall that Zuckerman (1973) found redundant relevant cue facilitation in his maintained discrimination task but that Maki et al. (1976) failed to find such facilitation. Zuckerman's experiment is an example of process limited discrimination in that the performance was limited by the similarity of the stimuli to one another. Only the values farthest from the criteria could be discriminated with anything near 100% accuracy. Consequently, the fact that the compound cues facilitated discrimination when redundant and perfectly correlated is consistent with the suggestion that this treatment produced integrality of the compounds through the experience of regular and consistent pairing. The abandonment of the regular pairing procedure in the third phase may have reduced the integrality of the cues so that the redundant relevant cues no longer resulted in facilitation of performance. Chase and Heinemann (1972) also ran pigeons on a bidimensional discrimination task similar to that run by Zuckerman. When the stimuli were redundant, relevant, and consistently paired, the pigeon's choice curves fitted a model consistent with Garner's analysis of integral stimulus compounds. That is, the discriminability of each dimension alone formed the leg of a right triangle, the hypotenuse of which corresponded to the discriminability of the compound. Further research is necessary to determine whether other features of an integral stimulus also occur as a result of repeated training with consistent correlated pairs. For example, Garner (1970) reported that for humans, redundant relevant integral compounds facilitated discrimination when the difficulty of discrimination was due to a process limitation (as in Zuckerman's experiment), but not when the difficulty was due to a state limitation. At the present time, there are no animal data relating state limitation to integrality and separability. Consider now the redundant relevant cue experiment by Maki et al. (1976, Experiment 3). There are several differences between their Experiment 3 and Zuckerman's Phase II. Zuckerman used a maintained discrimination in which the difficulty was produced by a process limitation - the stimuli were too similar for perfect discrimination by the subject. Maki et al. (1976), on the other hand, used a matching-to-sample procedure in which the difficulty was produced by a state limitation - the samples were presented too briefly to allow perfect performance. Garner found that redundant relevant integral cues do not facilitate discrimination when the difficulty of discrimination is produced by a state limitation. One would expect, therefore, that even if the birds used by Maki et al. did manage to integrate the compound stimuli presented in Experiment 3,

260

DONALD A. RILEY AND H. L. ROITBLAT

they still would not have shown facilitation because their performance was restricted by a state limitation. We find this to be an unlikely explanation, however. It does not seem reasonable that the matching-to-sample birds formed integral stimulus compounds in the same situations in which it is necessary to assume that Zuckerman's birds failed to preserve integrality, that is in his Phase III. Furthermore, if the birds did treat the samples in a manner similar to that in which humans treat integral compounds, one would expect compound performance to be facilitated to the level of element performance. Humans treat integral compounds as if the compounds varied along a single dimension different from either of the physical dimensions in the compound. If pigeons behaved this way, the memory load would therefore be reduced from two dimensions to one single new dimension: a reduction which should result in the bird treating the compound as a single element. The data from Maki et al. clearly do not support this prediction either, so it is unlikely that they did treat the compounds as integral. Finally, there is another possible explanation for the lack of facilitation observed in Maki et al.'s third experiment: Matching-to-sample, as opposed to maintained discrimination, requires at least two stages. First, when the sample is presented it must be detected and encoded for later use because it will not be present at the time of choice. Second, a recognition phase occurs during which the subjects are required to match at least one test stimulus against the representation of the sample in working memory (cf. Honig, this volume, Chapter 8) in order to select the correct choice. The two dimensions presented in the sample do not become redundant until the test is presented. Indeed, as we described earlier, on alternate days the two dimensions being followed by element tests were not redundant at all. It may be the case, then, either that the birds in the experiment of Maki et al. never learned to use the redundancy when it was present or that it was not useful at the necessary point in the matching sequence. To summarize, there is evidence suggesting that pigeons can treat ostensibly separable stimuli as integral. This analysis is consistent with the finding of redundant relevant cue facilitation in the second phase of Zuckerman's experiment, and with the shape of the postdiscrimination gradients Chase and Heinemann (1972) found after redundant relevant compound training. No such effect appeared in the third phase of Zuckerman's experiment or in Experiment 3 of Maki et al. These findings suggest that pigeons may integrate stimuli when two dimensions are consistently and uniquely paired. Such consistent pairings appear to produce facilitation while nonconsistent pairings appear not to produce it. Finally, we have suggested that configura! conditioning may reflect the perceptual integration of otherwise separable stimulus compounds. B. Selective Attention

One consequence of the concept of limited capacity is that an increase in processing of one set of stimuli reduces the capacity for processing another set. If then, an animal could be trained or instructed to attend to one of the dimen-

9.

ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

261

sions present in a compound task in which its capacity is exceeded, the animal ought to do better on that attended dimension, and at the same time, ought to do worse on the neglected dimension. Presumably, the improvement on the attended dimension ought to be roughly equal to the decrement on the neglected dimension. This has been called the inverse hypothesis (Thomas, 1970). Employing pigeons in a maintained discrimination task, Blough (1969) found what appear to be selective attention effects which correspond to this descrip· tion. He trained pigeons to perform in a bidimensional discrimination involving seven similar hues and seven similar auditory frequencies. Of the 49 possible pairs, one pair consisting of an extreme value from each set was designated S+. All other pairs were designated S-. When the task was changed so that discrim· inations were required on only one dimension (the other dimension being held constant on S+), performance on the varying dimension improved. When tested again with both dimensions varying, Blough found a deterioration on the dimension which had been represented by a single value that corresponded to the improvement on the dimension that had been varying. With further bidimensional training, this difference disappeared and performance returned to its baseline level. As has been pointed out elsewhere (e.g., Leith & Maki, 1975), it is possible to interpret the loss of control on the neglected dimension with mechanisms other than attention: That is, one can apply the laws that govern the strength of association in other situations and still produce a loss of control. The corresponding improvement on the varying dimension and its subsequent loss with further bidimensional training is more difficult to explain however. Blough was careful to ensure that the density of reinforcement to S+ on the varying dimension was the same in all three conditions. It therefore seems unlikely that simple changes in habit strength could account for the obtained improvement and subsequent decrement in performance on the varying dimension. Furthermore, the discriminative stimuli were presented for only a brief time - 1.5 sec, during which the bird had to perform a difficult discrimination. The combination of the time pressure from brief stimuli (a state limitation) and the difficulty of the discrimination (a process limitation) probably produced a large load on the bird's system, hence the demonstration of selective attentional effects. C. Stimulus Degradation

Earlier, in discussing the work of Maki and Leith (1973), we suggested an alternative to the generalization decrement and information load hypotheses. This stimulus degradation hypothesis suggested that element-compound differences found by Maki and Leith were due to a degradation of the quality of each dimension in a compound because of reduced contrast. This hypothesis does not readily apply to Blough's experiment, but for a task with visual stimuli such as Maki and Leith's, it offers an alternative to the selective attention hypothesis. The stimulus degradation hypothesis would be ruled out if matching-to-sample performance were to improve on the attended dimension while deteriorating on the neglected dimension for such a change would demonstrate that the element-

262

DONALD A. RILEY AND H. L. ROITBLAT

compound difference does not reflect poorer quality stimuli, but divided attention effects. A strong stimulus degradation hypothesis would require that selective attention training have no effect on the element-compound difference. On the other hand, a failure to fmd a selective-attention effect could reflect either less than perfect selective attention or stimulus degradation. Leith and Maki (1975) conducted an experiment analogous to Blough's but using the matching-to-sample task rather than a maintained discrimination. The task and the stimulus displays were basically the same as those already described in Maki and Leith's (1973) study. Two birds were first tested in the divided attention condition already described, and performance on both element sample trials and compound sample trials was measured. Next, prolonged training occurred during which element samples from only one dimension were used along with compound samples containing both dimensions. First, one dimension was tested for a month, then the other dimension was tested for a month with intervening periods of bidimensional retraining. During the single dimension training condition compound performance improved relative to element performance. That is, performance on compound trials approached but did not equal performance on element trials. This reduction in the element-compound difference for the selected dimension found during single element training was roughly matched by an increase in elementcompound difference for the neglected dimension upon the return to the two dimension condition. These results show a direct correspondence to Blough's (1969) data. Training intended to produce selective attention to line, caused line performance to improve and color performance to suffer. Training intended to produce selective attention to color, caused color performance to improve, and performance on line to suffer. Because these data are consistent with a selective attention interpretation, they are inconsistent with a strong stimulus degradation hypothesis. Although the failure to completely eliminate the element-compound difference may reflect either stimulus degradation or incomplete selective attention, stimulus degradation is not sufficient to account for element-compound differences. The data presented by Leith and Maki are of particular interest in that they appear to offer direct support for the inverse hypothesis. Following compound samples, performance on the attended dimension improved and, performance on the neglected dimension deteriorated. Such deterioration is not required by the constraints of the experiment itself; in principle, there is no reason why the animal could not improve in its performance to one set of cues while remaining at the same level of performance on the neglected set of cues. D. Trial-to-Trial Selective Attention Instructions

The demonstration of an inverse relationship between processing a selected input and processing the other inputs is one aspect of the demonstration of selective attentional effects. Usually, attention is also seen as being flexible

9.

ATTENTION AND COGNITIVE PROCESSES IN PIGEONS

263

(Kahneman, 1973). It would strengthen the argument that the preceding experiments demonstrated selective attentional effects if it were possible to demonstrate flexibility in the selective performance of pigeons. In an experiment very similar to that run by Leith and Maki, Leuin (1976) investigated the effect of cueing the to-be-tested dimension on matching-tosample performance. On every trial, sample presentations on the center key were preceded by one of three stimuli also on the center key. On half of the compound trials on which the bird would be tested for line orientation, the sample was preceded by a white cross on a dark surround. On half of the compound trials on which the bird would be tested for hue, the sample was preceded by a blue and a red spot, one above the other. On the other half of the compound trials with either dimension, the sample was preceded by a plain white disk signalling that either dimension could be tested. These presample stimuli were called line cues, color cues, and warning signals, respectively. They were each presented for a fixed 2.5 sec duration, and a single peck was necessary to initiate LINE

COLOR 100

~a:

90

a:

0

(.)

!z

w

R

.

..

80

ffia.

Cued Element Cued Compound Non-Cued Compeund Probes

25 26 28

70

!:_,;._,..._!- - ' ' - - - ' - - . . . . . . _

ln

I

0 ,

Q)

.4

(j)

c

ro

0..

(j)

.2

Q)

0:::

...0

0,._

o_

20

0

+-'

...0

30

(j)

_j

:

40

0

0

4

8

12

Time (sec)

16

10

15

30

Time (sec)

FIG. 1. Temporal-discrimination data from an estimation method (left panel) and a production method (right. panel). The psychometric function in the left panel shows the probability of a "long" response as a function of time; the response gradient in the right panel shows response rate as a function of time.

trained to press one lever following a 4-sec signal and the other lever following a 16-sec signal. This figure shows the mean probability (relative frequency) of a long response for the two extremes and for various intermediate values that were not differentially reinforced during ten 50-min sessions. The probability that the rat estimated a stimulus to be longer than a criterion increased as the signal duration increased from 4 to 16 sec.

4. An Example of a Production Procedure. In a production experiment for human subjects, a person might be given instructions to respond when the time interval appears to equal a criterion. The instructions could indicate that premature responses were of little importance, but that the person should try to respond as soon as the interval was complete. Knowledge of results (i.e., whether or not the produced time interval is shorter or longer than the criterion) is given following each response. An equivalent experiment for rats could be as follows: On each trial a single lever is inserted into the box, accompanied by a distinctive signal. Responses during the first 30-sec of the signal have no effect, but the first response after 30-sec results in the delivery of a food pellet, the termination of the signal, and the withdrawal of the lever. After a variable intertrial interval, another trial is begun. This, of course, is simply a discrete-trial fixed-interval schedule. As in the case of the estimation procedure, knowledge of results consists of a pellet of food, and the animal learns the instructions through experience.

282

RUSSELL M. CHURCH

Typical results from such a procedure is shown in Fig. 1 (right panel). The response rate of one rat averaged over five 3-hr sessions increased as the 30-sec criterion approached, i.e., the probability that the rat judged the interval to be complete increased as the time approached the criterion. II. EXPLANATIONS OF TEMPORAL DISCRIMINATION

Studies of temporal discrimination provide ample evidence that there is a relationship between time and behavior. In the estimation procedures, different time intervals lead to qualitatively different responses, e.g., a response on a left or right lever. In the production procedures, different time intervals lead to quantitatively different responses, e.g., more or less rapid responses or responses differing in latency. In this section, I will review some of the alternative explanations for such temporal discriminations. Some of them attempt to account for the behavior without resort to the concept of time as a stimulus. Some of them consider time to be an adequate stimulus for differential behavior, but do not involve the concept of an internal clock. Still others explain temporal discrimination in terms of an internal clock. A. Explanations Not Involving Time as a Stimulus

One of the major objectives of animal psychologists, especially during the last half century, has been to account for the behavior of animals without violating Lloyd Morgan's canon of parsimony (Morgan, 1894). A second major objective has been to account for behavior of animals without reference to unobserved cognitive or physiological processes (Skinner, 1950). With these objectives, it is easy to see why where would be a serious search for explanations of response gradients that did not involve time discrimination. The ability to keep track of time should not be assumed if a more parsimonious account is available (by Lloyd Morgan's canon). And time, and especially an internal clock, lacks the physical attributes of the standard stimulus and leads to explanations in terms of processes which are not directly observable. For these reasons there has been a search for explanations of response gradients that do not involve any time discrimination. For example, some behavior that appears to provide evidence for a time discrimination may be based upon reactive responding alone (Church & Getty, 1972). If an animal reacts to the delivery of food by stopping to eat and then responding again at a constantly accelerating rate, the resulting cumulative curve could be characterized as an "FI scallop." Although such a reaction is undoubtedly involved in some response gradients seen under fixed interval schedules (Killeen, 1975), it obviously does not explain all cases. Most simply, it cannot account for the different gradients seen with different fixed intervals between reinforcements, nor the evidence for the discriminative stimulus status of reinforcement (Staddon, 1972).

10.

THE INTERNAL CLOCK

283

Another explanation for response gradients involves mediating behavior. The general notion behind the mediating behavior explanation of fixed interval behavior is that the animal does not attend to time per se, (i.e., it does not engage in mental timing), but does attend to its own behavior which is correlated with time. Such mediating behavior might be easy to observe - for example, the animal might slowly turn in a circle and respond on the basis of where it is pointing at the end of the interval. Even if an animal used such mediating behavior, this fact would not be evidence against time discrimination. Instead, it would be a case in which it was possible to identify the hands of the clock. We have never observed this kind of mediating behavior during a time interval. We have, however, observed animals engaging in repetitive actions, e.g., biting at the front panel, mouthing both paws, rapid head movements in and out of the food cup. Typically during the signal in a time-estimation procedure, a rat engages in some stereotyped rhythmic behavior which is different from its behavior in the intertrial interval (Church, Getty, & Lerner, 1976). Perhaps this is related to an oscillator that causes a clock to advance. Alternatively, the rhythmic behavior might have no relation to the timing performance. Direct evidence against the mediating response position has been reported by Dews (1962) in a series of articles beginning in 1962. In these "interruption" experiments, a discriminative stimulus was turned on and off in successive segments of a fixed interval. The response rate was much greater in the presence of the discriminative stimulus than in its absence and response rate increased in successive segments of the fixed interval in which the discriminative stimulus was present. Dews (1970, p. 46) concluded that there was "a progressive increase in tendency to respond throughout the interval even when no responding is actually occurring." The interruption experiments have demonstrated that an animal can continue to time an interval independently of its actual responding. Apparently, and these are now my words, the animal has an internal clock which continues to advance whether the animal has the opportunity to respond or not. The general notion of the "delay of reinforcement" explanation of fixed interval behavior is that responses that occur shortly after a reinforcement are reinforced after a long delay; those which occur long after a reinforcement are reinforced after a short delay. If response strength is related to delay of reinforcement, the latter should be strengthened more than the former. Therefore, a response gradient will emerge (Morse, 1966). This explanation does not dispose of the problem of temporal discrimination. It transfers the problem from the simple relationship between one reinforcement to the next, to the more complex temporal relationship between response and reinforcement conditional upon the time of the response since the last reinforcement. In any case, the attempt to explain the fixed interval gradients without involving the animals' ability to discriminate one time interval from another does not seem to be a worthwhile exercise, since the estimation methods have demonstrated that animals do possess this capacity.

284

RUSSELL M. CHURCH

B. Explanations Not Involving a Clock

Of course, it is possible to accept the notion that the time interval serves as a discriminative stimulus without postulating an internal clock. There is a conservative tradition within behaviorism, less fashionable now than 20 years ago, that avoids the use of all intervening variables. If a relationship can be established between time and a response, why introduce the concept of an internal clock? This would require two sets of relationships (between time and the clock, and between the clock and response) to replace the single set of relationships between time and a response. Although the concept of an internal clock is not required to explain the results of any of the experiments described in this chapter, there are three types of arguments in favor of the concept. First, it is possible that there is a physiological reality to the internal clock, and that we are more likely to find it if we know its properties. Second, there is the argument of theoretical elegance. An intervening variable may simplify the input-output relationships. That is to say, the list of input-clock and clock-output relations may be shorter than the list of input-output relations. Finally, there is the pragmatic argument. It is possible that the concept of the internal clock will lead to the discovery of capacities of animals that otherwise would not have been identified, and that the concept will guide the investigator toward the discovery of general principles of animal cognition and behavior. C. A Clock Model

When we began research on timing behavior of rats several years ago, the concept of internal clock was, for us, simply a metaphor. As our research progressed, however, we found ourselves searching for the properties of the internal clock. After we discovered some characteristics of the internal clock, our attitude toward it gradually began to change. The concept was no longer a metaphor; we began to believe that the clock actually exists. Results from both the estimation and the production procedures suggest that animals may have an internal clock. Our assumptions are that the animal has an internal clock which advances as a function of time from (or during) a welldefined event, that the animal adopts a temporal criterion, and that there is a response rule which relates the probability of a particular response to the clock setting and the criterion. In the case of the estimation procedure, the signal starts the clock. When the signal terminates and the levers are inserted into the box, the rat reads the value of the clock and makes a decision to respond on the left (short) or right (long) lever, based upon the relationship between the clock setting and the criterion. If the clock setting is less than the criterion, the rat makes a left response; if the clock setting is greater than the criterion, the rat makes a right response. In the case of the production procedure, the signal starts

10. THE INTERNAL CLOCK

285

the clock, and the rat continually makes decisions whether or not to respond based upon the clock setting and the criterion. This type of model contains three basic terms: clock, criterion, and response rule.

1. Clock. For any given time the clock will have a distribution of settings, and this distribution changes in a regular way with time. If t equals time and s(t) equals the subjective representation of this time (clock setting), then the problem is to specify the relationship between s(t) and t. Of course, the clock may be flexible. That is to say, such factors as the nature of the stimulus and the conditions of reinforcement may affect the relationship between s(t) and t. 2. Criterion. If there is differential reinforcement of a response as a function of time from an event, the animal will adopt a temporal criterion for the response. The problem is to specify the criterion as a function of the conditions of reinforcement. Of course, it is possible that some of the variability in performance could be a result of variability in the criterion as well as variability in the clock. The source of variability could be in the clock, in the criterion, or both. In many situations, it is impossible to distinguish among these alternatives. 3. Response Rule. The probability of a response can be predicted on the basis of clock setting, criterion, and conditions of reinforcement. The problem is to specify this rule. A model for the estimation procedure (Fig. 1, left panel) is as follows: (1) The clock advances as a function of time during the signal of a given trial. {2) The rat gradually adopts a criterion somewhere between a clock setting characteristic of the 4-sec signal and a clock setting characteristic of the 16-sec signal. In the case of a symmetric two-choice situation, the criterion would be halfway between the clock settings characteristic of the two extremes. And (3) if the clock setting on a given trial is greater than the criterion, the rat responds with the long response; otherwise it responds with the short response. If the distribution of clock settings at any time is normal, then a cumulative normal distribution would describe the relationship between clock setting and the probability of a long response. A model for the production procedure (Fig. 1, right panel) is as follows: (1) The clock advances as a function of time during the signal of a given trial in a manner similar to that described for the estimation procedure. (2) The rat gradually adopts a criterion related to the duration of the fixed interval. Because of the asymmetry in the payoff and the fairly effortless responses typically used, errors for premature responding are probably less serious than errors of waiting beyond the criterion. Thus, the criterion is probably set low. And (3) the response rate, or probability of a response in a short interval of time, is related to the discrepancy between the clock setting and the criterion. The response rate increases as the clock setting approaches the criterion, i.e., there is some monotonic relationship between response rate and time.

286

RUSSELL M. CHURCH

If the three terms, clock, criterion, and response rule, have a status as separate concepts, then there should be operations which change one of them while leaving the other two unchanged. That is to say, some operations should change the relationship between time and the clock setting, some should change the criterion, and some should change the response rule. A change in the clock setting, or in the criterion, would displace the curves in Fig. 1 horizontally, on the time axis; a change in the response rule would displace the curves vertically on the response axis. For example, consider a drug such as amphetamine at a dose which increases the mean response rate on a temporal discrimination. Does it affect the operation of the clock, the response criterion, or the response rule? One test would be to determine whether the temporal gradient following the drug is displaced in time or in response rate. Our greatest interest will be in manipulations that change the operation of the internal clock. The first step is to describe the properties of the internal clock, and this requires that the basic terms be firmly connected to stimulus conditions and observable behavior, and that the relationship among concepts be expressed by quantitative rules. Ill. SOME PROPERTIES OF THE INTERNAL CLOCK

A. Properties of External Clocks

All clocks, by definition, change with time in a regular way. This is the only property that they all share. A sundial, a digital watch, a capacitor circuit, a stopwatch, and a computer clock all change with time in a regular way, but they differ in many other ways. They operate at different rates; some are more accu-

Proportiona I

Absolute Up

0

15

30

0

.5

1

Down 30

15

0

1

.5

0

FIG. 2. Four types of clock. The clocks in the left panels time in absolute units (e.g., seconds); the clocks in the right panels time in proportional units. The clocks in the top row time up; the clocks in the bottom row time down.

10. THE INTERNAL CLOCK

287

rate than others; some are continuous and others are discrete. Some are cyclical, and others change only in one direction. Some are driven by a specific external event, while others depend on internally stored energy. Some time up and others time down. Figure 2 is a diagram of four types of clocks. An absolute, up timer increases in units related to seconds toward a criterion (upper left panel). An absolute, down timer decreases in units related to seconds (lower left panel). It begins at a criterion and decreases toward a fixed value, conventionally defined as zero. Thus the state of a down timer is specified by its clock setting alone, but the state of an up timer requires the specification of the criterion as well as the clock setting. A proportional timer changes in units related to the duration of the interval to be timed (right panels). Such a timer runs half as fast for an interval of 2c sec (say, 60 sec) than for an interval of c sec (say, 30 sec). In the next section, I will examine some of the properties of the internal clock of the rat to determine which of the properties of various external clocks it possesses. More specifically, I will attempt to answer the following three questions: (1) Does an animal use the same internal clock to time the duration of a light and a sound? (2) Are the units of the internal clock absolute (related to seconds), or are they proportional to the duration of the interval being timed? (3) Does the internal clock time up toward an adjustable criterion value, or does it time down toward a fixed criterion? B. The Proportionality Result

The shapes of different fixed interval gradients are very similar when the dependent variable is the percentage of the maximum rate, and the independent variable is the percentage of the interval. For example, if an animal on a 30-sec fixed-interval schedule has achieved about 50% of its maximum rate after it is 75% of the way through the interval, the same will be true of an animal on a 60-sec fixed -interval schedule. We refer to this result as "proportionality." There are many examples of such a proportionality result. One of the more impressive examples was presented by Dews (1970) for fixed-interval behavior of pigeons that was equivalent for intervals varying from 30 sec to 50 min when scaled in proportional units. Some qualifications in the generality of this result are suggested by the data of others (e.g., Catania & Reynolds, 1968 Fig. 1b), but the result has been reported under a wide variety of conditions. Gibbon (1972) has shown that the relative response rate of rats on Sidman avoidance in a lever box is a function of the proportion of the interval completed, and Libby & Church (1974) have replicated this result for Sidman avoidance in a shuttle box. Many other experiments, including several from our laboratory, support the idea that the response rate, relative to the mean rate, is a function of time expressed as a proportion of the interval, rather than time expressed in seconds. For example, LaBarbera and Church (1974) found this principle to hold for temporal

288

RUSSELL M. CHURCH

50

40 c:

2:

..... Q)

0.

30

"' Q)

"'c:0

0.

20

"' Q)

a:

10 0

0

30 Time (sec)

60 0

.5

1.0

Proportion

FIG. 3. Response rate as a function of time (left panel) and proportion of the interval (right panel) of a rat in a 30-sec fixed-interval (solid circles) and a 60-sec fixed-interval (open circles) condition.

conditioning. The psychometric functions of the estimation procedures are also similar with time scaled in proportional units (Stubbs, 1968). The generality of the conclusion is increased by the great differences among the procedures under which it is found. The proportionality result means that a person could predict the relative response rate much more accurately from a knowledge of the proportion of the interval than from either (1) the number of seconds since the last reinforcement or (2) the number of seconds until the next reinforcement. Typically, the different treatments in an experiment examining proportionality are given to different animals, or they are given to the same animals in different phases of the experiment. Figure 3 shows a case in which the proportionality result holds even when the different treatments are given to the same animals during the same sessions. In this example, Seth Roberts and I trained rats on two discrete-trial fixed-interval schedules. Specifically, they were given discrete trials of 30-sec fixed-interval reinforcement signaled by a light, and other trials of 60sec fixed-interval reinforcement signaled by a noise. When an interval was completed, food was primed, and following the next response (1) the food was delivered, (2) the signal was turned off, and (3) the lever was withdrawn. Mter a variable intertrial interval, another trial was given. The respone rate of one rat is plotted as a function of time (left panel) and as a function of proportion of the interval (right panel). It is clear that the response rate is more closely related

10. THE INTERNAL CLOCK

289

to the proportion of the interval that was completed than to the number of seconds since the interval began, or the number of seconds until the next reinforcement was due. The proportionality result suggests how animals may be timing the interval. They may advance the internal clock at a rate r that is proportional to the duration of the interval to be timed t, i.e., r = k/t, where k is a constant. For example, the clock may run twice as fast while timing a 30-sec interval than while timing a 60-sec interval. If each clock setting corresponds to a particular mean relative response rate, then the proportionality result would occur. Of course, the proportionality result does not prove that the units of the internal clock are proportional to the interval being timed, since the assumption that each clock setting corresponds to a particular mean relative response rate may be in error. The next section will show that the units of the clock are absolute, not proportional. C. Three Properties of the Internal Clock: The Shift Experiment

To determine whether or not rats were timing in proportional units, Seth Roberts and I performed the following experiment (Roberts & Church, 1976; 1978). We trained each rat in a manner similar to that shown for the rat in Fig. 3. By the final day of fixed-interval training, responding increased in a fairly regular manner as the time of food approached under both the signal for the 30-sec and 60-sec fixed intervals. During the critical part of the experiment, rats continued to receive trials with the 30-sec light signal and with the 60-sec noise signal. In addition, on about one-third of the trials, the 30-sec light signal shifted abruptly to the 60sec noise signal. Shifts could happen 6, 12, 18, 24, or 30 sec after the trial began. If the rats used the same internal clock to time the light and noise, whether they were timing in absolute or proportional units and whether they were timing up or down could be inferred from their response rate during the trials shifted to the 60-sec noise signal. For example, after 12 sec of light (the 30-sec signal), the light could go off and the noise (the 60-sec signal) could go on. The four possibilities shown in Fig. 2 are as follows: If the animal used an absolute up timer, 12 sec have passed, so performance would be equivalent to an FI-60 after 12 sec. If the animal used an absolute down timer, 18 sec are left, so the performance would be equivalent to an FI-60 when 18 sec are left (i.e., after 42 sec). If the animal used a proportional up timer, 40% of the time has passed, so performance would be equivalent to an FI-60 after 40% of the time has passed (after 24 sec). If the animal used a proportional down timer, 60% of the time is left, so performance would be equivalent to an Fl-60 when 60% of the time is left (i.e., 36 sec). Thus the empirical queston is to determine whether the performance of a rat switched from an FI-30 to an Fl-60 after 12 sec is equivalent to its performance on an FI-60 after 12, 24, or 42 sec.

290

RUSSELL M. CHURCH

2.0 0

'.j:j ctl

a:

1.5

(!)

en

c: 0

c. en (!)

a:

10

20

30

40

50

60

Time (sec) FIG. 4. Mean response ratio during the 60-sec signal as a function of time since the trial began. Open circles are for trials that began as 60-sec fixed intervals; solid circles are for trials that were shifted to 60-sec fixed intervals. Points from trials that had the shift at the same time are connected.

On shifted trials we primed the food at the time appropriate for an absolute up timer for six rats (absolute group), and primed the food at a time appropriate for a proportional up timer for the remaining six rats (proportional group). The main result on the first day after shifts began is shown in Fig. 4. Since the two treatment groups (absolute and proportional) were not yet distinguishable, we have combined the groups in this figure. The data are reported as response ratios, defined as the response rate during a 6-sec interval divided by overall response rate. For example, a rat with aresponse rate of 40 responses per minute during the last 6 sec of a 60-sec interval, and an overall response rate of 20 responses per minute during the entire 60-sec interval has a response ratio of 2.0 in that interval. We used this measure to compensate for the large differences in overall response rate from rat to rat. The open circles show the mean response ratio for trials which began with the 60-sec signal as a function of time since the trial began; the closed circles show the mean response ratio for trials which were shifted to the 60-sec signal at various times (6, 12, 18, and 24 sec after the 30-sec signal began), also as a function of time since the trial began. The major point is that the response ratio in the 60sec signal was a function of the time since the trial began. It did not matter

10.

291

THE INTERNAL CLOCK

whether the trial began with a 60-sec signal, or if it was shifted to a 60-sec signal after 6, 12, 18, or 24 sec. To predict the response ratio in the 60-sec signal, one needed to know only the time since the trial began; it did not matter how much of the time consisted of the 30-sec signal and how much of the time consisted of the 60-sec signal. What do these results tell us about the internal clock? First, they provide evidence that a single clock was used in this task. If the rat used separate clocks to keep track of time during the 30-sec visual signal and the 60-sec auditory signal, then the length of time the animal spent in the visual signal prior to the shift to the auditory signal should have no effect on its behavior. (In Fig. 4 all five functions would begin with the same response ratio.) If, on the other hand, the rat used the same clock to keep track of time during the 30-sec visual signal and the 60-sec auditory signal, the response ratio in the 60-sec signal should be related to the time spent in the 30-sec signal. Figure 4 shows this is the case, so apparently a single clock was used in this task. Since the same clock is used for auditory and visual signals, the clock must centrally located, i.e., more central than the auditory or visual pathways. Our second conclusion is that the clock advances as a function of absolute, not proportional time. This is clear from the fact that the response ratio was a function of the duration of the combined signal (30 sec and 60 sec). Since it did not matter how much of the combined signal was the 30-sec signal, there is no evidence that the clock timed at a faster rate in the 30-sec signal than in the 60PROPORTIONAL GROUP

ABSOLUTE GROUP

_.

0

i= -

I-

(.)

w a:: ~

.17

4

(.)

IJ..

0

a:: w

__J

CD ------0

CHANCE PERFORMANCE IS 0

f-

m

z

CONTROL

DELAY

FIG. 5. The effects of delay produced by confinement in the center platform on choice accuracy as indicated by the mean (histogram) and range (bar) of the number of correct responses in the first eight choices.

error by repeating a choice made late during each day's test. To obtain evidence about primacy and recency, we determined for each correct choice the relative probability of an error by repeating that choice, divided by the number of opportunities to make an error by repeating that choice. Scores can range from 0, indicating that no errors were made to the choice in question, to some positive number, indicating that some errors were made to that choice. The more positive the score, the greater the relative probability of an error to that choice. (Further details are given in Olton & Samuelson, 1976). The results of this analysis are presented in Fig. 6. For both Days 1 to 10 and Days 11 to 20, the relative probability of making an error by repeating a previous choice declined slightly as a function of the sequential position of the correct choice. These data indicate that when an error was made it was least likely to be made to the just previous choice. This type of result is usually interpreted as evidence for a "recency" effect; i.e., the items placed in memory most recently are remembered best and thus fewer errors are made to them. The figure provides no evidence for a primacy effect, i.e., a decrease in the probability

12. SPATIAL MEMORY

351

of an error to the first choice of the day. Thus memory processing in the C8 procedure is characterized by a small recency effect but no primacy effect suggesting that the items most recently entered into the memory store are remembered best.

3. Confusion Among Items. The analysis of the probability of a correct response (see Section C.l) describes the likelihood that any given response will be an error and the analysis of the probability of repetition (see page 350) describes the sequential distribution of these errors among previous choices. The . question remains as to the spatial distribution of errors. In particular, when errors occurred, did they tend to be made by repeating arms close to ones which still contaihed food, suggesting that there was some generalization or confusion between chosen and unchosen arms? We calculated an error-location score (see Olton & Samuelson, 1976 for details), which reflected the spatial distribution of the remaining correct arms with respect to each arm on which an error was made, and then compared this score to the value expected by chance. In every case, the error-location scores closely approximated the chance value, demonstrating that when errors occurred there was no tendency for them to be made to arms close to correct ones, and suggesting that these errors did not arise because of generalization or confusion among arms. 0

0 70

Q

~

>
!::

DAYS I- 10

\

----.

DAYS 11-20 o----"

\ \

\

1--

0:: LL.

\

\

0

G:i 0. w

I I

:J

0::

I

0.30

\

\

0.20

v

/

/

/

/

_J

CD

' I

21~/.

·

1:·

tft··:···l3e}::~;'· ..,;i,::chophysiological disorders, addictions, psychosis and disorders of childhood. In none of these categories are complete 'hi-fidelity' models of human psychiatric conditions looked for. The relevance and value of animal studies varies from case to case: with psychosis it is to do with drugs and hospital management; with neurosis, with learning; with childhood disorders, with maternal deprivation and development; and with addictions, the roles of conditioned withdrawal and tolerance in relapse after detoxification. Thus the present work is not an attempt to force animal models on to clinical psychiatry but to illustrate ways in which these models can have psychiatric value. A summary of such ways constitutes the final, concluding chapter. At this point I would like to acknowledge the secretarial assistance provided by Jackie Logan, Rischa Sidon, Rhonda Strasberg and Marilyn Weinper, whose contributions are distributed throughout the entire manuscript, and to thank those authors and publishers who have permitted me to reproduce figures and tables from their original works. Appropriate acknowledgments in these cases appear with each individual illustration. J.D.K. Toronto, 1985

xu

Part I

Domains of biological psychiatry

1 Psychopathology: the status of animals

Comparative experimental psychiatry Scope and origin A few years ago, Gay ( 1967) made a vigorous plea for expanding the domain of comparative medicine, and two years later Leader ( 1969) surveyed the long history of contributions to medical skills made by comparative pathologists seeking animal models of human diseases. Soon after that, three famous ethologists, Karl von Frisch, Konrad Lorenz and Nikolas Tinbergen received a Nobel prize for the lessons that their accounts of animal habits have taught medical practitioners about human diseases and behaviour disorders. In his acceptance speech, Tin bergen ( 1974) vividly recounts how the ethological methods of studying natural behaviours of animals enlarged his understanding of the behaviour of autistic children and contributed to new methods of treating this childhood disability. The psychological study of abnormal behaviour in animals does not have as long and impressive a history as that of the study of physical diseases by way of laboratory animal models, nor has it attained the eminence accorded to ethological research; but it has amassed a substantial body of information since its beginning with the work of Maria Yerofeeva (Erofeew, 1916) in Pavlov's laboratory in 1912. Following the pioneering work of Pavlov (1927) and his colleagues there have been numerous attempts to generate animal models of human behavioural anomalies. These include early studies of cats by Masserman (1943) and Wolpe (1958), of dogs by Gantt (1944) and of sheep and goats by Liddell (1944), all of which were directed to the creation of animal versions of human neuroses. It is now doubtful that these early efforts achieved their intentions (Broadhurst, 1973), but the number of human psycholo-

3

DOMAINS OF BIOLOGICAL PSYCHIATRY gical disabilities for which animal models have since been proposed include schizophrenia (Ellison, 1979), depression (Seligman, 1975), asthma (Ottenberg, Stein, Lewis and Hamilton, 1958), epilepsy (Gaito, 1976; Killam, Killam and Naquet, 1967), infantile autism (Scott and Senay, 1973), alcoholism and drug addiction (Gilbert, 1977), and many more (Keehn, 1979; Maser and Seligman, 1977). Some of these conditions are the results of deliberate attempts to create specific parallels to human problems, while others emerged unexpectedly from routine care and maintenance laboratory practices (Startsev, 1976), or developed in the course of investigations into basic behavioural processes. The classic example of experimental neurosis first appeared in this way, and a similar later case is recounted by Ellison (1979): I became interested in ... experimental neurosis ... when ... one of the dogs I was training in a difficult discrimination developed a classical case. This dog had merely to stand in a conditioning platform and would receive a piece of dog candy 16 sec after brief presentation of a high tone but no candy following a brief low tone. He gradually refused to come to the experiment and had to be carried, even though dogs being shocked will readily follow the experimenter. He also developed trembling attacks and refused to eat candy dispensed from the feeder although he was quite hungry and would eat from the experimenter's hand (unless the piece of candy had previously been dispensed from the feeder!). His salivation became quite erratic, and he would stare fixedly at spots on the wall. (Ellison, 1979, p.82) It may be improper to call this animal neurotic in the human sense (Hunt, 1964), or even to describe either animals or humans as psychopathological. Psychopathology, however, is a convenient shorthand for alluding to behavioural abnormalities that have been recognized over the ages, and is used here for contemporary convenience, like psychology and psychosomatics, with only token commitment to a medical model of madness. (Siegler and Osmond, 1974).

What must an animal model? Disease entities, diagnosis and clinical pictures A major function of psychiatry is the diagnosis of mental illnesses associated with particular patterns of abnormal behaviours, perceptions and emotions, although voices have been raised against 4

Psychopathology: the status of animals such undertakings. Thus, according to Frank (1975): What we truly need, in the study of psychopathology, is a revolution .... Without it, we ... are doomed to an outmoded ... inappropriate mode of conceptualization that contributed little if anything to the understanding of people. Our system of classification in psychiatry has proven to be: full of sound and fury, signifying nothing. (Frank, 1975, p.80) But others, like Kendell (1975), express another opinion. Many present day psychiatrists have lost interest in the whole issue of diagnosis while others have suggested that it is an unnecessary, even a harmful exercise. This book was born of the conviction that such attitudes are profoundly mistaken, and that the development of a reliable and valid classification of the phenomena of mental illness are two of the most important problems facing contemporary psychiatry. (Kendell, 1975, p. vii) These problems have been faced before (Menninger et al., 1963). Several solutions have been offered, including one by Carl Wernicke (1848-1905) who distinguished deficit, distortion and excess in sensory, motor and cognitive functions according to the following scheme:

Function Psychosensory Deficit Excess Distortion

Psychomotor

Anaesthesia Akinesia Hyperaesthesia Hyperkinesia Parakinesia Parasthesia

Intrapsychic Afunction H yperfunction Parafunction

This schema permits a nine-fold classification of possible psychological malfunctions, but it is not a diagnostic system for differential diagnosis of mental diseases in the modern fashion. Numerous other classifications have been proposed throughout the ages. These are summarized by Menninger et al. (1963) who placed them in two groups that derive, respectively, from the historic traditions of Hippocrates and Plato: l classifications of individuals into clinical types based on presenting clinical pictures; and · 2 classifications of disease entities. In the Hippocratic tradition, the emphasis is on case histories of individuals and on treatment formulations based on similarities in clinical pictures of different patients; in the Platonic tradition, the emphasis is on categorical diagnosis, based on the assertion that

5

DOMAINS OF BIOLOGICAL PSYCHIATRY disease entities are ideal universals of which particular patients are more or less imperfect examples. Modern institutional psychiatry is mainly Platonic, while psychoanalysis, behavioural analysis and allied movements are Hippocratic. The two major classes of modern arrangements of mental disorders in the Platonic sense, neuroses and psychoses, began to take their forms in the seventeenth and eighteenth centuries. That part of a botanical system to do with neurosis devised by William Cullen ( 1712-90) is summarized in Table 1.1. Table 1.1 William Cullen's classification of neuroses Neurosis -

Class:

disturbances of sense and motion without infection. Diminution of voluntary motion or sense with 2 genera:

Orders:

apoplexy and paralysis. II

Diminution of involuntary motions with 4 genera, including vomiting and dyspepsia.

III

Irregular motionJ· of muscles and spasms with I 7 genera,

including epilepsy, asthma, chorea and hysteria.

IV

Disorders of judgment without fever or coma with 4

genera, including amentia, mania and melancholia Adapted from Menninger, K., et al. (1963) The Vital Balance, New York, Viking.

In the same era Pinel (1745-1826) proposed a classification by clinical types that now would be attributed to psychosis or organic brain disorder.

Pinel's fundamental clinical types

l 2 3 4

Mania: acute excitement. Melancholia: depression. Dementia: incoherent thought. Idiotism: feeble-mindedness.

The basis of present classification systems was laid down by Kraepelin ( 1856-1926), and after several attempts to accommodate psychiatric disorders with international classifications of diseases and causes of death, the American Psychiatric Association (1952) proposed a complex diagnostic system based on psychiatric experience in World War II. A simplification of the Association's 6

Psychopathology: the status of animals Table 1.2 Classification of abnormal patterns of behaviour

Disorders

caused

by

or

associated with function

A Acute brain disorders B Chronic brain disorders

impairment

of

brain

tissue

C Mental deficiency I Familial or hereditary deficiency 2 Idiopathic deficiency

Disorders of psychogenic ongzn or without clearly defined physical cause or structural change in the brain Psychotic disorders A Involutional psychosis B Affective reactions C Schizophrenic reactions I 2 3 4 5 6 7 8

Simple Hebephrenic Catatonic Paranoid Acute undifferentiated Chronic undifferentiated Schizo-affective Childhood

D Paranoid reactions

Psychosomatic disorders A Skin reaction B Musculoskeletal reaction C Respiratory reaction D Cardiovascular reaction E Haemic and lymphatic reaction F Gastrointestinal reaction G Genitourinary reaction H Endocrine reaction I Nervous system reaction J Special sense reaction

Neurotic disorders A Anxietv reaction B Dissociative reaction C Conversion reaction D Phobic reaction E Obsessive-compulsive reaction F Depressive reaction G Other neurotic reactions

Character disorders A Personality pattern disturbance B Personality trait disturbance C Sociopathic personality D Special symptoms Transient personality disorders A Gross stress reaction B Adult situation reaction C Adjustment reaction of infancy D Adjustment reaction of childhood E Adjustment reaction of adolescence F Adjustment reaction of late life

Simplification of a classification of psychological abnormalities proposed by the American Psychiatric Association based upon experience during and after World War I I. ( Diagnostir and Statistical manual, Mental Disorders, Washington, American Psychiatric Association, 1952.) Unlike earlier classifications made by individual men, this classification was agreed upon by a committee of psychiatrists. The classification was expanded in DSM-II and broadened into a multi-axial system in DSM-Ill. See text for details (Reproduced from Kcehn,.J.D. (I 962) The Prediction and Control of Behavior: A Shorter Introduction to Psychology, Beirut, Khayat).

7

DOMAINS OF BIOLOGICAL PSYCHIATRY Table 1.3 Multi-axis diagnostic and statistic system adopted by the American Psychiatric Association (DSM-III)

AXIS I

AXIS II

Traditional psychiatric syndrome

Principally administrative diagnosis and allocation for treatment, epidemiology and statistics. Personality disorders (adults); Developmental disorders (children)

For comparisons with norms rather than with ideals.

AXIS III

Non-mental medical disorders

Usual physical medical history.

AXIS IV

Severity of stress in previous year

Rated on a 7-point scale.

AXIS V

Highest level of adaptive behavior in previous year

Rated on a 7-point scale. This multi-axis system has replaced the 'botanical' classification system employed in DSM-I (see summary in Table 1.2) and its successor DSM-1/. Already a committee is considering a revision that will become DSM-IV. A brief psychological appraisal of DSM-III appears in Schact, T., and "\athan, P.E. (1977) 'But is it good for psychologists? Appraisal and status of DSM-III', American Psychologist, 32, 1017-25. The system is described in full in Webb, L.J., et al. (1981) DS~i-III training guide, "\ew York, Brunner .\laze!. /

official classification (Diagnostic and Statistical Manual I) is shown in Table 1.2. Since that classification the World Health Organization has revised its International Classification of Diseases, and the American Psychiatric Association's system has passed through a second (DSM-II) to a third (DSM-III) revision. The third revision combines diag1,1osis and formulation by including the categorization of diseases and the characterization of persons according to the summary in Table 1.3. From it, treatment plans are based on personal case histories (in the Hippocratic tradition) rather than on diagnostic labels that mark universal Platonic-type ·disease entities. The third diagnostic and statistical manual of the American Psychiatric Association thus incorporates both historical aspects of diagnosis in mental illness. Axis I represents the traditional psychiatric diagnostic system, although revised and expanded to include disorders of social communication as well as the mental illnesses summarized in Table 1.2. This axis is relevant for animal models but may turn out to be less important than Axis II, which is to do with atypical childhood and adult behaviours.

8

Psychopathology: the status of animals As an alternative to the formal categorization of mental illnesses, Shapiro ( 197 5) proposes a list of ten manifestations of human psychopathology marked by a combination of distress, disablement, incongruence with reality and social inappropriateness. l Intense referential feelings, such as fear of dogs and depression at work. 2 Intense non-referential feelings, like generalized anxiety and depression. 3 Exaggerated or reduced drives, expressed as insomnia, impotence and obesity. 4 Strong irrational beliefs, such as jealousy and paranoia. 5 Cognitive dysfunctions, like inability to concentrate or remember. 6 Maladapted non-social behaviours, in such forms as persistent handwashing and housecleaning. 7 Maladapted social behaviours, characterised by abusive language, talking to oneself and withdrawal. 8 Disturbed perceptual experience, such as hallucinations. 9 Very intense somatic experiences, like intense headaches, asthma and gastro-intestinal pains. 10 Motor dysfunctions, including tics, tremors, dyskinesias and stereotypies. Plainly some of these manifestations of human psychopathology are closer to the psychology of animals than are others. Motor disturbances are easily observed, and very intense somatic disturbances may be readily il1ferred in animals, as may be reduced or exaggerated drives, but with intense referential feelings, disturbances of perception and cognition and strong irrational beliefs it is a very different matter, with maladaptive social and non-social behaviours somewhere between the extremes. Nevertheless, illustrations of some of these characteristic human psychological problems are available in animals. A case of bizarre posturing and possible hallucination by an isolation-reared monkey is reproduced in Chapter 3, where unnatural animal movements and postures are described in detail. In addition, phobias and fears arc often attributed to laboratory animals that succeed in avoiding electric shocks in standard laboratory situations. Animals that fail to avoid, on the other hand, after prior exposure to inescapable shock, are said by Seligman (1975) to fail through a learned sense of helplessness. The fear and the helplessness supposedly felt by the animals in these cases may or may not experientially resemble those that Shapiro (1975) identifies as common disabling and distressing features in humans, but these are matters of symptomatology, in which humans and lower animals cannot be expected to

9

DOMAINS OF BIOLOGICAL PSYCHIATRY be identical. Nevertheless, originating circumstances that lead to behavioural disturbances in animals may be similar to those that cause psychopathologies in people.

Symptomatology and origination Uncertainty about the function of diagnosis in psychiatry greatly complicates the search for animal models of human mental illness by failing to provide clear objects for the models to copy, but Broadhurst (1973) may have reached an unnecessarily harsh conclusion by taking one side of the disease entity versus clinical picture debate. In 1960 it seemed sufficient to explore the extent of the evidence for an experimental neurosis in the strict sense, that is to say, to decide if evidence for a clear-cut animal analogue of a clinical disease entity of some sort existed. The answer, not surprisingly, it now seems was 'No'. No animal preparation that convincingly mimicked such entity could be provided by experimental psychology for the study of etiology, prognosis or therapy. (Broadhurst, 1973, p. 745) As it happens, many of the early attempts to generate experimental neurosis in laboratories actually demonstrated animal clinical pictures. Audiogenic seizures in mice and rats (Krushinskii, 1962), tonic immobility in chickens, rabbits, lizards, toads and goats (Crawford and Prestrude, 1977), and fear-induced feeding problems in laboratory cats (Masserman, 1943; Wolpe, 1958) and dogs (Pavlov, 1927) are prominent examples. Evidence for an experimental neurosis may be non-existent for the reason that there is no natural object for an experimental image to depict. Recent studies of experimental neurosis from Russian origins (Miminoshvili, 1960; Startsev, 1976) focus more on specific psychophysiological disorders like gastric achylia, hyperglycaemia, sexual dysfunction, hypertension and cardiac insufficiency than on general neurotic emotionality, and this focus is generally true whenever modern animal analogues of human disabilities are sought. Thus not schizophrenia but the biochemical roots of stereotypy is the basis of some animal models of psychosis; not alcoholism but relationships between animal and human responses to alcohol; not anaclitic depression but the animal and human reactions to isolation rearing, are the foci of contemporary research on animal models in psychopathology. Even so, as with psychiatric indecision about whether to keep or discard traditional diagnosis, laboratory workers with animals still cautiously continue using traditional terms. Classification systems 10

Psychopathology: the status of animals as elaborate as those of the American Psychiatric Association and World Health Organization are too complex ever to be of use with animals, but a more manageable list of diagnostic categories is employed by Woodruff, Goodwin and Guze (1974). All of the twelve psychiatric diagnoses that they enumerate and describe affective disorders, schizophrenic disorders, anxiety neurosis, hysteria, obsessional neurosis, phobic neurosis, alcoholism, drug dependence, sociopathy, brain syndrome, anorexia nervosa and sexual problems- have counterparts in animal behaviours. As with humans, there are reports of animal affective disorders (Seligman, 1975), schizophrenic disorders (Ellison, 1979), anxiety neurosis (Pavlov, 1927), hysteria (Sanger and Hamdy, 1962), obsessional neurosis (Ellen, 1956), phobic neurosis (Wolpe, 1958), alcoholism (Falk, Samson and Winger, 1972), drug dependence (Jones and Prada, 1973), sociopathy (Ellison, 1979), brain syndrome (Auer and Smith, 1940), anorexia nervosa (Masserman, 1943) and animal problems with sex (Chertok and Fontaine, 1963). Nevertheless, comparative experimental psychiatry cannot progress on the basis of presenting symptoms alone, for as Liddell (1956) long ago remarked, a psychodynamics of animal behaviour can never be. He continues: Because of man's incredibly complicated cognitive machinery, his neurotic symptoms may exhibit a bewildering diversity. Nevertheless, our emotionally disturbed animals under careful observation show many of the same or closely similar symptoms. The physician is in a much better position, however, to explore and analyze his psychoneurotic patient's symptomatology than is the behaviorist in the case of his experimentally neurotic animal. When it comes to analyzing the originating situations ... the shoe is on the other foot. The behaviorist can create and rigorously control the situation in which experimental neuroses originate. (Liddell, 1956, p. 59) Liddell spoke only of experimental neurosis, but the point applies to any animal model of human psychopathology. In this and other chapters many spontaneous animal disablements are described with psychiatric labels, but the human and animal cases can never be exactly the same in symptomatology. However, research scientists in this field run not one risk but two: one that the disablements they create are specific to their laboratory subjects, non-existent in natural animal life; the other that their creations are inapplicable to human disorders. Let me address the first risk by way of examples of spontaneous abnormal animal behaviour, and the second by way of criteria for animal models of human psychopathologies. II

DOMAINS OF BIOLOGICAL PSYCHIATRY

The psychopathology of animal life Spontaneous animal psychopathologies In farms and zoos

Abnormal animal behaviours that correspond to psychopathologies in humans are not only found in laboratories as deliberate creations for the benefit of mankind. Behavioural abnormalities spontaneously exhibited by all sorts of animals have been observed by ethologists, zoologists, veterinarians, husbandrymen, farmers and animal keepers in circuses and zoos (Fox, 1968). Thus there are reports of intromission phobia in the bull (Fraser, 1957), hysteria in hens (Sanger and Hamdy, 1962), ulcerative colitis in the gibbon (Stout and Snyder, 1969), and sexual inversions in certain birds and fish (Morris, 1955). Chertok and Fontaine (1963), in an account of psychosomatics in veterinary medicine, include collective epilepsy among dogs, nymphomania in cats, post-emotional traumas in horses, dogs and cats, and impotence and pseudopregnancy in the rat. To these, Brion ( 1964) and Levy ( 1952) add a variety of animal tics, convulsions and seizures, while Fraser ( 1960) describes cases of spontaneous tonic immobility ('animal hypnosis') in the horse, the goat and the cow. Several more behavioural problems of farm animals, described in a standard textbook of veterinary medicine as 'functional nervous diseases in animals', are summarized in Table 1.4. Behavioural abnormalities of captive animals in zoos include non-adaptive escape reactions, food refusal, excessive aggression, stereotyped motor reactions, displacement scratching, selfmutilation, homosexuality, sexual perversions, perversions of appetite (coprophagia), apathy, and defective mother-infant interactions (Meyer-Holzapfel, 1968). Among other possibilities, Meyer-Holzapfel interprets these as results of interference with normal inter-animal social intercourse, and of thwarting of natural flight reactions when danger signals appear. These could explain the sudden deaths of zoo animals described below, and also the arrested development of Merlin, the wild orphaned chimpanzee that van-Lawick Goodall (1971) observed. In the wild

Merlin was a 3-year-old chimpanzee who was 'adopted' by his older sister Miff after his mother died. He was involved in a violent 12

Psychopathology: the status of animals Table 1.4 Functional nervous diseases in animals

I 2 3 4 5 6

7 8 9

Dizziness (vertigo) in horses and dogs. Epilepsy (falling sickness) in dogs. Tetany (muscular spasms) in various animals. Catalepsy, e.g. convulsions, muscular plasticity in cows. Neuroses of pregnancy, parturition and lactation; milk fever (coma). Chorea, complex involuntary arrhythmic muscular movements. Twitch spasms (tics). Muscular tremors. Psychoses: feeblemindedness traumatic dementia maniacal staggers in horses mass panic in herds (stampeding) hysteria in dogs (fright disease, running fits) degenerative psychopathic constitution (sexual perversions)

Condensed from Hutyra, F., Marek,J., and Manninger, R. (1949) Special Pathology and Therapeutics of the Diseases rif Domestic Animals, vol. III, 5th edition, London, Balliere, Tindall & Cox.

encounter with a dominant male whose charge he failed to avoid, and thereafter developed abnormal social behaviour and a bizarre autistic stereotypy. MERLIN When he was four years old Merlin was far more submissive than other youngsters of that age: constantly he approached adults to ingratiate himself, turning repeatedly to present his rump, or crouching, pant-grunting before them. At the other end of the scale, Merlin was extra-aggressive to other infants of his own age .... As Merlin entered his sixth year his behavior was becoming rapidly more abnormal. Sometimes he hung upside down ... suspended almost motionless for several minutes at a time. Hunched up with his arms around his knees, he sat often rocking from side to side with wide-open eyes. (van Lawick-Goodall, 1971, p. 227) Illustrations such as Merlin are uncommon because naturalistic observations of wild animals more often focus on herds or troops than on target individuals, and abnormality is normally an

13

DOMAINS OF BIOLOGICAL PSYCHIATRY individual, not a collective, phenomenon. Nevertheless collective behavioural pathologies have been seen in humans, e.g. St Vitus's dance and mob hysteria, and there are comparable situations with animals. Migrating lemmings and stampeding cattle are wellknown examples; less well known are the rampages of herds of elephants intoxicated through consumption of fermented grain and fruit (Carrington, 1959).

During care and maintenance routines As with Merlin, disturbances in social relationships can cause pathological responses by caged animals in laboratories and zoos, even when they are not exposed to deliberate stress. The cases of agitation in the rhesus monkey, Cupid, and the chimpanzee, Dennis, described in Chapters 2 and 7, respectively, are illustrations. Both were separated from their normal mates for routine laboratory purposes. Similar examples from zoos, leading to inexplicable sudden deaths, are reported by Christian and Ratcliffe ( 1952) for otters, minks, a cheetah, a serval and a lynx, and by Stout and Snyder (1969) for Saimang gibbons. From the Sukhumi Research Station in the USSR, Startsev (1976) reports the sudden death of a baboon following a routine cage reshuffle that brought the victim in view of a group of dominant males. ZAGREB On August 7, 1965 [Zagreb and Ambarchik] were simultaneously transferred to a large cage .... Outside the cage, separated by a transparent barrier, were several full grown males, who greeted their new young neighbours with threatening gestures and vocalizations .... In the very first minutes both showed a gait disturbance characterized by incoordination and a posture with the knees half flexed .... During the subsequent 3 days [Zagreb's] movements reflected a steadily progressive hypotonia: he sat continuously with arms extended and head sunk down between his knees . . • . [Later] one of the animal handlers attempted to catch him ... and he fell to the floor, struggled a short time, and died. A careful autopsy revealed no anatomical lesion to which one could attribute such severe motor impairment. (Startsev, 1976, pp.l31-2) Ewing ( 1967) reports a similar phenomenon in cockroaches, wherein the subordinate members of pairs may die with no signs of physical injury when they are repeatedly attacked by dominant partners who do not respond to the victim's submissive posture. 14

Psychopathology: the status

of animals

Table 1.5 Most common spontaneous neuropathological disorders observed in a group of thirty-nine hamadryas baboons held in captivity

Disorder

Tonic-clonic seizures Adynamia after and between attacks Vomiting Tonic seizures Fibrillary muscle twitches Tremors Myorhythms Hypertonia

After Startsev, V.G. (1976) Primate Models N.J., Erlbaum.

Number of animals

15 14 14 13 11 10

8 7

of Human Neurogenic Disorders, Hillsdale,

'Voodoo deaths' in wild rats after restraint are described by Richter (1957). Upon release into a water jar they drown in a few minutes instead of surviving for the normal several hours. Spontaneous behavioural abnormalities in animals are distortions from normal animal behaviours, not just analogues of psychopathologies in humans. Startsev ( 1976) observed convulsive attacks, paresis, paralysis, hyperkinesis and other motor disturbances in more than thirty baboons at the Soviet primate research laboratories at Sukhumi. The disturbances, further detailed in Table 1.5, were spontaneous inasmuch as they were not intended by the maintenance and experimental procedures used with the animals. Seventeen other baboons were thus employed in deliberate attempts to produce experimental models of these spontaneous baboon disorders, and the symptoms were found to result from irregular disturbances in the animals' living routines on the one hand, and forced physical restraint on the other. In all cases, convulsive attacks and other 'hysterical' motor disturbances were associated with stressful situations accompanied by intense motor activity, such as attempting to flee from a handler or struggling against physical restraint.

In pets When experimental animal abnormalities model spontaneous 15

DOMAINS OF BIOLOGICAL PSYCHIATRY animal abnormalities it is safe to look for common originating situations, and it is likely that the spontaneous motor disturbances in Startsev's colony of baboons originated in the same way as those of his experimental animals. When originating situations that disable animals also disable humans, such as those depicted in Table 1.6, it is reasonable to hope that remedial techniques applicable to animals will also have some applicability to humans. In the following case, a therapy devised with humans (Wolpe, 1958) is applied to a phobic response in an animal. The case is described by Tuber, Hothersall and Voith (1974) who have proposed a clinical psychology of animals for dealing with spontaneous abnormal behaviours in animals, particularly pets. They illustrate their proposals with cases where psychological principles and procedures derived from normal animal research served to alleviate abnormal animal distress. Higgins is one such case. Table 1.6 Comparative factors predisposing hysterical and motor disorders in humans and captive baboons

Humans

Baboons

Overprotection Psychological trauma Reaction to physical trauma Exhausting physical labour Chilling Requirements of adolescence Imitation Voluntary hysterical symptoms

Social deprivation Conflicting social relations Reaction to aggression Resistance to immobilization Chilling Sexual and social life changes Imitation Conditioned hysterical attacks

After Startsev, V.G. (1976) Primate Models of Human Neurogenic Disorders, Hillsdale,

N.J., Erlbaum.

HIGGINS Higgins is an affable four-year-old English Sheep Dog of Goliath proportions whose tranquil demeanor was breached only by an intense fear of thunderstorms. At the first indication of an impending storm, Higgins would begin an accelerating pattern of aimless pacing, profuse salivation and marked panting which was rapidly climaxed by the hurtling 16

Psychopathology: the status of animals of his 110-pound body against any obstacle in a futile attempt to escape. (Tuber, Hothersall and Voith, 1974, p. 763) Higgins was cured by the method of desensitization by counterconditioning developed from Wolpe's (1958) studies of experimental neurosis in cats. Tuber, Hothersall and Voith prepared a stereophonic reproduction of a thunderstorm and played it to Higgins at increasingly louder intensities as he learned to tolerate each intensity level without any signs of fear. As they describe it: Training was initiated in the laboratory in daily sessions lasting one hour. The beginning intensity for each session was always slightly less than the terminal level achieved during the preceding session .... We had ... progressed from a thunder intensity level of a meager 35 decibels to that of a resounding 75 decibels when a typical summer thunderstorm intervened to test our efforts. Happily, the owner reported that Higgins initially exhibited only a mild version of the original fear. (Tuber, Hothersall and Voith, 1974, pp. 763-4) Additional cases of fears in companion dogs are described by Hothersall and Tuber ( 1979).

Animal contributions to medical science Animal models in physical medicine Several of the cases described above and in later chapters bear marked similarities to human responses to similar situations and incidents - paralysis under extreme fear, destructive behaviour at losing a sexual partner, perceptual disorientation during prolonged isolation, and thunder and lightning phobias. Their study may therefore be profitable not only for the alleviation of unintended animal suffering but also for the light it may shed on human problems, that is, as models of human mental illnesses. In physical medicine, studies with animals for the benefit of humans are commonplace, and Gay (1967) has vigorously argued for further development of comparative medicine for the combined sakes of humans and animals. Leader (1967) illustrates how the study of anthrax and other natural animal diseases have contributed to human and animal welfare, sometimes in unexpected ways, and a document issued by the American Public Health Association (1967) lists benefits to animal and human health derived from animal studies in the areas of nutrition, surgical techniques, drug therapies and vaccine production. Beyond these, 17

DOMAINS OF BIOLOGICAL PSYCHIATRY animal studies contribute to knowledge of the basic physiology of pathological states. Among the nutritional benefits are the identification of the role of cholesterol in cardiovascular disorders, of the relevance of exercise in the prevention of myocardial infarction, and the value of fluoridation and vitamin therapy for deficiency diseases. In fact, except for ascorbic acid, thiamin, niacin, and Vitamin D ... the search for the identification of other vitamins and minerals was motivated largely by work done in agricultural colleges on experimental animals. (American Public Health Association, 1967, p. 1598) Animal experimentation also contributes to increasing the world's food supplies, both from the standpoint of raising food production by attention to farm animal welfare (Kilgour, 1978), and to the need for developing improved nutritional foodstuffs. In this respect: In addition to animal researches in nutrition providing better food for man, they have provided better food for animals. In fact, entire new industries of considerable economic importance have been developed in animal nutrition as, for example, the dog food industry and even a sizable industry providing food for cats (McCann and Stare, 1967, p. 1603) With respect to surgical techniques, Winterscheid ( 1967) details contributions of animal preparations for the understanding of haemorrhagic shock ('a state of depressed consciousness; pale, cool, moist skin; increased respiratory rate, and a thin weak fast pulse') and for cardiac surgery and organ transplants: 'Thus ... children and adults whose lives have been incapacitated by cardiovascular lesions may now ... expect to have such defects corrected.' And concerning drugs, therapeutic actions of antibiotics and sulfonamides for infectious disease, antimalarials and others for parasitic diseases, diuretics for kidney disorder cases, as well as drugs to combat arthritic conditions, which occur in horses, dogs and cats, as well as humans, were all discovered or developed through investigations using laboratory animals (Robinson, 1967).

Animal models in psychological medicine Animals as replacement Perhaps it is in psychiatry that the most dramatic recent advances in drug therapies have occurred. For neuroses and mild emotional disturbances the use of minor tranquilizers has reached a point of 18

Psychopathology: the status of animals

abuse, and animal studies are required to discover the modes of action and addicting potential of such drugs. As for more serious, hospitalized, patients, the mental hospital patient population in the United States has declined by about two-thirds since antipsychotic drugs were introduced. The principal drugs employed in psychiatric institutions are phenothiazines (e.g. chlorpromazine) with schizophrenics, monoamine oxidase inhibitors (e.g. iproniazid) and tricyclic compounds (e.g imipramine) with depression, and lithium salts with mania. In many cases the psychiatric value of such drugs was discovered accidentally, and animal studies are now conducted for the discovery and classification of superior compounds, and also for the determination of the pharmacological basis of psychosis (see Chapter 7). In addition to the above, the establishment of animal models is necessary for the study of side-effects attributed to psychiatric drugs. One such side-effect is tardive dyskinesia, which is a condition of 'protruding, twisting and curling movements of the tongue; pouting, sucking, or twisting lip movements; bulging of the cheeks and various forms of chewing movements' (Tarsy and Baldessarini, 1976, p. 30) that appears after prolonged phenothiazine treatment. To combat the problem, numerous studies of tardive dyskinesia with mental hospital patients have been reported, often with no mention of consent by the patients or by relatives. Such studies violate ethical principles of valid consent (the subject properly understands the ·experiment) and of prohibited subjects (experiments should not be performed on the mentally sick, the aged, or the dying) proposed by Pappworth (1967) for medical experiments with humans, so the problem must be solved through animal experimentation of the kind reported by Weiss and Santelli (1978). These workers employed two monkeys as models for the investigation of tardive dyskinesia and obtained similar results to those obtained clinically with humans. Figure l.l shows one of the monkeys after three months' administration of weekly doses (0.25 mg/kg at first, then 0.50 mg/kg) of haloperidol in fruit juice. With the lower dose, dyskinesias in the forms of tongue and perioral mov'ements and twitching, crouching and writhing appeared after 9 weeks on the drug regime, only to disappear with two further weeks of treatment. When the drug dose was raised, however, as is sometimes the case with humans, the original dyskinesias returned in exaggerated forms. Such effects might be acceptable in severe human clinical conditions that respond only to pharmacological treatment but they are ethically unacceptable for human medical research. With drugs, especially, Pappworth proposes a third ethical principle, the 'principle of previous animal experimentation'. These ethical principles of Pappworth underpin much of the 19

DOMAINS OF BIOLOGICAL PSYCHIATRY

Figure 1.1 Monkey 42 after the 90th weekly dose ofhaloperidol (0.5

mg/kg). During this episode the animal grimaced, yawned, twisted into unusual postures, and, as shown, protruded its tongue as it made incipient chewing movements. Note the slight ptosis ofthe left eyelid and the peculiar clasped position ofthe front paws. (From Weiss, B., and Santelli, S. ( 1979) 'Dyskinesias evoked in monkeys by weekly administrations ofhaloperidol', Science, 200, 799---801, Copyright 1979 by the American Association for the Advancement of Science. Reprinted by permission ofthe author.)

experimental work performed on animal replacement for humans. (Ethical principles for animal experimentation are discussed in Chapter 2.) There are many contributions of experiments with animal replacements for humans in the study of psychopathology, including those on symptomatic equivalences, fundamental proces­ ses, reactions to environmental stresses, and attitudes to treatment. Concerning primates in particular, Mason (1968) provides three distinct perspectives for experiments: from an evolutionary com­ parative standpoint to enlighten ourselves about human disabilities through comparison with animal similarities and differences; for pure research out of direct concern about the animals; and as

20

Psychopathology: the status of animals permissible substitutes when humans cannot ethically be employed for examining simple, specific behavioural functions. From the third perspective, which generalizes Pappworth's principle of prior experimentation, there are certain varieties of experimental procedures possible with lower animals that would not be permitted with humans; screening for dangerous drugs and the creation of 'experimental neurosis' are obvious examples. Others are experiments that have focused on selective breeding for obesity in mice, on susceptibility to ulceration in rats, on nervous instability in dogs, and on addiction to opiates in monkeys. Experiments on decorticate and decerebrate animals for the study of brain functions also cannot be done on humans, nor can many of the experiments in medical laboratories, including some of those described above. In these cases, ethical and rational considerations must be balanced against whatever benefits the experiments may have. These matters are addressed below in Chapter 2.

Animals and mental illness As Henri Ey ( 1964) has said: Le concept de 'Zoopsychiatrie', representerait bien cc scandale logique et moral que certain 'cartesians' seraient enclins a denoncer, si n'etait acceptee depuis longtemps l'idee qu'il y a une 'Psychologic animale' dont I' objet et le psychisme ou, si I' on vent, Ia psychoide animale. Even if normal animal psychology is accepted, not in the literal sense of psyche-ology but in the sense of the scientific study of behaviour, its possible contribution to human psychopathology is not obvious. Nevertheless animals as well as humans display a variety of abnormal behaviours, some of them stressful and disturbing. \Vhen these abnormalities are created experimentally, discovery of the relation of cause to effect is possible in principle, but when behavioural abnormalities occur spontaneously in nature their origins and sustaining sources can only be surmised. One such supposition was the medieval belief in demoniacal possession, and another is the modern belief in mental disease. The question of possession of animals by devils has not been debated for centuries (Evans, 1906), but there is a question of mental illness in animals, and the question is said to be unresolved (Zubin and Hunt, 1967). Resolution of the question depends on what mental illness is intended to mean. ILL~ESS AS CAUSE The question of the existence of mental illness in animals implies its existence in humans, and the answer depends

21

DOMAINS OF BIOLOGICAL PSYCHIATRY on the status of the assertion with humans and the characteristics that differentiate humans from lower animals. Frequently, mental illness is taken to mean the cause of abnormal behaviour. This is the sense in which behavioural abnormalities are diagnosed as symptoms of psychiatric diseases. However, as a cause of abnormal behaviour mental illness, like demoniacal possession, is not an empirical discovery but an assumption embedded in a set of fashionable beliefs. Its existence in either animals or humans cannot be discovered by observation or experiment because it is not a causal agent like a toxin or a germ. It is mistaken to assert that mental illness is the cause of abnormal behaviour, animal or human, because the only evidence of mental illness is ~he behaviour it is supposed to cause. In this sense mental illness does not exist in animals because it does not exist in humans. ILLNESS AS DESCRIPTION However, in the sense of describing or categorizing the disturbing and distressing behaviour frequently seen by clinical psychologists and psychiatrists, mental illness may be said to be demonstrable in humans. In this case it is an effect (clinical picture), not a cause, and psychopathology is a possible synonym for extraordinary distress so long as its status as synonym, not determinant, is not overlooked. In this descriptive usage, mental illness is a private experience of the organism in distress, and can only be inferred by an observer from public verbal or non-verbal behaviour of the disturbed individual. Animals cannot say if they are mentally ill or well but this is not a disadvantage unique to animals, for neither can many humans. Claims by psychiatrically diagnosed psychotics that they are actually sane are often disregarded as mistaken, and pleas of insanity by legally convicted criminals are frequently contested by prosecuting lawyers and expert psychiatric witnesses. Thus mental illness may be attributed with doubt or confidence to animals and humans alike according to how public signs of private distress are interpreted. Qualification of illness in this case by 'mental' serves the useful purpose of separating statistically normal from unusual responses to private events, but the practice is dangerous because the qualification is easily mistaken for a separation of psychical from physical sources of distress. Distress responses with organic origins must be differentiated from those without them, but the difference is not physical-organic versus psychic. Physical and mental illness (in the descriptive sense) both have physical origins; organic damage in the one case and environmental contact with behaviour in the other. All the examples of animal behavioural abnormalities described earlier are physical behaviours established and main-

22

Psychopathology: the status of animals tained by physical environments. They are not instances of illnesses if the mind or pathologies if the psyche; they are instances of disordered behaviours that cou1d have distressing accompaniments. Mental illnesses apply to the accompaniments not to the causes of disordered public behaviours. Finally, the question of mental illness in animals pertains to the status of abnormal animal behaviours as models of mental diseases in humans. In 1960, P.L. Broadhurst thoroughly reviewed early Russian and American descriptions of experimental neurosis and concluded that 'no animal .analogue of neurosis as a disease entity had been demonstrated', a conclusion he repeated in 1973:. Broadhurst's negative answer in 1960 applied to the existence of neurosis as a disease entity, not the existence of behavioural disorders in animals. That answer is not now surprising because 'the terms in which the question was put have altered'. The old terms took a disease model of human neurosis for granted and looked for structural equivalence in animal and human disorders. It now seems that aspirations for structural . equivalence were never realistic, and even though Chamove, Eysenck and Harlow (1972) report comparable personality factors in macaque monkeys and normal humans it does not follow that abnormal personality structures of monkeys and humans will be identical. Even normal apes, which in some ways mentally represent the world like normal humans (Mason, 1976), obviously do not think in human terms, so how could an abnormal thinker like an obsessional neurotic or a paranoid schizophrenic be modelled by an abnormal ape? The answer may be found by way of a functional analysis of the public elements of obsessional and paranoid behaviours. The functional approach differentiates between behavioural disorders and unlawful behaviour, and dictates functional analysis of behavioural disorders on the basis of normal behavioural laws. Etiology, prognosis and therapy are equivalent to the development, maintenance and modification of behaviour, and Ullman and Krasner (1969) claim that: ANIMALS AS MODELS

Abnormal behavior is no different from normal behavior in its development, its maintenance, or the manner in which it may eventually be changed. The difference between normal and abnormal behaviour is not intrinsic: it lies in societal reactions to them. It is more fruitful to ask how a person develops arry belief than how he develops a false belief. The principles of the development of 'proper' beliefs are the same 23

DOMAINS OF BIOLOGICAL PSYCHIATRY as the principles of the development of 'false' beliefs. (Ullman and Krasner, 1969, p. 92) Arguably there are intrinsic differences between some normal and abnormal behaviours, but the argument need not revive the dogma of mental illness in its causal form. The attribution of mental disease to humans has led to a futile search for its nature in humans and also, as Broadhurst (1960) discovered, in animals. The question of causal mental illnesses in animals as models of causal mental illnesses in humans is not a question that is open; it is a question that is mistaken. The question is whether laws of human behaviour are visible in animals, and the answer of modern experimental psychology is 'Yes.' Claims for animal-human similarities in pathological processes or outcomes encompass psychoanalysis, psychiatry and ethology as well as experimental psychology. From ethology, Hinde (1962) analyses the relevance of animal studies to human neuroses in three stages: I Immediate responses to stress; 2 How these responses become habitual and distorted; 3 How they come to occur in new contexts;

and lists as discoveries from ethological and laboratory investigations: approach-avoidance reactions, displacement, sexual inversions, regression and tonic immobility as stress reactions; sensitive periods and early experience as fixation mechanisms; and superstitious behaviour and experimental neurosis as examples of transference to novel contexts. Similarly, from psychiatry, Jones (1971) cites analogies between animal and human behaviours in the areas of stereotypies, aggression, conversion hysteria, attention-seeking hysteria, sexual behaviour, and separation syndromes. Concerning this latter, Jones and Barraclough ( 1978) relate self-aggression (auto-mutilation) observed in mammals raised in partial isolation to the self-mutilation - 'scratching, biting, hair pulling and head banging' - that human children sometimes show. With infants, Bowlby's (1976) list of animal contributions to the hospitalism syndrome is given in Chapter 8. Similarities between animal and human psychologies may be examined with respect to origination and symptomatology, and from functional and structural perspectives. Similarities in functional processes between animal species, at least at the level of the law of effect, are generally acknowledged by psychologists, and B.F. Skinner, in particular, has constructed a psychology of humanity on the basis of experiments with animals. Desmond Morris and Konrad Lorenz have done much the same from an evolutionary

24

Psychopathology: the status of animals point of view, especially stressing animal-human parallels in the areas of aggression, courtship and sex. It is work of this nature that led to the Nobel awards to von Frisch, Lorenz and Tinbergen mentioned· at the beginning of this chapter.

Animals and diagnosis Ellenberger ( 1960) illustrates a value of animal studies for psychiatric diagnosis. He distinguishes between the disease processes for which a human patient may be hospitalized in the first place, and the reactions of the patient to the hospital after admission. These he likens to behaviour syndromes exhibited by animals confined in zoos. Unlike Seligman (1975), who claims a parallel between animal and human depression as a disease process, or Ferster ( 1966), who proposes similarities in environmental causes of retarded behaviours in animals and humans, Ellenberger draws a parallel between animal and human reactions to a common environment - confinement, or loss of individual freedom. Four animal-human parallels in confinement syndromes are suggested. For zoo animals, Ellenberger lists: 1 Captivity trauma, which appears in the form of acute depression, prolonged stupor and refusal to eat. 2 Nestling, or the establishment of the cage as a home, and the demarcation of a personal territory within it. 3 Social competition and frustration, expressed as dominance competitions and age and sex conflict interactions. 4 Emotional deterioration, in the form of repetitive stereotypies like pacing, rocking and swaying, coprophagia and faeces smearing, and signs of apathy and depression. The comparable syndromes in humans are: l Anger, withdrawal and negativism as the human captivity trauma reaction. 2 Nestling, with seating and eating locations as human territories and the hospital ward as the home. 3 Frustration manifested as pettiness and interpersonal jealousies. 4 'Alienization' in the form of infantile regression, bizarre catatonic gestures, aggression, agitation and delusions, and infantile behaviour characteristics as equivalent to emotional deterioration in animals. This parallelism is of value because it differentiates reactions to hospitalization (which would affect a psychiatric diagnosis made in

25

DOMAINS OF BIOLOGICAL PSYCHIATRY an institution) from behavioural manifestations responsible for hospitalization, upon which an original diagnosis would be made. To the extent that diagnosis is a desirable psychiatric enterprise then this is an essential differentiation for the improvement of diagnostic reliability, but more importantly still it reinforces the pleas of many recovered patients for reform in asylum management. From the time of John Perceval, son of a British Prime Minister (Bateson, 1961), to that of the New Zealand novelist, Janet Frame (1961), at least, indignities suffered by mental patients have been vividly but ineffectively described. It would make a fitting irony if the parallel drawn by Ellenberger between animal and human responses to confinement were to succeed in effecting reforms in mental hospital management where autobiographical and fictionalized accounts by articulate writers of patients' humiliations have mostly failed.

Animals and treatment: therapeutics, prosthesis and prevention In physical medicine, treatment by therapy aims to cure an unhealthy organ, whereas prosthetic treatment replaces or bolsters an imperfect organ with an artificial aid. Prosthesis remedies a disability without affecting the disabled organ; therapeuticseliminates disease. Thus, bone-setting and surgery are therapeutics while wooden legs and wheelchairs are prosthetic devices. Supportive drugs like phenothiazines and benzodiazapines used in psychiatry are more prosthetic than they are therapeutic. The medical model of mental illness is almost entirely oriented to therapeutics, so that even critics of the disease model of psychopathology propose only alternative methods for cure. Behaviour therapy is substituted for psychotherapy, but treatment still is aimed at cure. Yet if behaviour pathologies are maintained by abnormal interpersonal transactions it is not the cure of an intrapsychic malfunction that would be the objective of treatment but the prosthetic rearrangement of the disabling transactions. Among experimental psychologists, Lindsley (1964a; 1964b) has emphasized the importance of prosthetic environments in education and geriatrics, and Ayllon (1974) has defended the requirement of total control of a hospital environment for the maintenance of non-psychotic behaviour. The defense was in response to a challenge by Sherwood and Gray (1974) that some classic cases of behaviour modification treated by Ayllon and his colleagues had relapsed. While such a challenge is meaningful in the context of a medical cure model for mental illness, it is out of place with reference to a prosthesis approach to treatment, because only in a proper supporting environment can normal behaviour be main26

Psychopathology: the status

of animals

tained, and normal mental hospitals are arranged to support abnormal, not normal, behaviour, according to Rosenhan (1973) and others. In psychological medicine the dividing line between treatment and prosthesis is a fine one. Consider the hypothetical illustration of meanings of abnormality by Ferguson (1968). The term 'abnormal' in relation to the behavior of domestic birds is susceptible to different usages by husbandrymen, veterinary clinicians, and behaviorists. Thus the commercial egg producer who has learned to expect consistency of performance from commercial laying strains, producing in excess of230 eggs per annum, may insist on classifying as alarmingly abnormal any primitive behavior, such as marked broodiness, which seriously compromises egg production targets. The behaviorist accords greater respect to the biological significance of this normal maternal trait and may be inclined to regard the degree of suppression of broodiness, sought for in commercial layers, as prejudicial to preservation of the species and, therefore, classifiable as abnormal. Similarly, increased exploratory pecking activity ... shown to occur in calcium-deprived birds, is normal by a behaviorist's definition, since it has a sound biological purpose aimed at the eventual recognition of the corrective nutrient and thus the preservation oflife. The same activity, since it represents a deviation from that of birds receiving adequate calcium, would appear to the clinician as an abnormal behavior resulting from the abnormal nutritional status. (Ferguson, 1968,p.190) With an abundant supply of calcium the bird is essentially normal; without it, the bird is a compulsive peeker. From one view the bird is cured by a ration of calcium and relapses when the supply is exhausted; from another, calcium is a piece of prosthetic equipment, like spectacles, that remedies a behavioural disability without repairing a defective organ. At a more complicated behavioural level are the different effects of maternal separation on bonnet and pigtail macaque monkeys. In the bonnets, maternal loss produces minimal behavioural changes, whereas pigtail infants react to separation from their mothers with agitation, stereotyped postures, withdrawal and depression. The differences could be that relative to bonnets, pigtails are more prone to anaclitic depression, but it is more likely that the difference is the result of different mother-infant interactions in the two species. Bonnet mothers are permissive with their infants, who interact freely with other adults; pigtail mothers guard and restrain

27

DOMAINS OF BIOLOGICAL PSYCHIATRY their infants from social intercourse. In the absence of their mothers, pigtail infants are social isolates, but bonnet infants soon find guardian 'aunts'. Such aunts provide bonnet infant monkeys with prosthetic environments that pigtail infants are denied. A simple prosthetic apparatus for a baby rhesus monkey is a piece of soft material. Infant monkeys are difficult to rear in barren metal cages, but when an upright cone covered with terry-towelling is placed in the cage the animals easily survive and spend much of their time clinging to the cone. When given the choice of clinging to a doth-covered cone ('cloth mother') or a bare-wire cone fitted with a feeding bottle ('wire mother'), infant monkeys cling mostly to the cloth mother, visiting the wire mother only briefly to feed. The terry cloth is an important prosthesis for the isolated developing infant monkey, but even more important is its function for preventive psychiatry. Monkeys raised in such a manner do not show the psychiatric disorders of adulthood found in monkeys reared in isolation, a finding that has had ample application to the paediatric care of orphaned human children.

28

2 Animal experiments and animal welfare

Animal suffering The purpose of medical science is to alleviate suffering, human in the case of general medicine and animal in the case of veterinary medicine. Psychiatry is the specialized branch of general medicine that attends to suffering attributable to mental illnesses in humans, but there is no comparable specialization in veterinary medicine. Nevertheless, as we saw in Chapter I, there are numerous animal analogues of human functional disorders, and there are also anomalous animal behaviours that will be listed later in this book. It is necessary to make some assessment of animal suffering in all of these cases, not only as a basis of therapeutics but also for minimizing suffering in experimental psychiatry. Criteria for judging animal suffering

Some animal afflictions are laboratory creations (which raises the ethical questions discussed below), but many occur under accidental circumstances. In farm animals, for instance, tail biting in pigs (Colyer, 1970) and hysteria in hens (Hansen, 1976) are counterproductive results of complex interactions between space, temperature, humidity, nutrition and population density conditions. These and other disturbances in farm animals' behaviour have fostered the scientific study of animal welfare on farms, from which Dawkins ( 1980) has derived six criteria for judging suffering in animals. What are the conditions under which the animals are kept? 2 Are the animals physically healthy? 3 Does the behaviour, physiology and general appearance of the animals differ from that of genetically similar animals in less restricted conditions?

29

DOMAINS OF BIOLOGICAL PSYCHIATRY 4 Is there evidence of severe physiological disturbances? 5 What is the cause of the behavioural differences established under 3? 6 What conditions do the animals prefer? These criteria might be applied to zoos and laboratories and other circumstances wherein animals are subject to human control, although the pica and stereotypies frequently displayed by zoo animals would not constitute suffering by most of Dawkins' criteria, nor would the bizarre behaviour of some chicks that were used to examine the effects of learning and maturation on development. The chicks were taken directly from the incubator to a warm well-lit 60 X 50 X 30 em isolation box where they were watered and fed without ever seeing the keeper or an older chicken. After an isolation of no more than a week [they] ... started passionately chasing flies .... Sometimes they succeeded in catching one, but equally often all their efforts were in vain. As if possessed, they ran behind the fly aimed at and went on chasing as if after a phantom when it escaped them. (Katz, 1937, pp. 21~17, italics added) As with zoo-confined animals, some chicks also emitted stereotyped ritualistic movements uncharacteristic of normally reared chicks. Such side-effects of an experimental procedure, which also occur with isolated humans, serve to alert investigators to contaminating effects in laboratory studies. Although it is impossible to judge if there was suffering by the chicks, there cannot be much doubt in the following cases, which are examples of abnormalities in laboratory animals accidentally caused by routine caretaking procedures unconnected with formal experimentation. As a result of these cases, caretaking routines have been modified to reduce the risk of similar incidents in the future.

Animal distress in unexpected situations Cupid and Dennis were primates used respectively in the laboratories of Tinklepaugh ( 1928) and Ferster ( 1966) for psychological investigations with nothing to do with distress. Both animals, however, exhibited highly distressed behaviour when their normal mates were replaced by substitutes in the course of investigations. Cupid was a young male rhesus macaque monkey originally caged with an older female named Psyche. Some time later Psyche was removed and Topsy, another female, took her place. Cupid's response was not anticipated by Tinklepaugh, who wrote: 30

Animal experiments and animal welfare Cupid eyed her from a far corner of his cage. Topsy strolled nervously about the cage ... with her tail up so that her genitals were exposed to him. After about a minute of this behavior Cupid leaped upon her, seized her by the small of the back and hurled her across the cage. Some months after this, with Psyche, Topsy and another female intermittently in and out of Cupid's cage, on an occasion with only Topsy present, Cupid began 'biting his hind feet much as he had formerly done in play'. On account of this he was taken for medical attention, on the way passing Psyche's cage. He looked back at the cage where Topsy still remained and then toward the cage containing the other two females. Psyche, who had seemed to be much upset by Topsy's presence in Cupid's cage, was now on the side of her cage, shrieking threateningly across at the other female. Suddenly, and with no previous sign of anger or particular emotion, Cupid lurched to the end of his chain and began to bite himself. In a few seconds, he tore huge jagged rents in his ... legs : .. a three inch gash in his hip, ripped his scrotum open ... and mutilated the end of his tail. (p. 230) A comparable, though less dramatic, event occurred with Ferster's chimpanzee, Dennis. It happened during a long-term experiment involving Dennis and a female companion in a large complex multiunit experimental space. At one point the female was replaced by another, whereupon Dennis pummelled, kicked and pushed her, and for several weeks went partially off his food. 'We could,' Ferster observed, 'have described Dennis as angry and depressed, and we would not have been too far off the mark' (Ferster, 1966, p. 55). It is not difficult to find reasons for agreeing with Ferster's belief in Dennis's suffering, and for thinking that Cupid suffered also. These reasons are by analogy with ourselves: analogy by response, as when an animal squeals or limps; and analogy by stimulus, when we imagine ourselves in the same situation. We might confidently employ these analogies in assessing suffering in Dennis and Cupid, but there are many opportunities for error when the animals, like Katz's 'possessed' chicks, resemble humans less closely than do chimpanzees or monkeys (Dawkins, 1980). Other examples of spontaneous behavioural signs of distress in primates kept for experimental purposes are reported by Startsev (1976). The incidents involved the removal of some juvenile hamadryas baboons from an indoor to an outdoor larger cage, where the juveniles were exposed to threats from adult males in an 31

DOMAINS OF BIOLOGICAL PSYCHIATRY adjoining enclosure. Two animals 'showed a gait disturbance characterized by incoordination and a posture with the knees half flexed' which in one, the case of Zagreb described in Chapter 1, resulted in death with no evidence of organic impairment. Startsev ( 1976) describes another case. A hysterical motor disorder was produced in the baboon Azov under similar circumstances. He was transferred from a small home cage to a larger cage, which contained a female and was located within view of several other mature males. He had shown brief spasmodic attacks previously under the stress of being driven from one side of the home cage to the other; but now he developed frequent and prolonged convulsive attacks with paresis and paralysis of quite a different character. (Startsev, 1976, p. 135) The disorders of Cupid, Dennis and Azov were not results of experimental procedures as such, but of incidental care and maintenance procedures. Similar problems can arise when cages and cagemates are routinely rearranged in zoos (Stout and Snyder, 1969). The case of the isolation-reared chicks is different, however, because it is an example of an experimentally obtained effect, even though the effect was not a pre-planned aspect of the experiment. The classic prototypical accounts of experimental neurosis discussed in Chapter 5 are similar, for the experiments of Yerofeeva and Shenger-Krestovnikova were about stimulus generalization and discrimination, not attempts to create animal models of human neuroses. Laboratory accidents, household accidents, and traumatic events in the wild can all produce similar reactions in animals. The case of the wild chimpanzee, Merlin, was described in Chapter 1, and that of the home-reared chimpanzee, Lucy, appears in Chapter 8. Both cases show disturbances of posture that Mason (1968) includes in the primate deprivation syndrome and that are identifiable in human autistic children. A major characteristic of such children is head-banging, a behaviour that occurred in a male rhesus monkey· following a laboratory misadventure (Levison, 1970). When the subject was I year old the experimenter began training him to enter a transfer cage .... However ... the door of the transfer cage accidentally dropped on the subject, glancing off his head and shoulders ... his response was to race away from the transfer cage and crouch in the left rear corner of the cage, where he sat huddled in the corner. After this the introduction of the transfer cage i'nto his home cage

32

Animal experiments and animal welfare

was correlated with ... refusal to enter ... crouching, rocking and headbanging. He would sit in the left rear corner of the cage and rhythmically bang his head against the plastic wall of the cage. Chapter 8 contains a further account of the circumstances of this animal which, were it human, would be a clear candidate for psychiatric or psychological treatment. Apart from farm, zoo and experimental animals, ordinary household pets also show spontaneous behavioural disorders. The case of Higgins w:as described in Chapter 1. Hothersall and Tuber ( 1979) give other examples, including fear of thunderstorms by a German shepherd called Cindy and a Labrador retriever named Major. In Major's case there was a possible origin for the fear because an arc welder once exploded on a bench where he was chained, but in the other case, no ready explanation of the fear was available. Cindy was a gentle 4-year-old German Shepherd, who was acutely afraid of sudden loud noises and of thunderstorms. Long before the storm became apparent Cindy would begin to pant, whine and pace. As the storm became imminent, she became increasingly disturbed and would begin to discharge from the nose and mouth; each thunderclap would elicit strong and uncontrollable trembling [ending in] a final collapse into a flaccid trancelike state after which she would remain completely unresponsive for periods up to 24 hours. (Hothersall and Tuber, 1979, p. 24 7) These illustrations are sufficient to suggest, at least, that lower animals can experience suffering parallel to that of humans, and that the situations that occasion it are similar in both cases. They document the need for psychiatric investigations of animals not only for their possible relevance to the understanding of human aberrant behaviour but also for their practical possibilities for alleviating animal suffering that stems from intentional or unintentional sources. Naturally, such investigations cannot proceed on the basis of nineteenth century anthropocentric psychology, in which 'it is no worse from an ethical point of view to flay the forearm of an ape or lacerate the leg of a dog than to rip open the sleeve of a coat or mend a pair of pantaloons' (Evans, 1898, p. 99). Nowadays this is an unthinkable opinion and considerable attention is paid to the questions of ethical (Marcuse and P(;ar, 1979; Sechzer, 1981) and humane (Russell and Burch, 1959) procedures in experiments with animal subjects. At the heart of these questions is the one of correspondence in animal and human natures.

33

DOMAINS OF BIOLOGICAL PSYCHIATRY

Ethics and animal experimentation Animal nature and human nature

If human psychopathology is to be studied by way of animal psychopathology, then some correspondence between animal and human natures must be evident. A few animal-human similarities have been illustrated above, and elaborate generalizations from animal to human nature in the areas of territoriality, aggression and sex are claimed in popular works by Ardrey (1978), Lorenz (1966) and Morris (1967). Walker (1982) has convincingly documented by way of comparative brain anatomy the steps from animal to human thought, and Thorpe (1974) shows several ways in which animal natures are similar to those of humans - animals too can learn, plan ahead, form concepts, use tools, communicate with each other, count, and show visual and musical appreciation. Griffin ( 1976) uses several of these as criteria of consciousness in animals, and Harlow, Gluck and Suomi (1972) give examples of intellectual, motivational and pathological generalizations between human and non-human animals. 'There is', they claim, 'only one way to test the limits of interspecies generalizations and that is by experiment.' At the same time, however, the authors recognize that evaluation of such generalizations 'may be more an art than a science.' The essence of the art is identified by Frey (1976). Generally speaking, a statement about similarity reflects an experimenter's claim that investigated phenomena have common characteristics. The term similarity also indicates that the experimenter has observed characteristics in which these phenomena differed. In classifying phenomena as being similar, however, the experimenter implies that he considers the differences he observed as negligible. (Frey, 1976, p. 7) For some, the difference between humans and animals is not negligible. Bannister ( 1981), for instance, arguing that animal experimentation in psychology is a fallacy, acknowledges that animal-human similarities exist but cites economic and ethical reasons as the principal justifications for animal studies. He attacks reductionism and naive realism as the bases of scientific psychology, and rejects the image of psychology as a science under the influence of Darwinian biology. To Bannister, psychology as a study of behaviour is the study of man as an animal. Against this he stresses the uniqueness of humankind and argues for psychology to be the study of experience - the study of human experience alone. This_itselfis an economic argument if resources for the study of both are unavailable, and is anthropocentric in ignoring the 34

Animal experiments and animal welfare possibilities, first, that a psychology of animal experience could exist and, second, that experimental psychologists could be interested in animal suffering. Beyond that, as the study of human experience, psychology is improperly and unnecessarily restricted. Experience can only be communicated by means of public behaviour, so at least psychology must concern itself with human behaviour as well as with human experience, and human experience is normally expressed through verbal behaviour acquired via the medium of inherited gestures, some of which humans and animals share (Darwin, 1872). Human nature is anything but clearly understood on the basis of experiment or of self-examination by humans. In a recent symposium organized by the British Psychological Society on Models of Man (Chapman and Jones, 1980), Jahoda asks the question, 'One model of man or many?' and answers, 'I have come to only one firm conclusion: a rejection of the idea of one unitary model of man for psychology. I do not believe that one should or even could strive for it.' And by the same token, there is no reason to demand a unitary animal model to cover all varieties of human psychopathology, or to assume that animal psychology cannot be generalized to humans. The examples I have given already, strongly suggest that it can.

Scientific and moral judgment Apart from arguments against animal experiments in psychology on the grounds of dogma there are also ethical reasons given against experimentation with animals. These usually consist of reasons for not conducting experiments, but they may take the more positive form of guidelines or directives for the employment of certain experimental precautions. Precautions and guidelines are necessary because non-human animals cannot control their degree of participation in experiments and so must be protected by ethical principles and procedures adopted by experimenters. Thus there are two kinds of questions involved in the use of animal subjects in psychiatric or psychological research for the benefit of humans. One pertains to the validity of the results for humans - the rational, or result generalization, reason; the other pertains to the ethics of data collection - what can be done experimentally with animals that cannot be done with humans. The ethical question arises only if the rational question is answered affirmatively first, for if the results do not generalize then there is no reason to conduct the experiment. If the rational question is answered affirmatively then the ethical question only arises when there is 35

DOMAINS OF BIOLOGICAL PSYCHIATRY opposition between equally desirable values; the need not to cause suffering in animals and the need not to allow suffering in humans. The essence of the ethical problem can be expressed by what I call Singer's contradiction and its Wisconsin converse.

SINGER'S CONTRADICTION The researcher's central contradiction exists in an especially acute form in psychology: either the animal is not like us, in which case there is no reason for performing the experiment; or else the animal is like us, in which case we ought not to perform an experiment on the animal which would be considered outrageous if performed on one of us. (Singer, 1975, p. 49) THE WISCONSIN CONVERSE There is only one way to test the limits ofinterspecies generalizations and that is by experimentation .... If nonhuman data do not generalize to data derived from human beings, can human data be used to predict within reason the supposedly homologous behaviour of nonhuman animals? (Harlow, Gluck and Suomi, 1972, p. 709) How, that is, can we know how much an animal is like us without experimentation, and how, without knowing how much an animal is like us, can we judge when the animal is suffering? By analogy with ourselves, argues Dawkins (1980), and the analogies are discovered by experiment. Singer's contradiction and the Wisconsin converse argue logical dilemmas. They differ from the personal dilemma of Davis (1981) who laments that he must sometimes. conduct experiments with aversive stimulation of animals that he likes to keep as pets. He writes: I have run experiments using aversive stimuli since 1966. In both my training and my own research the use of animals has been axiomatic. On the other hand, I have an abiding fondness for animals: not simply for conventional pets, but for the kind of animals that one is likely to use as an experimental psychologist. Over the past five years I've trapped, fed and released as many as 200 mice who managed to find their way int9 my house. I've had the rare opportunity to rear a red squirrel from weanling to mature adult, and currently have a domesticated pet rat who, on most days, has the run of the house. (Davis, 1981, p. 63) At the risk of exaggeration, the Davis dilemma might be compared to that of a hangman. Bassford (1981), explaining role 36

Animal experiments and animal welfare differentiation in the ethics of psychological research, begins: If you or I were to put a noose around someone's neck and then hang them, we would rightly be considered moral monsters. But if the royal hangman were to do this while performing his official duties, his action would not be morally culpable. Indeed, until recent times, it would be morally laudable. (Bassford, 1981, p.27) Thus, as a scientist in an officially approved educational institution, Davis could appeal to a duty assigned to him by society as a reason for conducting an experiment that might be questionable to him morally as a private citizen. (In this respect, Reed ( 1981) makes a useful distinction between personal moral standards and codified ethical principles: 'Moral behavior is the practice of virtue . . . . Ethics is concerned with abstract moral principles and their codification.') So Davis could resolve his personal dilemma by changing his profession from psychology to law or moral philosophy, just as the hangman is free to change his. However, this still leaves society with the same dilemma as it faces with hangmen: does it require practitioners of the profession, or does it not? And if so, to what extent? The answer cannot come from illustrations of experimental successes or failures, for these are post- not pre-experimental events, but from the relative orders of priority that society gives to competing values. Diamond (1981) expresses the Davis dilemma in the context of two extreme views on animal experimentation. The context is broader than Davis's lament, which concerns itself only with the infliction of pain, for Diamond considers the justification for using animals in experiments at all. According to the first view: Within certain limits, experimental animals may be regarded as delicate instruments, or as analogous to them, and are to be used efficiently and tared for properly, but no more than that is demanded. (Diamond, 1981 p. 341) By this view, the standards of welfare for animals would be based on scientific judgment and scientific judgment alone. Possible cruelty to animals in experiments would be controlled by the disapproval of scientific colleagues on the one hand and the likelihood of producing unreliable experimental results on the other. Decisions would be based on the acquisition of knowledge as the ultimately desirable value for cultural survival. But, says Diamond, there is a second view, that: Within certain limits, animals may be regarded as sources of

37

DOMAINS OF BIOLOGICAL PSYCHIATRY moral claims. These claims arise from their capacity for an independent life, or perhaps from their sentience, but in either case the moral position of animals is seen as having analogies with that ofhuman beings. (Diamond, 1981, p. 341) The second view is based not on scientific judgment but on moral judgment, for which the scientist qua scientist has no more expertise than anyone else. Criteria for moral judgments rest on contributions to the welfare of the powerless, concern for others, about which scientific judgments are in principle neutral. These first and second views, scientific and moral judgments, present the dilemma facing experimenters using animals, for between them the cultural values of desire for knowledge and concern about others are likely to come into conflict (Marcuse and Pear, 1979). The category of judgment, scientific versus moral, establishes how the views differ, but, Diamond argues, there are matters common to both views hidden in the phrase 'within certain limits.' These are as follows. AN ANIMAL'S LIFE IS LESS IMPORTANT THAN A HUMAN LIFE

Neither view is opposed to this and neither view would take the extreme position of nineteenth-century anthropocentric psychology. Less dramatically, however, a group of psychology students surveyed by Keehn ( 1982) demanded less strict requirements for experiments with animals than for experiments in which human subjects were involved, but the differences were not large. On a 5point scale the statement 'The subject must be protected against any foreseeable injury' was rated 4.9 on average for humans and 4.0 for animals, and the largest human-animal difference was 4.4 for humans and 2.9 for animals on the statement, 'Drugs with unproven effects must not be employed in the experiment.' The sample that produced these values may not be representative of society at large, but it is unlikely that any group of individuals would reverse the order of stringency between animal and human experimental subjects' rights. Perhaps the fairest modern expression of what the 'certain limits' are in the general population is Dawkins' remarks that: Governments are under pressure to change the laws on the treatment of animals. Scientists are on the defensive over their experiments on animals. Farmers are criticized. But most people go on eating animals, demanding that the products that they eat or wear are tested, wanting better drugs or transplants or vaccines to save their lives. (Dawkins, 1980, p. 11) 38

Animal experiments and animal welfare ANIMALS

HAVE

MORAL

CLAIMS

IN

SOME

CIRCUMSTANCES

Whereas the first area of agreement between view one and view two may represent a softening of the moralist's extremism of banning all use of animals for human ends, the second region of agreement represents a retreat from the extremism of nineteenth century anthropocentric psychology. The retreat takes two independent lines, one differentiating moral obligations to animals in experiments from moral obligations in normal life, the other distinguishing between types of experiments and the authorizations required for their performance. In the first of these cases, Table 2.1 shows four hypothetical individuals, A who concedes no moral obligations to animals at all, B who accepts moral obligations in normal life but not in experiments, C who insists that moral obligations are indispensable always, and D who would deny moral obligations to animals in normal life but treat them with respect in the course of experiments. Of these imaginary individuals, A would be the absolute representative of anthropocentric psychology, B could be the animal-lover but uncompromising scientist, C is the champion of animal rights in life and in experiments, and D is the scientist who only respects animals as instruments for research. In absolute categories, A, B, C, and D are ghosts, but they represent compass points from which the 'some circumstances' under which animal experiments can be performed can be approached. Table 2.1 Categories

of moral obligations

In normal life

NO

YES

NO

A

B

YES

D

c

In Experiments

Concerning the case where different experiments require different kinds of authorization, Ross (1978) describes a procedure adopted by Sweden. By this procedure, which is summarized in Table 2.2, research laboratories and institutions must establish ethics committees composed of five each of research scientists, laboratory technicians and lay persons. The committee must be

39

DOMAINS OF BIOLOGICAL PSYCHIATRY Table 2.2 Categories of experiments and their control in Sweden

Experiments requiring committee notification

Experiments requiring committee approval

l No pain involved

4 Like 3 but with post-operative pain

2 Anaesthetized animals not revived after experimentation

5 Experiments expected to cause illness in anaesthetized animals

3 Anaesthetized animals revived without post-operative pain

6 Experiments using curare to cause immobility without anaesthesia

Adapted from Ross, M. W. ( 1978) 'The ethics of animal experimentation; control in practice', Australian Psychologist, 13, 3 75-8.

notified of experiments in categories l, 2 and 3 in Table 2.2 but, in addition, its authorization is also required before an experiment in categories 4 through 6 can be performed. In that case, a scientific or medical experiment on animals cannot, in principle, be conducted on the basis of scientific judgment alone. This is no guarantee of expert moral judgment, however, for qualifications of 'lay persons' in this respect may be no more than their qualifications as scientists. So who in the last analysis is the final moral judge? Lane-Petter ( 1976) provides an uncompromising answer. Not one of us can shrug off matters ... such as this ... we have a right and a duty to challenge the decisions that are made on our behalf. What the individual does not have the right to do is condemn, or refuse to listen to, those who have made moral or ethical judgments that differ from his own. (Lane-Petter, 1976, p. 121) Whoever makes it, the ethical decision is frequently in the end a cost-benefit affair in which, by the nature of experimentation, the benefits are hypothetical while the costs are crystal clear. Were the results of an experiment to be fore-known the performance of the experiment would be a waste; and were the procedures of the experiment not disclosed the results of the experiment would be of no use. So resolution of the ethical dilemma necessarily involves balancing the unknown against the known, which is why there cannot be a categorization of permissible experiments after the

40

Animal experiments and animal welfare manner of Table 2.2 based on the value of experimental results. It might seem that such a table could be constructed according to the practical potential of expected results (Seligman, 1975), but this is a retrogressive criterion because the power of scientific discoveries is to create practical possibilities, not to follow them. Had this criterion been adopted earlier, much applied medical and psychiatric research would never have begun, and its human beneficiaries would not be around to defend it. In any event, as a factor in an ethical equation the dangers of the practicality criterion outweigh its gains. Experiments on the effects of high altitudes on the human body, on effects of freezing on warm-blooded creatures, on treatment for injury caused by mustard gas, on the efficacy of sulfonamides in treating gas gangrene war wounds, on factors affecting bone, muscle and nerve regeneration, on effects of drinking sea-water, and on possible vaccines for typhus, were all conducted on German concentration camp prisoners in World War II, and were all defended on the practical grounds of likely benefits to other humans - high-altitude flyers, shipwrecked sailors and wounded soldiers in the Nazi forces (Cohen, 1953). Those experiments were unethical in principle, regardless of their outcomes, practical or otherwise. As Pappworth ( 1967) points out, an experiment is ethical or not before it begins, not after it is over. 'Morality', he claims, 'rests on what is right in itself ... not on justification by result, even though that may possibly benefit a great many others' (p. 185). If we knew absolutely what is right in itself there would never be doubts about what is moral and ethical, but unfortunately, as Marcuse and Pear ( 1979) point out, that which a society calls right is that which it values most for its own survival. In the case of experiments that might do harm to animals, where the desire for knowledge and the importance of kindness to animals are especially valued, the practical question becomes that of formulating principles of humane experimentation, not of prejudging the practicability of likely outcomes or results. It is widely judged that an animal's life is less important than a human's and it is widely agreed that animals have the rights to human protection. It is the responsibility of scientists to preserve these rights to the limits of their abilities through the adoption of the most humane experimental procedures available at any time.

Humane experimental procedures Ethical questions only arise in cases of imbalance of power and conflicts of interests. With laboratory animals, the imbalance of power is between the laboratory subject and the human experi41

DOMAINS OF BIOLOGICAL PSYCHIATRY menter, and the conflict of interest occurs when the animal is put in pain, discomfort or stress. These conditions are relatively common in experiments relevant to psychiatry, as in studies with aversive stimulation, infant-mother separation or social isolation rearing, because psychiatry specifically exists for individuals in distress. Such stressful conditions are not only found in laboratories, however. In farm management, lambs and calves are artificially reared, and painful electrical devices are used for fencing or for forcing cows to defecate in pre-arranged locations (Kiley-Worthington, 1977). For farmers, guidance for improvement of the welfare of their animals appears in connection with efficiency and productivity (MacDonald and Dawkins, 1981); guidance for experimenters is offered in connection with the humane arrangement of experiments, specifically with the replacement or reduction of animal subject populations and the refinement of experimental procedures (Russell and Burch, 1959). Replacement involves either the use of less sentient for higher animals in experiments likely to cause distress, or the employment of alternatives to live animal experiments. Both of these raise the problem of generalizability of results that was discussed above, and each has its use in different particular areas. Replacement of animals in biomedical research is discussed by Smyth (1978), who notes that the principal alternatives - bioassay isotope tracing, chemical analysis by chromatography, and computer simulationwere all started as the purest of fundamental research, not in the remotest way related to the alleviation of human suffering. Some of these substitutes for animals may be useable for testing biological mechanisms of psychiatric drugs or for toxicity screening, but substitution of lower for higher animals in behavioural research only complicates the question of generalizing the results to humans. Although it is a laudable objective, it is unlikely that inanimate materials can soon substitute entirely for living animals as replacements for humans. Where research is on psychiatric -problems of living, living organisms must necessarily be employed, and where the living organisms cannot be human for ethical or practical reasons, animal alternatives have to be found. If animals cannot be replaced by inanimate materials in behavioural experiments to do with stress, at least the number of subjects employed can be kept to a minimum (Hume, 1957). The reduction of the number of subjects does not in itself make an unethical experiment ethical, for by Pappworth's (1967) principle of equality in human experimentation 'if it is unethical to submit many to a proposed experiment, it is equally unethical to expose only one person'. But reduction in subject numbers is a humane experimental procedure that can be combined with other refine-

42

Animal experiments and animal welfare

ments in procedures for animal experimentation. Among such refinements, Russell and Burch ( 1959) recommend selective breeding of experimental animal subjects to minimize genetic contributions to variability in experimental data; This might serve some useful purposes, although it cannot substitute for improvements in experimental techniques and procedures. Psychology became an experimental science at a time when its capabilities of controlling a stimulus exceeded its capacity to control a response, so that precise manipulations of independent variables (stimuli) frequently produced imprecise dependent variable (response) data. The result of this was that experimental psychologists accepted experimental error (uncontrollable variance) as an inevitable characteristic of their discipline, and resorted to the employment of large subject populations to randomize individual variability and differentiate it from the true measures under investigation. Psychologists were forced into statistical control of error and opted to make descriptive and inferential statistics the cornerstones for the design of experiments. This option is still widely employed in psychology, with the result that many experiments use large numbers of subjects just for the production of stable averaged data. However, this is no longer necessary because, with the precise control of behaviour obtainable with modern operant experimental techniques, error can be minimized by experimental instead of statistical refinement. In that case animals in experiments are employed not as groups for statistical control of inevitable experimental error, but as individuals, one by one, for the refinement and replication of experimental findings. When replications are made systematically by testing the generality of results under different conditions in successive experiments, fewer animals are necessary than when one experiment exactly replicates another. Thus by systematic replication the reliability and validity of experimental results can be assessed with the minimum number of experimental animals. Just as experimental refinement can reduce the number of animal subjects in a particular study that are exposed to stress, so it can reduce the amount of stress to which each animal is exposed. In one common psychological procedure for assessing conditioned fear, for example, Davis and Wright (1979) found that a wide range of shock intensities was employed by different experimenters in arriving at much the same conclusions. On the other hand, with two similar procedures for studying shock avoidance learning by rats, Keehn ( 1967) found that the animals quickly learned to run in a wheel and avoid the majority of shocks, but failed to avoid the same level of mild intensity shock by pressing a bar. Thus if avoidance learning is a model for human anxiety neurosis (Levis,

43

DOMAINS OF BIOLOGICAL PSYCHIATRY 1979) then it can be studied more humanely in rats with the running than with the bar-pressing response. Better still is the suggestion of Russell and Burch ( 1959) to look for a natural stressor for the animal concerned. In the screening of psychiatric drugs, for example, they illustrate how anxiety and fear-reducing agents can be assessed by observations of naturally opposed flight and courtship behaviours in fishes and birds, rather than by arbitrary laboratory stressors. In practice, Ross (1981) has suggested the use of control procedures similar to those for Swedish experiments set out in Table 2.2 for behavioural research with higher animals. He recommends that wherever experiments with higher animals are proposed there should be a fifteen-member committee made up of five animal care technicians, five scientists and five other individuals. The committee should be readily available to make rapid decisions, and should be charged with two prime responsibilities: seeing that animals are not subjected to unnecessary harm or use; and seeing that research is designed to employ the fewest number of animals to the greatest advantage. A categorization of behavioural experiments in order of stressfulness to animal subjects proposed by Ross (1981) is summarised in Table 2.3, where studies in categories 4, 5, and 6 would automatically require full committee approval, while a research proposal falling in a lower Table 2.3 Categories of experiments on higher animals

I

Painless experiments

2

Some distress is involved, including experimentation with nonoptimum rearing conditions

3

Psychopharmacological trials and experiments employing aversive stimulation or non-laboratory reared animals

4

Aversive conditioning experiments and minimally painful experiments, CNS lesions or electrode placements which will cause minimal pain

5

Unavoidable pain, painful CNS lesions or stimulation, isolation studies

6

Long-term stress leading to animal neuroses and psychoses

Adapted from Ross, M.W. (1981) 'The ethics of experiments on higher animals', in Keehn, J.D. (ed.), The Ethics of Psychological Research, Oxford, Pergamon.

44

Animal experiments and animal welfare

category could be authorized by fewer than all members of the committee. As Ross (1981) concludes: The fact ,that such a research proposal has to be considered by such a committee frequently in and of itself has the advantage of sharpening the experimental design which is provided, and ... the ethics of humane experimental technique may be administered without going to the extreme of either having a rigid set of rules or imposing control from outside the scientific community. (Ross, 1981, p. 59) Given the size and significance of mental illness in human populations, it is essential for every effort to be made to bring the situation under control. Among these efforts is that which concentrates on the creation of animal models of the human illnesses. It is imperative that these creations do not themselves create more distress than the ones they are intended to eliminate, and it is encouraging to observe that experimental and administrative procedures are moving inexorably towards this end, fostered jointly thougp independently by the scientific and animal-loving communities.

45

Part II

Animal clinical pictures

3 Abnonnal movements and convulsions

Stereotypies and bizarre postures Definitio!ls and examples Stereotyped behaviours appear in autistiC children, retarded children, psychotic adults and amphetamine addicts. In animals, stereotypies arc induced by confinement, social and sensory deprivation, reinforcement schedules, and drugs. With humans, stereotypy pertains to symptomatology and diagnosis; with animals it relates primarily to origination. Animals emit stereotyped behaviour in a variety of situations, of which pacing by zoo animals and rocking by a wild chimpanzee were described in Chapter l. Other examples come from animals in laboratories and on farms. Kiley-Worthington (1977) observes that dogs, cats and horses develop stereotypies more frequently than cattle or deer, and lists variations of the kinds of situations in which stereotypies occur; restrictions on movement and sensory stimulation, confrontation with novelty, and conditioning. She defines a stereotypy as 'an aberrant behaviour repeated with monotonous regularity and fixed in all details'. However, not all stereotypies need be aberrant. So-called ethological stereotypies are fixed action patterns typical of a species. In the rat, for example, Barnett ( 1963) lists respiration, locomotion, ingestion, gnawing, hoarding, grooming, crawling under, fighting, coitus, parturition, nursing, retrieving and nest-building as stereotyped activities. These are not atypical activities of an aberrant individual but typical behaviour patterns of species members under specifiable conditions. Nevertheless, typical action patterns may occur atypically according to the form, frequency, consequence and mode of elicitation of the behaviour. Stereotypies often appear to be purposeless, but as Kiley-Worthington (1977) asserts, purpose is in the eye of the beholder and is best omitted as a defining

49

ANIMAL CLINICAL PICTURES characteristic of aberrant stereotypies. Even the characteristics of stereotypies included in KileyWorthington's definition are not applicable to stereotypies of all kinds. Monotonous regularity is a particular characteristic of stereotypies induced by restrictions on territory or freedom of movement, but stereotypies that result from deprivation rearing are as likely to appear as non-repetitive stereotyped postures as they are to take the form of monotonous repetitive movements. Likewise the imperative that stereotypies are fixed in all their details is not universally applicable. Amphetamine-induced stereotypy, which is widely employed as a bridge to link animal responses to chronic amphetamine administration with secondary schizophrenic symptomatology, takes more the form of 'fragmented actions' than it does of monotonous movements. We must beware, then, of fitting all stereotypies into a single universal mould, and recognize their existence in several forms, not all of which are aberrant behaviours (e.g., ethological stereotypies). The particular categories of aberrant stereotypies I propose to describe are: 1 Cage stereotypies, 2 Deprivation stereotypies, 3 Aberrant postures, 4 Fragmented actions.

Originations in animals Stereotyped behaviours have been differentiated according to form and according to origin. Differentiation by form is usually made with humans (see below), and the major categories are repetitive and non-repetitive movements. With animals, Berkson ( 1967) distinguishes between cage stereotypies and deprivation stereotypies according to the source of the stereotypy, and Robbins (1982) has introduced a category offragmerited actions to describe repetitive stereotypies induced by stimulant drugs. CAGE STEREOTYPIES Sources of cage stereotypies may include food frustration (Kiley-Worthington, 1977), imminence of feeding time, or thwarting of natural flight reaction when danger signals appear (Meyer-Holzapfel, 1968). Pre-feeding stereotypies may be conditioned 'superstitious' responses (Skinner, 1948), or variants of sign-tracking behaviours described below. Keiper ( 1970) describes such a case of 'spot pecking' in caged canaries. This stereotypy is not affected by cage size, in contrast to a stereotyped 'route tracing' movement, which is. Laboratory studies with monkeys show that cage stereotypies

50

Abnormal movements and convulsions

are responsive to the size of the cages the animals are kept in (Draper and Bernstein, 1963) aml. also to the quality of the animal's environment (Berkson, Mason and Saxon, 1963). Draper and Bernstein (1963) studied three male and nine female wild-born 3-year old rhesus monkeys in three outdoor cages of different sizes: small (3ft by 3 ft by 3 ft); medium (4 ft by 3 ft by 8 ft); and large (48ft by 24ft by 8ft). Several categories of behaviour in the different cages are listed in Table 3.1, which shows that most stereotyped behaviour occurred in the small cage, that significantly less (p < 0.01) occurred in the medium cage, and that there was no stereotyped behaviour in the large cage at all. Table 3.1 Most frequently observed stereorypical behaviours of twelve adolescent rhesus monkrys as function of size of holding cage

Small cage

Medium cage

Self-directed activities

Stereotypies: bouncing pacing twirling jumping somersaulting (backwards)

Large cage

hanging sitting

grooming self-clasping self-biting genitalia manipulation

Cage manipulation: shaking biting After Draper, W.A., and Bernstein, I.S. (1963) 'Stereotyped behavior and cage size', Perceptual and Motor Skills, 16,231-4.

The stereotypies, which differed from animal to animal, took the forms of: rapid bouncing on the floor with all four feet, bouncing using only the front legs, predictable circular pacing, pacing with a head thrust at regular intervals, regular pacing and recoiling from one corner of the cage, rapid pacing developing into an exceedingly fast spin or twirl on the hind legs in the. center of the cage, twirling holding onto the roof, backwards somersaults, unique awkward vertical jumping, and touching one leg to a particular place on the side of the cage as the animal travelled in a fixed pattern. (Draper and Bernstein, 1963). 51

ANIMAL CLINICAL PICTURES Some of these behaviours, particularly awkward vertical jumping, are plainly adaptations to the tinyness of the small cage, and resemble the distorted intention movements of buntings kept in cages with perches too close to the ceiling (Hinde, 1962). The forms of other stereotypies, while responsive to cage size, were not specifically determined· by the cage. For example, one female described by Draper and Bernstein showed continuous backward somersaults in the small cage, 'regular pacing that involved throwing up the forelegs and tossing back the head as if to begin the somersault', but without completing it, in the medium cage, and no sign of somersaulting in the large cage. Thus, somersaulting did not occur when it was easily possible, and occurred most when other movements were hampered. In a set of four experiments, Berkson, Mason and Saxon (1963) observed variations in the stereotyped behaviour of four male and two female laboratory-reared adult chimpanzees in situations that differed in novelty, available space and opportunities for alternative activities. Five classes of behaviour (repetitive stereotypies, non-repetitive stereotypies, manipulation of the environment, selfmanipulation, and locomotion) were recorded in an outdoor 39 ft by 57 ft home enclosure, a 69 in by 72 in by 85 in barred cage, and an 81 in by 79 in by 84 in enclosed wooden cubicle. The repetitive stereotypies observed were mostly rocking and swaying, but included also head nodding and shaking, and twirling. Nonrepetitive stereotypies were abnormal limb postures, lip contortions, eye poking and thumb sucking. Self-manipulation occurred in the forms of scratching and rubbing. Most stereotypy, particularly of the repetitive kind, occurred in the cubicle, which was the smallest and the most isolated with respect to social and sensory communication, of the three environments. In a comparison of the cubicle and home enclosure environments, when manipulatable objects in the forms of a broom handle, a clothes line and a piece of burlap were and were not available, the availability of the objects was found to reduce non-repetitive stereotypies and self-manipulation by statistically significant amounts. Davenport and Menzel ( 1963) compared stereotyped behaviours of sixteen chimpanzees raised from birth in various kinds of restricted environments at the Yerkes Laboratories with three wildborn chimpanzees brought to the laboratory at between about 4 and 7 months of age. The wild-born animals were kept together in a large open cage enriched with toys and exercise equipment, but the laboratory-born animals were housed individually and separately, except for two pairs whose individual cages were separated only by bars. Observations were made until the chimpanzees were

52

Abnormal movements and convulsions over 40 months old, but almost no stereotyping was seen in the wild-born animals. By contrast, fhree classes of stereotypies were t;xhibited by the deprived chimpanzees: rhythmical rocking, swaying or body-pivoting; repetitive movements of head, hand or lips; and posturing in an awkward position. Seven animals that were observed daily from birth to 21 months showed four major stereotypies that emerged at different ages: swaying, rocking, pivoting and thumb-sucking. The times of occurrence of these behaviours are indicated for each animal in Table 3.2.

Table 3.2 Month of onset and number of months up to 28 months of age in which each of seven chimpanzees raised individually in captiviry was observed to displqy swaying, rocking, pivoting and thumbsucking stereorypies.

Stereorypy

Sway

Animal

No. No. No. No. No. No. No.

200 169 171 190 173 196 188

Mean onset Mean number

Rock

Pivot

Onset

Number

Onset

Number

8 17 11 13 9 10 8

16 4 10

8 12 8 17 5 9 12

5 1 8 1 19 1 6

3 3

15 19

11.0

3 1 4 5 4 5 6

Number

Onset

Number

15 18 17 18 25 15 lO

2 5 2 3 2 3 2

7 2 15 10 4 4 12

4.0

10.0 10.0

Onset

Suck

7.2

2.7 17.0

7.7

Compiled from Davenport, R.K., and Menzel, E.W. (1963) 'Stereotyped behavior of the infant chimpanzee', Archives of General Psychiatry, 8, 99-104.

53

ANIMAL CLINICAL PICTURES Table 3.3 Frequencies rif oraliry, disturbance and aggression stereorypies exhibited by eighry-jour socially isolated rhesus monkeys under passive and stimulated conditions (see text for details)

Passive frequency

Active frequency

Oraliry: Sucking orality: Toe sucking Other digit sucking Self-sucking Thumb sucking

1472 196 141 50

684 84 36 72

Chewing orality: Chewing Cage biting Nail biting

3044 440 166

2795 52

Other orality: Self-licking Cage licking Finger in mouth Mouth rubbing Tongue pulling

395 238 12 3 I

24 4 2 2 0

Disturbance Vocalization Grimacing Head lowering and leg clasping Rocking Self-clutching Convulsive jerking

553 50 464 358 96 48

1879 1946 524 708 528 248

Aggression Externally directed: Threat Cage shaking

162 146

2507 382

Self-directed: Displaced threat Self-biting Teeth grinding Head slapping

220 294 40 13

1188

Category of stereorypy

10

1092

472 156

Condensed from Cross, H.A., and Harlow, H.F. (1965) 'Prolonged and progressive effects of partial isolation on the behavior of macaque monkeys',joumal of Experimental Research on Personality, I, 39--49.

54

Abnormal movements and convulsions DEPRIVATION STEREOTYPIES Stereotyped activities also occur in laboratory-reared monkeys that are socially isolated either as a matter of routine care and maintenance or for specific experimental purposes. Table 3.3 lists stereotypies observed by Cross and Harlow ( 1965) under passive and stimulated conditions of observation. They studied orality, disturbance, and aggression categories of stereotypies in eighty-four socially-isolated rhesus macaques ranging from l to 7 years of age. The table shows the number of times that each listed stereotypy occurred among all of the animals during ninety half-minute observations per animal over a period of three weeks. In the passive condition, observations were made unobtrusively; in the stimulated condition a threat stimulus in the form of a large black glove, usually used for handling the animals, was slowly moved across each animal's cage while it was under observation. Behaviours that differed significantly according to sex and age are marked in Table 3.4. Among the abnormal behaviours observed by Cross and Harlow ( 1965) were stereotyped pacing and back-flipping, self-clasping, finger-sucking and body-rocking, which are typical stereotypies induced by captivity and confinement. Other observations, of withdrawn-activity, self-biting, hair-pulling, self-mutilation and apparent aissociation, are more characteristic of social deprivation. As well as with monkeys and chimpanzees, infant social and sensory deprivation has been studied extensively in rats, mice and dogs (Beach and Jaynes, 1954). Mostly these studies have been statistical probes for critical periods in the development of social and problem-solving capabilities (Scott, 1962), but occasional reports of specific aberrant behaviours are available. In one, Melzack and Scott (1957) described how 9-month-old Scottish terrier puppies raised for seven months in almost total sensory isolation repeatedly burned themselves on lighted matches held before their noses. The dogs 'moved their noses into the flame as soon as it was presented, after which the head or whole body jerked away ... but then they came right back to their original position and hovered excitably near the flame.' In another study, Thompson, Melzack and Scott (1956) report 'whirling behavior' in eight of eleven other puppies raised under similar conditions. The whirling stereotypy consisted of 'very rapid, jerky running in a tight circle' accompanied by yelping, barking, snarling and tailbiting sometimes lasting as long as ten minutes. ABERRANT POSTURES Aberrant bizarre postures such as fixed staring, rigidity, awkward mobility and huddling do not qualify as stereotypies in the monotonous repetitive sense, but they are not uncommon in human clinical types in which repetitive stereotypies

55

ANIMAL CLINICAL PICTURES Table 3.4 Stereotypy categories showing statistically significant age and sex differences by eightyj'our socially isolated rhesus monkeys

Greater in:

Category

Older

0 rality Sucking orality: Toe sucking

Chewing orality: Chewing Cage biting Nail biting

Males

Females

X X X

Other orality: Self-licking disturbance Vocalizing Grimacing Head lowering and leg clasping Rocking Self-clutching Convulsive jerking Over-all disturbance aggression Externally direct.ed: Threat

Younger

Greater in:

X

X X X X X X X

X X

X

X

X X

X X X X

Self-directed: Displaced threat Self-biting Teeth grinding Head slapping Over-all self-aggression

X

Condensed from Cross, H.A., and Harlow, H.F. (1965) 'Prolonged and progressive effects of partial isolation on the behavior of macaque monkeys',joumal of Experimental Research on Personality, I, 39-49.

also appear. In an animal, Merlin (p. 13) is an illustration of the co-existence of a bizarre posture (hanging upside-down) and a repetitive stereotypy (rocking). 56

Abnormal movements and convulsions The most common origination of such aberrations in laboratory animals is social deprivation or inadequate rearing, and Mason ( 1968) lists the first of four characteristics of a primate deprivation syndrome: abnormal postures and movements; motivational disturbances; poor motor integration; and deficiencies in social communication. A vivid illustration of bizarre posturing in a monkey raised in isolation is a case recounted by Mitchell (1970). The animal could also be exhibiting a delusion, a hallucination or dissociation. One ... male slowly moved his right arm toward his head while in a rigid seated pose and, upon seeing his own approaching hand, suddenly appeared startled by it. His eyes slowly widened and he would at time fear grimace toward, threaten, or even bite the hand .... If he did not look directly at the hand or did not bite it, 'it' would continue to move toward him .... As 'it' approached him, his eyes became wider and wider until the hand was clasping his face. There he would sit for a second or two, with saucer-sized eyes staring in terror between clutching fingers. (Mitchell, 1970, p. 228) Social or sensory isolation is not the only condition that occasions bizarre posturing in animals. The co-occurrences of aberrant postures and repetitive stereotypies is reported by Davenport and Menzel ( 1963) in non-isolate laboratory-reared chimpanzees. These authors categorise rhythmic whole-bo4J movements (rocking, swaying, somersaulting, chest pounding), part-body movements (head banging, thumb and toe sucking, hand clasping) and posturing (head rolling, staring at the hand before the eyes) all as stereotypies. They report that these related not only to rearing but also to developmental status as summarized above in Table 3.2. For Davenport and Menzel (1963), as for Meyer-Holzapfel (1968) in connection with zoo animals, space restriction is the most striking contribution to stereotypies and bizarre posturing in animals. In such cases, as with animals on farms (KileyWorthington, 1977), scientific discoveries are put to use in improving the care and welfare of animals as well as in providing clues for improving the lot of humans. FRAGMENTED ACTIONS Robbins ( 1982) coined the term fragmented actions to explain a theory that he and Lyon (Lyon and Robbins, 1975) had proposed to account for the action of amphetamine-like drugs, and to explain the correspondence between certain behavioural results of abuse of these drugs and schizophrenia.

57

ANIMAL CLINICAL PICTURES The theory stems from three observations. The first is that in humans amphetamine overdose produces a clinical picture that can be confused with paranoid schizophrenia (Connell, 1958). The second is that stereotypy is a salient secondary characteristic of schizophrenia (Bleuler, 1950; see below). And the third is that animals given high doses of amphetamine and similar drugs exhibit perseverative stereotyped responses. These responses are not the characteristic pacings and rockings of caged or socially deprived animals but stereotyped intrusions into behaviour sequences that, if completed, would lead to a satisfactory end. That is, animals trained to perform a sequence A-B-C to secure food under ordinary circumstances might, after continuous amphetamine injections, perseverate in early portions of the sequence and emit AAAABAAABBABC instead. The stereotypies need not be productive parts of the behaviour chain but species-specific druginduced competing responses, like sniffing and headshaking in the rat, that stop the chain part way and start it over from the beginning. At low doses amphetamine behaves as a stimulant by elevating gross motor activity in all parts of the sequence, but as drug use and dose increase 'the increasing acceleration of behavioral initiation results in an increasing repetition within a decreasing number of response categories' (Lyon and Nielson, 1979, p. II 0). This is behavioural fragmentation. The response categories into which amphetamine-induced stereotypies fall are dose-dependent. As Randrup and Munkvad (1975) report them for the rat: l mg/kg of d-amphetamine given subcutaneously to rats thus produces selective stimulation of sniffing, locomotion and rearing while there is little grooming activity .... Gradual increase of the amphetamine dose in the intervall-10 mg/kg leads first to a further decrease and eventually to disappearance of grooming activity; then locomotion and rearing are also decreased and finally disappear so that only sniffing remains. At 10 mg/kg the maximal stereotypy is reached. Sniffing, often accompanied by licking and biting, is performed continuously and usually covers only a small area at or near the bottom of the cage. (Randrup and Munkvad, 1975, p. 759) This account of amphetamine-induced stereotypies points to dosedependent selective suppression of ethological stereotypes in the rat (Barnett, 1963, see p. 49), 'so that only sniffing remains'. This is the stereotypy to which Robbins (1982) attributes maximum behavioural fragmentation in that animal. Amphetamine-induced stereotypies reported for other species are looking from side to side 58

Abnormal movements and convulsions

in cats, circling or running back and forth in dogs, twittering and posturing in chicks, biting and chewing in the tortoise and the lizard, and pecking in several species of birds (Randrup and Munkvad, 1975). Going from rodents to humans, stimulant-drug stereotypies take more and more complex forms. Randrup and Munkvad (1975) classify these forms as ethological, operant, emotional and social stereotypies, with 'hung up' repetitious human thoughts and acts as the most complex. A study with forty-six cats by Ellinwood, Sudilovsky and Nelson (1972) assessed movements, postures and 'attitudes to the environment' as a function of methamphetamine exposure and dose. Concerning movements and postures, they recorded head-neck, shoulder-foreleg, hip-hindleg, tail and trunk bodily regions, and assessed overall coordination and synchrony between the regions. They found that over eleven days of chronic methamphetamine administration, movements in the respective regions became less coordinated and more dysynchronous. The dysynchronies in movement, where, for example, frozen hindlegs occurred concurrently with hyperactivity in head and forelegs, eventually led to a disjunctive posture wherein 'the active forelegs at times would back up against the relatively resistant hind legs to produce the hunch or camel-back phenomenon'. Beyond that, Ellinwood et al. continue: 'it appears as if the cat had forgotten where a leg is positioned ... [it] would remain in an awkward disjunctive position while the cat goes about other activities.' The 'attitudes to the environment' category included awareness, indifference, normal and abnormal interest, normal and abnormal investigative behaviour, reactivity, and normal and abnormal focus (apprehensively 'hooked'), all subjectively assessed on the basis of an animal's bodily posture in relation to its surroundings. In this category Ellinwood et at. record that methamphetamine induced a compulsive investigative fixated interest by the animals to a constricted sector of the environment. All in all, Ellinwood et al. (1972) argued that their amphetamineintoxicated cats show many resemblances to human catatonics, and conclude that: Following chronic amphetamine intoxication, components of behavior became relatively fixed over time and showed a loss of cohesive flow among different initiatives with their relative priorities. In addition ... there appeared to be islands of separate organization, each establishing its own autonomy or anarchy without integration into the larger behavioral symphony. (Ellinwood et al., 1972, pp. 227-8) The fragmented actions that accompany amphetamine intoxication 59

ANIMAL CLINICAL PICTURES in animals and man are offered by many investigators as evidence for an amphetaf'line model of human schizophrenia founded on the distribution of catecholamine neurotransmitters in the brain. Comparative evidence in support of the model from acute, progressive and residual effects of high-dose administration of stimulant drugs with rats, cats, monkeys and humans is summarized in Table 3.5 (Ellinwood and Kilbey, 1977). There are, however, critics of the model (Kokkinidis and Anisman, 1980).

Symptomatology in humans Four classes of humans exhibit stereotyped behaviours: retardates, autistic children, schizophrenics, and amphetamine addicts. In these cases the origination of the disorder is a matter of conjecture, and the behaviours are employed as criteria for differential diagnosis rather than as indicators of deficiencies in conditions of living. In these humans, aberrations that in animals are attributed to short- or long-term environmental circumstances are more usually attributed to personality defect, organic defect, or mental illness. Stereotypies in humans are more complex and varied than those that appear in animals. RETARDATES Berkson and his associates (cf. Berkson, 1967) report extensive studies of stereotyped movements in mentally defective humans. In children and adults they identify repetitive movements of body (swaying, rocking, twirling), head (shaking, nodding, banging), face (tics, grimaces), and hands (fiddling), and non-repetitive postures and self-manipulation (rubbing, scratching, poking, biting and picking). With respect to repetitive movements, some individuals appear mostly as rockers and others as fiddlers. These behaviours also appear in normal children but they do not persist. With the retardates, general arousal level and opportunities for competing behaviours affect the prevalence of stereotypies although, as Berkson (1967) concludes, the effects are ephemeral. Stereotypies in retardates have been studied extensively for therapeutic and institutional ward management purposes, and have been shown to respond to sensory and social environments. Luiselli (1975), for instance, reports on an institutionalized 14year-old boy with a 50--60 IQ range who engaged in frequent rhythmic rocking wherever he happened to be sitting. The boy's rocking was systematically observed and counted on a number of daily occasions (baseline) and was then subjected to a number of treatment modes. In the first, the boy received social praise, popcorn and candy when he was sitting still without rocking, while rocking was ignored; in the second, the attendant walked off and

60

Abnormal movements and convulsions Table 3.5 Characteristic acute, chronic and residual effects of high-dose administration of stimulants to humans (clinical reports) and animals (experimental reports).

Acute efficts

Stereotypy

Human

.Movement

Attitude

Suspicious

Repetitive movements Compulsive acts

Monkey

Visual scanning

Cat

Head movements

Dysjunctive

Investigatory

Rat

Sniffing Licking Gnawing

Restricted

Chronic effects

Stereotypy

~Movement

Attitude

Human

Oral dyskinesias Facial tics

Choreiform movements Catatonic immobility

Paranoia Delusions ofparasititis

Monkey

Oral dyskinesias Constricted, bizarre Self-grooming

Dystonic posture

Reactive Socially inadequate

Cat

Paw shaking Head shaking

Akathesia Dystonic post~re Ataxia

Hyper~reactive

Rat

Jerky Increased intensity

Backward walking Jumping

Reactive

Residual iffects

Human

Low dose thresholds for induction of psychosis or dyskinesias.

Monkey

One or two doses reinstate chronic end-state behaviours.

Cat

Low dose reinstatement, conditioned behaviour to setting.

Rat

Persistent augmented response to usual dose.

Adapted from Ellinwood, E.H., and Kilbey, M.M. (1977) 'Chronic stimulant intoxication models of psychosis', in I. Hanin and E. Usdin (eds), Animal Models in Psychiatry and Neurology, New York, Pergamon Press.

61

ANIMAL CLINICAL PICTURES ignored the boy at the onset of rocking and returned with praise and candy when the rocking stopped. In the third treatment condition proper sitting continued to be socially reinforced, but the boy was sent to stand in a corner for 3-minute time-out periods whenever he rocked; that is, reinforcement and punishment procedures were combined. The maximum 'therapeutic' effect occurred with the last procedure, and suppression of rocking was maintained in subsequent baseline and time-out plus reinforcement phases. As with animal stereotypies, stereotypies of retarded humans are amenable to environmental manipulation, although whereas with animals the manipulations are for analytical and experimental purposes, for humans they are for the purposes of treatment and ins ti tu tional management. AUTISTIC CHILDREN Early infantile autism is a syndrome identified by Kanner ( 1943) and distinguished from childhood schizophrenia on the basis of early age of onset, resistance to change in the environment, and intellectual-type personalities in the parents. Others have proposed different criteria of autism (see Chapter 8), and the syndrome is not reliably diagnosed. However, bizarre gestures and stereotyped repetitive movements are uncontested characteristics of autistic children. Epstein,- Doke, Sajwaj, Sorrell and Rimmer (1974) describe vacant staring and grimaces, rocking, swaying, and object twirling as characteristics of autistic children, and classify autistic stereotypies as:

1 inappropriate foot movements Uumping, hopping, running); 2 inappropriate hand movements (pounding, spinning, flapping, rubbing); and 3 inappropriate vocalizations (echolalia, talking to self, mumbling). Interpretations of these behaviours usually stress self-stimulation and self-reinforcement. They are generally unresponsive to environmental control, but some stereotypies, such as head-banging, can be suppressed by contingent punishment (Lovaas, Schaeffer and Simmons, 1965), and others may be maintained by social reinforcement (cf. Lucy, p. 50). A little 4-year-old girl, for instance, whom I shall call Dorothy Rocker, was hospitalized for infantile autism on account of the usual criteria (see Chapter 8). She performed an occasional repetitive stereotypy of crouching on all fours and making pelvic thrusts of an unmistakable sexual nature. The behaviour was selected for analysis, and for brief daily periods she was placed in a room with some toys and observed

62

Abnormal movements and convulsions through a one-way mirror. Her stereotypy began soon after she entered the room, and it was duly counted, but it soon became evident that, as in the case of the chimpanzee, Lucy, the stereotypy was situation-dependent: Dorothy constantly looked toward the mirror while she was rocking, and as soon as the door opened at the end of the session she stopped rocking and ran into the hall. In such events it seems preferable to analyse autistic stereotypies in humans in the way they are analysed in lower animals - case by case on an individual basis rather than in terms of a general characteristic of a clinical syndrome of autism. AMPHETAMINE ABUSERS Soon after World War II reports from Japan drew attention to instances of 'incomprehensible, odd and very unnatural movements' that are 'constantly, identically and energetically repeated' by amphetamine abusers (Randrup and Munkvad, 1967). In similar vein, Randrup and Munkvad describe cases in Sweden where abuse of an amphetamine-like compound, phenmetrazine, produced 'compulsive or automatic continuation for hours of one aimless activity, such as sorting objects in a handbag, manipulating the interiors of a watch, polishing fingernails to the point that sores are produced, etc.' Other reported illustrations of amphetamine-induced stereotypies in humans are teeth-grinding, chewing, lip movements and 'hung-up' repetition of simple thoughts. Similar symptoms appear in schizophrenic patients (see below), which has caused misdiagnosis of amphetamine intoxication as paranoia (Connell, 1958). From this observation has developed the amphetamine model of schizophrenia. SCHIZOPHRENICS Many of the aberrant behaviours described above for autistic children also appear in clinical accounts of schizophrenic adults. Tics, pacing, repetition of words and movements are characteristics of schizophrenics, as also are stereotyped thinking and echolalia. Aberrant postures like adoption of a foetal position or imitation of a Christ-like posture with arms outstretched as if on a cross are textbook illustrations of schizophrenic behaviours, and Bleuler (1950) includes stereotyped postures and repetitive movements as among the striking secondary characteristics of schizophrenia. The correspondence between amphetamine-originated stereotypies and symptomatogical stereotypies in psychosis is the basis of the dopamine theory of schizophrenia reviewed by Silverstone and Turner ( 1982). This theory addresses the problem of schizophrenia indirectly by noting first the equivalence (in part) of symptomatologies of unknown origin (clinical stereotypies) and of known origin

63

ANIMAL CLINICAL PICTURES (amphetamine stereotypies), and second, the role of dopamine in the action of amphetamine in the brain and central nervous system. The second is the step expected to expose the biochemical bases of human psychoses with animal models such as those described in Chapter 7.

Abnormal fixations and seizures Compulsion and insoluble problems IN IMPOSSIBLE DISCRIMINATIONS Although well known in farm animals for ages, reports of stereotypies in laboratory animals are relatively recent, and consideration of these animal behaviours as models of human psychopathology is an even later occurrence. In the case of so-called experimental neurosis, however, while farm animals were among the early subjects of investigation (Liddell, 1944), psychological interest quickly turned to the laboratory rat (Finger, 1944). At the time, behavioural disorders were primarily attributed to conflict, and conflict became the focus of laboratory studies with the rat. Such studies did not generate an acceptable animal model of human neurosis, but they nevertheless did generate behavioural abnormalities in the rat worthy of independent investigation. So-called neurotic disablement caused by an insoluble problem was discovered inadvertently by N.R.F. Maier (1949) in the course of investigations of discrimination learning by rats with an apparatus devised by K.S. Lashley. In this apparatus, subjects must jump from a platform through the one of two doors that hides a food reward. The incorrect door is locked and mistakes give the animals a bumped nose and a short fall into a net. When the doors are discriminable rats jump promptly to whichever side is correct on a training trial, but when discrimination is impossible they show reluctance to jump and, when pushed fixate their responses on a single side. These responses are called 'abnormal fixations', although under the circumstances they are the most efficient reactions the rats can make. When jumps to either side are reinforced half the time at random, the rate of reinforcement is maximized if all the jumps are made to one of the sides. In that way half the jumps are reinforced for certain, a fraction that is heatable only by accident with random jumps. The reinforcement schedule for stereotyped 'fixated' left or right jumping is variable ratio two meaning that on average every second response secures reinforcement.

64

Abnormal movements and convulsions Impossible discriminations were originally conceived as insoluble problems, not as reinforcement schedules, and the rats' fixations were attributed to frustration-induced disablement. Some animals seemed to be upset by the experience, but, Maier (1949) says, 'the fixated group develops some kind of adjustment to the test situation and is therefore able to prevent emotional tensions'. With soluble after insoluble discriminations, fixated rats continue to make stereotyped position choices, but they jump differently to each of the discriminable stimuli: 'When the rat jumps to the rewarding window, it does it head-on, and when it jumps to the punishing window, its long axis becomes parallel to the window so as to avoid a painful blow' (Liberson, 1967). The fixated rat, like the neurotic pati:ent, seems to make its responses against its wili, which testifies to the power of the varible ratio schedule of positive reinforcement as much as to conflict-induced frustration. IN REINFORCEMENT SCHEDULES A monkey trained by Findley and Brady ( 1965) showed even more compulsive-looking behaviour than that exhibited by Maier's rats. Hour upon hour the monkey sat pushing a button, a compulsion with no seeming effect. However occasional responses turned on a light-bulb and after 40,000 presses the monkey received a meal. The behaviour did not develop by accident but was carefully trained with schedules of fixed ratio reinforcement that got longer and longer day by day. The light served as a conditioned reinforcer and originally signalled a meal, but with training, fewer and fewer light onsets were followed by feeding till finally the ratio was I in 10. By the end, it took 4,000 responses to produce a light-flash, and after every lOth light the monkey was fed. A diagnosis of compulsive pushomania is an appealing evocative description of the monkey's behaviour, but it does not categorize the monkey, nor does it indicate how the 'pusho' is caused. Even a patient observer, seeing the monkey only after button-pushing developed, might never discern the origin of the animal's compulsion or how it was being maintained by a fixed ratio 40,000 schedule of reinforcement. He could hardly be unimpressed by the amount of energy expended by the animal for so very little reason. IN A SOCIAL SITUATION Calhoun (1967) describes a compulsive behaviour that also has an element of sadism, although its genesis is perfectly normal. The experiment was one of social cooperation among rats. In one group (COOP) two rats were required to cooperate in operating a water dispenser before either rat could drink; in another group (DISOP) no rat could drink if more than one approached the dispenser at a time. Calhoun describes one rat

65

ANIMAL CLINICAL PICTURES from the DISOP group that kept climbing a barrier separating the pens of the two groups. He would enter and approach the lever [of the COOP group water dispenser, and] one of the COOP rats would come over and enter the other side. To this invading DISOP male, his COOP companion's behavior was all wrong. He would immediately back out, grasp the 'offending' COOP rat by the tail or hind feet and pull him out .... During the following weeks he macerated the tails and hind feet of all the COOP rats. Most lost all their toes. Seven died from these wounds. And yet the invading male was never attacked. To the COOP rats this invading DISOP male was always behaving correctly ... and their ethical standards dictated that they came to his rescue. (Calhoun, 1967, p. 20). All these illustrations of compulsive behaviours in animals have parallels in human counter-productive compulsive actions. The animal cases may not have the full quality of the human 'symptom' but they do have the counter-advantage of illuminating originating situations: random reinforcement, ratio reinforcement, and a change in the social reinforcing context. These kinds of situations are not uncommon in human experience and they undoubtedly contribute to human compulsive behaviour. Such behaviours are not necessarily 'neurotic', however; that depends on the consequences of the act on society, and the reaction of society to the actor.

Superstition and reinforcement schedules Experimental distress or disablement can be operantly conditioned by suitable reinforcement schedules. These are rules about contingencies between specified responses and environmental events like delivery of food or electric shock. Operant conditioning with a variety of schedules of reinforcement is described by Ferster and Skinner (1957). The formal procedure involves individual animals working in isolated 'Skinner boxes' with reinforcers programmed for specified responses on scheduled occasions, such as every other response or the first response in every minute. In addition to those mentioned in the section above, several schedules of reinforcement have disabled animals or caused them to behave strangely or with maladaptive emotion. The following are the principal examples: 1 Fixed time, in which reinforcers are delivered at short predetermined intervals independently of an animal's behaviour. 66

Abnormal movements and convulsions 2 Fixed or variable ratio, in which reinforcers are delivered after an animal makes a regular or irregular number of responses. 3 Fixed or variable interval, in which reinforcers are scheduled for the first response that is emitted after regular or irregular time intervals. 4 Multiple schedules, in which two or more simple schedules alternate such that reinforcement density differs in the various components, or such that punishment and positive reinforcement are mixed. 5 Conditioned emotional response schedules, in which periodic predictable or unpredictable electric shocks are imposed upon other reinforcement schedules, usually variable interval reinforcement with food. 6 Escape schedules, in which responses terminate aversive events like electric shocks, loud noises or bright lights. 7 Avoidance schedules, in which animals can avoid or postpone aversive events that would otherwise recur with or without warning. These schedules may maintain abnormal behaviour either by indirect induction of emotional behaviours, (schedule-induced behaviours, see Chapter 4), by deliberate reinforcement of specific unusual responses, or by accidental response-reinforcement contingenci'es (superstitious behaviours). Laboratory-conditioned operants are usually prosaic efficient responses, like bar presses by rats and monkeys or key-pecks by pigeons, but Skinner ( 1948) once conditioned bizarre helplesslooking behaviour in pigeons by delivering reinforcers at short regular intervals regardless of what the birds were doing. The schedule is one of fixed-time reiriforcement and Skinner describes the ensuing behaviour as superstitious. He writes; One bird was conditioned to turn counter-clockwise about the cage, making two or three turns between reinforcements. Another repeatedly thrust its head into one of the upper corners of the cage. A third developed a 'tossing' response, as if placing its head beneath an invisible bar and lifting it repeatedly. Two birds developed a pendulum motion of the head and body, in which the head was extended forward and swung from right to left with a sharp movement followed by a somewhat slower return. The body generally followed the movement and a few steps might be taken when it was extensive. Another bird was conditioned to make incomplete pecking or brushing movements directed toward but not touching the floor. None of these responses appeared in any noticeable strength during adaptation to the cage or until the

67

ANIMAL CLINICAL PICTURES food hopper was periodically presented. (Skinner, 1948, p. 168) Although the experimental subjects seem to behave strangely in comparison to other birds, their oddities are not neurotic oddities in the usual sense. They are specific oddities that Skinner claims are learned, and the origination of the oddities in the experiment seems clear. However, the superstitions may not look as weird to other pigeons as they do to human experimenters, because hungry pigeons waiting for food often do the things that Skinner observed.

Audiogenic seizures IN BEHAVIOUR-GENETIC ANALYSIS Audiogenic seizures in rats and mice have been elicited by hissing airblasts, buzzers, highpitched tones, bells, and jangling keys. Finger ( 1944) characterizes the audiogenic seizure as a three-part phenomenon in the rat; a fore-period followed first by an active and then by a passive phase. In the fore-period, after an initial startle to the noise, the rat exhibits crouching, burrowing, and a state of heightened sensitivity to noise. Sometimes, especially when a seizure does not occur, the rat performs 'substitute activities' of nose rubbing, ear scratching, teeth chattering, grooming and yawning; otherwise there are pivoting movements of the head and body, jerkiness and short quick runs that lead to a full-blown active seizure. In thi~ active phase there occurs an explosive bout of running after which the animal falls down and emits a series of rapid clonic twitches. There may also be spasmodic bursts of running and jumping along with ejaculation, defecation and squealing. Occasionally with young animals this phase ends in death, otherwise the animal enters the third, passive phase. In this phase the animal remains comatose with no spontaneous activity and depression of normal reflexes for a period from two to ten minutes. After this, Finger and Schlosberg (1941) report, subnormal activity may persist for another twelve hours. However, not all rats exhibit the phenomenon, and a particular rat may not convulse on every exposure to the stimulating noise. With a high-pitched whistle as the eliciting stimulus, Auer and Smith ( 1940) reported repeated convulsions in 30 per cent of a group of over 400 rats. Some quantitative features of typical convulsions of ten randomly selected animals are shown in Table 3.6. The table summarizes latencies and durations of convulsive running, and durations of jumping and rigidity during the periods of stimulus onset shown in the right-hand column. Audiogenic seizures have served animal experimentalists as a

68

Abnormal movements and convulsions Table 3.6 Quantitative characteristics of the different phases of the convulsive pattern produced in rats by a high-pitched stimulus whistle. Data are from a random selection of ten rats from seventy-five convulsive animals. Stimulation was usually terminated during the convulsive jumping stage, so durations of this stage are arbitrary (time given in seconds)

Subject

1 2 3 4 5 6 7 8 9 10

Latency of convulsive running

Duration of convulsive running

44 50 46 45 5 7 35 5 8

15 40 14 7 43 29 23 18

10

Il

10

Duration of rigid state

12 43 29 25 5 8 33 21 20 17

Duration of convulsive Duration of jumping stimulus

175 147 70 125 168 155 100 160 !55 330

246 278 289 220 213 239 213 2II 400 373

From Auer, E.T., and Smith, K.U. (1940) 'Characteristics of epileptoid convulsive reactions produced in rats by auditory stimulation', journal of Comparative Psychology,

30, 255-9.

convenient phenotype for the study of behavioural genetics (Fuller, 1979), and also as models for epilepsy and the alcohol withdrawal syndrome in man. Concerning behaviour-genetic analysis, Fuller ( 1979) distinguishes between two kinds of phenotypes: somatophenes, to do with physical structures; and psychophenes, to do with behavioural processes. The psychophenes he divides into an ostensible class of observed behaviours and an inferred class of generalized tendencies or traits. The audiogenic seizure is a member of the ostensible psychophene class inasmuch as it is an observable, countable behaviour. Particular genotypes need not have identical phenotypes. Fuller points out that inbred strains of mice, for instance, of identical genotype may have the same pigmentation but differ in size, growth and ability to learn. The colouration he calls phenostable, the others, phenolabile. Phenolabile traits or characteristics have a reaction norm and a reaction range which are dependent on the nurturing environments of the group and of the individual. Thus

69

ANIMAL CLINICAL PICTURES the location of a phenotype on the dimension of phenostabilityphenolability will determine the degree to which that phenotype is amenable to genetic and environmental control. Detailed behaviour-genetic analysis of seizures with rats have yet to be reported, but the general role of heredity in audiogenic seizures has been known almost from the discovery of the phenomenon in that species. Thus Maier and Glaser (1940) found 74 per cent, 52 per cent and 0 per cent of 3-month-old offsprings of seizure susceptible-susceptible, susceptible-normal, and normalnormal crosses, respectively, to be classifiable as seizure-susceptible. Likewise there are strain differences in susceptibility to seizures among mice, where the susceptibility is higher among DBA/2] strain members than among members of the BALB/CJ and C57BL/6J strains (Fuller, 1979). REFLEX EPILEPSY Extensive studies of audiogenic seizures in rats selectively bred for susceptibility to seizures are reported by Krushinskii (1962), who coined the term 'reflex epilepsy' in 1949. Krushinskii and his colleagues conceive of audiogenic seizures according to the Pavlovian conception of higher nervous activity as interactions between excitatory and inhibitory processes (see Chapter 5), and they employed, among other things, inhibitory and excitatory drugs to manipulate the inhibitory-excitatory cortical balance in their subjects. They used a loud-pitched noise (14--16 kHzs; 70--112 db) to engender convulsions, which were scored on a five-point scale from zero for no reaction to 4.0 for a complete tonic convulsion within one and a half to two minutes after noise onset. With sodium bromide to strengthen inhibitory cerebraJ processes, Krushinskii reports a weakening of convulsions with increasing bromide dose. Conversely, with caffeine and strychnine, employed to increase excitatory processes, Krushinskii reports a statistically significant shortening of latencies of convulsions in comparison to normal non-drug measures. On the basis of such findings, Krushinskii recommends the audiogenic seizure response of rats as an experimental preparation for the rapid evaluation of drugs newly synthesized for the treatment of epilepsy. IN ALCOHOL WITHDRAWAL The alcohol withdrawal syndrome is described in Chapter 6, where correspondences in the syndromes of rats, dogs, primates and humans are detailed. In animals, the stages and complexities of the syndrome are less definable than they are in humans, but they all include 'tremors, hyperexcitability and convu_lsions in the last degree. In alcohol-withdrawn rats, audiogenic convulsions have been elicited by the noise of jangling

70

Abnormal movements and convulsions keys. Falk, Samson and Winger (1972) offer this as evidence of physical dependence on alcohol by rats made polydipsic by the spaced-feeding technique (sec Chapter 4, p. 81). Their animals were fed small food pellets at 2-minute intervals for 1 hour periods every 3 hours, day and night, with only alcohol to drink. After about 2 months, four animals were subjected to tests for audiogenic seizures 3 or 4 hours after alcohol was withdrawn. Falk et al. ( 1972) give this description of the result. A shaking of keys near the top of the cage for 1 to 2 seconds resulted in a tonic-clonic convulsion in rat No. 8. For the next hour, tremors, spasticity, and clonic head movements occurred, and finally, a second seizure ended in death. When keys were shaken (2 to 5 seconds) for the first time after 9! hours of withdrawal, a clonic running episode was produced in rat No. 2, followed shortly by death from a tonic--clonic seizure. Rat No. 7 showed all the preconvulsive symptoms, but keys shaken (up to 20 seconds) after 15 hours of withdrawal had no effect. (Falk et al., 1972, p. 813) The fourth rat was spared the key-shaking test, which had no effect on normal rats of the same strain that had not been subjected to the alcohol intake training and withdrawal routine beforehand.

The kindling if.fect AS A MODEL OF EPILEPSY The kindling effect refers to the gradual appearance of stereotypies (behavioural automata) and convulsions in rats,- monkeys and cats after a number of brief low-intensity electrical stimulations of the amygdala or other regions of the brain. Stimulations can occur from a few minutes to several days apart, and the effect generally appears in three successive stages (Gaito, 1979). Stage I is the appearance of apparently normal exploratory behaviour (ethological stereotypies) during the first few stimulations; Stage II is the emergence of preconvulsive stereotypies of chewing, salivation and eye closure; and Stage III is the final convulsive stage in which the rat rears up and begins to clonically convulse with the forelegs. The convulsions do not stop immediately the stimulating current is terminated, and the effect of kindling may appear in spontaneous convulsions in rats subjected to many stimulating trials spread over four to five months. Thus the effect is relatively permanent, a conclusion supported by the fact that after a period of up to several months without stimulation, re-kindled convulsions occur after only a few trials, or even on the first.

71

ANIMAL CLINICAL PICTURES Kindled seizures in animals and epileptic seizures in humans bear some resemblance to each other, although that need not make them analogous. Nevertheless, the kindling procedure is capable of stimulating an epileptic-like condition, and Gaito ( 1976) describes two ways in which the kindling effect might contribute to solving the problem of epilepsy: first by revealing brain chemical and structural features involved in kindled convulsions; and second by uncovering possible therapeutic agents. Chief among such agents are treatment measures by anti-convulsant drugs (chemical inhibition), and preventative measures by means of non-convulsive Table 3.7 Kindling mechanisms and chemicals that affect each

Mechanism

Chemical iffect

Convulsive mechanism Retard

Effect on rate of development Interanimal retardation factor Taurine Atropine Ll 9 - tetrahydrocannabinol Phenobarbital Diazepam

Facilitate

Triggering mechanism Suppress

Potentiate

Reserpine 6-hydroxydopamine Handling (reducing stress and lowering level of norepinephrine?) Effect on trigger Ll 9-tetrahydrocannabinol Phenobarbital Diazepam Acetazolamide Lidocaine Methamphetamine Footshock (increase levels of norepinephrine?) Convulsion producing chemicals (e.g. metrazol)

From Gaito, J. (1976) 'The kindling effect as a model of epilepsy', Psychological

Bulletin, 83, 1097-109. Copyright 1976 by the American Psychological Association.

Reprinted by permission of the author.

72

Abnormal movements and convulsions electrical stimulation of the brain (electrical inhibition). Table 3. 7 lists groups of chemicals that retard, facilitate, suppress or potentiate convulsions and thus provide models to follow or to avoid in the synthesis of pharmacological agents for the control of clinical epilepsy. Brain stimulation seems a less likely therapy for clinical epilepsy than does the use of drugs. Nevertheless, with the kindling preparation Gaito (1979) reports that under some conditions successive or simultaneous lowfrequency (3Hz) stimulation can inhibit an established normal effect (60Hz) either by reversing the elicited behaviour from Stage III to Stage I or II, or by necessitating higher current intensities at the 60Hz frequency for convulsions to be produced. FOR THE MOLECULAR STUDY OF LEARNING Kindling and learning share common characteristics. They are both relatively permanent effects of experience that must involve the occurrence of changes across synaptic junctions, and they both show positive transfer effects and retroactive interference (Gaito, 1974). As such, Gaito suggests that the kindling effect might provide a paradigm

Behavioural

Behavioural automatisms

Normal behaviour Stage 1b

Clonic convulsions

Stage 2b

Stage 3b

0 Neurophysiological

"i::.J:::Q

~ ~ :g_

0£ 0

~ .2:

:=

Stage 1n Chemical

Cl:2 c 0

roE;::::>- cCi5

?

Q)(/)