The Sparrowhawk 9781472597380, 9781408138342, 9781408138335

Citation preview

THE SPARROWHAWK

Also by Ian Newton: FINCHES

(Collins)

POPULATION ECOLOGY OF RAPTORS

(Poyser)

The Sparrowhawk by IAN NEWTON

Illustrations by KEITH BROCKIE

1"' & A D POYSER Calton

First published 1986 by T & AD Poyser Ltd Print-on-demand and digital editions published 2010 by T & AD Poyser, an imprint of A&G Black Publishers Ltd, 36 Soho Square, London W1D 3QY Copyright © 1986 by Ian Newton The right of Ian Newton to be identified as the author of this work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988. ISBN (print) 978-1-4081-3834-2 ISBN (epub) 978-1-4081-3832-8 ISBN (e-pdf) 978-1-4081-3833-5 A CIP catalogue record for this book is available from the British Library All rights reserved. No part of this publication may be reproduced or used in any form or by any means — photographic, electronic or mechanical, including photocopying, recording, taping or information storage or retrieval systems — without permission of the publishers. Visit www.acblack.com/naturalhistory to find out more about our authors and their books. You will find extracts, author interviews and our blog, and you can sign up for newsletters to be the first to hear about our latest releases and special offers.

Contents

Acknowledgments

15

Background, study areas and methods

17

2 The Sparrowhawk

28

3 Nesting habitat

44

4 Nest spacing and breeding density

56

5 Ranging behaviour

69

6 Population trends

86

7 Hunting and feeding behaviour

101

8 Food

III

9 The Sparrowhawk as a predator

124

10 Breeding season

138

11

150

The breeding cycle: early stages

12 The breeding cycle: eggs and incubation

164

13 The breeding cycle: growth of young

176

14 The breeding cycle: parental care

190

15 The breeding cycle: post-fledging period

205

16 Seasonal trend in breeding success

214

17 Nest failures

224

18 Age and breeding

233

19 Moult

245

20 Dispersal

255

21

270

Territory and mate fidelity

22 Migration

281

23 Mortality

290

24 Effects of pesticides

301

25 Conclusions

323

6

Contents Appendix 1: Finding nests

335

Appendix 2: Persecution ofSparrowhawks

337

Appendix 3: Procedure for analysis ofnest spacing

342

Appendix 4: Causes and diagnosis ofnestfailure

344

Bibliography

349

Tables 1-63

357

Index

389

List of plates

between pages 48 and 49 1 Cock with prey; hen at nest 2 Pair at nest; hen withyoung 3 Views in Annandale: Ae Forest)" westfrom Wamphrey 4 Views in Eskdale: upper Esk valley)" lowerEsk valley 5 Birch nestingplace)" nest in birch wood 6 Nesting habitat: young and old pine plantations 7 Nesting places. in Edinburgh: cemetery and park 8 Methods: bloodsampling)" feather patterns between pages 160 and 161 Portraits at a pool Related species: Goshawk withyoung)" Goshawk nesting place Related species: Shikra; Levant Sparrowhawks migrating Studies offood: cock at plucking post)" prey remains on plucking post Studies offood: inspecting a nest; prey remains from nest Nest in pine tree; broken thin-shelled eggs Variation in eggs)" pellets Cock guarding eggs: hen about to incubate Hen brooding; henfeeding small chicks Hens with small chicks Feedingyoung: swallowing prey remnant 20 Henfeeding young; henabout to brood 21 Hens with largeyoung 22 Behaviourin rain 23 Fledglings 24 Tawny Owl)" deadSparrowhawk 9 10 11 12 13 14 15 16 17 18 19

Lis t of figures

1 2 3 4 5 6 7 8 9 10

11 12 13 14 15 16 17 18 19 20

21 22 23 24 25 26 27

List ofstudy areas Annual cycle ofSparrowhawks Feet ofSparrowhawk Sexes ofyoung in Sparrowhawk broods Numbers offemales per male Sparrowhawk in different habitats Mean weights ofadult Sparrowhawks each month through theyear Breeding and wintering ranges ofSparrowhawks Use ofwoodlandareas with different tree spacing by nesting Sparrowhawks Use ofwoodland in two areasby Sparrowhawks and Goshawks Numbers ofold nestsat nestingplaces Regular spacing ofnestingplaces in Upper Speyside Spacing ofnestingplaces in continuous woodlandin different areas Spacing ofnestingplaces in relation to land productivityand elevation Spacing ofnestingplaces in relation to indices ofprey bird population Home ranges ofa particular cock and hen in different seasons and in differentyears A Sparrowhawk nest site in Ae Forest, and the places where cock and hen were located in the late nestling period Occurrence ofradio-tagged Sparrowhawks at different distances from the nest Activity ofradio-tagged Sparrowhawks at different times ofyear Overlapping homeranges ofSparrowhawks studied in winter Relationship between elevation and range size in cock Sparrowhawks Home ranges ofseven different cocks studied in Ae Forest Locationsofhens in the late nestlingperiod in relation to breeding performance Population trend in Eskdale Percent change in Eskdale population in relation to population level the previousyear Annual recruitment ojnewfemales to the nestingpopulation in Eskdale Frequency ojuseoJ54 Sporrotcluuok nestingplaces in Eskdale during a 13-yearperiod Overwinterloss ojpre-breeders in relation to population levelthe previous year

20

25 30

36 37 39 40

46 53 57 59 60

62 65 72 74 75 77 78 80 81 83 88 89 90 92 94

List offigures

28 Population trendin Annandale 29 Trend in various aspects ofbreeding performance in Annandaleand

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Eskdale Main prey species during the breeding season Main prey in different habitats in southScotland, April-August Sizes ofhrey items taken during incubation, and during the late nestling post-fledging stages Proportions ofthree tit species in the diet Proportions ofvarious prey species in the diet Main prey ofSparrowhawks in southScotland, September-March Vulnerability ofdifferent prey species to predation by Sparrowhawks Proportion ofHouseSparrows in the diet ofdifferent Sparrowhawk pairs, according to distance ofthehawk nestfrom a humansettlement Percentage ofyoung tits in thefood ofSparrowhawks shown in relation to meanfledging dateoftits Breeding ofSparrowhawks in relation tofood supply Start ofbreeding amongSparroiohaioks in relation to theappearance offledglings in the diet Relationship between meanlaying date in differentyears and mean April weather Mean weights ofadult Sparroiohaioks in each 5-day periodthrough the breeding cycle Growth rates ofyoung Weight gain ofyoung in an Ae Forest brood whichexperienced poor growth and died Survival ofnestlings through thefirst 25 days Brooding andguarding by hensat two nests in Ae Forest Fooddeliveries at three Sparrowhawk nests in Ae Forest Relationship between growth rate andfood consumption in individual nestlings Proportion ofvisits to nests on which theadult hen was detected Post-fiedging survival in relation to weight nearfledging Post-fiedging survival ofyoung in relation to rank within broods Seasonal trend in breeding performance Number ofclutches deserted through the season Relationship between meanbreeding production in differentyears and April weather Pre-lay weight increase in three groups 0.1females with different laying dates Causes ofcomplete nestfailure

9

97 99 114 115 117 118 119 120 126 128 132 140 142 143 145 179 180 184 192 197

199 203 207 211 215 217 217 219 227

10

List offigures

57 Nest success in different areas in relation to themean level ofDDE 58 59 60

61 62 63 64 65 66 67 68 69 70

71 72 73 74 75 76 77 78 79 80

81 82

231 237 239

residue in the eggs Mean breeding performance infemales ofdifferentages Mean number ofyoung raised by different mating combinations Lifetime productions of142females Lifetime production offemales in relation to theirlongevity Relationship between start ofmoult and start ofegglaying in individual females Start 0.1 moult in relation to start ojegg laying infemales ojdifferent ages Moult scores ojindividualsexamined at differentdates Locations ojSparrowhawks recovered in the breeding season, shown in relation to birthplace Dispersal distances ojSparrowhawks recovered in different periods after becoming independent Recoveries outside the study areas ofSparrowhawks ringed as nestlings within the study areas Numbersofrecoveries at differentdistances from the birthplace The path taken by a dispersing juoenile after leaving its natalnest area Median dispersal distances ofAnnandaleyoung hatched in differentyears Distances between hatchplace and recovery placefor cocks and hens raised in upland and lowland habitats Distances moved by siblings between hatchsite and recovery site Turnover ofoccupants onparticularnesting territories Full residence periods ofindividuals on territories Distances moved by Sparroiohaioks between territories usedin different years Migration ofSparrowhawksfrom different areas to showparallel movement patterns Migration ofS'parroiohatoks in relation to themigrations oftheirmain prey species Population trends, as revealed by annualmigration counts Proportions ofBritish ringedSparroiohaioks recovered in various ways Seasonal pattern ofrecoveries from British ringed Sparrowhawks Shell-thickness index ofBritish Sparrowhawks, 1870-1980 Mean shell-thickness index in relation to mean DDE level in thecontents ojeggs from different areas Main mechanisms ojpopulationdecline resulting from organochlorine use

307 311

useand organochlorine use

312

83 84 Changes in the status ofSparrowhawks in relation to agricultural land-

240

241 249 250

251 257 258 259 260

261 263 265 267 271 272 275 283 284 287 295 297 306

List offigures 11 85 Population recovery ofSparroiohawks in different areas 86 Organochlorine levels in livers ofSparrtnohaioks fiom different localities 87 Levels ofDDE and HEOD in the livers ofSparroiohaioks 88 Trends in HEOD levels in Sparroioharok liversfrom different zones 89 Numbers ofgamekeepers in Britain at different times 90 Index ofoutput ofyoung Sparrowhawks in Britain

314 316 319 321 338 339

List of tables

1 2 3 4 5 6 7 8

9 10

11 12 13 14 15 16 17

Wing lengths ofSparrowhawks ofdifferent sex and age S'porroiohaiok iris colour in relation to age and sex Some measurements ofS'parrouiiuuoks Proportion ofwoodsin Annandale occupied by breeding S'parroiohaioks Frequency with which nestingplaces wereusedin woods- ofdifferent size Frequency with which nestingplaces wereusedin woods ofdifferent type Heights ofnests in relation to height and type oftree Spacing and densities ofSparrowhawk nestingplaces in different districts Densities ofSparroiohawks in various parts ofEurope Age composition ofdifferent components ofthe Sparrowhawk population Numbers ofestablishedfemales and newfemales in the Eskdale breeding population eachyear Lack ofevidence for density-dependent regulation ojbreeding in Eskdale Breedingsuccess) persistence ojfemale) and proportion ofyearlings on nestingplaces usedin different numbers ofyears Annual survival ojpreviousbreeders and recruitment ofnew breeders in two study areas Attack success ojSparrowhawks seen hunting while on migrationat Falsterbo, Sweden Number o.fprey of different sizesfound nearnestsand aioayfrom nests respectively Frequency ojdifferent prey species in thefood afSporroiohaioks

357 358 360

359 359 360

361 361 362 361 363 363 363 364 364 365 365

12 List of tables 18 Main pr~y species in thefood ofSparroiohaioks in different regions 19 Predation by Sparrouihaioks onfour song birdsin theNetherlands 20

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41

42 43 44 45 46 47 48

Effect ofS'parrotohatoks on nesting tits Various aspects 0.[a Great Tit population nearOxford during periods whenSparrowhawks were absent andpresent respectively Relationshipbetween laying date and spring weather Weights ofhen Sparrowhawks at differentstages ofbreeding Effect ofsupplementary food on thelayingdates and clutch sizes of Sparrowhawks Breeding success ofSparrowhawks in differentdistricts Frequency ofdifferentclutch sizes infirst and repeat layings oj52,fenzales Incubation periods Spreadofhatchin clutches ofdifferentsize Proportion oJrecords of-young radio-tagged Sparrowhawks at different distances from thenestduring thepost-fiedging dependency period Survival in relation to brood size Survival in relation to sex composition ojbroods Proportions ojbirds recovered, according to date ojfledging Seasonal decline in breeding performance Survival 0.[adultfemales in relation to theirlayingdates Analysesofvariance on laying dates offirst eggs Summary ofoverall breeding performance Mean number ofyoung producedJrom clutches ofdifferent sizes Proportions ofnests whichproducedyoung in different periods Evidence for assortatiue mating betweenyearling and adult Sparrowhawks Percentage oJyearlings among breeders in different habitats Calculation ojproportion ojbreeders among females 0.1 different age groups Productivity ojJemalesin relation to ageoffirst breeding Breeding performance in pairs in which bothpartners could beclassed asyearling or adult NumberoJyoungproduced by individualfemales in successiveyears Survivalojfemales in relation to number oJyoung raised theprevious year Lifetime productions ofresident and immigrant females The start ofmoult in thesecondary and tailfeathers in relation to the stageofmoult in theprimaries Numberofprimaryfeathersper wing in growthat once at different stages ofmoult

368 369 370 370

371 371 372 372 373 373 373 373 374 374 375 375 376 376 377 378 378 378 379 379 380 380 381

381 382 382 382

List of tables

49 Dispersal distances from natal nestand subsequent performance 50

51 52 53 54 55 56 57 58 59 60

61 62 63

Correlation between distances from natal nestmoved by closely-related individuals Turnover ofbirdson territories ofdifferentgrade Frequency ofterritory changes in adults, according to nest success the previousyear Proportions ofhens which changed territory, according to ageand nest success thepreviousyear Breeding performance in relation to previous experience ofmateand territory Median dates ofmigration for Sparroiohaioks trappedat two sites, showingsex and agedifferences Mortality estimatesfor Sparroiohatoks obtained in thepresent study Mortality estimates for Sporroiohaioks obtained in previous studies Survival afjuoenilesin relation to weight in August-October Recovery circumstances ofBritish-ringedSparroiohaioks in different periods Estimated annual usage ofsome organochlorine pesticides in British agriculture/ horticulture duringdifferentperiods Extent ofchange in Sparroiohaiok eggshellsfrom differentregions of Britain Shell thickness and breakage in 127Sparrotohauik clutches Proportion ofSparrowhawk nesting places usedin different periods in an area with about30 traditional territories, Cumbria, NW England

13

383 384 384 384 385 385 386 386 386 387 387 388 387 388 388

This page intentionally left blank

Acknowledgments

In some ways the 14 years of research described in this book is the result of a team effort, and it is a pleasure to acknowledge the help of several friends and colleagues. Herman Ostroznik - the best tree climber I know - introduced me to the Annan Valley in south Scotland, and taught me many a skill in finding Sparrowhawks. The late Willy Murray also gave valuable help in the early years. In 1974 Mick Marquiss joined me, and until 1980 we worked together, so that many of the ideas arose from our joint discussions. Mick put in a particular effort on autumn trapping and radio-tracking work. For three years as a PhD student, Dorian Moss studied the growth of nestling Sparrowhawks, made censuses of prey-birds in various woods in the Annan Valley and elsewhere, and later excercised his statistical and computing skills in analysing some of the more complex of the information gained. Other statistical help was given by David Brown, Ken Lakhani and Peter Rothery. When I moved to England in 1979, Andy Village helped in Scotland for a while, while Ian Wyllie and I started fresh work in Northamptonshire. Robert Kenward introduced me to radio-tracking, and collected some of the information given on home range. Mac andJanet Hotson offered frequent hospitality in Langholm, and to Mac I am also grateful for help and company in the field. Several other raptor enthusiasts, working independently in other parts of the country, have made their findings freely available for incorporation in this book. They include Brian Etheridge in Morayshire, Neville Bousfield in Aberdeenshire, Brian Little, Alan Hcavisides, Eric Meek, Steve Petty and others in Northumberland, Ted Robson and Geoff Shaw in Derbyshire, Peter Davis in Dyfed, and Ted Green and others in Berkshire. To all these people

16 Acknowledgments I am grateful for their help, and I hope that they will derive some satisfaction from this synopsis of our efforts. I also owe a considerable debt to David Jenkins, for some years my senior officer in the Nature Conservancy, for helpful criticism and continuing interest in the work, to Derek Ratcliffe and lanPrestt for initial help and encouragement; and to Jack Dempster, my senior officer in the Institute of Terrestrial Ecology. I am grateful to the Forestry Commission, the Buccleugh Estates and other landowners in Dumfriesshire for permission to work on their ground, and to the British Trust for Ornithology for making the national ringing recoveries of Sparrowhawks available for inclusion in this book. Mick Marquiss and Ian Wyllie read the whole book in draft, and suggested many improvements, while Alistair Dawson, DavidJenkins, Robert Kenward, Dorian Moss and Steve Petty commented helpfully on particular chapters. Margaret Haas did some of the calculations and diagrams for the original papers on which much of the book is based, while Nigel Westwood drew the remaining diagrams and printed some of the photographs. Robin Chancellor kindly shared the onerous task of correcting the proofs. Finally, I thank my wife, Halina, for encouragement throughout, and for checking the book in draft. I. NEWTON

CHAPTER I

Background, study areas and methods

This book is a bo u t th e eco logy of th e Sparrow hawk Accipiter nisus: a bo u t its ha bitat a nd food need s, its breedi ng , movem ent s and m or talit y. The book is we igh ted un a sh amedly towa rds th ose as pects whi ch ha ve caugh t my own int er est , es pecia lly th e con trol of number s a nd breeding. The rem arkable size differen ce between th e sexes ad ds a n extra int erest , as m al es a re only hal f th e we igh t of femal es. This gives the Sparrowha wk sexes a differen ce in eco logy as grea t a s th at o fte n found between di stinct spec ies. The sexes div erge to so me d egr ee in habit at and di et , a nd a re affected differ ently b y th e sa me a dv ersities. The book is based primaril y on information gained during a 14-y ear study in so u the rn Scotland , in which I a tte m p ted eac h year to find a ll the nests in tw o study areas , to assess th e breeding success, and to tr ap a nd ring for id entification as m an y of th e birds as possible. Thus most of th e data are fro m known marked individuals , some of whi ch were stud ied for their en ti re lives . Add iti on al information has been includ ed from pa rall el studies in othe r a reas, and al so fro m th e lit er ature, a s th e Sparrowh awk has been stud ied exte ns ively, both in Brita in a nd in parts of con ti ne n ta l E urope . Those read ers wh o are un familiar with th e Sparrow hawk will not find it hard to ide n tify, but th ey m ay not see o ne ofte n . As th e m ajor pred ator

18

Background, study areas and methods

of woodland songbirds, it is the commonest diurnal raptor over much of its range, except for the Kestrel Falco tinnunculus, a bird of more open land, which lives chiefly on mice and voles. The two species are of similar size, and both have longish tails, but whereas the Sparrowhawk has the short broad wings of an accipiter, the Kestrel has the long pointed wings of a falcon. Their lifestyles also differ. The hawk is secretive, and can be glimpsed only briefly, as it skims low along hedgerows, alternating a few quick wing flaps with long glides, and putting up flocks of birds as it goes. The Kestrel, in contrast, sits conspicuously in the open, on tree tops or telegraph poles, or hovers over some grassy patch, scanning the ground for prey. The closest relative of the Sparrowhawk in Europe is the larger Goshawk A. gentilis. This species breeds in more mature woods than the Sparrowhawk, and takes larger prey, again chiefly birds. I t was persecuted to extinction in Britain during the last century, but has recently re-established itself in some areas, numbers of imported birds having escaped from falconers. It is so much larger than the Sparrowhawk that there is little chance ofconfusion. If you see an accipiter and, on the basis of size, are unsure which of the two, then rest assured it is a Sparrowhawk.

Somepersonalcomments My encounters with Sparrowhawks began at an early age, and ever since I found a nest, these birds have held a special fascination for me. In the nineteen-fifties, most of the boys in our Derbyshire village collected birds' eggs, and those of Sparrowhawks were especially prized possessions that only a minority of us could expect to get. Compared to other birds, Sparrowhawks were decidedly scarce; their nests were hard to find and often in difficult trees. My first nest was in a typical situation, about half way up a spruce tree, by a stream in a small mixed wood. To this day, I can still picture those eggs, beautiful pale blue objects, richly blotched in chestnut brown. As a teenager, I got to know Sparrowhawks better, searching ever further from my home, and in more and more woods, until I was finding up to ten nests each year. By this time, I had lost interest in egg collecting, which in the meantime had become illegal, but I enjoyed tracking the birds down, and watching the course of events through a breeding cycle. I read everything I could find on the species, but especially liked the articles written in the early part of the century by J. H. Owen in the journal British Birds. Owen was a schoolmaster, who enlisted the help of pupils to watch Sparrowhawks in the woods near Felsted in Essex. At practically every nest he found, the adults were shot by gamekeepers, although he did persuade the keepers to leave a few for his studies. He was one of an earlier generation who evidently did not see the need to count or quantify, so his work lacks the precision

expected today. Nonetheless, he knew a lot about the species, and raised endless questions which fired n1Y imagination at the time, and which I could attempt to answer later in life. Then came the time when Sparrowhawks throughout Britain suddenly

Background, study areas and methods

19

declined in numbers, disappearing altogether from certain districts where only a few years previously' they had been common. This followed the widespread introduction in agriculture of some new synthetic pesticides - the so-called 'organochlorine' compounds, which included DDT and dieldrin. The last nests I saw in my home area, near Chesterfield, were in 1958. I found eight nests that year, all in long-used sites. The following year there was none, and the woods of that area remained empty of breeding Sparrowhawks for nearly 20 years. For it was not until the mid-1970s that the birds began to return, increasing and spreading again following restrictions in organochlorine use. Through my work I was able to return to Sparrowhawks in later life. In 1971, while employed by the Nature Conservancy, I was given the opportunity to study birds of prey. At that time there was great concern because the Sparrowhawk and other species had declined through recent pesticide use, while the longer-standing problems created by land-use changes and gamekeeper activities were still with us. The Sparrowhawk seemed an ideal subject for study, because it was affected by all these factors, yet was still sufficiently numerous in some districts to provide the large samples necessary for statistical treatment. This last aspect was crucial because, in order to investigate population regulation, samples had to be large enough to reveal the roles of different limiting factors. No-one had previously studied a diurnal raptor in this degree of detail, and I hoped that my findings might be applicable to other species too. Let me say right away, however, that no-one who is interested solely in bird population problems would choose to study Sparrowhawks. Compared to many other birds, Sparrowhawks are scarce and highly secretive, spending most of their time in cover. They are thus impossible to count directly or to observe in the field for any length of time. Finding nests in numbers entails endless hours ofsearching, and inspecting nest-contents requires many difficult climbs. For me, however, it was the challenge of getting to grips with these elusive birds which gave them their appeal.

STUDY AREAS

In 1971, Sparrowhawks were still scarce around Edinburgh, where I then lived, and the nearest area with a substantial population was Dumfriesshire, where large-scale ringing of nestlings had been done for some years previously by members of the local ringing group. This county, with its mixture of hills and valleys, of sheepwalk, forest, farmland, heath and sea coast, provided a varied environment for raptors. It seemed a good place for study, and I settled on two areas, cen tred on the valleys of the Annan and Esk respectively, two rivers which run south from the Southern Uplands to the Solway Firth (Fig. 1). The Annandale area contains the towns of Moffat, Lochmabcn and Lockerbie, while the Eskdale area has Langholm and Canonbie.

20

Background, study areas and methods

0® @

f) {)

0 CD

CD

Fig. 1. Location of study areas. The main areas in south Scotland were Annandale (9) and Eskdale (10). Other numbers show locations ofsubsidiary study areas from which further information was gained: (1) Windsor, Berkshire; (2) Oundle(Rockingham Forest), Northamptonshire; (3) Tregaron, Dyfid; (4) Glossop (Peak District), Derbyshire; (5) Carlisle (south Solway Plain), Cumbria; (6) Slaley, Northumbria; (7) Kielder Forest, Northumbria; (8) Dinnington, Northumbria; (9) Annandale, Dumfriesshire; (10) Eskdale, Dumfriesshire; (11) Galloway; (12) Aberdeen (lower Deeside) , Aberdeenshire; (13) Banchory (mid Deeside), Kincardineshire (14) Ballater (upper Deeside), Aberdeenshire; (15) Mar Forest, Aberdeenshire; (16) Aviemore (Speyside), Inverness-shire; (17) Huntly (Clashindarrock Forest), Aberdeenshire; (18) Forres, Morayshire.

Background, study areas and methods 21

Annandale The Annan Valley begins above Moffat, and broadens as it spreads southwards, giving a wide expanse of gently undulating farmland, with scattered villages, and woods of varying size. The farmland was mixed, with grassland for cattle and sheep, and some arable, including root crops and cereals. Most of the woodland in the area, in both hill and valley, was managed for softwood production. The spruces Picea abies and P. sitchensis were the commonest species grown, followed by larch Larix deciduus and pine Pinus sylvestris. Birch Betula spp. and other broad-leaved trees remained in a few semi-natural woods on the valley floor, and as strips along some of the steeper streamsides in the hills. Overall, about 30% of the area was under mature woodland. When I began, the hills on either side of the valley were mainly grasscovered, grazed by sheep, but there were many small copses and several large plantations, the latter established 30-40 years earlier by the Forestry Commission. The Forest of Ae, on the west side of Annandale, was the largest expanse of trees in the area, at that time covering 55 krn", though much of it was too young or otherwise unsuitable for Sparrowhawks. In this forest, up to 27 nests were found in anyone year. Other forests in the study area were Greskine in the northwes t corner (with up to 11 nests per year), Craigieburn in the northeast corner (with up to 5 nests per year), and Auchenroddan on the eastern side (with up to 5 nests per year). As the years passed, more and more of the hill ground was planted with trees and, although these remained too young for nesting Sparrowhawks, most of the hill country changed in appearance during the years I was there. Throughout the 1970s, changes also occurred on the valley floor. Hedges and scrub patches were progressively removed, damp fields were drained, pastures were ploughed, and broad-leaved woods were increasingly cleared and replaced by conifers. These changes had obvious effects on the local bird life, a progressive replacement of open land species by woodland species in parts of the hills, and a steady decline of bird life in the valley, especially songbirds and waders. Almost certainly, the Sparrowhawks of Annandale experienced a progressive deterioration in food-supply during the decade that I was studying them. On the other hand, the use of organochlorine pesticides diminished, and breeding failures from these chemicals became rare. Eskdale The Esk Valley to the east was narrower than the Annan Valley, and the farming was less intense, with very little arable. Woodland was present in smaller proportion than in Annandale, covering about 22% of the total area, and was distributed mainly in strips along the Esk, and up the various tributary valleys. The area was generally better for Sparrowhawks than Annandale, as it held more prey, small birds being favoured by the particular landscape, in which small woods alternated with small fields, with no large

tracts of open land except on the hills. The woods themselves, managed by the Buccleugh Estates, were more varied in structure and tree species

22

Background, study areas and methods

than those in Annandale, but were still mainly coniferous. There was also more Pheasant Phasianus colchicus rearing in Eskdale than in Annandale. 'This was an important activity because the feeding of Pheasants in winter provided additional food for smaller birds, and almost certainly increased their numbers. The landscape in Eskdale changed little during the years I was there, and so provided more stable conditions for birdlife, including Sparrowhawks, than did Annandale. From time to time, I also worked near the top of the Esk valley, in the forest at Castle O'er, and in the neighbouring valley to the east, in Liddesdale. The Annandale study area covered about 700km 2, and held up to 110 known Sparrowhawk nests in anyone year, while the Eskdale area comprised about 200 km 2 and held up to 39 known nests in anyone year. The two areas were about 15 km apart at their nearest points, and were separated mainly by high ground. For some purposes of analysis, it was appropriate to distinguish three main habitats in each area: (1) low ground (mainly below 250 m), comprising mixed farmland with scattered woods and villages; and high ground of (2) open sheepwalk with scattered small woods, and (3) large plantation forests. Intensive work in Annandale spanned ten years, 1971-80, and in Eskdale 13 years, 1972-84, incorporating 14 years in all. Further information was gained from other areas, scattered between Windsor in southern England and Forres in northern Scotland (Fig. 1). These areas were worked by various friends and colleagues, who made their findings freely available for incorporation in this book. For convenience I refer to the birds in each study area as a population, but in reality each area's birds formed a segment of a much larger population, spread more or less continuously throughout Britain. In no sense were the birds in anyone area confined and isolated.

METHODOLOGY

With the help of colleagues, I searched the woods of the south Scottish areas systematically every year in an attempt to find all the nests. Our methods are described in detail in Appendix 1. For each nest, we noted the laying date (of first egg), clutch size, brood size at hatching and fledging, together with details of nest site and prey remains. Parent Sparrowhawks were caught at their nest sites each year, in order to ring them or to identify those ringed in previous years. As hens spent more time in the nest areas, they were much easier to catch than cocks, and thus provided larger samples. Both sexes were caught in baited cage traps, placed near nests in the pre-laying and late nestling stages, or on hunting areas at other times of year. Other hens were caught in noose traps placed on nests during incubation. With this last technique, we could often catch a bird within 20 minutes and, once the system was perfected, we could go from nest to nest and catch up to eight hens in a day, or up to a hundred in a season. It was never possible

Background, study areas and methods

23

to capture every known breeding hen in the study areas, because it took so long to get round all the nests that many birds had failed and moved away before we could get to them. Nonetheless, in some years we caught up to 80% of known breeding hens. Only rarely were cocks caught in this way, because they seldom visited the nest at this stage. We had few problems with trapping, and no nest desertions that we could definitely attribute to it. Some trapped birds did abandon their nests, but the proportion was the same as among birds that were not caught, and moreover many birds which did desert showed signs (such as fresh moulted feathers) of having incubated for long after they had been trapped. Thus in the case of the few nest-trapped birds which did desert, trapping was probably not the cause. Over the years, trapping and ringing provided information on the age structure of the population, on the history and longevity of individuals, on their fidelity to territory, on movements between territories, and on mate choice, as well as on moult, weights and measurements. Effective trapping was thus central to the study. The difficulty ofactually seeing Sparrowhawks made it necessary to capture the same individuals year after year, and identify them from their ring numbers. There was no point in using wing-tags or other conspicuous markers, because we seldom saw a bird well enough to identify it by such .means. For the same reason, we fitted radio-transmitters to birds in order to track their movements. Radios gave us detailed information on behaviour, on habitat use and home ranges, which could not be obtained in other ways. While breeding, Sparrowhawks begin their annual moult, and the shed feathers from the hen could be found on the ground near the nest. The large flight feathers were collected to build up a reference set for each hen, which could be used in individual recognition. Although. the ·flight feathers of all Sparrowhawks have the same basic pattern, with alternating light and dark bars, the feathers from different individuals vary in the proportions of light to dark, in the exact shape of the bars, and in other markings (see Plate I). After the first year of life, each bird showed a consistent feather pattern from year to year, so that, providing equivalent feathers could be compared, they gave a means of identification, analagous to human fingerprints. The wings were symmetrical in this respect and any particular feather in one wing had the same pattern as its counterpart in the other wing. The use of feather patterns to identify individual Sparrowhawks was first proposed by Opdam and Miiskens (1968) in Holland, but we were also able to check the method's reliability with known ringed birds. We used it chiefly to assess whether or not the bird on a particular territory was the one that was present the year before, but only on the minority of territories where we failed to catch the occupant. The same method was equally applicable to males, but, as they shed their feathers over a wider area, we seldom obtained enough for individual recognition. The method has also been used successfully on Goshawks, and would probably work for other raptor species which have patterned flight feathers.

24'

Background, study areas and methods THE SPARROWHAWK'S YEAR

In south Scotland, the daily light period changed from about 17 hours on 21 June to about 7 hours on 21 December, and average temperatures were highest in July and lowest in January. Snow usually fell between December and February, but in most winters it lay for only a week or two at a time. The winter of 1981-82 was especially hard, however, and snow lay for several weeks. Most of the small bird species that comprised the Sparrowhawk diet were resident in the area all year, and bred mainly during April-August. In consequence, their populations reached a peak in August, and then declined again until the following April. Thereafter, warblers and other summer migrants began to arrive and breed, boosting the local bird population even further. The departure of these summer migrants in August-September was followed by the arrival of large numbers of Fieldfares Turdus pilaris and Redwings T. musicus in October. Most of these birds passed through, and their numbers in south Scotland remained relatively low until the return passage in April, evident especially in Fieldfares. Overall, the total prey population of the area followed closely the changes in numbers of resident birds, from a low in April to a peak in A ugus t. The Sparrowhawks themselves were resident in the area all year. There was no evidence from ringing that any local birds migrated for the winter, or that birds from northern Europe came in. Continental Sparrowhawks have been recorded in Britain, but mostly in districts further south or east than the study areas. I t was between late April and August, the period when easily caught fledgling songbirds were available, that the local Sparrowhawks themselves bred (Fig. 2). Most pairs laid in May, so that their young hatched in June, left the nest in July and became independent of parental care in August. The young hawks dispersed mainly in their first few weeks after becoming independent, and remained relatively sedentary thereafter. The adults moulted mainly in summer, as indicated above, hens beginning in May, around the time of egg laying, and cocks in June. Both sexes continued the moult into September or October, beyond the date when their young became independent, but while food was still plentiful. Hence, the two most demanding activities occupied the most favourable season, and for the rest of the year the hawks were free, if necessary, to hunt only for maintenance; however, in early spring many pairs began nest building at a time when prey was still scarce.

HIS'TORY IN BRITAIN

As with most birds in Britain, changes in the status of the Sparrowhawk over the centuries have resulted chiefly from human activities, coupled in the case of the Sparrowhawk with a dependence on woodland for nesting

Background, study areas and methods

25

Nest.building - - - - - - - - - - - - - Egg-laying

------

Incubating

------

Feeding young in nest

- - -----

Feeding young out of nest

-------

Young dispersing

------

Adult males moulting Adutt females moulting

Jan

Feb

Mar

Apr

May

Jun

Ju I

Aug

Sep

Oct

Nov

Dec

Fig. 2. Annual cycle of Sparrowhawks in south Scotland. Solid lines show the period when more than halfthe population was engaged on the activitiesconcerned.

and on a supply of small birds for food. Some 2,000 years ago, when most of Britain was forested, Sparrowhawks were presumably much more abundant than today. However, their former numbers cannot be reliably estimated from densities in small modern woods, because densities in continuous forest were almost certainly lower. In addition, to judge from present day findings, the birds would not have bred uniformly through forest of all ages, but mainly in the young, scrubby stages, while more mature forest would be occupied by the Goshawk (Chapter 3). Sparrowhawk numbers in primeval Britain were therefore likely to have varied according to the state of the forest, being largest in periods when the greatest proportions of the total forest were at young stages. They probably numbered many hundred thousand pairs. With the progressive clearing of forest by man, the extent of potential breeding habitat was much reduced, reaching its lowest level towards the end of the 19th century, when less than 5% of Britain was wooded. The nineteenth century also saw the development of efficient firearms, which in turn led to game preservation, initiating an era when Sparrowhawks were heavily persecuted because of their depredations on young game birds. Considering the large numbers of hawks killed, it is remarkablethat the species survived this era so well. Evidently there were still enough unkeepered woods, where the hawks could breed undisturbed, and their high reproductive rate enabled them to maintain their numbers despite relentless killing. However, in many districts, notably East Anglia, their numbers seem to have been kept well below the level that the habitat could support. An account of this persecu tion is given in Appendix 2. Not only rap tors were destroyed in the interests of game, but mammalian carnivores too. Thus the two mos t important predators of the Sparrowhawk were eliminated, the Goshawk from the whole country and the Pine Marten Martes martes from all but northwest Scotland. In recent decades the Goshawk

26

Background, study areas and methods

has re-established itself in some localities, as mentioned earlier, and the Pine Marten has spread slightly, but the fact remains that for more than a century Sparrowhawks in Britain have been free of their main natural enemies. Of the minor predators, the Tawny Owl is probably the most significant. The further decline in gamekeeping during the Second Worid War enabled Sparrowhawks to increase quite substantially, and towards the end of this period they were probably more numerous than at any time since 1800 (Newton 1972). Soon afterwards, however, the birds were affected by a new threat, the organochlorine pesticides, which were introduced at that time in agriculture. DDT was used from the late 1940s, and the more toxic cyclodiene compounds - such as aldrin and dieldrin - from the late 1950s. In the space of a few years, these pesticides achieved what a century of gamekeeping had failed to do, and eliminated the Sparrowhawk from large parts of its British range. With successive restrictions on organochlorine use, the species began to recover in numbers, and is now back in strength over much of the country. At the time of writing, however, it is still largely absent from parts ofsoutheast England, where organochlorine use was heaviest. Otherwise it breeds over the whole country, wherever there is woodland, and can be seen anywhere except the highest mountain tops. Because of the decline in numbers, the Sparrowhawk was given full legal protection in Britain in 1961. This may have reduced the numbers shot, but was of little or no significance in the population recovery. Now that Pheasants are reared mainly in intensive units and released when well grown, the incentive to kill Sparrowhawks has largely disappeared, and persecution is negligible. The national reafforestation programme has also favoured Sparrowhawks in recent decades, as the increasing forest areas have provided much new breeding habitat, chiefly in the uplands. These new conifer plantations are managed on relatively short rotations, mostly of 40-60 years, so that soon after they become too open and mature for Sparrowhawks, they are felled and replaced. With different sectors being regularly felled and replanted, there always remains a high proportion of forest at a suitable stage. In fact, the present management of commercial woodland in Britain is ideal for sustaining a large Sparrowhawk breeding population (Chapter 3). On the other hand, the amount of prey available has almost certainly declined. This is especially so in the lowlands, where agricultural changes have rendered farmland much less suitable for small prey-birds than it was 50 years ago. The progressive destruction of hedgerows, scrub and wasteland, the increasing conversion of pasture to arable, and the widespread use of chemicals can only have greatly reduced the small bird populations of these habitats, though it is impossible in retrospect to judge by what extent. When I began in 1971, Sparrowhawk numbers in the Dumfries study areas had fully recovered from the low imposed by organochlorine pesticides, though breeding success nlay still have been somewhat reduced. Throughout the study, the birds continued to lay thin-shelled eggs - a symptom of con tamination with DDT residues - but in the majority of eggs the thinning was insufficient to cause breakage or to prevent eventual hatching. In fact, organo-

Background, study areas and methods

27

chlorine pesti cid es probably ca used only a sm a ll reduct ion in br eed ing success, an d no depl eti on of br eeding popu lations, in th e main study area s during th e yea rs conc erned (see later ). Nor wer e th e ac tivities of ga mekee pe rs of a ny significan ce. I n Annanda le, up to three fem ales wer e sho t on th e nest each ye a r, a nd in Eskd ale only 1- 2 in th e ea rly yea rs. A few othe r Sp arrowha wks wer e kill ed a t Pheasant-rearing pens , or during ga me shoo ts, but a lmost certa inly th ese form ed a sma ll pr oportion of th e tot al deaths. The hawk popu lat ions of the study areas could thus be cons id ered as littl e influenced by direct human activities. Much of th e mat eria l in thi s book has a lre ad y appeared in th e form of scie n tific pap ers , publish ed a t va rious dat es during th e stud y. For th e book, I have mostl y use d th e data a na lysis in th e pap ers, a nd ha ve incorpora ted more recent informa tion only whe re it wo uld a ppreciably modify or expa nd th e picture. Th e sa m ple sizes used to esta blish particul ar point s thu s va ry grea tly from on e cha p ter to a no th er. Some rep etition bet ween cha pte rs has proved in evit able, as certain topi cs had to be rai sed mor e than on ce, but I hope that the ind ex will direct th e read er to a ll that is wr itt en on an y one tonic.

C H A PT E R 2

The Sparrowhawk

The Sparrowhaw k is one of th e smalles t birds of prey in Europe. Th e fema le mea sures abo u t 37 cm from head to ta il, a nd has a wing span of aro und 74cm , whi le th e smaller mal e measu res abou t 34 ern by 62 cm. In size, sha pe a nd co lour, th e species is stri king ly well-a da pted to its wood lan d ha bita t a nd to its so ng -bird di et. It can be recognised by its small head, slim bod y, sho rt broad wing s, lon g tail , lon g thin legs a nd ch a ract eristi c m od e ofhuntin g. The sho r t wings a nd lon g ta il give the Sp arrowh a wk grea t m a noeu vr abilit y. The bi rd ca n thread its way wit h d ext erit y a nd b reath-taking spee d a mo ng trunks a nd branch es, a nd twist and turn with a mazing eas e. Altho ug h larger a nd hea vier , it is a m at ch for a ny fleein g song-bi rd. In th e hand th e a d ult Sparro whawk a ppears quite colourful, bu t in th e natural habitat th e bird ca n be re mar ka bly hard to see . Severa l features help to co nceal it. T he most o bvious in a flying bird is th e co un te r-s ha di ng, wit h d a rk blue-grey d orsa l surface and pale , barred und ersid e, in clud ing wings a nd tail. This dark-l ight pattern help s to d isrupt th e profil e. From below th e p al e undersid e is a lso less o bvious agains t th e sky than a d a rk one wou ld be, wh ile fro m above the dark upperparts are less visible agai nst the ground . M a rked co unter-s hadi ng is show n by raptors which purs ue birds a nd o ther fast -moving prey, a nd has p robably arisen to facilitate h unting ,

The Sparrowhawk

29

but it would also help to conceal the hawk from its own predators. Countershading is slight or absent in raptors which eat slow-moving prey, but the pattern is frequent in other animals, including mammals and fish. Some of these features of shape and colour occur in other birds of prey which share similar habitat or diet. Throughout the world, all woodlanddwelling raptors have short wings and a long tail, whether they be accipiters, other types of hawks, forest-eagles or forest-falcons. Not all such species are closely related to one another, so presumably they have evolved this same shape by convergence, in adaptation to a similar environment. It is evidently the habitat rather than the diet which is important, as long tail and short wings are found to some degree in all woodland species whatever they eat. Similarly, horizontal barring on the underside is uncommon among birds of prey (though many species have vertical streaks), and is almost entirely restricted to woodland dwellers. It helps to break the outline when the bird is perched among twigs and branches, and when it is in flight. On the other hand, the blue back is probably linked with the diet, and not with the habitat. Blue is a rare colour among raptors, and is found only in bird feeders, including the open country Peregrine Falco peregrinus and Merlin F. columbarius, as well as in several woodland accipiters around the world. Perhaps, therefore, this colour is of some advantage in bird feeding. All bird-eating raptors also have relatively long legs and central toes, though in none are these features so marked as in the Sparrowhawk and similar small accipiters. Such legs give a long reach, and can be shot out to seize the prey the moment it comes within range. The behaviour of Sparrowhawks also conceals them. Except during aerial display, they seldom show themselves, spending almost all their time in cover. When in the open, they usually fly close to the ground, taking advantage of any slight cover or irregularity in terrain to keep themselves hidden. The flight consists of brief series of rapid flaps alternating with long glides. The bird is more noticeable when flapping, and during a surprise attack on prey the final approach is usually made on a glide. The hawk is then extremely hard to see, as I was able to confirm many times when a bird flew towards me at the nest; the body is so slim and the wings so still and razor thin that the front profile is practically invisible until the bird is almost on top of you. While a person is at the nest, the female usually stays nearby, flying from tree to tree, but even then it is hard to get a view, as the bird keeps itself hidden by foliage. In contrast to Kestrels, which often sit conspicuously on tree tops or power poles, Sparrowhawks perch hidden within the canopy. The foot is well fashioned for its main tasks of grasping and holding the prey (Fig. 3). The three front toes differ in structure. The outer one is fairly long and slender, the central toe is very long and slender, and the inner is short and thick, like the rear one. The front central toe is long enough to extend the reach appreciably; near the end of this toe, on the underside, is a long fleshy protruberance onto which the claw can be closed, to form

a pincer. A hawk in a cage can reach between the bars, stretch its leg and toe to full extent and pick up an object with this pincer. Normally, the pincer

30

The Sparrowhawk

Fig. 3. Feet ofSparrowhawk, showing the structure ojthe different toes, and the protruberance on the central and outer toes onto which the clara can be closed.

probably serves to grip the feathers of a prey just too far from reach to be grasped in the foot itself. The outer toe has a similar structure, but less well developed. The protruberance also serves to close the gap completely when the claw is bent onto the toe, another advantage when a victim is caught by its feathers. No other European raptor has these toe protruberances as well developed, so that in these other species the claws cannot be closed onto the toes without leaving a gap. The claws themselves are curved almost to the extent of a semi-circle, and are needle sharp. Those of the slim central and outer toes of each foot are fairly slender, but those of the inner and rear toes are large and powerful. These latter two claws oppose one another, and are responsible for holding, and in many cases killing, the victim, once it is caught. They can each penetrate more than one centimetre into the victim's body. For feeding, the hawk uses the front inside claw on each foot to clamp the prey to a perch while it is plucked and torn apart with the bill. The bill itself is small for a raptor; it is bluish in colour, inside and out, and darker towards the tip. It is not used for killing prey, but for plucking and rending flesh. In this last action, the upper mandible acts as a hook, as the bird pulls at the prey held under its feet. Only the bill tip is sharp, not the edges, so that the bill is a poor cutting tool. The bill opens to reveal a mouth nearly two centimetres across (almost the whole width of the head), enabling the bird to swallow relatively large food pieces and awkward bones.

Sex and age differences in plumage Plumage differences between the sexes are slight. As a rule, cocks are bluer (less brown) on the dorsal surface than are hens, more orange and

The Sparrowhawk

31

less distinctly barred on the underside, and have less white on the head in the form of eye stripes and crown patch. In most males the white patch at the back of the crown is evident only when the feathers are raised (as in display), because the individual white feathers are edged with blue. In most females, however, the crown patch is visible, even when the feathers are flat, because the white extends almost to the tips. In both sexes, some individuals are generally darker than others, having more orange on the underside, and broader breast bars. On the underside, the feathers range in background colour between greyish-white, through cream, to orange, in different individuals. Each feather has several horizontal lines, variously patterned in brown and orange, which continue from feather to feather to form the bars. The undertail coverts are white and rather fluffy; at times of display they can be fanned on each side of the tail. The legs and feet are bright yellow, and the claws are black. The cere, brow and eyelids are yellow or greenish-yellow. Sparrowhawks don the bluish adult dress when approximately one year old. Until then, they have the brownish 'juvenile' plumage, acquired in the nest. The basic pattern of the two plumages is similar, but feathers which are blue-grey in the adult are dark brown with buff-red tips in the first-year bird. The juvenile underside is also different, with a background colour varying from whitish, through cream to medium brown, and dark brown bars, again varying greatly in width from bird to bird. In most juveniles, on each chest feather, the terminal bar is transformed to a spot or heart-shaped mark, with a paler centre. The general effect is to darken the chest, and make it appear more streaked than in adults, though this streaking is less marked injuvenile Sparrowhawks than in thejuveniles of some other accipiter species. Despite great individual variation, there is little consistent sex difference in the plumage ofjuvenile Sparrowhawks. Overall, the plumage of the juvenile gives even better camouflage than that of the adult. This has parallels in many bird species, and is presumably because the young birds, lacking experience, have greater need for camouflage, while the adults, being more concerned with territorial and mating behaviour, have greater need for conspicuousness. These two requirements are conflicting, and the balance is drawn differently in the two age-groups. On a general view, the juveniles appear the same size and shape as the adults. They are effectively full grown when they become independent of parental care, at about two months of age. However, the wing-length, measured from the carpal joint with the feathers flattened and straightened on the rule, is still slightly shorter, on average, in juveniles than in older birds. In our samples this trait was statistically significant only for hens, which increased between year 1 and year 2 of life, and again between year 2 and year 3 (Table I). When the feathers arc measured individually, the juveniles are found to have slightly shorter primaries and slightly longer secondaries, so that their wings are not only shorter but broader and less pointed than those of adults, and have a larger surface area. The juveniles also have slightly longer tail

32 The Sparrowhawk feathers by about 2% in males and 5% in females. These features, combined with reduced weight, give the juveniles more lift than the adults, enabling them to fly more slowly without stalling, and with less energy. On the other hand, juveniles probably cannot develop such high speed as adults. Similar morphological differences between juveniles and adults occur in some other raptors, and are presumably adaptive, the one shape being appropriate to juveniles with their inexperience, and the other to the more practiced adults. If there is any difference in hunting techniques between the two age-groups, this has not yet been described.

Eye colour As in many other birds of prey, the eyes in Sparrowhawks are especially striking. Together with the well-defined brow, they give the birds their fierce, intense appearance. Moreover, the eyes change in colour during the hawk's life. At hatching the iris is brownish-black, but it soon becomes paler, changing through olive-brown around day 14 to greenish-grey by day 20, the then to pale greenish-yellow by the time of fledging at around day 28. A month later, in the first August of life, the eyes are pale lemon-yellow, sometimes with a faint greenish tinge. Then, over the next few years, the eyes of most individuals darken, in males changing to orange or even to wine red. To gain more information, we matched the iris colour of every full-grown Sparrowhawk trapped to a standard colour chart. Then, comparing one age group with another, we found that the colour changes occurred earlier in males than in females, so that at any given age, males on average had darker eyes than females. However, the variation within age groups was so great that iris colour "vas of no practical help in ageing individuals precisely. The most consistent changes occurred between the first and second years, when the birds could anyway be aged more reliably on plumage. In a small proportion of males the eyes remained yellow throughout life, in the majority they turned to orange, and only in a few did they turn to red. In most females, the eyes remained yellow, even in the oldest individuals, though in a few they became pale orange (Table 2). The significance of these changes remained a mystery, but Snyder & Snyder (1974) suggested for American accipiters that eye colour might serve in individual recognition and mate selection. We had no way of testing this view in Sparrowhawks.

SEXUAL SIZE DIMORPHISM

In almost all birds of prey, the female is bigger than the male, but the Sparrowhawk is extreme in this respect. It shows a greater weight difference between the sexes than any other raptor in the world. In winter, adult males weigh about 150 g, on average, and females almost twice as much, at 290 g. At the start of breeding, when females put on extra weight for egg laying, the difference is even greater. In body length (bill tip to tail end), the female averages some 15% more than the male, and in wing span about 20% more.

The Sparrotohatok

33

These and other measurements are given in Table 3, which shows that the degree ofdifference between the sexes varies with the feature that is measured. During the breeding season, the separation of duties between cock and hen is also more marked in the Sparrowhawk than in most other raptors. The cock provides the food from before laying until the young are about halfgrown, while the hen incubates the eggs and tends the young; this includes brooding and sheltering the chicks, dividing up food for them, and attending to nest hygiene. Only when the young are large enough not to need brooding does the hen start hunting again, initially in the nest vicinity, and later further afield. In most raptors, the male incubates the eggs for part of each day, but the cock Sparrowhawk does not, probably because he is too small to cover the clutch effectively. Two questions have been asked about size dimorphism in birds of prey: first, why are females larger than males, and second why does the extent of dimorphism vary between species? Among raptors in general, dimorphism is linked with diet. Those species which feed on immobile or slow-moving prey show little or no size dimorphism; those which feed on faster prey, such as reptiles or mammals, show somewhat more; while those which feed on the fastest and most agile prey (birds) are the most dimorphic of all (Newton 1979). This link with diet is undisputed, but there is wide speculation on the reasons for it. As the differences in size and behaviour are sexual, they are presumably linked to some aspect of breeding, as well as to diet. The most obvious way in which raptors as a group differ from other birds is in the problems they face in getting food: they take relatively large items which can fight back or can move quickly and escape. The capture of alert, fast-moving prey, such as birds, demands unusual speed and agility on the part of the predator. Agility is relative, and is helped by the predator being as close as possible in size to its prey, and carrying the minimum possible surplus weight, in terms of body fat. Andersson & Norberg (1981) showed that, in five out of six aspects of flight, a small bird is faster and more efficient than a large one of similar shape. These included maximum acceleration and speed in horizontal flight, maximum rate of climb, maximum angular roll acceleration and minimum turning radius. Only in terminal diving speed is a large bird slightly faster. In general, therefore, small size gives greater facility in prey capture. On the other hand, body reserves are a help in survival and breeding, and the accumulation of reserves is greater in a large bird and also more efficient, because a large bird uses less energy per unit body weight than a small one. Like some other bird species, raptors accumulate huge body reserves in preparation for breeding. Indeed, the size of fat reserves accumulated by female Sparrowhawks in spring largely determines the number of young that they produce (see later). These two requirements, of extreme agility and large body reserves, are conflicting, as the one interferes with the other. In the Sparrowhawk, as in most other raptors, the compromise in body size is drawn differently in the two sexes, because the male retains the hunting

34

The Sparrowhawk

role throughout breeding, and the female adopts the storage role. The male hunts to feed both sexes and the female stores the body reserves on which the breeding of both sexes depends. The reason that the female, rather than the male, adopts the storage role is presumably that she has to produce the eggs anyway, which in itself entails a weight increase and associated reduction in agility. Also, if the female attacks prey at this stage she risks damaging the eggs inside her body, a risk which is greater in raptors than in other birds, because of a raptor's particular methods of obtaining food (Walter 1979). These problems are overcome by the female raptor ceasing to hunt before laying, and becoming reliant on the male. Being large also helps in the production and incubation of eggs (less energy per unit body mass). In addition, by being confined to the nest during the egg and small chick stages, the female is inevitably more involved in nest defence than the male. Nest defence is another task in which size may help, and which is more developed in raptorial than in most other birds. For these various reasons, therefore, if one sex is to be larger than the other, it is advantageous in birds of prey that it is the female. * In addition, as explained in Chapter 25, females compete for males at breeding time, so larger female size may also be favoured by 'sexual selection' (Darwin 1871). These ideas can be extended to other raptors. The needs of agility and food storage conflict in almost all species,' but to different degrees. Those species which exploit slower moving prey have less need for agility than the bird-feeders. In consequence, the compromise in body size between these opposing needs may be drawn differently in sexes ofdifferent species, resulting in the link between dimorphism and prey type mentioned above. Whatever the evolutionary history, the great size difference between the Sparrowhawk sexes has a number of consequences. One is that the male catches mainly smaller prey than the female (Chapter 8). This presumably reduces competition between the sexes, and perhaps enables a pair to live together in a smaller area than they might otherwise need. With different prey, moreover, the total food supply available to the two sexes is unlikely to be equal. Thus in some areas, depending on the prey-supply, one sex might persist in greater numbers than the other, leading to locally biassed sex ratios. In general, the smaller males are more vulnerable to various mortality agents. They are quite often killed by females, and possibly more by other predators too. They are also more prone to food shortage, as they can survive without eating for only about half as long as females can (Chapter 23). As a percentage of body weight, the fat contents of males are similar to those of females in winter, but as males have relatively higher energy needs (resulting partly from body size-area relationships), their body reserves deplete sooner. Males are thus likely to die at times when females might live.

* Any ideas on the evolution of reversed size dimorphism, are impossible to test properly. There is a considerable literature on the subject, of greatly varying quality; for further reading consult Earhart & Johnson 1970, Reynolds 1972, Snyder & Wiley 1976, Newton 1979, Walter 1979, Anderson & Norberg 1981, Wheeler & Greenwood 1983, Mueller & Myer 1985 and Temeles 1985.

The Sparrowhawk

35

SEX RATIOS

In view of the size difference between the Sparrowhawk sexes, which develops in the young while still in the nest, the sex ratio is of special interest. Several writers have suggested that male Sparrowhawks might survive less well in the nest than females because males are smaller and supposedly lose in competition for food. It was therefore relevant to examine the information available in detail, and find what holds in practice (Newton & Marquiss 1981) . Young Sparrowhawks can be sexed at a glance by size even before they are half grown, so when the young were ringed each year, cocks could be distinguished easily from hens, with no risk of error. Some 651 broods examined at this stage contained about equal numbers of each sex, a total of 1,102 (50.95 % ) males and 1,061 (49.05 % ) females, giving a ratio not significantly different from unity. Individual broods contained from one to six nestlings. In each size of brood, a number of sex ratios was possible. Thus a family of five might contain five males, or one male and four females, two males and three females, and so on through to five females. Whatever the brood size, however, the frequency with which broods ofdifferent sex composition were encountered in the wild did not deviate significantly from the frequencies expected if the chances of each young being male or female were equal (Fig. 4). Among broods of five, for instance, there was a theoretical 3% chance of all young bcing male or female, and this was the frequency with which all-male or all-female broods were in fact found. The situation did not differ in Sparrowhawks from that prevailing in human families. Moreover, no significant variation in the overall sex ratio at fledging occurred between Sparrowhawks in different areas, nor between years in the same areas. By the time these figures were obtained, near to fledging, some mortality of eggs and young had occurred. I t was possible, therefore, that the sex ratio in the newly laid eggs was unequal, and that mortality then fell more heavily on the surplus sex to produce the equal ratio at fledging. I could deduce the sex ratio in the eggs retrospectively from nests in which all eggs laid gave rise to surviving young. On this basis, the ratio at laying was 200 males: 205 females, again not significantly far from equal. Such a sample might not have been representative, because it took no account of nests in which mortality of eggs and young had occurred. But taking the data at face value, they gave no evidence for differential mortality between the sexes at any stage from cgg to fledging, nor that the overall sex ratio at fledging was other than equal. The mortality which occurred among nestlings usually fell on the latest hatched, and not selectively on one sex or the other (Chapter 13). Studies in continental Europe, involving smaller samples, have also reported an approximately equal sex ratio among the young at fledging (e.g. Gedeon 1983). Whereas it was simple to record the sex ratio of nestling Sparrowhawks in an unbiased manner, for free-flying birds it was practically impossible.

36

The Sparrowhawk

60 50 40

1

30

N

YOUNG

= 75

20 10 M

50

YOUNG

40

2

30

N = 99

20 10 MF

MM

FF

50 40

3 YOUNG

30

N = 161

20

en

10

0

a a

Cl:

co

"*'

MMF

MMM

FFF

MFF

50 40

4

30

N= 184

YOUNG

20 10 MMMM

MMMF

MFFF

MMFF

FFFF

40 30 20

5 YOUNG

10

N= 121 MMMMM

40

MMMMF

MMMFF

MFFFF

MMFFF

FFFFF

30

YOUNG

20

6

10

N= 11 MMMMMM

MMMMMF

MMMMFF

MMMfFF

MMFFfF

MFF FFF

FFFFFF

BROOD-COMPOSITION •

OBSERVED

D

EXPECTED

Fig. 4. Sexes ofyoung in Sparrowhawk broods. The frequencies with which broods of different sex composition were observed in the wild did not differ from the frequencies expected on the assumption of an overall 50: 50 ratio. M: male; F: female. Drawn from data in Newton & Marquiss 1979.

The Sparrowhawk BIRDS SEEN

4·0

E

37

2.0

CD 0.. (fl

1·0

CD

E

BIRDS

~ 2·0

TRAPPED

E

z 1·0

CF

FW

OF

Fig. 5. Numbers offemales per male Sparrowhawk seen and trapped in different habitats, south Scotland. CF: continuous forest; FW: farmland with small woods; OF: open farmland, with no woods. Males predominated in forest and females in open land. From Marquiss & Newton 1982. A ratio of more than two females to eve'(y male had previously been noted b.-y Ash (1960) in an area consisting mainly offarmland in southern England.

This was because the two sexes used habitats differently, the males hunting mainly in forest and the females largely over farmland and other open country. This was evident from birds seen in different habitats, from birds trapped in different habitats (Fig. 5), and also from radio-tracking work (Chapter 5). This habitat difference was of interest in its own right, but it meant that the overall sex ratio among adult Sparrowhawks could not be found by observation or trapping, because the results depended on landscape. However, independent evidence for an unequal sex ratio among adult birds came from ring recoveries which revealed a higher death rate in males, especially in the first year of life (Chapter 23). Given an equal ratio at fledging, this could only result in a surplus of females among Sparrowhawks of breeding age. The estimated first-year mortality of 69% for cocks and 51 % for hens, would give a cock: hen ratio in the second year of 1 : 1·6. The annual mortality of cocks in later life also seemed greater, but the sample was too small to be sure. I t meant, however, that the true sex ratio in the adult population as a whole might be tipped even further in favour offemales.

The habitat difference Why males hunted more in woodland than females was not certain. As males were smaller and more agile, they may have been better able to catch prey in enclosed situations. This alone might make prey more available for males in woodland, but in addition the prey species favoured by males were

38

The Sparrowhawk

probably more numerous there, while the larger prey of females were D10re abundant in open areas. In addition, by keeping in cover, the cocks probably stood less chance of being preyed upon by the larger hens (I recorded D10re than ten such instances). Irrespective of the reasons, males and females clearly exploited a landscape differently, and so could perhaps co-exist more easily than if both hunted the same habitats to the same degree. A similar habitat difference was noted among Sparrowhawks in the Netherlands, so it may well be widespread (Opdam 1975).

WEIGHTS

The routine trapping of Sparrowhawks in south Scotland provided enough weight values to examine the diurnal and seasonal trends. Many birds were caught in baited traps, so the samples for certain months could have been biased in favour of hungry or weak individuals, though I had no way of checking this. Few birds had previous food in their crops, but some may have been in a cage trap an hour or more before being examined. Other birds were caught in noose traps on nests, and were weighed within minutes of capture. In whatever way the birds were caught, their weights showed little variation through the day, apart from a slight depression in the middle hours. 'This uniformity presumably resulted because Sparrowhawks took their food in 1-3 large meals each day, separated by fasts lasting up to several hours. The weight of anyone individual might increase with each meal, and then decline to the next, but as different individuals ate at different times, depending on their hunting success, the mean trend of the trapped sample would show little diurnal pattern. The main variation in weight was from month to month through the year. In cocks even these changes were relatively slight, however, from a low in August (mean 143 g) near the end of one breeding season, to a peak in March (mean 155 g), near the start of the next (Fig. 6). In all months, first-year males were lighter than adults. This was especially marked in August-September, soon after the young had become independent of parental care, but over the year as a whole they averaged 6·5 g (4 % ) less than adults. They also had slightly shorter wings and other measurements (see above), so part of the difference may have been due to slightly smaller body size, and not only to poorer condition. Whereas males fluctuated in weight over the year by about 8% of the lowest monthly mean, females varied by as much as 33% (Fig. 6). This was associated with the females' role in breeding, and the accumulation offat, as discussed above. Females weighed least in August-September (mean 260 g), near the end of one breeding attempt, and most in May (mean 325 g) around the time of laying at the next attempt. Many individuals exceeded 360 g at laying time, and one even reached 430 g, perhaps 38% greater than her weight in the preceding August. Hens which did not breed did not achieve

The Sparroiohatok

39

Females

360

340

•

320

300

280

260

:§ _240

I I i i

I I

I I

•

)i

•

I

i

I

i

I

i

•

I

i

s:

en

Q)

~220 200

'--~------'T'"---,.--.--.,----..,.--~---r-----r----r--r---r---

Males

180

160

14

I

i

Aug

I • I Sep

• I

Oct

•, I

Nov

l • I I

Dec

I

I

• I

Jan

Feb

l

I

•II

Mar

Apr

I

i

May

I

•I

Jun

I

T Jul

Fig. 6. Mean weights (±SD) of adult Sparrowhawks each month through the year) south Scotland. F:'om Newton et a11983.

such a high weight in May, and most remained around 300g. Weights of first-year hens followed a cycle similar to those of adults but, as in cocks, remained lower throughout. Again, the difference was especially marked in August-September, in the first two months of independent. life, and lessened thereafter, averaging about 11·2 g (4 % ) less than adults over the whole year. The weights of the many individuals caught at different times of year followed the mean trend of the population, described above. In neither sex did weight increase with age beyond the second year, despite the slight increase in wing-length of females. Many small bird species of temperate regions put on extra weight in winter,

~

~

~ ~

~ ~

~

Bj ~

s ~

~

..~;>-."

?

?

~

Breeding range

~

Wintering range

?

Uncertain wintering range

Fig. 7; Breeding and wintering ranges of Sparrowhawks. Numbers indicate races: (1) A.n.nisus in west Palearctic; (2) A.n.nisosimilis in east Palearctic; (3) A.n.melaschistas in the Himalayas; (4) A.n.punicus in North Africa; (5) A.n.granti on Madeira and the Canary Islands; (6) A.n.wolterstorfi on Corsica andSardinia.

The Sparrotohatok

41

largely to help them survive the long cold nights (King & Farner 1966, Newton 1969). Sparrowhawks, being larger, can presumably more readily withstand short periods of deprivation, and showed no sign of a mid winter peak in weight. Most carcasses that were analysed in winter were not particularly fat, with an average of 5 g in males and 10 g in females, though some individuals had up to 25 g. Nor was there any indication that moult influenced weight. Both sexes were heavier in the early stages of moult than towards the end, but in most individuals moult had finished for several weeks before weight increased again in November. Undoubtedly, the major weight changes were associated with breeding, particularly in hens, and these are discussed further in Chapter 10. For the rest of the year, there was presumably no point in the birds being heavier than necessary for survival, because any surplus weight would lower flight efficiency, perhaps increasing their own predation risks.

RACES AND RELATIONSHIPS

The Sparrowhawk has a wide distribution, breeding in woods and forests across the Palearctic region from Ireland to Japan, and from the tree-line in the north, southwards into the Mediterranean region, including North Africa (Fig. 7). The northernmost populations are migratory, and in winter the species reaches southwards into desert scrub in Africa, the Middle East, India and southeast Asia, but is not known to extend south of the equator. Despite its extensive range, the Sparrowhawk shows little geographical variation in size or colour, and populations hardly vary from one corner of the Eurasian land-mass to another. This is true with respect to both adult andjuvenile plumages. What slight variation exists is mostly clinal (gradual), though six subspecies are recognised, including four which are geographically separate in the breeding season (Fig. 7). The two island subspecies are smaller and darker in all plumages than the others, while the far eastern Asiatic birds tend to be larger than the rest. Over most of its range, the Sparrowhawk breeds alongside its larger relative, the Goshawk, in which species males weigh about 800 g and females about 1,200g. Together, the two species form a graded size series, from male and female Sparrowhawk to male and female Goshawk. Between them these four sizes of accipiters exploit as food almost the entire bird fauna of wooded parts of the Palearctic, excluding only the very largest species. All of them can kill prey heavier than themselves, but the Goshawk eats more mammals than the Sparrowhawk. It also nests in more mature, more open forest, and thus differs both in diet and in habitat from the smaller species (Opdam 1975, van Beusekom 1972). In the south-central part of the breeding range, the Sparrowhawk overlaps with two other small accipiters, the Levant Sparrowhawk A. brevipes of southeast Europe and the southern Caspian, and the Shikra A. badius of southern Asia and Africa. Both inhabit somewhat drier, warmer, and more sparsely

42

The Sparrowhawk

timbered country than A. nisus, and in the regions of overlap nisus usually breeds in the mountains and breoipes or badius in the lowlands. These other species also differ in diet from nisus, taking mainly lizards and large insects, but some small birds and mammals. Linked with this, they have shorter, thicker tarsi and toes than nisus, less developed toe-protruberances, and tougher scales on their legs, a feature usually associated with reptile-eating. The females are about the same size as female nisus, but the males are bigger than male nisus, so that size dimorphism is less. In plumage pattern, they are similar to nisus, with paler heads, but in the field they can be distinguished by their different flight silhouettes, with narrower, more pointed, dark-tipped wings. The juveniles have thicker chest. streaks, including a pronounced median streak, and are not barred like the adults. The two species breoipes and badius are so similar to one another that they were formally classed as one species, but as they differ in wing formulae and certain plumage details, they are now considered separate, though clear ecological counterparts. A notable feature of A. breoipes is that it is gregarious on migration, occurring in flocks of up to 500 individuals, as it moves between Eurasia and Africa. One of the best places to watch this migration is at Eilat in Israel. In the east of its breeding range, the Sparrowhawk overlaps with three other small accipiters. One of these, the Frog Hawk A. soloensis, eats what its name implies, so is ecologically quite distinct from nisus. It also has a more conspicuous lifestyle, frequently perching in the open to swoop on prey. The two other eastern species, the Japanese Lesser Sparrowhawk A. gularis and the Besra A. virgatus, are both bird feeders. They occupy similar habitats and have similar plumage patterns to A. nisus. However, they breed mainly sou th of nisus, and are also smaller, taking smaller prey. They too have been classed as two forms of the same species, occupying different regions, but the most recent view is that they are distinct (Mees 1975). The situation in eastern Asia, with three bird-eating accipiters of different size, namely A. gentilis, A. nisus and A. gularis (virgatus), parallels that in North America, where again three species, rather than two, exploit as food the local bird population. These include the Goshawk (slightly larger than the European form), the Cooper's Hawk A. cooperii (larger than the Sparrowhawk), and the Sharp-shinned Hawk A. striatus (smaller than the Sparrowhawk). These species divide food and habitat between them in the same way as the three species in eastern Asia and the two in Europe, differing in the size of favoured prey, and in the type of woodland that they nest in (Storer 1966, Moore & Henny 1983, Reynolds et all982, 1984). The colour pattern of the Sparrowhawk, with bluish back and barred breast, is typical of almost every other accipiter occupying boreal, temperate, Mediterranean and some tropical habitats throughout the world, including all the species mentioned above, except the Frog Hawk. As argued earlier, such a widespread similarity of plumage pattern suggests strongly that it is adaptive to this way of life. I t is mainly the tropical and sub-tropical forest species which deviate from it. Most have similar plumage patterns

The Sparroiohaiok

43

to the northern forms, but with darker dorsal colours and paler ventral ones. The whole plumage is thus more contrasting than in the northern forms, as befits birds living where the bright sun produces more intense light-dark foliage patterns (Wattel 1981). The Black Sparrowhawk A. melanoleucos of Africa is extreme, for the dorsal plumage is wholly black and the ventral, through loss of the barring, is pure white. In other species, the orange and red components of the plumage are intensified, spreading in some to form a collar round the neck, but the net result is the same, producing a much more striking and highly coloured accipiter, which is nonetheless cryptic in its particular environment. Together the various accipiter species comprise the largest genus of all diurnal raptors. I t contains about 50 species in all, which between them occupy forest and scrub habitats throughout the world. They include small to medium-sized raptors, all of which have a similar shape to the European species. The genus is placed in the subfamily Accipitridae.

SUMMARY

In shape, with slim body, long tail and short broad wings, the Sparrowhawk resembles other woodland raptors; in its long legs and foot structure with long central toe, it resembles other bird-feeders. The barred underside is also common among woodland birds of prey, while the blue dorsal surface is typical of bird-eaters. These features can thus -be regarded as adaptive, either to the woodland habitat or to the bird-diet, having evolved independently in raptors of different genera by convergence. Plumage differences between the sexes of the Sparrowhawk are slight, but in first year birds the upper parts are brownish instead of blue-grey. Eye colour changes during the bird's life; in most adult males the iris is orange and in most females it is yellow. In males, body weight alters little through the year, being lowest in August at the end of one breeding season and highest in March near the start of the next. In females, the weight cycle is similar, but the increase before breeding is much greater, reaching a peak in May, with the accumulation of body reserves for breeding. The size difference between the sexes is greater in the Sparrowhawk than in any other bird of prey, with females about twice the weight of males. The sex ratio among well-grown nestlings is equal, but later swings in favour offemales, as males have a higher mortality, especially in their first year. The Sparrowhawk breeds across the Palearctic region, in boreal, temperate and Mediterranean zones. In the south-central parts, it overlaps with the Levant Sparrowhawk and Shikra, which eat more reptiles and insects, and in the southeast with the Japanese Lesser Sparrowhawk and Besra, which take smaller birds, and with the Frog Hawk which eats amphibia. Over much of its range, the Sparrowhawk also overlaps with the Goshawk, a much larger species, which eats larger birds and mammals.

CHAPTER 3

Nesting habitat

Sparrowhawks will not nest in all kinds of woodland, but are selective. It is not entirely a question of tree species or of size of wood, but of internal structure. Ideally a nesting wood should ofTer plenty of cover, yet at the same time be open enough to permit easy flight between the trunks and branches. Woods can be unacceptable, either because they are too thick

Nesting habitat

45

or too open. Young woods, before they have been thinned, are in the 'too thick' category, while old woods, subjected to several thinnings, are in the 'too open' category. The age of a wood, and its management, are thus the main factors influencing the period over which it remains attractive to Sparrowhawks. During this period, birds may nest there year after year, usually building a new nest each time in a tree near the old nests, forming a 'traditional' nesting place. In the Scottish study areas, almost all the woods were coniferous, and managed for timber production. At the outset, the trees were planted close together, at 1·5 X 2 m or 2 X 3 m, and for some years they were too small and dense for Sparrowhawks to nest in. Most woods were occupied for the first time within 1-2 years after their first thinning, at about 20 years of age. At this age the trees were around 8-l0m high. With thinnings every five years or so, woods remained suitable until about 40-50 years of age, when they became too open or were clear-felled. Thus each wood was acceptable for some 20-30 years of its life, though this period might differ in other regions, depending on tree growth and management. Managed conifer woods typically consisted of even-aged stands, with trees of uniform height and spacing. Because of the closeness of planting, the trees had narrow crowns, in which only the upper third of branches were alive at anyone time, the lower being dead and leafless. Once thinning had begun, the bare ground became littered with dead branches and logs. Only towards the end of its life did a stand become open enough to permit the growth of continuous ground cover and shrubs, though this occurred earlier under pine andlarch than under the thicker-foliaged spruce or fir. The few unmanaged woods in the study areas were mainly of birch, where natural regeneration had been allowed. They had more varied tree species and structure than the planted woods, and Sparrowhawks nested wherever the structure suited them.

Tree spacing The most useful single measure of the suitability ofa wood was the density of trees, or, in practical terms, the distances between them. In each wood or sector of forest, I paced out the distances from trunk to trunk as I walked through. Then by noting in which sectors Sparrowhawks nested, and in which they did not, produced the pattern in Fig. 8. Nesting occurred chiefly where the mean distance between trees at ground level was in the range 2-4 m, but seldom where trees were closer than this or further apart. Little difference occurred between tree-species in this respect, although pine was favoured at a slightly younger, denser stage than larch or spruce. Preferences seemed similar in broad-leaved woods, though few were available for study. In some of the woods concerned, Sparrowhawks nested only once during a ten year period, and in others they nested more than once. The preference for 2-4 m tree spacing was even more marked when account was taken of nesting frequency, for areas with spacing closest to the preferred were used most often. Some areas with wider tree spacing had been used in former

46 Nesting habitat Pin e Cf) (])

o co 0.

~l

•

rm~

~

I

D

0

Nests

DNa nest s

m

0

La r c h

OJ

c

n ~ :OO.~~_-+I11UL~~ 10 I 000

0I

n

q

o

Spruce (])

£2

E

3

21

n ~1IlIIIlIlI..-1~_I...-LLO_------._---L...LO------.JJ,-_-----, I

3 4 5

Mean

distance

between

7

trees

(m)

Fig. 8. Use of woodland areas with different tree spacing by nesting Sparrowhawks. The thickest and most open areas were avoided. Data from south Scotland, in areas lacking Goshawks. Excludedfrom the Figure are a few young woods whosegeneral canopy was too thick for Sparrowhawks to penetrate, but where the birds gained access by a streamor path, and nestedthere.

years, before the trees had been thinned out. So it was not that these localities were unsuitable for nesting, only that the stands concerned had become too open. Tree spacing was a convenient measure of acceptability, but other features varied in parallel. The further apart the trees, the- older and larger they tended to be, the higher and more open the canopy, and the more luxuriant the shrub and ground vegetation. Anyone of these features could have reflected the suitability or otherwise of a wood for nesting Sparrowhawks, To the practiced eye, however, no measurements were necessary, as it was at once obvious from a glance inside whether a wood was suitable.

Size and type ofwood Of 290 woods in the Annandale area, about a third were occupied by breeding Sparrowhawks at least once in the ten year study. Whether a given wood was used depended on its size and type, as well as its structure. Only 14% of woods less than one hectare in area were used for nesting, but this proportion rose with increasing size of wood, so that all woods exceeding 20 ha held breeding Sparrowhawks (Table 4). In addition, more of the conifer than of the mixed woods in the area were used for nesting, and more of the mixed woods than the wholly broad-leaved woods. Most of the broadleaved woods were small, but their usage was still less than that of conifer

Nesting habitat

47

woods of similar size (Table 4). Evidently, Sparrowhawks preferred large woods to small ones for nesting, and coniferous to broad-leaved stands. Whereas small woods contained only one nesting place, large woods often contained more than one. However, the individual nesting places in large woods were also used more often than those in small woods. This was true for woods up to 100 ha in area, but in larger woods, the mean usage of individual places declined (Table 5). This was due to places in the middle of large woods being used infrequently, compared to those near the edges. The preference for conifers still held (Table 6). The smallest wood in Annandale to hold two pairs of Sparrowhawks was 20 ha in area, the smallest which held three pairs was 40 ha, the smallest which held four pairs was 60 ha, and five pairs 160 ha. On these figures, the area per pair rose with increasing size of wood, but little significance could be attached to this, because birds hunted partly outside the woods. Also, other woods much larger than these minima held two, three, four and five pairs respectively, the number in anyone wood depending partly on how much of it was suitable and partly on its shape, a long narrow wood holding more pairs than a square wood of similar area. Findings in Annandale could not always be applied elsewhere. This was because the size of wood acceptable for nesting varied with availability. In all the districts studied, Sparrowhawks were more likely to be found breeding in large woods than in small ones. If a district had enough large woods, no small woods were used. This was the case in Deeside, Aberdeenshire. So from this district alone, one might conclude that Sparrowhawks 'required' large woods. In other districts, where large woods were lacking, the birds could only use small woods. This was the case near Dinnington, Northumbria. So it is that in some parts of the British Isles, Sparrowhawks are found only in extensive forests, whereas in others they are found in such unpromising places as overgrown orchards and quarries, rough old hedgerows, overgrown railway lines, and belts of trees along streams. Such sites are the commonest nesting places in some sparsely-timbered parts of Ireland, for example. Someone experienced in finding Sparrowhawks in one part of the country might thus have quite the wrong search image elsewhere, if the habitat differed. Even so, I know of no instances of Sparrowhawks nesting in isolated trees, apart from one in a city, mentioned later. Regional differences in nesting sites arise because the numbers of Sparrowhawks in anyone area may be determined mainly by the food supply, and then the hawks settle into whatever patches of nesting habitat happen to be available locally, forest or scrub patches as the case may be. There may be enough such cover to provide all the hawks with a chance to breed, or only a proportion of them. The nesting places themselves may supply a large part of the prey, or practically none, compared with other habitats in the vicinity. With tree species, too, there is an order ofchoice. The preference for conifers is so strong that when conifers are freely available, you may not find a single nest in a broad-leaved tree. But in districts where conifers are lacking, the

48

Nesting habitat

birds have no option but to use broad-leaves. Areas of scrub-oak and birch are preferred, but almost any species is accepted. Old bird books give the impression that Sparrowhawks could breed almost anywhere. On a national scale, the choice is indeed wide, but in anyone area the birds are usually narrow in their choice of nesting habitat, tree spacing being the most consistent feature. The preference for woods over small clumps and single trees is presumably because of the greater protection that woods provide. A nest in an isolated' tree is visible from afar, whereas one in a wood lies hidden among hundreds of similar trees, some of which contain old, empty nests. A predator therefore has much searching, and perhaps many fruitless climbs, to obtain a meal from a nest in a wood. The adults themselves also gain cover from the trees, and can there perform the various activities associated with nesting much less conspicuously than in the open. As discussed later, even the choice of the tree, and of the site within the tree, seems to be aimed at concealment.

Topography Sparrowhawks also prefer low points in the local topography. In large forests in undulating terrain, they favour the valleys, hollows or other places which are lower than average. Even on hillsides, the nest is usually nearer the bottom than the top, and is often in a gulley or other dip in the land surface. Again, this is a preference rather than an absolute need, and where no choice is available, Sparrowhawks occasionally nest in woods on hilltops. The preference for low points could again reflect the need for cover and seclusion, and is anyway in keeping with the general 'low profile' lifestyle of the species.

Nesting in cities Although most Sparrowhawks nest in rural areas, they are not averse to breeding in towns, or even in large cities, where prey are plentiful. Small woods, tree clumps and scrub patches on waste land or in parks, cemeteries, hospital grounds, large gardens and university campuses are frequent sites. In some cities the birds have accepted sites with much wider tree-spacing than is usual in rural areas and where nothing better was available, the birds have even nested in single rows of trees, and in one instance in a broad conifer 20 m from other trees. In recent years, records have increased, and nests have been found regularly near the centres of Bristol, Birmingham, Glasgow and Edinburgh. The colonisation of Edinburgh occurred mainly after 1980, and within five years there were probably more than 20 pairs breeding' within the city limits (G. Carse). These hawks fed on the "vide range of birds available in parks and gardens, but with a greater proportion of House Sparrows and feral pigeons than is usual in rural areas (Chapter 8). The breeding in 'unlikely' sites in towns reaffirms how the accepted nesting habitat varies with what is available. Sub-standard sites may not be acceptable to all Sparrowhawks, however, and there may have been a change

1 Sparrowhawks. (Upper) Cock with prey. (Lower) Hen at nest. For much of the breeding cycle , duties are divided, the cock providing the food and the hen tending the eggs and young. Photos: R.J. C. Blewitt.

2 (Upper) Pair at nest, the cock in flight and the hen incubating. (Lower) Hen with young. Photos: R.J. C. Blewitt.

3 Views in the Annandale study area . (Upper) Ae Forest, a large upland conifer plantation. (Lower) Looking west across the Annan valley from near Wamphrey, showing the small woods in a farmland setting. Photos: I. Newton.

4 Views in the Eskdale study area. (Upper) Esk valley, north of Langholm. (Lower) Esk valley, south of Langholm . Photos: I. Newton.

5 (Upper) A birch nesting place at Tarras, near Langholm . (Lower) Nest in mixed birch-sallow wood. south Solway Plain. Photos: (upper) I. Newton; (lower) D. A. Ratcliffe .

6 Nesting habitat. (Upper) Young pine plantation, Annandale. (Lower) Checking a plucking post in an older pine plantation, Speyside. This is about as mature and open a plantation as Sparrowhawks use for nesting in rural Britain . Photos : (upper) I. Newton; (lower) D.A. Ratcliffe.

7 Nesting places in Edinburgh city include cemeteries (upper) and public parks (lower) . In both photographs the nest was in the large centre tree. Photos: I. Newton .

8 Methodology. (Upper) Taking a blood sample from the brachial vein of a live Sparrowhawk for analysis of organo-chlorine pest icide residues. (Lower) Feather patterns. Individual Sparrowhawks can be identified from the ir flight or tail feathers, providing equivalent feathers are obtained each year. The top row of feathers are second primaries from different females, showing the variation in patterns, while the bottom row shows pairs of feathers from the same four females in two different years . The feathers are moulted at the nest, and can be found on the ground nearby, enabling a reference collection to be compiled for each female. Photos : I. Newton.

•

•

I>

L

"

11

I

r

"

'J

J.

"

I

"

"

Nesting habitat

49

in what the birds will accept in towns following the development of a more tolerant human attitude. The key resource is the food-supply, and an abundance of small birds nearby is essential, whatever the choice of nesting place.

Proximity tolerance There is a limit to how close a Sparrowhawk pair will tolerate another for nesting. About 250 m is the shortest distance I have found between pairs, but at least three records describe much closer nests. Young (1973) reported two nests in two successive years about 22 m apart in a Dumfriesshire wood, and claimed that in at least one year a bigamous cock was involved; while Beer & Morgan (1978) found two nests 30 m apart in a Devon wood, and saw separate pairs at each nest, both of which raised young. Normally, however, nests ofdifferent pairs are much further apart, at 400 m or more, depending on region (Chapter 4). Thus anyone pair will usually render unavailable much habitat which is structurally suitable for nesting and could, if Sparrowhawks were colonial, support larger numbers.

OTHER FEATURES OF THE NESTING PLACE

In the woods of south Scotland, the nest tree was often near a stream, path, wind blow or other opening which provided easy access for the birds. At the same time the tree was usually some metres in from the opening, so that the nest remained sheltered and hidden. Only in young woods, where the canopy itself was too thick to provide flyways, was the nest placed on the very edge of an opening. Within about 50 m of the nest, and where possible upslope of it, were the regular plucking places where prey birds were stripped of their feathers before being eaten: old stumps, upturned tree roots or fallen logs that provided a firm perch just above ground level. However, where ground vegetation had become tall, or where the trees themselves were high, plucking took place in the canopy, on horizontal branches or on old nests. Near the current nest, the hen also had regular perching places, which gave her a view of the nest and plucking posts. Such perches were obvious to the human observer from the white droppings, and occasional pellets and moulted feathers below. These various features were typical of Sparrowhawk sites in all the areas that we visited.

NESTS AND NEST TREES

Of the nests found in south Scotland, about 45% were in spruce, 33% in larch, 14% in pine, 3% in fir and 6% in broad-leaved trees, mostly birch (3 % ) , but occasionally oak, rowan, sycamore, hawthorn, hazel or sallow. Such figures meant little, unless related to the availability of the different trees in the area. In many planted woods, Sparrowhawks had only one tree

50

Nesting habitat

species in which to nest, but in others they could choose between two or more species, which were grown intermixed or in adjacent stands. In mixed larch/spruce areas, 55% of nests were in spruce and 45% in larch; in mixed larch/pine areas, 67% of nests were in larch and 33% in pine; in mixed pine/spruce, 52% were in pine and 48% in spruce; in mixed spruce/broadleaved, 77% were in spruce and 23% in broad-leaved; while in mixed pinel broad-leaved, 92% were in pine and 8% in broad-leaved. From this, the different trees could be ranked by preference: larch> pine> spruce> broadleaved. This order was related to the growth form and foliage of the different trees, which influenced the situations that the tree offered for nest placement, and the degree of concealment and shelter. In mixed conifer woods, preferences changed as the stand matured. The birds initially chose larch or pine, possibly because of their more open structure, but with successive thinnings, as the trees became more open, the birds nested increasingly in the neighbouring spruce. In general, tree preferences in our study areas were similar to those in other parts of Europe, where the subject has been studied (see HaldMortensen 1974 for Denmark). It was interesting that the broad-leaved trees, in which most Sparrowhawks in Britain nested until the present century, were the least used. Larch and spruce are recent introductions, and pine was much more local before the days of extensive softwood culture. All three conifers offered good situations for Sparrowhawk nests, and larch had the additional advantage of being a superb building material. I ts twigs were easily broken off and, with nodules along their length, locked together to form a resilient slip-proof structure. Pine and spruce offered better cover early in the season, before eggs were laid, but as neither were preferred to the deciduous larch, this was presumably not an important factor. Twigs of both species were used for nest building, but pine twigs made a firmer, longer lasting structure than did spruce. Nests were usually built in the lower parts of the tree canopy, near the main trunk. Thus placed, they were in the best compromise position in the tree for overall concealment, as they were not too obvious from the ground or from the air. Most nests in south Scotland were in trees 10-20 m high, about 5-15 m off the ground, in the lower green branches (Table 7). Nests in pine and larch were usually placed further up the tree than those in spruce (Table 7). The lowest nests which I have seen were about 2 m off the ground in 6-10 m trees, and the highest were about 20 m off the ground in 25 m trees. In conifers, the usual site was where 3-4 branches emerged at the same level to provide a base, but in old pines, the nests were occasionally out on a branch, a metre or more from the trunk. Somewhat more nests were on the south than on the north side of the trees, probably because, with better light, branches were stronger on the south side, providing both firmer support below and better protection above. In broad-leaved trees, most nests were built in a crotch, either beside the main trunk or on a major side limb. Some, however, especially in sallows, were placed where two or more thin branches crossed one another, forming a horizontal platform. Of 1,389 Sparrowhawk nests found in south Scotland during 1971-84,

Nesting habitat

51

93% seemed to have been built from the start on no foundation, other than the supporting branches. I say 'seemed' because we could seldom be sure that no twigs remained from any earlier nest in that site. However, the remaining 7% had clearly been built over a previous structure, including eight (0.6 %) on Squirrel drays, four (0·3 %) on old Crow nests, fifteen (1·1 %) on old Woodpigeon nests, and no less than 72 (5.3 % ) on former Sparrowhawk nests up to five years old. In some cases, the hawks built a completely new nest on top of the previous one, in others they added little new material, while in one case they renovated the same nest in four different years. Elsewhere Sparrowhawks have occasionally built on old nests of Magpies and Jays, but Magpies were absent from our Scottish areas andJays were scarce. Sparrowhawks used nests of other species only if they were in an appropriate position, both within the wood and within the tree. On this count, many nests of Woodpigeons and Squirrels were eligible, but few of Crows, which were normally placed too high in the canopy. After Sparrowhawks finished with their nests, these in turn were sometimes used unaltered by Tawny Owls or Long-eared Owls, or built on by Woodpigeons, Squirrels, Buzzards or Goshawks. Twice we found a Sparrowhawk nest built in exactly the same position as a previous one, which had fallen out. Several times birds used the same tree in different years, but built a new nest each time, higher than the previous. This gave us two trees each with two nests, one tree with three nests, one with four and one with five. Different females were involved each year. Most such trees were either unusually good compared to surrounding trees (a single spruce in a broad-leaved stand), or in good situations for access in otherwise dense thickets. Others were preferred for no obvious reason. Owen (1916) also noted the repeated use of particular trees: 'In one wood of about 60 acres, there is usually only one pair of Sparrowhawks. There are hundreds of trees in that wood that look eminently suitable for them to use. In spite of this, a particular tree has been used by at least five different hen birds, new nests having been built by all of them, for the old nests were always poked down.' Nonetheless, the commonest procedure is still to build a new nest in a different tree each year. This is unusual among larger raptors, and in the small Sparrowhawk can best be explained as an anti-predator tactic. In conifers the branches die from the bottom up, as the tree grows, so a nest which is concealed among green branches when new may be exposed on leafless branches a year or two later. This is why it is generally easier for a human observer to find old nests than new ones. In addition, some predators, such as Pine Martens Martes maries, may remember the positions of nests and check them from year to year (Sonerud 1985). They may even use some as regular sleeping sites. For this reason, too, new nests are more likely to escape attention than old ones, already known. New nests are also clean of parasites, unlike old ones which may harbour fleas and mites from a previous year. The re-use of old nests by Sparrowhawks is probably often an emergency measure, when the birds have insufficient time to build a new

52

Nesting habitat

one. They usually use old nests which have not contained young, and in which parasites are few or absent. The situation differs among large raptors, because they build a much larger nest, requiring greater effort, and can also defend it more effectively against predators. Often they have 2-3 alternative nests, used in turn, which may reduce the risks from parasites.

COMPARISON WITH GOSHAWK

The re-establishment of the Goshawk in parts of Britain enabled me to compare its nesting habitat with that of the Sparrowhawk. I studied two areas which had substantial populations of both species (Fig. 9). In Area A, with large conifer plantations, all the stands were relatively thick, none having trees averaging more than 3·8 m apart. Here the Sparrowhawks bred in stands with trees up to 2·4m apart, and the Goshawks in stands with trees 2,4-3,8 m apart. At least four of the Goshawk sites had previously been used by Sparrowhawks, at the same tree spacing, but before Goshawks had become established. In Area B, conifer and broad-leaved woods were available with trees up to eight metres apart, on average. In this area, Sparrowhawks bred in stands with mean tree-spacings of 1,6-3,0 m, and Goshawks in stands of 2,5-8,0 m. Thus in both areas A and B, Goshawks used the more open of the stands available, but because of the narrower range present in A, had apparently confined Sparrowhawks to a similarly narrower range in A than in B. In south Scotland, where Goshawks were lacking, Sparrowhawks occupied an even wider range of tree spacing than in either A or B (Fig. 8). Shifts of Sparrowhawks into thicker stands, following an influx of their main predators, has also been noted in Denmark (Bomholt 1981). By using different forest types, the two species divide the nesting habitat between them, and reduce the chance of mutual interference. The Goshawk prefers the older, more open stands, presumably because it needs wider flyways. The smaller Sparrowhawk, more vulnerable to predation, prefers more concealment and can make do with narrower fly-ways. Similarly, whereas the smaller species has its nests hidden within the canopy, the larger species often builds in an exposed position below the crown. This too could help with access but, being so large, the Goshawk is also better able to protect its brood physically against predators and withstand more exposure to weather than is the smaller Sparrowhawk. Its need for concealment and shelter is therefore less. In addition, the Goshawk builds a huge solid nest up to a metre across, often supported by thick living branches. This facilitates the continued use of the same structure year after year, and reduces the time spent on nest-building each spring when bad weather may cause delays (Hald-Mortensen 1974). The preference for different habitats may have evolved partly through past interaction between the species, and ifSparrowhawks attempted to breed in old woods, they could easily have fallen prey to Goshawks. Through preda-

Nesting habitat AREA

A

en (])

i ~lLi

II"~

---r-

___

(J)

c

53

0

Sparrowhawk

•

Goshawk

.•.

en Q)

AREA

c

B

o ~ ~~-.....L..+-------r-------,--------r----. . . . . . .- --_---. j

.0

~ :1,L

-..--__"---__ • • • 3

Mean

I

4

..

I~~.-"-i I• •

distance

5

between

6

trees

I ---1--1---.------..11I-..---...... ~

(m)

Fig. 9. Use of woodland in two areas by Sparrotohaioks and Goshawks. The range ofavailabletreespacingsdiffered in the two areas, but in each areathe Goshawksoccupied themore openstands and the Sparrowhawks the thicker ones.

tion, a strong preference for younger and thicker cover could have become fixed in Sparrowhawks by natural selection. Although Goshawks were absent from Britain for more than a century, Sparrowhawks did not in that time expand into the full range of old woods favoured by Goshawks, despite, in many areas, a shortage of younger woods. But since Goshawks have returned, Sparrowhawks have become further restricted in their choice. The implication is that Sparrowhawks have an inherent preference for dense woods, apparent even where Goshawks are lacking, but are further confined by the presence of Goshawks. Apart from the difference in habitat, the two species show no strong nest spacing behaviour, and nests of the one have been found within 200 m of the other. Whatever the reason for the species difference, it is most unlikely that Sparrowhawks choose their nesting places because of the food they offer. The uniform plantations preferred for nesting in our Scottish study areas were much poorer, in variety and numbers of prey, than were other woods in the vicinity which were not used for nesting. Moreover, radio-tracking showed that individual Sparrowhawks obtained much of their food away from nesting places, in other areas where prey were more numerous. Perhaps Sparrowhawks built their nests in places poor in prey in order to avoid areas which were attractive to other predators, including other hunting Sparrowhawks, which were not averse to taking unguarded young. Avoidance of predation and interference is probably the crucial factor in choice of nesting place. This is discussed in further detail in Chapter 11.

8

54

Nesting habitat

C O M P A R I S O N W IT H OT HE R B I RD S

Spa r ro wha wks a nd Goshawks are not unusu al in preferring breeding plac es wi th a particular veg eta tion structure . The sa me is tru e of m an y oth er woodland birds, fro m large ra pto rs an d ga me birds to small warblers , a ll of which breed ch iefly where the trees a nd shru bs ha ve the pa rti cular heights, spacings a nd other characte ristics which the bi rd s like. Each sp ecies seems to respond selec tively to th ose s tr uc tura l features which sig n ify a ppro pria te nestin g habita t, a nd is scarce or a bse nt as a breed er from woo dla nd with differ en t struc ture. Am on g th e raptors, accipiters all over the world seem to select th eir nesting places on the basis of woodl and str uc tur e, with the smaller species occupy ing th e yo ung er a nd den ser stands a nd th e larger species occup ying th e olde r a nd more ope n ones. SU M MA RY

In the th ree m ain facets of nestin g habi tat - the size of wood , the int ernal struc tu re as influe nced by tr ee spacing, a nd th e tr ee species - Sparrowh awk s

Nesting habitat

55

showed marked preferences, but in some localities they used 'sub-standard' sites, if nothing better was available. Thus they preferred large woods to small ones, but where large woods were scarce, they nested readily in small woods, scrub patches or even small clumps of trees. Similarly they preferred young, fairly thick, woods, with trees 2-4 m apart, but accepted m.ore open woods where necessary, including scattered trees in city parks. As nest trees, they preferred conifers, but where these were scarce, they built readily in broad-leaved trees. The type of nesting habitat thus varied to some extent between districts, depending on what was available. Compared with Sparrowhawks, Goshawks usually nested in older, more open stands, perhaps because these larger birds preferred wider fly-ways through the trees. In parts of Britain recently colonised by Goshawks, the Sparrowhawks have become more restricted to the thicker stands, though the actual range of stands used by the two species again differed between districts, depending on what was available. Sparrowhawks usually built a new nest each year, in a concealed position, whereas Goshawks often built on an old nest, in a more exposed position. Avoidance of predation and interference was probably the main factor in the choice of nesting place by Sparrowhawks. All important aspects of choice could be interpreted in this way: the preference for woods over single trees, for dense woods over open ones, for conifers over broad-leaved trees, and also the positions of the nest tree within the wood and of the nest within the tree, together with the building of a new nest in a fresh site each year.

CHAPTER 4

N est spacing and breeding density

Thi s cha p ter is co nce rn ed with th e m ark ed differ en ces in Sp arrowhawk br eeding den sit y whi ch occ ur from on e d istrict to an other. As th e birds nest primarily in woodl and , th eir numbers in a nyo ne di st rict d ep end largely on the a mo un t of suita ble woo d la nd presen t. But even within suc h wood la nd, th eir natural d en sities va ry betwee n di stricts, and th e probl em is to expla in why. Th e d istri but ion of Sparrowha wks ca n be ass ess ed most read ily fro m th eir nest s. With in suitable wood la nd , th e birds breed in th e sa me rest ricted local ities over severa l yea rs. T hey usu all y build a new nest each tim e, close to previou s ones (C ha pter 3). As each nest ca n last for several yea rs, esta blished sites are ea ch marked by a grou p of nests of vary ing ag es , which with pra ctice ca n be eas ily di sti ngu ish ed from nests of o ther spe cies. They a lso ena ble th e obs erver to find the nest in g pla ces at a ny tim e of yea r. The number of nest s in a group d ep ends partly on how often th e place has been used in previ ou s yea rs, a nd partly on the lifesp an of th e individual nest s. In o ur st udy a reas , nests d isin tegr ated slow ly over th e years , but ofte n befor e a nest co uld fall a par t, th e tree itsel f was felled during th e regul a r th inning regim e to which most woods we re su bjected . I n th e a bsence of tree thinning, nest s mad e ofl a rch twigs lasted longest. When I began in Du m friessh ire in 19 71- 2, the n umber ofnest s in particular nest in g p la ces vari ed between one a nd 18, bu t was mos tly in th e ran ge 1-7

Nest spacingand breeding density

57

(Fig. 10). At most places, the nests comprising a group lay within a circle of 50 m radius, but in some areas, where the distances between groups were greater, the nests within groups were also more scattered. Rarely, as shown by subsequent experience, a group was divided between two separate locales, where the birds concerned had alternated between different sites in different years. This was analagous to the alternative nesting sites often held by Peregrines Falco peregrinus and other raptors (Ratcliffe 1980).

Relationshipofnests to pair distribution When occupied, Sparrowhawk nesting places could equally be called 'nesting territories', because they were both exclusive and defended, either being a commonly accepted criterion of a territory. They were exclusive in the sense that each was normally occupied by no more than one pair at a time (except in rare cases of bigamy), and they were defended in the sense that any other Sparrowhawks which were detected near the nest were chased away. I saw several vigorous chases through the trees, in which one hen drove another from the nest area, to the accompaniment of loud alarm calling. The owners did not succeed in keeping away all comers at all times, however, partly because the owners themselves were not always near their nests, and partly because in such thick habitat it was apparently impossible, even for a keen-eyed Sparrowhawk, to detect all other hawks which passed through. Intrusions also came to light during work with radio-transmitters, which we placed on some individuals in order to follow their movements (Chapter 5). Some of these radio-tagged birds had nests of their own, yet during the course of a day's hunting, they often passed through the nesting places of other pairs. This was especially true of hens which, in the pre-laying and

ANN AND ALE

20

(1971)

ESKDALE (1972)

en 15 c

c

10

1

1,--,

10

1

__

. .

Eskdale Oxford

.,

.

-.L... .

_

Lowland Gall-oway

.-.--------..------.

Upland Galloway

~~~~~~~~-------------SI~eY Lower Deeside

~_.

•

----I

_

Mid Oees ide

•

m

.

Clashindarroch Forest

.

.

-

.

• -.

1-

Nearest

Upper Deeside

-

• I

. .,..

2·0

-

K ielder Forest

I

3·0

neighbour distance (km)

Fig. 12. Spacing 0.1 nesting places in continuous icoodland in different areas. The places tcereregularlyspaced in an] one areabut at different distances a/Hut in different areas.

Nest spacingand breeding density

61

3, and, since the initial analysis, information from seven other areas has been incorporated.

Nest spacing and landscape As a measure of spacing for each area, we took the mean nearest neighbour distance for nest groups in continuously suitable woodland, ignoring nests separated by extensive open country. As suspected, such distances were highly correlated with elevation, the nest groups becoming further apart with increasing height above sea-level. They were also correlated with land productivity, with nest groups becoming further apart as land fertility declined (Fig. 13). They were related to a lesser extent to the proportion of farmland (closer together in woods in areas with much farmland), and to the proportion of other open land (further apart in woods in areas with much moorland and sheepwalk). These various environmental features were themselves correlated with one another. As is obvious to anyone familiar with the British countryside, as the ground rises, land productivity and proportion offarmland decline, while proportion of moorland and sheepwalk increase. On the other hand, no correlation was apparent between nest spacing and latitude: fertile lowlands in the north of the country held Sparrowhawks at just as high a density in woodland as did similar areas further south. Statistical procedures enabled us to formulate precise relationships between nest spacing on the one hand and environmental features on the other. Comparing areas, 79% of the variation in the mean nearest neighbour distances within suitable woodland could be explained in terms of elevation alone, and 69% in terms of land productivity alone. A multiple regression analysis, incorporating all habitat measures, did not explain appreciably more of the variation in nest spacing. Within the range of conditions in our areas, the distance between nest groups in continuous woodland increased by 0·1 km for every 20 m rise in elevation, and for every 0·35 point drop in land productivity, measured on a 1-10 scale. By comparison of areas, we thus found precise relationships between Sparrowhawk nest spacing on the one hand, and one or two easily measured landscape features on the other. Four areas in the Dee Valley in northeast Scotland were especially instructive, as nest groups were closer together in woods on the lowest and richest ground near the coast (0,9 km), became further apart in two areas around Banchory (1'1 km) and Ballater (1,4 km) part way up the valley, and still further apart on the highest and poorest ground at the top, in Mar Forest (2'1 km). Likewise, in Ae Forest (Dumfriesshire) nest groups became progressively further apart from the lowest and richest ground on the east and south sides to the highest and poorest ground on the north and west (hence the considerable spread in nearest neighbour distances for this area in Fig. 12). In the Upper Speyside and Mar Forest areas, nests were found up to 500 m (1,650 ft) altitude, but in the last area suitable woodland extended in places up to 610m (2,OOOft). There may therefore have been an altitude limit on nesting which was well below the tree line. Where small nesting woods were scattered amid large stretches of open

62

Nest spacing and breeding density 2·5

2·

1·5

Q)

o c

• ••

1·0

...en

(\1

:c 0·5

•

• fit • • 1 0

•

•

•

•

•

••

•• •

200

Elevation

(rn )

2·5

en Q) "(\1

2·

•

Q)

c

•• •

• •

•

o-

2

4

Land

• • • • • • • • •• 6

productivity

•

• 8

10

index

Fig. 13. Spacing of nesting places in continuous woodland shown in relation to land productivity and altitude above sea level in 19 different areas. Nearest neighbour distances increase (so density decreases) with increase in elevation (linear relationship) and decrease in land productivity (probably curvilinear relationship). Regression equations: Mean nearest neighbour distance (km) = 0·0046 X altitude (m) + 0'280, r = 0,89, P< 0-001. Mean nearest neighbour distance (km) = -0.263 X land productivity index + 2,428, r = -0'79, P

population

index

5

6

(numbers)

(1)

c

2·5

en

(1) '-

••

ctS 2·0 (1)

c

•

• • ••

1-0

• •

0-5

25

Prey

•

•

•

50

population index

•• 75

100

(biomass)

Fig. 14. Spacing of nesting places in continuous nesting habitat in relation to indices of prey bird population in 14 different areas. Nearest neighbour distances increase (so densities decrease) with declines in prey densities and biomass. Relationshipbetween spacing and prey numbers: r = -0'77, P< 0-01: between spacing and prey biomass: r = -0'61, P

.~

- - Prelay 1976

.0

E

40

::I

c:

____e_____

e~'~"""""'O .

e--e~

e--e

···········0

· .O

~

~.~::::---e~ ·····

········0

0

0 .. ·········

e--e--e ·0··

· .. ··0

-; 20 C1>

z

1974

1972

Fig. 23.

1976

1978

1980

1982

Population trendin Eskdale, south Scotland, 1972-84. • - nests, 0 - clutches.

year. In Eskdale, nest numbers remained fairly stable during the 13 year study, 1972-84 (Fig. 23). Numbers fluctuated about a mean of 35, reaching as low as 29 in 1977, and as high as 39 in 1974 and 1981. These fluctuations stayed within about 16% on either side of the mean, but the greatest decline observed between successive years was 18%, while the greatest increase was 24 %. Compared to the fluctuations which were theoretically possible, these figures reinforced the idea of stability. Although not all nests subsequently contained eggs, over the years the number of clutches fluctuated almost in parallel with the number of nests, so either figure could be used as an index of trend. The role of density-dependent factors was apparent from the fluctuations, in that the extent and direction of change each year was correlated with nest numbers in the preceding year (Fig. 24). In general, years of lowest population were followed by the greates t increases, whereas years of highest population were followed by the greatest declines, so that numbers tended to return each year close to the mean level. Now, if density-independent factors had prevailed instead, the direction of change from anyone year should have been independent of population. This was clearly not the case. Although the overall stability of the Eskdale nesting population could be attributed to the action of density-dependent factors, some of the fluctuation about the mean could have been due to density-independent factors, such as weather. In fact, the three years of lowest population in the 13-year period followed the three coldest winters and springs. 10 further explore the role of weather, several analyses were done, but only one gave significant results. From Fig. 24, one could work out the average change in numbers expected from any given population level the previous year. This is depicted by the trend line in the Figure. However, points were scattered widely around this line. In some years the population changed more than expected from its level the previous year, and in other years it changed less than expected. These discrepancies were associated with the weather in March-April. When these months were unusually cold and wet, declines were more marked, or increases less marked, than expected from nest numbers the previous year; conversely, when these months were

1984

Population trends

89

unusually warm and dry, declines were less marked, or increases were more marked, than expected. On a statistical analysis, some 58% of the variation in nest numbers between years could be explained in terms of nest numbers the previous year, whereas 84% could be explained in terms of previous nest numbers plus the numbers of raindays in March-April. This relationship between nesting population and spring weather made sense in terms of Sparrowhawk biology. In March-April, food was at its scarcest, as the numbers of resident prey reached their lowest point of the year, most winter visiting prey had left, while most summer migrants had not yet arrived. So these were months when Sparrowhawks would be expected to experience greatest food-shortage. In March-April more hawks were found dead, and reported through the national ringing scheme, than in any other month (Chapter 23). They included many birds which had starved to death. Moreover, as judged from food deliveries at nests later in the year, Sparrowhawks could not hunt effectively during rain (Chapter 13). It was not surprising, therefore, that nest numbers were related to rainfall in the months when prey were scarces t. On the other hand, no relationship was apparent between population changes and the residues of organochlorine pesticides in blood samples and eggs from different years. Such residues were evidently present in the birds

30

• 20

~ 10~

o

0

~ Q>

O'l C ctl

s: o

'*

-10-

• -20

--

3'4 numbers in

• • previous

Fig. 24. Percent change in Eskdale population (Y) in relation to population level (X) the previous year. The graph reveals the existence of (density-dependent' regulation of numbers, because the extent and direction of change was correlated with the population level in the preceding year. In general, years of highest population were followed by the

greatest declines, and years of lowest population werefollowed by the greatest increases. Regression relationship: Y = 117-21 - 3·36X, r = -0' 76, P < 0·01.

90

Population trends 1·4

•

c (1)

E

1·0

:::J L

U Q)

L

0·6

•

co :::J

c c

0-2

- 80-

-; 70 o

.D

*-

60

• 50 -4-----r-----~---_r_---__w----~---__, 110 12 a 130 140 150 160

Post-breeding

population

in

previous

year

Fig. 27. Annual loss of Sparrowhawks in the pre-breeding sector in relation to the population level the previousyear. The relationship was significantly density-dependent, in that losses weregreaterafteryears ofhigh population than afteryears oflow population (r = 0'87, P< 0'01).

each year and the number of new breeders added to the population in the following year gave a measure of loss over winter from the pre-breeding contingent. In practice, this loss did not refer solely to locally fledged young, but was a net loss, representing the balance between all gains to the population in this period (including immigrants) and all losses (deaths plus emigrants). The extent of overwinter loss each year was inversely related to population density; in other words, it was density-dependent (Fig. 27). Years when numbers were high at the end of breeding were followed by heavy losses over winter and low recruitment the next year, whereas years when numbers were low at that time were followed by light losses over winter and high recruitment the next year. The net change in the pre-breeding sector each year thus contributed the density-dependent component which tended to stabilise breeding density in the long term. One further inference could be made. The fact that the only density-dependent loss identified acted each year on birds which had left the nest implied that, up to a point, breeding failures could be compensated by reduced losses over winter. In other words, Sparrowhawks could withstand substantial reduction in breeding success without it being reflected in the subsequent numbers of breeders. This was presumably why, in the past, Sparrowhawks could suffer heavy losses in the breeding season from gamekeepers or egg collectors, yet still maintain their breeding numbers from year to year. Any losses in the nest would have been compensated by improved survival of the remaining birds after fledging, or by recruitment at an earlier age.

Population trends

95

With the spacing system of Sparrowhawks, it was almost inevitable that the losses from each new cohort were density-dependent. Each year new birds entered landscape in which a large proportion of possible home ranges were already occupied by established breeders. Those new birds which managed to become breeders were mainly dependent on the vacancies created by the deaths and movements of established birds. Thus, in years when new birds were numerous, losses among them must generally have been proportionately greater than in years when new birds were few. In other words, density-dependence in this sector of the population was a consequence of an established sector occupying a relatively fixed number of potential home ranges, itself set by the carrying capacity of the habitat. It was not dependent solely on competition among the new birds, regardless of the rest of the population. To summarise this section, the Eskdale nesting population remained fairly stable during the I3-year study. Density-dependence in the net losses among the pre-breeders, and hence in their recruitment, was sufficient to stabilise the nesting population in the long-term, returning the number at the start of each breeding season to a level near to that of the year before. No densitydependence was detected in reproduction, or in the losses of established breeders from year to year. Competition for home ranges in prime habitat resulted in density-dependent control. This was modified by weather in sprIng.

Furthercomments on habitat quality Sparrowhawks in Eskdale may have favoured certain localities for nesting because these localities offered better prospects for breeding or survival. Evidence was obtained for both. At favoured places, nests produced young relatively more often than at less-favoured places (Table 13). The comparison was not straightforward, however, because the occupants of less-favoured places included a greater proportion of first-year birds than did those of favoured places. First-year birds may have bred less successfully because of their inexperience, rather than because they were in poor habitat. However, when the comparison was restricted to adult birds, nests in favoured places were still the most successful, suggesting that habitat quality was indeed involved. The extent to which breeders remained in particular localities from year to year was checked at nesting places where occupant females were identified in successive years (Table 13). At favoured localities, occupants were more likely to remain from year to year, whereas at less-favoured localities they were more likely to change. Some of this turnover was due to mortality, and some to movement (Chapter 21), but whatever the cause, individuals were clearly more likely to persist in favoured localities than in others. In conclusion, habitat gradation was implied from the frequency of use of nesting places. An association was apparent between use of nesting place, nest success, age of occupant, and tendency to stay. The latter points were established for females only; they probably held for males, but too few were

96 Population trends trapped to be sure. Similar patterns in the use of nesting places have been noted in other raptors, such as the Peregrine, in which Ratcliffe (1980) graded territories as 'regular', 'irregular' or 'occasional', depending on usage.

POPULATION TREND IN ANNANDALE

Stability of Sparrowhawk populations would be expected only in areas where the habitat and other conditions remained stable. It would not be anticipated in areas where the amount of suitable woodland was changing greatly, through felling or planting programmes. Nor would it be expected where numbers were being depleted by organochlorine pesticide use, or where they were recovering from previous depletion. The second population, which I studied in south Scotland, occupied the Annan Valley, about 15-20 km to the west of Eskdale. In Annandale, nest numbers declined markedly during the study period, from 110 in 1971 to 61 in 1979 and 1980. Apart from a slight increase from 1973-74, the decline was fairly steady throughout, and numbers fell by up to 13% in successive years (Fig. 28). As the two areas were close, and had similar weather, I could rule out weather as responsible for the difference in long-term trend. I could also exclude organochlorine contamination: residues of each chemical did not differ significantly between eggs collected in the two areas, and, although residue levels fluctuated from year to year, in neither area did they parallel the changes in nest numbers. The most obvious difference between areas was in land-use trends. In Eskdale, the landscape changed very little during the study period. A few small woods were felled, but their loss was compensated by the growth of other younger stands, so that the total area of woodland remained almost constant. Nor did any great change occur on the farmland, either in the size of fields, or in the crops that were grown. In consequence, the habitat of the Eskdale area remained fairly stable during the study period, and with it, by inference, the food-supply for Sparrowhawks. In Annandale, by contrast, marked changes occurred in the wooded and in the open parts of the area, which reduced the abundance both of nesting and feeding places for Sparrowhawks. During the ten years, about 20% of the suitable woodland was felled, and relatively little matured to take its place. This entailed the net loss of 27 nesting places, and a considerable food supply, because prey birds were about ten times more numerous in woodland than in the clear-felled habitat which replaced it. Planting of new woodland during the study period greatly exceeded the losses, but by the end of the study the new woods were still too young to support either Sparrowhawks or large prey populations. Nonetheless, the decline in Sparrowhawk nest numbers was disproportionately greater than the decline in area ofwoodland; and an apparent surplus of nesting places was available throughout the decline, though mostly in poor habitat. Changes also occurred on the

Population trends 120

100

'- 80

e______ 0 .....

e

' ~. _ _'-

'~

OJ ~;~

'0

~;.=

5 '----------''f------.,...------,--------r------,-------.,..-------,

OJ

c: >III

1972

1974

1976

1978

1980

1982

Fig. 29. Trends in various aspects of breeding performance in Annandale and Eskdale. Most aspects of breeding in the two areas varied almost in parallel from year to year. The two areas were under similar weather (which influenced breeding), but differed in population trend (which did not) .• - Eskdale, 0 - Annandale.

In a Danish area, where nests were counted between 1974 and 1982, their numbers varied between 56 and 89 in different years, with increases up to 31 % of the mean level and declines up to 18 % of the mean. They again showed a fair degree of stability compar.ed to trends in many other birds. Other published censuses from "Europe occurred mainly over the period when populations were either declining from the effects of organochlorine contamination or were recovering from such effects in the past (e.g. Opdam 1975 for Holland, Dyck 1983 for Denmark).

1984

100 Population trends CONCLUDING REMARKS

In general, the idea of long-term stability in breeding population, mentioned at the start, seemed to be borne out by the data available, except in areas where habitat change was marked or where change in pesticide use was marked. The stability was achieved by density-dependent losses in the pre-breeding sector of the population, and hence by density-dependent recruitment of new breeders. In stable environments, Sparrowhawks certainly show more long-term stability in numbers than do some other raptors, such as Kestrels, and the difference is linked with food supply. As Sparrowhawks have a wide range of prey species, they are buffered against the temporary shortage of anyone. In contrast, Kestrels in most areas have restricted diets based on rodents. These small mammals fluctuate greatly in abundance from year to year, often in regular 3-5 year cycles, and promote similar fluctuations in the numbers of their predators (Cave 1968, Village 1982).

SUMMARY

In Eskdale, nest numbers remained fairly stable during 1972-84, fluctuating only between 29 and 39, around a mean of35. This stability was achieved by competition for a relatively fixed number of prime home ranges, which meant that entry to the breeding population was largely dependent on gaps created by the deaths or movements of previous occupants. The proximate mechanism was a density-dependent loss in the pre-breeding sector of the population, and hence a density-dependent recruitment of new breeders. The weather in early spring intervened to prevent a perfect density-dependent system: cold wet conditions in March-April led to greater decline, or less marked increase, than expected from nest numbers the previous year. Annual losses of established breeders was not density-dependent, and nor was production of young. No evidence was found that organochlorine pesticides influenced the year-to-year changes in breeding numbers or breeding success during the period concerned. Whereas in Eskdale the nesting population remained fairly stable during 1972-84, in nearby Annandale the population declined by 45% during 1971-80. Land use remained stable in Eskdale during the period concerned, but in Annandale extensive tree-felling and agricultural changes lowered the amount of nesting and hunting habitat. The breeding populations of the two areas were thus ultimately limited by resources for nesting and feeding. Breeding performance was similar in the two areas and tended to fluctuate in parallel from year to year, according to spring weather.

C HA PT E R 7

Hunting and feeding behaviour

The Sparrowhawk is the only woo d la nd raptor in E urope whic h p rima ril y ea ts small birds . Its fo ra gin g be havio ur is hard to study, becau se most hunting occurs in cover , a nd the individu al prey ca pt u res occ u py no more than a m oment , so a re easily mi ssed . The ord inary obse rve r sees chie fly th ose typ es of hunting which occu r in th e open, glim ps ing th e hawk for only a few sec onds in a lon ger seq ue nce . Wh en we put radio-trans m itters on wild Sparrowhawks, we soo n di scover ed th a t th e m ai n hun tin g techniqu e had no t even been d escribe d, a nd wh at wa s recorded in textbook s form ed only a small part of th e tot al rep ertoire. But eve n with th ese radio-tagged bird s, so me of whi ch ca ug h t a ro u nd ten prey per d ay for th eir you ng, we saw very few ca p tu res, d espite foll owing some individu al s for seve ra l d ays a t a time. Falcon ers with tr ain ed hawks ca n witness m ore a ttacks, but here the situa tion is to som e ex te n t con tr ived (Mavrogordato 1960 , Fox 1981 , K enward 19 78 , 1982). In th is chapter, th erefo re, I ca n give no more than a broad id ea of hunting be havio ur, a nd d escribe pa rti cul ar incid ents wh ich help to com ple te th e pi cture.

102

Hunting and feeding behaviour

Although for much of the year the countryside abounds in small birds, Sparrowhawks often have difficulty in feeding themselves. This is because small birds have various defences, some of which are highly effective. Only a tiny fraction of the birds that a hunting hawk encounters can be attacked with any chance of success, and even then they usually detect the hawk in time to escape. A warning call from just one individual is enough to send all the small birds in the vicinity scuttling for cover. As the hawk moves through woodland, its passage is marked by a succession of alarms, and as it crosses fields, flock after flock rises before it, while still safely out of range. In fact, after detection by the prey, the hawk has no more than about three seconds to grab that prey before it escapes. Little wonder that Sparrowhawks often go hungry, and that shortage of food is a major factor limiting their numbers and breeding success. To catch its prey, a Sparrowhawk has the benefit of a keen eye, great stealth and manoeuvrability. But it is not particularly fast in level flight. The usual speed is 30-40 km per hour, reaching 50 km per hour in short bursts. The hawk is thus somewhat faster than most songbirds, which fly at less than 35 km per hour, but slower than Swallows, waders and larger prey, such as pigeons. For its size, a Sparrowhawk can twist and turn with remarkable ease, but smaller birds can still out-manoeuvre it. To avoid an attacking hawk, small birds normally dive into cover, but larger and faster birds take to the air. Flocking species in open country may bunch together, and perform fast zig-zag flights which hinder the hawk in singling out a victim. Thus to be successful, the hawk has to approach closely without detection, and then dash in and seize its prey as quickly as possible.

HUNTING TECHNIQUES

Short-stay-perch-hunting In the commonest prey-searching technique, the hawk makes short flights from perch to perch, pausing on each to scan the surroundings before moving on. In thick woodland, the bird may fly as little as a few metres between successive perches; in open woodland, where the view is wider, it may fly 50 m or more; and in open country the bird usually proceeds from one tree or patch of cover to the next, whatever distance is required. At each stop, the bird pauses for a few seconds or up to several minutes. I t usually sits well hidden in the top third of a tree or bush, and peers out. On leaving, the bird dips down, flies low over the ground, and at the last moment swings up to perch in another tree. This flight is both energy saving and concealing. As the bird leaves, it uses the momentum of the fall to build up speed along the ground and slows itself by rising abruptly to the next perch. The bird is concealed because for most of the route it is close to the ground. This was the usual searching technique used by all but one of our 57 radio-marked birds. It would have been impossible to record this properly without the help of radios.

Hunting and feeding behaviour

103

Attacks on prey vary according to circumstance. If the hawk gets close to its quarry, it may approach in a directflying attack, travelling at full speed straight towards the target; the hawk makes no attempt to conceal itself, and relies on overtaking the prey before it can escape. Distant prey are approached more stealthily, in a indirect flying attack, in which the hawk takes advantage of any cover or ground contours to hide itself. As it leaves cover, it typically gives a few quick pumps with its wings to clear foliage and build up speed, and then makes as much of the approach as possible on a glide, the still wings further reducing the chance of detection. The victim may be snatched from the ground or a perch, but more often it sees the hawk at the last moment, and is seized in mid air as it tries to escape. It is at this stage that the quick reactions and agility of the hawk are put to good effect, as it twists and turns to grab its fleeing prey. The concentration by the hawk at the time of attack must be, extreme, because every move in the prey seems to be anticipated and countered. Small prey may be taken in one foot, but both feet are needed to subdue large prey. Three outcomes of an attack are possible: either the prey is caught, or it flees and the hawk follows in a tail chase, or the prey escapes and the hawk gives up. In tail chasing, falconers have observed that experienced accipiters often follow in a bird's blind spot, below and just behind, then swing up to seize the prey after a short burst of acceleration ack 1971, Fox 1981). If the first attempt to grasp the prey fails, the hawk may swing up for a second stoop, or even a third, but it seldom follows its prey for more than about 30 m into the air. In this sense, the open sky is as much a refuge for the prey as is cover.

a

High soaring and stooping One radio-marked hen hunted mainly by stooping on prey from a high soar. She would circle up to just beyond the range of human vision, and then stoop almost vertically down, with closed wings, to attack a Starling flock passing below. She did this repeatedly during the week I was following her, but I only once saw her kill. In this case she dived through the flock, then swung up again, turned upside down and grabbed a Starling from below. This enabled the hawk to take advantage of her upward momentum, and avoid the impact of a full-bodied strike. It was another technique which I might not have seen without the use of a radio transmitter, but it has been described by other observers (Rudebeck 1950). Several times I saw different Sparrowhawks circle up to more than 100 m, and then suddenly leave on a long downward glide. Each time the descent was swift, and possibly enabled the hawk to get close to prey in open country, away from perches or cover, as trained Sparrowhawks occasionally attack in this way. Diving approaches are thus of two types: the almost vertical stoops at airborne prey, and the slower and shallower dives to reach prey on or near the ground. In both types the hawk proceeds headlong without flapping its wings, and throws its feet forward at the last moment to seize the prey; no attempt is made at concealment but, because of the height

104

Hunting and feeding behaviour

and speed of attack, some degree of surprise is presumably gained (Fox 1981).

Contour-hugging flight Sparrowhawks fly fast and low along wood edges, riversides and hedgerows, continually slipping from one side of a hedge to the other, ready to take any small bird by surprise (Owen 1932, Pounds 1936). This is the hunting technique which is seen most often, but it was used relatively little by our radio-marked birds. Probably in many such hedge-hopping flights, the hawk has already seen its victim beforehand, perhaps from as much as a kilometre distant. Only this would explain why the hawk manages to emerge from cover just at the spot where the intended victim is sitting. Nonetheless, the hawk seems also to attack small birds which are encountered by chance as it flies from one perch to the next. Similar flights occur at the level of the canopy in woodland, especially where tree-feeding birds, such as Crossbills*, abound. The hawk moves slowly over the trees, and makes a dash at any bird caught unawares. Flights of this type were seen when Sparrowhawks attacked tit flocks near Oxford (Morse 1973). After first appearing above the canopy, the hawk approached the flock from below. The tits were then higher than the hawk, but the moment the alarm was sounded, they all shot downwards to low vegetation. Of five attacks seen, one was successful, and another probably so; two failed because the tits flew down before the hawk could strike, and the fifth because the tit concerned darted behind a tree trunk at the last moment. After each attack the hawk left immediately, and the tits resumed feeding. Still-hunting The hawk sits concealed near pools or other places visited by small birds, and waits for one to approach. In my experience this is not a frequent method of hunting. When our radio-marked hawks were feeding young, and presumably hunting at near maximum capacity, they moved around almost all the time in active search, and spent little time in one place. Still-hunting is often hard to distinguish from resting, especially as a bird may do both together. However, still-hunting does seem to be used as a technique in its own right in open areas with no cover, where the best chance of the hawk getting close to prey is to remain still. On extensive coastal flats, Sparrowhawks sit on the ground for long periods, leaping into flight to attack small birds which pass by. Low quartering The hawk flies slowly, just above stalling speed, 1-3 m above the ground, with head angled down, apparently searching, before suddenly dropping on prey, with legs extended. I have seen Sparrowhawks use this flight along a hedge top where fledglings were perched, and above rough grass where voles occurred, and also when trying to re-locate adult birds which had gone into cover. Indeed, Sparrowhawks probably use this technique mainly when

* Scientific names of prey species are given in Table 17.

Hunting and feeding behaviour

:]05

they are searching for something they know is there, because the flight is often repeated over the same spot, and interspersed with sitting and watching. They are less good than harriers at low quartering, for their wing-loading is greater and they cannot fly so slowly without stalling; nor can they hover in one spot like a Kestrel, at least not for more than a moment or two.

Hunting by sound Sparrowhawks can recognise sounds which indicate food. Both they and other accipiters can be attracted to bird distress calls, or to artificial sounds which resemble them. They sometimes appear at mist nets when the birds which are being extracted begin to call. Fox (1981) described how his captive hawks approached prey which they could evidently hear but not see. Typically, the hawk approached the noise slowly in short flights, appearing alert (supposedly listening) at each stop. It did not necessarily take the most direct route, but moved so as to remain in cover and get as close as possible without being seen, before moving in to attack. At other times, the hawk dived at a patch of cover where it could hear prey, causing the prey to fly, or at least to flinch, revealing its position to the hawk, which then swung in to attack. Hunting onfoot Although most attacks are probably made on the wing, Sparrowhawks also approach prey on foot. In thick spruce woods, they will hop and run from branch to branch apparently searching for nestlings, and in open fields they will run along the ground to approach prey hidden behind low vegetation. Naunton (1973) saw a male travel on foot for 50m up a path in a weedy field, apparently to get close to a finch flock. Some general points The technique used at anyone time depends on the terrain and the prey, but the overriding feature of most of the Sparrowhawk's hunting behaviour is self concealment. This can range from such obvious tactics as keeping in cover until the last moment, if necessary taking a circuitous route, to more subtle tactics, such as flying close to the ground, and keeping a tall plant between itself and the prey, or positioning itself with the sun behind it in its flight towards the quarry (Fox, 1981). Some of these tactics have been seen only with falconers' birds, as there would be little chance of observing wild birds in such detail. Sparrowhawks sometimes persist in chasing their prey into buildings or bushes, and indeed many individuals die from collisions with obstacles (Chapter 22). I have several times seen juveniles fluttering at thick hedges, reaching in and trying to seize some songbird which had taken refuge there. One bird sat for 20 minutes on an isolated bush, where a thrush had taken retreat. I eventually scared the hawk away, but even then the thrush was reluctant to leave. On another occasion, Brian Etheridge caught by hand a juvenile Sparrowhawk which was spread-eagled on a hedge top, holding

106

Hunting and feeding behaviour

at legs-length a screaming Blackbird which it was unable to pull out. The hawk was so preoccupied that it did not notice his approach. Normally, however, if a small bird can reach cover which is impenetrable to a hawk, the bird is safe whether or not it is hidden. Barnard (1979) described how a House Sparrow, chased into a hedge, instantly 'froze' and became silent. It remained thus 'until the hawk struck, whereupon it quickly dodged a short distance (4-5 em) to either side of the line of strike and froze again; this behaviour was repeated during all strike attempts'. Although the hawk and Sparrow were only about 30 cm apart, the Sparrow was protected so long as it stayed in the hedge. According to falconers, it is in the art of self concealment that experienced adult hawks clearly excel over young ones. The latter are much more likely to make a direct flying attack at the prey from too great a distance or in full view, so that the prey escapes. Experience also teaches a hawk when to give up. Young hawks, after a failure to catch prey, may persist with strike after strike in what, to the human onlooker, seems a hopeless situation, whereas adults usually leave immediately to try elsewhere (Barnard 1979). Young hawks flown by falconers also learn how best to attack a particular prey without injuring themselves. They learn, for example, to grip a young Rabbit at the head end, so as to avoid a kick from the hind legs. In captive hawks, Fox (1981) found no difference between adult and young in the percentage of attacks that were successful, but thought that the adults, because of their greater skill, could exploit a wider range of opportunities. This is a plausible idea, but hard to test objectively, because of the problem ofdefining an 'opportunity'. Ifit held in wild birds, however, adult hawks could survive in situations where young could not. Although Sparrowhawks usually forage alone, hunting in pairs has been reported, with one bird driving prey towards another, or both chasing the same item (Bernt 1970, Naunton 1973). Such instances may have resulted from genuine co-operation, or from chance events with two birds in the same place simultaneously. For where prey birds gather in large numbers, they often attract more than one hunting hawk. Thus at the large Starling roost at Leighton Moss (northwest England), up to five Sparrowhawks have been seen in the air at once, and on some occasions seemed to benefit from one another's presence. Attack success

Little information is available on attack success, apart from that obtained by Rudebeck (1950-51) at Falsterbo in Sweden. At this bird observatory, hundreds of Sparrowhawks and other birds pass each autumn on migration. Over several years, Rudebeck saw 190 attacks by Sparrowhawks on potential prey, of which only 12% resulted in capture. Attacks were not always easy to discern, as some consisted of no more than a slight deviation in flight. They may therefore have been more frequent than recorded, and attack success correspondingly lower. The commonest type consisted of a short fast chase after prey taken by surprise; but some long-distance attacks, long

Hunting and feeding behaviour

107

chases, long stoops, and attacks on mobbing birds were also seen (Table 15). Most attacks were on birds which were perched when first encountered; very few were on birds already in flight, and these almost always failed, partly because such birds had freedom of movement in the airspace, and could outmanoeuvre their pursuer.

SPECIALISATION IN DIET

Individual raptors are often said to specialise on particular prey, taking many more than the average for their species in the area. This idea comes partly from falconry experience and partly from the differences in diet which can sometimes be found between neighbouring pairs. Such differences in Sparrowhawks were always linked with local variations in the abundance of particular prey, so that the hawks were not selecting differently from the same spectrum of prey, they were merely taking whichever species were most available locally. However, two types of behaviour may lead to apparent specialisation, at least in the short term. First, individual Sparrowhawks often return to where they made a previous capture, and if only one species is available there, this will result in a temporary run of that species in the diet. Secondly, individuals sometimes develop particular techniques of prey capture, which make certain prey more available than otherwise. The radiotagged female which hunted by stooping at flying prey took many more Starlings than other females breeding in the same wood. Similarly, one individual among the several studied by Tinbergen (1946) took many nestlings, including those of hole-nesters. This bird had probably developed for itself a special technique of nest finding. In general, however, any local, seasonal or annual variations in the diet of Sparrowhawks could usually be attributed to parallel variations in the range of prey available.

KILLING AND EATING

Small prey items are probably killed by the impact of capture, or by being squeezed in the hawk's foot, especially by the two large claws which can each penetrate more than a centimetre. Nonetheless, some birds are not killed in this way, and twice I scared a hawk into releasing its prey, which to my amazement flew away. If the victim persists in struggling, the hawk will continue to 'knead' it, opening and closing its grip on the body, so that the victim is repeatedly squeezed and stabbed. You can feel this for yourself if you take a hungry falconry bird onto your (gloved) hand. Keep your hand still and the bird will perch quietly, but wriggle your hand and the hawk will continually shift position and squeeze with its feet. This is presumably how larger quarry are often killed. Otherwise, Sparrowhawks have no specific action for despatching large prey. They seem merely to attempt to bring the victim to ground, hold it

108

Hunting and feeding behaviour

down, and start eating. The long legs of the hawk enable the prey to be held at long range, an advantage in dealing with thrashing wings or jabbing bill. In a prolonged fight on the ground between a male Sparrowhawk and a Mistle Thrush, the two kept jumping up at one another, the hawk attempting to grasp the thrush in its feet and the thrush pecking with its bill. Eventually the hawk managed to stand on its victim and start eating. Some minutes later, I scared the hawk off, and found that the thrush was still alive with its flight feathers removed, and part of its wings and back eaten. Another time, I disturbed a hen Sparrowhawk from a pigeon, which after a short time flew away, with its back naked and bleeding, and some wing feathers missing. In these cases, the prey would have been killed by the act of feeding, though the early removal offlight feathers may have helped to prevent escape. To the Sparrowhawk, it does not matter whether the victim is killed immediately, only that it stays reasonably still while being eaten. By these actions, Sparrowhawks are able to subdue prey larger than they can carry or eat in one meal. Females often kill Woodpigeons, but they cannot lift them well, only drag them for short distances along the ground. This must limit their ability to conceal carcases, and keep them safe from scavengers. Also, being smaller than many other predators in the countryside, Sparrowhawks have little chance of protecting carcases against marauders in the daytime, and still less against Faxes at night. They may often lose to other animals portions of the large prey items which they kill, or in winter as a result of them freezing solid overnight. Nonetheless, hawks often do return to feed on carcases which escape the attention of scavengers, and Mick Marquiss (1981) saw a hawk kill a pigeon, feed on it, and then stand on it (possibly guarding it) for about 20 minutes, before feeding again. Observations indicate the maximum weights of prey that Sparrowhawks can carry. A female which killed a feral pigeon was unable to carry its prey in level flight and still air; it did manage to lift the pigeon, with difficulty, for three metres over a level surface, and to carry it down a 25 m slope, but was unable to clear a 1·5m wall at the bottom (Weir 1981). The hawk was trapped and found to weigh 340 g, with the remains of a previous meal in its crop, and the pigeon carcass weighed 430 g. Weir watched another female carrying a young Capercaillie Tetrao urogallus, estimated at 350 g, over a distance of one kilometre to the nest. I t seems, therefore, that females can carry, with relative ease, items as heavy as themselves. Such birds as Jays and Lapwings are well below this range. Females cannot eat such prey in one sitting, but they can at least get them to a safe caching site for consumption later. Remains of Woodpigeons and other large prey are often found on hawk nests (Chaper 4). As Woodpigeons normally weigh more than 500 g, Sparrowhawks probably could not carry even severely emaciated ones. So the remains on nests must presumably come from individuals which are killed nearby, and partly eaten or dismembered beforehand. Male Sparrowhawks, which weigh about 150 g, can probably also carry prey as heavy as themselves. Their range certainly includes Mistle Thrush and Fieldfare, which weigh around 120 g. When in flight Sparrowhawks carry prey tucked

Hunting and fteding behaviour

109

close under their belly near the centre of gravity. This is aerodynamically most efficient. A hawk normally plucks its prey on a stump, log or other low mound, or occasionally on a horizontal branch or old nest in a tree. The hawk stands on its victim to hold it down, and pulls out bunches of feathers in its bill. On the ground, the site of action often appears afterwards as a circle of feathers, around where the hawk stood. From small prey, the hawk eats all parts of the carcass, and leaves only feathers, but from large birds it often leaves bill, legs or other remains, and occasionally the gizzard or a piece of gut. From really large prey, a hawk leaves other parts of the skeleton too, including the shoulder girdle with wings and primary feathers attached, and occasionally also the sternum, often notched where pieces of bone have been bitten out. Sometimes whole eggs are discarded from the bodies of victims. From mammals, the hawk leaves fur or pieces of skin with fur attached, and occasionally the guts and tails of small rodents and the feet and legs of young Rabbits and other bone remains from larger mammals. When a Sparrowhawk is feeding a mate or small young, it normally plucks a prey completely before handing it over, but when feeding itself, it plucks and eats at the same time, removing the feathers as it comes to them while eating. This is another time when Sparrowhawks are vulnerable to foodrobbing, either by other Sparrowhawks or by other predators, so this rapid eating probably serves to get as much as possible into the crop in a short time. Both .the tendency to take prey to cover, and the mantling behaviour, in which the hawk crouches over its prey, with wings and tail drooped and fanned, serve to shield the prey from the eyes of potential pirates. Even large young in the nest do this, facing away from their brood-mates. Sparrowhawks always seem to bolt their prey as fast as possible, and the plucking movements are sometimes almost too quick for the eye to follow. Larger raptors, with less to fear from food robbers, eat in a more leisurely way. In the depleted raptor fauna of Britain, it is easy to underestimate the importance of food-robbing in the time between catching and eating, as it is seldom seen, but I have watched Sparrowhawks being robbed by other Sparrowhawks, as well as by Kestrels and Crows. In North America and Africa, where the raptor fauna is more intact, there are records of individual prey passing through the clutches of up to five different raptors before being finally consumed. No doubt food-robbing has been important in the evolution of Sparrowhawks, and may account for much of the eating behaviour we see today. Occasional large meals, separated by long fasts, are typical of many raptors, in striking contrast to most other birds, which take numerous smaller items and spend much of each day eating. To cope with large meals, Sparrowhawks have a capacious crop, which in the male can hold up to 35 g (35 cc) of food, and in the female up to 45 g. With a further 10 g in the gizzard, a male can hold up to one-third of its body weight in fresh food, and the female up to one-fifth. Having eaten, a hawk then has to wait until the food has passed down the gut before it can feed again, and a bird with

110

Hunting and feeding behaviour

a full crop usually does not even attempt to hunt. Gorging is common among predators and allows them to make good use of their larger kills. Sparrowhawks seem to eat most of their prey near where they caught it, but in the breeding season they carry items to the nesting place. Plucking posts there soon become covered with feathers from prey, but considering the number of birds that are killed during the breeding cycle, only a small proportion are plucked near the nest. When providing food for females or small chicks, male Sparrowhawks remove the head from each victim, as well as the feathers, so that the female is presented with a naked, headless food item. I do not know whether the male does this because he likes to ea t the heads himself, or because they are difficult for the female with small chicks to deal with. Either way, considering the number of items killed for small chicks, the males themselves at this stage could live chiefly on heads.

SUMMARY

To catch their prey, Sparrowhawks rely largely on concealment, getting as close as possible before making a short swift attack. The main searching method is short-stay-perch-hunting, in which the bird moves through woodland, from perch to perch, at each pause briefly scanning its surroundings for prey. Other hunting methods include the 'contour hugging' flight in which the hawk flies swiftly along hedges or other lines of vegetation on the chance of surprising a victim, and stooping from a high soar onto prey in the open. Slow quartering (like a harrier) is seldom practised, and still-hunting is probably practised mainly in open habitats. Both sexes can carry items as heavy as themselves; they pluck prey before eating, and from larger victims they leave some bones and other parts. Some aspects of feeding behaviour, including choice of feeding sites, caching, fast-eating and gorging, probably help to reduce the chance of food-robbing.

CHAPTER 8

Food

In so uth Scotl and , Sparrowhawks hunted partly in woo d la nds a nd partly in ope n habitats, in clud ing th e villages a nd fringes of towns. They took all th e small bird sp ecies available locall y, a nd yo ung ind ividual s of most large spec ies. In con tra st to th eir hunting beh aviour, th e di et was easy to study from plu ckin gs and other prey remains. M an y previou s stud ies of die t ha ve be en made in continen ta l Eu rope. All sho wed th e importance of av ia n prey, whi ch ever species of a pp ro pria te size we re av a ila ble locall y. They a lso confir med that th e hen Sparrowhawk took bigger prey than did th e smaller cock. Aver age food cons um p tion in ca p tive birds was found to be a bo ut 40-50 g per da y for coc ks a nd 50- 70 g for hen s, d ep ending largel y on ac tivity (T in bergen 1946, Fri edmann 1967). The number of carcass es need ed to provide thi s ration vari es with th eir size a nd with th e proportion ea te n, but it is eq uiva len t to a bo u t two a nd three sparrows p er d ay resp ecti vely for each sex . Consumption is not the same fro m d a y to day, however , for on some days th e hawk s ca tc h nothing , while o n o the rs th ey stufT the mselves. O ver th e year as a wh ole, to ta l int ak e a moun ts to abou t 16·5 kg of meat for a coc k a nd 22 kg for a hen . Adde d to th e needs of br eeding , a successful Spa rrow hawk pair could account for 55 kg of meat

112

Food

per year. This is equivalent to about 2,200 House Sparrows, or 600 Blackbirds or 110 Woodpigeons.

METHODS AND BIASSES

Pluckings of Sparrowhawk prey are easy to recognise, and provide a ready source of information on diet. The absence of Goshawks from our main study areas left no confusion over which predator was involved. Most pluckings were found near Sparrowhawk nests, but some were found elsewhere during searches of woodland. Other prey-remains were found on the nests themselves, especially in the post-fledging period when the young still fed there, leaving bony remains to accumulate. Whenever possible, full-grown birds, fledglings (young that hadjust left the nest), and nestlings were distinguished among prey remains, according to the growth-stage of the feathers. Any unidentified feathers were later compared with museum skins. Three types of bias seemed likely. If feathers from large or pale birds, such as pigeons, were more easily noticed or longer-lasting than those from small birds, then large birds might have been over-represented in recorded kills. To check for this, pluckings were classed when found as fresh (feathers dry and blowing in the breeze), medium (damp and small feathers stuck to the substrate), or old (damp, stuck down and partly covered by litter). Little difference in the size composition of items was apparent between the three categories, but very 'Small species seemed under-accounted among old pluckings (Newton & Marquiss 1982). To assess diet, therefore, only items noted as fresh or medium were included. These would usually have been less than a week old. A second bias would have arisen if Sparrowhawks took to their nests an atypical cross-section of the prey they killed. In particular, the size of items may have influenced whether they were taken to nests. I therefore compared for size-composition those items found near nests with other items found away from nests in the same months (Table 16). The former were probably taken entirely by the breeding birds concerned, whereas items away from nests would have been taken partly by breeders and partly by non-breeders. In general, relatively fewer large items were found near nests than away from nests, and this difference increased as the season progressed (Table 16). To a degree, this was expected, because breeding females did not hunt during the incubation and early chick stages, whereas failed and non-breeders did. And as females took larger prey than males, this alone could account both for the difference in prey sizes between sources, and for the widening difference through the season, as failed females increased in the population.

I could not resolve whether breeders selectively brought prey of certain sizes to the nest, but had to accept that prey recorded at nests was not wholly representative (in size) of that killed by the hawk population at large. A third bias applied to skeletal remains found on nests in the post-fledging period. When these mainly bone items were compared with plucked feathers

Food 113 nearby, larger species predominated among bones. This was probably because the bodies of most small prey were eaten completely after plucking, whereas bones were left from large prey. Tinbergen (1946) found a similar bias, but noted that large prey were present among pluckings in the same proportion as they were seen to be fed to the young during nest watches, so pluckings gave the most reliable picture. Because of the bias in bone remains, I used them only for limited purposes. Another method of study used elsewhere entailed cageing the young on the nest (Sulkava 1964). The parents left food items on the cage, which could be checked periodically and the prey fed to the young. Providing the young were beyond the age of brooding and could feed themselves, they could be kept caged long after they would normally have left the nest, to provide information on food and feeding frequencies in the 'post-fledging' period. Analysis of gut contents has also been used to study diet, revealing differences between the sexes, which were hard to assess in other ways. On the other hand, regurgitated pellets were of limited value, because they contained little or no bone material (which Sparrowhawks digested), and consisted mainly of greyish feather fragments which were hard to identify as to species. Nor was radio-tracking much usc, because before we could find the hawk, it had usually finished its meal and moved on. In summary, therefore, fresh to fairly fresh pluckings gave the most abundant and reliable data.

DIET IN THE BREEDING SEASON

During April-August, birds comprised 97% of the ten thousand prey items found, the remaining 3% being mammals. The birds were mainly species that were resident in south Scotland all year (91 % of total prey), with a few winter visitors which had not yet left (1 %), and some summer visitors (5 % ) . In all, 72 species were recorded in these months. About 57% of all birds were taken as full-grown, 41 % as fledglings, and at least 2% as nestlings. The mammals were mainly voles, with a few young Rabbits and various others, making nine species in all. Six bird species were pre-eminent in the summer diet, namely Chaffinch, Song Thrush, Blackbird, Robin, Starling and Meadow Pipit. Each of these formed more than 5% of all items recorded, and together they comprised about 56 % • Individual prey ranged in weight from less than 5 g to more than 500 g, so a somewhat different picture emerged when the diet was expressed by weight (Fig. 30, Table 17). On this basis, the Woodpigeon formed 21 % of the food taken in these months; the Blackbird, Song Thrush, Starling and Chaffinch also formed more than 5% each of the diet by weight and the five species together formed 60 % • However, large items (such as full-grown Woodpigeons) were probably over-represented by weight, because of wastage and loss to scavengers (Chapter 7). Among the mammals recorded, young Rabbits were taken mainly in early spring, when bird prey were scarcest, and also in summer when the hawks

114

Food

Starling

Chaffinch

o

Meadow Pipit

•

%

by numbers

%

by weight

Song Thrush

Blackbird

Robin

Woodpigeon

5

15

10

% in

20

25

diet

Fig. 30. Main prey species by number and weight during the breeding season (AprilAugust) in south Scotland. Only species whichformed at least 50/0 ofthe diet by number or weight in this period are listed. Total number of items = 9390. From data in Table 17.

were feeding young. Among rodents, Bank Voles were the most frequently taken, perhaps because they occurred in woods and were active by day. They were eaten throughout the year but chiefly in early spring. Field Voles were taken from sheepwalk, clear-felled forest areas and young plantations, mainly in years with high numbers. Diet differed between habitats in line with availability (Fig. 31, Table 17). Song Thrush, Blackbird, House Sparrow, Mistle Thrush, Blue Tit and

Great Tit were most important in farmland woods; Skylark, Wheatear, Meadow Pipit, Snipe and Lapwing on sheepwalk; and Chaffinch, Robin, Jay, Siskin, Wren, Goldcrest, Coal Tit and Field Voles in large conifer plantations. Nonetheless, the hawks nesting in one habitat often foraged in other habitats nearby, and the six main species were important throughout.

Food 115 ----LJay

Blackbird

Starling

Wheatear

Siskin

Robin

Linnet

Goldcrest

Redpoll

Willow Warbler

Bullfinch

o unnock

Chaffinch

Blue Tit

House Sparrow

Great Tit

Skylark

Coal Tit

MeadOW Pipit

Wood-pigeon

Wren

Woodcock

Mistle Thrush

Snipe

Song

Lapw ing

. . Farmland woods

Thrush I 10

I 15

I

20

%in

CJ 10

Sheepwalk woods

15

diet

Fig. 31. Main prey in diffirent habitats in south Scotland, April-August. Only those species which formed at least 1 % of the diet by number in any of the three habitats are listed. From data in Table 17.

No great differences in summer diet were found between the years 1971-80, so records for all years were pooled (Table 17). However, summer visitors (mainly warblers) seemed especially numerous in the environment and in the diet in 1980, and Goldcrests were extremely scarce in environment and diet in 1979, after a hard winter. In addition, young Siskins and Crossbills were especially prevalent in the food of some hawk pairs in conifer plantations in spring 1977, when good cone crops favoured these species. Some fast or agile prey were taken, including Swallows, Sand and House

20

116

Food

Martins and one Swift. Also, six male Sparrowhawks were recorded in the breeding season, including two adults at the pre-egg stage, and two yearlings and two fledglings in the late nestling stage. At both these stages, females were hunting, and may have been responsible, but the deaths in spring may have resulted from fights between males, with the victor eating the loser. These records of cannibalism excluded the fairly frequent cases where birds ate one or more of their own young after these young had died. One adult and two fledgling Kestrels were also found among the prey. Otherwise, prey items included little that was unexpected, and virtually all the common birds in the area were taken at some time.

Sex difference in diet Among prey remains, items killed by the male were usually indistinguishable from those killed by the larger female; however, a difference in the prey sizes taken by the two sexes could be inferred from comparison at individual nests of items taken during the incubation period (when only the cock hunted) with items taken during the late nestling/post-fledging periods (when both sexes hunted). The cocks evidently took prey up to 120g in weight (the larger thrush species) and seldom larger, whereas hens regularly took prey up to 500 g (Woodpigeons) and occasionally larger (Fig. 32). Overall, the smallest prey-birds (such as Wren and Goldcrest) were taken somewhat more frequently by cock than by hen; birds of sparrow and lark sizes were taken in similar proportion by both sexes; those of thrush and small dove sizes were taken by both sexes, but more by hen than cock; while larger birds, such as full-grown Woodpigeons , were taken almost entirely by the hen. These general findings were in line with those from three nests at which prey deliveries by cock and hen were witnessed (Chapter 14), and with the earlier findings ofTinbergen (1946). During incubation, almost 2% of recorded prey items exceeded the weight of a cock Sparrowhawk. Such items included young Rabbits (200 g), young Lapwings (up to 200g), young and adult Woodcock (up to 280g), one Partridge (380 g) and no less than five full-grown Woodpigeons (400-500 g). As these items were fresh, they were unlikely to have been taken as carrion. Cocks may well have killed them, at up to three times their own weight, because in other parts of the breeding cycle prey of similar relative weight were recorded for (presumed) hens, including female Pheasant (normally 600-800 g) on three occasions, female Red Grouse (550 g) once, and young Black Grouse (less than 500 g) once. I t was also possible that some hens left the nest to hunt during incubation, accounting for the large items found then, but we never recorded this in our nest watches, and nor apparently did other observers (Owen 1916, Holstein 1950, Tinbergen 1946). Apart from Woodpigeons, however, such large kills were clearly unusual for either sex and in number formed less than 1% of all items recorded. At nests watched at Oxford and various places in Holland, the cock brought chiefly small birds to the nest, and only seldom brought items as large as young Song Thrushes and Blackbirds, which were often brought by cocks

Food 117 40

30

:u Q)

c:

i

20

;!. 10

I

2

«20g)

(21- 50g)

:5 (51-1209)

4

5

(121·· 350g)

(> 350g)

Weight class

Fig. 32. Sizes of prey items taken during incubation, when only the male hunted (open columns), and during the late nestling/ post-fledging stages, when both sexes hunted (shaded columns). The difference between the two sets of data was statistically highfy significant (J:¥/= 180'9, P < 0'001), and indicated that in general males took smaller prey species thanftmales.

in south Scotland (Geer 1979, Opdam 1975). The hawks themselves did not differ in size between these areas, but the Oxford and Dutch areas were richer in prey than the Scottish ones, so perhaps the hawks there could exercise more choice, taking chiefly those species they could cope with easily. Increase in size of prey meant increase in risk to the hawk, and as cock Sparrowhawks had difficulty in subduing Mistle Thrushes (Chapter 7), they might not have tackled such 'high risk' prey when easier species were common. Perhaps for the same reason, female hawks in south Scotland took more Wood pigeons in hill woods, which were poor in prey, than in valley woods. The pigeons themselves were commoner in valley woods.

Predation offledglings Each spring fledglings appeared on plucking posts several days before I began to see them in the environment, and in late summer fledglings ofseveral species continued to appear among pluckings long after the main breeding seasons of the species concerned. Evidently, Sparrowhawks found fledglings with particular ease, even when these were extremely sparse in the prey population as a whole. Having just left the nest, fledglings were of course highly vulnerable, as they could hardly fly and were very noisy. Those of one species or another were available to Sparrowhawks from April to September each year, and inJune-July formed more than half the total prey. Each prey species experienced heavy predation around the time its young left the nest. Then the hawks increased the proportion of that species in their diet, mainly with fledglings, and took fewer again once the young prey

118

Food

5 4

(0 )

3 2

(b)

:.0

Cii

4

.s

3

VI

E

~

2

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

;!.

(c)

4

3 2

................. ~ May

Ju~e

. July

August

Fig. 33. Proportions of three tit species in the diet, April-August. (--): all individuals; (- - -): fledglings only. (a) Coal Tit, (b) Blue Tit, (c) Great Tit.

could fly well (Fig. 33 for three tit species). Through the season, the hawks continually switched emphasis from one species to another, as each produced its young (Fig. 34). In general, various thrushes predominated near the start of the season, Chaffinches and tits in mid season, and hirundines, wagtails and Starlings near the end. Those prey species which raised more than one brood each year suffered heavy predation over a long period, whereas singlebrooded species experienced heavy predation over a short period. Since many of the prey birds classed as full grown (on wing and tail feathers) would also have been juvenile, the concentration of predation on the young individuals of each species was even more marked than Figs. 33 and 34 suggest. This was evident in the Starlings, in which full-grown juveniles could be distinguished from adults by their brown flight feathers; once such juveniles became available, extremely few adults were taken. Few fledgling Starlings were recorded, too, probably because, when they emerged from nest-holes, their flight feathers were almost full grown. Predation on Starlings

became heavy from late] une, when the young began to flock in fields. In contrast to fledglings, nestling birds apparently formed only a tiny part of the Sparrowhawk diet (2% by number). More than half of those found were W oodpigeons, which built fairly open, easily-seen nests. Others included Blackbird, Song Thrush, Chaffinch and many unidentified. Perhaps Sparrow-

Food

119

~O)

;

2 I

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

4

.'

j

..

(b)

3

/\----

2 I 4

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

.. ... :

....

(e)

3 ~

~

2

.S

I

E ~ 0~

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

~

IJ')

5

----'_ _

--w~

.r....-___

.

(d)

4

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

3 2

.

I

.

.

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

5 (e) 4 3

2 I

...................... April

May

........ .

June

.. July

August

Fig. 34. Proportions of various prey species in the diet, April-August. Each species increased in the diet temporarily, whilejledglings were available. (--) all individuals; (- - -) jledgings only. Mistle Thrush (a) reached a peak in the diet during May, followed in order by Song Thrush (b) in late May, Robin (c) in early June, Chaffinch (d) in lateJune andJuly and Starling (e) in lateJuly andAugust.

hawks were not good at finding nests, but records may have been biased against nestlings, if some had been eaten completely, leaving no remains.

DIET OUTSIDE THE BREEDING SEASON

During September-March each year, we noted any pluckings found incidentally, but about half the records came from the winter of 1975-76, when all Sparrowhawk nesting places in the study areas were checked. In all, 36 species of birds and two of mammals were recorded in these months (Table 17). Only six species formed more than 5%

of the diet then, in order

of numbers: Redwing, Blackbird, Fieldfare, Chaffinch, Woodpigeon and Goldcrest; together these comprised 62% of all items recorded. The Starling

120

Food

Starling

D%

Chaffinch

•

Fieldfare

%

by numbers

by weight

Redwing

Blackbird

Goldcrest

Woodpigeon

I

10

I--------r----------- T--15

% in

20

25

I

30

I

35

diet

Fig. 35.

Main prey of Sparroiohaioks in south Scotland, September-March. Only those species which formed around 5% or more of the diet by number or weight in this period are listed. Total number ofitems = 412. From data in Table 17.

provided almost 5% of items (Fig. 35). Sixteen species formed more than 1% of the diet, and together these comprised "87% of all items. Expressed by weight, the Woodpigeon again emerged as the most important prey, providing no less than 34% of the total recorded (though again not all such large items would have been eaten completely). Three thrush species, Blackbird, Redwing and Fieldfare, provided 17% , 1 4 % and 10 0/ooffood by weight, and together the four species provided 76 %

•

Little unexpected occurred

among the food in these months. Three Sparrowhawk prey were juvenile males, found plucked in September, a few weeks after leaving their natal territories; again we guessed that they had been killed by larger females. Rabbits were not recorded in winter, perhaps because Sparrowhawks could kill only smalljuveniles which were not available then.

Food 121 PREDATION ON GAME BIRDS

The main reputation of the Sparrowhawk as a predator on gamebirds came from years ago, when Pheasant chicks were hatched under broody hens and raised in coops in open fields. Under these circumstances, the small chicks were easy prey. But with the change in rearing methods, whereby the poults are not released until they are half grown, predation is much less, because they are then well above the preferred size range of Sparrowhawks. Throughout the study areas, many thousands of Pheasants were raised and released each July when about half-grown. In Eskdale alone, some 10,000-20,000 were put out in different years, augmenting the 'wild' stock which bred naturally. Yet in ten years, only two downy chicks were recorded among Sparrowhawk prey items, 50 partly-grown young and three adult females; together these Pheasants comprised 0'6% of all items found during April-August. None was recorded in September-March. As most summer records were from nests, they may have under-estimated the proportion of poults in the diet. Put out at six weeks of age, such birds were vulnerable to Sparrowhawks for only about three weeks and, as they were heavy, most of those killed were probably eaten on the spot. On the few occasions when hawks were seen at pens, they took only one poult, and not the large numbers killed by some other predators at pens. Game-keepers in the study areas did not regard Sparrowhawks as major predators of poults at pens, but thought that hawks sometimes scared poults and caused some to injure themselves against the sides. With such large numbers of Pheasants available in the study areas, the proportion removed each year by the Sparrowhawk population could only have been negligible, probably less than 1% of those raised. This was much less than the numbers taken by mammalian predators, or killed by farm machinery and road traffic. Partridges were relatively uncommon in our areas, and in the ten years only two young and six adults were found among prey, formed 0'08% of all items during April-August. Red Grouse and Black Grouse were more localised, and only two individuals of each of these species were recorded as prey. Once they were more than half-grown, the last two species were really too large for Sparrowhawks to deal with easily.

CONCLUDING REMARKS

Despite the wide range of prey taken, relatively few species emerged as important. The ten most frequent species in April-August provided 68% of all items and 67% of all food by weight, and the equivalent ten in September-March provided 78% of all items and 68% offood by weight. The Black-

bird, Woodpigeon and Chaffinch were important at all seasons, with the

122

Food

addition of Song Thrush, Robin, Meadow Pipit and Starling especially in the breeding season, and Redwing and Fieldfare in winter. The Woodpigeon was so much heavier than these other species, that it could have provided up to one-fourth of the total meat consumed during the breeding season, and up to one-third during winter, if none had been lost to scavengers. Fullgrown Woodpigeons were taken almost entirely by hen Sparrowhawks, but the nestlings were taken readily by both cocks and hens. No bird species of suitable size occurred commonly in the study areas without being taken, and the lack in our food list from the breeding season of species which featured prominently in lists from other regions was attributable to the extreme scarcity or absence of such species in south Scotland. In winter, on the other hand, with only 412 items recorded, some fairly regular prey could have been missed by chance. Almost certainly, the majority of prey were taken live, and the only evidence that Sparrowhawks would take carrion (apart from their own previous kills) was the fact that they were occasionally poisoned when they took meat baits put out by gamekeepers. Seven such victims that came to our notice were all females. I wondered whether I was recording all the types offood eaten by Sparrowhawks from prey remains alone, and whether other types of animals were eaten, which left no remains. However, examination of the guts of dead Sparrowhawks revealed only birds.

Comparison with previousstudies In Europe as a whole, more than 120 bird-species have been recorded as taken by Sparrowhawks, with big regional variations depending on availability (Table 18). However, certain widespread species were important in most of the regions involved. Of nine studies in the breeding season, the House Sparrow was important (>5% ofdiet) in eight, the Chaffinch in seven, the Song Thrush in five and the Yellow-hammer in five. Of three winter studies, the House Sparrow was important in all three, forming up to 300/0 of all items recorded. In general this species was much more prevalent among prey remains from Holland and Germany than among those from Britain or northern Europe. Game birds (mainly downy chicks) formed no more than 2% of the summer diet in any area, and less than 1% in most. Mammals formed less than 3% of items in most studies, except during vole plagues in northern Europe, when they reached up to 15% of the winter diet (Sulkava 1964). At least 17 species were recorded in all, including various small rodents and shrews, young Rabbits and young Hares, Moles, Weasels, rats, squirrels and various bats. The only other items reported were occasional lizards, killed but not always eaten. Furthercomments on the sex difference Cock and hen Sparrowhawks may have different diets partly because they select different places to feed, but mainly because they select different species from the range available in anyone place. Thus cocks may take more tits than hens do, partly because they feed more in woodland, where tits are

Food

123

plentiful, and also because within woodland they are more prone than hens to attack any tits encountered. Such selection was studied experimentally in the North American Sharp-shinned Hawk, a close relative of the Sparrowhawk. To catch migrating hawks, Mueller & Berger (1970) used live pigeons, starlings and sparrows as decoys. The incoming hawks could see all the decoys, so could choose what to attack. By noting the sex and age of all hawks caught at each decoy, preferences emerged. Both sexes were attracted to all three species, but males showed a stronger liking for small sparrows than did females; hawks of neither sex actually struck a pigeon, but both landed beside it. In addition, juveniles more often attacked inappropriately large prey than did adults. Hence, given the same choice of prey, the sexes and age groups selected differently; and young birds probably learned by experience not to tackle prey that were too large for them. Both Sparrowhawk sexes gain from reducing hunting time to a minimum, not just to save energy, but also because it is while hunting that accidents occur (Chapter 22). So one might expect that, in conditions of abundant prey, both sexes would take the largest birds they could manage, and increasingly switch to smaller birds as larger prey became scarce. The largest prey are not the optimal, however, partly because they can fight back and put the hawks themselves at risk. Also, as large prey cannot be consumed in one meal, they are more likely than small items to be found by scavengers, so the amount of food they yield does not increase in direct proportion to their weight. For these and other reasons, the foraging models developed in recent years for other birds cannot be applied unaltered to Sparrowhawks and other raptors.

SUMMARY

Birds formed almost the en tire diet (the rest mammals), the smaller male hawk concentrating on smaller prey species (5-120 g, especially 5-80 g) than the larger female (chiefly 20-120 g, also up to 500 g or more). All bird species of appropriate size which occurred in the study areas were at some time taken. In April-August, the most important prey species in terms of numbers were Chaffinch, Song Thrush, Blackbird, Robin, Starling and Meadow Pipit; and in terms of weight they were Woodpigeon, Blackbird, Song Thrush, Starling and Chaffinch. Throughout the spring and summer, fledgling birds were taken in large numbers. Each prey species increased in the hawk diet for a short period after its young had left the nest; and as different prey species produced their young at different dates, the hawks continually switched emphasis from one prey species to another. In September-March, the most important prey species in terms of numbers were Redwing, Blackbird, Fieldfare, Chaffinch, Woodpigeon and Goldcrest, and in terms of weight

they were Woodpigeon, Blackbird, Fieldfare and Redwing. The diet varied with area, year and season, depending on the bird species available.

CHAPTER 9

The Sparrowhawk as a predator

In thi s cha p ter , I shall co ns ide r th e ex te n t of Sparrowha wk pred ation on differ ent prey s pec ies , th e rol e of thi s predation in the overall mo rt a lity of th e prey, a nd its effect on th eir population s. Over the yea rs , d eb at e ha s co n tin ue d o n whethe r, in th e lon g ter m, pred at ors ca n reduce the numbers of th eir prey, o r whe the r th ey m erely p ro vid e one form of morta lity whi ch , if re moved, would be repl aced by others, suc h as di sea se or sta rva tion . If thi s last view held , th e predator would have no lasting effect on prey numbers . It is not ea sy to assess th e impa ct of a ny wild predator on its prey, because this involves so m an y different measurements, some of wh ich a re diffi cult to m ak e. However, on e of th e ea rl ies t a ttem p ts at thi s ge ne ra l probl em conce rned Sp arrowh awks and so ng birds. With th e help of stude n ts, Tinber gen (1946) a tte m p ted through sa m ple co un ts to ass ess th e wh ole so ng bird population in an a rea of wood s, farms a nd villages in H oll and . H e a dj usted his ce ns us techniques to suit th e prey sp ecies involved , a nd rep eat ed the work ove r th e three su ccessive su m me rs, 1941-43. From th e ratios of adult to young birds at different d ates, he al so esti m a ted th e breeding production of eac h prey sp ecies , and, from rin gin g results and o the r inform a tion, he calcula ted th eir variou s mort alities . In ad d ition , he found a ll th e Sp arrowhawk nest s in th e a re a, a nd recorded th e number of yo u ng p rodu ced at each. From m ea surements of th e food co ns u m p tion of adult ha wks, a nd th eir brood s a t different ages, he was a ble to ca lcula te th e tot al meat cons u m p tion

The Sparrowhawk as a predator 125 of the hawks during each summer. Then, from prey remains and prey weights, he could work out the contribution to the total hawk food that each prey species provided, and from this the number of individuals of each that were taken. These numbers could then be viewed in relation to the total mortality of each species. Several interesting results emerged.

Predation on different prey species Even within the preferred size range, different prey species were not taken in relation to their abundance in the environment. Certain species were preyed upon more than expected from their numbers, and others much less so. In general, the most vulnerable species were either conspicuous or easily caught, both depending partly on their behaviour with respect to cover. Thus species which ventured into the open were caught relatively more often than those which remained in thick foliage. The vulnerability of a species to Sparrowhawk predation could be quantified by comparing its proportion in the total prey population with its proportion in the hawk diet. If a species formed 5% of the local population, and 5% of the hawk diet, that species was taken in exact proportion to its abundance; if, on the other hand, that species formed 10% of the hawk diet, it was taken twice as often as expected from its abundance, and if it formed 2'5% of the hawk diet, it was taken half as often. The figures from different prey species reflected the relative vulnerability of each. Only species living in the same habitat could be, compared, because the hawks hunted more in certain habitats than in others. Among species which were centred on villages and farmyards, the House Sparrow Passer domesticus and 'free Sparrow P. montanus were taken most frequently relative to local abundance (Fig. 36). These species were at risk, probably because they often left cover to feed in the open, and formed into noisy groups, which were easily detected by hawks. When close to cover, sparrows were good at escaping from hawks, but not good enough to offset the effects of their other behaviour. In contrast to sparrows, Swallows Hirundo rusticola were captured far less than expected from their numbers. This was presumably because of their fast and agile flight, which more than compensated for their conspicuousness. Spotted Flycatchers Muscicapa striata were also seldom taken, a finding which Tinbergen attributed to their motionless surveying attitude when foraging, and their general alertness. Icterine warblers Hippolais icterina were taken even less often, apparently because of their liking for dense cover. Among woodland birds, Tree Pipits Anthus trivialis were taken much more than expected from their abundance, and were the most vulnerable of the various species studied. This was linked in the first place with their more exposed lifestyle, particularly when singing, but in addition, they were numerous in places where hawks liked to hunt. Willow Warblers Phylloscopus trochilus were also taken more than expected. This was surprising for a scrub-dwelling

species, but may have been due to the males singing for most of the day in the open parts of bushes, where they could be seized. In contrast, tits

126 The Sparrowhawk as a predator WOODS

VILLAGES

VI LLAGES & WOODS

2·2 Tree Pipit

2·

1·

)(

Q)

'0

1·4

c

Willow Warbler Tree Sparrow

:-1·2

Great Tit Turtle Dove, Jay Redstart

House Sparrow

::J

>

0·8

Coal Tit

Blue Tit, Chaffinch Blackbird

06 Goldcrest Crested Tit 0·4

Swallow Spotted Flycatcher

0·2 i~terine

o

Warbler Wren

Fig. 36. Vulnerability of different prey species to predation by Sparroiohauiks. An index of 1 indicates that the species concerned is taken in exact proportion to its abundance in the environment, more than 1 that it is taken more than expected from its numbers, and less than 1 that it is taken less than expected. Taken from a study in Holland by Tinbergen 1946.

were taken less than expected, but Coal Tits Parus ater somewhat more than Crested Tits P. cristatus. Coal Tits left cover more often than Crested Tits, and had more contrasting plumage patterns, both of which may have put them at greater risk. Goldcrests Regulus regulus showed a similar low vulnerability. Turning now to the species found both in woods and villages, the Great Tit Parus major appeared about twice as prone to capture as the Chaffinch Fringilla coelebs. This difference was hard to explain, for both were about equally conspicuous, though the Chaffinch was more agile on the wing. The difference between the Great and Blue Tit P. caeruleus was as expected, for the less predated Blue Tit spent more time in foliage, was smaller and had less contrasting colours. The high vulnerability of the Redstart Phoenicurus

The Sparrowhawk as a predator 127

phoenicurus could be explained by its habit of choosing exposed places when singing and food seeking. The percentage of Blackbirds Turdus merula taken seemed rather low, possibly because in the area concerned this species stayed in cover during summer. Two larger species, the Jay Garrulus glandarius and Turtle Dove Streptopelia turtur, were taken much more than expected, presumably because of their bright colours and conspicuous behaviour, and (in theJay) slow flight. Finally, the Wren Troglodytes troglodytes, an inconspicuous denizen of thick cover, was almost untouched. While the figures obtained reflected the relative vulnerabilities of these different species, the explanations offered were little more than reasoned guesses. The species differed in many characteristics which could not be quantified, some of which may have influenced vulnerability more than was at first apparent. Nonetheless, as already stated, the most important elements influencing vulnerability seemed to be conspicuousness and relationship to cover. The initial detection of the prey by the hawk also seemed to be important, as prey which were not easily noticed were seldom captured. Only in a few species - such as Swallows - was conspicuousness offset by some other factor, such as superior flying skill. Inconspicuous species were generally overlooked. Thus, although the Wren was as abundant as the Robin, only two Wrens were taken, compared with 76 Robins. Interestingly, body size was not the most important factor affecting vulnerability, perhaps because, although larger prey offered larger meals, they also posed greater risks, as discussed in Chapter 7. In anyone situation, the degree of hunger in the hawk, and its general experience and ability, may also have influenced both its tendency to attack and its persistence. Probably no species was preyed upon uniformly throughout the study area, because the hawks concentrated their hunting within a limited distance of their nests. This was apparent in the proportion of House Sparrows taken by different hawk pairs. Such sparrows were present in numbers only around human settlements, and their frequency among prey declined rapidly with increasing distance between hawk nest and settlement. From Fig. 37, one might surmise that the hawks seldom hunted beyond 2·5 km from the nest, and that prey living further away were relatively immune from attack. In other words, wherever hawk nests were more than 5 km apart in this area, predation on all prey species would be patchy. The places where the hawks hunted most depended partly on overall prey availability. In the Dutch areas, prey were more numerous in the villages than the woods, and for much of the summer the hawks hunted much more in the villages. In our Scottish areas, in contrast, prey were scarce in the villages, and the hawks usually hunted in woods. This is another reason why the vulnerability of anyone prey species will vary between areas, and why caution is needed in extrapolating from one study to another. An abundance of sparrows in a village will put all small birds in that village at risk, whereas a scarcity of sparrows may effectively protect the others, which collectively may not reach numbers sufficient to attract much hunting from Sparrowhawks. Hence, the vulnerability of anyone species depends not only

128

The Sparrowhawk as a predator

60

•• .~ 40 CJ)

~

e ctl

Cl. (f) Q)

~ 2

o

I

•

••

• • • ••

•

•

• •

•

•• •

•

•

• ••

•

1 2 3 Distance (km) between Sparrowhawk nest and nearest human settlement

Fig. 37. Proportion of House Sparrows in the diet of different Sparrowhawk pairs) according to the distance ofthe hawk nest from a human settlement. Redrawn from Tinbergen 1946.

on its own numbers, but on the numbers of other prey in the same place. In theory, certain uncommon species could be eliminated from some localities simply because other prey, which are themselves less vulnerable to Sparrowhawk predation, are nevertheless sufficiently abundant to attract the hawks in numbers.

Predationcompared to othermortality Turning now to the contribution that Sparrowhawks make to the total prey mortality in different months, it will help first to give some orders of magnitude. Over the season as a whole, Sparrowhawks in the Dutch area were estimated to account for half the total deaths in House Sparrows, onequarter in Chaffinches and Great Tits, and only a very small percentage in Coal Tits. Less accurate estimates were made for other prey species, but overall it seemed that Sparrowhawks caused more than 25% of the summer mortality of nine of the 17 species considered. For each prey species, the figures varied greatly from month to month. This was due partly to the hawks switching hunting emphasis from woods to villages during the course of the breeding season, and partly to a change in the hawk to prey ratio. Within each summer the prey multiplied more rapidly than the hawks did, so that, as the months passed, the number of prey per hawk increased. In consequence, the percentage mortality inflicted on the prey by the hawks declined through the season (Table 19). In addition, different species showed differing overall mortalities, which influenced the proportion due to hawks. The Great Tit and Coal Tit had higher overall mortalities than did the House Sparrow and Chaffinch, but the contribution of Sparrowhawks to the total mortality in summer was greatest in the House Sparrow, intermediate in the Chaffinch and Great Tit, and very low in the

The Sparrowhawk as a predator

129

Coal Tit. The figures obtained for different species also varied somewhat from year to year. In one year, Goshawks preyed more heavily than usual on Sparrowhawks, and thus indirectly reduced the predation on all songbirds in that year. From a knowledge of the diet of other predators present in his study area, Tinbergen was able to draw further conclusions on the importance ofpredation to different species, and on the collective contribution of all predators to the total summer mortality of certain birds: (a) (b) (c) (d)

(e)

in the House Sparrow, all predators together were important, and exceeded all other mortality agents; in the Chaffinch and Great Tit, predation formed no more than half of the total mortality; in the Coal Tit, the role of predators was practically negligible, and most mortality was due to other factors; in the Jay, which was also taken frequently by the Goshawk, the two accipiter species together probably accounted for a large part of all the mortality. The same was true, but to a lesser extent, for the Turtle Dove. in the Swallow, about the same numbers were taken by Hobbies and Sparrowhawks. Individual Hobbies took more Swallows than did individual Sparrowhawks, but were less numerous in the area.

These findings were probably applicable to some other parts of the Netherlands at that time, and clearly indicated the differing importance of predation to different species. In some species predation accounted for more than half the total mortality, but in others practically for none. Hence, other mortality agents, such as disease and starvation, must similarly have varied between species, and between habitats for the same species, achieving greater importance wherever predation was slight. The same could be said of the egg and nestling stages of songbirds, because some species experienced heavy predation at these stages, and others practically none. Consequently, the role of predation varied not only between species and areas, but also between different stages of the bird's life. Where predation forms a large part of the overall mortality, predators are clearly an important proximate factor influencing numbers. However, it would be another step to say that they are actually limiting the numbers of their prey, because if they were removed, some other mortality agent may well take over this role, so that no long term increase in numbers would occur. Furthermore, predators need not have a wholly negative effect on a population, but could in some cases promote an increase in breeding numbers. Imagine a population, such as that of a seed-eater, which has a fixed stock of food to last the winter. High numbers in autumn may cause the food supply to be depleted quickly, and few birds to survive the winter; but lower numbers in autumn (resulting from predation) could ensure that the food supply lasted for longer, so that more birds survived the winter than might otherwise have done so. Similarly, if predators selectively removed

130

The Sparrowhawk as a predator

diseased individuals as they appeared, they might help to suppress epidemics, and thus contribute to the maintenance of numbers. In this instance, they would also be taking birds which were destined soon to die, and so have no negative effect on numbers. However, little or no information is available on these aspects. Considering when it was done, Tinbergen's was a remarkable study, involving a great deal of tedious census work and calculation. Unfortunately, it was not possible to measure all the important parameters, and some could only be estimated roughly. Nonetheless, no-one has attempted such a mammoth exercise since, involving the repeated censusing of several different prey species, and to this day the study remains a classic in its field. The whole work has a ring of credibility about it, and was obviously written by someone who knew Sparrowhawks and their prey extremely well. Not surprisingly, it was not possible to state conclusively whether Sparrowhawks were in the long term depressing the breeding populations of their prey. This question could be properly answered only by experiment, in which the Sparrowhawks were removed for some years, and the prey populations monitored, along with appropriate controls. As it happens, this experiment was done crudely and unintentionally over much of Europe around 1960, when Sparrowhawks were largely wiped out by use of organochlorine pesticides. No large increase in songbird breeding numbers occurred, though there were changes in the population dynamics of some species, as discussed later in this chapter.

Predationon tits More recently, Sparrowhawk predation on Great Tits and Blue Tits has been studied in particular detail by Tim Geer and Chris Perrins in the 320-ha Wytham Wood near Oxford. The situation was ideal, because the majority of tits bred in nest boxes, and were already subject to a long-term study, in which adults and young were ringed each year. It was thus possible, not only to assess the proportion of tits taken by the hawks, but also, from the rings retrieved at hawk nests, to learn which individuals had succumbed. Initially, most of the rings which were recovered came from pellets regurgitated by the nestling hawks and collected by the observers. Later, the observers learnt that, if they sat in a hide near the nest, they could reach out with a pair of long tongs each time the hen arrived with prey, take the item from her, snip off the leg with the ring, and then return the item to the nest. The hawks soon became used to this procedure, and became increasingly reluctant to surrender their prey, sometimes engaging in a prolonged tug-of-war with the tongs. The birds also became accustomed to waiting for the return of their prey before they could feed their young. By this method, all the rings brought to a nest could be acquired, instead of only the proportion found in pellets. In all, nine hawk nests were studied intensively during 1976-79, out of the total hawk population in the wood of 6-8 pairs per year. Early in the breeding season, the predation on adult tits was considerable.

The Sparrowhawk as a predator

131

Geer (1978) calculated the proportions of nest boxes that were occupied by tits, and the proportion of these occupied boxes which produced young, at varying distances from each hawk nest. Within 60 m of each successful hawk nest, both the occupancy and success ofboxes were depressed, compared with boxes further away. In other words, in the area immediately around each active hawk nest, relatively fewer of the available nest boxes were used by tits, and of those which were used, relatively fewer produced young (Table 20). No such phenomena were noted around hawk nests which failed soon after eggs had been laid, nor around one successful nest on the edge of the wood, whose owners hunted chiefly in farmland. Results were consistent from one year to the next, even though the locations of the hawk nests changed. In particular localities, depressions in tit numbers occurred only in the years when hawks were present. Hence, the findings could not have been due to some other feature of the wood in these localities, which was inimical to tits, but could only have been due to the hawks themselves. Observations indicated that the areas around the hawk nests were in fact occupied by territorial tits in spring. The subsequent fall in the numbers breeding was, presumably, therefore due to locally heavy predation on the tits, before they could start nest building, rather than to tits avoiding settling near hawk nests. Similarly, the reduced success of tits which did breed near hawk nests was apparently due to hawks continuing to take adult tits during the season, so that growing numbers of nests were deserted. The female tit incubates the eggs and small chicks, so if the food-providing male dies at these stages, she can no longer continue with the breeding attempt, and abandons the nest. These findings ran contrary to the old belief that raptors did not kill close to their own nests; but they were thoroughly in line with results from our radio-marked Sparrowhawks in Scotland, which often hunted close to their nests in the early part of the breeding season. Again in agreement with our Scottish findings, the hawks near Oxford took their greatest toll from the young which had recently left the nest (Perrins & Geer 1980). As recovered rings showed, the young tits were especially prone to capture in their first five days out of the box, but then became much less vulnerable as they became better able to move around the canopy. The 'number of young taken varied between different hawk pairs, depending largely on how well the nestling period of the hawks coincided with the fledging dates of the tits. Early hawk pairs, which were feeding young when the tits fledged, took large numbers, whereas late hawk pairs, which hatched their eggs only after the majority of tits had fledged, took relatively few. This trend was evident even at single hawk nests, in that the later the date after tit fledging, the fewer the tits in the diet (Fig. 38). Not only did the hawks switch increasingly to other prey as the tits matured, but the total food deliveries by the male hawks declined, perhaps because the completion of growth in young tits marked the end of a period of easy feeding. It was about this time that the female hawk began to help with the hunting, but she took mainly larger prey than tits. The known Sparrowhawk population ofWytham Wood took an estimated

132 The Sparrouihaiok as a predator 50

•

40

•

Q)

""0

c C/)

3

20

•• • • • • • •

?fi?-

10 Days

20 after

• 30

mean

tit

•

,

•

•• 40

50

fledging

Fig. 38. Percentage ofyoung tits in the food ~f Sparrowhawks shown in relation to the mean fledging date of tits. The diagram is based on eight intensively studied hawk nests near Oxford, each of which provided two points, one for days 1-20 after hawk hatching, the otherfor 21 days onwards. Redrawnfrom Perrins & Geer 1980.

34%, 18% and 33% of young Great Tits in the three years, and 27%, 18% and 27% of young Blue Tits. This seemed a lot, but more than six young tits were raised per pair (including failed pairs) in each of these years, so with up to a third taken, there were still plenty left to maintain the numbers. Only about one chick per tit pair is normally needed to replace the losses among breeding adults in order to maintain a stable population. Hence, many young tits must have died at other times of year or from other causes. The proportion of young tits taken in any given year was dependent partly on the number of hawk broods in the area (3, 4 and 5 in 1976-78) and partly on the number of tit broods. Within years, the two tit species were taken in approximately the ratio in which they were present in the wood. For the first three weeks after fledging, the selection was slightly in favour of Blue Tits, shifting thereafter towards Great Tits (which were also favoured in Tinbergen's area). However, compared to the 18-34% of young taken, only about 14% of adult tits were killed in summer. This again revealed the greater vulnerability of the young, but was statistically significant in only one of the three years studied (Geer 1982). Among the adults, males were taken more often than females. Possibly the males were more vulnerable as they sang on exposed perches or foraged

to feed their brooding mates, safe within the boxes. To judge from ring numbers, young were taken from all parts of the wood, and not just from near the hawk nests. However, a succession of young from the same nest were often brought in, suggesting that a given hawk returned repeatedly to where it had found a brood at the right stage. When young

The Sparrowhawk as a predator

133

which were taken were compared with those which were not, the only selection apparent was for late fledged young over early ones, when both were available. This was because the hawks took late fledged young which were still in family parties in preference to early fledged young which were foraging independently in the tree-tops. The early young had suffered greatest predation at an earlier date, when they were still in families. In terms of weight, brood size and age of mother, young tits killed by Sparrowhawks did not differ from those which were not killed (Geer 1982). Many of the tits, both young and old, would no doubt have lived longer if they had not been taken. Hence the hawks not only reduced the size of the post-breeding population in tits, but they also changed the seasonal pattern of mortality from what it might otherwise have been. Predation was not studied at other times of year, but some indication of the overall effect of Sparrowhawks on the Great Tit population could be gained by comparing results obtained when hawks were absent from the Oxford area (due to high organochlorine use) with results from later years when hawks were present. The Great Tit breeding population fluctuated from year to year, but in general the numbers of nest boxes occupied in each year were no higher in the period when hawks were absent than in the period when hawks were present (Table 21). These results thus gave no evidence that Sparrowhawks depressed the Great Tit breeding population in the long term. The hawks must, however, have reduced the breeding tit population very slightly, because they continued to eat tits in the prebreeding stage, and caused the local depressions in tit numbers around their own nests, reported above. Evidently, the effects of this were so slight as not to show in the overall data. As practically all the tits which fledged in the wood were ringed, and most of the breeders were identified, it was possible to distinguish immigrants from locally hatched birds trapped in the boxes each year, and also to examine the survival of breeders from one year to the next. Comparing periods, two tendencies emerged. First, the Great Tit breeding population contained slightly more immigrants in years when hawks were present than in years when hawks were absent; and second, the annual survival of adult tits was slightly lower in the years when hawks were present. Neither trend was significant statistically. The only other difference noted between the two periods was that the proportion of unexplained nest failures was 5 0/0 greater in years when hawks were present. Perhaps many of these unexplained failures were due to hawk predation on the breeding adults, leading to nest desertion. These findings for tits could probably be generalised to many other prey species, in that the main effects of Sparrowhawk predation are to: (1) change the seasonal pattern of mortality; (2) reduce the size of the post-breeding peak in population; and (3) change the main agents of death; all without causing any noticeable long-term decline in breeding numbers. In practice, this means that, instead of dying mainly in winter from food shortage, for example, prey birds may die at all seasons (especially just after fledging), and largely from predation. Sparrowhawk predation may also result in 1110re

134

The Sparrowhawk as a predator

local movement than might otherwise occur, if it depresses populations in some localities, which are then made good by immigrants from elsewhere, where predation is less. In Britain, prey populations in woodland receive more attention from Sparrowhawks than do those in gardens and hedgerows, so that the latter. may produce enough young to offset losses in the former. Nonetheless, caution is needed in extrapolating from tits to other prey, or even to tits in a different area, for, as explained earlier, the pressure on anyone species depends partly on the number of alternative prey available, and this will vary from one locality to another. Predation on wintering waders

In a recent coastal study in southeast Scotland, the impact of raptor predation on overwintering waders was estimated from frequent counts of the waders themselves, together with searches for signs of kills (Whitfield 1985). Almost all the recorded mortality was due to predation by raptors, chiefly Sparrowhawks. As in song-birds, species varied greatly in vulnerability. The proportion of the total taken was estimated over two winters for Redshank Tringa totanus at 20% and 16 %, for Turnstone Arenaria interpres at 4% and 5% and for Knot Calidris canutus at 1·5% and 0.5 %. Over one winter, the same proportion was estimated for Purple Sandpiper Calidris maritima at 1.5 %, for Dunlin Calidris alpina at 4 %, for Grey Plover Charadrius squatarola at 2% and for Ringed Plover Charadrius hiaticula at 19%. These figures were minima because some kills were almost certainly missed. In at least four species, juveniles were taken in greater numbers than expected from their proportion in the population, probably because they tended to feed on the edges of flocks and were last to fly when a hawk approached. No mortality from hawk predation was found in Curlew Numenius arquata, Oyster-catcher Haematopus ostralegus or Bar-tailed Godwit Limosa lapponica. Presumably the large size of these species was sufficient to protect them from Sparrowhawks, in the presence of much smaller prey. In conclusion, avian predators - mainly Sparrowhawks - caused the bulk of recorded winter mortality in. the small-medium sized waders studied. This mortality was substantial in one species, but small or negligible in the others. Most previous studies had concluded that virtually no mortality in coastal waders was due to raptors, but these studies were done before the late 1970s, at a time when raptor populations were much depleted by organochlorine pesticide use. The situation is now different, as Sparrowhawks and other raptors are being increasingly seen on sea coasts. However, it is unlikely that their predation is sufficient to depress the breeding populations ofwaders, although this needs checking.

SOME FURTHER OBSER V ATIONS

Evidence for the lack of effect of Sparrowhawks on the breeding populations of most prey species came during the years when hawks were virtually absent

The Sparrowhawk as a predator

135

from large parts of the country. No great upsurge in prey populations occurred then, and similarly no decline in such populations was noted when Sparrowhawks returned. The Common Birds Census of the British Trust for Ornithology showed that most songbirds fluctuated in numbers over this period, but in a manner independent of Sparrowhawk numbers. No other predator in Britain could take on the role of the Sparrowhawk so, as in the Oxford tits, other mortality agents among small birds must have achieved greater importance in the years when hawks were lacking. This is not to say that Sparrowhawks could not depress or even exterminate a species in certain circumstances. What we observe now are the stable relationships, involving prey species which have maintained their numbers over the years, in spite of Sparrowhawks. Any species which could not withstand this pressure would have long since disappeared, and even today occasional species may be restricted in distribution by Sparrowhawk predation, if this is too severe outside the existing range or habitat to enable the species to expand. In particular, rare but easily caught species may be eliminated locally if Sparrowhawks are attracted and maintained in the area by some much commoner but less vulnerable prey, as described earlier. Thus I was surprised at how many escaped budgerigars appeared among prey remains in south Scotland; so many in fact that, if ever budgerigars escaped in greater numbers, their establishment as a breeding species would almost certainly be curtailed at an early stage by Sparrowhawks. This would only be possible, however, because Sparrowhawks were not dependent on budgerigars, but were supported by much commoner prey. If they depended on budgerigars alone, the hawks would soon be extinct. We have no way of knowing whether the distribution of any wild species is being checked by Sparrowhawk predation, but one species which did expand when Sparrowhawks were lacking was the Bullfinch Pyrrhula pyrrhula. At that time Bullfinches started to feed in more open situations, further from cover, than had formerly been usual. The absence of Sparrowhawks may have encouraged this, and thus greatly increased the amount offood available to these finches, in turn enabling their numbers to rise (Newton 1967). Since the return of the Sparrowhawk, the Bullfinch population has dropped to a somewhat lower level. There is no way of testing for cause-and-effect, however, and so the changes in Bullfinch numbers remain open to other interpretations. Much discussion has centred on the extent to which predators remove the sick members of a prey population, thus 'keeping the population healthy' and at the same time having minimal effect on numbers. Only if sick prey were more conspicuous, or less wary, would they be more vulnerable to Sparrowhawk predation than healthy prey. Those raptors, such as the Merlin, which chase their prey over long distances would have greater scope for selecting weak individuals, for it is chiefly in situations where the prey are taxed in some way that the weak are disadvantaged. The Sparrowhawk, taking most of its prey by short surprise attacks, is likely to be much less selective in this respect, but no usable information is available.

136

The Sparrowhawk as a predator

While there is no good evidence of Sparrowhawks having a widespread effect on the breeding numbers of their prey, there is circumstantial evidence for the reverse, that Sparrowhawk numbers are affected by prey numbers. Not only do hawks breed more densely in areas where their prey are most abundant (Chapter 4), but in anyone area the numbers of hawks fluctuate from year to year according to changes in prey numbers (Chapter 6). This evidence is entirely correlative, however, so is not wholly conclusive. The numbers of many prey species seem to vary from year to year, depending partly on the winter weather and food supplies. If the numbers of Sparrowhawks are similarly determined by food supplies, then both prey and predator could be said to be 'resource limited'. By increasing in numbers when and where prey become abundant, the hawks show the so-called 'numerical response' to their prey. By increasing the proportion of certain species or age groups (such as fledglings) in their diet at times when these are plentiful, the hawks also show the so called 'functional response'. These terms are used by some ecologists to distinguish the two ways in which a predator may increase the pressure on its prey, in response to an increase in prey numbers. Such 'density dependent' responses of a predator have a stabilising influence on the numbers of the prey, as proportionally more prey are taken when prey numbers are high than when they are low. In this sense, if the density-dependent response is strong enough, the predator may be said to 'regulate' the numbers of its prey. As we have seen, however, if the predator is removed, the prey may not necessarily increase in the long term because other con trolling factors take over. A reduction in one form of mortality is compensated by an increase in others, so that the prey population stays near the level that resources will support. Evolutionary aspects

Sparrowhawks have another longer term influence on their prey species, besides affecting the pattern of mortality. By continually removing the more vulnerable individuals from a prey population, Sparrowhawks act as an important agent of natural selection, gradually moulding the appearance and behaviour of the prey, and thus influencing the course of evolution. As the Sparrowhawk is the main predator on woodland songbirds throughout its range, it must presumably have played a major role in the evolution of all these species, having influenced much of the behaviour and colour pattern which we see in these birds today. While the Sparrowhawk, in the long term, encourages the evolution of prey which are ever more effective at avoiding it, the hawk itself is under selection to become ever better at outwitting the prey. Most predator-prey systems may be stable, with neither predator nor prey becoming extinct, because the prey are always one step ahead in the arms race. Consider the effects of the so-called 'life-dinner principle' (Krebs & Davies 1981). When a song-bird evades a Sparrowhawk, the song-bird saves its life, but the hawk loses only its dinner. The cost of a mistake is clearly greater for the prey. Admittedly, a hawk which never caught a prey would starve to death, but

The Sparrowhawk as a predator

137

the fact remains that natural selection will have been much stronger on improving the ability of prey to escape than on the ability of hawks to make successful captures. We may ask why the prey have not become so efficient at escaping as to drive their predators extinct. One hypothesis is that, as predators become rare because of increased prey efficiency, they exert little selection on the prey for further improvement. A second possibility is that the more efficient the prey become, the less scope there is for further improvement. Thirdly, the prey may be compromised, and able to improve their escape ability only at the cost of some other aspect important to survival. There would then be limits to the evolution of escape behaviour.

SUMMARY

Sparrowhawks do not take their various prey species in proportion to numbers in the environment. Species which feed far from cover, or which are conspicuous by appearance or behaviour, are taken more often relative to their numbers than are scrub-dwelling, inconspicuous, or fast-flying species. The more vulnerable species include Great Tit and House Sparrow, and the less vulnerable include Wren and Spotted Flycatcher. Sparrowhawks probably account for more than half the total mortality in certain prey species, and only a negligible proportion in others. The extent of their predation varies greatly from one place to another in the same general area. Through predation, Sparrowhawks near Oxford depleted the breeding population of tits nesting within 60 m of their own nests, compared with those further away. They preyed heavily on young tits, the proportion in the hawk diet declining from the date of tit fledging. Although Sparrowhawks took up to one-third of all young tits fledged each year within the tit fledging period itself, the hawks had no noticeable long-term effect on the size of the breeding tit population. Their main effects were to alter the seasonal pattern of mortality in the tit population, reduce the size of the post-breeding population peak, and reduce the numbers dying from other causes. When Sparrowhawks were largely absent from south-east England during the height of the organochlorine era, no obvious increase appeared in songbird breeding numbers (except in one species) and no obvious decline occurred when Sparrowhawks returned. This implied that Sparrowhawks did not determine the level of the breeding populations of their most commonly taken prey.

CHAPTER 10

Breeding season

In common with other birds, Sparrowhawks breed during that part of the year when th eir food is most readily availabl e. The populations of songbirds , on which the hawks rely, reach their lowest level of the year in early spring, increase through breeding to reach a peak in late summer, and then decline again over winter. This pattern of availability is modified to some extent by the arrival and departure of migrant songbirds, which serve to swell the prey supply, both in summer and, by different species, in winter. Young songbirds, wh ich have rec ently left th e nest, provide especially easy pickings , and these are available from late spring well into summer. In addition, longer summer days giv e the hawks more tim e to hunt, but on the other hand the growth in vegetation gives th e prey more cov er as the summer progresses . Taking all factors into account, however, it is not surprising that Sparrowhawks breed wh en they do -late April to late Au gust in most of Britain. In anyone region , Sparrowhawks lay their eggs on about th e same date each yea r, but as on e moves from sou th to north in Europe, th ey lay progressively later, corresponding with th e later spring. In the M editerran ean region, they lay mostly from mid April to mid May, in middle Europe and Britain mostly in M ay , and in northern Europe mostly from mid May to mid June.

Breeding season

139

Also, corresponding with the shorter northern summer, the breeding season is itself shorter than in the south. This is achieved by a greater synchrony in the starting dates of individuals and by a shortening of their individual breeding cycles, especially of the pre-laying period, when nests are built. Everywhere, Sparrowhawks raise at most one brood each year, but over much of their range the season is sufficiently long for some pairs to have time for a repeat clutch if their first fails at an early stage.

Some general points In order to make the best use of its food supply, any bird species should lay its eggs so that the period of peak food demand, when the young are being fed, coincides with the period of peak food supply. Because of the time needed to build a nest, to lay and incubate the eggs, the bird m us t start preliminary breeding activities long before the food supply reaches its peak and, in the case of Sparrowhawks, long before it even begins to increase. Many bird species in temperate regions have been found by experiment to respond to the lengthening days of spring, and to begin to develop their gonads, re-establish territories and build nests at a suitably early date in preparation for subsequent breeding. In matters of timing, therefore, it is useful to distinguish between 'ultimate' factors, such as food supply, which provide the reason for breeding at a particular season, and 'proximate' factors, which bring the birds into condition at an appropriate early date. Some proximate factors, such as daylength, provide the primary stimulus, while others, such as food availability at the time, may bring about minor modifications in laying dates. In this case, food acts as an ultimate and as a proximate factor in the timing of breeding. So far, only three raptor species (though not the Sparrowhawk) have been subjected to experiments involving light. When exposed to increasing artificial daylengths at the 'wrong' time of year, all responded by achieving breeding condition at the 'wrong' time of year (Newton 1979). In such species, increasing daylength probably serves not only to stimulate gonad growth, but also to raise the female's appetite, so that she responds appropriately to a sufficient food supply, and lays down body reserves prior to laying. Without primary control by daylength, the bird might respond to a temporary food abundance at an inappropriate season, when it would have little chance of raising young. The reproductive development of the hen in early spring entails the growth of both the ovary and the oviduct. About two weeks before laying, the individual eggs within the ovary begin to swell, as a result of yolk deposition, and, as laying begins, the hen loses feathers from her lower ventral surface and develops a brood patch. This is an area of bare, well vascularised skin, through which heat is transmitted to the eggs during incubation. The equivalent development of the cock entails the growth of the testes and accessories, but no extra fat deposition and no brood-patch formation, as the cock does not incubate. In both sexes, these various physiological changes are accompanied by appropriate behaviour, such as nest building.

140 Breeding season 100 c .2

(a)

80

0 S

00

0-

.x ~

0

:x:

f/.

60

40 20

(b)

60

4i

:.0 .S

40

CJ)

0>

.5 "01 -0

e>

Li:

;t.

20

April

May

June

July

August

September

Fig. 39. Breeding of Sparrowhawks in relation to food supply in south Scotland. (a) Proportionofthe breeding hawk population laying, incubating orfeedingyoung at different dates through the spring-summer. (b) Proportion offledgling prey-birds in the diet at different dates through spring-summer. The breeding season of the hawk population coincidedwith theperiodwhen vulnerablejledglings wereavailableand taken. BREEDING AND FOOD SUPPLIES

In any population of prey birds, the easiest individuals for a hawk to catch are the fledglings, young birds which have just left the nest, and can move around, but cannot fly strongly. Each young songbird is at this vulnerable stage for only 1-2 weeks, after which its feathers have grown and it can fly well. However, because songbirds of one species or another breed throughout the summer, fledglings are available over several months (Chapter 8). In south Scotland, fledglings first appeared in the Sparrowhawk diet in late April, increased through May to reach a peak in mid June to mid July, and then declined through August (Fig. 39). Their frequency in the diet thus closely paralleled changes in their abundance in the environment. For several weeks in mid season, fledglings formed more than half of all the prey eaten by Sparrowhawks, the rest consisting of older juveniles and adult birds. Of course, the total prey population in the countryside increased through the summer, as fledglings matured, to reach peak numbers in August. In south Scotland fledgling songbirds seemed crucial to Sparrowhawk

Breeding season

141

breeding. Their sudden appearance in the bare countryside marked the end of a seasonal food shortage for the hawks. In most years, the first fledglings to appear were those of Mistle Thrush, Blackbird and Song Thrush in that order, but in two years these were preceded in the Sparrowhawk diet by nestling Woodpigeons, and in one year by fledgling Siskins and Crossbills. However, these last species were too irregular or local to provide a reliable food source for any but a few Sparrowhawk pairs in certain years, and it was the three thrush species which were important to the majority of hawks. Following the thrushes, Robins, Chaffinches and various other birds began to breed, producing a succession of fledglings through the summer. The latest to occur in the hawk diet each year were mainly those of various finches in August, but young Woodpigeons continued into September. The Sparrowhawk breeding season in south Scotland coincided almost exactly with the period when fledgling songbirds were available, and the hawks continually switched emphasis from one prey species to another, as each produced its young (Fig. 39). Individual hawk pairs took about 14 weeks from the start of laying until their young became independent, but there was a spread of up to six weeks in the laying dates of different pairs, giving up to twenty weeks for the breeding season of the whole population. In most years, the first hawks to lay eggs started within 5-10 days after the first fledglings had appeared in the diet, both events occurring later on high ground in the study area than on low ground (Fig. 40). Other hawks followed, as the supply of fledglings increased. Moreover, the period of peak food demand, when the majority of hawks had young, coincided with the period of peak fledgling supply, midJune to midJuly (Fig. 39). As mentioned earlier, Sparrowhawks needed extra food, even to start breeding. The majority laid only after the flush of vulnerable fledglings became available. They may therefore have been dependent on this increased food supply to produce their eggs, and it seems unlikely that in most years the first hawk pairs could have laid any earlier than they did despite, in some cases, having completed their nest weeks previously. Laying dates and spring weather

In south Scotland, Sparrowhawks began laying between late April and early June, with the majority in the first two weeks of May (Fig. 40). During 1971-81, the annual variations in the dates that the earliest clutches were started spanned about ten days (22 April-2 May), as did the annual variations in mean dates (5 May-15 May), and in latest dates (23 May-3June). Earliest and mean dates each year tended to vary in parallel, but the latest dates varied independently, giving a total spread for first egg dates of 25-39 days in differen t years. Because most songbirds began breeding earlier in mild springs, their young were available to Sparrowhawks earlier than in cold springs. Other work showed that Sparrowhawks had more difficulty in obtaining prey in wet weather than in dry. So, if egg production by Sparrowhawks depended on food supply, their laying dates could have been influenced by rainfall as

142 Breeding season

!

{ a}

40

20

I,

15

30 ~

10

i

~

20

I,

5

-r I

!

'"0

~

Q)

(;

en

:.c

(b)

.s

ct'

cv

J::.

20

t.>

'"5 U

CI)

01

.S;;

0'"0

Cl)

~ 0

t;

40

15

30 10

20

10

i-

5

.> April

May

June

Fig. 40. Start ofbreeding among Sparrowhawks (columns) in relation to the appearance offledglings in the diet (lines): (a) upland; (b) lowland, south Scotland. well as by temperature. I therefore investigated the relationship between laying date (of first eggs) on the one hand, and mean temperatures and number of rain-days on the other, during the periods of February-April, March-April, and April each year. The three periods were included in order to explore how long, prior to laying, weather may have been influential. Over the years, laying was indeed later in the colder and wetter springs (Fig. 41). For April, the number of rain days gave the strongest relationship, and accounted for 40% of the annual variation in average laying dates; mean temperature accounted for 32% of the variation in laying dates, and the two factors together for 50 % • This latter value was higher (83 % ) when the weather over March-April was considered, and about the same (85 % ) for February-April (Table 22). This implied that weather in the 2-3 months

~ 0

Breeding season

143

before laying was important, as well as that in the one month before laying. Overall, it seemed that about four-fifths of the total variation in mean laying dates between years was associated with the weather features examined, and the rest with other factors. Spring weather thus provided a good predictor of Sparrowhawk laying dates. Temperature could have been involved in two ways. It may have influenced the energy needs of the hawks, so that in warm weather more of their daily food intake could go towards improving body weight than in cold weather. Secondly, because it influenced the dates when the usual prey species began to breed, it also influenced the dates when their young first became available as food. Earlier laying in warm springs has been well documented in many songbird species, including various thrushes, tits and finches (Myres 1955, Lack 1966, Newton 1964). This relationship with prey probably caused some of the association between spring temperatures and Sparrowhawk laying dates. Rainfall may have influenced laying dates through its effect on hunting success and food intake. During nest watches, we found that prey deliveries were reduced by one third during rain periods, compared with dry ones (Chapter 14). Any reduction in food intake in early spring could have delayed the date at which hens could produce eggs. The later laying in upland, mentioned earlier, was associated with colder temperatures and higher rainfall than in nearby lowland, which would have delayed laying by Sparrowhawks and their prey species. 15

(0)

15

•

•

(b)

>-

0

~

10

.s 10 cv

0

"'0

>-

~ c 0

~

~

5

OL...-

6-0

5

• .....L-

7·0

----..

8·0

Mean temperoture °C in April

•

~

5

10

Rain-days in April

Fig. 41. Relationship between mean laying date in different years and (a) mean April temperature; (b) number of April raindays, south Scotland. Egg laying was earlier in warm dry springs than in cold wet ones. Between laying date and mean April temperature, r = 0,57, P < 0'05; between laying date and April raindays, r = 0,63, P < 0·05.

144

Breeding season

There remained the question of why the start of egg laying in the hawk population as a whole varied in spread between 25 and 39 days in different years. Laying seasons tended to be slightly shorter in years when laying began late, but other factors were probably involved too. By analogy with other birds that have been studied, daylength change is the primary controlling factor, which ensures gonad regression in summer, and prevents the birds from laying at an ecologically inappropriate time but, as with starting dates, the finishing dates may be modified slightly by other factors.

WEIGHTS AND BREEDING

Within the normal breeding season, weights of hen Sparrowhawks served to emphasise the dependence of reproduction on food supply (Fig. 42). In the three weeks before laying, the hens gained 40-50 g (15 % ) on average, and became heavier then than at any other time of year. This was due not only to ovarian and egg development, but also to the accumulation of extra body reserves, mainly fat, as was evident in females handled at this stage. Moreover, hens which were caught at nesting places in May, but which did not lay (some of these places had new nests and others did not) showed no weight increase. I t seemed, therefore, that high weight was necessary for laying. Birds which did lay eggs lost, on average, only about 20 g in weight over the laying period, yet the clutch itself weighed up to 180 g, so that, even allowing for the extra water in eggs, their formation must have been dependent on a sufficient food supply at the time. Reserves could, however, have accounted for the yolks in the eggs, the 'whites' resulting from food consumed during the laying period. The ability of the hen to acquire enough body reserves with which to begin incubation may have been at least as critical in influencing whether eggs were laid as an ability to produce the eggs themselves. Weight was maintained at high level into incubation, and in most individuals was not reduced until after hatch. The extra weight evidently acted as a reserve, helping to tide the hen over any spells in incubation when the cock was unable to maintain food deliveries. The hen would not then have to abandon the clutch and seek her own food. Hens which had hatched their eggs lost weight markedly through the nestling period. This was because, as we learnt from nest watches, they gave almost all the food brought by the cock to their young, taking for themselves only the feet and other parts of prey items which the young refused. Thus, another function of the body reserve, if not used during incubation, was to enable the hens preferentially to feed their young at the expense of their own condition. Small, rapidly growing chicks were more vulnerable to temporary food-shortages than their mothers were, so at this stage of breeding the hens body reserves seemed even more useful than during incubation. Furthermore, hens may have benefited from some loss of weight at this stage, because from about the time their young were half grown, they began to

Breeding season

145 Females

12

15

17

14

"i7

15

(19:3,

88

68

36

21

17

13

13

16

13

1

360

I

r

340

320

I .b 300

II1 Ii r

I

! I

I

TC

•II •I I •I

• 280

I

•I

I

260

• Males

11

15

18

12

15(12)6

170

160

150

I •

•I

·r

I • • • • •I I •I •.9 • I

I

I

I

I

I

I

I

I

•

•II

I

• • I •I I •I

. t

I

• •II

140

130

Pre -lay

Fig. 42.

lay

Incubation

Nestling

I

Post-fledging

Mean weights (±SE) of adult Sparroiohaioks in each 5-day period through the breeding cycle, south Scotland. Numbers show sample sizes in each period, and open circles refer to non-breeders. Females above, males below. The variation in female weight through the cycle was statistically highly significant, and breeders were significantly heavier than non-breeders during the laying/incubation periods. Significant differences were found between periods labelled: a and d, a and e, band e, c and e, e and 1, e and h, f and h, g and h, d and j, e and j. Mean male weights did not differ significantly between any two periods. Re-draum from Newton et al1983.

146

Breeding season

help with the hunting, and may have flown more efficiently if they weighed less. Indeed the loss of weight at this stage may have been the stimulus which caused the hens to hunt again. The hens reached their lowest weight of the year around the time the young left the nest, and only after the young became independent, some weeks later, did the hens begin to recover weight (Chapter 2). These were average trends, and from a comparison of different individuals, weight emerged as even more clearly related to breeding performance. Thus hens which gained weight rapidly in the spring laid early, while those which gained weight slowly laid late (Chapter 16). Second, the number of eggs laid correlated with body weight (Table 23). Hens which laid large clutches had higher weights in the immediate pre-lay period than did hens which laid small clutches, and a similar difference was even more marked in the last stages of laying (6-10 days after the first egg) and in the early stages of incubation. Comparing birds examined in the early and late stages of laying, the weight loss seemed less in birds which laid large clutches than in those which laid small clutches. Hence, large clutches were associated with a high weight before laying and with a relatively small weight loss during the laying period, whereas small clutches were associated with lower weights before laying and with a bigger reduction during laying. It was thus perhaps a decline of body condition below a threshold during the laying period which curtailed egg production in some hens, resulting in a small clutch. Following from this, weights at the start of incubation were also related to clutch sizes, being highest in birds which had produced 6-7 eggs, and lowest in birds which had produced only 1-2 eggs. Furthermore, the larger the clutch, the more likely it was to survive till hatch. Desertion was the main cause of failure, and its incidence decreased progressively in clutches of increasing size: all one-egg clutches were abandoned, compared with 74% of two-egg clutches, 26% of three-egg clutches, 8% of four-egg clutches, 4% of five-egg clutches, 2% of six-egg clutches, and no seven-egg clutches. Other types of failure were independent of clutch size. During the later stage of laying, females which subsequently succeeded in their breeding were significantly heavier than females which failed, the means (± S.E.) being 337 ± 10 g and 306 ± 5 g respectively. Only at this stage was such a significant difference found, and most nest failures occurred soon afterwards, as clutches were deserted early in incubation. The ability to maintain high weight through the laying period may therefore have been important for continuing incubation. These various findings suggested that an ability to reach and maintain a high weight was crucial, not only to the date of laying, but to the number of eggs laid, and to their subsequent success, in enabling the hen to remain at the nest. Presumably, it was an insufficiency of food which prevented some hens from achieving or maintaining a high weight, and which caused them to produce few or no eggs, and subsequently to desert their clutches and start hunting.

Breeding season

147

From before egg laying until the chicks were partly grown, the cock provided the food for the hen, so it was upon him that the hen's body condition depended. It was he, through his own hunting success, who determined the fate of the breeding attempt. The mean weights of cocks which we studied remained relatively constant through the breeding cycle, but showed lows at two crucial times. The first occurred in the pre-laying period, coinciding with the time of maximum weight gain in the hens. Presumably the cocks gave as much food as possible to their hens at this critical stage, even at their own expense. The mean weight of cocks recovered during incubation, but fell again in the late nestling/post-fledging stages, coinciding with the period of peak food demand by the young. These weight changes were relatively slight, however, compared with those in the hens. It was the physiological condition of the hen, associated with the development of breeding condition, which pre-disposed her to lay down body fat, given enough food from the cock. At no stage did cocks accumulate such large body reserves as hens. This could be due to their having to maintain an active hunting role throughout breeding, in contrast to the larger hens which - providing the cock could feed them - could adopt a passive storage role for much of the cycle. The advantage of body reserves may have contributed during evolution to the clear division of labour between the sexes, and to the female having the storage role. She has to put on some extra weight for egg production anyway, and, being larger than the male, has a potentially greater storage capacity. With these conditions, it is a small step to cessation of hunting and to accumulation of even greater reserves to serve during the incubation and nestling periods. The one essential is that the male should be able to supply enough food for both. This reasoning would apply to other raptors, too, and may help to explain the evolution of the female's larger size (Chapter 25).

Feeding experiments To some extent, we were able to test the above correlations between weights and breeding by providing some Sparrowhawks with additional food. In particular, hens which were given extra rations in the pre-lay and laying periods produced earlier and larger clutches than did other hens which had only natural food at this stage. We did this feeding experiment in the food-poor habitat of Ae Forest, where Sparrowhawks had previously bred with poor success. We firs t assembled in deep freeze a stock of pigeon and quail carcasses for the purpose. Then, in the pre-lay period, we placed the unfrozen carcasses, one every 1-4 days as necessary, on plucking posts near the nests. Thus food remained available each day on the territory, should the hen wish to eat it, but each hen would also have obtained an unknown amount of food in the normal ways. The carcasses were fixed firmly, so that the bird could not remove them and had to eat them on the spot. Only in this way could we be sure of finding the remains to judge whether a Sparrowhawk, and not some other animal, had eaten the carcasses. On different territories, feeding was started on dates between 23 and 30 April, depending on when

148

Breeding season

signs of occupation were first found, and continued until a few days after clutch completion, about the end of May in the latest nest. The experiment was done in four different years, and in each year the performance of fed pairs was compared with that of unfed pairs in the same forest. As far as possible, pairs on different territories were fed each year, giving a total of 13 fed and 22 unfed pairs for comparison (Newton & Marquiss 1982). Effects were found on body weights, laying dates and clutch sizes (Table 24). Three fed hens, trapped in the pre-egg stage after feeding was begun, were all heavier than the average at this stage in Ae Forest, and one at 386 g was the heaviest yearling, and the second heaviest Sparrowhawk, that we had ever handled. On average, fed birds laid about five days earlier than unfed ones, and produced clutches about one egg larger. Non-laying (after having built a nest) was frequent in Ae Forest, and on average affected about 27% of pairs. But none of the fed pairs was in this category, as all produced at least four eggs. One even laid seven eggs, the only clutch of this size recorded in Ae Forest in a ten year period. The yearling hen laid five eggs, which was again an unusually large clutch for a bird of this age in such poor habitat. Overall, the increase in clutch size was greater than expected solely from the advance in laying dates (for earlier clutches were normally larger), indicating that extra food affected egg number independently of laying date. Although feeding was stopped after clutch completion, we expected that fed birds (especially considering their greater weights) would have been better able to resist temporary shortages during incubation, and thus more often to have hatched and fledged young than unfed birds. They showed such a tendency, but it was slight and not statistically significant. Evidently, good feeding in the early part of the cycle carried only limited benefit in the later stages. Another experiment, done for different reasons, had further bearing on the effect of food supply on clutch size. On three occasions we removed the cock from a territory a few hours after the hen had laid her first egg. This was in prey-rich habitat where experience had led us to expect clutches of 4-6 eggs. Yet in each case no further eggs were laid, although for several days the hens remained near their nest, and one called persistently for food. Perhaps the shock of losing their mates stopped them from laying, but such a prompt response would also be expected if the hens were relying day to day on food from the cock; it would not be expected if the hen was committed to laying a certain egg number by her initial body condition. These findings were therefore consistent with the view that food consumption during the laying period (as well as before it) influenced egg number. A second experiment involved the provision of extra food in the nestling

period to find whether this would reduce the weight loss in the hens, which normally occurs at this stage. Bird carcasses were placed on four nests each day through the nestling period, and near the end each hen was trapped and weighed for comparison with other hens which had not been given extra food. The four hens concerned were well above average weight, and two

Breeding season

149

were the heaviest of 64 individuals examined at this stage. The young also grew well and in each nest all survived. Thus we concluded that food supply in the nestling period influenced the hen's weight, and that the usual weight loss in hens at this stage resulted partly from them feeding their young in preference to themselves. If food supply was maintained, the hens lost less weight.

SUMMARY

In south Scotland, the breeding of the Sparrowhawk population each year coincided with the period that fledgling songbirds were available. The first hawks laid within 5-10 days after fledglings first appeared in the environment and in the hawk diet. The appearance of songbird fledglings ended the season of lowest food supply for Sparrowhawks, and fledglings probably provided the extra food needed by female hawks to increase their weight and produce eggs. Moreover, the period of peak food demand, when most hawks had young, coincided with the period of peak fledgling supply, mid June to mid July. Sparrowhawk clutches were started during the period late April to early June, but mostly in the first half of May. On average, egg laying was earlier in warm and dry springs, when fledglings were available earlier, than in cold and wet ones. Breeding hens needed extra food not only to form eggs, but also to lay down body reserves, chiefly fat; hens which did not accumulate such reserves did not lay. The ability of the hen to achieve good body condition probably influenced the date that she began to lay, and her ability to maintain high weight probably influenced the number of eggs laid, and whether or not she deserted the clutch. Thus the heavier the hen in the laying period, the larger the clutch and the less the chance of desertion. For much of the breeding cycle, the weight of the hen depended on the ability of the cock as a food provider. It was therefore the cock who largely determined the fate of the breeding attempt. Hens which were given extra food experimentally in the pre-lay and laying periods laid earlier and larger clutches than did control hens which had only natural food. Other hens, which were given extra food in the nestling period, were heavier at that stage than were control hens, which had natural food alone. Breeding at all stages was thus dependent on the food supply and body condition of the hen.

CHAPTER 11

T he breeding cycle: early stages

R ep roduct ion is th e mo st significa n t eve n t in th e life of a ny a nimal, because on ly throug h this p ro cess ca n th e a nimal pe rpetua te its genes in fu tu re ge nera tio ns . A Spa r row ha wk's suc cess in br eeding d ep ends la rgely on its choice of nesting pl ace a nd mate, its skill in nest construction , and its ability to provid e food a nd parental ca re . In this a nd su bseq ue n t chap te rs , I shall d escri be th e va rio us stages of br eeding, from pai r formati on to fledgin g of you ng . T h e ea rl y s tages a re th e hardest to stu d y, a nd the acco un t below has been pieced together mainly fro m incid ental obse rva tions . For me, the first sign of im pe nding breed ing a mo ng Spa r row ha wks eac h sp ring wa s the sudd en ap pe a ra nce of fres h white d roppings on th e dead bro wn of th e woodl and floor in th e vicin ity of the nesting pla ces. Su ch pla ces had no t necessa ril y been deserted ove r winte r, but ge nerally the bird s spe nt so littl e time ther e that they left few, ifany, sig ns of th eir p resence. As spring a p p roac hed , a nd th e birds began to prep are for br eeding, th eir d roppings and plu ckings became incr easingl y evide n t a t nesting places. Som etime after th ese sig ns first a ppeared , nests were st arted . O ccasion a l pa irs began building from Febru ary, bu t th e majority not until March or Ap ril, espe cia lly in th e hills. During M arch a nd A pri l, th e ha wks themsel ves a lso became more eviden t, displa yin g in flight a bove th e nesting woo d s. H owever , watching for ae ria l acti vities was a time-con su ming a nd not wholly reliable way oflocating breeding pairs. This was becau se Sp arrowhawks occasiona lly di spl ayed over plac es

The breeding cycle: early stages 151 where they did not nest, and apparently sometimes went for days without displaying over places where they did nest. Also, the amount of aerial activity varied greatly from pair to pair.

Preparation for breeding It will help to state the main points at the outset, and give the detail later. To breed, a Sparrowhawk needs a nesting place and a mate. Nesting places in good home ranges provide the main limiting resource for males, and competent males form the main limiting resource for females. This is because females are surplus to males in the population at large, making established males a scarce commodity for females in this primarily monogamous species. In both sexes, the acquisition of nesting places and mates is in turn dependent on individual foraging success. To retain a female, the male has to feed her, in addition to himself, otherwise she will wander away and attempt to settle elsewhere. The female has to obtain food readily, with some help from the male, to enable her to spend as much time as possible on the nesting place itself. Only in this way can she maintain ownership, and prevent her position being usurped by another female. Moreover, as nesting places and mates vary in quality - that is, in the chance they offer for successful breeding - both invoke heavy competition. Little wonder that early spring is the peak period for territorial defence and aggression among Sparrowhawks, and that many would-be breeding attempts fail at an early stage. Each year in the study areas, birds were present at some nesting places in March-April, as evident from fresh droppings or prey remains, but apparently not thereafter, and no nest was built. At other places nests were partly built and then deserted before laying (Chapter 6). In both instances we sometimes trapped birds in April, occasionally both members of a pair. Some such birds were not seen again that year, but certain females were later found breeding at other nesting places nearby. It seemed, therefore, that some places were occupied in early spring by birds which did not reach the egg laying stage, and remained non-breeders that year, while other sites were occupied temporarily by birds which later moved elsewhere to breed. The use of radio-transmitters on individual birds helped to show what was happening, and gave further insight into pair formation. Pairformation As indicated above, the success of a resident cock in attracting a hen in spring centred largely on his ability to obtain food. This depended partly on the quality of his home range and partly on his hunting skill. If he could catch prey easily, he had more time to spend at the nesting place, and could also afford to feed anyone female which lingered there. Of six radio-tagged cocks which were studied at this stage, four spent most of their time hunting and very little at the nesting place. These four evidently had difficulty in getting sufficient food. They formed no more than temporary associations with females, and none bred that year. The other two cocks spent less time

152

The breeding cycle: early stages

hunting, yet were well fed, and more time at the nesting place. Both formed stable pair bonds and subsequently bred. Hens in early spring behaved differently. To begin with, they had much larger and more overlapping ranges than cocks. The range of anyone hen overlapped the ranges of several cocks, and the nesting place of anyone cock lay within the ranges of several hens. Those radio-tagged hens which we studied each visited the nesting places ofseveral cocks, and most eventually settled on one for breeding. One hen spent more than a week at one nesting place and then shifted suddenly to another where she subsequently bred. Other hens, although mainly resident at one nesting place, made occasional visits to others. Yet other hens continued to wander from place to place for much of April, and then failed to breed at all that year. One hen was trapped twice in three days at nesting places 15 km apart, but a move of this distance was unusual. The extent to which a hen remained at one nesting place depended at least partly on whether she could obtain enough food there, either from her own efforts or, additionally, from the cock. Hens in good habitat spent more time at the nesting place in the pre-lay period than did hens in poor habitat. Moreover, when we provided food (in the form of pigeon carcasses) for hens in poor habitat, they became as sedentary as hens in good habitat, and hardly ever left the immediate nesting place (Chapter 5). We could thus infer that any inability of the cock to provide food resulted in the hen wandering, and in some cases shifting for that season to the nesting place of another cock, or not breeding at all. An important activity of the 'settled' hen was defence of the site against rivals, and any such aggression by settled birds would have been another factor keeping unpaired hens continually on the move until they found an unmated male or gave up attempts to breed that year. Given this system of pair formation, and a surplus of potential breeding hens, it is little wonder that the literature contains so many records of hens shot at nesting places being quickly replaced by others (e.g. Owen 1916). We ourselves removed what we presumed to be the resident hen from one nesting place, and then over the next four days trapped another three hens there, releasing each in turn. Without removing the first hen, it was seldom possible to catch a second at the same place. However, we three times trapped two hens at a place which was subsequently occupied by only one of them. The implication was that the presence of one hen at a nesting place normally kept out others, or at least prevented them from spending any time there. Pair formation in Sparrowhawks thus seemed to involve: (I) a contraction in the range of the hen, from overlapping the ranges of several cocks to spending most of her time at the nesting place of one cock, but occasionally visiting those of other cocks; (2) the provision of food by the resident cock, allowing the hen to spend more time at his nesting place and to ward off other visiting females. By remaining at the nesting place in the pre-lay period, the hen could also obtain any food provided by the cock, for it was only to the nesting place that the cock brought prey.

The breeding cycle: early stages

153

By this simple response of staying when fed and leaving when hungry, the hen is at once assessing the quality of both the cock and his range, and is reacting accordingly. There is little point in a hen remaining in a place where she is short of food, either through an inadequate cock or an inadequate hunting area. The prey supply is crucial because the hen is dependent on food from the cock from before the eggs are laid until the chicks are at least half grown. By defending the nesting place, therefore, the female is in effect protecting the food supply which will enable her to reproduce. In choice of mate and nesting place, some hens may be compromised by competition from other hens; and similarly the cock could in theory lose a hen to a cock on another nesting place if the second cock was a superior food provider. I do not know whether the cock exercises any choice in mate, but he could presumably withhold food from any female which he found unacceptable. In the population at large, yearling-adult pairs were much less frequent than would be expected ifmating were entirely random between these age-groups, while yearling-yearling and adult-adult pairs were more frequent (Chapter 18). This implied some selection of mate, but whether this involved food or some other mechanism I could not discover. The impression we gained in south Scotland was that cocks spent time at their nesting places and, as a matter of course, were visited by hens. Perhaps, however, cocks were able to take more active steps to attract a mate. I heard of one instance where the female was shot from a nest, and the cock then spent long periods in the next few days circling over the nesting place, often with prey held conspicuously in its talons. This sort of behaviour would probably soon attract a female, but may not be widely used. While monogamy was apparently usual in the Sparrowhawks we studied, about 1% of nests in south Scotland were occupied simultaneously by two laying hens which were not related to one another (see later). This may have occurred when a male was able to feed more than one, and when two equally matched hens occupied a nesting place, so that neither could displace the other. Most such instances occurred in one year, after a hard winter, when males may have been scarcer than usual, and competition among females especially strong. In addition, some cocks may have attended more than one hen at different nesting places. None of 15 males tracked by radio did so, but of 240 occasions when cocks were caught at nesting places, five had previously been caught at nearby places that year. No other cock was caught at these places, but this of course provided no more than a suspicion of bigamy.

Defence ofthe nesting place In early spring, the cock is normally established in a home range well before breeding begins, and in most cases he will have been there from the previous year. His most frequent territorial contacts will be with neighbouring cocks, whose ranges will also be familiar to him, and who will be concerned with their own nesting activities at this time. Other, mostly young, males in the general area tend to avoid the ranges of established males, and do

154 The breeding cycle: early stages not advertise themselves. For these reasons, defence by the established cock is probably only slightly more necessary in spring than at other seasons. For the hen, it is a different matter, and once she has settled on a nesting place, defence against other hens must become paramount. With more hens than cocks among potential breeders, competition is intense. Any intruders which are detected at the nesting place are chased vigorously, and any which pass overhead may elicit vigorous threat in the form of 'slow flighting' and 'undulating' displays (Jones 1974). In slow flighting, a bird flies at a constant level to and fro above the trees, on slow exaggerated wing beats. The tail and flight feathers are not separated, but at close quarters, the white undertail coverts can be seen to be fanned at each side. The display may stop at this, and the bird may start to soar or disappear into the trees. At other times display is intensified, and the bird begins to dip up and down in flight, gently at first, then with increasingly marked undulations. From slow flighting, the bird will suddenly close its wings for 5-10 m; it will then spread its wings and tail for a split second to check the dive and swing back upwards at high speed, almost vertically, with wings and tail closed. As upward momentum is lost, the bird tilts gently forward and the slow wing beating is resumed until the next dive, so that in the airspace the bird follows an undulating path, gradually losing height. The display may end with a spectacular stoop into the trees. Undulating displays occur in other raptors too, and in similar context (Newton 1979, Cramp & Simmons 1980).

Some such displays may be spontaneous, but most occur when there is another Sparrowhawk in the air nearby. Usually one bird leaves on seeing the slow flighting or undulating displays of the other, but on two observed occasions when neither left, one bird stooped at the other, and in one instance the two birds locked talons and fell through the air for several metres before one bird left. Such escalated fights are clearly rare, however, and most contests are settled at an earlier stage by display alone. After any such encounter, the remaining bird stays in the air, slow flighting for a while, turning its

The breeding cycle: early stages

155

head from side to side, apparently scanning the surrounding area. Most such incidents last no more than a few minutes and so are easily missed. They seem to represent defence of a nesting place by the occupant, and in one instance in which the occupant was radio-marked, this was the case. But with unmarked birds it is seldom possible to be sure which bird (if any) is occupant and which intruder. In my experience, each bird defends only against others of the same sex. Such aerial display, with its eye-catching manoeuvres, makes a normally secretive bird extremely conspicuous. It is the most effective way in which a Sparrowhawk can reveal its presence. While in cover, the bird remains unseen, but once in the air it is visible to other Sparrowhawks over a wide area. Display by the defender signals occupation, while display by an intruder provides a way of testing the situation, of inviting a challenge. If the area is vacant, there will be no response to the challenge, so a Sparrowhawk can find a gap in the territorial system. Some other territorial birds behave in a similar manner. Aerial displays can be seen in any month from October to June, but are most frequent in March and April, as birds prepare for breeding. They may be given by cock or hen to others of their sex, but most displays which I have seen were from one hen to another. This was probably because females were surplus to males, and some moved from one male range to another before eventually settling. Owen (1916) wrote that such territorial displays were accompanied by a 'sibilant whistling note', but I have never heard this. Other aerial activities consist merely of soaring above the canopy, a few brief wing flaps alternating with long periods on still wings and spread tail, as the bird climbs in tight spirals, sometimes to the limits of human vision. The undertail coverts may be spread to form a conspicuous white patch below the tail, suggesting that flights also have an advertising function. They are performed by either sex, in my experience again more by females than by males, but occasionally by both sexes together. They may be seen at any time of day, but more especially in the morning, on bright breezy days. They may end with the bird stooping abruptly, sometimes into the nest wood, or leaving the area on a long glide. I can think of no obvious function for such flights, except to signal the occupation of a nesting place. In some other accipiters, pairs on territory call frequently and their sounds carry long distances. This may provide another means of territory proclamation, equivalent to singing in passerines. The situation is not strictly parallel, however, for whereas passerines sing most at the time of settlement, accipiters are noisiest at a later stage, near the time of egg lay. Sparrowhawks also are much less vociferous than their relatives, and their calls do not carry beyond the immediate nesting areas. They are unlikely to be more than a short-distance deterrent. Benefits and costs ofterritorial behaviour

I t is worth reviewing at this point the benefits and costs of territorial

156 The breeding cycle: early stages behaviour in the Sparrowhawk. Lying within the home range of an established male, the nesting place is not a clearly defined area with sharp boundaries. But it is where the birds show two important aspects of defence behaviour, namely strong site attachment and strong aggression. Other Sparrowhawks are likely to be attacked there, the likelihood decreasing with distance, and varying from one part of the breeding cycle to another. The increased aggression and display shown by Sparrowhawks in spring are concerned primarily with the acquisition of a nesting place and mate, rather than with a foraging area, which the birds already have. The vicinity of the nest area may, however, provide an untapped food source for the female at the establishment stage, and later when the chicks are small, these being the only times when the female hunts near the nest. The nesting place also provides a meeting point for the pair, where they can be sure of making contact, and where the male can deliver food. This is an important aspect in otherwise solitary birds, which forage independently over wide areas. The main costs of territorial behaviour are the time and energy used in threatening and attacking other birds, and the loss of feeding time while on guard. In spring, these costs are borne mainly by the female, at a period when other demands on her are much reduced. Without the food provided by the male, however, the female would have difficulty in sustaining her vigil. In this sense, both sexes contribute to defence at this time, but the female more actively. As the main need is to be present on the territory, probably less than an hour each day is spent in display and other active defence. Thus, the costs of being territorial do not seem great, especially as they precede egg laying and other demands of breeding. Yet they are costs which many Sparrowhawks cannot sustain, and these birds do not breed that year. Territorial behaviour, with its clear net benefits to the individual in terms of reproduction, has important consequences at the population level, namely the spacing out of breeding pairs and the limitation of their density, as discussed in Chapter 4.

Interactions between mates Pairs which progress to egg laying show six phases in behaviour before the eggs appear, namely: (1) the attraction of birds to nesting places and of potential mates to one another; (2) mutual roosting on the nesting area, mutual calling in the early mornings, and aerial display; (3) the feeding of the hen by the cock; (4) nest site inspections and stick carrying; (5) nest building proper; and (6) copulation. These activities appear roughly in the sequence listed, though some may start almost simultaneously and, once started, most continue until a later stage in the breeding cycle. Pre-laying behaviour thus consists of a succession of stages, involving the progressive incorporation of additional activities. Some pairs pass through particular stages more rapidly than others, and pairs may break up and fail to breed at any stage, especially if food is scarce. Occasional pairs are so much later than others that they can be seen engaging in preliminary

The breeding cycle: early stages

157

breeding activities, such as mutual calling and stick carrying, in mid-summer, when successful pairs are already feeding young. Such late and temporary associations do not normally lead to nest building. Most activities involving the pair take place in the early mornings, perhaps partly because dawn is a time when both sexes are together at the nesting place, having roosted there the previous night. When the partners first begin to roost together, they use separate trees, but call to one another at dawn and dusk. Both give the 'kek ... kek ... kek' call, the male at once recognisable by his higher pitched voice. The female occasionally gives a wailing 'kee' or 'kee-oo' call, which I came to associate with food begging. In their various aerial activities at whatever time of day, the partners may circle above the nest area for a while, first one and then the other uppermost, and then stoop headlong with closed wings one after the other into the canopy. The male may chase the female fast among the trees, or repeatedly dive at her as she sits on a tree-top. There is at least one record of a male stooping down to the airborne female, who flipped over and the two birds grasped talons and cartwheeled down to the trees (Dawson 1978). Holstein (1950) said that such aerial displays are often accompanied by a call which he described as 'kirr ... kirr ... ', likening it to the cry of a tern, but I have never heard this. Such mutual activities are usually termed 'courtship displays', but this puts an interpretation on them which may not be justified. All one can say is that they involve two birds of opposite sex above a nesting place and, although such displays must be noticeable to other Sparrowhawks, they are not obviously aggressive, like the territorial displays described above. Some pairs seem to display in the air much more than others, and it is possible that others never do so. To feed the hen, the cock brings prey to the nest vicinity, one item at a time. Usually he has already plucked the prey, but otherwise he may pluck it in full view of the hen, with exaggerated movements. The hen takes the item from him, either on a perch or in the air, in a foot-to-foot pass. Typically, the cock brings the prey to a bare branch below the canopy; he holds the prey under one foot, and slightly away from his body, not mantling it, while he calls to the hen, a soft 'kew ... kew ... kew ... '. The hen flies towards him, aiming straight for the food and as she reaches it, the cock flies away leaving the prey, which a split second later the hen has secured in her feet. Throughout the transfer, the white undertail coverts of the cock may be fluffed out, and the hen may give the food begging call, described above. Alternatively, the cock may fly to the nest area, and hover momentarily a few metres above ground, holding the prey in one foot stretched down to its full extent. The hen then dives beneath him, turns up and grabs the food in her feet as she passes, effecting the whole transfer in a split second. Thirdly, the cock may fly very slowly below the canopy, calling, with the prey dangling, and the hen flying fast merely intercepts him, again seizing the food as she passes. She flies to a perch to eat the item. This behaviour is commonly called 'courtship feeding', but it continues for much of the breeding cycle, and well beyond the 'courtship' phase. It is initially important

158

The breeding cycle: early stages

in enab ling the h en to acc um ula te th e bod y reserv es necessary for egg productio n. Some d ays before egg laying, th e hen apparentl y becom es wholl y d epend en t on th e cock for food , if she is not already, and may th en receive several item s each day.

-

Th efo od pass

~-

The breeding cycle: early stages 159 Nests and nestbuilding The main function of the nest is to provide a protected platform for the eggs, chicks and brooding adult throughout the long breeding cycle. The placement of the nest high in a tree precludes access by ground dwelling mammals, and its hidden position in the lower canopy gives some further protection against arboreal mammals and birds, as well as shelter from weather. The selection of a site, and the construction itself, are thus important activities on which the success of breeding may depend. As discussed in Chapter 3, Sparrowhawks normally build a new nest in a fresh position each year. Most of those in conifers are on the south side of the tree, where the branches are stronger, providing a firmer base below and better protection above. Nest building may begin with 'inspections', with the male flying from one potential site to another, including any old nests in the vicinity. He does this in full view of the female and with undertail coverts fanned. The hen may watch from a nearby tree, or follow him to each site, where both perch for a while, looking intently. This may be followed by twig-carrying behaviour, in which either bird may break off a twig and carry it around for some time before dropping it. The next stage occurs when twigs are carried to specific sites, where some merely fall to the ground, thus eliminating certain sites as unsuitable. Sometimes a few twigs may be put on each of several good sites, including old nests, before the birds begin to concentrate on one site, signifying the start of building proper. In rare instances the birds may almost complete a nest and then start again in a different, betterprotected, site. The impression to the human observer is that various sites are tried, and great care is taken over the final choice. The completed Sparrowhawk nest is a fairly flat, bulky structure, with a deep cup. The nest is usually about 4Q-SOcm across, IQ-30cm deep, and the cup is about 15-20cm across and 5-10cm deep. The whole structure weighs around 1-2 kg. From above, the nest may appear round or oval, bean-shaped if pressed close to the trunk of a conifer, or slightly hour-glass shaped if wedged in the fork of a broad-leaved tree. The nest is made of twigs, some of which may encircle the trunk, holding the nest firmly in position. The central cup serves to keep the eggs together, making them easier to incubate and keep warm. It also keeps them partly hidden and reduces heat loss when the female is off the nest. The Sparrowhawk nest is distinguished from other raptor nests by its size, its light airy structure, and by the virtual absence of green sprays, which are a regular feature of many other raptor nests. While I occasionally found one or two leafy twigs on Sparrowhawk nests, at least some had probably been picked bare, and burst into leaf later. However, in areas where lichens occur on trees, these plants may cover the twigs used by Sparrowhawks, and the whole nest has a greyish-green aspect, which adds camouflage, especially in larch or broad-leaved trees. Where building is spread over several weeks, it takes place in spurts, during mild weather. A few days (mainly early mornings) of intense activity may

160

The breeding cycle: early stages

be followed by days or weeks in which little is added until the next bout. Long thick twigs (up to 60 em long) are first arranged to form a rough outer circle; on these are placed progressively thinner and shorter twigs until the nest is complete. The cup itself is lined with fine twigs, with the addition in most nests of bark flakes. These are not normally added until the eggs are due, beginning a day or two before the first egg and continuing during the laying period into incubation. The birds collect the twigs one at a time from the ends of branches , breaking off small twigs with the bill and gripping large ones in the feet, then hanging and flapping vigorously until they break. The birds prefer twigs which are recently dead, but still sufficiently flexible to position without snapping. They seldom collect twigs from the ground, probably because fallen twigs are more decayed and liable to break. Both sexes build, but the contribution of cock and hen varies between pairs. The cock seems to bring much of the material to begin with, and the hen brings most later on. The material is usually obtained nearby, but larch twigs are so favoured that they are often carried 70 m or more to nests in other trees. These twigs are ideal, because they are easily broken, enabling larger twigs to be used for the nest base, and they also have small nodules along their lengths which stop the twigs slipping against one another and help to bind the structure together. Pine twigs are commonly used but are less easily broken and less flexible. Spruce twigs are too flexible, enabling only small ones to be used, and they do not interlock. Several spruce twig nests in our areas disintegrated before the young could fly, and this happened frequently in the nearby Kielder forest (Petty 1979). Such twigs were used only when nothing else was available in the vicinity. In one young pine wood, where only live twigs were available which the birds could not break, they built a tiny nest of thin heather stems picked from the ground; this nest also collapsed, causing the young to die. Among broad-leaved trees, birch seemed the most popular building material, again probably because freshly dead twigs were easily broken. Birch twigs were occasionally used to line nests built of other materials. Bark flakes used for nest lining are picked directly from the trunks of pine or larch trees. Some are taken from the nest tree itself, and just above the nest the trunk often appears to have been stripped of all loose flakes. The individual flakes are about 2 ern square, and number only one or two to begin with, but accumulate to 20 or 30, occasionally up to 60, by the end of incubation. These flakes prevent the eggs from slipping down into the nest structure, and reduce the airflow from eggs to nest, conserving warmth. The female pays great attention to positioning the bark neatly under the eggs, and if you upset the arrangement she will instantly re-do it. After hatch, the chips tend to work their way down through the nest as the young move around. If pine and larch are not available, oak or birch bark or bits of dead wood may be used instead, and one nest in south Scotland had been lined with about 30 small flat stones from a nearby track. To many nests, however, particularly those in pure spruce stands, no bark or other material is added, and the eggs rest on fine twigs. Owen (1916) wrote that

9 Portraits at a pool. (Upper) Two adult males (Lower) Two females, first-year on left , adult on right. Photos : R.J.C. Blewitt.

10 Related species . The Goshawk is a larger version of the Sparrowhawk and is also its main predator. (Upper) Female with young. Exterminated from Britain during the last century, the Goshawk has recently re-established itself, as a result of escapes of falconry birds, imported from Europe . This female, at a nest in the English Midlands, carries the remains of a leather jess on her left leg, as testimony to her earlier life in captivity. (Lower) Nesting habitat. Compared to the Sparrowhawk, the Goshawk nests in older, more open woodland and builds a large nest, which is often very conspicuous . Photos: R.J.C. Blewitt.

Related spec ies. (Up per) Shikra at nest in India. (Lower ) Levant Sparrowhawks migrating over Eilat in Israel. These two species are similar in size to the Sparrowhawk, but eat mainly reptiles. in addition to birds and mammals. Photos: (upp er) R. Naoroji; (lower) G.A. Vaucher.

11

~

-~

.,

~

"'"

-, -I

.4

.,

.,

-

~

-t

.....

r:

-I

~

12 Studies of food. (Upper) Cock Sparrowhawk at plucking post. (Lower) Prey feathers on plucking post. Photos : (upper) R.J.C. Blewitt; (lower) I. Newton.

13 Studies of food. (Upper) Inspecting a nest. (Lower) Prey remains obtained from nest in post-fledging period. They include Black-headed Gull, Homing Pigeon with ring, two Jackdaws, Jay, several thrushes, pipits, Starlings and a Crossbill. Photos: (upper) W.M. Hotson; (lower) I. Newton.

14 (Upper) Nest in pine tree, well protected in the canopy. (Lower) Eggs with shells thinned by residues of the pesticide DDT often break and produce no young . Photos: (upper) D.A. Ratcliffe; (lower) I. Newton.

15 (Upp er) Single eggs from 12 different clutches to show some of the variation in size, shape and markings. (Lower) Regurgitated pellets of undigested prey rema ins. Unlike owls, diurnal raptors digest bone , so their pellets usually conta in no skulls or other bone material, and are of very limited use in studies of food . Photos: N.J. Westwood .

)

16 (Upper) The cock does not incubate, but is seen here guarding the eggs, while the hen eats the food he has brought for her nearby. (Lower) Hen about to incubate. Photos: R.J.C. Blewitt.

17 (Upper) Hen brooding small chicks. (Lower) Hen feeding small chicks. Photos : (upper) G. Yates; (lower) J.F. Young.

18 Females with chicks. The upper nest was in a typical site, high in a pine, but the lower was in a gorse bush, only one metre from the ground in a narrow scrub-filled gulley in open moorland . Photos: (upper) N.J. Westwood; (lower) G. Yates.

19 Feeding young is not always easy (upper), but the hen swallows the feet and other parts which the young cannot deal with (lower). Photos: (upper) J.F. Young; (lower) R.J.G. Blewitt.

20 Hen about to feed chicks (upper) and about to brood them (lower). Note the clenched foot . a tactic which the hen adopts to prevent damage whenever she puts her feet among eggs or chicks. Photos: (upper) N.J. Westwood; (lower) WAlston.

21

Females with large young . Photos : (up p er) R.J.C. Blewitt; (lower) G. Yates.

22 The nest in rain. When the chicks are small the hen can cover them all (upper) , but when they are large she can cove r only one or two at a time (lower). Photos : (up p er) R. Brewer; (lower) R. T. Smith.

23 Fledglings about to leave the nest (upper) and after first flight (lower) . Photos: (upper) R.J. Westwood; (lower) WAlston.

24 (Upper) In the absence of the Goshawk, the Tawny Owl was the main avian predator of nestling Sparrowhawks in the south Scotland study areas. (Lower) Among adults, natural mortality is seldom recorded, because most birds which die are quickly removed by scavengers. However, a few birds were found dead in the woods of the study areas, including this ringed female. Photos: (upper) R. T. Smith; (lower) I. Newton.

The breeding cycle: early stages

161

he found some nests in Essex which had been lined with dried leaves, but I have never seen this. The process of nest building involves little more than the positioning of each twig with the bill, and tucking the ends of twigs between existing ones. I t varies according to whether the bird is starting afresh or adding to an existing structure. A dilapidated nest may be built up considerably, first with large sticks and then with small ones. But to a nest in good condition little may be added, apart from a few surface twigs and a new lining. Yet, if a nest is blown down during building, the birds will start again in the same place, and thus bring much more material than is needed for a single nest. Such events imply that nest building is not an autonomous process which follows a fixed pattern from start to finish, but that feedback from the structure already built influences the subsequent course of the birds' behaviour. No doubt stimuli from the completed nest, perhaps detected by the hen through her developing brood patch, serve to bring building to an end. I t may then bea matter of days or weeks before the first egg is laid. As with most other aspects of parental care, the effort expended in nest building involves a cost to the adults, offset by the benefits to be gained by successful reproduction. Nests started afresh contain, when completed, anything from about 700 to 1,500 twigs, and, as each twig is carried individually, building absorbs considerable time and energy. Owen (1916) observed that, in bouts of continuous building, the birds took 4-5 minutes to collect and place each twig. On this basis, a large nest would take more than 100 hours to complete. Nevertheless, Sparrowhawks can construct a sizeable structure in a few days, as happens occasionally when nests are destroyed shortly before laying. At the other extreme, the addition of a small number of twigs to an existing structure could be completed within a few hours. In general, however, nest building represents a considerable investment, and it is not surprising that it is normally spread over several weeks. Started well before egg-laying, moreover, the energy costs of building are normally well separated in time from those of egg production. No-one has yet measured the insulative properties of Sparrowhawk nests, but large bulky nests are presumably better in this respect than are small shallow ones. Thus high investment at nest building is likely to lessen the subsequent thermoregulatory costs to the sitting bird. Large structures also provide a firmer platform for the growing young, and are less likely than small ones to fall apart before the young have flown. In south Scotland, more substantial nests were generally built by adult pairs than by young ones, and in early springs than in late ones, when the time available for the nest building was reduced. Pairs which began building earliest in any one year were also among the first to produce eggs, but in the population as a whole, nests were started each year over a longer period (about 13 weeks) than clutches were started (6 weeks) . Sparrowhawks continue to add sticks to their nest during the incubation and nestling period. And as the young stand and walk around the nest, the whole structure becomes flattened, providing a broad platform on which

162 The breeding cycle: early stages the young can feed and exercise their wings. The shape of the nest thus changes during the breeding cycle, to fit its changing function.

Copulation and egglaying The cock Spaxrowhawk often seems reluctant to approach the hen closely, perhaps because he could so easily be killed by his larger mate. This reluctance is apparent during food passes, especially early in the season, and during copulation, the only action in which the pair achieve close body contact. Mating seems to be initiated by the hen, who crouches and calls to the cock, a 'pee-co' note similar to food-begging. With body feathers sleek, she fluffs out the feathers of her lower abdomen and flanks, and also her white undertail coverts and neck feathers, thereby exposing her white nape spot and presenting an unambiguous signal to her mate. She is usually sideways to the male or facing away. She leans forward at an angle of about 30 degrees, with wings slightly open and tail moved slightly to one side. As the cock settles on her back, he clenches his feet, thereby avoiding hurting the hen, and rests on his tarsi, forcing his tail down beside hers to affect cloacal contact. The process takes only 5 to 10 seconds, after which the male departs promptly to another perch, while the female fluffs and shakes her feathers. The female may solicit for some days before the male responds by mounting. The initial attempts often fail, but the number and frequency of successful matings increase as the date of laying approaches. They can occur several times each day, mainly in the early mornings, for two or more weeks before egg laying, and continue until around the time the clutch is complete. Mating often follows a food pass, the female soliciting while she crouches over the food that the male has just brought. This may help to link a crucial aspect of breeding with food provision, and thus reduce the chance of the female being fertilised by a non-committed male. Copulations may also follow certain other activities between the pair, including nest building, chases or diving displays. They mostly take place on the nest, or on a thick branch nearby or on a low stump, but one observer saw Sparrowhawks mating on a hedge outside the nest wood. As the cock has to leave the nesting place for part of each day in order to hunt, the hen could be visited by other cocks. At some nesting places, in the pre-egg stage, we trapped neighbouring cocks and also recorded intrusions by radio-tagged birds with no nests of their own. It is perhaps to reduce the chance of cuckoldry that the cock spends more time near the nest in the hen's fertile period than at any other time. In addition, the relatively high frequency with which copulations take place then might help to lessen the impact of any 'non-marital' mating which might have occurred (Birkhead et al 1985). From a few days before the eggs are due, the hens remain constantly near their nests, moving only when disturbed. Signals from our radio-marked hens pulsed monotonously from the same spot for hour after hour. The hens are heavy and lethargic then, and only move short distances with laboured flight. Almost certainly, they do not normally hunt at this stage, though

The breeding cycle: early stages

163

occasionally we caught them in baited traps set near the nest. In this state, the hens would probably not possess the agility to catch free-flying prey, and might risk damaging the developing eggs within their bodies. There is thus a strong advantage in being inactive at this stage. The eggs are laid at any time of day, mostly before noon, at intervals of about two days, but occasionally three days or more. For the human observer, the laying of the first egg is the most specific point in the whole breeding cycle, and thus affords the most concise measure of the timing of reproduction for each pair.

Geographical variation I have described the events that lead to egg laying as they occur in south Scotland. In more favourable areas, further south, breeding partners may associate on nesting places at any time of year, and indulge in preliminary breeding activities, such as mutual calling, throughout the non-breeding season. Some may even do some stick carrying and nest building in the autumn, equivalent to the breeding activities occasionally seen in various other birds at this season. There is also some resumption of display in autumn, including slow flighting. Thus, events which are restricted to a briefpre-breeding period in northern regions, where food is scarce in winter, may extend over more of the year further south, giving less of a break between breeding seasons. Some variation in the timing of events occurs within regions, moreover, depending largely on the local food supply.

SUMMARY

For male Sparrowhawks, the limiting resource permitting reproduction is the nesting place in a home range offering abundant prey. For the female, the limiting resource is an established male, able to provide food. As females were surplus to males in south Scotland, they competed strongly, and once a female was settled on a nesting place with a male, her main initial function was to ward off other females. Special slow flighting and undulating displays were used as warnings to signal occupancy of a nesting place, and these occasionally escalated to chases and fights. Other activities between the pair at the pre-egg stage included aerial display, food provision by male to female, nest building and copulation. The nest of each pair was constructed in a site which provided a firm foundation and a degree of seclusion from predators. It was made of recently dead twigs, with a central cup, usually lined with bark flakes or other materials. Both sexes participated in building. The food provided by the male initially helped the female to stay on the nesting place, and defend it against other females. I t later helped the female to accumulate the body reserves necessary for egg production. Some days before laying began, the female became heavy and lethargic, and wholly dependent on the male for food.

CH AP TER 12

The breeding cycle: eggs and incubation

To ensure that the eggs hatch , the hen Sp a rro wh awk m ust con tro l th eir temperatu re a nd hum idity throu gh th e lon g inc ubation peri od , a nd pr ot ect th em against pred ation. The egg s form a n a tt ra ctive meal for man y a nimals, including mammal s suc h as Sq uir rels an d M artens, and bird s such as Jays a nd Crows . The position of the nest gives only partial protection, a nd th e surviva l of th e eggs d ep ends la rgely on th e con tin uin g p resen ce of th e fem al e to wa rd ofTpr edators. She, however , ca n remain with th e eggs o nly so lon g as the mal e provid es her with food .

EGGS

M ost Sparrowhawk eggs a re lon ger th an they a re broad , and have on e end mor e rounded than th e othe r. Som e are alik e a t both ends , however , whil e o the rs tend towards eithe r piriform or sphe rica l. They va ry gr eatly in size, ran gin g in len gth 33-4 7 mm (mean 40 mm ) a nd in breadth 28-36 mm (mean 32 mm ). When fres h , norm al eggs weigh 20-30 g, or on average ab out 26 g. This is eq uiva lent to abo ut 8 % of th e we ight of th e laying fem al e, so th at a clutch of six could weig h hal f as mu ch as th e hen herself. In eac h egg , th e co nte n ts com prise a bou t 26 % yolk a nd 74% a lb umen, o r in term s of che mica l compon ents, a bo u t 82 % water, 6% lipid , a nd 12% of ot he r

The breeding cycle: eggs and incubation

165

solids, mainly protein. These values are similar to those in other birds of prey. To many people, Sparrowhawk eggs are particularly attractive. They are basically pale blue or pale green, more or less covered with red-brown blotches and spots ofvarious shapes, sizes and intensities. The number and positioning of markings varies greatly, and no two eggs look exactly alike. At one extreme, the shell may be almost covered with pigment, rather like a falcon's egg, while at the other the shell may be wholly devoid of markings. On some eggs the blotches form a band around the middle or, alternatively, a cap at one or other end. On yet others, the marks appear skewed, all in the same direction, as though they were smudged as the egg was rotated in the bird's uterus, when pigment was applied. In addition to the red-brown marks, some eggs have paler lilac markings, often referred to as 'shell marks'. Such colour variations depend on the amount of pigment deposited, and its vertical distribution in the shell. Within most clutches, the markings become fewer and paler with successive eggs, as if the bird runs out ofpigment, and the last egg may be almost featureless. The last egg is also often the smallest, suggesting a shortage of other materials too. As incubation proceeds, the eggs lose their beauty. Their initial bloom goes quickly, the background colour changes first to a paler blue and then to white, and the red-brown markings become duller and browner. The shell absorbs oil from the brooding bird's feathers, and gradually takes on a slight gloss instead of the original matt texture. Any addled eggs which remain in the nestling period are usually quite polished, faded and dirty, poor relics of their former selves. Some egg collectors claim that they can distinguish the eggs of particular females. I t is certainly true that eggs from the same clutch resemble one another in size and shape, if not in colouration, more so than eggs from different clutches, and eggs from the same female in different years resemble one another more than do eggs from different females. But only in the case of females whose eggs depart considerably from the norm can one be reasonably certain of identification. The collection of Mr G. Tomkinson, now in the safe-keeping of the Wildfowl Trust, contains some beautiful series, spanning up to nine years, which are clearly from the same females, but even here a slight shift in egg type is apparent over the years, towards smaller and paler eggs with increasing age. The usual explanation for markings is that they break up the outline of the eggs and afford a degree of crypsis. In the hand, it is hard to believe that Sparrowhawk eggs are camouflaged, but in the nest, in the dappled sunlight of the treetops, it is a different matter. This was made clear to me whenever I viewed a nest from a distance, as for example when I was unable to climb the nest tree, and had to check the eggs from a nearby tree. In the flickering pattern of sunlight and shadow which falls on the nest surface, it is not always easy to see the eggs, let alone to count them with the naked eye. The spattering of white down which accumulates during incubation on the dark nest surface adds to the difficulty. In the natural

166 The breeding cycle: eggs and incubation setting, therefore, one is left in no doubt that Sparrowhawk eggs are camouflaged. For this purpose, the position and shapes of the blotches matter little, so it is not surprising that no two eggs look exactly alike. It is only the small species of accipiter, such as the Sparrowhawk, which have marked eggs. In the larger species, such as the Goshawk, the eggs are unmarked. This may be partly because the larger accipiters are better able to deter predators than are the small ones, but also because in the larger species the males can incubate, so that the eggs can be covered during the periods when the hen is off the nest, eating food the male has brought. In the Sparrowhawk the eggs are often exposed at such times, and depend for safety chiefly on their cryptic colouring.

LAYING DATES

Most Sparrowhawks in Britain start egg laying in May, but some begin in late April and others in early June. If the first clutch is taken, and a repeat is laid, this can be started as late as mid-June, at least in southern England. Beyond this date, Sparrowhawks would not have time to raise young before the autumn was upon them. During the 1970s, laying dates (of first eggs) were obtained for 12 different districts scattered between northern Scotland and southern England, and from earlier this century for a further district in England (Table 25). These various records revealed no clear geographical trends in mean laying dates within Britain, either from south to north or from west to east. Over a wider span of latitude, however, laying becomes progressively later further north, corresponding with the later spring. Regional bird books give the laying dates for southern Europe as mid-April to mid-May and for northern Europe as mid-May to mid-June (Chapter 10). Any such trends which might exist within Britain are evidently masked by the effects of more influential factors, such as elevation. The effects of elevation could be assessed within south Scotland, where nests were found each year from sea level to more than 330 m. Here there was a one-day delay in mean laying date for every 60 m rise in elevation, or about a five-day delay for the whole 330 m range, but the variation at anyone elevation was enormous. A difference of this order was probably applicable to other regions, and the districts with the latest laying dates in Table 25 were the highest in elevation. Within our south Scotland areas, mean laying dates also varied by up to ten days (5-15 May) between the warmest and the coldest spring experienced during 1971-84, and by up to five days according to female age, but these aspects are discussed in detail in Chapters IOand 18. As the eggs in a nest are laid on alternate days, clutches of 3, 4, 5 and 6 eggs usually take 5, 7, 9 and 11 days respectively to complete. If the first clutch is taken at this stage, the repeat clutch - if one is laid - is usually started on an old nest 12-15 days later, giving a total of up to 26 days between the start of first and second clutches. Among the many clutches

The breeding cycle: eggs and incubation

167

collected earlier this century by Mr G. Tomkinson in Worcestershire, the mean date for start of first clutches which were followed by repeats was 9 May and the mean date for repeats was 4 June (Table 2). Further south in England, a few records were obtained of third clutches being laid after two previous ones had been taken (Owen 1916).

CLUTCH SIZES

Most Sparrowhawks in Britain lay 3-6 eggs. Clutches of one and two eggs are rare, as are those of seven. The one-egg and most two-egg clutches tend not to be incubated. Most occur when the hen is unable, through food shortage, to produce a clutch of more normal size (Chapter 10). As in other birds, the mean clutch size in the population varies slightly between different regions, different habitats and different years, and according to female age. Clutches become larger from south to north in Europe. In Mediterranean regions, most clutches contain 2-4 eggs, but in northern boreal regions 5-6. Within Britain, any geographical gradients which might exist in mean clutch sizes are small and, as with laying dates, are overridden by the effects of other factors, such as elevation and habitat. In the different areas mentioned above, means varied between 4·0 and 5'0, but not in parallel with either latitude or longitude (Table 25). Within the south Scottish areas, mean clutch size declined by about 0·01 eggs for every 60 m rise in elevation, or by 0·05 eggs over the whole 330 m range from sea-level. This was very little, and was not significant statistically. As habitat varied with elevation, one could re-state this trend by saying that clutches were slightly larger, on average, in woods on low farmland (4'62) than in woods on upland. Within the upland they were smaller, on average, in large forest plantations (4'50) than in smaller woods on neighbouring sheepwalk (4'57). In south Scotland, the mean clutch size also varied between 4·5 and 4·9 in different years, dropping to 3·9 in 1979 and to 4·1 in 1982, after hard winters. Like laying dates, clutch sizes in a given year or habitat varied with female age, increasing up to age 3-4 years, and declining thereafter; and in addition they declined progressively the later in the season that laying began. These latter trends are discussed in greater detail in Chapters 17 and 18. In so far as any comparison is valid from studies in different areas at different times, no difference in clutch size was apparent between coniferous and broad-leaved habitats. Repeat clutches of Sparrowhawks are usually smaller than firsts, and the individual eggs are often smaller and less well marked. In Worcestershire, Tomkinson obtained both first and second clutches on 52 occasions during the years 1899-1918. On 40 occasions the repeat clutches were smaller than the first, on nine they were the same size, and on only three were they larger (Table 26). The mean size of first clutches was 5·3 eggs, and of repeats 4·2. In general, individuals which laid the largest first clutches also produced the largest repeats, though the variation was great. The records gave no

168 The breeding cycle: eggs and incubation indication of what proportion of birds laid repeats, but as egg collectors regarded them as normal, it was presumably the majority. However, Tomkinson in his area obtained no repeats from birds whose first clutch was started after 20 May, or contained fewer than four eggs. In south Scotland, I recorded only three naturally occurring repeats after the failure of first clutches, but, like an egg collector, was able to induce others by removing the first clutch within a few days after completion. Of twelve pairs robbed in this way, seven (58 % ) produced repeats, in five cases smaller than the first. In only five nests could success be followed, and four produced young. The scarcity of natural repeats could have been because many of the females which failed in south Scotland deserted their eggs after losing weight. If these birds failed at one attempt through food shortage, they were unlikely to be able to regain weight soon enough to try again. Others failed later in the cycle, by which date the season was too advanced for repeat laying to be worthwhile and their gonads would have regressed. In addition to repeat clutches, egg collectors also found that, if the eggs were removed one at a time as laid, some females would continue laying every second or third day until they had produced many more eggs in total than would normally comprise a clutch. Numbers up to ten eggs were common, and Walpole-Bond (1938) claims to have obtained no less than 23 eggs from one nest in this way. The usual procedure was to leave one egg throughout, in order to encourage the female to stay on the same nest. After a certain number, the eggs became progressively smaller and devoid of markings. We tried this progressive removal with several pairs in Northamptonshire, and found that, whereas some produced no more than the usual 3-6 eggs, others produced up to 14. In general, pairs which were in the best habitat, and began earliest in the year, produced the most eggs. Evidently, clutch size was not fixed at the start of laying, but in some individuals could be modified by events during the laying period. In biologicaljargon, Sparrowhawks were 'indeterminate' layers, as opposed to 'determinate' layers, such as pigeons, in which clutch size is fixed from the outset. In our study areas, the larger the clutch, on average, the more young it produced, not only to fledging age, but also to breeding age (Table 37). The individual young survived better in small clutches than in large ones, but this difference was not sufficient to offset their smaller number, so that large broods were the most productive overall. This was true for broods raised in the early, mid and late parts of the season, for those in both lowland and upland habitats, and for those reared by yearling and adult hens. In other words, in all these circumstances, the largest natural clutches were the most productive of future breeders, and should have been favoured by natural selection. Six was the usual maximum, and seven-egg clutches formed less than 1% of the total. None of the seven-egg clutches produced more than the sixes, so there seemed little advantage in birds laying more than six. Equally, it seemed worthwhile for birds to lay as many eggs, up to six, as they could manage to produce. I argued in Chapter 10 that food shortage was the main factor which prevented most females from achieving the opti-

The breeding cycle: eggs and incubation

169

mum number. It appeared, therefore, that while some females were unable to produce six eggs, others could produce many more than this, as indicated by the experiments reported above. The natural limit at 6-7 was presumably again set by natural selection, as clutches larger than this would have been wasteful, and no more productive of young. The main constraint was the amount of food needed to feed such large broods, but large clutches may also have been difficult for the female to cover effectively. Although I had no instance of an individual female laying more than seven eggs in a clutch, the Tomkinson collection has one set of eight eggs, which appear to be the product of the same hen, and Brian Etheridge found a similar clutch in Moray. However, seven (0.6 % ) nests in south Scotland contained 8-10 eggs which were of two distinct types. At each nest, we trapped or otherwise identified two different hens. In one case, two hens were put off the nest together, but in the others, the second hen went onto the nest while we were handling the first, having just trapped it. We were unable to find how many cocks were involved, but, because only one nest was used in each case, we suspected bigamy. None of these eggs produced young, as in each case some of the eggs were broken or pushed aside, and the remaining ones were then deserted. In Derbyshire, Ted Robson also trapped two hens on a nest containing eight eggs, the only instance amongst 106 nests he found during 1976-82. This nest failed to produce young. Similarly in Berkshire, Ted Green found that four (1 %) out of 239 clutches examined during 1974-82 contained 8-9 eggs; he did not trap the hens, but in at least some of these nests the eggs were of two types. In his area, three of these double clutches produced young, but fewer than the average six-egg clutch would give. Four double clutches, containing 8-10 eggs, had been previously reported by Jourdain (1928) and others, and at each nest two hens were shot or seen together. In a fifth case, two nests had been built side by side, with a hen on each. In no instance was there evidence of more than one cock.

INCUBATION

After laying has begun, the female Sparrowhawk normally remains on the nest or on a nearby branch. She may even lie beside the eggs of the incomplete clutch or stand over them, but without deliberately warming them. It is presumably the need to guard the eggs against predators which discourages the hen from wandering far at this stage. However, some females do apparently leave the nest area at times during the laying period, but the evidence is indirect. Owen (1916) noted at one nest that each egg was eaten by a Jay as laid, and I once disturbed a Jay at one nest and a Red Squirrel at another, only to find on climbing the trees that the eggs had been partly eaten. This was unlikely to have happened if the hens were close by, and I could only conclude that they were far away. In both cases, further eggs appeared, showing that the nest had not been abandoned. These

170 The breeding cycle: eggs and incubation instances accounted for less than 0'2% of clutches in our study areas, suggesting that absences by the hen were rare. This was born out by the radio-tagged females which, during the laying period, were at the nest whenever I checked. Sparrowhawks seem to start incubation either suddenly with the laying ofa particular egg (any egg after the second), or gradually, sitting for progressively longer periods with each successive egg, and reaching full incubation at around the time of clutch completion. I did not discover the frequency of these different systems, but either way, the eggs usually hatched over a period of two or more days. In general, the larger the clutch, the greater the spread in hatching dates, and the greater the initial size variation among the chicks (Chapter 13). During incubation, the hen covers the eggs almost continuously. Her behaviour is characterised by intermittent dozing, frequent changes of position and turning of the eggs. In dozing, the hen closes her eyes and nods her head for several minutes at a time; but she wakens at the slightest sound and is instantly alert to any disturbance. Every hour or so the female rises, peers at the eggs, and sweeps her bill gently between them towards her belly. In this way the eggs are both turned and shifted in position relative to one another, thus helping to ensure a more even distribution of warmth. The hen also adjusts her behaviour according to ambient temperature, from merely shading the eggs from the sun on hot days (rare in our areas), to sitting tight on cold days. By taking thermo-electric measurements at four nests, Holstein (1950) claimed that the incubation temperature rose slowly from about 23°C to 36°C in the first 18 days of incubation and thereafter kept a fairly steady 37°C. In an incubator, however, I successfully hatched Sparrowhawk eggs with the temperature at 36°C throughout. Over the incubation period, the eggs normally lose about 17% in weight, due to loss of water through the shell pores (such pores are necessary for the embryo to breathe). It is not known to what extent the hen controls water loss through incubation, but it depends partly on the humidity in the nest cup, under her brooding body. Periodic standing would reduce such humidity by allowing damp air to escape, while brooding with wet feathers may help to increase it. In species monitored throughout incubation, nest humidity fluctuates much more than does temperature (Howey 1982). The sitting hen periodically preens herself. She stretches back to reach the oil gland at the base of the tail, and then works over her dorsal plumage. To reach her underside, she may rise, or step to the nest edge for a few minutes to preen more vigorously. As incubation proceeds, the hen moults heavily, and feathers and down emerge at every preening. Some of the down is swallowed, but most becomes caught upon the nest and surrounding twigs, and in late incubation it is a beautiful sight to see the hen sitting amid a well-flecked nest. I have often wondered why the hen does not clean off this down: to the human observer it makes the nest more conspicuous, but the eggs within the nest less so. After preening out a large wing or tail feather, the hen may play with it for a while in her bill before tucking it into the nest structure, or releasing it to drift to the ground. While sitting, she may

The breeding cycle: eggs and incubation

171

occasionally adjust the nest material, or nibble at nearby twigs. Perhaps the nibbling is done merely to relieve boredom. Females are very careful when moving around close to their eggs; they usually shuffle on their tarsi and, like some other raptors, they hold their feet in a special position, with the back toe turned forwards between the front toes, which are bunched limply together. Birds also turn their eggs with the bill slightly open. Both actions reduce the risk of spiking the shells. A hen with six eggs normally keeps them in two rows of three and, if they are disturbed, she puts them back into that arrangement as she covers them. A hen with five eggs usually keeps them in rows of three and two, and one with four eggs keeps them as two and two. When a Sparrowhawk is flushed from the nest, an egg may be accidentally knocked clear of the cup towards the rim, or a small chick may be flung over the edge. If not replaced by the observer, a displaced egg remains on the nest edge indefinitely (unlike some other birds, Sparrowhawks do not retrieve displaced eggs), while fallen chicks would soon chill and die. On the other hand, a Sparrowhawk which leaves the nest in its own good time does so slowly, removing its feet carefully from the eggs and stepping to the nest edge before flying. After returning to the nest edge, the hen walks forward, placing her feet carefully among the eggs, and as she lowers herself, she fluffs out her breast feathers, to apply the brood patch to the eggs. There is then a rapid shuffling, as the bird finally lowers herself with another shuffle to the level position of incubation. The feathers then become generally loosened, sleeking only when the bird is on the alert. Throughout incubation, the cock brings food to the hen up to five times per day, presumably depending partly on his hunting success. The hen eats the prey away from the nest, possibly for reasons of hygiene. The cock has normally plucked each item beforehand, enabling the hen to bolt it quickly, and return promptly. The eggs are thus exposed for only 5-10 minutes and on some occasions the male may even stand on the nest while the female feeds, leaving when she returns. The cock does not incubate, probably because he is too small to cover the eggs effectively. In other raptors in which the male can incubate (such as Peregrine), the female eats in more leisured fashion, and may spend an hour or more off the nest at a time, while the male takesher place. Apart from feeds, the hen Sparrowhawk leaves only for brief periods to defaecate or preen nearby, where she can keep an eye on the nest. When she returns, she often brings a twig to add to the nest. I know of no certain cases of hen Sparrowhawks killing prey for themselves during incubation. I could not rule out the possibility that hens may have left the nest to attack a bird which ventured near, but could be certain that they did not normally leave for more than minutes at a time. During thousands of visits to successful nests during incubation, the hen was almost always seen, and the eggs were always warm. The only cold eggs proved subsequently to have been deserted. Also, I never caught a breeding female in a baited trap during incubation, even when the trap was placed for hours in view

172

The breeding cycle: eggs and incubation

of the nest. I t seems, therefore, that hunting behaviour is suppressed in the brooding female, in favour offood from the male.

ANTI-PREDATOR BEHAVIOUR

The chance of almost any form of mortality to eggs or chicks can be lessened with greater parental care, but nearly always at the cost of greater risk to the adults. If a Sparrowhawk loses its eggs, once incubation is well under way, its sole chance to reproduce that year is lost, so not surprisingly the adults take steps to defend their eggs against predators. This entails sitting tightly on the approach of a predator, engaging in mobbing displays, settling near the predator, or even attacking it. To a Sparrowhawk, any human intruder is a potential predator, so much can be learned by watching the reactions of the adults to your own presence. The 'tightness' with which a hawk sits, and the speed with which it returns after leaving, vary with the individual and with the state of incubation. Some females are hard to dislodge from eggs when incubation is well advanced, and will not respond to hand claps or to raps on the tree trunk. A few wait until the observer has climbed within a metre or two of the nest, and I have known two females which have merely retreated in threatening posture to the back of the nest as I poked my face over the side. Andy Village managed to grab one of these birds by hand in order to read its ring number, and after release the bird was soon back on the eggs. In general, the birds sit especially tightly around the time of hatch, and are reluctant to leave, even when the male arrives with food. Tight sitting in the presence of a predator confers maximum protection to the eggs, but carries the risk that the hen herself may be killed. The hen may sit more tightly around the hatching stage because the chicks within the chipping eggs cheep loudly, and would soon attract attention. On a still day, they can be heard 30m away. They may also be especially vulnerable to desiccation then. Once they have been put off the nest, some females leave silently and watch a human intruder from a distance, but most fly from tree to tree around the nest, uttering loud alarm notes, 'kek ... kek ... kek ... ' repeated rapidly. This is a form of mobbing, and at the least serves to distract attention from the nest to the hen. If the observer climbs the tree, occasional females will dart at him, and perch in the tree, calling loudly, presumably in an attempt to drive him away. Natural predators, such as owls, elicit diving attacks and may be struck, as mobbing escalates to outright attack, in attempts to protect the nest contents. Loud calling by the female in the presence of a predator has two additional effects. It may be heard by the male, who is usually away hunting, and summon him to help with the defence. It may also attract another predator, a situation in which the Sparrowhawk has little control, but may occasionally gain. At a nest in Holland, for example, as a noisy Sparrowhawk mobbed

The breeding cycle: eggs and incubation

173

an owl near the nest, a Goshawk suddenly appeared and took the owl. Sparrowhawks respond differently to different predators, according to the risks they pose. In general, natural predators, such as Tawny Owls, are attacked much more vigorously than a man. This may be associated with the difference in body size but, in addition, the behaviour of Sparrowhawks towards men has probably been moulded by more than a century ofpersecution. Those individuals which in the past put up the boldest defence were most likely to have been shot. If the defence behaviour was inherited, over the years the more timid individuals would have been favoured by selection. The behaviour of the related Sharp-shinned Hawk and Goshawk in North America lends support to this view, for in areas (such as Alaska) where they have not been molested, these species put up a very vigorous nest defence, and will strike a man repeatedly. Yet in several thousands ofvisits to Sparrowhawk nests in Britain, I have only once been hit. When trapping hens on the nest, we had plenty of opportunity to check how quickly birds returned to the nest once the disturbance had passed. Most returned within 15 minutes, but some with hatching eggs were back in less than five minutes, and occasionally while we were still descending the tree. In my experience, Sparrowhawks do not desert their eggs as a result of people visiting their nest, even when this occurs daily, as in the case of a biologist taking records. In most years we had pairs which bred successfully on nests over public footpaths, and heard of several Sparrowhawks hit by gunshot as they left the nest returning to be shot at again, or die later on the eggs. Those desertions which I could attribute to disturbance usually involved prolonged activities near the nest, which would have kept the bird off for hours on end. Some females continued to incubate in the midst of tree-felling activities. As the trees crashed around them, such birds often sat tight or, if disturbed from the nest, they managed to sneak back while the men retreated for a tea-break. On four occasions, when a wood was being clear-felled, I persuaded the cutters to leave the nest tree until later, so that a tree which began the season in the midst of a wood ended as a solitary pole in a bare plain. The two birds which had eggs at this stage did in fact desert, but the two with young reared them successfully to independence. Another two nests that were brought down by felling were moved to adjacent standing trees, and produced young. Otherwise, where desertion did follow some obvious human disturbance, this was often where the bird was underweight and probably on the point of giving up for reasons of food shortage.

DURATION OF INCUBATION

As I could not always be sure when incubation started, incubation periods were best calculated as the interval between the laying and the hatching of the last egg in a clutch, assuming that incubation started, at the latest,

174 The breeding cycle: eggs and incubation with the laying of this last egg. In south Scotland, intervals were calculated only from nests in which all eggs hatched, and not from nests in which addled or broken eggs were found, for only in this way could I be sure that the last egg was included. Some 72 periods ranged between 31 and 37 days, with a mean of 33 days (Table 27). This was a wider range than the 32-35 days given by most authors, but it was based on a larger sample. Longer periods tended to be associated with larger eggs, but this trend was not significant statistically. No relationship was apparent between incubation periods and time of year or clutch size.

FLEXIBILITY OF PARENTAL BEHAVIOUR

Once the nest is completed, the hen may spend some time sitting there before laying. If the eggs are addled, she will sit on them for up to twice as long as is normal. On the other hand, the completion of incubation is not necessary before hens will show parental behaviour. They will accept and feed young given to them at almost any stage of incubation, as we found on several occasions when, for one reason or another, we had to swop eggs and chicks around. Clearly, the integration of the various activities that comprise a breeding cycle involves not only changes from within the bird, but also a continuous interaction between the bird and its environment. The fact that Sparrowhawks will accept strange eggs, or even chicks at a different stage of development from their own, suggests that they lack any ability to discriminate against offspring which are not their own. This is perhaps not surprising, because there would normally be no opportunity for eggs and young from different pairs to intermix, so no reason for such discrimination to evolve. Individual recognition of eggs and chicks is chiefly a feature of colonial seabirds, such as penguins, which are crowded together and have young that wander from the nest. In some seabirds, even the young. can distinguish their parents from other adults (Evans 1970).

SUMMARY

Sparrowhawk eggs are pale blue with red-brown blotches. In the natural setting of flickering light and shadow, they are cryptic. Laying occurs from late April to early June, but mainly in May. It is slightly later on high ground than on low. Most clutches contain 3-6 eggs. On very rare occasions, two females lay in the same nest, giving 8-10 eggs of two types. Comparison of mean laying dates and clutch sizes from various parts of Britain reveals no geographical trends, but over a wider span of latitude from Mediterranean to boreal regions of Europe, laying becomes later and mean clutch size increases. Throughout incubation, the female covers the eggs almost continuously, and is fed by the male, who brings prey to the nest vicinity. Females usually call loudly or mob human intruders at the nest, whereas they attack

The breeding cycle: eggs and incubation

175

and strike natural egg pr ed ators. They are espec ially relu ct ant to lea ve th e nest a ro und the time of hatch , a nd ifput off, normall y re turn within minutes.

CHAPTER 13

The breeding cycle: growth of young

This cha p ter is the first of three whi ch deal with th e lives of young Sparrowhawks from hatching to indep enden ce. This is a period of eight weeks , of whi ch the firs t half is sp ent in the nest and the second half close by . During this time th e young progress from helpless hatchlings, dependent upon their parents for warmth , food and protection, to free-living individuals, able to regulate their own body temperature, obtain their own food , avoid predators and take care of th eir own plumage and body hygiene.

THE HAT CH

Some days before an egg hatches, th e chick can be heard tapping and che ep ing inside, and up to two days before hatching th e first chip appears in th e sh ell. Many hours ela pse before the chick begins to work its way round, cu tting th e sh ell in half, and emerging with th e help of its mother. The female th en eithe r carries the shell pieces away and drops th em, or places th em on the nest rim and gradually nibbles them away over th e next day or two . The removal of th e larger shell pieces from the nest cup is important, because they can 'cap' any remaining eggs and prevent them from hatching. Smaller pie ces of shell get trampled into the nest material.

The breeding cycle: growth ofyoung 177 The newly hatched chick is wet, and lies limp and exhausted, too weak even to raise its head. Over the next few hours, the chick dries under the warmth of its brooding mother. At this stage, it weighs some 14-18g, about two-thirds of the fresh egg weight. About 17% of the weight loss is due to evaporation of water through the shell during incubation, and the rest to the shell, membranes and other waste products discarded at hatch. The remaining yolk is absorbed over about 36 hours, and if the chick is not fed during this time it will die, weighing 10-12 g, about two-thirds as much as at hatch. When dry, the newly-hatched Sparrowhawk chick is, as Owen (1916) wrote, 'one of the most beautiful nestlings in existence. It is covered with short, thick, pure white down, which is thicker and shorter on the head than on the body ... the leg is bare from the ankle to the foot on the underside, but the down comes below the ankle on the upperside. The legs and toes are flesh coloured, the talons light ash grey. The iris is very dark brown, sometimes with a tinge of grey in it, the pupil a very deep indigo. The culmen is short and curved with the base flesh-coloured, the point black, and a small white excrescence at the bend (the egg tooth). The lower mandible is flesh colour with a dark ash stripe on each side, the tip slightly darker. The mouth and tongue are pink flesh-colour.' In hatching with a complete covering of down and with their eyes partly open, Sparrowhawks resemble other birds of prey, but differ from songbirds, which are born naked and blind. The young hawk is also fed differently, for instead of craning its neck and opening its bill upwards for food to be placed within, the young hawk faces its mother and takes food morsels directly from between her bill tips, an action which requires more visual perception and co-ordination than a young songbird has at this age. Although at first the hawk chick's eyes appear bleary and unseeing, they must be able at least to focus roughly on close objects and to recognise from other colours the red bits of meat which stimulate the pecking response. The chick communicates its hunger by cheeping softly, and occasionally by pecking instinctively towards its mother's bill. The hen orientates her head carefully so that the young can reach the food. She normally feeds her young for the first time within a few hours of their hatching. Because Sparrowhawks start incubating part way through laying, the eggs hatch asynchronously, and in general the larger the clutch the longer it takes for all the eggs to hatch. At nests where full details were obtained, three-egg clutches hatched over 1-2 days, four-egg clutches over 1-4 days, five-egg clutches over 2-6 days, and six-egg clutches over 3~6 days (Table 28). So the larger the clutch, the greater the range of sizes among the chicks, at least initially. The advantage usually claimed for an asynchronous hatch is that it leads to a grading in size among young, so if food is short, the smallest chick quickly starves, thus reducing the brood to the number that the parents can feed (Lack 1954). If all young hatched on the same day, and if food was scarce, it would take longer for the weakest to die, with a correspondingly

178

The breeding cycle: growth ofyoung

greater waste of food. Asynchronous hatching can thus be regarded as an adaptation to an unpredictable food supply, enabling all young to survive in times of plenty, but ensuring rapid reduction of the brood to an appropriate level in times of scarcity. The degree of asynchrony is less in Sparrowhawks than in some other raptors and owls which incubate in earnest from the first egg. But the food supply is probably more predictable for Sparrowhawks than for these other species, so it is advantageous in Sparrowhawks to put only part of the brood at greater risk than the rest.

GENERAL DEVELOPMENT

As is usual among birds, young Sparrowhawks do not develop all parts of their body at the same rate, but put the emphasis on different parts at different stages, depending on the needs of the time (O'Connor 1984). The first need is for effective food processing, because on this all subsequent growth depends. The digestive system is well formed at hatch, but develops further in the ensuing days. The legs and feet are the next obvious parts to show most growth, probably because they help the young to position themselves at feeds, and thus compete effectively with their nest mates. Then follows a period of great weight-gain, which gradually gives way to a period of increased plumage development. The body feathers, which help in insulation, are complete before the large wing and tail feathers, which do not need to function effectively until the bird leaves the nest. Behavioural development depends to some extent on morphology, but is also sequential, according to the needs at different stages. The chick's ability to take food from its mother improves rapidly within days after hatch, and it later develops the ability to peck at meat, to stand, to pluck prey and to fly, in that order. These patterns of development have presumably evolved for reasons of economy, directing the resources available for growth where they are most needed at the time, and ensuring that structures and behaviours used in the nest develop more rapidly than those needed after fledging. The bird is thus equipped to meet the functional needs at each stage. When the young leave the nest, they are approximately the same weight as adults of the same sex. They can fly, but their feathers are still in growth.

GROWTH RATES

This section is based largely on the work of Dorian Moss (1979) who, over a 3-year period (1973-75), made daily visits to nests in south Scotland to weigh and measure the young. He marked the chicks in each nest on the day they hatched in order thereafter to distinguish them individually. He visited the nests at about the same time every day, taking measurements on each young at roughly 24-hour intervals. Individual young could be handled only up to 22-25 days, because after this age they tended to flee the

The breeding cycle: growth ofyoung 179 nest prematurely if disturbed. The nests concerned were in the Annan Valley and Ae Forest, representing habitats that were rich and poor in prey respectively. Over the three years, 49 broods, containing a total of 194 young which survived for at least two days, were studied. The measurements taken from each chick included body weight, the length of the outer large primary (number 10) as a measure of feather growth, and tarsus length as a measure of bone growth. The marked size difference between male and female developed mainly in the first three weeks. Hatched from eggs of similar weight, the sexes diverged rapidly and,after 24 hours, females were already heavier than males to an extent that was significant statistically. However, the young could not be sexed with certainty until a later age, because of some overlap in early weights between the sexes. In tarsus and primary length, the sexes did not diverge until 9 and 18 days respectively (though males had significantly longer primaries at 6-13 days). As the young grew, the sexes diverged further, and from about half way through the nestling period, say from 12 days old, could be easily distinguished by eye withou t risk of error. Females were at once recognisable by their larger size, especially of their legs and feet. One of the most convenient comparisons, I found at ringing, was the greater width of the tarsal joint in females, which could easily be seen by turning the young on their backs. If the young called, the deeper voices of the females were also evident. As the chicks in the study nests were marked individually from hatch, they could be classed retrospectively, once their sex became apparent. Young of both sexes increased in weight slowly for 4--6 days after hatching, then rapidly for about 10 days, and then slowly again, giving the usual sigmoid growth curve typical of many birds (Fig. 43). Calculating the daily averages 280

~ E if E 60

60

200 CJ)

E

cf'

~ 160 C)

..... s: Q)

CJ) CJ)

~

~

80 40

~

E 50 E

~

~

>

ro

E

8 12 16 Age days

20 24

40

0

E 20

20

0

50

.~ 30 ..... CJ)

30 -

2 ;:,

10 4

~ if

s:

s: C, 40 c: ~

.~120

0

70

70

~

240

0

4

8 12 16 20 Age days

24

10 0

4

8 12 16 20 Age days

Fig. 43. Growth rates ofyoung. Lines show mean weights, tarsus lengths and outermost primary lengths of nestlings at different ages, with 95% confidence limits. Based on 49 broods studied by Moss 1979.

24

180

The breeding cycle: growth ofyoung

from the weights of many young, tended to smooth the curve, and eliminate the sometimes large day-to-day fluctuations which occurred within broods. These fluctuations occurred whether or not the young had just been fed at the time they were weighed. In illustration, Fig. 44 shows the growth 10

c

.~

'0 ~

:J 0

I

250

-: 1

200

/

:\

•···••..•..ODEAD

.:

'bDEAD

·e:

c/

~ · · · · ·O·· ....

150 C)

0)

~~ ~ ,

...

.

~.......

:.. :

~

~P

/-..,

€f

.0 /

1.-.@'

/

.

/~DEAD

•

._.-.-.-.

~

:

...•.....•.....•....

·

DEAD

··· ·...,:..0 . /•..•.•....•.

50

::

~/./.

/ . . / . ,':-A--

0/.

..£.

• ..........-

/

.:'

A_

A-A ............A--ADEAD

-;:/

A~'

---~i~./" O-i,;:::::·--

o

j

JULY

~

Fig. 44. Weight gain ofyoung in an Ae Forest brood which experienced poor growth and eventually died. Growth rates ioere .related to hatching order with thefirst two chicks growing best, and the last three grouting poorly. Notice how the growth of the chicks was reduced on days of heavy rain. Four out offioe deaths occurred on wet days..From dataprovidedby D. Moss.

i

15

The breeding cycle: growth ojyoung

181

of one particular brood, in which the young suffered poor growth and eventually died. As was usual, weight gain in this brood was related to hatching order, with the first hatched chick showing the best growth, and the last chick the worst. The mean growth curves for tarsus and primary feather lengths were also sigmoid in shape, though somewhat shortened at the left and right-hand ends respectively (Fig. 43). Comparing individuals, these measures fluctuated less about the mean values than did weight. It seemed that, when food was scarce, young continued their bone growth at the expense of feather growth, and bone growth and feather growth at the expense of weight gain. Similar tendencies have been noted in passerines and other birds. They reflect priorities in growth which enable any underweight young in a brood to develop in competence at about the same rate as their better-fed siblings. One of our objectives was to compare the growth of young in different broods. For this we needed a single measure for each young which reflected its overall growth. Because of (a) the sex difference in growth, (b) the sigmoid pattern, and (c) the large day-to-day fluctuations in weight gain, obtaining such a measure was not straightforward. However, the central parts of the mean growth curves for weight, tarsus and feather length were linear over periods of about 10 days. For these periods it was possible to calculate a statistic of growth - the slope of the line (or linear regression coefficient) for the weight, tarsus and feather lengths of each nestling which survived through the period concerned. These coefficients are hereafter termed 'growth rates'. Each growth rate, in grams or millimetres per day, was also expressed as a percentage of the mean growth rate calculated for all males and all females respectively. Each young could then be compared directly with others of its sex, and different broods could be compared without bias due to sex composition. One point was obvious in both sexes from the start: the less the day-to-day fluctuation, the better the overall growth. This was most evident in weight, but was also apparent in tarsus and feather length. Birds which showed the largest day-to-day fluctuations in growth were also those most likely to die as the weight losses became severe. Observations at nests confirmed that good, even growth resulted from frequent feeding, while poor and irregular growth resulted from infrequent feeding (Chapter 14). Similarly, in broods in which the mean growth rate was high, the variation between the young was much less than in broods in which the mean growth rate was low. This implied that, under good food conditions, when growth rates were high, all the chicks in a brood were fed equally well, but under poor food conditions a much wider spread in growth rates resulted, as some chicks went hungry. Broods which gained weight rapidly also showed good bone and feather growth. Factors affecting growth

Growth rates (relative to the mean for each sex) were found to vary according to brood size, habitat, hatching date, and age of mother. Among the

182

The breeding cycle: growth ofyoung

broods studied by Moss, the number of young after hatching ranged from 3 to 6, with 4 the most frequent. In the food-rich habitat of the Annan Valley, where few chicks died, mean growth rates did not differ between broods of 3, 4 or 5 young. In these areas, the parent hawks evidently found it just as easy to raise 5 young as 4 or 3, and this held whatever the sexes of the young. In such habitat, therefore, growth rates were independent of brood size. This was not true in parts of Ae Forest, however, where chick starvation was frequent, reducing most broods to three or less. In this area, comparisions ofgrowth between brood sizes were complicated by the frequent depletions of broods which occurred, so the effect of brood size on growth couldnotbe properly examined. Nevertheless, the most appropriate brood size for poor areas (such as Ae Forest) seemed to be three, compared to five in the better areas (such as the Annan Valley). In poor areas, the effect of the asynchronous hatch was apparent, and the earlier hatched young in each brood were at a clear advantage over their younger siblings. In general, the growth of nestlings was poorest on the highest ground in remote parts of Ae Forest, intermediate in central Ae Forest, but best at the lower edge of the Forest and in the Annan Valley. The difference in this respect between the Ae Forest as a whole and the Annan Valley was statistically highly significant in both sexes, and in all three measures (P < 0·001). That this was due to a difference in prey supply between areas was supported by an experiment in which extra food (sparrow carcasses) was provided daily to two broods in the remote part of Ae Forest. Four and five young were reared respectively, and the five had growth rates well above the average. These were the only occasions, in a ten year period, when such large broods were raised in this impoverished area, despite many pairs starting with five. On average, owing to differences in laying dates, eggs hatched about four days earlier in the Annan Valley than in Ae Forest. In both areas, the later in the season a young hatched, in general the poorer it subsequently grew. Overall this trend was significant only for primary feather growth (P < 0,05), which showed a decrease of 0·35% for each day delay in hatching, but in the Annan Valley it also held for weight (P < 0·02), where the decline in rate of gain was 0'5% per day. Evidently, Sparrowhawk pairs which hatched young early in the season found it easier to supply food than pairs which hatched young late. This was probably because the earlier pairs were more efficient foragers than the late ones (Chapter 16). One other tendency was apparent: namely that, in the same habitat, young with adult mothers showed better growth than young with yearling mothers (Newton et al 1979). As adult hen Sparrowhawks usually paired with adult cocks, and yearling hens with yearling cocks (Chapter 18), most of the broods

with yearling mothers probably had yearling fathers as well, though we were unable to check this at every nest. Adult parents of either sex would, on average, be more experienced at raising young, and also at foraging, so in this aspect, too, food-supply may have been involved. In no aspect of growth were differences found between years, either in Annan Valley or in Ae Forest,

The breeding cycle: 'growth ofyoung 183 so all the analyses discussed above were based on the pooled data for all three years. To summarise this section, in good food areas nestling growth was best and was also as good in large broods as in small ones, but in poor food areas growth was generally poorer, especially in the larger broods. Within areas, growth was better in early broods than in late ones, and better in broods with adult mothers than in those with yearling mothers. The artificial provision of extra food at two nests in a poor area improved the growth and survival of the young over those in control nests with no extra food. Evidently, the food-supply for young was a key factor limiting the overall reproductive rate, at least in the hill areas with low prey numbers.

Effect ofwet weather ongrowth In addition to the factors discussed above, wet weather depressed growth (Fig. 44). For each day on which rain fell for more than four hours, the mean weight, tarsus and primary length increases of individual nestlings were calculated, and compared with the expected mean increases ofan identically aged set of nestlings growing at the average rate for the population. This revealed that, in each year of study, weight gain was depressed significantly on wet days. Feather and bone growth were much less affected, evidently being maintained selectively, as discussed above. However, the chicks often developed fault bars on wet days. These were fine lines across the flight and tail feathers where the rachis was pinched and the barbs deficient in barbules. The feathers were later liable to break at such points, affecting the bird's flight until the next moult, a year later. Observations at nests showed that, during rain, young were fed significantly less often than normal, probably because the hunting ability of the parents was impaired (see later). MORTALITY OF NESTLINGS

Daily visits to nests enabled Moss to find exactly when chicks died, and to elucidate the causes in a way not possible with less frequent visits (Fig. 45). Most deaths occurred within two days after hatch. This was partly linked with the asynchronous hatch, and as the last chick to emerge was the smallest, it was at a disadvantage compared with its nest-mates. Such young sometimes starved, presumably when their yolk reserves were exhausted. Of the young which survived more than two days, 21 % died before fledging. This mortality was higher in Ae Forest (33 % ) than in the Annan Valley (6%). Most young which died disappeared quickly, being eaten by their mother or siblings, so unless one was present around the time of death, there was little chance of finding the cause. All one could usually record was a reduction in the number of young present. At one nest under observation from a hide, the hen was seen to pick out her smallest nestling when it had become weak, and feed it to the rest of her family. This was the only case of infanticide observed, but it may have occurred more often.

184

The breeding cycle: growth ofyoung 100----------------------,

80

60 C)

c: > > I-

::J

en

40

O-Q

20

H

5

10

15

20

25

Days

Fig. 45. Survival of nestlings through the first 25 days, as determined from daily visits to 55 broods. Data mainlyfrom D. Moss.

The most common cause of chick mortality beyond the second day was starvation, which accounted for 15 (38 % ) of all deaths in the broods studied by Moss. In each case, death was preceded by several days of poor growth, and ultimately by a loss in weight. Nine of the nestlings to die in the twelve broods concerned were the last of the brood to hatch, a higher proportion than expected by chance. In three broods, two young starved in succession. Eight of the 15 young died at 12-16 days of age. The second most frequent 'cause of death was exposure to rain. Ten (25 % ) deaths in four broods occurred in prolonged wet spells. Such deaths were not preceded by weight loss (unlike starvation), but some of the young which died during rain had previously shown poor growth. All the young concerned were aged 18-23 days, at a stage when their developing feathers gave them little protection from water and their down became sodden, leading to death by chilling. They included two complete broods, of three and four, which probably received no protection from the adult hen, and three single nestlings in other broods. Such deaths were all in the remote or central parts of Ae Forest. These were areas with little prey, where we knew from radio-tracking that the mothers often hunted several kilometres away. One brood died during a torrential downpour, which came on so suddenly that it prevented the hen (which we were tracking) from returning to shelter her brood. This hen was 8 km from the nest at the time. In contrast, among nests which were at risk at the same time in the Annan Valley, all the young survived. The third most important cause of mortality was predation by Tawny

The breeding cycle: growth ofyoung

185

Owls, which accounted for a further seven deaths (18 % ) in three broods. "This was inferred from owl feathers which were found on the nest, apparently lost during a struggle with the young hawks. These chicks were large, and too old to be brooded by the hen, so had probably been taken while the hen was away hunting. They also were at nests in Ae Forest. The remaining six deaths (15 % ) were due to the desertion of two broods. In neither case was prey left on the nest by the cock following the hen's disappearance, so desertion by the hen may have followed the desertion or death of the cock. One of the hens was trapped alive in the following year, so her desertion had evidently not been due to her own death. These were the only cases of nestling desertion recorded during the whole study, except where a parent was known to have been shot. Among these various causes of mortality, only starvation was greater in larger broods, and only in the area of Ae Forest where prey was in short supply. Other deaths were independent of brood size, but were again more frequent in Ae Forest. They were associated with female absence, which was in turn linked with poor prey supply (Chapter 14).

OTHER ASPECTS OF NESTLING DEVELOPMENT

As the young grew, the white down, with which they were clothed initially, became too sparse to cover them effectively and at 7-8 days was replaced by a longer, slightly greyish down, a prelude to the feathers proper (Meissel 1937). The first real feathers to emerge were the large flight feathers of the wings (from day 5), followed by the large tail feathers (from day 9), and then the smaller feathers of the body and lastly the head. These various feathers pushed out the down, which stuck to their tips until the feathers were more than half grown. The primary feathers broke sheath from about day 11, the secondaries from about day 12, and the tail feathers from day 13. In general, males were ahead of females in plumage growth. The first body feathers appeared on the flanks on day 14 in males and day 16 in females, and the growing scapulars formed brown stripes down the back by day 18 in males and day 20 in females. Despite its later emergence, the body plumage developed faster than the flight and tail feathers. It helped to insulate and camouflage the young, giving them a mottled brown appearance to replace the white, and coincided with their being left unguarded for longer periods. As the various body feathers grew, down was shed from different parts of the body in turn, but it did not go completely until 2-3 weeks after the young had left the nest, with the crown the last to clear. At 14 days the eyes began to change from dark brown to greenish-grey. The skin between the eye and bill became bare and greenish-yellow in colour. By the 17th day it became covered with the dark hair-like filaments typical of the full-grown hawk. The eye itself gradually became more yellow, but

still had a hint of green at fledging. The behavioural changes were equally striking. Within an hour or two

186 The breeding cycle: growth ofyoung of hatching, the chicks could not only take prey fragments from their mother, they could also alter their position within the nest cup, and excrete outwards towards the nest rim. They soon became more proficient at adjusting foodpieces to facilitate swallowing, and within a few days could gulp down the feet and other bony parts of small prey, as well as the softer parts. They began to sit up at feeds, and group themselves in front of the hen, grabbing the food from her bill. As the young grew, they also became able to eject their droppings further. At first the faeces were in the form of small white blobs, but soon became a liquid stream, as in adults. Earlier attempts fell on the nest, but by five days, the first spots of whitewash began to appear on the ground below, and by the time the young flew a wide area, up to 4 m radius under the nest tree, was well splattered - a sure sign to the practised eye of a well-grown brood above. The inaccessibility of the nest gives some protection from ground-hunting mammals, and Sparrowhawks - like other raptors - do not remove their droppings from the area. The young developed feeding skills gradually, and again males were ahead of females. They were observed to peck at prey from day 9 (11 in females), put their feet on it in a sitting position and pull off strands of meat by day 10(13), obtain some food in a standing position by day 13(15), eat complete prey items by day 18(20), and do this proficiently by day 24(26). As to other skills, the young began opening and flapping their wings on day 10 (12 in females), standing on day 13(15), obtaining lift with wing-flapping on day 16(21), moving onto branches near the nest on day 19(21), and made their first flight on day 26(30), before their feathers were quite full grown. The young made their first preening movements on the breast on day 9(10), and were proficient at preening the whole body by day 22(23). All these figures are averages from three broods watched from hides, and, where growth was poor, most aspects of development were delayed somewhat. When the young were able to stand, they greatly increased their feather preening. The wings and tail received most attention, as each large feather was drawn through the bill at each session, and the wing coverts were continually smoothed and re-arranged. The head received least attention, because it had to be done with the foot, and the other foot was not strong enough to support the bird at this stage. In consequence these early efforts were entertaining to the concealed observer. The young also cleaned their feet and bill in the same way as adults. An amusing performance in young over three weeks old was what Owen (1916) termed the 'flap dance'. A bird would extend its wings fully and slightly above the horizontal and, with very rapid quarter beats, flap its way around the nest, rising a few centimetres above the surface at each jump. Such flapping often took place suddenly in the midst of preening, and when one bird began, others often joined in. As the young got older, it was not uncommon for all to go through this procedure after a meal. It presumably helped them to 'feel their wings'. Another exercise was used for working the leg muscles. A leg was lifted and extended forward and,

The breeding cycle: growth ofyoung

187

while rigid, the toes were opened and closed slowly several times before being brought down onto the nest. The young also began to mantle the food at this stage, to crouch over it with wings spread, and the first young to grab an item from a parent usually got it all. The gradual acquisition of various body skills by the ybung was accompanied by an increase in their general awareness. The chicks spent their early life huddled together in the nest centre, mainly being brooded, with occasional feeds and long sleeps. But in the late nestling period, they spent most of their time standing and looking intently around, with occasional spells offeeding, preening, wing-flapping and walking about the nest looking for prey scraps. They showed intense interest in any moving creature, whether small birds in the canopy or deer on the ground, and spent long periods staring at flies around the nest.

Sex difference in development The sex difference in development was striking. The females gained weight faster, but the males grew feathers and developed skills more quickly. Probably this was adaptive, enabling small males to survive in the presence of larger sisters. By remaining one step ahead of their sisters throughout, the young males could acquire their share of food and parental care in a way not possible if they were smaller but no more advanced. Also, by the time of their first flight, the young cocks weighed about 10 g (15 % ) more than adult cocks, largely because of extra fat reserves. In contrast, the young hens at this stage weighed rather less than their mothers, and did not carry the surplus fat. The males probably had greater need of a body reserve then, because it was in the post-fledging period, while they were still being fed by their parents, that their sisters caught up with them in development. Anti-predatorbehaviour When newly hatched, the young gave faint begging calls at feeds, audible for at most a few metres, but later they were normally completely silent at feeds. In this respect they contrasted with songbirds, which cheep loudly while being fed. The only food-begging I heard in Sparrowhawks was from nests where all the young were starving. Such loud begging, while perhaps exhorting the parents to greater efforts, must also put the brood at greater risk from predators. On a still day it can be heard 100 m away. Until about half grown, the young have little defence against predators, except to remain quiet, and crouch flat on hearing the alarm call of their mother. At about this age, however, they themselves develop the loud 'kek ... kek ... kek' alarm call. On your approach, the young first flatten themselves on the nest, but as you climb the tree they begin to call. This usually brings the hen if she is not already there. When you reach the nest the young draw away from you and lay on their sides with their beaks open, making a faint hissing note caused by the exhalation of air. The tongue collects moisture which drips or shoots out on exhalation. Larger young, that can stand, retreat to the back of the nest, and face you with plumage fluffed,

188

The breeding cycle: growth ofyoung

wings extended behind their backs, with their mouth working as before, and striking with their feet at your outstretched hand. Lined up in this ,,yay, the young must provide a formidable barrier to any predator. If handled, they grip the nest surface, which makes them difficult to carry away, an action of no avail against the human observer, but more effective against an owl or other small attacker. When older, the young walk out onto branches or flutter from the nest if approached, as do the young of many other birds. This premature dispersal reduces the chance of a predator getting the whole brood.

FIRST FLIGHT

In three broods studied from hides, in which the young left of their own accord, the cocks flew first at 26 days after hatch and the hens at 30 days. This would have been hard to determine without a continuous watch because each young left on its own, in the absence of the parents, and stayed away for only an hour or two before returning. And at about this stage the young perched increasingly on the branches near the nest rather than on the nest itself. Up to 32 days, the parents continued to bring all prey to the nest, and only later did they begin to pass some items to the young nearby, though the young themselves often returned to the nest to eat. Hence, an observer making infrequent visits to a nest could gain only a rough idea of the so-called 'fledging dates'; on some occasions he would find the young on the nest and on other occasions off it. The fact that the young would fly prematurely if disturbed only added to the difficulty. Five other nests, which we checked from a distance 2-3 times daily, contained hens only, and each brood was first seen away from the nest at 28, 29, 30, 31 and 32 days respectively. A single male from another nest left at 27 days. Averaging the figures for all young, irrespective of brood, gave 26·2 days for cocks and 30·0 days for hens. At three nests watched by Owen (1916), the young cocks left at 28 days and the hens at 29. A four-day difference in the fledging ages of males and females was of the order expected from the size difference between them, as it fitted the same trend that emerged from comparing different species of raptors, the larger fledging at a later age.

SUMMARY

Young Sparrowhawks, like other raptors, hatch with a complete covering of down and with their eyes open. From the start they are able to take food morsels from their mother's bill, and turn to excrete towards the nest edge. They grow rapidly, and the gains in weight, tarsus length and primary length follow the sigmoid pattern typical of many birds. Day-to-day weight gain is. more variable than feather growth, and this in turn is more variable than

The breeding cycle: growth ofyoung

189

tarsus growth . H en ce, wh cn food is scarce, growth of bone and feather is maintained pr efer entially over weight gain . Young fem ales gain weight faster , and achi eve a higher fledging weight, than do males. However , plumage and behavioural developmen t is fast er in males, an adaptation helping the smaller males to survive in the presence of bigg er sisters. When undisturbed , young males first fly at a bo ut 26 days and young females at about 30 days . In south Scotland g rowth was better (a) in small broods than in la rge ones (in a poor food area only) , (b) in a good food area than in a poor one, (c) in early nests than in la te ones, and (d) in nests of adult than of yearling mothers . Weight gain wa s d epressed in wet weather , du e to lower food intake. Many chick d eaths were du e directly to food -shortage, which caused starvation , whil e others wer e due indirectly to food -shortage and included ch illing by rain and pred ation by owls . The experimental provision of extra food to two nest s in poor habitat resulted in im p roved growth and surviva l of young, com pa red to other broods in the area . Food supply was thus a factor lim iting production of young, at least in a poor prey area .

fu reui!« maleStmrroicha u-«

CHAPTER 14

The breeding cycle: parental care

Parental ca re involves not only feeding th e young, but providing them with warmth and shelter, and protecting th em from enemies. It also entails some nest maintenance and sanitation , and other behaviour which helps the young to becom e ind ependent. These va rious activities are not without hazards . By spending more time hunting, th e adults increase their own chance of fatal accidents; by giving food to th e young at the expense of th eir own body condition, they risk starvation; by protecting th e young from rain th ey risk chilling; and by attacking predators , they increase the danger to themselves . However , by investing time and energy on parental care, even at some risk to the ir own survival, th e adults enha nce their genetic fitness by increasing the survival chances of their offspring. In this chapter, I shall describe parental behaviour in Sparrowhawks and how it varies with prevailing conditions, affecting the growth and survival of young. For close study from hides , I selected three nests in Ae Forest which differed in position relative to good feeding a reas. Nest A was on low ground at the forest edge next to farmland , wh ere prey were plentiful; nest B was on higher ground further into th e forest ; while nest C was on th e highest ground a nd further removed from good feeding areas. All three broods began with 4 you ng, but they ended with 4, 3 and 2 survivors resp ectively . Using teams of observers , th ese broods wer e watched for six hours ea ch day during most of the nestling stage. At ea ch nest , th e obs erv ation period was changed from day to day, so that, by combining records from 2-3 successive days, th e whole daylight period was cov er ed . Observations wer e continued until each

The breeding cycle: parental care

191

brood was 30-32 days old, after which the young were being fed away from the nest and were hard to watch. Other nests were observed for shorter periods, mainly to study behaviour.

GENERAL PARENTAL BEHAVIOUR

For the first week or more after hatch, the hens at our nests did little except brood the young to keep them warm, and almost all the activity occurred during brief periods when the cock arrived with food. For the first day or two, each hen was often reluctant to leave the chicks, thus forcing the cock to bring the prey to the nest itself. Perhaps in this way the cock learnt that the hatch had occurred. Subsequently the hen received the food away from the nest, and was off the moment she heard her mate's call. Usually the cock had plucked and beheaded the prey beforehand, but if not, the hen plucked it on the nest edge. Then, holding the item between her feet, clamped firmly by the inside claw of each foot, she pulled off tiny pieces in her bill and offered them one at a time for the young to take. If a chick rejected a morsel the hen would offer it to another of the brood. If it was again rejected, she might offer it to a third or bolt it herself. For the first day or two, the young were satiated after receiving only a few small pieces of muscle or red viscera (liver, heart or lungs), and the hen ate the rest herself. As the young grew, they took other parts of the carcass as well, and within a few days the hen ate only the pieces which the young refused or were unable to swallow, such as the legs. From now on the hens lost weight, living largely on their body reserves, and giving most of the food provided by the cock to the chicks (Chapter 10). Individual feeds did not last long. On average, each hen took only four minutes (range 1-8) to distribute prey of 5-15g (such as a Blue Tit), ten minutes (range 5-16) for prey of 16-30 g (such as a Chaffinch), 15 minutes (range 8-32) for prey of 60-100 g (such as a Blackbird), and 30 minutes (range 16-45) for bigger prey (such as a Lapwing). The larger carcasses took relatively longer to eat than the small ones, because they were pecked over for longer, as the young became full. After a feed, each hen swallowed any scraps remaining on the nest, cleaned her feet with her bill, and then wiped her bill on a branch, before settling down to brood. Fluffing out her breast feathers, she bent forward, edged over the young a bit at a time, and then settled down with side to side shuffling movements. As the young grew and produced more heat, the hens brooded them less and less, and after a time began to supplement the food-supply by doing some hunting themselves. In fact, over the nestling period as a whole, three phases in female behaviour were apparent, though the transition between them was gradual, and occurred at a different stage in each nest (Fig. 46). In the first phase, lasting up to 7 or 8 days, the hen spent almost all her time brooding, except for the periodic feeds. At this stage, the young were small and defenceless, and unable to keep themselves warm. In the second

192 The breeding cycle: parental care N EST

C>

N EST

A

B

40

c .....

:o~

~ c 20

""'c

(/)0

I

~

i

100

80 C>

.::

60

"0

o o

.0

40

20

L 10

I

I

15

I

20

25

II.

30

Nestling

10

15

20

days

Fig. 46. Brooding and guarding by hens at two nests in Ae Forest. Nest A was near a good food supply, and the hen spent more time at the nest. Nest B was further from a good food supply, and the hen spent less time at the nest, and more time away hunting. Observationsfrom day 3 at A andfrom day 5 at B.

phase, up to age 12-15 days, the hen brooded progressively less, and spent most of her time perched within 50 m of the nest. She continued to take prey from the cock and feed the young, but she also caught some prey herself, depending on whatever happened to be available nearby. At this stage, the young were able to call loudly if a predator approached, which immediately brought the hen to their defence. In the third phase, from about 16 days, the hen no longer brooded her young in the daytime, except during rain, but she left the nest area completely for part of the time to hunt elsewhere. By now the young were becoming feathered, and were better able to keep warm, and to defend themselves against predation by striking out with their feet, or later by fluttering from the nest if necessary. They were left by themselves for long periods and continued to call when we visited the nests, but the hen was often too far away to hear them. The hen continued to protect her young from rain until they left the nest, except during showers which fell when she was far away. In this third phase, the roles of the parents became less distinct. Both spent most of their time hunting, but if the hen was near the nest when the cock arrived with prey, she took it from him and fed the young herself. Otherwise he took the prey to the young, spending only a split second on the nest edge, and making no attempt to feed them. As the young saw him

25

The breeding cycle: parental care

193

coming, they rushed towards him and, in his haste to be gone, one cock sometimes did not place the item far enough onto the nest, so that it rolled off before the young could grab it. This cock did not retrieve the fallen food, which was therefore lost. Perhaps cocks behaved in this hurried way because they were smaller than their hungry daughters, making it unsafe for them to linger. In fact, the visits of cocks to large young were so brief that we could never get a photograph, even with the camera set ready. A similar reluctance of adults of both sexes to approach large young has been noted in Peregrines and other raptors, and several times young Peregrines were seen to grab their parents (Sherrod 1983). Food delivered by a cock in the hen's absence was usually eaten unaided by the young themselves. They were slow to begin with, and a backlog of prey from the cock sometimes accumulated on the nest before the hen returned and distributed it. For the last half of the nesting period, most prey were plucked only partly or not at all, especially when small. This gave the young the opportunities to take some roughage and to practice plucking prey themselves. Well-grown young ate the entire carcasses of small prey (apart from some feathers) but, like their parents, they often left the bigger bones from large prey. Our findings on the roles of cock and hen at different stages agreed with those of other observers, who watched nests in other parts of Britain and Europe (Owen 1916, Tinbergen 1946, Sulkava 1964, Opdam 1975, Geer 1981). Almost certainly, therefore, the recorded behaviour was usual for the species, and in most respects it was similar to that in other raptors (Newton 1979). While the cock Sparrowhawk acted as food provider throughout, the hen had several roles which changed in importance as the young grew, namely brooding and sheltering the young, tearing up carcasses and feeding the young, keeping the young clean by removing food scraps from their bodies and from the nest, protecting them from predators, and latterly hunting for food as well. Considering the size difference between the sexes, this seemed an appropriate division of parental duties, but some facets of the cock's behaviour were puzzling. Why, for example, did he remove the heads from prey in the early stages - were these the best bits that he wanted for himself, or were they difficult for the hen to feed to the chicks? Either way, heads could have formed the entire diet of the cock at this stage (Chapter 7). I have already suggested why the cock did not stay close to well-grown young, but why did he never feed smaller chicks? The hen was so intent on this herself that, until she began hunting away, the cock had little opportunity. According to Tinbergen (1940), at a nest where the hen was shot, the cock was indeed seen trying to feed the young. To begin with, he was so inept that the young would have died if the observer had not intervened, but after several days the cock began to master the technique. In my experience, whenever a hen was shot, the cock continued to dump prey on the nest, but did not feed the young. If the chicks were too small to feed themselves, they starved to death surrounded by piles offood. This confirmed the earlier findings of Owen (1916).

194 The breeding cycle: parental care Defence against predators was similar at the chick stage as at the egg stage, but tended to be more vigorous. This was perhaps because the parents had invested so much in the breeding attempt by then that they were prepared to put themselves at greater risk. At that stage they had no time to raise another brood in the same year, if the first was taken. In addition, the alarm calls and attacks of the parents on intruders may have helped the young to learn their enemies. As the female spent most time at the nest, most defence continued to fall on her, but if the male was present, he also mobbed predators.

Nest sanitation Until near the end of the nestling period, each hen paid great attention to nest hygiene. This had the twin function of keeping the young clean and of removing food scraps which might have attracted predators. As the nest sites were fairly inaccessible, Sparrowhawks could afford to be less particular about the disposal of waste products than some other birds could, and simply flung some unwanted items over the side, with the faeces. On the other hand, the hens took care to remove right away from the nest any unfinished carcasses which might have attracted martens or other climbing mammals, or flies. Such carcasses were cached away from the nest and brought back later. The down which flecked the nest during incubation was removed soon after the young had hatched, along with feathers from any prey which had been plucked there. Some of the down and feathers were swallowed, but most were thrown overboard. At first the hen also picked out the small white faeces of the chicks, which could yet not defaecate over the nest edge. From an early age the chicks regurgitated pellets, consisting largely of undigested feather remains. Initially the hen swallowed these, but as the young grew and produced larger pellets, she picked these up and flicked them clear of the nest. After most meals the hen examined the nest cup carefully, sometimes gently pushing the young aside to get a better look, and swallowing any food scraps remaining. Throughout the nestling period, each hen continued to add nest material. Perhaps this also helped with hygiene, covering any pellets, faeces or meat scraps that had slipped below the nest surface. Long thick twigs were put on the outside, and many short thin ones in the centre. The deep-cupped nest thus became flattened, and often several centimetres higher, by the end of the nestling period. At this stage, the young could preen and keep themselves clean and could also fly from the nest if disturbed. From then on the hens made less attempt to keep their nests clear of prey remains, which steadily accumulated along with the pellets which the young regurgitated once or twice a day. The young also lost a lot of down at this stage, so

once again the nest and adjacent twigs became flecked with white. This down was not cleared by the hen, and remained visible for weeks after the young had flown. In general, therefore, all forms of nest sanitation finished when the young could fly, even though they would continue to use the nest for 3-4 more weeks.

The breeding cycle: parental care

195

Behaviour in extreme weather During June and July, when young Sparrowhawks were in the nest, they were sometimes exposed to rain storms or to extremes of heat. This was particularly so when they were well grown and left alone while the hen hunted. However, if rain came on, the hen usually hurried back and pushed her way over the young. She stood with her back to the rain, crouched over the young, with her wings out from her body and depresse.d stiffly onto the nest. Her primaries were separated, and water drained down each one from the upper portions of her wings. She kept this up for hours, if necessary, and the chicks, too, did not stir for long periods. If a hen was slow in coming to the nest in a downpour, the young huddled together with their heads in the middle. When they got wet, the dampness extended initially only to the down and feathers on their backs. This dorsal plumage soon became matted, and then the water coursed off it quite freely. When the young got really soaked, they usually waited for half an hour or so after the rain had stopped before they attempted to preen themselves. The young could survive showers, but if they were left exposed to long periods of cold, pelting rain, the weaker young in some nests died, and in others the whole brood (Chapter 13). In heavy rain the hen herself became totally bedraggled, and as soon as it stopped, she left th.e nest to perch nearby, sunning herself with spread wings and tail, with steam rising from her plumage. In various positions, she would stand motionless for a minute or more, exposing herself to the breeze, and continually moving between sun and shade. When she had dried to some extent, she began to preen, paying particular attention to the flight feathers. She took an hour or more to dry completely. In hot sunny weather, if the nest was exposed, the hen also shaded the young by partly expanding and depressing her wings, to throw a big shadow. If her body was not in direct line from the sun to the young, she would gradually edge round the rim of the nest until it was. While the young were small, she would move slowly forward into the cup of the nest and crouch there so that the young were completely shaded. As the young grew, her behaviour changed, and with large young she simply turned her back to the sun, so that if a chick wished to make use of her shadow, it could. Usually only one did so at a time, but as the heat kept the young shuffling round the nest, the one in the shade was continually changing. The hen's endurance was amazing, and sometimes she would hardly alter her position for periods of 2-3 hours, panting all the time, until the cock arrived with food. These various observations were similar to those of Owen (1916), made 60 years previously, again suggesting that such weather-induced behaviour is general in Sparrowhawks. It has parallels in other species too. FEEDING RATES AND FOOD CONSUMPTION

In one sense, young Sparrowhawks - like their parents - have an ideal diet, as it approximates in composition to their own body tissues. All their

196 The breeding cycle: parental care 1500

A

B

1250

~

1000

'0

750

.e

...

III

E

to

~ n;

500

....

250

(5

.....'. '

.-..../

24

20

>-

Q)

a::

16

12

. . · A· · . I .-...1 .-./ \/. /\~ ....

.~

\

\

.-e-e

\

~\

16

Nestling

24

.

e

\

/\

e

days

liquid needs come cither from the body fluids of prcy or from fat, the metabolism of which yields an almost equal weight of water. Further, as the young can digest bone, they presumably have no shortage of calcium, which is often deficient for other birds. However, they cannot digest feathers, so this may represent the only nutritional imbalance they have to cope with. Feathers contain large amounts of sulphur amino-acids, so to grow their plumage the young have to eat excess other protein simply to get enough of these sulphur acids, but this is a problem shared by practically all birds. Otherwise, the main food stress that young Sparrowhawks face is shortage of total food, rather than dietary imbalance. As the parents bring only one prey item to the nest at a time, the total food intake of a brood depends on the frequency of deliveries and the size of individual items. The items probably vary little in composition. At the three nests studied from hides, prey deliveries averaged less than one per hour, with around ten per 16-hour day (Fig. 47). Pooling the records from different days revealed no diurnal rhythm, as items were brought at the same rate throughout the 16 hours, whether by cock or hen. However, at two nests, the cock slackened his rate of delivery part way through the nestling period. This coincided with increased hunting by the hen, so that, by the late nestling stage, the two parents contributed in total about the same number of prey items per day as the cock alone had brought previously.

But as the hen brought larger items, the food consumption by individual young increased. This was even more marked in broods depleted by mortality, as the food was shared by fewer young. Despite these average trends, the number of prey items (and the total weight of food) brought to each nest fluctuated greatly from one day to the next. Individual young thus experienced

32

The breeding cycle: parental care

J97

c

1\-, /I? j.\·

/

•

•

..

.-.. /\vJ\

'e-./ '.

\

16

24

.-. 32

Fig. 47. Food deliveries at three Sparrowhawk nests in Ae Forest. Observations started on day 3 at A, day 5 at B, and day 9 at C. The large amount oj food delivered every few days at B was due to the hen at this nest killing Woodpigeons.

widely varying rates of intake (Fig. 47). Part of this variation was linked with weather, and, at all nests, prey deliveries were depressed by up to onethird during periods of rain, compared to dry spells. This was probably due to the reduced hunting efficiency of the adults in wet weather, and in the late nestling stage also to the hen having to cover the young instead of hunting herself. As discussed in Chapter 13, the young grew less on wet days. The caching of uneaten items enabled food to be stored at times of glut for use at times of dearth, but such stored food provided no more than a small part of a day's needs. Prey delivery rates were not obviously influenced by the number of young in the nest. Deliveries were not more frequent to the largest brood than to the smallest; nor was there any reduction in delivery rates to broods depleted by mortality. Perhaps at our nests the parents hunted to full capacity throughout, irrespective of the number of mouths to feed. The same was later found by Perrins & Geer (1980) at several nests near Oxford. Probably all these pairs were hard-pushed, and if food had been more plentiful, they may well have varied their hunting rate to correspond with the needs of the young. Given enough time, noting prey deliveries was a simple matter, but it was less easy to assess the sizes of prey. From a hide placed several metres away, we could recognise only a fraction of the species brought in. Field guides were of little help for birds which were plucked and headless. We identified a few prey specifically from their feet, and others by quickly climbing the nest tree when the hen had left and inspecting the item before the young

could eat it. Mostly, however, we had to rely on estimating the sizes of prey from comparison with the hen's foot, and placing them in broad weight

198 The breeding cycle: parental care classes. It was at once apparent, not only that cock Sparrowhawks generally brought smaller items than did hens, but also that the individual cocks and hens themselves differed in the prey they brought. On average, the cock and the hen in the best habitat each brought smaller prey than the cocks and hens in the poorer habitat. Perhaps the latter took more high-risk prey, or were hunting further afield, so brought back chiefly the more profitable larger items, eating the small ones themselves. The estimated total food consumption of our three broods (including the mother's) during 32 days was about 5·9 kg at nest A, 13·8 kg at nest B, and 5·9 kg at nest C. Paradoxically, the growth rates of these broods were in the order A> B > C, but the trend in consumption differed, and B received most food in the late nestling period, when growth had nearly finished. The hen at this nest brought some exceptionally large items, including pigeons. On days when large prey were not brought (including the first 12 days), the daily intake of brood B was usually less than that of other broods (Fig. 47). The overall consumption of this brood increased from start to finish, while those ofother broods (and of the individual young) were highest around the middle of the nestling period, which fitted their sigmoid growth pattern (Chapter 13). In general, the weight gain of individual young in all nests during the period of linear growth correlated with the amount of food eaten (Fig. 48). Food consumption has been estimated in other studies, but only for a captive pair was it measured through a whole breeding cycle (Hurrell 1973). This pair raised four young and were fed as much as they wanted, mainly day-old chicks from a hatchery. Daily intake approximately doubled to 250 g after hatch, then increased to reach 450 g by day 26, and fell slightly once the young had left the nest. The total consumption over the nestling period was about 5·5 kg, and over the whole breeding season about 11·5 kg, including the needs of both parents. Observations on other wild broods with 3-4 young gave similar results to ours, with total intake in the first 32 days ranging between less than 6 kg in some broods to more than 13 kg in others (Meissel 1937, Tinbergen 1946, Geer 1981). Noting such variation, Tinbergen (1946) commented that Sparrowhawks might sometimes 'overfeed' their young, and give them more than needed for mere growth and maintenance. The young might then simply pass the food through more quickly and digest it less thoroughly; they might also lay down body fat, to serve as a reserve for wet days, or for the post-fledging period. Not only total consumption differed between the various broods studied, but the trend in consumption over the nestling period. Both aspects were probably influenced by prey availability, and did not necessarily reflect the changing needs of the young. In Hurrell's captive pair, consumption increased during the nestling period as a result ofincreased feeding frequency. Similarly, prey deliveries in the wild pairs watched by Tinbergen (1946) and Holstein (1950) increased from 6-7 per day in the first week to 9-10 in the last week. At our nests, deliveries stayed at the higher level throughout, and any increase in consumption was due to the hen bringing larger items than the cock.

The breeding cycle: parental care 110

199

o 0'

'0 9 ~

«S

a:

o 8~----....,...----~-----r--------,

10

15

20

25

Mean number of food pieces per hour

Fig. 48. Relationship between growth rate andfood consumption in individual nestlings, both measured over the linear growth period. Growth rates (weight) are expressed in relation to the mean values (100) for each sex. • - male, 0 - female. Weights were taken daily, and food consumption was assessed by counting the number of food pieces swallowed by each young in standard watches. Significance of relationship: r = O· 76, P< 0·05. The same was true of the Oxford broods, in which prey deliveries were up to ten per day, yet total in take doubled or trebled during the time the young were in the nest (Geer 1981).

Foodconsumption ofthe sexes As female nestlings gained weight more rapidly than males, it was ofinterest to compare the food consumption of the two. To distinguish the chicks individually, we marked them on hatching with a spot from a marker pen, each with a different colour. Then to assess food consumption, we counted the number of food pieces taken by each chick during watches throughout the nestling period. To our surprise, at each nest, male and female young ate the same number of food pieces, and this was true at all stages of growth. No indication was obtained that hens took larger pieces than cocks, and increased their intake in this way. The pieces did become larger as the young grew, but at anyone period they were similar in size for all the young. Thus, the marked weight difference which developed between male and female young in all three broods was associated with no difference in food intake between the sexes. However, whereas females had a faster weight gain, males developed more rapidly in both plumage and behaviour (Chapter 13). In other words, the sexes utilised the same food difTeren tly between the various aspects of nestling growth, at least initially. Partitioning offood From other observations at the Ae Forest nests, we learnt how young were affected by feeding routines. Food was shared more evenly among the young when the hen fed them than when they fed themselves. When they

200 The breeding cycle: parental care were small, the hen would point her head towards whichever individual was receiving food from her at the time, but periodically she changed her stance, to enable poorly positioned young to reach her beak. Only the last hatched young, smaller and hidden beneath the rest, often missed out. When the young were larger, each piece of food was bolted by the first young to grab it; the young thus competed over individual pieces, but no one young got them all. On the other hand, when the mother did not feed large young, the first to grab a prey item when it was dropped onto the nest usually ate most of it, leaving little for its siblings. The size of the prey had no influence on whether the hen partitioned it, but when she did not, it was usually only from large carcasses that all the young could feed. In other words, participation by the hen at feeds meant more even shares for all the young. In some birds of prey, fighting among nest mates is usual, and often the latest hatched dies from attacks by its larger siblings (Newton 1979). In the Sparrowhawks which we studied, aggression was rare; it occurred only between young which were hungry, and most often developed when the young were feeding themselves. I t was thus most frequent at nests Band C in poor habitat, where the hen seldom distributed the prey herself. At nest A, in good habitat, the hen continued to feed the young on food she had caught herself right to the end of the nestling period, and thus ensured more even shares among her brood. The main reason why the hens in poor habitat seldom partitioned prey was that they spent more time hunting, and left immediately after each delivery. There may have been more to it than this, however, and further observation might clarify matters. Look at it this way. Before the young could tear up food for themselves, the hen had no option but to distribute it; but later she had the choice of either distributing it, or of dumping it on the nest for the young to eat themselves. Distribution by the hen led to more equal sharing among the young, whereas dumping often resulted in fighting and in one or more young going hungry. Thus the best tactic for the hen, if food were plentiful, might be to distribute it herself to discourage runting, but if food was scarce, it might be better to dump it and encourage runting, so that the brood was more rapidly trimmed to an appropriate level. In this way, less food would be wasted on young that would die anyway. I do not wish to imply that the hen made a conscious decision, but that her reaction to the prevailing food supply favoured runting only when conditions were appropriate. The cock always dumped prey, but if the hen was present she often distributed it, and thus still controlled the supply. The failure of the cock to partition food may be partly to allow more time for hunting, and partly, as mentioned, because of the risk to him in feeding daughters

up to twice his weight.

Comparisons between nests That the growth and survival of young varied with parental behaviour was evident from comparison of different broods, and of the same broods

The breeding cycle: parental care

201

at different stages. Good even growth rates in all the young of a brood were associated with: (a) frequent deliveries of relatively small prey items; (b) minimum depression of deliveries during rain; (c) good attentiveness by the hen, including almost continuous brooding in the early nestling phase, and presence near the nest to hunt in the late nestling phase; (d) distribution of prey by the hen until the young fledged, ensuring more equal shares among them; and (e) minimum aggression between young and, when feeding themselves, the ready surrender of prey from one young to another. In contrast, poor growth and survival were associated with the opposite conditions, namely: (a) infrequent deliveries of prey, including some very large items; (b) marked depression of deliveries during rain; (c) minimal brooding by the hen in the early stages and her frequent absence on prolonged hunting trips in the late stages; (d) the young feeding themselves as soon as they were able on items dumped on the nest, resulting in unequal shares between them, and (e) frequent fighting among nest-mates over food. Comparison with other nests showed that the frequent absence of the hen on remote hunting trips had other consequences, namely the occasional chilling and death of young in rain storms, and predation by Tawny Owls, neither of which happened when the hen was on hand to protect her young (Chapter 13). One might attribute some of these differences to variations in maternal behaviour as such, some hens being 'good mothers' and others 'bad mothers'. This may have been involved to some extent, but the fact that, at our three nests, maternal behaviour varied with the food supply, suggested that food supply influenced the quality of parental care. The more the cock was able to meet the food needs of the family, the more the hen could stay with the young. Her behaviour in the late nestling period was thus a compromise between the need to see the young fed adequately, which entailed absences, and the need to protect them. The hen clearly required to assess the food needs of the chicks at all stages, and either communicate these needs to the cock or begin hunting herself. When the chicks were.small, they indicated their hunger by faint begging cries and by pecking at their mother's bill, and she in turn called to the cock with a drawn out 'kee-oo' note, indistinguishable to my ears from the food-begging call given earlier in the cycle. A hen with chicks gave this call during long inactive periods of brooding and at food transfers. Schnell (1958) heard a similar call from a hen Goshawk which had chicks, and named it the 'dismissal scream', because he thought that it encouraged the male to leave on another hunt. For much of the nestling period, the hen Sparrowhawk thus acted as an intermediary between the young and the cock, and apparently modified her behaviour according to circumstances. At one of our nests, where the cock maintained his rate ofprey delivery, the hen stayed near the nest and provided extremely little food, but at the other nests where the cocks reduced their rates of delivery, the hens contributed increasingly to total food supply. I had no way of knowing why these cocks began to bring less food part way

202 The breeding cycle: parental care through the nestling period, but the same happened at several nests observed by Geer (1981) near Oxford, and in this area it coincided with known reductions in the availability of prey suitable for cocks. Other factors, such as loss of body weight and start of moult, may also have been involved. As in south Scotland, however, this decline in provisioning by the cock resulted in increased hunting by the hen, and less protection for the young. At a nest watched by Owen (1916) in a rich broad-leaved wood, the hen stayed with the young, even in the post-fledging period. At this nest, the cock evidently continued on his own to meet the food needs of the family, enabling the hen to stay and guard the young, an arrangement of clear benefit to both parents.

Observations at othernests Further information on female behaviour was obtained during daily visits to other nests, when Dorian Moss (1979) noted each time he saw or heard the hen. At nests with newly hatched young, the hen was seen on 90% of observer visits, but with the increasing age of the young, the hen was seen progressively less often. This decline was first apparent around the mid nestling period, when the hens began hunting, and was more marked in the poor food area of Ae Forest than in the Annan Valley (Fig. 49). Thus, with young aged 20-23 days, the hen was seen on only 32% of observer nest visits in Ae Forest, compared with 66% of nest visits in the Annan Valley. In addition, the young showed the best growth at nests where the hen was most often present, and the poorest growth at nests where the hen was most often absent. These findings on a large n urn ber of nes ts were consis tent with the detailed observations at three nests, and could be interpreted as follows: where feeding conditions were good, the cock could provide most of the prey required for raising the brood, supplemented if necessary by the hen's hunting near the nest. The hen was thus present and observed on most observer visits. In addition, the growth rates of young were high. By contrast, in poor areas, the cock could not supply enough food, and the hen had to hunt for much of the time, once the young were old enough to be left unbrooded. Places where prey were sufficiently abundant were at some distance from the nest, so the hen was observed infrequently at observer visits. This was again supported by radio-tracking, which showed that hens at remote forest sites regularly flew up to 9 km from the nest for prey, while those nesting in or near valley woodlands spent more time at the nest, and hunted only up to 2-3 km (Chapter 5). Further support for these ideas came from the feeding experiment mentioned earlier (Chapter 13). When extra food was provided daily at two nests in the remote part of Ae Forest where prey were scarce, the females spent most of their time at the nest, until after the young had left. The young themselves showed better growth and survival than usual in this area, as four flew from one nest and five from the other. There was little doubt, therefore, that food supply influenced the quality of parental care.

The breeding cycle: parental care 203 100

•

•

o Volley

0

• Forest

0

80 0

~ ~

u

c

0 -0 C

60

•

0

•

0

•

0

0

••

•

~

0

o

6

Q)

L

rn c =3

0 >.

4

"+-

0

L

CD ..0

2

E ~

z

0

U 1

2

2

3

2

Rank ord e r

Fig. 51.

3

4

of

young

2

3

4

5

Post-fledging survival of )'oung in relation to rank within broods. Taking

~ccount of the sex difference, young are ranked by fledging weight in broods of 2-5 respect-

ioely. Includes only broods of more than one young from which at least one young was recovered. No broods of6young represented.

POST-FLEDGING SURVIVAL

Survival after fledging was examined from individuals ringed as nestlings and later recovered (i.e. found and reported). Any bird reported more than one month after leaving the nest, whether dead or alive, was assumed to have survived the post-fledging period, and any bird recovered after April of the following year was assumed to have reached breeding age. Some birds were reported by members of the public through the National Ringing Scheme, while others were trapped as breeders at nests in the study areas. For analysis, it was necessary to separate the sexes, because females were more likely to be recovered than males, regardless of their survival (see Chapter 23).

Survivalin relation tofledging weight The Sparrowhawk broods studied by Dorian Moss gave a sample of 140 ringed young whose weights were known on days 19-21, close to the age of peak weight. The weights of these young were enormously variable, ranging in cocks 118-180g and in hens 160-284g. Yet no trend was apparent for the heavier young to yield more recoveries than the lighter ones (Fig. 51). Some very light-weight young were reported months after becoming independent, and occasional ones bred in the study areas. Similarly, when the young in individual broods were ranked according to their weight at fledging, allowing for the sex difference, the lighter young in a brood were as likely to be recovered subsequently as were the heavy ones (Fig. 51). There was a slight tendency, in broods of five, for heavy young to yield more recoveries

212

The breeding cycle: post-fledging period

than lighter ones, but the trend was not significant statistically. had no clear evidence, either in the population as a whole or in broods, that the weight of young Sparrowhawks near fledging their subsequent survival. In this respect, Sparrowhawks differed other species which have been studied (Perrins 1980).

Hence we individual influenced from some

Survival in relation to otheraspects ojearlylife Young from smaller broods were recovered in slightly greater proportion than young from large broods, but in neither sex was the trend significant statistically (Table 30). In other words, no marked variation in the subsequent survival of young was found that could be related to brood size at fledging. The implication was that, the larger the brood, up to six, the greater the number of eventual survivors. In contrast, the sex composition of broods seemed to affect the proportion of individuals recovered. Males from all-male broods were recovered in slightly greater proportion than males from two-sex broods, and females from all female broods were recovered in slightly greater proportion than females from two-sex broods. For males this difference was significant in the largest broods (5 young), and for females in the overall sample for all brood sizes combined (Table 31). It seemed, therefore, that the presence of the opposite sex in a brood lowered the survival chances of the other, especially in large broods. I have no idea why this should be so, and possibly it was a chance result. These various trends were apparent among the earliest autumn recoveries each year, as well as among the subsequent ones. This implied that mortality associated with brood size and composition was largely over by the time that any recoveries came in. It must therefore have occurred within a few weeks after fledging, possibly while the young were still dependent on their parents. Combining all sizes of broods, relatively more recoveries were obtained from broods raised in lowland habitats than from those raised in upland, and relatively more recoveries were obtained from broods reared by adult mothers than from those reared by yearlings. Both trends were apparent among the earliest autumn recoveries each year, as well as among subsequent ones. So, like mortality associated with brood composition, that associated with habitat and maternal age occurred during the first few weeks after fledging, possibly again while the young were still dependent on their parents. In another respect, however, the first-winter and subsequent recoveries showed different patterns (Table 32). Relatively more of the late-fledged young were recovered in their first winter, and relatively more of the earlyfledged young were recovered subsequently. This implied that, in general, young fledged early in the season survived longer than those fledged later, many of which died in their first winter. It was mainly the young from the early part of the season which lived long enough to enter the breeding population. The difference in trend between the two sets of recoveries was statistically significant in both sexes, and implied that mortality associated with fledging date occurred, not in the post-fledging period, but later in life. Likely reasons

The breeding cycle: post-fledging period

213

for the trend were not far to seek because, after becoming independent, early young had a longer period of food abundance in late summer, during which to gain experience in hunting, before prey populations began to decline; they were also able to establish themselves in suitable habitat weeks before the latest young tried to do the same. I t was not surprising, therefore, that more of the late young died in their first few months of independent life.

SUMMARY

After their first flight, at about four weeks of age, young Sparrowhawks remained in the nest vicinity for another 21-31 days, fed by their parents. During this period the young completed their growth, and developed the flying skills necessary to catch prey. At the end of the period, the young dispersed, in at least some cases probably following a reduction of food provision by the parents. From then on, the young had to feed themselves, unless they could attach themselves temporarily to another, still dependent, brood. No relationship was apparent between the weights of young before fledging and their subsequent survival, or between the number in the brood and subsequent survival, but males in all-male broods and females in all-female broods survived slightly better after fledging than others of their sex in mixedsex broods. In general, young of both sexes from lowland habitats or from adult mothers survived better than young from upland habitats or from yearling mothers. These survival differences were manifest during the post-fledging dependency period or soon after. In addition, young reared early in the season survived, on average, longer than those reared later, a difference manifest mainly during the first winter of life.

CHAPTER 16

Seasonal trend in breeding success

The start of egg laying among Sparrowhawks in south Scotland was spread over a 4-6 week period, between late April and early June each year. On average, laying begain up to ten days earlier in warm dry springs than in cold wet ones, a few days earlier on low ground than on high, and a few days earlier with adult hens than with yearling hens. These variations were slight, however, compared with the seasonal spread within years, habitats or age groups , so examination of the spread itself was important. This was the more so, because laying date was linked with breeding success , and the earliest pairs to lay each year were very much more productive than the latest (Newton & Marquiss 1984). This raised two questions : first why did nest success decline with advance in laying date; and second, since the early pairs did best, why did any pairs breed late? One might have expected that natural selection would have favoured early breeding by all pairs.

Performance and laying date To begin with , I examined breeding performance in relation to laying date by pooling all the data from different years, habitats and age groups . Four aspects of performance were included , namely clutch size, brood sizes at hatching and at fledging, and the proportion of clutches which produced young. All these aspects declined significantly with advance in laying date (Fig. 52, Table 33). The later a clutch was laid, th e smaller it was and the less its chance of producing young. Between late April and late May , the period over which most clutches were started, their mean size declined from 5·3 to 3'1, and the mean brood-size at fledging in these nests declined from 3·7 to 2·3. In early nests , the difference between the initial clutch size

Seasonal trend in breeding success

215

and the final brood size was 1·6, and in late nests it was 0·8. This was equivalent to losses of 30% and 26% on initial clutch sizes, implying no greater loss of eggs and young from successful late nests than from successful early ones. In other words, the seasonal decline in brood sizes could be largely attributed to the decline in initial clutch sizes, and egg and chick mortality was no more frequent among late broods than among early ones. Turning to complete failures, between late April and late May, the proportion of clutches which produced young declined from 91 % to 38%, and then to 20% for nests begun after 30 May. This trend clearly contributed more to the seasonal decline in overall productivity than did the declines in clutch and brood sizes. Overall, Sparrowhawks starting in late April produced an average of 3·4 young per clutch, whereas those starting in late May produced only 0·9. There was thus a four-fold difference in production associated with laying date. For most of the season the decline in performance was fairly steady (Fig. 52). When pairs failed, we could usually tell the immediate cause from signs left at the nest (Chapter 17). From various causes identified, only clutch desertions increased through the season, while the other types of failure declined in proportion. Desertions were caused mainly by birds suddenly ending the breding attempt and abandoning their eggs, rather than by their being

6

•

Q)

N CJ)

l=

4

•

s: o

• •

s:

~ CJ)

,:,

:=l

c

«S

4

Q)

N

~ o

6

«S

0 0

2

.0

Q)

c

~

«S Q)

~

•

Q)

OJ "0 Q)

~ Q)

N

10 :=l

~

6

::J

CJ)

CJ)

o

U;

-0

CJ)

o

.D

• •

CJ)

Q)

g

3

Q)

c

1

*-

C «S

•

2

Q)

~

April

May

June Date

that

April

laying

May

began

Fig. 52. Seasonal trend in breeding performance. The points are based on average values for successive 5-day periods. All aspects ofperformance declined significantfy with advance in laying date. For statistical details) see Table 33.

June

216 Seasonal trend in breeding success killed by predators, because many of the individuals concerned were retrapped at a later date. In most cases, desertion occurred within a few days after clutch completion, and before midJune (Fig. 53). These overall trends emerged from the pooling of all nest records, but they also held in individual years, in different habitats and in females of different ages (Newton & Marquiss 1984). In other words, the seasonal decline in performance was a general feature of Sparrowhawk breeding, which held regardless of year, habitat or age group. In addition, the later the population as a whole bred in a given year, the poorer the production of young (expressed as mean young raised per clutch started, Fig. 54). In fact, mean laying dates accounted for 30% of the annual variation in production over the study period. As mean laying dates were in turn related to April rainfall and temperature (Chapter 10), so too was the production of young. Some 46% of the variation in production between years was linked with variation in the number of rain-days in April, and 26% with variation in the mean April temperatures. Hence, a sequential pattern emerged, starting with spring weather, which influenced laying dates and then breeding success. The relationship between spring weather and production held, probably, because in cold wet springs the prey species themselves were scarcer, and started breeding later, than in warm dry springs. This would have led to a later and less abundant supply of fledglings for Sparrowhawks to feed on.

Post-fledging survival The above findings showed a decline in breeding success with advance in laying date, so that pairs laying early had a clear advantage over those laying late. However, it was important to check that poor production in late nests was not offset by improved survival of young after fledging, or by improved survival of parents. If the lifetime production of early and late laying parents was similar, there would be no strong advantage in Sparrowhawks laying at any particular date in the season. The post-fledging survival of young was examined in Chapter 15, with the help of ringing recoveries. It was shown that, on average, young from early nests survived longer than young from late nests, and that early young were more likely to reach an age at which they could breed. These findings were statistically significant in both sexes. They thus lent no support to the idea that the poor performance of late nests was compensated by improved post-fledging survival of the young. Rather they indicated a continuation of trends at the nest stage, with early young surviving best. This result was not unexpected because, after becoming independent, the early young had a longer period with abundant food in which to learn to hunt than did later ones, and may also have occupied the better habitats before the later birds came along.

Parental survival The notion to be checked was whether those parent Sparrowhawks which

Seasonal trend in breeding success

217

Layi ng per iod


0-

J 16J

•

]

]

J '0 12J J CD

•

L-.

.0

E ::l

z

] ] 8J

] J ] 4J ]

]

J

°

7 Longevity

in

8

9

10

years

Fig. 61. Lifetime production offemales in relation to theirlongevity. Each spot represents the number of fledglings raised by an individual female. Significance of relationship: b = 1'46, r= 0'67, P< 0.001. From Newton 1985.

above, in which 36% of yearling cocks and 7% of older cocks paired with yearling hens while 53 % of yearling hens and 13 % of older hens paired with yearling cocks. Thus the average age of mate was higher for cocks than for hens, in line with the later start and greater longevity of hens.

Proximatecauses ojvariations in lifetime production The main source of variation in lifetime production was female longevity (Fig. 61). Lifetime productions of up to five young were recorded from females which survived for only one breeding season, productions up to ten young were recorded for birds which survived up to three or four years; while all productions greater than ten young were from those which survived more than four years. The three productions greater than 20 young were for birds which survived for 6-8 years. Within females of anyone age group, however, variation in productivity was great, and some individuals up to eight years old raised no more than did other individuals which lived to only one year. The one individual which survived for ten years produced only 12 young. Other factors, besides longevity, which influenced the lifetime production of young, included the proportion of a lifetime over which breeding occurred (in turn dependent mainly on age of first breeding), the number of eggs laid at each breeding attempt, and the proportion of these eggs which gave

242

Age and breeding

rise to fledglings. It was this last factor which was most important, after longevity, in influencing lifetime production. As Sparrowhawks raised up to six young in a brood, individuals which lived for ten years could in theory produce 60 young in their lifetime. Only individuals which lived for one year produced close to the maximum possible for their age, and with increasing age of bird the difference between actual and potential production widened. This might have been expected, for while an individual might by chance achieve maximum reproduction in anyone year, it became progressively less likely to continue to achieve this the more years that it bred. Not until the fifth year of life had all birds in the sample produced young.

Ultimate causes ofvariation in lifetime production I attempted to relate the individual variations in lifetime production of fledglings to features of individual anatomy, previous history and habitat. The only useful anatomical feature which I measured was wing length, with the feathers flattened and straightened on the rule. This value varied among females between 226 mm and 247 mm, but within individuals was fairly consistent (±3 mm) beyond the first year of life. Longevity was significantly correlated with wing length (n= 113, b=0·09, r=0·20, P 0·7). (b and c) Relationship between Annandale population density in late summer (expressed as % of mean) and median dispersal distance of cocks (b) and hens (c) in differentyears. No significant relationship for eithersex (cocks, r = 0·17, P< 0·7, hens, r = 0·12, P< 0·8). FromNewton & Marquiss 1983.

that they had produced. This figure gave an underestimate, because it took no account of non-breeding birds, but it was probably comparable from year to year. In anyone year, only small numbers of recoveries were obtained, and at widely varying distances from the birthsite; the annual variations in recorded distances were not statistically significant. The population ofAnnandale declined during the ten years of study, while that of Eskdale fluctuated irregularly. In neither area were correlations apparent between population density and dispersal distances, and in Annandale no significant change in the dispersal distances of either sex was evident over the years, as numbers declined (Fig. 70). Moreover, the years when many cocks moved long distances were not always those when hens did likewise. I t seemed, therefore, that Sparrowhawks did not disperse further in years when their numbers were high. However, density may have combined with some unmeasured factor, such as food supply, in influencing dispersal distances. As was the case with Goshawks in northern Europe (Sulkava 1964), Sparrowhawks in Britain may well have moved further in years when food was scarce. In part, the sexes ate different prey species, so years with long movements by males would not necessarily coincide with years with long movements by females. As I had no measure of food supply in late summer, this question remained unresolved. The idea that long-distance birds were 'inferior' was hard to test directly, but if so, one would expect that (a) they might have experienced different conditions during their early life from short-distance birds, or that (b) they might subsequently survive or breed less well. Regarding early experience, for each young recovered, I examined the natal habitat, the grade of natal territory, the natal brood size, and the fledging date in the season, all in relation to the subsequent dispersal distance. Only one significant trait emerged: birds born in upland habitats dispersed further in general than

160

264

Dispersal

did birds in lowland. This held both for cocks and hens, and was evident among birds recovered dead and among retrapped birds (Fig. 71). It gave some ground for suspecting that the early experience of individuals influenced their dispersal distances. None of the other aspects examined even approached statistical significance. The lack of relationship between fledging date and dispersal distance was surprising, for I had expected that young from late broods would disperse further than young from early broods, but in fact they did not differ. For the second point, (b) above, on subsequent performance, I examined the dispersal distances of individuals in relation to their breeding habitat, grade of nesting territory, laying date of first egg, clutch size, brood size near fledging, and age at death. The latter could be examined only for birds reported dead by members of the public, because many of the birds which we trapped ourselves were still alive at the end of the study. But all other aspects could be examined only in the trapped birds, for I had no data on breeding in the majority of birds reported dead. Breeding performance was of course influenced not only by the birds recovered, but also by their respective mates. Birds which had moved long distances seemed, on average, to do less well in all respects than did birds which had moved short distances. This tendency was statistically significant only in hens, and only with respect to habitat, laying date and clutch size (Table 49). Thus birds which moved long distances from their birthsites subsequently bred in poorer habitat (upland), laid later in the year and produced smaller clutches. These aspects were probably inter-related, because the later and smaller clutches could have resulted from the poorer habitat. Nonetheless, the findings gave some grounds for linking long-distance dispersal with poor performance, and hence for thinking that long-distance birds were 'inferior' in some way compared with short-distance birds.

DISPERSAL AND LACK OF INBREEDING

One of the suggested functions of dispersal is to reduce the risk of inbreeding, to spatially separate young from their parents and brothers from their sisters, and thus reduce the chance of close relatives pairing. The harmful effects of inbreeding have been known for years in domestic animals and humans, but have only recently been demonstrated in a wild bird (Harvey et al 1976). In all these species, matings between close relatives produced fewer viable offspring than did normal matings. Thus natural selection might be expected to favour behaviour which reduced inbreeding, including dispersal itself and, perhaps, also a difference in dispersal distances between the sexes. It was of interest, therefore, to assess in Sparrowhawks the effectiveness of dispersal in separating close relatives. Almost certainly, spatial separation was more important than individual recognition in preventing inbreeding, for while brood-mates might recognise one another in later life, they could

Dispersal 4

265

Upland (N=20)

2

--.180

0 (f)

.x: U

0

u

6 Lowland (N =45)

4

--.185

2 0 4

Upland (N=35)

--.265

2 0 (f)

c

(l)

I

6 4

2 0

10

20

30

40

50

60

70

80

90

100

110

Distance (krn)

Fig. 71. Distances between hatch.place and recovery place for cocks and hens (all ages) raised in upland and lowland habitats respectively. Based on recoveries provided b] members of the public. In general, within each habitat, hens dispersed further than cocks; and within each sex, birds raised in upland dispersed further than those raised in lowland. For statistical analysis, geometric mean distances were calculated and compared using t tests. Comparing cocks and hens: lowland, t lO7 = 3·52, P < 0·001; upland, t28 = 1·43, P < 0·2. Comparing lowland and upland: cocks, t63 = 2·12, P < 0·05; hens, t97 = 2·18, P < 0·05. From Newton & Marquiss 1983.

not be expected to recognise brothers and sisters raised in other years, whom they had never known. Most Sparrowhawks, once they had bred, kept to the same nesting territories, or moved only short distances from one territory to another in subsequent years (Chapter 21). Hence, the initial dispersal distances (in Figs. 69 and 70) were similar to those that separated offspring from their parents by the time the offspring started to breed. Considering the concentration of recoveries (especially of cocks) in the vicinity of the birthplace, birds might by chance quite often breed on the very nesting places where they were born. In fact, however, of 32 cocks raised in the study areas and later found breeding on 33 territories (one bird occupied two territories in different years), not one bred on its natal territory, and only three bred in adjacent territories. Similarly, of 67 hens raised in the study areas and later found breeding on 75 territories (several occupied differen t territories in different years), only one bred on its natal territory, but only after first breeding on another territory 5 km away. Two other hens bred on adjacent territories. In most cases, failure to settle on the natal territory could not be attributed to the parent still being present, and keeping out the young. With a 34%

120

266

Dispersal

annual mortality, and frequent movements of adults between territories, the majority of birds were present on particular territories for only one year, and the average bird for only 1·5 years (Chapter 21). Because most young did not breed until their second year or later, their parents would usually have vacated the original territory by then. Perhaps, therefore, Sparrowhawks actively avoided breeding on their natal territories, the combination of avoidance and dispersal accounting for the virtual lack of such records. We were certain that none of the 32 males of known history paired with their mothers, sisters or daughters, and of 29 occasions when mates of the 67 females were identified, none was a father, brother or son. We also had no cases of birds which shared a known father or mother from different years subsequently pairing together, or of birds reared on the same territory in different years pairing together. In other words, no cases of breeding with close relatives were found, even among hawks which bred near to their own birthplace. Taking these findings as a whole, therefore, the dispersal system ofSparrowhawks gave generally good spatial separation between closely related individuals. There was little opportunity for inbreeding to occur, and no recorded cases.

Comparison between siblings From some broods, two or more young were subsequently recovered. To my surprise, the distances moved by nest-mates were significantly correlated, the young from some nests making short movements and those from other nests long movements. This was apparent both among recoveries provided by members of the public and among our own re-traps in the study areas (Fig. 72). It was not generated by the response to habitat, mentioned above, because the majority of records - including the long-distance ones - were from lowland young. Yet from none of the broods concerned did one of the young make a much shorter or longer movement than its siblings. So whatever factors influenced the dispersal ofyoung, they probably affected whole broods. However, the sex difference was still apparent, and in all cases where the reported siblings were of opposite sex, the cock made the shorter journey. On the other hand, there was no consistency in directions within broods, and recoveries of some nest-mates were from opposite points of the compass. This confirmed that (a) the young in a brood travelled independently of one another, rather than as a group, and that (b) landscape features around the birthplace were not paramount in influencing directions. The distances and directions moved by siblings determined the distances that separated them when they were recovered. Such distances ranged between 0 km, where brood-mates were found on the same territory in different years, and 270 km. The two cases of brood-mates on the same territory concerned two brothers in different years, and a brother and sister in different years. These were the only such cases recorded, and otherwise we had no instances of brothers and sisters breeding even in adjacent territories, let alone in the same ones.

Dispersal 267 100

]

(0 )

75

C

•

Q)

E Q) >

0

50

a

E

0

Q)

a

( b) 0

in .c

30

25

25

(j)

0

t?

.~~ 0

.....t

.00

25

50

75

100

125

150

175

200 0

,

15

Longes t movement (k m )

Fig. 72. Distances moved by siblings between hatchsite and recovery site: (a) recoveries from members of the public: (b) recoveries from birds trapped at nests in the study areas. In each case, the shortest movement (y axis) was plotted in relation to the longest (x axis), and in both cases, movements of siblings were significantly correlated (see Table 50). e, cock-cock: 0, hen-hen: ~, cock-hen. From Newton & Marquiss 1983.

Comparison between offspring andparents The similarity between the movements of brood-mates could have been due to the effects of inheritance or early experience. Either factor could have caused siblings to be more similar to one another than to young from different nests. To explore this aspect further, I compared the dispersal distances of young from the same mother in different years (usually the father was not known), and ofoffspring with their parents (Table 50). In the first comparison, significant relationships were found, with young from some mothers moving short distances and from other mothers long distances (Fig. 72). This too could have been due to either inheritance or to early experience, because most mothers used the same nesting places in different years, or nearby places in the same habitat. In the comparison between parents and offspring, the relationship was weak and not statistically significant. I twas based on only nine parent-offspring pairs, however, so I could not rule out inheritance on this result, especially as parents and offspring had sometimes been raised in different habitats, which would have reduced the correlation.

CONCLUDING REMARKS

The findings as a whole suggested that early experience was one factor which influenced the distances moved by young Sparrowhawks. Success in obtaining food was probably involved, and in turn depended on the hunting skill of the individual, the availability of prey in the local environment, and the extent to which good areas were already occupied. Thus the longer movements of young from uplands were associated with a relative scarcity of prey there, especially after late summer. On average, therefore, young. raised in

30

268

Dispersal

upland may have had to move further to find a place to live, compared with young raised in lowland. A food-linked dispersal was consistent with the finding that, later in life, long-distance birds bred less well than shortdistance ones, for both the long movement and the poor breeding could have been manifestations of poor hunting skill or poor social status. I twas also consistent with the similarities in dispersal distances among siblings, which could have resulted from similar experience and abilities. So, while both genetic and environmental factors probably influenced dispersal, the latter - the particular conditions experienced by individual birds in their first few weeks - probably accounted for some of the variation in recorded distances. I should stress, however, that while the main findings were suggestive of this view, crucial evidence on hunting success and other aspects was lacking. The idea that the sex difference in dispersal helped to avoid inbreeding did not explain why females moved furthest and not males. However, female Sparrowhawks also had larger day-to-day ranges than males (Chapter 5). In fact the home range diameters for cocks and hens (1'5 and 4· 5 km in low ground) were not dissimilar in ratio to the median dispersal distances (14 and 27 km respectively), or to the median distances moved by adults between nes ting places in differen t years (0'8 and 1·5 km). Thus a male: female ratio in the order of 1: 2 or 1: 3 held for all types of movement made by Sparrowhawks in south Scotland. Sparrowhawks normally fly from one tree or patch of cover to another, so inevitably make longer flights in open country than in woodland. This could be partly responsible for the longer movements of hens, which hunt in open country more than cocks do. Among Goshawks in northern Europe, moreover, it is the males that move further than females and the males that spend most time in open country (Haukioja & Haukioja 1971, Marcstrom & Kenward 1981). Habitat cannot account for the whole difference, however, for a sex difference in movement patterns occurs in other birds, including species in which male and female show the same use of habitats. So far, I have considered dispersal in terms of the individual, but what is the net effect of the process on the population? On a dictionary definition, disperse means 'to scatter in all directions'. The term is clearly appropriate to the members of individual Sparrowhawk broods, and to some extent to the juvenile population as a whole, which spreads over the countryside in the late summer, occupying areas previously deserted. During the winter, birds may disappear from some of these areas, and the population as a whole again becomes more confined. In the process, however, there has been a re-distribution of young birds, and a mixing of individuals from different broods and from different localities. And this process takes place year after year with each crop of young. The process is somewhat different in migratory Sparrowhawks from northern Europe, where the young make a long southern journey in their first autumn; however, most of those which survive still return to breed in the general region of their birth, bu t again mixing with young from other broods and from other localities.

Dispersal

269

SUMMARY

To judge from ring recoveries, most British Sparrowhawks settled to breed within 20 km of where they were raised, and only a minority beyond 50 km. Dispersal occurred soon after the young had become independent, in AugustSeptember, and in any direction from the nest. Very few individuals made long movements in later life, and no migration was apparent. Of birds raised in the study areas, hens dispersed further than cocks, with median (and maximum) distances from birthplace of 27 (and 265) km for hens and 14 (and 185) km for cocks. These distances could be compared with a minimum between nesting places 0[0·4 km. Young raised in upland habitats tended to move further than young of the same sex raised in lowland. Also, young which moved long distances bred less well in later life than did young which moved short distances (significant only in hens). In distance, movements by brood-mates were highly correlated, the young from some nests making short movements and from other nests long movements (but the sex difference was still maintained). Distances moved by young of the same mother in different years were also correlated, but not the distances moved by parents and offspring. Experience in early life, especially hunting success, was probably important in influencing the distances moved by individuals. Dispersal gave good spatial separation of closely-related individuals, and no parent-offspring or brother-sister matings were recorded.

CHAPTER 21

Territory and mate fidelity

Once they had bred, some Sparrowhawks stayed on the same nesting territories with the same mates in subsequent years, while others changed their territories or mates . This chapter is concerned with the factors that promote fidelity to territory and mate , and with those that lead to changes. It is based on information obtained in south Scotland from the repeated trapping of individuals at as many nests each year as possible. Throughout the term 'territory' is synonymous with ' nesting place', and refers to that part of the male's range which is used for nesting each year, and from which other breeders are excluded .

TURNOVER OF TERRITORY OCCUPANTS

On territories where occupants were identified in successive years , no less than 57% of cocks and 50% of hens had changed between one year and the next (Fig. 73). This remarkably high turnover gave average periods of residence on territories ofonly 1·4 and 1·5 years respectively. So in this species, whose nesting places were used over several decades , individual birds stayed for only short periods. The fewer records from territories wh er e birds were

Territory and mate fidelity

271

caught at intervals of 2-3 years gave roughly similar estimates of turnover. Evidently the long-term use of territories in the Sparrowhawk was the result of many different individuals occupying the same territories in quick succession , but each staying for a relatively short time. On low-grade territories , changes of hens were more frequent than on high-grade ones (Table 51). On low-grade territories 63% of hens changed between one year and the next, giving a mean residence period of 1·3 years , while on high-grade territories th e equivalent figures were 43% and 1·6 years. Even on the best territories, residence periods were often short, and the more continuous occupation of such territories was due largely to different hens occupying them in quick succession without gaps . Some such territories were occupied in ten successive years, but by a different hen each year. The cocks showed no difference in turnover between territory grades, but too few were identified on low-grade territories to examine the point properly . In neither sex was there any significant variation in turnover between years . The figures above give only th e mean turnover of birds from year-to-year changes, but many individuals were identified in several different years, enabling their full tenure periods to be calculated (Fig. 74) . The findings confirmed the shortness of the periods involved . The majority of birds retained the same territories for merely a year or two , and only occasional individuals stayed for up to four consecutive years (cocks) or six years (hens). Two other cocks had been present for four and five years when the study ended , so their residence periods may have been even longer. Similarly, three other hens had be en present for six years, and two others for at least five years, when the study ended. On these records, a Sparrowhawk which began to breed in its first year and lived to the average age for a breeder (2' 7 years) would occupy two territories in its breeding life. The slightly reduced figure for males was probably due to the small sample, and also perhaps to a shorter mean lifespan (Chapter 23).

60

Mal es

I N = 5 81

Fe m a l e s

(N =2 9 61

z:

'"

.; 40 >

o 20

Fig. 73. Turno ver of occupants on par ticular nesting territories. The data were from territories where occupants were trapped in successive years , and indicate the proportions of territories where the bird was the same or diffirent in the second year. On 57% of territories the male had changed, and on 50% thefemale had changed.

272

Territory and mate fidelity 50

MALES 25 -

17 5 rJ)

F E MAL ES

I1l ::J

D

150

> D C

.-

12 5

0

ClJ

10 0

.0

E ::J

Z

75

50

25

_

1

2

R e si de n c e

_

3

~=~~~

llilf@l)§]""","'-."

4

period s

5

6

( y e a r s)

Fig. 74. Full residence periods of individuals on nesting territories. Most birds were present for onlyayear or two, but others for severalyears. The only other information on residence periods that I have come across for Sparrowhawks is based on eggs collected over several years from what were obviously the same females . These clutches, taken in Worcestershire in the early years of this century by Mr G . Tomkinson, indicate that one hen stayed on the same territory for six years, two for seven years, two for eight years and one for nine. The same collection contained many series collected over fewer years, which may not have covered the full residence periods of the hens involved. However, to find this number of long periods in such a sample suggested generally longer tenure periods in this earlier study than in mine. Possibly th e average lifespan was longer then .

Territory and mate fidelity

273

CHANGES OF TERRITORY

High annual turnover in the occupants of particular territories was due partly to deaths and partly to movements. With a 34% annual mortality, one would expect in a stable population that this proportion of places would come up for re-occupation each year. But some surviving birds changed territories between years, thus adding to the turnover. Of males which were identified in more than one year, 78% were on the same territory both years, compared with 69% of re-trapped hens, the remainder of each sex having moved to different territories. These figures suggested that cocks were more sedentary than hens, but the difference was not statistically significant. Both sexes more often changed territories if they had failed in their breeding the previous year than if they had succeeded, a trait significant only in hens (Table 52). This tendency was especially marked between the first and second year of life in hens, and became less marked with age (Table 53). Older hens showed a stronger tendency to stay on the same territories, whether successful or not the year before. Among cocks, the number of one-year-olds in the sample was small, but among older ones the same trend held as in hens, as they became more resident with age. Territory grade also influenced the movements. Both yearling and adult hens tended to stay on high-grade territories after a successful breeding, but to move away from low-grade territories, whether successful or not the previous year. This was especially true of yearlings which, as just mentioned, moved more than adults. On the only occasions when yearlings stayed on the same territory, it was ofhigh quality and the previous attempt was successful (six out of seven birds stayed in these circumstances). From poorer territories, yearlings always moved (ten birds), whether successful or not, and from high-grade territories they always moved after a failure (four birds). Too few cocks were identified to examine their behaviour with respect to territory grade. To summarise, most Sparrowhawks which survived from the previous year used the same nesting territory. Changes of territory were most frequent among hens which failed the previous year, which were yearlings the previous year, or which were on poor territories the previous year. Thus site fidelity was related to the age and previous success of the bird, and to the quality of the territory. Distances moved In recording movements between territories, the study areas themselves imposed constraints, because any birds which moved outside would have been missed. The two areas had maximum north-south and east-west dimensions of 40 X 20 km and 20 X 12 km respectively, and were 15 km apart at their nearest points. The longest movement that could have been recorded was 120 km, if a bird had moved from the corner of one area to the opposite corner of the other. The shortest was 0·4 km because this was the closest distance between nesting places.

274 Territory and matefidelity Many birds made only this minimum move, and beyond this, progressively longer moves became steadily less frequent (Fig. 75). In general, hens which changed territories moved further than cocks. The median distance moved by hens was 1·5 km, and the maximum was 27 km, whereas the median distance moved by cocks was 0·8 km, and the maximum 19 km. Most of such movements were short compared to those that could have been recorded, and only one bird moved between study areas. In both sexes, movements became shorter with increasing age (Table 53). So not only were old birds less inclined to change territory than young ones, when they did change, they moved less far. No difference in recorded distances was apparent between birds which succeeded the previous year and those of the same sex which failed, nor between upland and lowland habitats. On this last point, however, findings were not strictly comparable. The distances moved by individuals were affected by the local landscape, and as most woods in the uplands were larger, with more territories, than those in lowland, there was more opportunity in upland for short moves. In lowland, birds more often had to cross wide stretches of open country to reach neighbouring territories. Without this difference in landscape, movements may well have been longer in the upland, for within the woods themselves the nesting places were further apart. In both habitats, most of the movements were to adjacent territories, either in the same or in a different wood; some were to the next territory but one, while others were further afield. Cocks more often moved to adjacent territories in the same wood, and less often to other woods, than did hens. These overall trends were well illustrated by events in one wood where for five years we obtained an unusually complete picture of what went on. The wood held four nesting territories. Over the years these were occupied by more than seven different cocks and by more than nine different hens. During their periods in the wood, each cock kept to the same terri tory for breeding, except for one bird (present for four years) which moved to a neighbouring territory for one year, and then back again the next. The hens moved around much more, and the four that were in the wood for more than two years each used 2-3 different nesting territories in that time, mating each year with a different cock. At least one of these hens had previously bred in a nearby wood. The remaining five hens were in the main wood for only one year each. Moreover, in other parts of the study area, four hens were known to return to a terri tory they had used in a previous year after breeding for one or more years in a different territory. In the interim, their original territory had been used by a different hen. We naturally wondered whether our trapping increased the amount of movement. I believe that this was most unlikely, because movements were just as frequent among birds that were identified on their feathers, as they were among trapped birds. Secondly, when traps were deliberately left near nests, some individuals entered them repeatedly, apparently not greatly upset by the experience. Further, there was no obvious reason why trapping should have accounted for the differences in moves found between territory grades and age groups.

Territory and matefidelity

275

COCKS ~18·5

I ..a '+-

o

"-

(1)

.c

E ::::>

Z

6

HENS

4 ~16·5 ~27·0

o

4

6

7

8

9

10

Oi stance (km)

Fig. 75. Distances moved by Sparroiohatoks between nesting territories used in different years. In general, hens moved further than cocks, with median distances of 1·5 km and 0·8 km respectively. No territories werecloser than 0·4 km.

Othercauses ofterritory changes Other factors, besides those considered, which might have caused birds to change territory included: (J) being ousted from the original territory by another bird of the same sex; (2) an opportunity to move to another territory which had been vacated; or (3) the lack of a suitable mate, if the original mate died or was replaced by a less desirable one. I t was impossible to tell from observation whether any birds were ousted from territories but, if this were common, one would expect to find that a large proportion of terri tories vacated by one bird (' the loser') would have been occupied by another ('the winner'). In fact, for cocks only two (29 % ) out of seven such territories, and for hens only 29 (43 % ) out of 67 such territories, were found to hold a new bird in the second year, the rest remaining vacant. So in the majority of cases displacement seemed an unlikely reason for change, though it may have happened occasionally. If birds often moved to territories from which they had previously been excluded by other individuals, most movements should have been to territories that were occupied the year before. This was in fact the case in 90% of moves by cocks and 85% of moves by hens, a far greater proportion than expected from occupancy levels in general. Hence, it could be argued, such birds might well have occupied their new territories in an earlier year, had such territories been available then. On this view, competition was important in excluding birds from favoured territories, and its removal stimulated movements into such territories. Regarding the role of mate loss in territory changes, suffice to say that, in territories which were occupied in successive year-s, movements by neither

276 Territory and matefidelity sex were associated with the loss from the territory of the original mate (Newton & Marquiss 1982). Mate loss thus seemed of little importance in stimulating moves, providing that a lost mate was replaced by another individual. Summarising, while some Sparrowhawks may have been ousted from territories by others, this seems not to have been a major factor in causing individuals to change territories. On the other hand, many birds may have been kept from territories in certain years by the mere presence ofother individuals there, moving to these territories when they became vacant.

Consequences ofchanging territory The majority of birds which moved did in fact acquire a better territory. Of the moves made by cocks, only one was to a territory of lower grade than the previous one, two were to territories of similar grade, and four were to territories of higher grade. Of the moves made by hens, 19 were to territories of lower grade than the previous one, 14 were to territories of the same grade, and 34 were to territories of higher grade. Thus no less than 73 % of all recorded moves were to territories of similar or higher grade. These figures were hard to test statistically because we could not say what range of territories was available to each individual. But if it was assumed for sake of argument that birds had equal chances of moving to lower, similar or higher-grade territories, then the tendency towards higher-grade territories emerged in both sexes, though samples were large enough to show statistical significance only in hens (X 2 = g.7, P < 0·0 1). Birds which changed territory may also have obtained a better mate in the process, but how could onejudge mate quality? As Sparrowhawks tended to breed best at 4-7 years old (Chapter 18), the best mate for anyone breeding attempt could be considered as the one nearest this peak-age range. For each bird which moved, the choice assumed was between the new mate and the mate it might have got that year if it had stayed on the territory of the previous year. For only six birds (two cocks and four hens) was this comparison possible, and five obtained 'better' mates after their move and one obtained a 'worse' mate. These results were suggestive, but as the numbers were small and age perhaps not the best measure of quality, I could draw no firm conclusions on this point. Changes in territory and mate would have been of particular value if they led to improved nest success. It was of course not possible to tell whether particular birds which changed territories bred better than they would have done if they had not changed. I could only compare their performance on their new territories with that of other birds which stayed, and with their own performance in the previous year. In the first comparison, hens which changed territory after failing the previous year did no better (56% successful) than those which stayed on the same territory after failing the previous year (70 % successful) . Similarly, among previously successful hens 59% were successful again after changing territory, compared with 71 % which remained on the same territory. In each comparison, it seemed that birds which changed territory did less well than birds which stayed, but in neither case was the

Territory and mate fidelity

277

difference statistically significant. The records gave broadly similar conclusions when analysed separately for poor and good territories. In the second comparison, hens which changed territory did no better, on average, than they themselves had done in the previous year (of 57 birds which changed territories, 32 raised young in the season before their move and 33 after it). Insufficient records were available for a similar analysis on cocks. So to summarise, many Sparrowhawks which changed territory seemed to have obtained better territories or mates, but neither tendency could be firmly established on the facts available; hens which moved did not breed more successfully than others which stayed on the same territory, nor than they themselves had done the year before. But neither of these last comparisons could tell us how the birds which moved would have performed if they had not moved. It seemed likely, however, that most birds which changed territory did not 'lose out' in these respects as a result of the move. They also may have survived differently as a result of a move, but this was impossible to judge.

MATE FIDELITY

Only small numbers of cocks were caught on nesting territories in more than one year. On most such occasions, the hen was identified as well, so it was possible to assess fidelity to mate, according to whether the cock had retained or changed his territory. All possible combinations were recorded, in which a cock: (a) stayed on the same territory with the same mate (13 cases); (b) stayed on the same territory with a different mate (14 cases); (c) changed terri tory and kept the same mate (one case); or (d) changed terri tory and found a differen t mate (11 cases) . Moreover, in some cases where a cock had changed mates, the original hen was known to be alive and breeding elsewhere with a different cock; and in other cases the original partner was known to be dead. A freq uen t pattern (13 cases) was for cocks to retain both terri tory and mate from one year to the next. Four years was the longest period that a pair was known to stay together in this way. The one ins tance of a pair staying together over a change in territory entailed a move of only 0·6 km in the same wood; such behaviour was probably infrequent, as I had plenty of opportunity to record it if it were common. In other words, fidelity to mate was usually associated with fidelity to territory, and a change of territory usually entailed a change of mate. In some cases where the cock changed territory, the hen stayed on the original territory and re-mated there, and in other cases the hen also moved and re-mated. As a result both of movements and deaths, many birds which stayed on the same territory for more than a year or two inevitably had more than one partner during their time there.

Up to three mates were recorded for an individual cock, and up to four for a hen, but the number may at times have been even greater.

278

Territory and mate fidelity PREVIOUS EXPERIENCE AND BREEDING SUCCESS

A common pattern was for surviving Sparrowhawks of either sex to retain both territory and mate from the year before. I therefore attempted to find whether previous experience of territory or mate influenced breeding success, and hence whether there was any advantage in birds remaining faithful in these respects from year to year. One might expect that a pair which had bred together on the same territory in previous years would do better than a pair which were new to one another or to the territory. Excluding the one pair which changed territory and kept together, there were three types of breeding attempt for comparison: (a) both partners had been together on the same territory the previous year; (b) one partner had been on that territory the previous year, but the other partner was new; and (c) both partners were new, both to the territory and to one another. On investigation, there did seem to be some relationship between experience and breeding performance, as birds with previous experience of territory and mate (category a) did better than those with previous experience of territory alone (b), and these in turn did better than birds without previous experience of either territory or mate (c) (Table 54). However, in some of the pairs in category (b), the new partner was a yearling, while in some of those in (c) one or both partners were yearlings. Such young birds may have performed less well than older ones through paucity of all kinds of experience, rather than merely of a particular territory and partner. However, when yearlings were excluded from the comparison, and only adults were considered, the difference between the three categories of pairs was still apparent, though only in laying dates was it significant (Table 54). Another explanation of these findings was possible, other than birds benefiting from experience, namely that it was on the best territories that both birds tended to stay, and on the poor territories that one or both tended to change. The difference in breeding performance between categories could then have resulted from different quality territories, rather than from differing degrees of experience. Categories (a) and (b) birds did in fact tend to be on slightly better territories than category (c). In conclusion, therefore, although previous experience of mate or territory seemed to be beneficial to breeding performance, this result could equally be explained on other grounds. I thus had no firm evidence that Sparrowhawk partners gained by remaining together from year to year; moreover, if they did sta y together, their gains in breeding success were small. In view of this, it was perhaps not surprising that changes of territory and mate were frequent.

TERRITORY CHANGES AND FOOD SUPPLY

When we fixed radio-transmitters to individuals, it became apparent that each nesting place lay within the range of an established cock, and that the number of nesting territories within anyone wood corresponded to the

Territory and mate fidelity

279

number of established cocks which lived there (Chapter 5). For most of the year, such cocks overlapped in foraging range with their immediate neighbours, but only seldom did they venture beyond their neighbour's range into the next. On the other hand, hens hunted more widely, greatly overlapping in range not only with other hens, but also with several different cocks. In good habitat, individual cocks seldom had ranges more than 1·5 km across, while hens on average had ranges 4·5 km across; in poor habitat both sexes ranged over correspondingly larger areas. Thus most of the moves between nesting places made by cocks or hens could have been within areas already familiar to them, as part oftheir existing ranges, the longer movements of hens reflecting their larger ranges. In fact, one radio-tagged hen was found to shift during the nest-building period of one year from the nest of one cock to the nest of another within her own range. Another moved in the second year to nest in what had been her main hunting area the year before when she bred with a different cock 7 km away. Some of the short distance 'territory changes' by hens may thus have depended partly on where in their range they found a receptive and unmated cock in spring. It perhaps mattered little if the part of their range where they bred the year before had already been taken by a different hen, especially if the advantages of re-pairing with the same cock were not great, as results indicated. On the other hand, a few birds of both sexes moved such long distances that they must have totally changed their hunting ranges between one year and the next. These longer movements were perhaps the result of local conditions, which caused the birds concerned to shift further than usual, and settle in a completely different area the following spring. As birds were likely to behave in ways that could maintain or improve breeding success from year to year, it was not surprising to find that they were more likely to change territory after a failure than after a success, and to shift from a poor territory than from a good one. However, the birds need not have made 'conscious decisions' on the basis of experience, as almost all the traits observed could have resulted from the same response to food supply. As radio-tracking showed, the better the food supply the more sedentary Sparrowhawks became, and the worse the food supply the more wideranging they became (Chapter 5). This response may have accounted for the longer residence on good territories than on poor ones, and similarly for the greater mate fidelity on good territories, neither partner needing to move away where food supply was good. For any given prey supply, yearlings may have had greater difficulty than adults in obtaining their food needs, which may in part have accounted for the more frequent and longer movements recorded from yearlings. The correlation between nest failure and tendency to change territory could similarly be explained in terms of local conditions, both the failure and the move being caused by poor food supply. Such a response to food would, on balance, result in many birds acquiring better territories (and with them better mates) during the course of their lives. I t would also result in vacancies in good feeding areas being more quickly filled than those in poor areas, leading

280

Territory and mate fidelity

to more continuous occupation of good habitat and to higher-grade territories there. Thus the main findings could mostly be explained as incidental consequences of a single response to food, and need not necessarily have entailed the birds making decisions on the basis of other factors, such as fidelity to mate or previous nest success. No doubt competition for the better places also contributed to the patterns that emerged, as it meant that many young birds occupied poorer habitat and moved to better places later in life. The greater residence of cocks than of hens may have had some additional basis, however. I t is widespread in other kinds of birds and in other birds of prey (Baker 1978, Greenwood 1980, Newton 1979), and is perhaps linked with the cocks holding the territories. Among Sparrowhawks, cocks were apparently the prime defenders of the nesting places, and - through their home ranges - of the food supply necessary to support the female and young during the breeding cycle (Chapter 11). Thus the advantages of continued residency were probably greater for cocks than for hens, as familiarity with the area may have helped in both defence and foraging. In addition, cock Sparrowhawks had greater mortality than hens (Chapter 22), leading to a surplus of hens in the population, which in turn may have caused hens to move around more in search of suitable mates.

SUMMARY

The turnover of Sparrowhawks on individual nesting territories in south Scotland was high, with 57% of cocks and 50% of hens changing between one year and the next. Mean residence periods were 1·4 and 1·5 years respectively, but some cocks stayed on the same territories for up to five years and some hens for up to six. Turnover was more rapid on poor territories than on good ones. It was due mainly to mortality (higher in cocks), but partly to movements (more frequent in hens). Changes of territory were most frequent among young birds, and among birds which were on poor territories or which failed in their breeding the previous year. Such birds did not move far, with median distances of 0·8 km (maximum 19 km) for cocks and 1·5 km (maximum 27 km) for hens. Birds of both sexes tended to move to better territories than their previous ones. Birds which stayed on the same territory in successive years often kept the same mate, whereas birds which changed territories also changed mates. The main findings could be interpreted in terms of a single known response to food supply, together with the greater land-holding role of the cocks and competition in both sexes for the better habitat.

CHAPTER 22

Migration

As explained in the previous chapters , Sparrowhawks which breed in Britain are relativel y sedentary. Aft er a period of po st-fl edging dispersal, most individuals settle, and remain in th e sa me ge ne ral a rea throughout th eir lives . They co n tras t with Sparrowhawks whi ch breed furth er north a nd eas t in Europe, wh er e th e w in te rs a re m ore seve re. These birds perform a regular migration , leaving th eir breeding areas a nd movin g ge ne ra lly so u th-wes t in autumn a nd returning north-ea st in sp ring . Some of th ese con tine n ta l hawks sp end th e winter in Britain , a ugme n ting the local population , while ot hers pass through on migrati on . Our kn owl ed ge of th e mov em ents of th ese birds is based chiefly on ring recov eries , but also on observations ofbirds on passage. Righ t ac ross th e wh ole Eurasian ran ge of th e Sparrowhawk, th e northernm ost populations a re migratory. The birds from som e reg ions move no more than a few hundred kilometres, but th ose from other regions ma y m ove more than 2,000 km. In winter, birds occupy th e so u the rn parts of th e breeding range, a nd a lso extend fur th er so u th into Afri ca and th e southern parts of Asi a . At thi s sea so n th ey a re found not only in woo de d a nd cultivat ed are a s, but al so in sa van na h and desert, wh er ever th er e a re tr ees or bush es .

282

Migration

They are little known in Africa, perhaps partly through confusion with 0 resident species, but are found commonly in the Sudan, south to 11 north, and occasionally in Ethiopia, Chad, Kenya and Tanzania (Moreau 1972). The west Palearctic birds winter primarily within Europe, the Mediterranean islands and northernmost Africa, and the birds which reach south of the Sahara are probably from the central Palearctic, though no ring recoveries are available to check this view. So far as is known, none cross the Equator. The bulk of the birds wintering in India and south-east Asia are probably of the Siberian race, A. n. nisosimilis.

MIGRATION AND FOOD SUPPLIES

The seasonal movements of Sparrowhawks, like those of other birds, relate to seasonal changes in food supplies. As one moves from southwest to northeas t in Europe, the winters become progressively harsher, and an increasing proportion of the small bird population, on which Sparrowhawks depend, moves out for the winter. The same is true of Sparrowhawks themselves, which are completely resident in the south and west of Europe, including Britain, and completely migratory in the extreme north and east, while in between some birds stay while others leave (partial migration). Thus the extent to which the hawk population of anyone region withdraws for the winter largely corresponds with the degree of reduction in food supplies. The net result is a redistribution of Sparrowhawks twice each year, which corresponds with the movements of prey species, giving a fairly uniform distribution over the Continent in summer, and a concentration in the southern and western parts for the winter. I t is in these latter areas that many prey species concentrate, including various finches, thrushes and Starlings. In general, Sparrowhawk populations maintain the same west-east distribution on wintering areas as they do on breeding areas, through parallel migrations. On average, Norwegian and Swedish birds winter further west in Europe than do Finnish birds, and Finnish birds winter mainly west of Russian ones, and so on (Fig. 76). This is more apparent in the middle latitudes of Europe than further south, where birds from a wide span of longitude inevitably become funnelled into Spain, Italy and Greece. Thus the ring recoveries in Spain in winter include Sparrowhawks from a large part of Europe, including the Soviet Baltic states, Finland, Sweden, Denmark, Germany, Poland, Czechoslovakia, Switzerland, Holland and Belgium (Bernis 1966). A curious aspect of the migration, both of Sparrowhawks and of some other species, is that, while individuals from the north come to mid-latitudes in winter, those that breed there move further south. In Germany and Poland, for example, Sparrowhawks occur all year, but there is partial replacement of populations each spring and autumn. This raises the question, if immigrants can survive in such regions in winter, why do many of the breeders leave? Perhaps it is partly a question of adaptation and competition, with

Migration

283

)

Fig. 76. Migration of Sparroiohmoks from different areas to show parallel movement patterns. Arrows show centres ofbreeding and wintering areas, as indicated by ring recoveries, and spots show trapping places for migrants. From Michelson & Viksne 1982.

the northern birds being competitively superior in mid-latitudes in winter, and the birds that breed there being superior in summer. The timing of Sparrowhawk migration is also linked with food supply, for in effect the hawks accompany their prey south-westwards in autumn and back again in spring. The precise relationship between the passage of Sparrowhawks over the Courland Spit, in the south Baltic, and the passage of their main prey species, is shown in Fig. 77. At this locality, in both spring and autumn, Sparrowhawks pass through at exactly the same periods as their prey species. Moreover, annual variations in the timing and duration of Sparrowhawk passage were correlated with similar variations in the prey passage, and extra Sparrowhawk movements in some years coincided with extra movements by Crossbills. Sex and age diffirences In some partial migrant populations, age and sex differences are apparent in the proportion of individuals which migrate, in the dates when they travel, and in the locations of wintering areas. In general, a greater proportion ofjuvenile Sparrowhawks than adults moves out for the winter. The juveniles

284

Migration

011

~

30

co

-0

20

~

10

o

~ 123123123123123123123123123

M

A

M

A

s

o

3

1

2

N

3

1

2

3

o

Fig. 77. Migration of Sparrowhawks over the Courland Spit in the Baltic in relation to the migrations of their main prey species. Shown as a percentage of all birds trapped in spring and autumn respectiuely, (1) thick line - Sparrowhawks; (2) thin line- Chaffinch Fringilla coelebs; (3) broken line - Brambling F. montifringilla. Redrawnfrom Belopolski 1971.

depart slightly earlier than the adults in the autumn, and return slightly later than the adults in spring; in consequence the adults spend longer on the breeding areas each year, and the juveniles longer on the wintering areas. This might be interpreted in terms of food supply, for as prey stocks decline in autumn, the juveniles are likely to become short before the more experienced adults, and similarly as prey stocks increase in spring, the adults are likely to be able to survive on the breeding areas earlier than the majority of juveniles. Within the sexes, the adults are presumably also dominant to the juveniles, so might gain priority at areas of good habitat. Within age groups, sex differences have been noted in the timing of passage, but these were slight and not consistent between sites, at least in the autumn (Table 55). From Sparrowhawks ringed on migration at the Courland Spit, on average the males in winter were recovered further south-west than were females, with mean distances of 1,328 km and 927 km respectively. The wintering areas of the two sexes evidently overlapped extensively, but did not exactly coincide (Belopolski 1971). The same was apparent for Sparrowhawks ringed on migration through Heligoland, but not for those caught at Signildskar in Finland (Moritz & Vauk 1976, Saurola 1981). As the sexes differ in size, and take different prey, they might be expected to winter in partly different areas. But there are no prey data to check for such a correlation, or to explore why the sexes migrate different distances in certain populations and not in others.

INHERITANCE OF MIGRATION PATTERNS

As mentioned above, the predominant migration direction for populations breeding across the western Palearctic is approximately north-east/southwest. To investigate the navigation skills of Sparrowhawks, Drost (1938) did a large-scale transplantation experiment. He trapped 209 Sparrowhawks that were on autumn migration at Heligoland, in Germany, and some days

Migration 285 later released them at a different site, some 600 km to the east-southeast. From among these birds, 36 recoveries were subsequently obtained. They showed that the young birds continued to migrate from the release point in the same direction which they would normally have taken from Heligoland. They thus followed a course roughly parallel to the normal, and wintered outside the range appropriate to their particular population. Although recoveries in the following year were naturally fewer, the experimental displacement was maintained on the breeding area. This implied the existence of an innate directional tendency. In addition, the recoveries ofdisplaced birds extended for a similar distance from the release point as did those for control birds from the trapping point. The young birds thus seemed initially equipped at least for a simple bearingand-distance flight to a hitherto unknown area, being capable, with no previous experience, of migrating in the right direction for the right distance. Provided no unusual circumstances were encountered, this facility alone would enable the birds to shuttle back and forth repeatedly between their summer and winter homes. The displacement of older birds, with experience of at least one return migration, produced different results. Instead of sticking to the standard direction, older birds tended to return towards their normal wintering and breeding areas, indicating a more advanced form of navigation which enabled them to correct for displacement. I t implied the existence in the birds of a map sense, enabling precise orientation. The young birds probably had this same sense too, but with no experience of a previous wintering area, they had no reason to take other than their inherently programmed direction. The results of this experiment were in all respects consistent with those from similar, but larger scale, experiments on other species (Matthews 1968), so the Sparrowhawk appeared no different to them in its navigation abilities.

PREPARATION AND MODE OF MIGRATION

Although the ultimate reason for migration is avoidance offood shortage, hunger need not necessarily stimulate the flight. Some songbird species have been found to respond to daylength changes, reaching migratory condition in the autumn before food becomes scarce, and while there is still time to accumulate fat for the journey. In spring, moreover, the food supply may actually be increasing when the birds set off. The situation presents a parallel with the timing of breeding, discussed in Chapter 10, the winter food shortage being the ultimate factor through which migration has evolved, and daylength-change the main proximate factor stimulating flight at appropriate dates in autumn and spring. As research has been mainly on songbirds, little is known of the proximate control ofmigration in birds-of-prey. However, in some such species, food shortage in autumn must act as a proximate, as well as an ultimate factor influencing the flight (Newton 1979). This is because more birds leave the breeding areas in years when food is scarce

286

Migration

than in years when it is abundant, the Goshawk being an obvious example (Mueller et alI977). In raptors, as in other birds, the situation seems most complex in populations which show partial migration, where some birds stay while others leave. Two possibilities suggest themselves. Either some birds are programmed to migrate, while others are not; or all birds are programmed to migrate only when they, as individuals, meet certain conditions. The latter seems more likely, for it could better account for why a greater percentage of birds leave in some years than in others, and why some individuals migrate in their first year but not subsequently. These ideas are speculative, however, and no information is available on the proximate control of migration in Sparrowhawks. In contrast to some songbirds, migrant Sparrowhawks seem not to put on extra fat for the journey, at least to judge from their weights. Individuals caught on passage in autumn and spring were not noticeably heavy, and in fact they were within the weight range of non-migrant British birds caught at the same season (Moritz & Vauk 1976). They probably carried some fat at this weight, but not a large amount. Extra fattening is probably not necessary, however, for Sparrowhawks can easily obtain food on the way. They have often been seen catching songbirds while migrating (Rudebeck 1950), and when trapped on migration, individuals often had food in their crops. Some Sharp-shinned Hawks, which were tracked by radio while on migration through the United States, were found to hunt for periods each day, mainly before and after the main flight, and to move up to several hundred kilometres in a day, resting at night (Cochran 1972). When on migration, Sparrowhawks tend to concentrate at short sea-crossings, but not to the same extent as the larger soaring raptors, such as eagles and buzzards. This is because Sparrowhawks migrate more by active flapping flight and although they take advantage of any updraughts and thermals available, they are not dependent on them. In general, the migration is broadfront, and what concentrations do occur seem to result merely from birds in the area taking advantage of local topographical features, rather than whole populations favouring particular routes. Well-known concentration points include Falsterbo in south Sweden, where more than 10,000 Sparrowhawks have been counted in recent autumns (Wallin 1984). On the other hand, Sparrowhawks are only poorly represented at migration points further south, such as Gibraltar (less than 1,000), the Bosphorus (less than 200), the eastern Black Sea (less than 700), Eilat in Israel (less than 200) and Suez (less than 50). This is presumably in part because the majority of west Palearctic Sparrowhawks winter within Europe and a relatively small proportion venture into Africa. Much more impressive concentrations than those for the Sparrowhawk occur in the Levant Sparrowhawk, which migrates in flocks. This species breeds mainly in south-east Europe (Chapter 2), and the whole population probably passes through Israel every autumn and spring, en route to Africa. Around 24,000 individuals were counted over Eilat in spring 1977, and around

Migration

287

23,000 over Kafer Kassem in autumn 1981, but presumably these formed only part of the total (Christensen 1977, Leshem 1984).

POPULATION TRENDS

Counting migrant raptors at concentration points has become a popular pastime for hundreds of birdwatchers. Although such counts give no more than a minimum estimate of the numbers migrating, when repeated over many years at the same site, they can give some idea of population trends. The figures in anyone season might be influenced by weather and other prevailing conditions, so little significance can be attached to changes from one year to the next, only to those that occur over a longer period. Counts at Ottenby and Falsterbo in Sweden indicated a marked depression in the passage of Sparrowhawks during the 1950s and 1960s, which coincided with the main period when organochlorine pesticides were used in Europe (Fig. 78). In this respect, the Sparrowhawk resembled the Peregrine and Merlin at these sites, while other species, such as the Kestrel, which was known to be less affected by these pesticides, showed hardly any decline. The birds passing over anyone point are of course drawn from a wide area, so by counting the migrants year after year, an observer can monitor population trends over a far larger area than would be possible from the same number of hours spent on breeding grounds.

200

•

>.

ct5

"'0

~

Q)

160

.,

c.

en c ~

120

0>

~.

E

'0 Q;

80

.0

E

::J C

Q) C)

40

/.

:

\i \t\....-•

•

,

~

..

-.-.-.

•

.! \i •

Q)

>

0

o Q)

CJ)

HIT WINDOW

:1

\:)

.o \:)

Q) 0)

c

0

:l

HIT VEHICLE

'FOUND

10

DEAD'

~

5

o

1931-40

1941-50

1951-60

1961-70

1971-80

Fig. 79. Proportions of British-ringed Sparrowhawks that were recovered in various ways in different time periods. Only those types of recovery which have changed over the years are shown. The decline in shot birds followed protective legislation in 1961; the increase in window casualtiesfollowed the rise in popularity ofplate glass; the increase in road casualtiesfollowed the rise in fast traffic; and the peak in numbers 'found dead' coincided with the main period of organochlorine pesticide use. For details of other British ringing recoveries, see Table 46.

were also often thin). Haemorrhaged birds showed internal bleeding around the brain, lungs or foregut, not obviously resulting from impact; they often had high levels ofHEOD or DDE in their tissues, so were probably pesticide victims. The same was true for many other individuals, in which the cause of death was classed, on post-mortem, as unknown. The proportion of haemorrhaged birds recorded in this scheme declined after 1975, when the uses of aldrin and dieldrin pesticides in agriculture were greatly curtailed. Otherwise, the frequency of different types of casualties varied little between years, seasons, sex or age groups. Recoveries of ringed birds provide less information than post-mortems, because the finders say only how they came to acquire the birds. Many birds are merely reported as 'found dead'. Nonetheless the British recoveries show some interesting points, especially on the changes in different types

296

Mortality

of mortality over the years (Fig. 79, Table 59). The scheme started in 1908, since when the proportion of Sparrowhawks recovered by any means has fluctuated around 14%, but rose to about 19% in the 1950s.The proportion of ringed birds reported as shot declined greatly from about 9% before 1940 to less than 1% after 1961, when legal protection was afforded. Whether this decline was wholly genuine, or due partly to shooters failing to report their birds after the ban, was impossible to say. On the other hand, the proportion of ringed birds reported as killed on the roads increased from the 1950s, associated with the increase in the numbers and speeds of vehicles. The proportion of birds killed against glass also increased in this time, following the rise in popularity of large plate glass windows and 'see-through' houses. The proportion reported as 'found dead' was especially high in the late 1950s and 1960s. This was the peak period for use of aldrin and dieldrin, which may have accounted for some of these deaths. The proportion recovered in other ways, through collisions with wire and other obstacles, and through chasing prey into buildings, has not altered greatly over the years. Interestingly, no less than six Sparrowhawks (of which five were males) were reported as killed by cats. Perhaps these birds had been taken on the ground, while they plucked prey. Similar results on mortality causes have been obtained from other ringing schemes in continental Europe (e.g. Saurola 1981). All indicated the importance of collisions, and a decline over the years in the proportions of birds reported as shot. The seasonal trends were also revealing (Fig. 80). Recoveries came in all months, but showed two peaks, in February-April and August-September respectively. Like the Monks Wood carcasses, these recoveries reflected the seasonal pattern in deaths that were noticeable by people. While these deaths may not have reflected the pattern of unseen mortality, the peaks coincided with known seasonal crises. The August-September peak was formed of newly-independent juveniles, which had recently dispersed from their parents' territories. The March-April peak was formed of adults as well as juveniles, and coincided with the period of greatest food shortage each year. In most schemes based on finds by members of the public, fewer cocks than hens were reported. Among our Scottish recoveries, as among the national recoveries, the ratio was 0·8 males per female, and among the Monks Wood carcasses, it was 0·7 males per female. As the sex ratio at fledging was equal, the recovery rate should have been equal, even though males in general died earlier in life. The bias against males was probably due to the habitat divergence between the sexes, with cocks predominating in large woods, and hens in small woods and open country. This in turn would result in more hens than cocks dying near houses and other places where they could be found by people.

Foodshortage, predationand disease It is the natural deaths that occur in the countryside which are usually missed. This is because most carcasses disappear before anyone can find

Mortality

297

2°l FI R S T

YEAR

' 5~

@

10 -

V>

5-

" ">

0

o

"

L

15

L AT E R

0,8

YE A RS

10

L

Aug

1L S ep

O ct

Nov

Dec

J an

Feb

Ma r

Ap r

Ma y

Jun

Jut

Fig. 80. Seasonal pattern oj recoveries from British-ringed Sparrotohatoks. First year birds showed a peak in recoveries in August-September, soon after theyhad become independent, and both first-year birds and adults showed a peak in M arch- April, the period oJg reatest food shortage. Total recoveries oJfirst-year birds = 813; total of adults = 504. th em , bein g promptl y rem oved by Foxes a nd othe r scave nge rs . During 14 yea rs, we found onl y twelve Sparrowhawk ca rc asses in th e woo ds of the study area s, a ne gli gible fracti on of th e birds whi ch mu st have di ed in that time. And with on ly about 14 % of a ll rin ged birds eve r rep orted by members of th e public, th e majority of d eaths went unrecorded. This is not just a feature of Sparrowhawks, but is true of mo st other bird s. It is in this majority unknown category that d eaths from natura l ca uses could pred ominate. Probably th e major natu ral ca use of mortality in Sparrow hawks is food shortage. Evidence suggested two main periods of starvation , on e during Au gust-September and restricted to juveniles, and other during MarchApril, which affected adults as well. T his was the period eac h year when the resid ent prey population rea ch ed its lowest point, wh en most wintervisiting prey had left, and before mo st sum m er visitors had arrived . Starvation in both peri ods was apparent from th e large spread in weights , a s many ind ivid ual s were in poor co nd ition . M or e importantly, as mention ed above, man y were found d ead in the cou n tryside th en , apparently having sta rved . Food sho r tage kill ed some birds directl y, a nd othe rs probabl y indirectl y, throu gh forcin g th em to hunt for lon ger periods, or nearer to human habitation , thus in creasing th eir cha nce of death from collision or othe r accide n ts.

298

Mortality

One bird which we fitted with a radio-transmitter in March was able to obtain all its food in only a few hours each day, but with the start of a cold spell it began to have difficulty, spending more and more of each day hunting, until it eventually hit some wires and died. I t is often stated that, in contrast to other birds, Sparrowhawks should thrive in hard winter weather, because their prey are starving then and easy to catch. If this is true at all, it is likely only in the first week or two of a cold snap. After this, many of the prey birds themselves have either died or moved away, and Sparrowhawks could have a much leaner time. Cold weather with food shortage could accentuate most kinds of mortality, not only starvation. By weakening the hawks it could make them more prone to disease, and by reducing prey numbers it could force the hawks nearer to human habitation, where they would be more likely to die from collisions. On these aspects we lack firm data, and the chiefevidence that Sparrowhawks are affected by prolonged hard weather is the drop in breeding population which sometimes follows cold winters (Chapter 6). On disease there is even less information. Raptors are supposed to select diseased individuals from among their prey species, so for this reason they may often risk infection. Although rap tors have been found with many types of disease, the incidence was no greater than in some other types of birds. Specific diseases diagnosed in Sparrowhawks include fungal ones such as aspergillosis, bacterial ones such as tuberculosis and coccidiosis, viral ones such as Marek's disease, and various parasite infections. Of the many Sparrowhawks which were handled during our work, none showed coughing or other signs of infection, and only two had obvious external lesions. Both of these had 'bumblefoot', a swelling on the underside of the foot caused by injury and subsequent infection. Two birds had broken legs which had healed, and several old females had swollen joints in their feet, reminiscent of arthritis. None of these particular birds was ever retrapped, so as a group they probably survived less well than the rest of the female population. Further information on parasites came from screening apparently healthy birds. Blood samples were taken from 119 full-grown Sparrowhawks in south Scotland during 1979-80, and no less than 113 were found to contain parasites of one or more kinds (Pierce & Marquiss 1983). Leucocytozoon toddi was found in III birds, Trypanosoma corvi in four birds, and Haemoproteus nisi (a species new to science) in 32. Examination of nestling Sparrowhawks showed that infection with at least the first two organisms occurred before the young were half-grown (Chapter 7). All these parasites are probably transmitted by blood-sucking insects, but apparently have no serious effects on their hosts. The various fungal and bacterial diseases identified have effects from slight

to fatal, and seem to appear as sporadic infections (Cooper 1978, Greenwood 1977). Certain parasites may be widespread, but do not normally kill their hosts. On present data, therefore, I would guess that disease is not a frequent primary cause of death in Sparrowhawks, but more information is needed, and its effects on populations are yet impossible to judge.

Mortality

299

The same is true of predation, but on this subject more information is available from diet studies of other raptors. To judge from extensive prey lists from central Europe, the main avian predator of the Sparrowhawk is the Goshawk, although Sparrowhawks comprised only 87 (1 %) of9,000 Goshawk prey items identified (0 ttendorfer 1952). Other predators included the Peregrine (three out of 6,400 prey items), the Eagle Owl (four out of 5,800 prey items), the Tawny Owl (six out of 58,500 prey items), and other Sparrowhawks (46, all young, out of 59,500 prey items). From records in the literature (including 0 ttendorfer's), Mikkola (1968) listed 20 Sparrowhawks taken by owls eleven by Eagle Owls and nine by Tawny Owls. These were among 482 birds of prey (totalling less than 5 % of the diet) killed by Eagle Owls, and among 54 birds of prey (totalling less than 1% of the diet) killed by Tawny Owls. For other large owls, such as the Great Grey S. nebulosa, and Ural S. uralensis, the few diurnal raptors recorded in the food did not include Sparrowhawks. Hence, Sparrowhawks have apparently formed no more than a negligible fraction of the food of any owl or raptor species. This is not surprising as Sparrowhawks are not much more abundant than some of the other predators and are considerably scarcer than Tawny Owls. Some of the recorded hawk deaths may have resulted from conflict over territories or nest sites, rather than from normal predation. In our south Scotland study areas, Goshawks were absent and the main predators were Tawny Owls (Chapter 13). In 14 years, we had only one definite instance of an adult female killed on the nest by this owl, but 21 cases of broods being taken (Chapter 17). At two other nests we suspected Buzzards Buteo buteo had taken young, but were unable to confirm it, and Richard Mearns (1983), working on Peregrines in south Scotland, found three Sparrowhawks among 3,600 prey remains. The only known mammal predators are martens, which can climb to nests, but again these animals were absent from our study areas. Other mammalian predators may kill the occasional Sparrowhawk as it is plucking on the ground, but of this I have no records, apart from the domestic cats mentioned earlier. Records such as these confirm that Sparrowhawks experience predation, but again they tell us nothing about the contribution of predation to the total mortality of the species, or of its role in population limitation. Of the various kinds of mortality mentioned, starvation, predation and certain diseases are the ones most likely to act in a 'density-dependent' way, claiming an increasing proportion of a population as its numbers rise. They are therefore the only killing factors that are likely to 'regulate' a population in the strict sense. Some diseases, and at least some accidental forms of death, are 'density-independent' in that the proportion of birds they claim is for the most part unrelated to population density. In the early days of bird population studies, much effort was devoted to assessing mortality causes, in the hope that these would reveal what determined population levels. But Sparrowhawks in many areas are clearly limited by the carrying capacity of their habitat, and the various mortality agents act mainly to trim down numbers each year to a level that resources will support. This

300

Mortality

suggests the over-riding influence of density-dependent factors (Chapter 6). Sparrowhawks in other areas are well below the environmental carrying capacity, as a result of overkill from pesticide poisoning, as discussed in Chapter 24. In the past, persecution from gamekeepers was apparently sufficient to reduce breeding populations in certain districts (Appendix 2), but I know of no part of Britain where this is still the case.

SUMMARY

The oldest male Sparrowhawk, among British ringing recoveries, died in its eighth year, and the oldest female in its eleventh year. By different means, the average annual mortality of cocks after the first year was estimated at 31-33 % , and of hens at 29-36 % • For birds ringed in south Scotland, the first year mortality of cocks was estimated at 69% and of hens at 51 %. If this last difference was genuine, it would lead to a substantial surplus of females among Sparrowhawks of breeding age (for which there was other evidence). Autumn trapping suggested a heavy mortality of newly independent juveniles in August-September, which was greater in males than in females, and due largely to starvation. This presumably occurred through inexperience, as prey were plentiful in the countryside then. A second period of food shortage, affecting adults as well as juveniles, occurred in MarchApril, when prey populations reached their lowest level of the year. Starvation, predation and disease were all recorded as causing deaths of Sparrowhawks, as were various collisions and other accidents, shooting, trapping and poisoning. The ring recoveries and post-mortem analyses, which provided most information on mortality causes, were biased towards deaths which occurred from human action or near human habitation. The proportion of ringed Sparrowhawks reported as shot has declined in recent decades, while the proportions killed on the roads and against glass windows have increased.

CHAPTER 24

Effects of pesticides

Pesti cid es have had such a major impact on Sp arrowhawk populations in recent yea rs that it wa s necessary to refer to this in a lmos t every previous chapter , wha tever th e asp ect under review . M y a ims here a re to dis cuss th e wa ys in whi ch th ese che m ica ls reac h and affect Spa rro whawks, and to sum ma rise th e changes in br eeding and populati on wh ich have occ ur red . In fact , only a min ority of pesti cid es affect Sp arrowh awk populat ions . Mo st belon g to th e 'orga nochlo rine' gro up, whi ch includ es DDT a nd dieldrin. The Sparrowh awk was not alone in being affected, however, for sim ilar populati on cras hes occurred in the Per egrine a nd so me oth er raptor s (R a tcliffe 1980). Declines in suc h spec ies wer e wides pread , occ ur ring ove r mo st of Eu rope and North A me rica, whe reve r organochl orines were in large-scal e use (Ne wton 1979). Besid es being toxi c, o rga noc hlorine pesti cid es hav e three main properties whi ch con tribu te to th eir effects. First, th ey are extre me ly sta ble, so that th ey ca n persist more or less un ch anged in the enviro nme n t for man y yea rs. Second , they diss olv e readily in fat, which means that th ey can accumulate in animal bodies, and pass from pr ey to predator, con centrating a t su ccessive ste ps in a food chain. Pr ed atory birds , near the top s of food cha ins, are esp ecially liable to ac cumulat e large amounts . Thirdly, a t sub-l ethal levels of only a few ppm in tissu es, organochlorines ca n disrupt th e br eeding of cer ta in sp ecies . They are al so di sp ers ed in wind and water curre nts , or in th e bodies of migrant birds and insects, a nd can thus reach remote regions. No Sparro wha wk in Britain is free from risk of cont amin ati on , a nd ind eed non e of th e SOO birds and 1,500 eggs a na lysed since 1963 was devoid of residues.

302

Effects of pesticides THE CHEMICALS AND THEIR USES

DDT cam,e into widespread agricultural use in the late 1940s. It has since been used against many insect pests, mainly on top fruit (against several pests), brassicas (against flea beetles and caterpillars) and cereals (against leatherjackets). In the environment, and in the animal body, most DDT rapidly degrades to DDE; both compounds are of low direct toxicity to birds, but DDE causes shell thinning (and hence egg breakages) and embryo mortality, thus lowering breeding success (Cooke 1973, Newton 1979). Another organochlorine, commonly known as gamma-BHC, HCH or lindane, came in to use at about the same time. It is much less persistent than DDE, and, although often present in bird tissues, it is not known to have affected populations. The more toxic cyclodiene organochlorines, including aldrin, dieldrin and heptachlor, came into wide use after 1955. They were used mainly against pests in the soil: as seed dressings on cereals and other crops, on brassicas against cabbage root fly, on potatoes against wire worms, and on various minor crops against other pests. In Britain, their usage has thus been greatest in those (mainly eastern) regions with the greatest proportion of arable land. They have caused large-scale direct mortality of wildlife, especially of seedeating birds and their predators. Many Sparrowhawks have died as a result of eating contaminated finches and pigeons. The active ingredient in dieldrin is known as HEaD, which is also produced in the environment or animal body from aldrin. So, on finding HEaD in a hawk, it is not possible to tell how much was from dieldrin and how much from aldrin. In the years since 1962, successive British government restrictions have greatly curtailed the use of these cyclodienes. In 1962, they were 'banned' from use on spring sown cereals; in 1965 from sheep dips and other minor uses; and in 1975 from autumn sown cereals. After that date, the remaining uses were chiefly on brassicas and root crops. Heptachlor was used hardly at all after 1964. The restrictions were not legal bans, but 'voluntary' agreements, involving manufacturers, distributors and users, and there was no legal comeback on any farmer who chose to ignore them. Most restrictions did not lead to sudden reductions in usage, but rather to steady declines over a period of years. Thus, although dieldrin had been banned in sheep-dips for 15 years, analyses of wool samples showed that some farmers were still using this chemical in 1980. Similarly, DDT was no longer recommended for use on top fruit after 1976, but orchard surveys revealed extensive usage after this date (Sly 1981 ). Information on organochlorine use in Britain has resulted chiefly from periodic farm surveys by the agricultural authorities (Table 60). These surveys indicated a decline in DDT usage during the 1960s, and a further slight decline in the early 1970s; in some regions, however, DDT use increased again between 1974 and 1977 (Cutler 1981). The main finding, however, was of continued substantial DDT use in Britain throughout this period. On the other hand, the same surveys revealed a progressive and marked

Effects ofpesticides

303

decline in the use of aldrin and dieldrin between the early 1960s and the late 1970s, particularly after the 1975 ban. A further set of restrictions came into effect in 1981-83, applicable to all EEC countries, and aimed to phase out completely the use of organochlorines in agriculture. Until then, DDT was still recommended in Britain against leatherjackets in cereals, chafer grubs and cutworms, while aldrin also had a few small-scale uses. These restrictions were again voluntary, so it remains to be seen how effective they will be on chemicals that can be bought and stockpiled. Meanwhile, dieldrin continues to be used in moth proofing and wood preservation. It is not just the usage pattern of organochlorines which affects the levels in wildlife, but also their persistence, the very quality that makes them so effective as pesticides. The persistence of chemicals in any medium is usually measured by their 'half life', the period taken for the concentration to fall by a half. The half life of DDE in soils has been variously calculated at between 12 years in some cultivated soils and 57 years in some uncultivated soils (Buck et al 1983, Cooke & Springer 1982). So even if use of DDT was stopped dead, soil-dwelling organisms would remain a source of residues for Sparrowhawk prey species for many years to come. HEOD is much less persistent, with a half life in soil estimated at 2·5 years (Brown 1978). From animal bodies, organochlorines can disappear more rapidly than from the physical environment, but again DDE lasts longer than HEOD. For example, in pigeons the half-life of DDE has been estimated at about 240 days, compared with 47 days for HEOD (Walker 1983). These rates vary between species, and with the condition of the individual. In general, birds have a less good detoxification system than mammals have, which partly explains why birds are generally more contaminated than mammals from the same area, and why avian predators of birds, such as the Sparrowhawk, have been more affected by organochlorines than have avian predators of mammals, such as the Kestrel. The Sparrowhawk also tops a longer food chain than the Kestrel does, giving more opportunities for residues to concentrate.

CONTAMINATION OF SPARROWHAWKS

Organochlorine chemicals reach Sparrowhawks by several routes. When used in seed dressings, the grain may be eaten by a finch or other seed eater, which in turn may be consumed by a hawk. When used as sprays, the contaminated leaves may be eaten by a pigeon, or the pesticide which reaches the soil may be taken up by worms or other invertebrates, which may be eaten by a Blackbird or Starling, which may in turn fall prey to a hawk. With such persistent chemicals, this process can take weeks, months or even years after spraying occurred. There is literally no prey species a

hawk could eat without risk of some contamination, and none of the bird species analysed in the last 20 years has been found to be free from organo-

304

Effects ofpesticides

chlorine residues. Levels vary between species, according to their feeding habits, and between areas, depending on local usage. Hence, certain food lines pose greater threats to Sparrowhawks than others, and individual differences in diet could well account for some of the marked individual differences in contamination levels in hawks. A Sparrowhawk breeding in some remote forest may be less directly exposed to organochlorines than one on farmland, but it cannot avoid them. At the time of spraying, the chemical will drift on the wind and reach places far removed from the fields being treated. And over the ensuing weeks, more of the chemical will evaporate and get carried elsewhere on air currents, or pass into stream water and get dispersed in solution or on particles. But perhaps most important, some of the birds eaten by Sparrowhawks in remote forests may themselves have fed on treated land. They could include migrants from other parts of the world, or local Chaffinches and Woodpigeons which move out of the forest to feed on nearby farmland. In these ways, persistent chemicals become widespread, and although they may be at greatest concentration on farmland, they are present in soils, water and wildlife throughout the country.

Distribution oforganochlorine within the body Sparrowhawks do not get their high organochlorine contents simply by storing the accumulated residue from all their prey, but rather because their average rates of intake are higher than in most other types of bird. Within the body, organochlorines usually settle in tissues in direct proportion to the concentration in the diet, the rest being excreted or broken down. A high dietary level results in a high tissue level, and a low dietary level results in a low tissue level (Lincer 1975). Moreover, because organochlorines dissolve in fat, the level they reach in anyone tissue depends on how much fat that tissue contains. The brain contains less fat (and so less organochlorine) than does muscle, liver or eggs, and much less than adipose tissue which is almost pure fat. Within the adipose tissue the organochlorine is fairly harmless, but if for some reason the fat is suddenly mobilised, the residue shifts to other tissues and may cause death. The brain is the most sensitive tissue. Thus, a bird in good fat condition can carry more organochlorine without harm to itself than can a bird in poor condition, and deaths from direct poisoning are most likely to occur at periods of food shortage, when fat is metabolised. Thus some birds may die weeks or months after accumulating organochlorine, if they are exposed to sudden hunger. The concentration in the female at the time of lay influences the concentration in her eggs. This is expected because, until the shell is added, the egg is merely another body tissue. The organochlorine is in the yolk, which

holds practically all the fat of the egg. The variation in organochlorine levels between the eggs of a clutch is usually small compared to the big differences between clutches (Newton & Bogan 1978). Hence, the levels in individual eggs could be taken as reflecting levels in the female at the time of lay, and as similar to levels in other eggs from the same clutch. However, levels

Effects of pesticides

305

in repeat clutches of given females are lower than those in firsts. By laying eggs, females rid themselves of residues. A hen Sparrowhawk killed after laying had put more than half her total body load of DDE into her six eggs (Bogan & Newton 1977). This avenue of excretion is not open to males, but females do not necessarily have lower body burdens in consequence. This is because other factors influence organochlorine accumulation, including total food intake and diet, which also differs between the sexes. Eggs from yearling females contain lower levels, on average, than eggs from older females in the same population (in common with body residues), but the older females show no consis ten t trend of decreasing or increasing levels in the clutches from successive years. Evidently the level of contamination of particular females changes, not only from season to season, but also from year to year (Newton et al 1981).

EFFECTS ON BREEDING

Shell thinning The field evidence for a link between DDT and shell thinning is circumstantial: (1) widespread shell thinning in wild populations immediately followed the introduction of DDT in agriculture; (2) Sparrowhawks in different parts of the country, exposed to different levels of DDE, showed correspondingly different degrees of shell thinning; and (3) different individuals within a single population also showed some correlation between shell thinning and DDE content in their eggs, with the thinner shells associated with the most contaminated eggs. In addition, experimental evidence on captive birds has repeatedly confirmed that DDT and its metabolites will cause shell thinning in various species, including raptors (Cooke 1973, Newton 1979). Shell thinning was first established by comparing recent shells with historical ones, housed in museums and private collections (Ratcliffe 1967). Some 2,000 clutches have now been examined for the period 1870-1984, from which the first sign of thinning can be dated precisely to 1947 (Fig. 81). This was the first year of widespread DDT use in Britain. On average, shells obtained since that date have weighed 17% less than those laid earlier, but with no change in egg-size. Some collected shells were up to 27% underweight, but beyond this level shells were so thin that they broke on laying, so were not obtained for study. In general, the degree of thinning was more marked in the intensively arable areas of south-east England than in the north and west, where DDT use was less (Table 61). There was thus coincidence in both time and space between shell thinning and DDT usage. If shell index is plotted against DDE content for a sample of eggs, the relationship is linear when the DDE scale is logarithmic (Fig. 82). This is similar to the dose-response relationship for many medical drugs, and means that each ten-fold increase in the DDE concentration is associated with a further doubling in the effect on shells. Like other raptors, Sparrowhawks respond much more strongly to a given dose of DDE than do some other

s ~

• 1·70

1'60

)(

Q) ~

.E

1~~

1-40

•

I.

•

•

• •

:.

•

",1 ••

~

~

1·70

....1; .r: i·.• ... ·1·.I I·.... ..:.II.:..1. ••• .... . I....... 1'.1 I I. • :~•.:.. ·1:·' Ii .. :...11 . .. I!! rPI·~:;:·I:II:·I:I. · '.::1,.1. ·tai·..:· 1 I :1. . !:'I:, ;11 • :.... • : '.. I I . ".:1 II ... . .•:.. . 1••1 I 1 1 1 :!I,I:, !):.:.::-. : :·!.I::JI•.il:ll,dl!:lII:II!.1 I _ I!III lUll · ':. 1:.,1:.•. 11: •. 1 ••• " • ··1.. ... .... I . 1I1III 1 •

Q)

c ~

•

..

. . : .. ... . ··

:

DDT in use

I

••

•

•

•

II .

1

1

1·30

•• • ••••

••

1)

,

0) 0)

.!

w

• I •••

J : ••• !

.1 I ,.

.s: III 1·20 I

••••• Ii

I..

1

••

I. ·

·

..

.. 1 • •

'...

I

I .• 1.

•

.1

1·00

I

I

I

,

,

,

;~~

I

I

,

I!

,

I

;8~ , , ,

I

,

,

I

;~

,

,

I

,

,

,

,

;911~ , ,

I

,

' , , ,1920 I I

I

'··.1 • •• I

,

,

, , , , , ,1930I,, ,

I

,

,

,

I

;9~~

,

I

,

,

,

I

,

;9~~

I

,

I

I

I

I

,

~9~

I

,

,

I

,

I

I

•• 1 I

I·.! •. ~9'~ , , ,, , , ;9'80 I

Fig. 81. Shell thickness index of British Sparrowhawks, 1870-1980. Shells became thinner abruptly from 1947, coincident with the widespread introduction of DDT in agriculture. Each spot represents the mean shell index of a clutch, and nearly 2,000 clutches are represented from all regions ofBritain. Shells available in museum and privatecollections. Shell index = shell weight (mg)/ length X breadth (mm). Updatedfrom Newton (1979).

•

~

~

;;;

.. t)

0

l'

0

w

J:

o-

--

0"1

0'5; in May, x:Z = 12'7, P < 0-05; in June, X2 = 48-3, P < 0-001; in July, X2 = 22-7,

TABLE 17: Frequency ofdifferent pre» species in thefood ofSparrowhawks, south Scotland. April-August by number in Forest plantation

All habitats

bv weight in all habitats

0-13 0-13 1-18 3-46 0-89 0-34 2-36 0-04 0-89 1-22 1-43 20-13 0-42

0-06 0-20 0-55 6-55 0-37 0-70 0-83 0-31 0-96 0-89 1-19 16·80 0·43

0-46 0-65 1·32 8-48 0-24 0-29 0·15 0-05 0-23 0-15 0-43 5-43 0·14

%

%

Farmland

Crow IRook Corvus coronejfrugilegus L_ Jackdaw C_ monedula L_ Jay Garrulus glandarius (L_) Starling Sturnusvulgaris L_ Crossbill Loxia curvirostra L_ Greenfinch Carduelis chloris (L_) Siskin C_ spinus (L_) Goldfinch C_ carduelis (L_) Linnet C. cannabina (L_) Redpoll C_flammea (L_) Bullfinch Pyrrhulapyrrhula (L_) Chaffinch Fringilla coelebs L. Reed Bunting Emberiza schoeniclus (L.)

0-06 0-23 0-46 6-96 0-10 0-96 0-25 0-40 1·13 0-71 1-15 15·49 0·44

Sheep walk

0-22 9-45 0-50 0-45 0-50 0-39 0-56 1-01 1-01 16-22 0·39

September-March % by by number weight in all in all habitats habitats %

0-24 4-85

0-46 4-23

1-46 0-73 0-97 1-21 0·24 1-46 7-77 0-73

0-44 0·10 0-19 0-24 0-03 0·38 2·03 0-17

~ ~

~

c..,

~ V)

TABLE 17: Contd. A/nil-August ~r number in

%) br

%

Farmland Ycllowharnrner H. citrinella L. Tree Sparrow Passermontanus (L.) House Sparrow P. domesticus (L.) Skylark Alauda aniensis L. Grey .\Vagtail Motaciila cinerea Tunst. Pied Wagtail J1. alba L. Meadow Pipit Anthuspratensis (L.)2 Trcecrecper Certhiafamiliaris L. \Yren Troglor{rtes troglodytes (L.) Mistle Thrush Turdlls ciscirorus L. Song Thrush T. ericetorum Turton Ficldfare T. pilarisL. Redwing T. musicus L. Ring Ouzel T. toiquatus L. Blackbird T. merula L. \Vhcatear Oenanthe oenanthe (L.) Whinchat Saxicola rubetra (L.) Stoncchat S. torquata (Han.) Redstart Phoenicutus phoenicurus (L.) Robin Erithacus rubecula (L.) \Yhitcthroat Svlua communis Lath. Blackcap S. atricapilla (L.) Goldcrcst Regulus regulus (L.) \"ilIow Warbler Phvlloscopus trochilus (L. ):~ Wood \Varbler P. sibilatrix (Bcchst.) Dunnock Prunella modularis (L. ) Long-tailed Tit Aegithalos caudatus (1..) Willow Tit Paws atricapillus L. Blue Tit P. caeruleus L. Great Tit P. major L. Coal Tit P. ater L. Spotted Flyca tcher Muscicatia striata (Pall.) Swallow Hirundo rustica I.. House Martin Delichon urbica (L.)

0·78

Sheepiralk

0·89

0·25

3·10 1·20 0·40 0·78 4·09 0'42 1·85 2·39 12·89 0·78 0·06 11·740'46 0·19 0·06 6·67 0·25 0·040·98 2·52 0·041·09 0·23 0·11 2'43 1·70 1·05 0·23 0·55 0·11

Forest plantation 0·25 (}OB

0·89 3·08 0·34 0·67 12·08 2·13 1·90 8·56 0·56 0·11 5·59 1'40 0·11 5·65 0·06 1·57 1·73 0·39 1'01 0·78 2·29 0·28 0·50 0·34

0·13 1·140·25 0·21 5·27 0·30 4,,68 1·90 6·75 0·93 0·13 0·08 4-·89 0·340·25 0·040'04 11'48 0·13 4v

~

~

~

~

e"

'rABLE 32: Proportions of birds recovered, according to date offledging, south Scotland. Birds .fledged early in the season were more likely than late ones tosurvive and breed, iohereas late ones weremore likely to die in theirfirst winter. From Neuiton & Marquiss 1.984. 6--10

1-5

Jusv

Ju£v

Periodichen)'oun,~ left nest 16-20 21-25 26-30 July' Jusv .Iu£v JU£l

3/-4 !1ugust

11-15

5-9 August

Slope ofregression of proportion recovered on .fledging period

/0-/6 August

Cocks Numbers

(~/o)

68

11

l\; umbers ringed

reported dead by general public]

Numbers (cYo) trapped alive in local breeding population

291

207

190

131

2:>

2( 18)

4(6)

13(6)

21(7'1

11(6)

II (8)

0

6(9)

12(6)

14(51

3(2)

5(4)

5

6

0

2(33)

+0·003

0

0

0

-0'012*

22

10

1(4)

Hem 72

210

273

220

N umbers ringed

G

Numbers (%) reported dead by general public]

0

5(7)

1.1(7)

32( 12)

22( 10)

9(8)

1( 18)

2(20)

0

+0·009

Numbers (0/0) trapped alive in local brccdina population

0

10(14)

16(8)

24(9)

12(5)

2(2)

2(9)

0

0

-O'(ll 7*

liS

5

The small numbers ringed in the earlier and later periods meant that the proportions recovered were subject to relatively greater error. The data were therefore analysed by a regression of proportions recovered (y) on periods (x), but weighted according to the numbers ringed in each period. Similar analyses, with the data for certain periods pooled, revealed the same trends. * p < 0·05. t Mostlv in first winter.

TABLE 33: Seasonal decline in breeding performance, uiithyears, habitats and agegroupscombined, south Scotland. From Neioton & Marquiss 1.984. Lavine started durinv:

(a) Clutch size mean ± SD(:'\)

21-5 :Ipli!

2f>-}().1!JlI!

6-0(2)

_~d±O'7(W)

:)"()(I)

-l'S ±

I 5

.\I(~)'

6- I() .\/~~·

II-I:> .\In.)

S·I ± ()·B( 11~»

l·t}::: n·BI2J:»)

Hi ±

l·3± I-:WlI)

,HI ± 1·21 l·lS)

3·; ± I-L( 11m

O·~I(

ltd'

16-20.\lm

,,2

~

I-O( lOti)

21 25 .\In.~ :~

2 ~ II ( -l2 ~

26-}f) .\1(/\ TI

±

l'O(Hi)

)1-4.1/1'11'

L·q ± 1-21 II)

(h) Brood siz« at hatch mean ±

SI)I~)

lc\ Brood size' at mean

± SI)(:\!l

% nests deserted

l l ± I·' (n)

2-7::: 07(20'

:2-6 ± I ± 1-:2(lr))

H7( I-H)

HO(222)

:n:::

I-:!~

12H\

:)"(}

Tl

:L?

2-B

711IRS) 2,,1

()

2

B

I:>

20

~~

I (.cd)

:H)~

1·1 (B2)

2-:> ±

Oq(2:~)

2.~

± 1·1 (7)

71(IHi\

ICI(:>1\

2·1

1·'2

f).q

'27

:n

IB

The downward trends in clutch sizr-, hrood sizes at hatching and fledging. and 'Yo r1111r1ws successful were all high" significant st.uist irallv (p < 0-001 \. Full details in

~B{:21

~ewl()n

)

& Marquiss I CHH.

-Hl('2' 20( I O~ O·B 30

~

~

TABLE 34: Survival of adult females in relation to their laying dates, south Scotland. iVO seasonal trend was apparent, and laying date apparently had no influence on the subsequent survival ~l the.female. FromNeioton & Marquiss 1984. 26 April-5 .H~}'

Number of birds trapped Number (%) retrapped in a later year

64 38(59)

Date offirst (~g 6-/5.Ha)' /6-25.Ha]

174 98(56)

74

~

~

~

~

~

26 JIa)L4 June 15

38(51)

9(60)

Note: The records for different vears were combined and individuals were included more than once if they were caught in more than one year, the unit of comparison being one 'bird year',

TABLE 35: Ana(yses of variance on laying dates of.first eggs, south Scotland. For each bird the la..ying dates (from mates, in the case of males) were from. different years, irrespective of ichether the bird had the same partner and territory throughout, or different partners and territories. All the males were adults, but for females the analysis was done separately, including and excluding yearlings. The analysis showed that variation in laying dates within individuals was significantly less than that between individuals. From Neuiton & Marquiss 1984. Degrees of freedom

Sum of squares

square

Fratio

Between individuals Within individuals Total

20 28

71·74 20·47 92·21

3·59 0·73 4·32

4-91***

Between individuals Within individuals Total

135 231

432·96 411·27

3·21 1·78

1,80***

844-23

4·99

Between individuals Within individuals 'fotal

130 217

1~33'57

3·34 1·71

Source ofvariation

Males

Females including yearlings

Excluding yearlings

***, P < 0·00 1.

371·63 805·20

Mean

.1·05

1,95***

Tables 377 TABLE 36: Summary o.! overall breeding performance, south Scotland, 1971-84. A. Ouerall success, in terms ofindicidual nestingattempts * Number of nests found Number (%) in which eggs were laid Nurnbcr (%) in which young were hatched Number (%) in which young were reared Mean clutch size (±SD) Mean brood size (±SD) at hatching Mean brood size (±SD) at Hedging Mean number of young raised per clutch laid Mean numberofyoung raised per nest huilt

1,389 1,176(85) 843(72) 783(93) 4'58 ± 1·08 3·85 ± 1·24 3·4-3± 1·26 2·3 1·9

B. Overall success in terms of total eggs laid by population* Total number of eggs laid Number (%) of eggs which hatched Number (%) of chicks which fledged

5,386 3,246(60) 2,686(83)

C. Causesoffailure

Failure to lay eggs Egg stage Total Desertion Predation I':gg disappearance Nest collapse Two females at nest Egg addling Egg breakage Addling and breakage Female shot Human egg-robbing Tree felling Unknown

Number

°h> o.fall nests

0/0 ofnil failures

213

15

35

333

21

55

122 8 15 3 7 13 72 2 23 5 10

53

Nestling stage Total Predation Starvation Nest collapse Female shot Tree felling Unknown

Notes: * Each

%

is of the previous figure.

= less than 1% .

t

I

t t

20 1 2

t

5

1 2 12

2

4

1

t t

t

1

1 4

2 9

4

10

60 21 27 1 5

2

15

1

1

t

9

2

3 4

t

t

t

t

t

1

2

378

Tables

TABLE 37: Mean number ofyoung produced from clutches of different sizes. In general, larger clutches produced mostyoung, and in successful nests partial losses were not significantly greater in clutches than in small ones.

A. Forall clutches, including those which produced noyoung

16 0 0

Number Mean brood size at hatching Mean brood size at

Initial clutch size 4 5

6

7

174 2·4 2·0

73 4·3 3·9

6 4·8 3·3

6

7

62 5·0 4·6

4 6·5 5·0

2

3

23 0·1 0·1

53 0·9 0·7

338 3·5 3·1

B. Forclutches whichproduced at leastoneyoung 2

Number Mean brood size at hatching Mean brood size at

1 2·0 2,0

0 0 0

Initial clutch size 4 5 3 14 2-6 2·5

120 3·2 2·9

272 4·2 3·9

periods, Sudlausitz; Germany. Note were in use. From Kramer 1973.

Number of nests found Number (%) which produced young

1916-20

1921-30

1931-40

1941-50

1951-60

1961-70

32 25(78)

106 84(79)

176 142(81)

193 135(70)

205 110(54)

199 108(54)

TABLE 39: Evidencefor assortative matingbetweenyearling (YJ andadult (AJ Sparrowhawks. Probability ofchance occurrence *

Year 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980

All

1

o

3 3

3 3 I 4 3 4 3 3 4

1

4

1

6 9 21 19 27 32 31 22 15 35

18

32

17

217

I 1

o 5 3

o

1

o 3

2 3 4

o 3

0·400 0·505 0-087 0·592 0-002 0·048 0-709 0-006 0·194 0·232 0'9). For the analysis in this table, all age groups were pooled. Females which were retrapped in a later year were classed as survivors, whereas those not retrapped were classed as 'non-survivors'. Individuals which were retrapped at a nest in more than one year were included more than once, for in each case the unit of comparison was one 'bird year'.

oyoung and birds which raised

382

Tables

TABLE 46: Lifetime productions ofresident and immigrantfemales. Number of females which produced thefollowing numbersofyoung O 1-10 11+ (i) Bred once in study area Residents (raised in study area) I mrnigrants (raised elsewhere)

18

17 27

0 0

(ii) Bred more than once in study area Residents (raised in study area) Immigrants (raised elsewhere)

2 2

13 38

8

For 'once only' birds, X2 = 4,12, P

2

= 0'02; for others, X~ = 2,98,

4·

P> 0·3.

TABLE 47: The start of moult in the secondary and tail feathers in relation to the stage of moult in the primaries. Details on secondaries from 114 birds of both sexes, and on tail from 151 birds ~f both sexes. rVide variation was apparent in the relationship between moult timing in these differentfeather types. ofbirds iohich started moult in secondaries

ofbirds which started moult in tail feathers

%

Last primm), to beshed

2

3 4· 5 6

%

12 3426 22

1 19 4·0 37

6

3

TABLE 48: Number 0.1 primary feathers per wing in growth at once at different stages of moult, south Scotland. Most birds had 2-3.feathers in growth at once, but slight differences uiere apparent between malesandfemales, and between agegroups. From Newton & Marquiss 1981a. Latest primary shed

I 0

1·0

3

2

4

5 3 1 3 1 4 10

1·5 2·2 1·3 3·0 1·7 2·0 2·3 2·0

I

2 5 6 7 8 9 10 !\

A~

Yearling male m

Iv

0-3 0-4 1-2 1-2

I 16 9 10 43 .)

Adult male m 1·0 1·40·3 1·0 1·3 1·7 1·8

i" 0-1 0-2 1-2 1-2 1-2

0 2··-3

0-3

I 4·

3·0 1·5

1-2

Yearling female m

Adult female lV m

3

1·0

8

4

2·0 3·0

42

7 13 13 17 33 2 3 10

3·8 4··2 1·8 1·7 2·0 2·0 0·9

2-4 1-5 0-6 1-2

0-2

= number of birds examined; m = mean number of feathers in growth; r = range.

72 94· 110 39 4 4· 1 3

1·0 2·0 1·0 1·5 2·5 1·3 1·8 1·8 2·0 2·0

0-4 0-5 0-6

0-2

1-··2 1--3

Ti\BLE 49: Dispersal distance from natal nest and subsequent performance, south Scotland. For each category thefigures show the number ~f indiciduals (N), the geometric mean dispersal distance (m}, and the range icithin one geometric standard error on either side (RSE). In general, females which bred in lowland habitats, laid early in the season and produced large clutches had moved less far from their birthplace than had females u.hidi bred in uplandhabitats,and laid latesmall clutches. From Neuiton & Marquiss 1983.

.v

Males Dispersal distance (km) m RSE

. S. ignificance ofdifference

Females Dispersal distance (km} .V m RSE

Sip, nificanee ofdifference

Birds which nestedin the studv areas Breeding habitat Lowland Upland Grade of territory Poor Good Laying date < 10 May > 10 rvlay Clutch-size 1-4 5-7 Brood size 0-2 3-5

20 12

4"26 6·09

3'61- 5·03 4-62- 8·03

t:30

~S

39 28

7·89 11·29

7·04- 8·81 12·1 1-16'88

t6 .'">

= 3,09, P < ()oOI

8 24·

5·65 1'{)'1

4-44- 7-18 3'87- 5·55

t:30 = 0,57, NS

30 37

10·05 9·54-

8-51-1 I ·H3 8·62-10,5,1-

t 4CJ

= 0,28,

14 13

5·144·96

3'91- 6·76 4·29- 5·74-

t 20

= O'11, NS

51 ,10

7·55 11·23

6·81- 8·37 10·06-12· 5:>

tIN = 2,62, P < 0·02

5 20

5·27 5·01

3·69- 7·51 4'13- 6·09

t

= 0,12,

NS

27 63

11·97 8·31

10,37-13,81 7'57- 9·11

tHB = 2,15, P < 0·01

8 18

7·97 4·:,7

6·46- 9·83 3·54- 5·39

t 2.t

= 1,73, l\'S

34 54

10·82 8·70

9·37-12,49 7·89- 9·60

tB6 = 1·3D, N'S

12·64 11·57

10·39-15·3H 9,10-14,72

0,25, NS

67 32

24·35

21 .55-27·51 18·10-25·43

t CJ 7 = 0,60, l\S

21

= 1,18,

NS

Birds recovered deadb] members~f the public Age on recovery I year

47 18

t 6:3 =

21"~5

Note: 32 males and 67 females bred in the study areas, but for some individuals data from more than one breeding attempt were obtained, and are included. Members of the public reported 65 males and 99 females. i\S - not significant.

~

e;:::,..

~

~

VJ

384

Tables

TABLE 50: Correlation between distances from natal nest moved by closely-related individuals, south Scotland. (1) and (6) based on re-traps in the study areas; (2)-(5) basedon recoveries provided by members of the public. For correlations, all distances were expressed as a percentage of the median distance for each sex, and then converted to logJ() values to give a distribution close to normal. In general,young from the same brood, orfrom the same motherin differentyears, dispersed similar distances; but no suchrelationship was apparent between the equivalent movements ofyoung and their parents. From Newton & Marquiss 1983. Comparison between:

Correlation coefficient (r)

Degrees of freedom

Statistical significance

0·84 0·92

14 16

P < 0·001 P < 0·001

0·68

9

P < 0·05

0·94

11

P < 0·001

0·91 0·33

22 7

P < 0·001 P>0·5

(1) Young from the same brood (re-traps) (2) Young from the same brood (recoveries) (3) Young from the same mother and territory in differen t years (4) Young from the same mother, bu t differen t territory, in different years (5) Young from the same mother, irrespective of territory (3 and 4 combined) (6) Paren t-offspring*

* Involved 4 mother-daughter, 3 mother-son and 2 father-daughter comparisons. To allow for the sex difference, I expressed the dispersal distance of each individual as a percent of the median distance for its sex, and could thus compare the sexes directly.

TABLE 51: Turnoverof birds on territories ofdifferentgrade, based on individual identification from successive years only) south Scotland. Figures show numbers (0/0) of cases. Turnover was greateston low-gradeterritories. From Newton & Marquiss 1982. Males Same bird

Different bird

Females Same Different bird bird

Low-grade territories! High-grade territories/

4 21

6 27

36 112

All territories

25(43)

33(57)

148(50)

62 86 148(50)

Note: 1. Used in 1-6 years out often. 2. Used in 7-10 years out often. Significance of variation: males. X2 = 0,02, ns; females, X2 = 9,53, P < 0·01.

TABLE 52: Frequency of territory changes in adults, according to nest success the previousyear, south Scotland. Birds were more likely to change territory if they failed in their breeding the previous year, than if theyraisedyoung. FromNewton & Marquiss 1982.

Males, numbers (0/0) in secondyear in Same territory Different territory After success After failure

19(86) 6(60)

3(14) 4(40)

Females, numbers (%) in secondyear in Same territory Different territory 125(78) 23(42)

Significance of variation: males, X2 = 1'47, ns; females, X2 = 21'5, P < 0·09.

36(22) 31(58)

TABLE 53: Proportions of hens which changed territory, according to age and nest success the precious )'ear, south Scotland. Birds tohich [ailed the preuious "year changed territor) more often than did birds which succeeded! but the tendency to change territory became less marked icith increasing age. From Neicton & Marquiss 1982. Comparison betlHenfo/!ou.'inp, ages (rears) 2 t03

1 102

On same territory

3+ tofolloicingvear

On different territory

On same territor»

On difjerent territorv

On same territory

On differmt territory

After success in previous year

6

5

IG

:>

103

2()

After failure in previous year

o

10

,1

:>

19

IG

Significance of variation within age groups

x2 =

~)'2! P < O'OS

x~ =

1·6. P