Hemlock: A Forest Giant on the Edge 9780300186772

Citation preview

HEMLOCK

HEMLOCK A Forest Giant on the Edge

DAVID R. FOSTER , Editor

BENJAMIN BAISER , AUDREY BARKER PLOTKIN, ANTHONY D’AMATO, AARON ELLISON, DAVID FOSTER , DAVID ORWIG, WYATT OSWALD, JONATHAN THOMPSON

STEPHEN LONG, Consulting Editor

Published with assistance from the Louis Stern Memorial Fund. Copyright © 2014 by President and Fellows of Harvard College. All rights reserved. This book may not be reproduced, in whole or in part, including illustrations, in any form (beyond that copying permitted by Sections 107 and 108 of the U.S. Copyright Law and except by reviewers for the public press), without written permission from the publishers. This volume is a contribution of the Long Term Ecological Research (LTER) program funded by the National Science Foundation. The largest and longest-lived ecological network in the United States, the twenty-five LTER sites encompass diverse ecosystems from Alaska and Antarctica to islands in the Caribbean and the Pacific. The Harvard Forest has been an LTER site since 1988.

Yale University Press books may be purchased in quantity for educational, business, or promotional use. For information, please e-mail [email protected] (U.S. office) or [email protected] (U.K. office). Frontispiece: Mist rising from the snow in a hemlock woods on the Harvard Forest. (David Foster) Excerpt from A Farewell to Arms by Ernest Hemingway reprinted with the permission of Scribner Publishing Group. Copyright © 1929 by Charles Scribner’s Sons. Copyright renewed © 1957 by Ernest Hemingway. All rights reserved. “Dust of Snow” and “On a Tree Fallen Across the Road” from the book The Poetry of Robert Frost, edited by Edward Connery Lathem. Copyright © 1923, 1969 by Henry Holt and Company, copyright © 1951 by Robert Frost. Permission granted by Henry Holt and Company, LLC. All rights reserved. Designed by Mary Valencia. Set in Adobe Garamond type by Tseng Information Systems, Inc. Printed in the United States of America. Library of Congress Cataloging-in-Publication Data Hemlock : a forest giant on the edge / David R. Foster, editor; Benjamin Baiser, Audrey Barker Plotkin, Anthony D’Amato, Aaron Ellison, David Foster, David Orwig, Wyatt Oswald, Jonathan Thompson. pagescm Includes bibliographical references and index. ISBN 978-0-300-17938-5 (hardcover : alk. paper) 1. Eastern hemlock. 2. Forest ecology—New England—Research.  3. Forest declines—New England—Research. I. Foster, David R., 1954– QK494.5.P66H46 2014 577.30974—dc232013042752 A catalogue record for this book is available from the British Library. This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 10 9 8 7 6 5 4 3 2 1

To our Harvard Forest predecessors and their legacy

Report on the Forests of North America Charles S. Sargent, Arnold Professor of Arboriculture at Harvard College 1884

Hemlock Tsuga Canadensis, Carrière Nova Scotia, southern New Brunswick, valley of the Saint Lawrence river and southwest to the western borders of northern Wisconsin; south through the northern states to New Castle county, Delaware, and along the Alleghany mountains to Clear Creek Falls, Winston county, Alabama. A tree 21 to 33 meters in height, with trunk 0.90 to 1.15 meter in diameter; dry, rocky ridges, generally facing the north and often forming extensive forests almost to the exclusion of other species, or less commonly borders of swamps in deep, rich soil; most common at the north, although reaching its greatest individual development in the high mountains of North Carolina and Tennessee. Wood light, soft, not strong, brittle, course, crooked-grained, difficult to work, liable to wind-shake and splinter, not durable; color, light brown tinged with red or often nearly white, the sap-wood somewhat darker; largely manufactured into coarse lumber and used in construction for outside finish, railway ties, etc. The bark, rich in tannin, is the principal material used in the northern states in tanning leather, and yields a fluid extract sometimes used medicinally as a powerful astringent. Canada or hemlock pitch, prepared from the resinous secretion of this species, is used in the preparation of stimulating plasters, etc. OTHER BOTANICAL NAMES

Pinus Canadensis, Linnæus Pinus Americana, Miller Pinus Abies Americana, Marshall Abies Canadensis, Desfontaines Picea Canadensis, Link, Linnæa

CONTENTS

Foreword by Robert Sullivanix Prefacexxi Acknowledgmentsxxvii 1. Hemlock’s Future in the Context of Its Past1 2. An Iconic Species11 Lessons from Harvard Forests and Ecologists: I. The Pisgah Forest25 3. Prehistory to Present44 4. Tree-falls and Tanbark64 Lessons from Harvard Forests and Ecologists: II. Bob Marshall’s Plot77 5. Hemlock as a Foundation Species93 6. A Range-wide Hemlock Decline105 7. Invasion of an Exotic Pest120 8. Cut or Girdle136 9. Modeling the Dynamics of a Forest Giant153 10. Reprise: Eastern Hemlock as a Foundation Species165 Lessons from Harvard Forests and Ecologists: III. The Earl Stephens Plot172 11. When Doing Nothing Is a Viable Alternative: Insights into Conservation and Management181 Lessons from Harvard Forests and Ecologists: IV. Three Views from John Sanderson’s Woodlot204 12. Lament224 Bibliographic Essays231 References269 List of Contributors293 Index295

This page intentionally left blank

FOREWORD Robert Sullivan

W

e have gathered together today, in a still-dark grove on the side of a hill, to say a few words about hemlock. We have gathered, in fact, in a hemlock forest to say these few words, though we are a small group and there are not many words, and nothing anyone can say can possibly sum up the long life of hemlock on the eastern seaboard of North America, which began shortly after the last ice age and thus included a long stretch of difficult-to-sum-up-time that runs from the late Pleistocene epoch through the Revolutionary War and up to the arrival from East Asia of a tiny sap-sucking bug, the hemlock woolly adelgid, which landed in the United States around the time Calvin Coolidge was elected. Now that the woolly adelgid is here, the hemlock is dead or dying, half of its vast range disintegrating, the most northern extent already well on its way out, as warmer weather has moved the bug’s habitat north. In just a few years, the forest we are standing in will no longer be standing. Which means that technically a eulogy would be a little ahead of things; the hemlock in the grove I am standing in are not all dead yet. And yet this forest in Petersham, Massachusetts, is dying and telling our little group as much—when we look around, but also when we listen. It is thus possible to hand over the eulogy to the hemlock, to listen to what the hemlock has to say. Don’t worry—I am not trying to say the trees are talking or, more worryingly, that I hear the trees talking, though I would be lying if I told you that I had not done quite a bit of talking to trees. What I am about to take away from my visit with a grove of dying hemlocks is the sound of the hemlock trees, as they stand dying. Hearing the hemlocks was a first for me, and it brought me back to all the questions that are addressed and examined in this book, the questions that we face at the moment of hemlock death: What do we do with the death of this species? Have we seen this before? And are we prepared to bear witness? I’d been to many hemlock groves on many occasions, but I’d never really listened to a hemlock forest before, at least not so intentionally. I was amazed, frankly,

ix

and I carried that sound all the way home with me like a catchy pop song or something important that someone had said. I heard it on the walk from the woods back to my car, on the drive back to bigger and bigger highways. Sure, I turned on the radio some, and I got out of the car and bought some coffee at a rest stop, and at some point I listened to the traffic report—passing through various eastern cities, small and large, seated on big rivers that had themselves heard, in a distant way, from little streams that had been cooled by hemlock groves to the north—but I could still hear the hemlock, and still can. Like traffic, hemlock is not sexy. It is not a celebrity tree. Hemlock is no sequoia, with a national park. To put it in economic terms, even when hemlock was thriving it was never expensive. It was not coveted by woodworkers the way a woodworker would have coveted walnut or oak. New England fisherman used its bark to dye their sails and nets, and John Josselyn, an Englishman who lived in the colonies for a few years and published New England’s Rarities Discovered in 1672, noted that hemlock’s turpentine was good for “any Ach.” Hemlock tea, it turns out is high in vitamin C, and, to return to our theme, anything in any way associated with the common cold is just not sexy. One theory as to why hemlocks are called hemlocks in North America suggests that colonists thought the hemlock trees’ drooping branches were similar to the hemlock herb in Europe, the poisonous plant that Socrates ingested at his suicide. Hemlock was separated from other conifers—christened Tsuga—late in the eighteenth century by Élie-Abel Carrière, an arborist remembered for his 1857 work Arbres et la civilisation, a book suggesting that cutting a lot of trees caused flooding, and that forests were good for civilization. “Books on the same subject are still being published,” notes nature writer Diana Wells in Lives of the Trees. “And land is still being cleared of trees.” The eastern hemlock’s humble reputation may also have to do with its habitat. If absence makes the heart grow fonder, the opposite may have applied to hemlock: it grows in swamps and rocky slopes and deep ravines north in Nova Scotia and New Brunswick and west into Québec and Ontario, and then stretches out to Michigan and Wisconsin. In the south, hemlock extends into Georgia and Alabama, as if following the Appalachian Mountains. Charles Sargent’s classic 1884 monograph on North American forests, whose entry on hemlock is reprinted opposite this book’s table of contents, marks the southernmost reach of the tree at Clear Creek Falls, which is now gone, having been flooded by Alabama Power in 1961 and subsumed by the current Lewis Smith Lake. Hemlock tends to grow in the faraway places that are close in. It is ubiquitous but at a distance.

FOREWORD x

As opposed to even old-growth pine, which to a woodworker is tight with rings, stable and hard, hemlock has always been considered knotty, coarse-grained, brittle, shunned even as a Christmas tree, the needles falling as soon as the tree sets up indoors. When woodworkers turned to hemlock, they used it for palettes and boxes, crating, shingles, lathings, or paper pulp—stuff we stepped on, shipped, jotted on, and crumpled. Loggers were mostly advised to sell it fast and cheap as railroad ties; I like to think of all the trains that went west on hemlock. Rayonier, a pulp mill in the state of Washington, helped developed rayon with the pulp of western hemlock, a distant relative of the eastern hemlock (back from when the continents were lined up differently) that is immune to the ravages of the hemlock woolly adelgid. The Amish used hemlock for framing, and a barn or house roughly made might be covered in hemlock, built with long-lasting hemlock floors, with hemlock siding on the outside. Given all the tannin in the bark, hemlock siding lasted. Looking back over the course of its interaction in the daily life of humans, the most famous use of hemlock was in tanning, thanks to its thick, reddish, tannin-rich bark. Anyone who has ever fly-fished for trout in the Schoharie Creek in the Catskill Mountains has considered the shade-scarce banks of the creek in comparison to what must have been before Zadock Pratt moved to Schohariekill to build what was, by 1840, the largest tannery in the world. (The big trout, it is said, were even bigger pre-Pratt.) Pratt chose Schohariekill precisely for its hemlock—old majestic groves surrounded the town, supplying the tannery—and as Prattsville grew, it was soon referred to as “the Gem of the Catskills.” In The Catskills: From Wilderness to Woodstock, Alf Evers describes the procession of furs brought in from around the world—from California and Argentina, from Honduras and Uruguay—to Pratt’s swamp. (Today Prattsville is in a ski area, which gives you an idea of the kind of terrain that hemlock loves.) The trees were skinned in the woods by workers called “peelers,” and entire slopes of barkless trunks were left to rot: un-clear-cuts. Pratt was an eccentric, described by his phrenologist as “extravagantly organized.” “He is from these causes, consequently eccentric,” the phrenologist continued, “each action and motion bears the impress of his mind, which makes him somewhat peculiar, isolated and detached from his species.” Pratt married five times. Three wives died young, the fourth got a divorce, and he met the last woman to marry him in the office of the tanning industry’s newspaper, Shoe and Leather Reporter, the editor of which claimed that the final Mrs. Zadock Pratt had, in his words, “acquired that amiability and flavor of The Swamp that made her attractive to the old tanner.” Pratt claimed he wanted “to live with the local people and not on them,” though

FOREWORD xi

he eventually managed both. As the trashed trunks of debarked hemlock began to litter Prattsville, he hatched a plan for his life after hemlock based on the theory that “hemlock land,” as former hemlock forest was known, made for great butter. Orange County, a little farther down the Hudson River Valley, was already famous for its butter—Orange’s butter was considered slow to spoil and bought by the U.S. Navy. There was a theory that Prattsville’s hemlock land, when cleared and turned over to dairy production, would produce butter better even than Orange County’s. But the world’s largest high-quality butter town never happened. By 1845 the hemlock within a ten-mile radius was gone. The town quickly shrank. Pratt commissioned himself a hemlock coffin. In the Catskills it was well noted that a green hemlock log thrown on a fire crackled like a gun battle. “And when I die let me be buried in a hemlock coffin,” was an oft-heard refrain, “so I’ll go through hell snapping.” Shortly before he died, a flood washed Pratt’s coffin down the Schoharie Creek. Pratt commissioned a Shoe and Leather Reporter correspondent to write his life story—and had a shorter version carved into rocks that overlook the town, in an area that, prior to the adelgid, was experiencing a regrowth of hemlock. In 2011 the town once known as Schohariekill suffered disastrous flooding as a result of Hurricane Irene. It might not have been so damaged, it could be argued, had Pratt never set up shop, or moved to butter a little sooner. A little forest management goes a long way. In preparing to listen to today’s hemlock eulogy, it is important to remember that a striking characteristic of human civilization is its tendency to discount what is most important to it, the essential thing to sustaining its long-term existence. Swamps, for instance, are forever being filled in, or “reclaimed”—the old real estate term for dumping garbage in a marsh. Reclamation always happens in the name of civilization’s betterment, even though civilization would be better off if swamps were not filled in: they are a big key to civilization’s health, the watery guts that clean and nourish our rivers and streams, just for starters. In my mind, a hemlock forest is the nearly secret, upland equivalent of the undervalued salt marsh at the distant bottom of the stream, a place that gets along despite the encroachment of humans surrounding it, a place that is often a bit neglected by the powers that be (due to its economic unsexiness and its out-of-the way-ness), even though it is often cherished by its locals, the people who know it best. Like an old swamp in a city, it is a place that, when engaged with, helps the humans get along a little better. Forests, like salt marshes, offer us a chance for a long-term relationship with a landscape, especially when we tie ourselves to them as a resource—an economic, ecological, and, sure, even emotional resource—and

FOREWORD xii

especially when we allow them to be forests. What’s important to note about a hemlock forest is that hemlock is a foundation species, a species that orchestrates the architecture of a large community of species in the forest. Losing hemlock is like losing a conductor and the music, though rest assured the concert hall remains, and there are plenty of people waiting in line to play. The hemlock forest that is eulogizing itself today is quiet and dark, like a hemlock forest, and it began sprouting around the time that First Lady Dolly Madison was clearing out her things from the White House in preparation for the second British invasion, during the War of 1812. These are young trees, relatively speaking. Some of the oldest hemlock are close to 500 years old and are still standing in the least touched places in the East, such as the Berkshires and Adirondacks. But many of them have already fallen, like the old groves in the Great Smoky Mountains National Park. I hate to separate the country from the city, or call one thing natural and another thing not—to avoid doing so is a matter of personal temperament for me. So to put it in terms of urban ecology (related to rural ecology but with higher rents) these rare, old hemlock woods are the city trees in the abandoned lots, that old retired guy in the last rent-controlled apartment, surrounded by gentrification, holding on, until now. It had been hot in the city when I drove out in the morning to pay some final respects to the hemlock. It had been hot on the highway, of course, and even on the back roads as I wound through the Massachusetts countryside, but when I stepped into the hemlock forest, the temperature dropped; it was immediately and blessedly cooler, by at least ten degrees. The hemlock grove is a calm and cool oasis, like an old church in summer. The hemlock that I went to hear being eulogized is in the New England uplands, and I made the drive in part because the hemlock in the city where I live and in the coastal lowlands are already gone. The Hemlock Forest, a fifty-acre grove in the New York Botanical Garden—once a cool, quiet place along the Bronx River—is no longer hemlock. Likewise, most of the hemlock on Hemlock Hill at the Arnold Arboretum in Boston have fallen to the comma-sized aphid-like insect. The hemlocks on Hemlock Hill were famously visited by the transcendentalists, notably Margaret Fuller, the transcendentalist newspaper columnist—she called herself the most intelligent woman in America, and if you go back and look at her reports for the New York Tribune, you’ll see that she was probably right, or at least not far off. The Great New England Hurricane of 1938 knocked down dozens of the trees, yet the grove survived, and now, post–adelgid arrival, those few that remain are healthy Chinese hemlock, resistant due to an evolutionary history with the bug, some planted near the top of the hill. More recently, the hemlock-less peak

FOREWORD xiii

of Hemlock Hill—now a dry, rocky outcrop—is a place that attracts another kind of transcendental experience whose presence speaks, like ski resorts, to the habitat of the hemlock and its relationship to human ecology. The last time I was there, I saw evidence of a high school outing: empty beer cans. No surprise, as teenagers like to be just out of reach of the authorities, the way successful old-growth hemlock tends to be just out of reach of the ax. Another transcendentalist who stumbled on some old hemlock in his own backyard was Henry David Thoreau. He climbed Mount Monadnock, just over the border from Massachusetts, in 1844, 1852, 1858, and 1860, always botanizing, always taking notes. Monadnock wasn’t wild and unspoiled like the Maine woods, where he also explored; he liked it because it was a rocky old place practically just out back, a pocket of wildness. It was hemlock-y, I would say. I blame Thoreau for causing me to stumble into the hemlock grove at the Harvard Forest. By my measure, the world is still a pretty good place if you can call somebody up and get invited to walk in some woods, and that’s what happened when I first called up David Foster, the latest in a not-too-long-line of Harvard Forest directors. I was thinking a lot about Thoreau at the time, and Foster had written one of my favorite books on the forever misinterpreted Massachusettsian, Thoreau’s Country: Journey through a Transformed Landscape. Thoreau’s Country begins with Foster himself landing in the Vermont woods in 1977, just out of college and setting out to build a cabin. Taking Thoreau’s journals for company, Foster quickly discovers a discrepancy between the thick New England forest he sees surrounding him in 1977 and the forest described by Thoreau over the course of his journal keeping, 1837 to 1861. Thoreau, it turns out, lived at the peak of New England deforestation—not what we think of today when we think of Thoreau country. He lived among some woods but primarily ecologically productive farms and meadows. The pondside woodlot on which Ralph Waldo Emerson allowed Thoreau to build his cabin was an anomaly, and a precious one at that: ministers were paid in wood in Thoreau’s time, the trees at Walden Pond being a little like an oil field in twenty-first-century Iraq. Foster’s book helped me access that other time in the landscape, to imagine that whole forests came and went, and not for the first time. This is a trait that I admire about foresters: they think big. Some of the greatest thinkers about the American landscape have been foresters, including Gifford Pinchot, Aldo Leopold, and Benton MacKaye. MacKaye drew up plans to remap the entire United States during the Great Depression, and his Appalachian Trail is a model of locally administered regional planning. With Leopold, Bob Marshall founded the Wilderness Society, after spending time at the Harvard Forest (as this book details). When I am thinking about rush-hour traffic, a forester is likely to be thinking about the last ice age, or the birch and oak that will be along 100 years after FOREWORD xiv

a hemlock forest dies—or, in the case of the ecologists at the Harvard Forest, a regional plan called Wildlands and Woodlands that seeks to keep our vast (and often forgotten) northeastern forest healthy and productive. Maybe this bigger picture has to do with the roots of our definition of a forest. The forest was defined early on as the place outside the king’s garden, the place beyond. The Harvard Forest, I learned upon visiting, is a 3,750-acre experimental forest. Its scientists are using a forest, as their mission statement notes, to “investigate the ways in which physical, biological and human systems interact to change our earth.” I, on the other hand, consider it a great place to visit; first, because it is just a little nutty—no offense to the scientists working for decades on experiments that are changing our understanding of the landscape, giving us views into the organism that is a forest. There is a farm, the old Sanderson place, and postcard-perfect New England landscape almost everywhere you look, and there are trails for the day person. But it is also a mad place too, with pipes and hoses and towers where people check the vital systems of the trees, inspect a grove’s breathing. It’s a wired forest. Yet despite all the appliances, when I extol the place to friends, I start by praising the dioramas in the Fisher Museum: eight beautifully crafted depictions (trees made of copper wire, the smallest wire, the smallest limb) of a moment in the New England landscape, running from 1700 to 1930. Better than any 3-D Internet experience I know of, the dioramas, constructed in the late 1930s, offer the story of a forest that, since European settlement, was cut and cleared, that grew back, then was cut again, and now, in its latest rendition—targeted less by axes and more by invasive species and exurban sprawl—is a system we need more than ever in a warming global environment. When spring comes and the trees in New England burst forth their leaves, as part of the larger East Coast forest that runs from Canada to the Carolinas, from New York to Ohio, there is more carbon sucked out by the forest surrounding the Harvard Forest than the amount of carbon sucked out by the Amazon rainforest. As the woolly adelgid moves north with warmer weather, as the hemlock are dying, what will happen to the forest? This is the question that the Harvard Forest is poised to answer—to measure the devastation, the total loss of a foundation species, more fully than ever before. A hemlock forest is wired at the Harvard Forest, ready to be measured and analyzed in its collapse, which has begun. In ecological terms, this is a little like having a seat at the big bang, a view of a disruptive event, from which will inevitably come change, a new forest, new systems interacting in new ways. Understanding this death—as well as hemlock’s past and present, naturally—is crucial to understanding how we shall proceed with all the rest of life, in a world where species collapse is less and less a surprise. FOREWORD xv

What will happen when hemlock is gone? This book details the natural history of the hemlock and its loss, and also the interaction of ecologists with hemlock and with forests generally. It discusses lessons learned from past declines, and insights, primarily into the dynamism of the forest, or what Bob Marshall described as its “dynamic beauty”: “Then some ancient tree blows down and the long-suppressed plant suddenly enters into the full vigor of delayed youth, grows rapidly from sapling to maturity, declines into the conky senility of many centuries, dropping millions of seeds to start a new forest upon the rotting debris of its own ancestors, and eventually topples over to admit the sunlight which ripens another woodland generation.” There are predictions, of course. At the moment, the woolly adelgid has a competitor as far as hemlock death goes: in New England, the primary cause of hemlock death is preemptive logging, whereby hemlock is logged by humans prior to its death by the hemlock-killing bug. Studies of plots logged in the Harvard Forest to simulate preemptive cuts are predictions as to what will happen, showing us that a lot of carbon is released from the forest floor. Black birch, for instance, comes running in, happy to take hemlock’s place. A forest is a system defined in large part by change, and this you will hear again if you walk into a hemlock grove with David Foster and a couple of his fellow Harvard foresters. “In many ways you can just look at this as part of the natural flux of the forest,” Foster said to me on the day I went to hear hemlock offer up its own eulogy. “The trick in thinking of it that way is that this of course is an introduced species from outside the evolutionary history of the system . . .” That’s the woolly adelgid, which hails, as mentioned, from Asia. “And so it isn’t a particularly natural part of this system,” Foster went on. “But the way the forest is coping with it is the same way that it would cope if this were a native insect or a hurricane or a tornado or something else—the forest wouldn’t get upset, it copes. In fact, the deer are quite happy. Lots of new shoots.” “There are winners and losers,” said Steve Long, a longtime forest watcher and fellow at the Forest. He is the founder of Northern Woodlands Magazine, the author of More Than a Woodlot, and a student of the 1938 hurricane, an adelgid-like devastation whose damage is still visible in the Harvard Forest, just as the earlier chestnut blight is visible: dead but still standing chestnuts lean relaxedly against a terminal hemlock. “Black-throated green warblers are here now, and they will go. The conifer structure is what they like.” When I imagined jogging out to this place in the forest fifty years hence to describe to children the hemlock forest that had once been on this spot—the cool

FOREWORD xvi

shade, the darkness, the quiet—Foster notes that I would likely not be able to distinguish the onetime hemlock land from the rest of the forest on my jog—hemlock would be replaced by the forest around it. “Right now, I would just jog in place,” said Clarisse Hart, a forest researcher and a poet. “The ground is so soft.” “It’s sad,” she adds. She is talking about the death of hemlock, and I agree, it is sad. At this moment, Hart brings to my attention the sound of the hemlock grove. “We know that scientific experience is only a piece of the decline,” Foster says, eventually breaking our silence. “We’re documenting this in a way that is completely unrivaled, but we are also living and feeling it. The forest is dying, and yet it is still recovering from its last hit. This forest is old by our standards, but it has only been growing a couple of centuries. It takes 500 years to come back from a hit to anything like a steady state. We’re less than midway toward that point, and the adelgid is going to stop that. It’s going to divert it into another transition.” We all talked a little about watching for change in the landscape, and in a forest’s time, versus looking for steady states. “When you look at the geological record you tend to say, ‘Oh man, look at how much happened in 10,000 years. Hemlock was knocked back five times!’ But in between those events, you will have 1,000 years to 2,000 years where essentially nothing happened. And you could just as easily and profitably focus in on the long periods of relative stability. We’re all about change, but in fact the records are not just about change.” Just before I finally allow the hemlock to say their piece—the group that gathered here today is now standing still for a moment, about to listen—I’d like to mention a few poets who looked at hemlock in the Harvard Forest’s neighborhood. A good poet, after all, does what the forester does: takes in the myriad details to give us perspective, an entrance into a larger view of time. Robert Frost applies here for many reasons. In “On a Tree Fallen Across the Road (To hear us talk),” published in 1923, he writes: The tree the tempest with a crash of wood Throws down in front of us is not to bar Our passage to our journey’s end for good, But just to ask us who we think we are, Insisting always on our own way so. She likes to halt us in our runner tracks, And make us get down in a foot of snow Debating what to do without an axe.

FOREWORD xvii

Trees mark time, for the humans who notice them, and in this poem the tree interrupts, diverts us, into what a composer might call a fermata, in which, as the tree now directs, we hold a note—in this case, a note in time. Frost is exacting; his metaphor becomes one with the natural detail; figure of speech and the reality of the birch synch. I was once in a bar when the poet Seamus Heaney, extolling Frost’s “Birches,” went on and on in admiration about the bendiness of the birch branch, which almost gets the boy to heaven but does not, earth’s gravity pulling him back— he reaches a kind of beautiful limbo, simultaneously a heavenly earth and an earthly heaven. “The bending, the bending,” Heaney kept saying. “Dust of Snow” is what I would call a very hemlock-y poem that Frost published in 1923, while teaching at Amherst, which is, by the way, eighteen miles as the crow flies from Petersham and the Harvard Forest. You can read it as a crow startling a man, a note of morbid awareness. But the power of the poem builds, I would argue, as you delve deeper into hemlocks—as you enter into this book, for example, to understand the nature of the hemlock, considering all the various conservation and management plans, considering the time it takes for hemlock bark to decay, which is, in human terms, just about forever. When you ponder what it means for a subdivision to punctuate forever a northeast forest you perhaps hadn’t noticed, to poison a forest that has survived an earlier life as a meadow around the time of Thoreau, or when you discover the efficiency of a hemlock forest—the need for so little light, the characteristic denseness of its canopy, branches never dropping, even down low (as opposed to, say, a pine, which concentrates its branches toward the sky)—the poem grows on you, as if the hemlock-knowledged you were a nurse log, primed for succession. The way a crow Shook down on me The dust of snow From a hemlock tree Has given my heart A change of mood And saved some part Of a day I had rued. As you shall learn, the hemlock canopy is a canopy that traps and holds snow, the hemlock growing so slowly as to be on another clock from a birch tree or a fastgrowing pine, or a man beneath a crow. It seems impossible to separate the hemlock tree from the hemlock plant’s poison, for a poet to keep the death of Socrates out of the picture—for death is in the forest, especially a hemlock forest, especially now.

FOREWORD xviii

But the dust of snow falls from one timescape to another, bridging a gap, adjusting distances and, thus, adjusting endings. “You can hear it,” someone in our little group says, and at first I think they mean the quiet. The floor of a hemlock forest is the softest ground, and I never have a problem imagining a knight on a horse, or the opening of Macbeth or a film crew working with Game of Thrones disrupting the quiet—the risk of old fashioned B-movie time warp. I listen and at first think I hear the coolness, a stream of cold, rushing water, perhaps, the temperatures that stream life loves. We look up, through the dense branches and millions of needles and the smallest pine-like cones. Like the dust of snow, or a touch of gray, we see evidence of death: the eggs of the adelgid, white woolly patches, cotton-like eggs that cover the trees’ fine needles. Needles are brown, and on second glance more needles are white, the adelgids sucking the old trees’ life away. In certain places in the canopy we see what we would not see in a healthy hemlock forest—light from above. (In chapter 2, we learn, astoundingly, that the light on the healthy hemlock forest floor is one percent of the light at the top.) I keep listening to the sound I now know is not a stream, but it is not a stream at all. This is what the hemlock has to say, the last sounds it offers us, its final words. It is the sound of a thousand dying needles falling. It is the sound of a gentle rain, and it is steady, so that I can record it, memorize it, lock it in and carry it with me, as I mentioned, all the way home. A forest is leaving, gone forever, most likely, or for as long as anyone in our little group will be on this earth. And yet, when I play it back at night in my mind as I head off to sleep, I don’t feel sad, though it is a sad sound. Death is sad but death is not the end, even in a forest. Emily Dickinson— who was as close as Frost to the hemlocks I hear today—suggested that a part of us leaves when a friend dies, but she suggests an estuarine return. Each that we lose takes part of us; A crescent still abides, Which like the moon, some turbid night Is summoned by the tides. When I listen to the hemlock’s sad rain, play it over in my head, I feel thicker, like a hemlock, more worn. I feel the still-cool air of the hemlock woods, and I imagine the ground, the cool ground of a forest that will once again undergo a change, a shift, a dynamic transformation that is like so many it has undergone over thousands of years of different lives—a high or a low tide, depending on how you look at it.

FOREWORD xix

This page intentionally left blank

P R E FA C E

[Hemlock] . . . the sublime of trees, which rises from the gloom of the forest like a dark and ivy-mantled tower. —Thomas Cole, Essay on American Scenery

N

ow is the right time to celebrate eastern hemlock, a truly distinctive tree whose natural history has been so intertwined with human history. While it is no coastal redwood or giant sequoia in grandeur or Douglas fir in economic prominence, hemlock has no equivalent in the eastern landscape. Dominant but vulnerable, useful but undervalued, hemlock has inspired poets and intrigued ecologists. It has its own set of tales and lessons for humankind. For more than a century, scientists at the Harvard Forest in Petersham, Massachusetts, have been fascinated by this long-lived evergreen, and their efforts are the heart of this story of nature, conservation, and the pursuit of science. Of all the tree species in our eastern forest, hemlock suffered the heaviest decline from the landscape destruction that followed the arrival of Europeans in North America. It has had a reprieve because the intensity of logging and farming has abated for the past two centuries; in parallel with the reforestation of much of New England and the eastern United States, hemlock has been steadily recovering its revered place in our woods. This recovery should be cause for celebration, but instead it is a time of sorrow, for today the magnificent trees are once again imperiled. This time an insect—the hemlock woolly adelgid—may succeed in accomplishing what axes, saws, sheep, cattle, and fire never could: eliminating the eastern hemlock as a fully functional species across much of its range in eastern North America. Thus our primary task with this volume is to draw attention to a cataclysmic shift taking place in our landscape. In doing so, we recognize and offer tribute to a truly distinctive and powerful force in our forests, a distinctive species that shapes

xxi

our streams, wetlands, and broader environment, and we sing its praises now while we still have it in our woods to observe, reflect on, and admire. Hemlock provides us with a natural centerpiece for discussion because its survival is threatened today. But the rationale for focusing on hemlock is much greater than convenience, and hemlock’s rich story involves more than a singular species. Naturalists, writers, and artists of all stripes have found it magical in its effect in the woods and on the human spirit. For ages, hemlock has served as a metaphor in the work of novelists, poets, and artists alike. From Robert Frost’s snow dust to the home of James Fenimore Cooper’s Mohicans and Washington Irving’s Devil, hemlock boughs, glades, and ravines have lent an exhilarating and foreboding mood to scenes of raw wildness. Now, while it disappears from our view, hemlock serves as a new symbol of the losses to our woods as they are stripped of species after species— first chestnut, then elm and butternut, and gradually beech, now hemlock, soon white, green, and black ash, and, in growing numbers and places, oak. Species disappearance might seem like ancient history, as with the chestnut or passenger pigeon, but we will undoubtedly experience such losses with increasing regularity as the pace of climate and landscape change quickens and the role of exotic organisms increases. Thus through the travails of hemlock we can address a broad question that looms for all in an age of increasing extinctions: do solitary species matter? We believe that they do. There can be no better test for this proposition than hemlock, which plays such a foundational role to so many different woods and ecosystems. Its decline and loss offer us a natural experiment with which we can seek answers to this fundamental question about nature. Hemlock today prompts questions that extend even further, beyond the role of individual species to our very interpretation of and care for nature. As ancient hemlocks die, vast stretches of our eastern landscape are being transformed. Decades or centuries of seemingly stable or slowly changing conditions are upended in a few short years as big trees fall, warm sunlight streams into sites shaded for decades, and entire ecosystems disintegrate. To most eyes, this and any unfolding scene of physical damage to nature is one of doom, disaster, and despair. The loss of hemlock is a truly sobering event. Yet our natural world has always experienced fire, flood, gales, and tectonic upheavals. Is the present situation or any of these other events truly a disaster to the ecosystem of plants, animals, rock, and soil that we call a forest? After all, hemlock has fallen to the axe and plow over recent centuries, it nearly died off in earlier millennia from some ancient combination of drought and plague, and like all of our northern life, it returned to thrive in a land once buried deeply below a glacier’s snow and ice. Does nature completely unravel in the face of a new challenge such as the adelgid? PREFACE xxii

The natural experiment that we are witnessing yields an opportunity to assess not only how nature deals with upheaval but how we humans respond to such events. In our self-absorption we may assume that the damage we tally in dollars of timber or in beauty and solitude is paralleled by some enduring damage to nature. Further, in our hubris we may assume that we have the knowledge and capacity to assist nature in its recovery from the loss of trees attacked by insects, flames, or wind. There are good reasons to doubt such thinking and to challenge such inclinations. If we could adopt an ecocentric mode of thought, we might argue that such apparent destruction is not only part of nature’s history and therefore a transient phase, but also, in surprising ways, often beneficial to many parts of nature. We might see that, amid the death and apparent chaos, the different, new forest that begins to emerge is fully functional and offers new opportunities for all kinds of life. Hemlock’s demise allows us to evaluate how we and nature cope with apparent calamity. Can we help, and does nature need our assistance? Or is our interpretation of nature’s needs incorrect, and do our actions simply exacerbate conditions that we seek to repair? Are the values that we project on nature’s upheavals warranted for nature itself, or should we restrict our concerns to those features that directly affect us? We recognize that many practical as well as philosophical questions arise from our exploration of this great eastern conifer. Hemlock’s dilemma may provide insights that can aid society as we grapple with the many tumultuous changes that occur in nature and our land. Beyond the utilitarian lessons that may emerge, there are many personal motivations for this volume and its form. In central Massachusetts there lies a quiet and unusual institution, the Harvard Forest, a department of Harvard University dedicated to the study of New England and the global environment. It has served for more than a century as a base for students and scholars who live not in Cambridge but rather in the heart of the New England landscape that they study on a daily basis. The Forest was conceived in the late nineteenth century with the same mission that it carries today—research, education, and the demonstration of lessons gleaned from the natural world—and with the same singular model of a community of scholars who are a true part of the system that they study. Its 3,750 acres of woods, streams, wetlands, and farm fields are typical of the larger landscape, and provide a laboratory and classroom where researchers, administrators, students, Woods Crew, and other staff live and extend their work. Rather than episodically descending on a remote field station on weekends and summer months or restricting field studies to good weather, the team of collaborators at the Harvard Forest works daily—in fair weather and otherwise—at their administrative home and intellectual nexus in PREFACE xxiii

Petersham. The daily trip to work takes each through his or her own part of their larger New England laboratory. This unique model for environmental research and education was conceived, when American forestry, ecology, and conservation were young, by individuals who were among the first to advance the notion of hands-on experience for scholars through science field camps and summer schools. The Harvard Forest was established in a colonial farmhouse in Petersham in 1907. Its operational model was strengthened when a new building was constructed in the 1930s to serve students, scientists, and visitors and to house an intimate museum whose mission is to convey lessons learned in the acres of natural classroom outside. It was a model that encouraged natural immersion under all conditions—the brilliance of summer, the deluges of spring, and the ice and deep snows of winter. It also encouraged its participants to share their knowledge and passions on municipal boards, in town meetings, at schools, in local land trusts, and on the sidelines of playing fields, where they could participate in the cares and concerns that everyone in this region shares. It was in this setting and through our weekly meetings, hallway chats, and over breaks in field and lab work that this book emerged. We recognized the precarious position of eastern hemlock as an opportunity to explore many important issues, and we began to engage in an extended discussion of a collaborative volume. For each of us, hemlock has been part of our growth as scientists, and we have explored it in diverse realms from physiology and ecosystems to history and paleoecology and from ecological modeling to forestry. Collectively, we have the breadth and depth of scholarship, as well as the intimate personal experience with hemlock, to explore it quite thoroughly. Through these discussions it became clear that we could provide an unusual perspective on the species and our relationship to it. Moreover, we recognized that we could draw from an unparalleled asset in these efforts: more than a century of observations, musings, photographs, and intense studies by our predecessors at the Harvard Forest. These men and women recognized the many values of history to the interpretation and conservation of nature. They recognized that history could be read from many sources including nature itself, and that this information provided insights into nature today and the way it is changing. Consequently, they sought to document every clue to the past so that they might better understand the present. They also recognized that one of the greatest values of their work was the growing archive of records that they were contributing to about their forests and the broader landscape, charting the ways in which both changed over time. With this in mind, they established study points and plots as well as long-term experiments that they measured and remeasured and managed for the future. In both their pursuit of site history and their investment in intensive studies, they recognized that the only thing that made this typical landPREFACE xxiv

scape unusual and valuable was what they learned about it, and the detailed records that they produced and retained for future use. From the beginning in 1907, hemlock was a central part of their landscape and focus. Modest initially, both in stature and in information provided to scientists, these woodlands and the accompanying investigations grew. The inquiries of a constantly changing group of scholars and students led to invaluable insights, innovative approaches to science, and stories that we seek to share. As we began to write this tale of hemlock, we made some critical decisions about our approach. We knew we wanted to tell a full story of the species. We also committed to exploring broader themes, using hemlock as a metaphor and exemplar for many of the larger issues in ecology and conservation that are challenging us today. Given our own depth of involvement, the nature of the close community that we work in, and the wonderful lessons that the history of this place and institution has given us, we also resolved to tell a personal story. Rather than attempting to assemble everything that is known about the species all across its range, we chose to tell the story that we know intimately and have at hand. Thus, this is an exploration of hemlock through our own lives as scientists and those of our predecessors at the Harvard Forest. It is very much a New England– bound story and one rooted centrally in Petersham, Massachusetts. Yet we also cast a broad net that reaches across the region and extends far back in time. This personal approach extends to our mode of presentation. As scientists we are accustomed to writing dense, academic papers for other scientists. We’ve consciously avoided that style in telling this story of hemlock. For those who seek reassurance that our statements are grounded in solid science, we provide references to our sources in the bibliographic essays and accompanying citations. The principal authors of each chapter are also identified in the bibliographic essays. We also chose to highlight the people who lie behind much of this science. We have included intimate human stories because they reveal an important lesson: so much of science is dependent on the personalities involved. After all, individuals choose the subjects they study, they inject inspiration and ideas into their work, and they embrace or choose to ignore the new insights that emerge from their inquiries. Discussing science without this personal dimension ignores the larger and often intriguing story of how studies were undertaken and how lessons were actually learned. For that same reason, we do not hide ourselves unduly from the text. We choose to tell many of these human and historical tales in a series of “Lessons from Harvard Forests and Ecologists.” Although they often weave in and around the central theme of hemlock, we have strived to make all the strands fit within the book’s narrative and focus. Looking back, we realize that we started out thinking we were compiling a hisPREFACE xxv

tory, an overview, and a fitting farewell to a singular species. We penned much of the first draft in that vein. Yet as we discussed the emerging work with colleagues in our weekly Forest meetings and picked the brains of our Bullard Fellows (a cohort of a half dozen scholars on sabbatical leave each year at the Forest), we could see that we were engaged in something broader and more personal. We were telling a larger story of the New England landscape and its history, ecology, and conservation. The book was about hemlock, but it was also about an unusual institution and the peculiar way that it conducts long-term research, rooted in a single place yet extending much further. In delving into that place and its roots, we produced a story about the many people who preceded us and whose work we all carry on. Ultimately we could not downplay the emotional attachment that we share for the people, the place, and hemlock. While this is a story that emerges from science, we are individuals whose lives have been continually shaped by the distinctive tree that is the focus of this volume.

PREFACE xxvi

A C K N OW L E D G M E N T S

O

ver the years and decades many people have contributed to the research and ideas developed and presented in this volume. Among our wide-ranging science colleagues, the work referenced in the bibliographic essays provides a testament to the debt owed them. At the Harvard Forest, many individuals have contributed in key ways to our studies and this book, including Emery Boose, Jeannette Bowlen, Jessica Butler, Laurie Chiasson, Betsy Colburn, Brian Donahue, Elaine Doughty, Edythe Ellin, Ed Faison, Richard Forman, Lucas Griffith, Clarisse Hart, Jenny Hobson, Dave Kittredge, Oscar Lacwasan, Kathy Lambert, Ron May, Glenn Motzkin, Liza Nicoll, John O’Keefe, Julie Pallant, Manisha Patel, Lisa Richardson, Kristina Stinson, Mark Van Scoy, and John Wisnewski, head of our modern-day Woods Crew. Special thanks are due to Linda Hampson, who worked closely with us throughout this effort and then promptly retired at its conclusion, and to Brian Hall, who provided the maps and considerable help with field and historical studies. Bullard Fellows Andrew Bennett, Martha Hoopes, John Roe, and Duncan Stone provided valued broad perspective and joined with the rest of the Harvard Forest Lab Group in hashing through early chapters and contributing countless ideas and suggestions. Core funding for all of our efforts comes from Harvard University, greatly strengthened by twenty-five years of major support from the Long Term Ecological Research program of the National Science Foundation (NSF). Additional research support has come from the Andrew W. Mellon Foundation; the NSF Ecosystems, Ecology, Small Grants for Exploratory Research, and Populations programs; the U.S. Department of Agriculture’s National Research Initiative, Focus Funding, and National Institute of Food and Agriculture (NIFA) Agriculture and Food Research Initiative; the National Aeronautics and Space Administration (NASA); the Smithsonian Conservation Biology Institute, the Smithsonian Institution Global Earth Observatories, and the Smithsonian Tropical Research Institute’s Center for Tropi-

xxvii

cal Forest Science; the U.S. Department of Energy’s National Institute for Climatic Change Research; the Highstead Foundation; the Blue Hills Foundation; and the Arnold Arboretum. A major portion of our research activity in the past three decades has been provided by our largest educational program: the three-month-long Harvard Forest Summer Research Program in Ecology. Every year this program engages thirty or more undergraduates from Harvard and other educational institutions across the globe in hands-on studies with a mentor, other undergraduates, graduate students, and postdoctoral researchers. This program receives funding from the NSF Research Experience for Undergraduates program, NASA, and Harvard University. We extend special thanks to Jean E. Thomson Black, executive editor at Yale University Press, for her boundless enthusiasm and great advice, and to Sara Hoover for many insights and much assistance.

ACKNOWLEDGMENTS xxviii

HEMLOCK

ONE

HEMLOCK’S FUTURE IN THE C O N T E X T O F I T S PA S T We had lain in hay and talked and shot sparrows with an air-rifle when they perched in the triangle cut high up in the wall of the barn. The barn was gone now and one year they had cut the hemlock woods and there were only stumps, dried tree-tops, branches and fire-weed where the woods had been. You could not go back. —Ernest Hemingway, A Farewell to Arms

T

he long-term history of the forests of the northeastern United States is one of resilience, loss, and recovery. No species—not even American chestnut— captures those dynamics, or the associated lessons for ecology and conservation, like eastern hemlock. Long before chestnut was crippled by an exotic fungal blight, indeed thousands of years before chestnut migrated onto the New England scene following the last ice age, hemlock suffered a range-wide collapse and then recovered as a dominant species. By the time European explorers arrived in the New World, hemlock and American beech, its broad-leafed counterpart in shade tolerance and longevity, dominated much of the ancient forest that blanketed the Northeast. In the absence of extensive disturbance from fire or the small population of native people, forests were shaped by infrequent smaller disturbances—wind, ice, insects, and disease—and many trees grew to massive size and old age. This magnificent wild landscape presented both challenge and opportunity to the arriving colonial farmers and their industrially minded successors. From the shores of the Bay Colony inland, the woods were cleared for grazing animals and crops, burned to remove the staggering age-old accumulation of woody debris, and cut rapaciously for timber, firewood, and even for the tannin in hemlock bark, an indispensable ingredient in the production of leather. Swamps were drained; brooks, streams, and even mighty rivers were tamed. In the forests that were cut but not cleared for agriculture, the woodland com-

1

position shifted toward species resilient to repeated use and abuse. Hemlock and beech suffered disproportionately from this human onslaught. Unlike white pine, its coniferous counterpart, hemlock grows and matures slowly. It lacks the ability to sprout or expand rapidly through widely dispersed seed like its hardwood competitors oak, birch, and maple. Its roots are vulnerable to trampling by livestock, and it suffers mightily from fire, which kills it easily and can eliminate it locally. Hemlock declined precipitously in a wave that commenced at the coast and then over the next three centuries spread with human settlement into the eastern mountains and then north and west, eventually to the Great Lakes region. Yet, despite the depredations, hemlock’s woody resilience prevailed. Its saplings endured in the shade of more aggressive companions, and large groves persisted within those landscape refuges—ravines, mountain slopes, and swamps—less suitable for human exploitation and settlement. An eventual opportunity for recovery came in an unexpected fashion. The growth of the nation and its human population brought relief to the beleaguered eastern landscape. As agriculture and timber production shifted westward onto newly acquired lands with deep fertile soils and untapped virgin forests, New England and much of the expansive eastern United States was freed from intensive use and began a second transformation. New Englanders moved off established farms and down from rural hill towns into the newly forming valley mill towns and coastal cities, where water power and access to shipping and railroads began supporting an industrial economy and waves of new immigrants. Trees spread out across the hard-won pastureland and hay fields as the sheep and cows and mowing declined. Forests began to recover across much of the East. At the tail end of this wave of renewed green that spread across the landscape came hemlock. Gradually and relentlessly, the long-lived champion of the untamed woods emerged from its subordinate position in woodlots and forest fragments and began to reassert itself. Because of its slow growth and the limited dispersal of its seeds, hemlock is among the most sluggish of species to recover and expand across the landscape in the aftermath of intense disturbance such as land clearance or fire. Consequently, our newest young forests are largely devoid of hemlock. Wherever magnificent groves of large and old hemlocks exist today, the site has likely supported forests continuously, since long before colonial times. Hemlock’s current distribution and abundance are clear examples of ecological legacy: a feature in the modern landscape that reflects environmental conditions and processes that operated decades or centuries ago. Its presence can therefore be used to identify those sites that were spared from land clearance, the plow, widespread grazing, and fire. While they may have been cut, and even heavily, these areas remained wooded and retained most of their native species through centuries of tumultuous change. In contrast, forests having HEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 2

A hemlock forest on the Prospect Hill tract of the Harvard Forest, with its distinctive open and heavily shaded understory. (David Foster)

only small or young hemlocks in the understory most likely were cleared of forest and farmed historically. This history of forest clearance and recovery is often confirmed by adjacent stone walls, cellar holes, traces of ancient plowing in the soil, and other evidence of the agrarian past so prevalent in New England’s woods. Hemlock is also often the dominant species in the very few old-growth forests that remain in the region—areas of old and often immense trees that were never directly disturbed by human activity. In these ancient forests hemlock frequently fills the canopy and bides its time in its own deep shade. For the last 150 years and since the days of the Civil War, hemlock has reasserted its way into these expanding forests and back toward its former role of prominence in the northeastern landscape. The expansion of hemlock has had great consequences, because it exerts a strong influence over the local environment, other species of plants and wildlife, and human enterprise. By producing shady, cool conditions and laying down a thick carpet of needles on the woodland floor, hemlocks structure forest ecosystems from top to bottom, controlling the environment of the streams, ponds, and wetlands that they contain. By creating a unique suite of habitats, hemlock is a quintessential example of a foundation species, an abundant plant or animal that controls the characteristics of an ecosystem and exerts far greater influence than its simple numbers may imply. In our day hemlock has once again become ascendant in eastern forests. It is widespread, common, frequently dominant, and ever distinctive. Based on ecological characteristics and history, we should be able to expect this trajectory to continue apace, and hemlock should thrive and expand relentlessly as our forests age and mature. In our rewilding landscape, where beaver, bear, bobcat, and moose are now becoming familiar if not everyday sights, we fully expect the greatest arboreal symbol of wild eastern woods to thrive as well. However, like the remake of an epic movie (and some five millennia after its prehistoric collapse), hemlock is being crippled once again. An exotic aphid-like insect that weakens and gradually kills the tree has arrived from Asia, made its way through our coastal ports, and is spreading out across eastern North America. The hemlock woolly adelgid jumped Long Island Sound in 1985 and began to expand its way northward and then westward through New England. The insect’s range in 2013 reached from southern Maine through New York down to Tennessee and across to Georgia. Relentlessly, the adelgid is expanding its scope and taking control over the fate of hemlock. As this tiny insect spreads, it is sucking the life from this forest giant. By extracting this single tree from the landscape the adelgid is altering our forests in a most profound way, from backyard hedgerows and our maturing second-growth forests to the most magnificent ancient woods. The short-term impact of the adelgid on a hemlock forest varies considerably with location, but the long-term consequences HEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 4

Trees killed by the hemlock woolly adelgid gradually lose their branches and collapse to the ground. (David Foster)

are a consistent and sobering tale of decline and loss. Mortality has been swift in the Southern Appalachians, where the region’s mild winters and long growing season work to the advantage of the rapidly reproducing insect. In northern regions, occasional bitter winters and frigid cold snaps have episodically checked the spread and growth of the adelgid. In these areas the insect’s expanding impact has progressed in fits and starts, and the hemlock’s decline has been more of a prolonged spiral. Nonetheless, wherever the adelgid appears, the larger story line is the same. Hemlock shows little resistance to this novel foe, and mortality progresses across every site and among all sizes of trees. Mighty trees fall, just as stunted saplings drop their needles and wither away. As they succumb, the deeply shaded woods are inexorably converted to starkly open groves. And through this process, new ecological legacies appear: residual “ghost” trees with thin and sickly graying foliage and towering barkless skeletons, bleaching white and gradually disintegrating in the open sun. The world is losing hemlock and becoming more homogeneous. Across hill slopes, ravines, and coursing streams, landscapes once broken and diversified by the distinctive environment of hemlock are assuming a simpler appearance and altered function as hardwood trees take this evergreen’s place. Hemlock once interrupted the expanses of oak, maple, and birch, offering the deepest shade year-round and welcome snow-free islands of shelter in the winter. As hemlock declines, this subtle forest variation disappears. Our woods are now that much bleaker, poorer, and less verdant when deciduous leaves drop and the snows of winter fall. No effective means of combating this invasive scourge has appeared, despite substantial effort and many assertive claims to the contrary. In the news and press releases from public agencies, university media offices, professional conferences, and cocky researchers, we read optimistic claims that diligent science, good forestry practices, and American ingenuity will whip this exotic foe. Yet, as with the parallel claims for the American chestnut many decades ago, so far all are just so many words. We remain helpless in slowing the spread of the adelgid or constraining its impact in the region’s forests. Individual trees and backyards can be sprayed to control the insect, though this comes at considerable cost and with potentially grave and poorly quantified effects on other species and the environment. Major efforts have been launched to identify, rear, and release native predators from Asia, but over the decades these have had no demonstrable success. Unimpeded, the insect progresses in its relentless spread, and the hemlocks in our woods continue to disintegrate. As with the chestnut, the grim reality in the woods pays no attention to the promises of researchers, managers, or their spokespersons. The story line with hemlock is indeed reminiscent of that for American chestnut, but with one critical difference. Chestnut is a hardwood and among our most

HEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 6

The range of eastern hemlock and distribution of the hemlock woolly adelgid as of 2012. (Brian Hall)

prolifically sprouting trees; when its leaves, branches, and stems die, new shoots emerge from the surviving roots. This process can continue essentially indefinitely, ensuring that the chestnut endures in place. As a consequence, billions of chestnut continue to persist in the eastern woods. Each of these thriving sprouts grows for a decade or so until it reaches three to four inches in diameter; then the blight takes hold and chokes the stem dead, and the sprouting cycle begins again. Thus this once magnificent tree survives in a completely new status as a shrub or small tree in the

HEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 7

Illuminated leaves of the American chestnut in its new role as an understory tree. (David Foster)

forest understory. Hemlock, on the other hand, has no such recourse, for it cannot sprout. Like most conifers, when its foliage falls the tree dies completely. As a consequence, hemlock’s decline will be much more complete than that of chestnut, and its ongoing survival in the region will depend on plants that either escape the insect or persist in some as yet unwitnessed and substantially diminished capacity, tolerating the infestation. So far, some scattered trees have lived on for years—even decades—as graying ghosts whose long-term prospects seem all too clear. Overall, the future is indeed bleak for this grand species of the eastern forest. Now, then, is an appropriate time to learn all we can from hemlock’s past, present, and future. The poignancy of hemlock’s situation—reemerging to prominence following centuries of abuse and exploitation, only to be struck by an unforeseen and essentially invisible foe—brings the future of all of our forests into question in this time of great environmental changes. At the grandest scale, this new saga forces us to broaden our thinking about issues of forest conservation. We recognize from this tale the vulnerability of any of our species—plant, animal, and microbe—and the ecosystems they comprise. As scientists, we have concluded that the single best approach to ensuring that our tree and wildlife species and natural communities persist and can cope with future stresses and disturbance is to conserve vast quantities of forest in a diversity of landscape settings. This approach confronts these threats with as little environmental change as possible. Since so many resources and qualities of life are tied to forests, this broad conservation agenda is as vital to humankind as to nature itself. The advancing loss of hemlock before our eyes also provides a grand, albeit heart-wrenching experiment in forest ecology and conservation, where we can grapple with and better understand many issues that confront science and society. By concentrating on this singular species, we can achieve a focus that eludes broader, more holistic studies. As we examine hemlock today, we can address important questions and synthesize insights unknown to the early ecologists who lived through the decline of chestnut a century ago. We can use our position in ecological time to appreciate more fully this unique species, our forest ecosystems, and their connections and responses to the perturbations wrought by nature and humankind. We can compare the unfolding decline of hemlock to that of chestnut in the early twentieth century and of hemlock twice before: 5,000 years ago and then following European settlement. What makes this long-lived conifer so unusual, and with what specific attributes has it shaped our forests and landscape? To what extent and in what ways are ecosystems affected by the loss of such a singular and foundational species? How do other plants and animals respond and reassemble following such impacts, and to what exHEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 9

tent are the original conditions recovered? Most broadly, how does nature respond to what we characterize as catastrophic events? Are forests “impaired,” “crippled,” or “destroyed” when they experience wind, fire, ice, insects, or disease, or are these terms the emotional response of humans to changes that species and nature take in stride over time ? As we evaluate these questions with regard to hemlock, we can consider the changes our land has experienced over deep time, is undergoing today, and may confront in the future. And we can ask how best to apply the resulting lessons to current and future environmental challenges. Hemlock provides a compelling record of change. We can track its lengthy saga through pollen records many thousands of years old and in tree rings that capture its persistent waiting in the shadows for the opportunity to grow. Assembling and drawing lessons from diverse sources and perspectives affords more than a simple story of a species in decline. It sheds a powerful light onto the history of the American landscape, science, and conservation; it initiates and engages discussions of grand concepts in ecology; and it explores the past, present, and future of the distinctive landscape that we call New England.

HEMLOCK’S FUTURE IN THE CONTEXT OF ITS PAST 10

TWO

AN ICONIC SPECIES

This is the forest primeval. The murmuring pines and the hemlocks, Bearded with moss, and in garments green, indistinct in the twilight, Stand like Druids of eld, with voices sad and prophetic, Stand like harpers hoar, with beards that rest on their bosoms. —Henry Wadsworth Longfellow, Evangeline

N

o other tree species in our eastern landscape exerts such a widespread and profound influence on the environment and other organisms, including ourselves. When we enter a hemlock forest on a sunny day, we sense a change and immediately recognize that we are in a different and special place. The light dims perceptibly, the wind dissipates, and the temperature drops. The ground beneath our feet turns soft and spongy, and the understory opens up around us as the abundance of other plants declines. Stillness reigns. The wooded environment quiets, and singular sounds stand out clearly. Such is a hemlock forest. What is it about this tree that provides this memorable experience and shapes the woods in such a profound manner? The secret lies in the smallest part of the tree—the leaves—and the structure and qualities that these impart to it. The small, flat, and gently flexible needles are densely arrayed horizontally to the stem in a manner that optimizes the exposure of their deep green surface to the sun. The drooping, dark branches form successively overlapping patterns that are effectively designed to intercept light. As a result, almost every ray of incoming light—direct sunlight, sun flecks, and the dimly filtered and indirect glow—encounters a dark green surface to absorb it. The density of hemlock foliage is the consequence of a simple adaptation. Most trees drop their lower branches as a forest grows and the light lessens beneath them. Pines, oaks, and birches, for example, concentrate their foliage high toward the sky;

11

Needles and cones in the upper branches of a hemlock tree. (David Foster)

hemlock, on the other hand, does not need to self-prune because its leaves function so well in low light. As a result, hemlock leaves fill the view from below, magnifying the perception of hemlock dominance and allowing almost no sunlight to reach the ground. This deep shade, augmented by widely spaced columns of large tree trunks and dense canopy overhead, creates the distinctive, cathedral-like atmosphere of hemlock woods. Though seldom the tallest or most abundant tree, hemlock always shapes the scene and dominates the view. In this deep, cool shade, few other plants or animals busy the woods. A century ago, scientists recognized that hemlock’s great shade tolerance comes from its unsurpassed efficiency in light absorption, especially at low levels. In recent decades, this quality has been thoroughly quantified, as have its consequences for many aspects of the forest environment. Typically, the light level beneath hemlocks is only 1 percent of that above the forest. In this 99 percent shade the temperature drops as much as ten degrees in the canopy and another five or ten degrees near the ground. The layers of feathery branches also effectively intercept incoming rain and snow, leading much of it to evaporate back into the atmosphere and greatly reducing the moisture that reaches the ground. Consequently, a shadow of drought develops beneath every hemlock during a summer shower or powdery snowfall. Through the winter this effect is cumulative, and substantial differences in snow depth can develop within and beyond a hemlock stand. Yet, surprisingly, the snow that does accumulate often persists longer beneath hemlocks than in adjoining hardwood forests, due to the cool environment and slow rate of snow melt beneath the conifers. The characteristics of the needles also have important consequences. In contrast with the rigid, sharp needles of spruce, hemlock leaves are flat and bendable. This flexibility reduces breakage and lends a soft feel to the foliage; it was this quality that led many an early outdoorsman to seek hemlock for soft bedding at the end of a grueling day in the woods. Meanwhile, the distinctly refreshing smell that pervades these woods comes courtesy of a chemist’s delight of volatile compounds exuded by hemlock needles, including tannins, terpenes, phenolics, and essential oils. Armed with this knowledge of the physical and chemical signatures of hemlock, we can experience these woods with new eyes—and noses. The deep and spongy forest floor results from the progressive accumulation of a thick layer of needles that can be more than a foot deep and hundreds to thousands of years old. In our forests of oak, birch, or maple, only the thin layer of the previous year’s leaves persist. The thick layers beneath the hemlock are a consequence of the leaf chemistry and cool, moist environment that inhibit the organisms involved in decomposition. In its extreme shade tolerance, hemlock is rivaled only by two other eastern trees: American beech and sugar maple. Each is able to persist in the deep woods by AN ICONIC SPECIES 13

A deep layer of needles and organic matter on an uproot mound in a hemlock forest. The white disk is used to sample gases emanating from the soil. (David Foster)

capturing errant sun flecks, the infrequent short periods of three seconds to thirty minutes when direct light shoots down through gaps in the many layers of foliage. The challenges to such a growth strategy are immense. Each individual leaf must be able to increase its light-gathering activity rapidly for brief periods and then efficiently synthesize food resources and convert these into growth. Then, as the light dims, the tree must abruptly shut down most of its maintenance operations until the next bit of light abruptly arrives. For seedlings and saplings, this leads to lengthy periods of near stagnation interrupted by episodic bursts of capturing light. In this struggle to eke out an existence in dim light, the evergreen qualities of hemlock convey major advantages over its hardwood competitors. So do the physiological adaptations that protect hemlock leaves from freezing, allowing them to photosynthesize at extremely low temperatures. Even in deep winter, whenever liquid water is available and the air temperature exceeds freezing, hemlock can capture available light and produce additional food. This capacity yields opportunity throughout the dormant, leafless season of hardwoods, and it is especially valuable each spring when warming and lengthening days, abundant moisture, and high light levels yield productive conditions for growth—for those trees with leaves. By the time most hardwood leaves are emerging each year, hemlocks are already peaking in growth. Hemlock’s phenomenal ability to survive deep shade is readily apparent in the annual growth rings on almost any stump. We recently cut down a four-foot-tall hemlock sapling to estimate its age and compare its growth to that of the overhanging trees. The growth rings were so narrowly arrayed that we couldn’t possibly count them in the woods, so we cut a small cross section to examine back in the lab. After we sanded the one-inch-diameter piece of wood to create a smooth surface, we peered at it under a dissecting microscope. What a surprise: sixty rings—sixty increments of annual growth—in less than one-half inch. This remarkably slow and dense pattern of wood production demonstrated that this sapling was just barely capturing enough light and other adequate resources for survival. Over half a century, the plant had shown almost no significant growth. We often see a similar pattern when we use an increment borer to extract a pencil-thick core from older trees to examine the growth pattern without cutting the tree down. In most old hemlocks, there is at least one period in which the tree grew less than an inch in fifty to eighty years. In contrast, an oak, pine, or tulip tree growing in full sunlight will add up to one-quarter to one-half an inch of diameter each year and will seldom display such lengthy periods of slow growth. Hemlock also has a notable flexibility to switch abruptly, adding girth and height when light becomes available. Such plasticity in growth in both young and old trees is highly unusual. Other species exhibit some flexibility as seedlings or AN ICONIC SPECIES 15

A patch of shrubby hobblebush amidst sunflecks and saplings in a hemlock forest. (David Foster)

saplings, but most lose that ability as they mature, and quite a few will die when confronted with a sudden increase in light. In sharp contrast, hemlock can undergo multiple bouts of extreme suppression and rapid growth throughout its exceptionally long life span. Hemlock’s efficient use of light does demand significant trade-offs in growth rate. The metabolic qualities that enable hemlock to operate well in low light restrain its ability to capitalize optimally on full sunlight. In open sites with abundant water, red oak, red maple, and black birch photosynthesize and grow at two to three times the rate of hemlock. The net result of the various trade-offs among these species is surprising. The total annual growth in hemlock and hardwood forests is often rather similar, even though the seasonal details are strikingly different. One way of making this comparison is to measure the seasonal and annual pattern of carbon storage—the amount of carbon from carbon dioxide captured by photosynthesis minus that lost through the plant’s respiration. Different tree species have predictable patterns of growth. Each spring, hardwoods begin to photosynthesize aggressively shortly after they leaf out. They then store large amounts of carbon in the peak summer months and decrease their uptake as the leaves prepare to fall. Hemlock, on the other hand, may absorb and store as much as one-half of its annual carbon during the spring and conduct relatively little photosynthesis in July and August, a period when temperatures and drought stress are often high. Both growth strategies are effective, but they may lead to different annual totals depending on the particular conditions in a given spring, summer, or fall. In years characterized by a warm, moist spring and autumn or extremely dry summer, hemlock forests can store more carbon in a year than the rapidly growing hardwoods. Some years, the tortoise beats the hare. The low rates of summer growth for hemlock come with one decided upside: less water use and loss from the forest. To take up carbon dioxide, the leaves of all trees have to open their pores to the outside. While this exposes the leaf cells to carbon-rich air, it also enables water vapor to escape from inside the leaf. To minimize moisture stress during the warm summer months when soils are dry, hemlock shuts its pores. At the Harvard Forest we use sap-flow sensors inserted through the bark to measure the rate of water movement in individual trees. These studies reveal that the maximum rate of water loss is two to four times higher in hardwoods than hemlocks. We observe similar differences for entire forests when we use instruments suspended above the forest canopy to measure the fluxes of water, carbon dioxide, and other gases between the forest and the atmosphere. Although hemlock intercepts and evaporates more rain and snow than the hardwoods, its overall rate of water lost to the atmosphere is typically half that of nearby forests of oak, red maple, and birches. Hemlock’s frugal use of moisture results in more water entering AN ICONIC SPECIES 17

The tower in the Prospect Hill hemlock forest provides access for studies in the forest canopy. (David Foster)

streams and groundwater, so these watercourses in hemlock forests maintain more consistent flow and temperature than those in other forests, especially in summer. Thus, while hemlock’s presence along streams and in wetlands indicates a preference for moist sites, its biology ensures that these sites remain moist. Hemlock’s wide-ranging effects on its environment create unique habitats for various aquatic and terrestrial plants and animals. While hemlock forests often support few understory plants, a few herbaceous species tolerate the deep shade and cool conditions and are common in these woods, including partridgeberry, wintergreen, Canada mayflower, wood sorrel, and the saprophytic Indian pipe. Some shrubs—hobblebush, witch hazel, and mountain laurel—can handle the impoverished conditions, but few tree seedlings can. Only the shade-tolerant species, such as sugar maple, American beech, and red spruce, can successfully bide their time beneath hemlock. Though the diversity of plants growing in a hemlock forest may be low, the deep organic soils provide a long-term repository for seeds and maintain an enduring legacy of the plants that have grown on the site over time. Long-lived seeds remain dormant in this “seed bank” until a disturbance disrupts the canopy or ground, increasing light levels and temperature and stimulating germination. We have explored the composition of seed banks at the Harvard Forest by excavating the surface organic layers, incubating these in a greenhouse, and carefully documenting the number and variety of tiny seedlings that emerge. Black birch and light-demanding plants including sedges, rushes, and raspberries are the most abundant species. Deep in the soil of one stand, we found the seeds of grasses and shrubs characteristic of open disturbed sites. These can only be legacies from the nineteenth-century logging and agricultural history of this particular site. Certain animals exhibit an affinity for the hemlock environment. Red-backed salamanders and red efts—the distinctive orange juvenile form of the red-spotted newt—thrive in the cool microclimate and often live under fallen wood, where they feed on flies, beetles, and other insects. The sight of dozens of red efts crawling about a hemlock forest after a summer rain confirms that one benefit of these forests is abundant moisture and a fairly consistent environment. Many soil arthropods, including mites, centipedes, and millipedes, dwell among the hemlock litter, along with several ant species that scurry across the ground and make subterranean nests. Careful observation in the complex array of branches often reveals many webbuilding spiders, along with many other easily overlooked insect species, including mites, harvestmen, bark lice, flies, and moths. These insects provide food for many bird species that spend at least part of their life cycle nesting or feeding in hemlock canopies. During late spring and early summer, hemlock forests are occasionally alive with a chorus of neotropical songbirds, AN ICONIC SPECIES 19

many recently arriving from winter homes in Central or South America. Blackthroated green warblers commonly feed and nest in the upper branches, and their spring call is seldom missing from any New England hemlock forest. Other birds include the Acadian flycatcher, which nests in riparian areas, colorful blackburnian warblers that perch in the upper canopy, Canada warblers, and melodic hermit thrushes. Meanwhile, distinctive songs and calls announce black-capped chickadees, winter wrens, and red-breasted nuthatches. In winter, ruffed grouse roost among the hemlocks. Finally, barred owls, northern goshawks, and red-shouldered hawks frequently nest and choose hunting perches in the branches of mature and old-growth hemlock forests. Many mammals spend time in hemlock, and the white-tailed deer often depends on it in a harsh winter. Deer congregate under hemlock groves for food, cover, warmth, and reduced snow depths. These “deer yards” are marked by large patches of trampled snow and abundant droppings where the animals have bedded down for the night. With reduced snow depths, deer expend less energy, which enhances survival during harsh winters. Red squirrels and mice seek out hemlock seeds, and snowshoe hares eat hemlock foliage. Porcupines eat hemlock bark, often damaging saplings and trees by completely stripping the bark and dropping branches to the ground where they are eaten. The most successful predator of porcupines is the elusive fisher, a weasel that uses these deep woods as den sites for raising their kits. Hemlock’s impact on wetlands, streams, and creeks is advantageous for many organisms. Brook trout depend on the cool water temperatures, leading to their early description as “hemlock trout.” The vivid coloration of these fish—steel blue sides with rows of scarlet dots—was attributed to the dark pools common in brooks that wind through these forests. Hemlock contributes to the stable flows and temperatures during the summer that help these fish survive. On the other hand, these streams are relatively poor in food because the shady conditions reduce algal production in streams, limiting the food base for many stream invertebrates. In addition, hemlock needles provide a low-quality food source for invertebrates, which are generally more diverse and abundant in hardwood-dominated streams. Insects common in hemlock streams are generally collector-gatherers such as mayflies, midges, and crane flies. Hemlock grows in a remarkable array of sites and environments, from uplands and streamsides to deep ravines and rocky outcrops. Remarkably, despite the abundance of data available from satellites and geographic information systems, we still lack accurate, large-scale maps for our tree species. To rectify this situation we have begun mapping hemlock’s presence from aerial photographs. We can easily differentiate hardwoods from conifers if the photos are taken during fall and winter when AN ICONIC SPECIES 20

A young moose on the Tom Swamp tract of the Harvard Forest. (David Foster)

the leaves are down. And it’s not difficult to distinguish hemlock’s crown from that of white pine, the other dominant evergreen in southern and central New England. As a result, by working our way through image after image across every acre of forest, we can identify individual hemlock stands in aerial photographs and create detailed digital maps of its distribution. Working with collaborators at the University of Massachusetts, we created a distribution map for the nearly 3,000-square-mile region extending from Long Island Sound through central Connecticut and Massachusetts to the Vermont border. In the process, we identified more than 6,100 hemlock stands covering 253,000 acres, or slightly less than one-third of southern New England. The map highlights great variation in hemlock distribution, including an expected increase in abundance to the North. From scattered, individual small stands in southern Connecticut, hemlock increases in central Massachusetts and northward, where it constitutes the dominant tree cover in expansive tracts, often accompanied by white pine and northern hardwood species. When we overlay this distribution on a map showing topography, hemlock exhibits a clear affinity for sheltered locations. We found it most commonly on slopes facing north to northwest—sites that receive less direct sunlight and tend to be cool and moist. Field visits to these sites have told us more about their individual characteristics and histories. We have confirmed that hemlock commonly occurs in wetlands and riparian areas, and on ridge tops. Land use history reveals that it grows mostly in former woodlots. Thus a number of different factors seem to favor the presence of these forests, including moisture, poor soils, and a low intensity of historical disturbance by human activity and fire. Hemlock has preferences, but it is also perfectly capable of growing across a wide range of sites and conditions in the landscape. Despite its great abundance in New England, hemlock never attained the exalted economic status of white pine, a premier timber tree for four centuries. Hemlock wood is soft and light like pine, but notably weaker. It also suffers from a structural defect called ring shake, in which the wood separates lengthwise, leaving cracks in the resulting lumber. Shake reduces the strength and integrity of the wood and lessens its aesthetic quality and usefulness in finished products. Originally shake was thought to be induced by wind and repeated swaying of the tree. A Harvard Forest study from the 1950s corrected this notion by demonstrating that the cracks result from growth stresses within the wood that cause its longitudinal separation along the annual growth rings. That research examined sixty harvested hemlock stands in north central Massachusetts and concluded that shake more commonly afflicts large-diameter and older trees. Hemlock wood also tends to “check” or split along the radius of the trunk as it dries. Despite those shortcomings, hemlock is

AN ICONIC SPECIES 22

The distribution of hemlock in the study transect that runs from Long Island Sound in Connecticut to the Vermont border. (Brian Hall)

widely available at a low price, so it is commonly used in the construction of outbuildings such as barns, sheds, and garages, as well as custom post-and-beam homes. The wood is easy to work, holds nails and spikes well, and is well adapted for framing. It was also used for railroad ties. The single largest and sustained contribution of hemlock to the forest products industry is as pulp for the manufacture of paper. Even here, however, hemlock does not yield a premier product, due to the short length of its wood fibers. Instead of being used for magazines or books, hemlock pulp is largely employed for newsprint, wrapping paper, and similar lightweight products. Hemlock did, in fact, shine in one marketplace. Its bark had high value historically and was in great demand as a rich source of tannins, a critical ingredient in the process of tanning. The bark was stripped from the tree, ground up, and then soaked in water to leach out the valuable chemicals. Tanning was one of the oldest and most important colonial industries, with tanneries rivaling sawmills and gristmills in abundance on streams that provided water power. In the eighteenth cen-

AN ICONIC SPECIES 23

tury, most towns in New England had at least one tannery, and by the latter half of the nineteenth century tanning had emerged as a major industry along the rivers, fueling the growth of the large urban centers of the Industrial Revolution, including the towns of Lynn and Lowell in Massachusetts; Manchester, New Hampshire; and Rutland, Vermont. As the scale of the industry grew and chemical production from wood products emerged as a major industry, demand for hemlock meant that many forests across the northeastern United States were devastated. In many cases, the value of the tannin was so much greater than that of the lumber that the bark was simply peeled from felled trees and hauled away while the woody carcasses were left to rot. Archival photographs of horse-drawn wagons or sleds loaded with immense piles of bark, fifteen feet or greater in height, highlight the value of the bark to the burgeoning hide industry. Hemlock’s role in the leather industry peaked in the late nineteenth century, when synthetic tannins replaced the native product. Now, a hundred years later, the abundant tannins and slow rate of decay of hemlock bark make it valuable once again as a preferred source of landscape mulch. In the economic ranking of tree species, hemlock falls near the bottom, yet it was cut heavily. Thanks to its staying power, hemlock has recovered much of its former abundance. It bides its time, waiting for its opportunity to shine. Through time its influence grows until it reaches the point where it virtually controls all that lives within its reach. Few species in our woods can make such a claim.

AN ICONIC SPECIES 24

Lessons from Harvard Forests and Ecologists I. The Pisgah Forest

O

ne irony of ancient forests and wilderness areas—known for their absence of human imprint—is that the names of many of these great natural areas often conjure up associations with people. Generally, these are individuals who protected the landscape or documented its qualities in ways that enable us to appreciate them today. In these associations there is much to be learned, for although we value nature by itself as a source of inspiration and insight, many lessons and much enjoyment also come from the stories of the people whose lives are entwined with it. Nationally, there are many examples of the enduring connections between great natural areas and particular individuals. In the Sierras, for example, the Mariposa Grove of giant redwoods and the Hetch Hetchy Valley are forever associated with John Muir, who mused profoundly and passionately on both and helped to preserve the former while losing a grand but bitter fight to protect the latter. Gates of the Arctic, one of America’s largest wilderness areas, is inextricably linked with Bob Marshall, the zealous wildlands advocate who mapped this spectacular section of the Alaskan Brooks Range before he founded the Wilderness Society and was later memorialized in the Montana wilderness that is known to its users simply as “the Bob.” In New England many grand forests are inseparable from the people who explored, described, or protected them: Walden Pond and the Maine Woods are tightly associated with Henry Thoreau; Acadia and Marsh-Billings-Rockefeller National Parks with the Rockefeller family; Connecticut’s Colebrook Forest with Yale professor George Nichols and the studies and lament that he penned for the state’s last old-growth forest just before it fell to the ax; and the White and Green Mountain National Forests with Congressman John Weeks, whose 1911 bill allowed the designation of federal forests in the eastern United States. These places inspire in ways that are stronger and more enduring because of the human history, endeavor, and emotion that they evoke.

25

In Harvard Forest history, there are four great forests that were shaped, conserved, and interpreted by men and women whose actions and written words have linked their names with these woods in enduring ways. In the four “Lessons from Harvard Forests and Ecologists” distributed throughout this book, the insights and inspirations associated with the Pisgah Forest, the Bob Marshall Plot, the Earl Stephens Plot, and John Sanderson’s Woodlot are explored. Each of these forests engages and inspires visitors in different ways. Each individual and woodland also offers classic studies in ecology, environmental history, and conservation. I have never seen any more impressive and authentic old growth pine than stands on this tract, and I hope that a good sample may be somehow preserved. —Richard T. Fisher to Harlan Kelsey of the Appalachian Mountain Club, 1921

Pisgah. This magnificent forest has a social history as tumultuous as its natural one, which is saying quite a lot, given that the entire forest was flattened by the 1938 hurricane. But in the course of the last century, this grandest old-growth forest in New England has been saved from the logger’s ax twice, reconstructed in minute detail by three different graduate students, and rejected as a study site by the world’s leading wilderness advocate. As the relentless press of deforestation crescendoed in the middle to late nineteenth century, and New England became a domesticated agrarian landscape, a few artists, romantics, naturalists, and scientists began searching the region for an antidote to this civilized world—for nature untouched by human hands. With almost all of the southern woods despoiled, Henry Thoreau found his wildest adventures in the deep woods of northern Maine. Alternatively, many painters were drawn to the White and Green Mountains, the Catskills, or Adirondacks, where they were able to capture scenes of raw nature on canvas before the waves of logging and fire could sweep through those areas. But wilderness zealots also spread out through the already settled lands, searching for pockets of ancient woods in less inhabited corners and locales where they expected that rough terrain, thin soils, and the contrast with fertile intervales may have left a few stands untouched. There they found the occasional grove spared from saw and ax. The Berkshires proved to be the most fruitful and famous destination for such quests, and its rugged and wild qualities attracted figures such as Herman Melville, Nathaniel Hawthorne, and Daniel Chester French for inspiration. A second, lesser-known region that developed a reputation for magnificent trees lay along the Massachusetts and New Hampshire border. Here rugged hills rise east above Brattleboro, Vermont, and stretch beyond the Connecticut River toward

LESSONS: THE PISGAH FOREST 26

New England and adjacent New York state showing locations mentioned in the text. (Brian Hall)

Mount Monadnock through border towns including Hinsdale, Winchester, and Winchendon. In his never-ending search for local wildness, Henry Thoreau pursued the rumors of big trees and ancient groves in this region through minor explorations that were notable if casual. Four times he climbed Monadnock for its rugged splendor, and on at least one occasion he diverted his return to Concord through the well-rivered borderlands to gaze toward ancient forests and beyond the mill towns that were expanding to destroy them. Richard Fisher, the founding director of the Harvard Forest, grew up in the small town of Berkshire, Massachusetts, familiar with these tales of immense trees. An English major in the late 1890s at Harvard College and a passionate consumer of Thoreau’s writing, he spent summers on his uncle Abbott Thayer’s small hill farm in the town of Dublin, New Hampshire, just below the looming mountain, Monadnock. Thayer was an inspired and nationally renowned painter, self-taught naturalist, captivating storyteller, and spiritual eccentric. The three subjects that dominated his art captured these qualities effectively: the grand isolated mountain in its varied moods and perspectives; the natural camouflage of birds and other wildlife in their native habitats; and, rather disconcertingly and a bit garishly, angels in flowing robes with radiant visages and immense outstretched wings. Beyond his art, Thayer was a political and conservation activist and avuncular inspiration. He ultimately energized the effort to conserve and protect his favorite mountain from further ravages, and he shared his many talents genetically and through endless tromps and evening gatherings with his nephew. It was from Thayer that Fisher derived his keen interest in nature, and he turned this growing passion into a living after graduation by joining the national natural history survey headed by the great zoologist William Hart Merriam. Through continental excursions with the survey and meetings back in Washington, D.C., with Merriam and his federal forestry counterpart Gifford Pinchot, Fisher ended up heading to Yale for a forest education. It was also from Thayer that Fisher first heard of the ancient groves of forest in the Ashuelot, New Hampshire, region, west and south of Keene and centered on the towns of Richmond, Winchester, and Chesterfield. The date of his first explorations in that area is unclear; these may have begun with summer excursions in the late 1890s with his undergraduate chums the James boys, sons and nephews of William and Henry, who occasionally frequented the Thayer household with Richard. But within years of the establishment of the Harvard Forest in 1907, notes and photographs emerge from trips to magnificent stands of towering pines and hemlock. Though some of the greatest trees lay close to the river in Richmond or downstream near the village of Ashuelot, Fisher increasingly focused his attention (and that of his students and visitors) just north, toward Pisgah Mountain. In that LESSONS: THE PISGAH FOREST 28

The distinctive rocky summit of Mount Monadnock in southern New Hampshire. (David Foster)

rough massif lay dozens of untouched and unnamed old-growth groves in an extensive and uninterrupted forest landscape. The stands were diverse, ranging from immense beech, oak, maple, and birch on the slopes above North Round Pond to isolated patches of pitch pine on the most extreme rock ridges and outcrops. But the attention of the Harvard forester seemed rather single-mindedly riveted on the awe-inspiring areas of white pine and hemlock, as Fisher related years later in a letter to his buddy Harry James: The stand in question is perhaps the best sample remaining of absolutely primeval forest such as once covered the whole tract and still covers various areas aggregating perhaps 600 or 700 acres. Personally, I do not know LESSONS: THE PISGAH FOREST 29

of any other absolutely authentic original forest, except for small areas, left in central New England. Furthermore, the remoteness and wildness of the region makes this particular example uncommonly valuable and interesting, not to say secure. The trees are for the most part white pine and hemlock, the largest of which range from 200 to 300 years old. Some of them are approximately 140 feet tall and three feet in diameter at breast height. If there is to be any surviving specimen of the original forest preserved for the enjoyment and study of the present and future generations, I do not know where it could be secured except here. (November 2, 1925) Fisher’s interest was not primarily aesthetic, although his letters reveal that he reveled in these dark forests, while photographs of him sitting with his dog and constant companion, Johnny, at the base of towering conifers confirm his feeling of complete comfort in these woods. Rather, as a gifted naturalist turned forester, he used these stands to advance a remarkably prescient direction for his profession. Through his travels and work from the redwood country of northern California to the hills of New England, Fisher had become convinced that forest management must be guided by a deep understanding of nature and the natural and cultural processes that shape it through time. It was this appreciation for history, storytelling, and art that inspired the creation of the Harvard Forest dioramas and their captivating depiction of the changes in New England over the past four centuries. It was also Fisher’s emphasis on learning from nature that was captured in the largest of the dioramas and the centerpiece of the museum. This scene of an old-growth forest depicted Fisher and Johnny alongside the noted geologist and Harvard dean Nathaniel Shaler, sharing the moment and their thoughts on this natural and cultural landscape. The motivation for Fisher’s studies and the foresight with which he grounded his approach to management in an understanding of natural patterns and process is well described by Al Cline in his own publication on Pisgah: Studies of the virgin forest remnants in the Pisgah mountain section of southern New Hampshire began some twenty-five years ago when, under the guidance of Professor R. T. Fisher, first director of the Harvard Forest, students and staff members made yearly trips to see the veteran trees and to try to locate remnants of old stands not previously discovered. . . . At first these field excursions served chiefly to gratify the curiosity of those who had never before seen central New England virgin timber and to obtain rough notes and measurements on the character of the stands and the size of the trees—always with the accompaniment of picture taking. As time went on,

LESSONS: THE PISGAH FOREST 30

Richard Fisher and Johnny amid old hemlocks on the Slab City tract of the Harvard Forest (1925). (Harvard Forest Archives)

however, it became more and more evident that the soundest basis for the development of the art of silviculture lay in an understanding of the “natural” forests or the regional complex of trees, shrubs and herbaceous plants which would occupy a given piece of ground when free from interference by man. Thus, interest in the Pisgah old growth grew, and with it a feeling of urgency to make a thorough study lest the old stands be cut or accidentally destroyed. . . . This basic philosophy [espoused by Fisher] of working in harmony with nature has time and again proved its soundness and has become firmly fixed as the foundation of silvicultural policy at the Harvard Forest. (Cline and Spurr 1942) Pisgah furnished Fisher with a grand laboratory for the investigation of nature’s history and dynamics in the raw. It was a place where he could learn about the growth and death of trees over the centuries and in the absence of human activity. Once solidly in hand, these lessons could then be applied to an understanding and management of the working laboratory of the Harvard Forest. From there the resulting insights could yield publications, demonstrations, and graduates that would help guide the management of private and public forests across New England and beyond. Fisher and his students commenced their studies in the old-growth areas simply by exploring this natural laboratory and describing and tallying its assets of diverse stands and landscapes. They were struck by many features: the immense height, mass, and apparent ancient age of the trees; the variation in tree sizes and species both in individual stands and across the landscape; and the abundant evidence for many natural disturbances—wind, ice, insects, pathogens, and even small fires. Most striking were the features that were missing from the typical forests that cover New England. There were windthrow mounds ten feet high, immense dead snags that reached up so far that they were lost in the canopy, and the fallen carcasses of trees up to four feet in girth that seemed to molder on the ground for decades. These initial reconnaissance trips fostered an understanding of the New England landscape that stood in sharp contrast to the simple, deterministic view of nature that would dominate ecology for much of the twentieth century. Pisgah’s influence can be seen in Fisher’s writings, as well as in the design of the presettlement diorama in the Fisher Museum. In that scene we see a highly varied forest structure, great evidence of natural dynamics, and a few awestruck and beleaguered land surveyors, dwarfed by the towering trees and woods that engulfed them. If, however, we consider the great bulk of the original forest in central New England—a region of transitional climatic factors, and, as regards species, a

LESSONS: THE PISGAH FOREST 32

tension belt between the coniferous types of the north and the hardwoods of the central states—we find a relative instability in composition. The original upland forest, assuming again the exclusion of extremely dry or wet sites, contained many species of hardwood together with hemlock and white pine, the hemlock a more or less constant element and the pine variable in numbers and distribution. Although this forest over huge areas was apparently stable in percentage composition of species, there is evidence for believing that on smaller areas, and over periods of several centuries, there was a tendency to fluctuation in dominance between softwoods and hardwoods. The natural adaptabilities of species to site were intermittently upset by lightning, fire, windthrow, ice storms, and sometimes insects or disease. Thus a declining group of pine and hemlock would be replaced by one of hardwoods, and vice versa. This tendency is indicated both by the frequent groupwise distribution of the softwood and hardwood elements in the stand and by the relative arrangement of the size or age classes. (Richard Fisher, introduction to Evolution of Soils, 1930) With a solid grasp of the structural patterns in this natural landscape, the Harvard group sought a critical but missing element. They needed a temporal framework for interpreting the relationships among all these features. Exactly how ancient were these trees, and what initiated their establishment? How frequently did these different disturbances occur, and what was their role in determining the age, type, and placement of individual trees? How much did the various disturbances vary across the landscape? And do forests develop slowly and progressively, or episodically and in fits and spurts? Though the researchers could hazard guesses, based on information such as the sizes of trees or their degree of decay when fallen, they were rightfully insecure about these speculations. What they needed was to really dive into and disassemble the forest. The opportunity for just such a forensic breakthrough came with some calamitous news that simultaneously triggered the first step toward conservation activism in Harvard Forest history. In 1919 the sons of Ansel Dickinson, longtime principal of the New England Box Company, announced plans to log the old pine and hemlock on their 5,000-acre holding on Pisgah Mountain. New England Box held tracts across the region that it logged heavily, including the Adams Fay lot in Petersham. But the family owned Pisgah privately and had mined small pockets of the old growth judiciously. Fisher had worked closely with the company on occasion, and yet there is no evidence that he attempted to reverse their thinking about Pisgah or ever contemplated that a broad swath of that landscape might be conserved. But the Dickinsons’ decision did motivate him to launch a comprehensive study

LESSONS: THE PISGAH FOREST 33

of all the old stands and to count and measure the tree rings on every stump produced by their initial harvests, in order to reconstruct the history and dynamics of the old-growth forests. Their decision to log the ancient stands also prompted him to parallel this research effort with a fund-raising campaign to purchase and protect the most magnificent of the old stands. Fisher’s focus lay on a fifty-acre tract that wrapped over one of the higher ridges in the region and an intervening narrow valley. Here stood the forest he described in the letter to his old chum Harry James: pines and hemlock three to four feet in diameter and more than 140 feet tall. To save this stand he sought funds from Harvard alumni such as James and forged a campaign with Philip Ayres, president of the Society for the Protection of the New Hampshire Forests. The effort ultimately led to Harvard University’s peculiar ownership of a small parcel in the midst of what has become the largest state-owned property in New Hampshire—the 13,000-acre Pisgah State Park. By the time that Richard Fisher died tragically of a heart attack in 1932 at the age of fifty-four, his interpretation of the Pisgah landscape had been reinforced with solid data. Three of his students—W. C. Branch, Robert Daley, and Thomas Lotti— spent two years ferreting out and studying all the ancient stands and following the loggers’ fresh trails through those areas that had been cut. They studied sixty-eight old-growth patches that either had escaped logging or were located and sampled just before the axes and saws arrived. In twenty-eight stands that were harvested, the ages and growth history of the immense trees were recorded by laying a strip of paper across each fresh stump and then marking the tree pith and every ten years of growth out to the bark and outermost ring. The students examined each site for evidence of past disturbance and then placed their focused study in a broader context by laying out a few lengthy transects across the entire landscape and sampling at points along them to evaluate the variation in sites, forests, and history. Though their story would not become widely known until Al Cline and Steve Spurr mined it for their classic paper in the early 1940s, the three students’ joint thesis became a treasure in the Harvard Forest Archive and lore to all who worked and studied there. Cline and Spurr captured the major message from their work: “The primeval forests, then, did not consist of stagnant stands of immense trees stretching with little change in composition over vast areas. Large trees were common, it is true, and limited areas did support climax stands, but the majority of stands undoubtedly were in a state of flux resulting from the dynamic action of wind, fire and other forces of nature. The various successional stages thus brought about, coupled with the effects of elevation, aspect and other factors of site, made the virgin forest highly variable in composition, density and form.”

LESSONS: THE PISGAH FOREST 34

Graduate student W. C. Branch in the Pisgah old-growth forest, Winchester, New Hampshire (1930). (Harvard Forest Archives)

In the years since Fisher’s death, his successors occasionally failed to fully grasp his insights or to apply his vision. One striking example occurred in 1938, when the Great New England Hurricane swept the region, uprooting nearly three billion feet of timber from Long Island Sound to northern Vermont. Rather than recognizing this as a natural process to be studied and worked with, the man who attempted to fill Fisher’s shoes as director single-mindedly declared the event a disaster and promoted a massive federally orchestrated timber harvest and cleanup. Ward Shepard’s success in advancing this view played a major role in the formation of the New England Timber Salvage Administration, which organized the single largest forest salvage operation in U.S. history. Fortunately, it also quickly drew Shepard to Washington, D.C., to help administer the program and the ensuing timber effort in World War II, leaving the Harvard Forest in the decidedly more capable hands of Fisher’s old student and acolyte Al Cline. Shepard’s departure was fortuitous, because in his desire to focus federal attention and resources on New England forestry, he had perversely advanced the argument for the region-wide timber harvest by appealing to concerns over human safety. His most strident appeal painted a fear-mongering vision of firestorms sweeping the damaged land and roasting villages, homes, and citizens. Rather than reflecting on the insights from Pisgah and Fisher—namely that hurricanes had shaped New England for millennia and that the forests would regrow rapidly—Shepard cast the 1938 storm as a national environmental calamity and emergency. As a consequence, the salvage operation garnered support from federal, state, and local sources, crews began arriving from across the country and Canada, and ultimately a remarkable 50 percent of the downed timber was harvested. But the environmental consequences of the post-hurricane logging frenzy were severe. The New England landscape was left cut over, scraped over, and smoldering from slash fires, which occasionally escaped into intact forests. The zeal to sanitize the land and thwart the flame even threatened the virgin forest that Fisher had fought so hard to save. The most critical challenge arrived as pressure mounted to liquidate and clean up the voluminous material laying on the ground at Pisgah. Fortunately, with Shepard diverted by emerging wartime resource procurement efforts in Washington, Fisher’s insights were heeded at this critical moment. Acting director Al Cline was no visionary leader, but he revered Fisher and grasped the broad implications of his research and subscribed to his philosophy of management. Cline faced three powerful arguments from federal, state, and local voices for a cleanup operation on the Harvard Pisgah tract. Those arguments? Harvard had a civic duty to protect the region from devastating fires; a moral duty as a leading forestry institution to demLESSONS: THE PISGAH FOREST 36

View east across Harvard Pond after the 1938 Great New England Hurricane, with logs cut from uprooted trees stacked for storage in the water. (Harvard Forest Archives)

onstrate appropriate management; and an economic responsibility to contribute its valuable old-growth timber to local mills and markets. Cline’s response surprised the advocates of salvage. Despite his well-known role in overseeing other salvage operations on the three Harvard tracts in Petersham and as a member of the town and state timber salvage commissions, he stood resolutely behind the writings of Fisher and his three students when it came to Pisgah. In the face of intense political and social pressure, he refused to allow any cutting on Harvard’s small virgin tract. The civic and moral duty he embraced was to preserve the forest for future generations. As Cline recalled in his 1942 paper with Steve Spurr, he had argued that the magnificent trees should remain as nature had thrust them: “Even in their prostrate position, these veteran trees are an impressive sight. In the small openings among the fallen trunks and branches, the new generation is already becoming established. In accordance with the purpose of acquisition of the Harvard tract, no salvage of the down timber was permitted. It remains a place where one may study nature undisturbed by man.” When Cline made his decision, he was deep in the process of compiling the many Pisgah studies for his publication with Spurr. That had led to recognition that the old-growth forest had itself developed following some kind of “catastrophic” event back in the 1600s that was analogous to the 1938 hurricane. Based on that insight, Cline reasoned that the forest in its uprooted condition provided an invaluable natural laboratory in which recovery and the slow development of the next old-growth forest could be studied in great detail. Cline’s bold decision proved prescient. Though no longer filled with immense trees, this virgin forest came to attract even more research following the hurricane than it had received in the three decades under Fisher. It emerged from the apparent catastrophe as a healthy forest and unique laboratory for a series of studies. Cline’s synthesis was the first for Pisgah. His decades-long struggle to publish the thesis of Branch, Daley, and Lotti ended with a thick Harvard Forest Bulletin that included the new perspectives revealed by the impacts of the 1938 hurricane. His effort was significantly aided by the youthful energy and ecological background of Yale graduate student Steve Spurr. Following that initial joint effort, Spurr then turned many of these lessons into a truly insightful article with Cline in the Journal of Forestry, whose title coined a phrase that was reinvented by the forest profession nearly half a century later: “Ecological Forestry in Central New England.” Strongly reminiscent of Fisher’s arguments, the conclusions of Spurr and Cline were extremely innovative: “It is generally recognized that, in theory at least, silvicultural treatments should follow nature as far as possible. As our knowledge of the forests has increased we have come to realize that the more we correlate our forest practices with the natural factors operative in the forest the less expensive and hazLESSONS: THE PISGAH FOREST 38

ardous forestry becomes. Only by studying and following the ecological relationships of the forests of the region can we progress in this direction and finally achieve profitable forest management.” Spurr’s early success led him to devote his Yale Ph.D. thesis to collective studies of Harvard’s many forests and to capitalize on the history of natural and culturally induced changes in its varied landscapes to draw lessons for ecology and forestry. He was the only scientist at the Forest to set up a few plots and a lengthy transect at Pisgah in the late 1940s to evaluate the forest’s recovery. The results formed a nice chapter in his thesis and a separate 1956 paper in the leading journal Ecology. Collectively, this body of work initiated a steep trajectory of brilliant research and career development that would eventually thrust Spurr into national leadership in forestry and academia—as professor of silviculture at the University of Michigan and eventually president of the University of Texas—greatly surpassing that of his Harvard Forest mentors. His observations at Pisgah complemented the photographs of the place and underscored its distinctive qualities relative to the rest of the New England landscape: “Conditions following blowdown on this virgin timber tract differed very greatly from those on second-growth areas. The tangled blowdown of heavy timber (largely hemlock and white pine) was 4 to 10 feet thick over most of the area. Under this criss-cross of logs, the soil was moist and relatively undisturbed except around the uprooted stumps. The boles of the fallen trees covered over 30 per cent of the area, and inasmuch as most of these were suspended several feet from the ground, they affected practically all of the surface.” Spurr also drew a major conclusion that has been verified many times over the decades in ecologists’ studies of nature’s response to disturbance. He correctly recognized that natural disturbances often advance the development of mature forests rather than generating the scarred and rather weedy early successional conditions produced by salvage harvesting: “In this case where no logging followed the blowdown of old growth pine and hemlock, the following stand represents essentially a later successional stage than the preceding one, being characterized by hemlock and beech. In other words, the blowdown accelerated rather than set back successional development by felling the overstory and releasing the tolerant understory.” In the early 1960s, an unlikely pair teamed up at Pisgah to produce another classic paper that was also published in Ecology. David Henry, a young master’s student with a thin background in plant ecology, would go on to a distinguished career in Canadian wildlife biology. His advisor, Mark Swan, a brilliant British ecologist and Harvard assistant professor known for groundbreaking research in quantitative analysis, would quixotically depart academia shortly thereafter for Houston and a new career that began at the calculator company Texas Instruments. The Henry and Swan paper received great notice in this internationally visible journal for its three LESSONS: THE PISGAH FOREST 39

The Pisgah tract four years after the 1938 hurricane, showing the rapid regrowth of the old-growth forest. (Harvard Forest Archives)

major attributes: it convincingly advanced the argument introduced by Fisher that ecological interpretation was strengthened by historical understanding; it clearly laid out the Harvard Forest approach to forest reconstruction; and it presented a dynamic understanding of New England forests at a time when the field of ecology was still mired in rather static thinking. Henry and Swan’s paper reinforced the reputation of the Harvard Forest as a leader in forest history, and it continues to be cited as a key source of information on natural disturbance regimes in northeastern forests. When I arrived in Petersham in 1983, I was surprised and disappointed to learn that only a single permanent plot existed at Pisgah despite three-quarters of a century of research (Spurr’s plot data remained in the archives, but the field markers had been lost through the years). That plot was the solitary twenty- by twentymeter (sixty-six- by sixty-six-foot) area established by David Henry in 1967. After nearly two decades of neglect, it was barely defined by a few wire stakes. With the help of Peter Schoonmaker, who conducted his Ph.D. research at Pisgah, I added twelve similar-sized plots across the Harvard tract and reanalyzed all the historical data. This effort led to Schoonmaker’s thesis and his development of an approach to ecological studies that built on Henry’s work but also harkened back to Branch, Daley, and Lotti. Schoonmaker found all the former old-growth stands identified by that group in the 1920s and studied how they had fared during the hurricane and the past fifty years. He then turned to the Harvard tract and spent nearly two years establishing two permanent transects of plots across the breadth of the twentyfive acres and dissecting the forests to examine landscape-level variation in site and forest conditions. To complement the spatial perspective gained from these efforts, Schoonmaker applied pollen analyses and other paleoecological approaches to explore the deep history of the area. The results of this more recent work would have fascinated Professor Fisher, but it wouldn’t have surprised him. For while it presented many new wrinkles in the story that he was beginning to unravel when he died, these studies leave intact his understanding of Pisgah and the broader New England landscape. Fisher would certainly have anticipated the rapid way that the forest regrew after what appeared to many people to have been catastrophic destruction. Now, eight decades later, the new woods have thrived and emerged as a maturing forest landscape. What has surprised everyone and would have impressed even the first director, however, is that, while hemlock has recovered superbly and emerged as the dominant tree, there is but a single white pine on the entire tract. Pine, which once reigned here as the largest and most visible species, has disappeared. Natural processes alone caused this remarkable change, confirming how kaleidoscopic our forests can be. Over the course of time, the general appearance and broad composition of New England’s regional forest remain relatively steady, but the specific details regarding LESSONS: THE PISGAH FOREST 41

The Pisgah Forest in 2012, showing the remarkable persistence of oldgrowth trees blown down by the 1938 hurricane. (David Foster)

the size of trees and types and distributions of individual forests may change considerably. A dramatic shift in composition also has major consequences for the ecosystem. White pine grows faster and larger than most other species, and its great longevity allows it to reach superlative heights and girth. This meant that Fisher’s old-growth stands at Pisgah had among the greatest amounts of standing timber and organic matter of any forest in New England. In contrast, the new stand—with its smaller and slower-growing hemlock and beech and much shorter-lived birches and red maple—will never support such massive trees or staggering amounts of wood. The rapidly regrowing forest will become increasingly impressive and may one day reach old-growth qualities, but it will never be the one that Fisher saved. Another aspect of Pisgah that never fails to amaze even those of us who visit it frequently is the persistence of the immense downed trees, root masses, and associated pits and mounds of earth. Despite, and in large measure because of, the ravages visited to the site by the hurricane, the forest continues to support features and scenes once common in the pre-European landscape but rarely seen in New England today. A trip in 2012 to core and identify these downed trees revealed an even greater surprise to us all: every one of the immense remaining prostrate logs is white pine. The branches and trunks of all the other species that once constituted this ancient forest have decomposed, leaving only pine, whose old-growth wood is extremely resistant to the action of insects and microbes. These downed trees will persist well through the centennial of the hurricane as a legacy of the ancient stand and the storm. Although living white pine plays no role in this new forest dominated by hemlock, pine will remain an important but subtle part of the forest for decades to come.

LESSONS: THE PISGAH FOREST 43

THREE

P R E H I S TO RY TO P R E S E N T

Nail down the lid; caulk the seams; pay over the same with pitch; batten them down tight, and hang it with the snap-spring over the ship’s stern. Were ever such things done before with a coffin? Some superstitious old carpenters, now, would be tied up in rigging, ere they would do the job. But I’m made of knotty Aroostook hemlock; I don’t budge. Cruppered with a coffin! . . . Come hammer, caulkingiron, pitch-pot, and marling-spike! Let’s to it. —Herman Melville, Moby-Dick

S

ome of us never seem to lose the joy of playing in the mud and on the water. Picture a remote New England pond on a beautiful, breezy summer day. The sky is a deep blue and filled with those puffy but solid clouds that may get larger and angrier later in the day. An unlikely research vessel is anchored in the deep center of the pond and tethered in three directions, floating beyond the reach of the mosquitoes buzzing through the red maples and sedges along the shoreline. This makeshift catamaran, constructed by strapping a sheet of plywood across two canoes, holds three sun-baked scientists. The activity unfolding on the catamaran could be accomplished through the ice in the winter, and much of our work does happen that way. But there is nothing like spending a warm, blissful day on and occasionally in the water and then exploring the woods together to develop the camaraderie that makes for deeply collaborative research. Aboard the vessel, the researchers chat eagerly as they strain to pull a thick metal rod upward through a five-inch-wide hole cut in the center of the plywood. After every other heave, they stop so one of the team members can unscrew a sixfoot-long section from the top. After they haul up and set aside ten lengths of rod, a three-inch-wide aluminum tube, similar in length to the rods and attached to the bottom one, is brought carefully up through the hole and placed horizontally on

44

the raft’s wide wooden deck. Within the tube lies a core of sediment that has lain undisturbed at the bottom of the lake for many thousands of years. As the researchers position themselves for their final exertion of extruding the mud from the tube, their chatter diminishes and their focus sharpens. They yearn to see what their efforts have yielded. With a sheet of plastic wrap unrolled beneath the tube, two of them grab onto the metal cylinder, which has been cooled by its journey into the depths of the water and mud, slowly pulling it toward the third member of the crew, who sits at the end of one of the canoes, bracing the end of the coring rod against an upright two-by-eight board. As the tube slides toward the board, the mud emerges gradually from the other end, a glistening cylinder of sediment exposed on the plastic. The section has the color and consistency of chocolate pudding with an olive tinge, an unremarkable appearance for New England lake sediments. The scientists have been watching varying shades of similar sediment emerge from the tube for more than a half-dozen drives, and they can now relax with the knowledge that the day has been a success. But when this current cylinder of mud is exposed about halfway out of the tube, its texture and color change abruptly. The dark, browngreen organic mud switches to a four-inch-long layer of coarse, light brown sand, and then transitions back to the old brown mud for the remainder of the core. The sandy layer grabs everyone’s attention and prompts immediate discussion and conjecture concerning its origin (“Erosion from a hurricane?” “More likely a drought!”) and some speculation on its age. “It’s likely the mid-Holocene warm period and drought, about 5,000 years.” “No way, we’re deep into this record, clearly older, more like 12,000 years old.” A knife is skimmed across the surface the length of the core to remove the outermost film of mud and water that was smeared along the inside of the tube. With this closer look, they see a set of thin sandy layers that appear as light bands in the darker mud just before it transitions to the thick layer of grit. A grunt of satisfaction emerges from the advocate of the mid-Holocene drought, although the conviction of his exclamation is tempered by the low and questioning “Hmmm” that follows from the rest of the group. Then, recognizing that little more can be gleaned from such a cursory examination of a rapidly drying section of mud while sitting in the middle of a lake, the group wraps the core gently in the plastic, folds over the ends, and encloses the entire core in a layer of aluminum foil. They carefully label it with tape and a black permanent marker before laying it to rest next to other cylinders of sediment in a shallow plywood box referred to as “the coffin.” Over a lunch of sandwiches, gorp, and some delicious fresh-baked cookies contributed by a reliable veteran of the group,

PREHISTORY TO PRESENT 45

Our University of Wyoming collaborators, led by Bryan Shuman (right), coring Deep Pond in Falmouth Massachusetts. (David Foster)

the discussion of the sand layer picks back up. The researchers recount and compare stories of similar scenes and jot down a few additional details in a waterproof notebook. Images of vanished ecosystems and ancient environments run through their minds as they settle back and look out across the surrounding landscape, imagining the shifts in forests, environments, and people through time. Extracting a core from a lake, bog, or soil in this fashion is the first step in reconstructing past environments and ecosystems using the sediments accumulated over many thousands of years. If this account makes it seem like a straightforward process, in practice simply getting the cores is actually a major challenge, especially in lakes. Consider that the coring team must retrieve a continuous sequence of perhaps thirty or more feet of oozy mud from beneath an equivalent depth of water, all while positioned on a windy lake surface and blind to the whole procedure as it unfolds below. In fact, it is only since the 1960s that good, relatively inexpensive equipment has been available to carry out this kind of lake-based work. Because it is difficult to sample in water, these types of paleoecological studies began in wetlands and first on sites in Scandinavia, Britain, and elsewhere in northern Europe, where the deep layers in bogs were exposed by a long history of cutting vast areas of peat for fuel. The next step, that of coring bogs, was comparatively easy, because the material lay right at one’s feet and was compact and rather firmly held into place. But coring lake muds did not become manageable until an ingenious device, the piston corer, solved the vexing problem of collecting intact sequences of mud. Invented first by Dan Livingstone from Duke University (who earned additional fame by surviving an attack by an enraged crocodile in his early coring days in Africa), it was later refined by Herb Wright from the Limnological Research Center at the University of Minnesota, who was largely responsible for bringing the Scandinavian approach of paleoecology to North America. The coring device is a simple stainless steel or plastic tube, affixed to accumulating lengths of detachable rods as the depth increases. A rubber piston fits tightly into the tube and seals it. As we lower the corer down through the water column, we position the piston at the bottom of the tube to keep water and mud from entering until it is at the correct depth, which we measure carefully along the way. A steel cable is attached to the piston, and we secure it up in the raft when the corer reaches the desired depth. The cable keeps the piston in place while we push the tube farther down into the mud. As the tube descends, it collects a long cylindrical sample of intact mud and all the associated material that has accumulated over time on the lake’s bottom. Once the corer is driven its entire length and is full (typically three feet), the whole assembly—the rods, cable, and corer—is pulled to the surface. On the way up through the water the piston maintains suction in the tube to keep the mud from falling out, while the stainless steel tube itself holds the mud intact and PREHISTORY TO PRESENT 47

protects it from the water. On the next drive, the rigid metal poles enable the corer to be lowered precisely to the desired depth, often in the same exact hole, so that we can take a sequence of continuous segments until the bottom is reached. A typical ten-acre New England pond might contain ten to fifty feet of sediment that has accumulated in the central depths over more than ten thousand years. Wrapped in plastic and aluminum foil and secured in the coffin, the cores are brought to the laboratory and subjected to a wide range of treatments and analyses carefully designed to coax out a sequential story of how environments and ecosystems have changed. With many cores of mud to analyze, a joyous day spent coring on a raft with good friends under blue skies will result in many months and likely years of paleoecological laboratory work. Trying to understand present-day ecosystems and anticipate their future trajectories without a long-term perspective is akin to trying to make sense of a movie plot by examining just a few frames from a large reel. History understood through paleoecology places the present in context, because we can understand the factors that shaped the land over time and gave us the present, and we can evaluate the rates at which many things around us are changing today. But deep history provides much more than context—it allows us to examine important biological and geological processes that we could never study in real time, even through many lifetimes. We can observe phenomena that occur infrequently on a human timescale, such as major shifts in climate, intense natural disturbances, or the development and consequences of new modes of human activity. We can observe how the landscape and species that are well known today operated when stressed by factors and magnitudes of change that we have never experienced and may well have no record of. These lengthy records of the past embrace a breadth of conditions that are absent on earth today, ranging from unusual combinations of temperature and precipitation and the presence of immense continental ice sheets to the activities of people and species that disappeared long ago. Using the records that emerge from these archives of mud, we can address questions that are pertinent to contemporary environmental challenges: How do species adapt to and move over continental distances under a rapidly changing climate? How do forests respond when a new tree species arrives in a region and takes on a major role in the landscape, as has happened many times over past millennia? What are the cascading changes in species and ecosystems that occur when one dominant species suffers a major decline due to a disturbance that is specific to it alone? Hemlock has changed in abundance numerous times in the past, and it now faces an extreme threat from the hemlock woolly adelgid. As we seek to consider this new dynamic in perspective, we are fortunate that hemlock has left a remarkPREHISTORY TO PRESENT 48

able array of records that shed light on its ecology under a wide range of conditions. These historical and paleoecological archives inform the field studies, experiments, and modeling activity that we undertake in the woods and back in the laboratory. A look at hemlock’s fossil record helps us examine how hemlock has changed with the intense human activity in the past few centuries and allows us to assess how it might cope with the combination of insect onslaught, climate change, and ongoing human activity today and in the future. It also enables us to evaluate whether there is any hope that hemlock may stave off or recover from the population collapse associated with a new invasive organism. We use a variety of tools and techniques to reconstruct the historical dynamics of the forest environment and vegetation, as well as individual tree species. To reach back furthest, we study pollen, other microscopic fossils, and diverse signatures of past environments that are preserved for millennia in the sediments of lakes, bogs, swamps, and other wetlands. More recent centuries and decades come alive in historical land-survey documents, field studies of old-growth forests, and tree rings that yield insights into the composition and structure of forest vegetation from the time of European arrival forward. In some cases, the particular qualities of hemlock provide a record that bridges prehistory and history. For example, by carefully dissecting the deep beds of needles that accumulate on the cool, moist ground beneath hemlock, we find pollen and other plant parts that yield a chronological record connecting the postglacial period with the time since European settlement. From these distinctive soil layers comes a record of changes in the composition of individual forest stands that can be linked to the evidence from tree rings, uprooted trees, and the many other clues that are present in the hemlock forest itself. Those of us conducting retrospective studies at the Harvard Forest have employed this full array of approaches, exploiting every opportunity to reconstruct the distribution, abundance, and dynamics of hemlock across New England and going back thousands of years into the past. One hundred and fifty years ago, Henry Thoreau mused in his journal on what stories might be gleaned from the pollen grains accumulating in small pools and ponds, but it took nearly a half century more for the Swedish naturalist and geologist Lennart von Post to first take advantage of this phenomenon in studying the history of plants over long periods. He published a report in 1917 showing that the grains of pollen identified in Scandinavian peats told an astonishing story of dynamic changes in vegetation composition. Two characteristics of pollen make it a particularly useful tool for interpreting the past. First, pollen grains are remarkably durable because they are shielded by an outer layer of complex chemical compounds that protect the sperm cells as they PREHISTORY TO PRESENT 49

A 220-plus-year-old hemlock adjoining Hemlock Hollow. (David Foster)

get transferred from the stamens to the pistils of flowering plants, or from male to female cones in conifers like pine or hemlock. Second, the pollen of different species and genera of plants is different enough to allow us to identify them. It comes in a wide range of shapes, sizes, and surface markings, all of which allow palynologists—the meticulous and patient scientists who toil over microscopes, examining these minute fossils—to separate and identify the pollen or spores of particular plants. Some pollen can only be distinguished at the level of the plant family (such as roses, buttercups, or peas) or at the genus level (as is the case for oak, which has many different species but unfortunately only a single type of pollen). In other cases, finer distinctions can be made, such as with the pines, the maples, the hickories, and the spruces, where many but not all of the species can be separated. But in the case of two of our most important species—hemlock and beech—we are fortunate that they can be identified individually. Indeed, the pollen of each of these species is rather distinctive. Hemlock pollen grains look like rough spheres with a fringe along their equators. By contrast, each beech grain has three deep furrows with circular pores in the middle. Palynologists puzzle over these and many other distinctions through their microscopes, with the assistance of reference materials, photographic keys, and colleagues. Over time—many years to a lifetime—the many different types of pollen have become readily distinguishable. The different pollination strategies of individual species influence how reliably we’ll find a particular tree’s pollen in the cores we extract. Some species produce small amounts of pollen in an attractive flower to enlist the assistance of insects, birds, and even small mammals to transfer the tiny grains from the flower of one plant to that of another of the same species. The efficiency of this process and the characteristics of these pollen grains, which are often comparatively large, heavy, and sticky, ensure that very few errant grains end up in some sediment. That means that for many plants that use bright and showy flowers to attract the attention of pollinators, there is but a scant record in the mud. Among New England trees, the pollen of chestnut and maple, for example, is largely distributed by insects, so even though these species were or are often abundant, they are underrepresented in the pollen record. If, however, a plant relies on the wind to distribute its pollen grains— and most of our abundant trees such as oaks, birches, beech, and all of our evergreen species use this strategy—it’s a different story. These species produce prolific amounts of pollen each year, sending clouds of pollen aloft so that some lucky few might happen upon a female flower. The vast majority of these pollen grains miss their mark and end up in the sediments of lakes, wetlands, and forest soils. A large lake collects pollen not only from the adjacent vegetation, but also from plants in the landscape as far as ten to a hundred miles away. In contrast, pollen accumulating in vernal pools, small ponds, PREHISTORY TO PRESENT 51

bogs, or soils is much more likely to be derived from nearby plants, including those hanging immediately above it. This means that records from those types of small basins reflect the local vegetation. Paleoecologists need to take these factors into account in their interpretation of records. They can also apply this knowledge to choose sites that sample the vegetation at either local or regional scales. Regardless of the site, changes in the pollen grains found in successive layers of sediment indicate whether the composition of the vegetation has changed through time. The key to obtaining a good and continuous record is locating an environment with slow decomposition, in which such layers can accumulate gradually and remain undisturbed. We find such conditions in lakes, where fine-grained mineral and organic matter settle out as mud in the deepest areas and then are preserved in the cold and oxygen-poor environment. An alternative environment is wetlands, where waterlogged conditions inhibit decomposition, and the vegetation grows on a surface composed of the remains of previous generations of plants. In New England, where glaciers scoured the earth surface during the last ice age, the duration of both of these sedimentary archives is limited to the period since the ice melted, the land surface stabilized, and the climate allowed the growth of plants. Thus, the oldest lake records span about twelve to fifteen thousand years, and many wetlands only extend back five or six thousand years. Meanwhile, back in the lab, we slice the cores into thin sections, half an inch or less in length, and carry out a series of treatments and analyses of the material. It’s not just pollen grains that we seek. For instance, we want to know the age of the mud at different depths in the core, so we extract small samples of sediment or plant material and send them to a specialized (and expensive) laboratory that assesses the radiocarbon content of the material. We also measure the sediment’s organic and mineral content or particle sizes to determine changes in the lake environment, including past droughts, which are often registered as layers of sandy, inorganic material. In combination with other chemical analyses, these sedimentary characteristics provide a detailed record of past variations in climate. We isolate pollen grains as well as the spores from ferns and other early plants by subjecting mud samples to intense acid baths, washings, centrifuge spins, and sieving steps. It’s remarkable that these intense treatments remove most of the organic and mineral material but leave a tiny residual fraction that contains the concentrated and quite intact pollen, along with bits of insects, charcoal, and other miscellaneous detritus. The tiny pieces of charcoal and insect remains, both of which are as highly resistant to decay as pollen, are sieved, identified, and counted under a microscope to provide information about past wildfire activity and insect outbreaks. We mount the residue on microscope slides and examine them with highPREHISTORY TO PRESENT 52

The Harvard Forest paleo crew cores sediments from a pontoon raft in four inches of water on Hemlock Hollow. (David Foster)

powered magnification, carefully scrutinizing and identifying every pollen grain that is encountered. At any given level, a palynologist might identify 300 to 500 pollen grains through a painstaking process that can take anywhere from two to eight hours or more. Pollen data tell us the relative abundance of different species. If 50 out of 500 pollen grains at a given level are identified as hemlock, this would yield a value of 10 percent. Knowing whether or not a species is a prolific pollen producer helps us to assess how well the relative abundance of its pollen corresponds to its actual abundance on the landscape. The pollen of insect-pollinated trees such as maple and chestnut rarely exceeds 5 percent of the total, whereas pine, birch, and oak can easily reach 10 to 20 percent or more. Considering these factors, we would assume that 5 percent chestnut means a significant presence. At its very crudest, a pollen diagram will show at what point in the past hemlock or any other plant was absent, rare, or abundant. In most cases, it will also reveal fascinating curves depicting the long-term variation in these species in relationship to other species and many environmental factors. In well-studied regions such as eastern North America, many dozens of pollen records have been analyzed over the last few decades. In southern and central New England, the Harvard Forest group has analyzed cores from more than three dozen sites. We make the data available to everyone electronically on our website and collaborate with many people who use them. We also keep the cores from which samples have been taken in cold storage for our future needs and those of other scientists who may be interested in examining our records in more detail or for searching for other materials and clues in the mud. Our network of study sites enables us to understand how the environment and ecosystems have changed in certain places, and how geographic patterns of climate and vegetation have shifted through time. They also help us reconstruct the migration history of various trees, including hemlock, as they returned following the last glacier. At the height of the last glacial period, approximately 20,000 years ago, a milethick ice sheet covered the New England landscape, with its southern limit extending just to or slightly beyond the modern-day coastline. Pushing and carrying material southward like a combination of a bulldozer and conveyor belt, the immense glacier piled up linear landforms called moraines that today form the higher parts of Cape Cod, Martha’s Vineyard, Nantucket, other coastal islands, and Long Island. We use the term “sea level” as if it were a constant, but with vast quantities of water stored on land in these continental ice sheets, the sea level then had dropped more than 300 feet. New England and other coastal regions extended thirty-five miles or more outward on the exposed continental shelf. Pollen records show that, during PREHISTORY TO PRESENT 54

this peak of ice and cold global temperatures, hemlock thrived far south—in the valleys and hilly landscapes in the Southern Appalachians, where oaks, hickories, and tulip poplars thrive today. As the climate warmed and the ice melted back to the north, hemlock migrated northward, arriving in New England around 10,000 years ago. To get to the Northeast from the Southeast, populations of hemlock had to travel nearly 900 miles in approximately 5,000 years, a migration rate more rapid than we might expect based on our modern studies of the dispersal distances of the species in our forests today. Most estimates of migration are based on standard observations of the dispersal of a parent tree’s seeds and the establishment of new seedlings, which are then extrapolated over time. The small, winged seeds that drop from hemlock cones generally fall within 100 feet of the parent tree, and as a result hemlock moves more slowly across the landscape than most species. For example, in many New England forests today, hemlock has yet to travel the short distances required to return to stands from which it was extirpated two or three centuries ago. In contrast, the seeds of birches and pines may be dispersed 200 feet within a stand and more than 700 feet across an open landscape, enabling them to be highly successful at colonizing abandoned agricultural fields. Given these factors, we would expect hemlock to be among the slowest of species to have migrated north after the ice age. Indeed, the characteristic slow movement of hemlock initially led to predictions that, during its northward march, it would have lagged well behind the availability of suitable environmental conditions that developed as the climate warmed. Rather surprisingly, however, all current evidence suggests that hemlock and the other major tree species migrated fairly rapidly, effectively keeping up with the climatic conditions that were able to support them. Consequently, the order in which the species arrived in New England fits nicely with our general understanding of their individual environmental requirements, as well as their modern distribution. Open, treeless tundra occupied the harshest climates in the early postglacial landscape. As the climate became more hospitable, the tundra was invaded by northern boreal species—spruce, larch, and birches. With further warming, white pine followed, and then came the truly temperate tree species, including hemlock. Far to the north, the tundra continued to follow the receding glacier toward the pole, and, where they could, boreal forest trees then seeded into the tundra. Paleoecologists have struggled to reconcile the observed and expected rates of migration and have even given a name to this incongruity: Reid’s paradox. The issue has emerged as one of great importance today because of the looming likelihood of rapid climate change and the question of how plants will respond and cope with new conditions. We are employing all sorts of approaches—genetics, simulation PREHISTORY TO PRESENT 55

Young hemlock seedlings on a decaying trunk of white pine in the Pisgah Forest. (David Foster)

modeling, field and laboratory studies of dispersal, and pollen analysis—as we continue to grapple with the question. Have we overestimated the rates at which trees moved in the past, or are we underestimating their anticipated and potential future dispersal rates? One possible way to account for a more rapid past dispersal is to invoke a history of rare long-distance dispersal events, such as abrupt gusts and updrafts in wind that may loft a seed into the jet stream, or the rare flight of a bird in which it carries a seed for dozens of miles. In this way, a chance event can disperse seeds great distances. If such an event happened even once a decade, it may have been extremely important in shaping patterns of movement over centuries. We cite uncommon processes such as these in our modeling discussions when talking about the dispersal of insects like the hemlock woolly adelgid or the adaptations of plant species under future climates. As research on this dilemma progresses, the answers to these questions will have important implications for predicting the future shape of our forest ecosystems and for gauging the ability of many species to survive the expected changes in climate in coming decades. The long-term history of hemlock also reveals the extreme malleability of forest types and assemblages, including those that are familiar to us today. Hemlock arrived in the northeastern United States about 2,000 years after white pine and 2,000 years before American beech, even though today it frequently grows alongside both these species, and we often think of them as members of the same plant communities. Given beech’s similarity to hemlock in shade tolerance and suitability for forest canopies, and the manner with which they coexist in many places today, it is hard to imagine that hemlock grew in New England for 2,000 years without beech. Similarly, it was only with the arrival of hemlock that the New England landscape developed forests akin to the old-growth stands of white pine and hemlock studied by early ecologists and described in many Harvard Forest studies, including those by Richard Fisher, Bob Marshall, Tony D’Amato, and Dave Orwig. The contrasting histories of these various trees illustrate that species respond in highly individualistic ways to environmental change. Because conditions in the past were distinctly different from the present, we witness the species behaving in significantly different ways over time. The assemblages of plants and animals that are familiar to us today are actually quite ephemeral in deep time and space. It is through such understandings that we’ve developed an ecological theory that accepts and explains the separate though interactive behavior of species. One of the earliest and best articulations of this theory came from a noted northeastern botanist—Henry Gleason of the New York Botanical Garden—who developed the “individualistic concept of ecology” in the early 1900s. This simple but revolutionary theory posited that the makeup of vegetation on a site was determined by

PREHISTORY TO PRESENT 57

the actions of the many individual species, each of which operated quite separately from others and according to its unique ecological qualities. Although this concept was debated for decades, some of the strongest evidence that led to its conclusive support came from paleoecological studies that showed the highly disparate behaviors of different tree species in migration and in response to climate change and to natural and human disturbances. While this understanding of plant behavior and ecology emerged from the past and helps us explain our current landscapes, it should also prepare us for unanticipated combinations of species to appear under the anomalous conditions expected for the future. Coring dozens of ponds and bogs and examining tens of thousands of pollen grains preserved in their sediments has helped us outline the following picture of New England’s prehistory. After a lengthy dry period, from around 11,500 to 10,000 years ago, during which white pine dominated the landscapes of the northeastern United States, hemlock increased in abundance across much of New England, then reached its peak population levels during a relatively warm and moist interval from 8,000 to 5,500 years ago. Beech had arrived to join hemlock in the region at that point, and with oaks, birches, and maples also present, and white pine and pitch pine already well established, the overall composition of New England forests was quite similar to what we find in our landscape today. Although the environmental conditions of that earlier time appear to have been well suited for hemlock, some of our recent research suggests that brief periods of cold climate occurred every few centuries, with deleterious impacts on hemlock in some parts of New England. Various lines of evidence, including chemical analyses of lake sediment records, show that the generally warm, moist conditions were interrupted occasionally by a century or so of cold, dry climate. And while hemlock and other species did not always respond uniformly to these events across the region, some of our relatively detailed pollen records feature abrupt, short-lived declines of hemlock, including significant population reductions at around 8,000 and 6,000 years ago. Hemlock certainly didn’t disappear from the landscape during these events, but the pollen data do suggest that it became much less abundant during times of cold, dry conditions. Then, around 5,500 years ago, hemlock experienced an abrupt, range-wide collapse. For about two millennia it nearly disappeared throughout its entire range in the Northeast before it rebounded about 3,500 years ago. Although it recovered greatly across the region, at most sites hemlock never returned to its predecline levels. This hemlock decline is one of the most thoroughly studied aspects of the postglacial vegetation history of North America, yet we still don’t completely understand what caused it or sustained it. Conclusions drawn over the past three decades variously attribute hemlock’s decline to a species-specific disease, a massive insect outbreak, a sustained shift to drier climate, a series of drought events, and a combiPREHISTORY TO PRESENT 58

nation of these factors. It is now quite clear that climate was strongly involved and that in some ways the big decline was a larger version of the earlier declines witnessed during cold spells. If the trees weren’t killed directly by drought, then the associated environmental conditions either stressed hemlock in ways that made it more susceptible to insects or disease or facilitated an unusual outbreak of a pest or pathogen. (It was this record of minor events leading to the major drought and decline in hemlock that our colleague correctly surmised he was seeing in the various layers of sand we observed that day on the raft in the middle of the lake.) Hemlock eventually recovered, and pollen records reveal that it was again abundant in New England forests from around 3,500 years ago to the time of European settlement. Our studies of the sediments of Hemlock Hollow, a vernal pool hidden in the large hemlock forest on the Prospect Hill tract of the Harvard Forest, have yielded a detailed stand-scale record of forest changes over the last 10,000 years. The local nature of this record enables us to examine the fine-scaled ecological response of an individual forest to various changes in its environment. Here we can see that when disturbances occurred, including fires every 1,000 to 3,000 years, hemlock abundance dropped abruptly and then rebounded slowly, taking 500 years or more to recover to original levels. In the recovery from these major disturbances—intense events that we interpret to have killed most of the larger trees—the successional sequences brought back the species that we know so well and comply exactly with our understanding of the modern ecology of New England forests. For much of the pre-European period when hemlock declined, it was replaced around Hemlock Hollow by some combination of early successional and rapidly reproducing and growing species—white pine, birches, and other hardwoods—as well as more midsuccessional, long-lived species such as oaks. Everything changed when chestnut arrived. After spending the ice age in the southeastern United States, chestnut slowly migrated north and finally arrived in New England 2,000 years ago. At Hemlock Hollow we see chestnut employing its phenomenal ability to sprout and its rapid growth rate to become the dominant species when the populations of hemlock and other species were reduced by disturbance. This pattern occurred following fire and also after European settlement and the first episodes of logging in these forests. These disturbances affected both species, but chestnut bounced back quickly. Dead chestnut boles are a common sight in many hemlock forests today; it is clear from the fossil record at places such as Hemlock Hollow that the two species had a close and often reciprocal relationship in the more distant past. One other notable observation emerges from the longterm record at Hemlock Hollow: regardless of the nature of the disturbance or the successional species that followed it, in each case, hemlock recovered from the disturbance and eventually returned to dominance. PREHISTORY TO PRESENT 59

Dead chestnut trees, killed after 1915 by the chestnut blight, entwined in hemlocks dying from the hemlock woolly adelgid. (David Foster)

These records offer other instructive insights into the broader nature of the New England landscape and its forests. The low abundance of charcoal in lake sediments confirms that there was little fire. Meanwhile, the long duration of hemlock dominance confirms that the region was only infrequently affected by fire or any other major disturbances: drought, wind, and ice. Similarly, there is no direct evidence of disturbance to or use of these forests by the dispersed populations of largely hunting and gathering American Indians who inhabited central New England. Thus, while we may assume quite correctly that change is a prominent factor in forest ecosystems, the paleoecological perspective demonstrates that New England hemlock forests experienced lengthy periods of relative stability. We also have a detailed map of North American forests just before they were first cut and then cleared. For this we can thank a largely anonymous group of seventeenth- and eighteenth-century land surveyors. While walking the landscape and demarcating it into towns, sections, and ownerships, colonial surveyors recorded the presence of individual trees by their species and sometimes by their size. Ecologists have been using these accidental forest inventories to reconstruct presettlement forest composition for almost a century. By far the most common source for survey records has been the Public Lands Survey of the General Land Office, which was established by Thomas Jefferson and covered much of the midwestern and western states. But because southern New England was largely settled prior to the establishment of the General Land Office in 1785, its survey records are much less standardized. Survey-based reconstructions of New England forests typically rely on some type of town proprietor records. The English colonies deeded unsettled land in the form of regularly shaped towns, often about six miles square. In laying out the boundaries in these towns, surveyors identified and blazed “witness trees” as permanent markers at the corners of individual lots ranging in size from 1 to 160 acres. Longtime Harvard Forest collaborator Charlie Cogbill has spent decades amassing a comprehensive spatial database of these tree records from across the Northeast. The maps derived from his witness-tree data set have been analyzed by Jonathan Thompson to show how forest composition varied across the region. In northern Maine, spruce, balsam fir, and white cedar dominated the landscape. Moving slightly southward into the rest of Maine, New Hampshire, and Vermont, hemlock, beech, maples, and red spruce were common, even reaching down along the broad uplands of the Berkshires in western Massachusetts and Connecticut. Oaks, pines, hickories, and American chestnut picked up from there and were prevalent in the south and along the coast. In broad detail, this pattern closely parallels the regional environmental gradient, with cooler and moister conditions PREHISTORY TO PRESENT 61

to the north and warmer and drier conditions to the south. Hemlock became less common farther south and was found in increasingly smaller concentrations. Near the coast it would only have occurred in isolated stands in protected moist areas. Pollen records provide context for the witness-tree snapshot of New England vegetation patterns, including some perspective on the dynamics that were under way when the European settlers arrived. For example, we can see that American chestnut was the last tree species to reach New England from its glacial refuge in the Southeast, arriving here only in the last 1,000 to 2,000 years. Meanwhile, hemlock and beech appear to have already begun a slow decline a couple of centuries before colonial deforestation commenced. The timing of these declines seems to coincide with the Little Ice Age, a relatively recent climatic interval (A.D. 1550–1850) that triggered physical and ecological changes in many regions of the world, including glacial advances farther north. It may seem counterintuitive that two species common in northern New England would be bothered by a shift to colder climate. It is quite possible, however, that conditions became both colder and drier, with hemlock and beech suffering due to their relatively high moisture requirements. The latter part of the Little Ice Age coincided with the expansion of European colonists across New England, transforming the land. Region-wide, up to 60 percent of the land was cleared for agriculture and the rest was cut—repeatedly in some places—with a peak in harvesting occurring in the late nineteenth and early twentieth century. Although forest once again covers more than 80 percent of New England, these second-growth stands are not the same as those of presettlement times. When we compare the witness-tree data with present-day forest composition, we find that some species are more common than they were centuries ago, such as early successional birches, red maple, and pines, including the old-field white pines that invaded abandoned agricultural lands. These light-seeded, fast-growing, and light-requiring species spread and grew rapidly across heavily disturbed areas, thriving after the intense farming and logging subsided. On the other hand, some species are less abundant than they were before European settlement. Species of mature forests, including hemlock and beech, are much less common than they were in the witness-tree surveys. Throughout the Northeast, hemlock declined as much as 10 percent over the last 400 years. When we zoom back in from the region-wide scale to that of the individual landscape, we often see considerable evidence of land use in the characteristics of hemlock forests. In some cases, seemingly ancient hemlock stands have undergone much greater changes in their recent past than we might at first assume. These are the unexpected findings of a study led by Harvard Forest researchers Jason McLachlan and David Foster. They set out to reconstruct the histories of four old hemlock forests in central Massachusetts, using both tree-ring analysis of the largest trees and PREHISTORY TO PRESENT 62

centimeter-by-centimeter analyses of pollen grains preserved in the approximately six-inch-thick layer of organic matter forming the top layer of the soil. They found that the stands, dominated today by hemlocks 100 to 200 years old, had experienced a series of disturbances over the last few centuries, including logging, windstorms, fires, and pathogen outbreaks. Indeed, early and mid-successional trees such as oaks, pines, and American chestnut had occupied those same stands at different times in the past. In many of the forests, it appeared as though today’s dominant hemlocks may in fact owe their current good fortune to the removal of competing species by selective logging and the chestnut blight. Like many of our other retrospective investigations of hemlock, this study of second-growth stands obliges us to change the way we think about the species, the forests it forms, and the way that nature operates. On one hand, forests that appear to be unchanging may be relatively recent in origin and shaped by processes that the species has never experienced before. On the other, although hemlock forests have been dynamic at times, the history of the species in New England has always been one of long-term dominance interrupted by infrequent abrupt declines. With such a decline spreading across the landscape today, we can expect another lengthy period with little hemlock followed by—we can only hope—its gradual return.

PREHISTORY TO PRESENT 63

FOUR

T R E E - FA L L S A N D TA N B A R K

I cross an ancient brush fence and am fairly within the old hemlocks, and in one of the most primitive, undisturbed nooks. In the deep moss I tread as with muffled feet, and the pupils of my eyes dilate in the dim, almost religious light. —John Burroughs, Wake-Robin

T

he dark-crowned groves of hemlock stippling the landscape of southern New England form repositories of a critical forest history, yielding stories of the natural and human-induced destruction and rebirth of these majestic forests. Most tell a familiar second-growth story of hemlock’s devastation by the ax and subsequent resilience, but a select few yield old-growth tomes with lessons of how these forests may have looked and functioned prior to European settlement and how they have weathered the past few centuries of natural dynamics and climate change. The last century witnessed an amazing period of interpretation of both of these story lines and the manner in which these forests have developed in response to natural and human forces. These stories further our understanding of ecological processes and provide lessons for forest conservation and management. Sadly, these studies increasingly serve an additional purpose: they contribute to the baseline of understanding we can use to document and to understand a longstanding feature in our landscape that is beginning to disappear. Thus we turn to these forests for broad insights into natural and human history, for conservation lessons, and for a last deep view into the ecology of thriving hemlock forests. The reach of the lumberman’s ax and the impact of the flames and farm animals that often followed were extensive indeed. Together they left a landscape that by the turn of the twentieth century was a stark departure from the nearly unbroken sea of virgin forests that had greeted settlers just a few centuries before. Although traces of the great primeval forests of southern New England persisted in the accounts of

64

explorers, travelers, and early naturalists, little truly ancient forest remained by the time scientists began studying the region in earnest in the early 1900s. Scouring the landscape on foot and on horseback, they located many large, old trees and a few scattered old-growth tracts, left largely due to inaccessibility, serendipity, or deliberate protection by a succession of landowners. These remnant stands provided a generation of botanists, geologists, and ecologists with a fascinating window into the characteristics of the primeval tracts. Old-growth forests have one unique, defining historical quality: an absence of direct human disturbance such as logging, grazing, or farm clearing. The rarity of this condition in the southern New England landscape was strikingly apparent to the early Harvard Forest scientists at Pisgah. They were similarly recognized by the renowned Yale ecologist George Nichols, who made many forays to the Litchfield Hills of northwestern Connecticut seeking a natural resource that was rare, largely ignored by society, and therefore disappearing before his eyes. The efforts by these two college-based groups are emblematic of the mixed history of success in the conservation of great wilderness regions. Pisgah was ultimately protected by Richard Fisher and the Harvard Forest. In lamentable contrast, the last of the even more diverse ancient forests in North Colebrook, Connecticut, was harvested in 1912. Nichols’s eloquent descriptions and captivating photographs of these stands and individual trees serve as a posthumous homage to untouched nature in New England and to the man who argued in vain for its preservation. Fisher and Nichols left us with insightful portrayals of what were long believed to be the only remaining old-growth hemlock forests in the region. For decades it appeared fruitless to search for additional old-growth stands in southern New England. Frank Egler’s 1940 monograph on the vegetation of the Berkshire region, the most rugged and sparsely settled part of southern New England, stated the case rather conclusively. Egler, a preeminent, Yale-trained, and decidedly idiosyncratic ecologist who relished bucking tradition, nevertheless aligned himself with the views of his peers and most natural historians in New England when he declared that “no original pre-colonial forests remain on the entire Berkshire plateau.” Two decades later, this proclamation was refuted by Robert Livingston, professor of botany at the University of Massachusetts, and his graduate student, Paul Hosier. Deep in the ravines and hidden slopes of that rugged landscape of northwestern Massachusetts, the pair discovered a series of old-growth groves dominated by stunning hemlocks. These initial discoveries were the result of extensive exploration fueled by a passion for unaltered nature, grounded in a solid understanding of New England history, and emerging from grueling months scouring the steep and rocky terrain. The breakthrough by Livingston and Hosier was subsequently exT R E E - FA L L S A N D TA N B A R K 65

Tony D’Amato with a 300-plus-year-old hemlock in the old-growth forest on Mount Everett in western Massachusetts. (Dave Orwig)

tended by Peter Dunwiddie and Robert Leverett in the early 1990s and later broadened and confirmed through extensive field and archival research by ecologists at the Harvard Forest and the University of Massachusetts in the early 2000s. Remarkably, and despite the ravages of land clearance, logging, and fire, some old-growth forests did remain in the Berkshires. Totaling roughly 1,100 acres, spanning more than twenty locations, and consisting of trees exceeding 120 feet in height and three centuries in age, the collection of old forests uncovered by these ecological sleuths provided an unanticipated opportunity to pick up on the efforts of the early twentieth-century ecologists from Yale and Harvard. By comparing the ages, size, growth patterns, and histories of the ancient stands to the region’s much more common second-growth forests, it was possible to understand how natural ecosystems function and differ from those that recovered following the plow and ax. The first insight simply involved understanding how these forests managed to survive the ravages of New England history. Just like the tracts at Pisgah studied by Fisher and those at North Colebrook investigated by George Nichols, the stands of old-growth trees scattered across the Berkshire Hills and the Taconic Mountains of western Massachusetts occupied steep terrain that was often nearly inaccessible and challenging to harvest. The highest concentration of these ancient forests lies where widespread European settlement lagged behind that of the adjoining river valleys and gentler terrain by fifty to a hundred years. Surprisingly, it turned out that western Massachusetts was not the only part of the region that revealed some hidden old growth. Well to the east of the tightly grouped constellation of old-growth hemlock forests in western Massachusetts, an outlier occupies the steep slopes of Mount Wachusett, an isolated but rugged monadnock less than fifty miles from Boston. Richard Fisher, in fact, had recognized and described these gnarled and peculiarly distinct ancient forests in 1906. Having established a side business as a consulting forester to supplement his modest junior faculty salary and to develop contacts in state agencies and the forestry world, Fisher took on a job surveying and assessing lands for the Wachusett Mountain State Reservation Commission. He was scrambling all over Mount Wachusett the year before he founded the Harvard Forest and just months before his growing commitments in Petersham and Cambridge caused his consulting business to wither. When that point arrived, he put his Wachusett work aside and filed the description of his old-growth discovery deep in the archives of the Harvard Forest. Nearly a century later, ecologist Dave Orwig discovered the early notes while perusing old files in Petersham. Intrigued, he set out to find these stands, and he eventually confirmed the presence of the ancient forests. Orwig led a series of extensive studies across the steep slopes of the mountain, which by that time had become established as a state forest, commercial ski area, and focal point for major telecommunication towers T R E E - FA L L S A N D TA N B A R K 67

and regional tourism. In thick woods traversed by major hiking trails and steep downhill skiing trails and offering panoramas toward New England’s largest metropolitan area, he discovered a diverse array of old-growth forests covering more than 200 acres. Based on Orwig’s work, these stands are now recognized as the largest old-growth forests in southern New England. They also include some of the oldest living yellow birch and red oak in the world. It’s no coincidence that hemlock dominates many of the region’s remaining old-growth stands. Because the species has been perennially viewed as a low-value timber species, many areas that are “heavy to hemlock” were simply passed over for richer forests of white pine as the lumbering boom moved through the region in the late nineteenth and early twentieth centuries. Although vast areas of hemlock were felled and stripped of tannin-rich bark, its value did not outweigh the challenges of tackling the remote and rocky slopes that sheltered these ancient stands. After progressive waves of settlement, farming, and logging had swept through the region and past these remarkable woods, the western Massachusetts landscape was left dotted with a collection of surviving stands. With their ancient trees, boulders, and jumble of living and dead material, each is a living museum and archive of the natural dynamics and ecology of the region’s forests. Given the size of the trees and the dominance by hemlock, many of the early scientists who first examined the old-growth forests in New England conceived of them as representing the “climax forest”—a self-perpetuating stable condition that might develop across much of the region in the absence of major human or natural disturbance. Accordingly, the ancient woods on these apparently sheltered sites were interpreted as the product of the many natural processes of forest development—seed production, seedling establishment, growth, and competition— operating solely under the influence of climate. The tranquil scene that greeted visitors to these woods reinforced this notion. The deep shade, sheltered locations, and cathedral-like canopy of great trees conveyed the sense that quietude prevailed and stability was the norm. Gradually, however, this interpretation was undermined. The deeper that researchers explored these forest ecosystems and examined their many features, the clearer it became that the forests are dynamic and shaped by many different natural disturbance processes. Wind, ice, erosion, insects, and gravity helped to generate many of the distinctive structures commonly used to recognize and define old growth: large downed logs, immense uproot mounds, and a broad spectrum of tree sizes. Along with the ecological characteristics of hemlock itself, these disturbances helped to shape the structure and environment of the woods and their patterns of plant and animal abundance. These forests had never been static, had never marched inexorably toward a climax state. Nowhere was the role of disturbance in hemlock forest development more apT R E E - FA L L S A N D TA N B A R K 68

Old-growth forest of hemlock and hardwoods on Pisgah Mountain (ca. 1920). (Harvard Forest Archives)

parent than at Pisgah. In their early visits to this rugged region of southwestern New Hampshire, Harvard Forest scientists documented how common naturally disruptive processes had been—wind, snow and ice damage, lightning strikes, and insect damage—all underscored by the 1938 hurricane. That storm’s 120-plus mileper-hour winds had converted the landscape into a tangle of logs and terminated any lingering belief in the stability of old-growth hemlock forests. The climax paradigm no longer applied. Ecologists and foresters alike became focused on the role of disturbance in shaping the long-term development of forests. It’s not that scientists had ignored disturbance. Before the Great Hurricane, Fisher, Cline, and their students had taken what they’d learned at Pisgah and applied it to the broader New England landscape, stating that disturbances were critical in disrupting the shaded understory conditions created by hemlock and allowing entry by light-demanding species such as white pine and paper birch. A similar array of species and conditions were found among the hemlocks at the old-growth tract in North Colebrook, Connecticut, suggesting a similar legacy of disturbance T R E E - FA L L S A N D TA N B A R K 69

in a place that George Nichols had otherwise characterized as “self-perpetuating and permanent.” Chestnut blight was one disruption sweeping through the Pisgah landscape in these early studies. Chestnut generally formed a minor, albeit important component of old-growth and second-growth hemlock forests before the blight arrived from Asia and spread through the region around 1920, killing all the adult trees. Although the loss of chestnut did not dramatically alter the appearance or dynamics of most hemlock groves, its absence was recorded in the tree-ring records of many hemlocks as a notable increase in their rate of growth. The general pattern described for many old-growth forests of localized small disturbance followed by forest response and development has come to be called gap-phase replacement. Small openings in the forest created by the death of a single individual or small groups of overstory trees are subsequently occupied by seedlings and saplings of shade-tolerant species such as hemlock and beech that have patiently bided their time in the densely shaded understory for decades to over a century. In New England and much of the eastern forest, such small but frequent disturbances remove 5 to 20 percent of the canopy trees in a given decade, regularly adjusting the forest structure. But because many of these seedlings and saplings are of the same species that dominate the existing forest, this long-term pattern of local dynamics promotes landscape-wide equilibrium. Despite consistent evidence for this general pattern of forest growth and development, the tree-ring record and species composition of these sites all indicated that the forests were not at a perfect equilibrium. Studies at Pisgah and in the Berkshires also yielded evidence of larger, moderate-intensity disturbances. A 1921 winter storm in southern New Hampshire, for example, loaded trees thickly with ice that tore off limbs, toppled many groups of trees, and left sizable openings in the dense canopy. In a similar fashion, gusts from hurricanes in 1788, 1821, and 1938 reached the sheltered groves in the Berkshires with patchy but severe damage that left a legacy in our modern forests of trees released from oppressive shade. This kind of event has come to be known as an intermediate-intensity disturbance—greater than single dying trees but much less than the full-bore hurricane that laid the Pisgah forest flat. In many types of ecosystems, moderate disturbance encourages a greater diversity of plant and animal species than either extreme does. As a result, ecologists have proposed the intermediate-disturbance hypothesis, which states that species diversity is maximized when disturbances are neither too frequent or intense nor too rare. Disturbance and dynamics vary, but they are part and parcel of natural systems. The intensity of the forest disruption wrought at Pisgah and across much of southern New England in 1938 unveiled a third pathway and model for the development of hemlock forests that was comprehensively documented through Harvard T R E E - FA L L S A N D TA N B A R K 70

Forest studies. The clearest picture of it emerged initially through detailed studies by David Henry and Mark Swan in the late 1960s and was synthesized two decades later by David Foster and Peter Schoonmaker. Both efforts highlighted that, while small-scale canopy gaps could be the prevailing disturbance for centuries, the long-term development of these forest ecosystems was often punctuated by standreplacing disturbance events. These episodic disturbances may be extremely rare, and they are certainly impossible to predict, but they reorganize the forest and add considerable diversity in structure, habitat, and species distributions across a landscape. In their detailed work, Henry and Swan used the evidence provided by buried stems, charcoal, and tree rings to recount the history of Pisgah’s old-growth forest in a manner similar to that of archaeologists reconstructing past civilizations through an intensive dig. The forest artifacts they uncovered told a tale of catastrophe followed by quietude and light to moderate levels of disturbance. Lines of charcoal and decaying boles of ancient uprooted trees led them to reconstruct the story of an intense windstorm and blowdown followed by fire, which created growing space for the seedlings that would ultimately develop into the 300-year-old hemlock and white pine that had enthralled Harvard Forest researchers in the early 1900s. Downed logs that were less decayed had come down in more recent storms large enough to create opportunities for seedlings not only of beech and hemlock but also red maple and birch. The largest disturbance these forests had experienced in the century prior to 1938 was a tornado in 1921 that toppled about one-fourth of the canopy trees and generated gaps—within which many shade-intolerant trees would come to join hemlock and beech following the hurricane. Of course, the complete destruction of the ancient overstory in 1938 provided the enhanced light and exposed mineral soil that favored the widespread establishment of paper and black birch and red maple among the existing stems of understory hemlock and beech. While what Henry and Swan uncovered in their single plot captured one major line of history and dynamics at Pisgah, it does not pertain uniformly to the entire landscape of old-growth stands. A broader look at Pisgah shows that the severity of the Great Hurricane’s impact was tempered by topography. Even the most intense winds don’t uniformly destroy everything in their path; most leave a heterogeneous pattern of damage even within areas that on first glance may appear completely devastated. Given that the hurricane’s strongest winds were from the southeast, forest areas on level, south-, and southeast-facing slopes experienced the greatest damage. Comparatively little damage occurred on northerly and westerly lee slopes or the bottoms of deeper ravines and valleys. Coupled with the variation in soils and moisture that occur across the landscape, this topographic filtering of disturbance and response helped maintain a diverse mosaic of forest conditions. Immense white T R E E - FA L L S A N D TA N B A R K 71

pine and hemlock persist on sheltered slopes and broad ridges with deeper soils, while nearby narrow and rocky ridges that are more exposed feature smaller-statured forests of hemlock, pine, and disturbance-dependent species such as paper birch and even pitch pine. In both cases, these can represent old-growth conditions, even though most of the individual trees in the latter case don’t reach the extreme ages of hemlock. Like the majestic groves in the sheltered environs, these exposed oldgrowth forests reflect the dynamic linkages between disturbance patterns and tree species abundance and form. Old-growth hemlock forests remain a rare condition in the New England landscape. Much more common are second-growth stands that grew back following logging or the abandonment of old agricultural fields. If you substitute axes and plows for gale-force winds, the development pattern of these second-growth systems has some analogy to that observed at Pisgah following the 1938 hurricane. As in the stands at Pisgah, many of the areas harvested for lumber and tanbark contained myriad hemlock seedlings and saplings that were poised to take advantage of the demise of their elders in the canopy. In places that were cut hard, forests quickly regenerated as the residual hemlock were joined by faster-growing red maple, birch, and white pine. Continuously forested areas are referred to as “primary forests,” given that forests prevailed on these sites without interruption by farm fields or other land uses after settlers first felled the virgin forests. While these are not old-growth woods characterized by no direct human impact, the primary forests contrast with “secondary forests,” whose history include an intervening period of agriculture. Both are second-growth. In the long history of New England, primary forests often served for some time as woodlots, whereas the secondary forests once supported cows, sheep, hay, or other crops. When tracing the modern distribution of hemlock forests across the region’s landscapes, it is in these primary areas that hemlock thrives, highlighting the resiliency of the species to past lumbering. Its slow return to secondary sites shows its sensitivity to more intensive land use and disturbance. The resiliency of hemlock in primary forests was exemplified in the tree-ring patterns uncovered by Bob Marshall in his early studies as a graduate student at the Harvard Forest. By measuring the record of annual growth rings in cross-sections and increment borings from the trunks of 784 hemlock trees, Marshall concluded that most trees spent decades (and some spent a full century) in the forest understory before they emerged into the canopy after the harvesting of overstory trees. While many of these forests were viewed as functionally young given their recent history of logging, the trees making up these forests were often quite old. As highlighted by Marshall in his thesis, some of these trees even had a direct lineage on the T R E E - FA L L S A N D TA N B A R K 72

Cross-section of a large hemlock analyzed by Bob Marshall on the Adams Fay lot (1924). The dense rings in the center adjacent the knife show that the tree grew in the shade for 108 years until the overstory was harvested around 1840, when the rings widen abruptly. The wood exhibits cracks along the radii and shake between the rings. (Harvard Forest Archives)

site that extended back to the advent of settlement. Moreover, given the absence of plowing and fire in these forests, many of the areas contained intact soil conditions and understory plant communities rarely observed in forests with a more intense legacy of human disturbance. It was through examining the understory vegetation that George Nichols began to recognize the distinctiveness of North Colebrook’s old hemlock forests, relative to second-growth areas. Although lacking the spectacular spring floral displays and species richness of their hardwood counterparts on richer soils, the Colebrook tract held several species, including hobblebush, Canada yew, and twisted stalk, that were rarely encountered in second-growth stands. Similar trends were documented years later at the Pisgah tract by Gordon Whitney and David Foster using data collected by Fisher, Cline, and their students, and by Tony D’Amato among the scattered old-growth relics in the Berkshire Hills of Massachusetts. Many of these species rely on the rotting log seedbeds common to old growth and don’t flourish in second growth because of their sensitivity to the compromised conditions that often follow logging. The slow pace at which these species disperse to and recolonize an area mimics that of the hemlock that shades them. At the North Colebrook tract, Nichols recorded that the overstory hemlock averaged nearly three feet in diameter, with several of the largest trees approaching five feet in breadth. Trees of such a stature seem a physical impossibility within the context of our second-growth hemlock forests populated with much smaller individuals, but these living monoliths commonly dotted old-growth ecosystems throughout the region prior to settlement. The spectacle of witnessing and photographing the region’s largest and oldest trees provided much of the early attraction to old-growth areas. But the lasting scientific lesson emerging through decades of study in these hemlock forests is recognition of the structural complexity of the stands and the critical role that these conditions play in maintaining forest ecosystems. What’s missing from second-growth forests is exactly what old-growth stands have in abundance, beginning with a broad spectrum of tree girths and heights that produce a multilayered canopy through which only scant sunlight reaches the forest floor. In turn, the immense living trees generate an equally impressive and immense array of downed logs and standing dead trees that persist for many decades to centuries while they slowly decay. The standing and downed wood residing in the undulating landscape of hummocks and hollows, created by centuries of uprooting and tree death, provide a wide array of soil conditions and habitats unseen in the second-growth forests that replace them. In particular, large, decaying hemlock logs often resemble small garden plots, with seedlings of hemlock and yellow birch sown throughout on the moist, decaying wood. Meanwhile, the influence of the giant trees themselves extends well beyond T R E E - FA L L S A N D TA N B A R K 74

The complex structure of an old-growth hemlock forest in Mohawk Trail State Forest, Massachusetts. The black birch in the right foreground established within a large canopy gap. (Tony D’Amato)

the forest to the embedded and adjoining streams, ponds, and wetlands. As the immense trees fall, they add critical complexity to these aquatic environments—pools, ripples, and cascades in streams; islands of drier habitat extending into wetlands and vernal pools; and physical structure along the shores of ponds and lakes. Old-growth ecosystems encompass much more than dry land. Each of these features—complex structure from forest floor to canopy, pitand-mound topography, downed wood—was gradually erased by land clearance, repeated harvesting, and the hooves of the many thousand grazing animals that followed European settlement. As a result, the second-growth forests that surround us today are lacking major elements that are common in old-growth ecosystems; in their simplicity, perhaps they are more comfortable for many humans. Natural disturbance played an integral role in shaping the complex structure of these systems, creating tangles of logs and snags with each downburst, windthrow, and pocket of disease. The biological legacies—the living trees and logs left behind by disturbances—play a critical role in the way forest communities reorganize and T R E E - FA L L S A N D TA N B A R K 75

develop following disturbance. Furthermore, they provide the clues for distinguishing old-growth forest structural conditions from more simplified second-growth hemlock forests. Although the modern forests at Pisgah are quite similar to many second-growth hemlock forests in terms of living tree sizes and species composition, the legacies left by the 1938 hurricane still persist as fallen logs throughout the landscape, resulting in volumes of dead wood comparable to those found in the towering old-growth Douglas fir systems of the Pacific Northwest. This distinctive quality of old-growth forests was reinforced when a number of us returned to Pisgah in 2011. Accompanied by Shawn Fraver, a forest ecologist with the U.S. Forest Service in Minnesota, and armed with a large-diameter increment borer that Shawn supplied, we set out to take dowel-like cores of the wood by drilling into the center of the immense trunks that lay prostrate on the ground. Although we Harvard Forest researchers are used to witnessing the surprisingly intact conditions of logs that were stored but never retrieved from ponds following the salvage logging in 1938, we were shocked by what we recovered from these trees that had lain exposed in the woods for more than seven decades. The logs retrieved from ponds are solid and intact, stained yellow-brown by years in tannin-rich water, but dense and sound enough that we pull them out on occasion and run them through our sawmill to make special displays or gifts for retiring employees who have a deep regard for forest history. But the first core taken from the logs strewn through Pisgah yielded an unfamiliar pattern that we then witnessed repeatedly. Though the bark was long gone and the outer three to four inches of wood were decayed beyond any ability to decipher a consistent record of growth, the scene changed abruptly once we hit the heartwood. The first few inches of wood were brown and mottled, partially intact and with readily distinguished annual rings. But the inner six inches or more were astonishingly bright yellowish white. Though saturated with moisture and weaker than wood from standing living trees, the cores were strong and intact even when fully removed from the borer. Over the years we had marveled that the big trunks remained in place, continued to support our weight, and even bridged depressions in the ground ten to twenty feet across. These cores revealed why. Between the qualities of the heartwood—impregnated with resins that deter microbes—and the low supply of oxygen throughout the waterlogged wood, the inner portions of these old-growth logs resist decay. Reinforcing the early studies by Nichols and Cline, these cores shed light on an additional suite of qualities that help to distinguish old-growth from secondgrowth forests and make them functionally distinctive in our landscape.

T R E E - FA L L S A N D TA N B A R K 76

Lessons from Harvard Forests and Ecologists II. Bob Marshall’s Plot April 21, 1924 Mr. Robert Marshall State College of Forestry Syracuse, N.Y. My dear Marshall: Al Cline gave me your article about the Adirondack Forest Preserve, and I have read it with entire sympathy and agreement. There is no argument about the proposition that to furnish the highest kinds of enjoyment a forest should be left strictly alone. With so little real primeval forest now left, sparing the remnants that still exist in the Adirondacks does not seem too much to ask. Your argument gave me a thought about a project, which I have long had in mind, and which might interest you for your investigative work next autumn at Petersham. Not far from there is at least one considerable tract of virgin forest which is gradually being cut in small areas. How would you like to make a detailed study of the origin and maintenance of the virgin forest, with special reference to what might be called the chronological order in which the several species and elements of the stand came in? This sort of study I think will be very useful as a background for the forest management of the future, and unless we make it soon the opportunity will be gone . . . Very truly yours, Richard T. Fisher When Bob Marshall rejected Richard Fisher’s intriguing offer to study the ancient old-growth white pine and hemlock forest on Pisgah Mountain, the Harvard

77

Forest director must have been dumbfounded. But however surprising the decision was, given Marshall’s already clear passion for wildland forests, the choice was consistent with the young man’s life plan. Ironically, the alternative path that he followed as a graduate student—working on a timber harvesting study—led to a research approach that others would later apply successfully at Pisgah. Marshall’s work also did as much as any study to reinforce Fisher’s belief in the value of forest history to ecology and scientific forestry. Marshall’s answer shocked Fisher, because the professor had every reason to expect that his new graduate student from Syracuse University would be enthralled to spend his days amid the ancient trees in the rugged New Hampshire landscape. Through their prior meetings and correspondence leading up to his acceptance in the graduate program at Harvard, Marshall had shared with Fisher many thoughts on forestry, conservation, and the value of forest reserves. In his essay on the Adirondack Reserve for a course in silviculture, Marshall had stated that the “finest formal parks, the most magnificent artificially grown and cleaned woods, cannot compare with the grandeur of the primeval woodland. In these days of over civilization it is not mere sentimentalism which makes the virgin forest such a genuine delight.” The paper had earned the eager and innovative student an “A,” along with a measure of scorn from some in the more timber production–oriented sector of Syracuse’s Forestry School. Moreover, Fisher knew a bit about Marshall’s extraordinary background as the son of one of the leading civil rights attorneys in America, a man who led the largest Jewish community in the nation and fought to protect the poor, the immigrant, and the defenseless in venues ranging from the local courthouse to the U.S. Supreme Court. Among the defenseless and voiceless clients championed by Louis Marshall was nature. Through the years visiting his spectacular Adirondack camp, Louis had witnessed the devastation driven by greed and wrought by ax and fire to the lakeshores and mountain slopes of upstate New York. Outraged by these insults to the environment, armed with knowledge derived as a founder, major funder, and board chair of the Syracuse School of Forestry, and equipped with an orator’s skill and expansive legal mind, he had championed the defense of one of the most important conservation documents in America: Article 14 of the New York Constitution. This legal document included the “forever wild” clause that made the Adirondack State Park the first designated wilderness in the country and kept its land free from logging. Fisher certainly knew that Bob Marshall had wild forest running in his veins. Fisher also had keen reasons for disappointment. Bob had shown the potential for greatness, graduating first in forestry at Syracuse and supported by rave recommendations from his professors and dean. Fisher was keen to initiate genuine scientific research at Pisgah, and this new candidate promised to bring considerable LESSONS: BOB MARSHALL’S PLOT 78

Bob Marshall (center) with Harvard Forest researchers at the Adams Fay lot experiment (1924). Faculty include Rupe Gast (second from left), Al Cline (with machete) and Richard Fisher (far right). (Harvard Forest Archives)

woods skills to Petersham, along with proven writing and computational abilities. Since the successful campaign to protect the ancient forest, Fisher had conspired with John Phillips, a national leader in forest and wildlife conservation, to establish an endowment to pay the annual taxes on the tract and support a scholarship for research into the history, dynamics, and ecology of the old-growth forest. Though the fund was growing, its income remained inadequate to cover the desired field studies. But Marshall had means; he was the sole member of the incoming class at the Harvard Forest who took no salary and required no scholarship. He could freely attach to any project and might jump-start the old-growth effort. And he had one last trait that augured success in the rugged, remote New Hampshire mountains. Bob had a passion for hiking and a knack for navigation. With brother George and guide Herb Clark, he was well on his way to becoming the first of the “46ers”— individuals who had reached the summit of each of the forty-six Adirondack peaks exceeding 4,000 feet. Marshall’s fanaticism for distance hiking was already well LESSONS: BOB MARSHALL’S PLOT 79

Bob Marshall (right) and fellow students with their field vehicle at the Adams Fay lot (1924). (Harvard Forest Archives)

established, and his life list of twenty-, thirty-, and forty-mile treks was lengthy and growing. Bob would not need any hand-holding in the confusing topography around Pisgah Mountain. The match between the man and the Harvard Forest’s newest project seemed perfect. Yet, unbeknownst to Fisher, by age nineteen Bob Marshall had already developed a remarkably well-defined life plan that did not allow for wilderness studies at Harvard. His scheme had emerged in general terms on trips to the family’s Adirondack camp and had been first articulated in the starkly simple prose of a high school writing assignment. Over the years the plan was honed through endless letters and discussions with his father and became elaborated into a well-defined course of action. As he’d written in that high school essay, Bob wanted to become a forester more than anything else. As passionate as he was about wilderness, Marshall had precocious insight; he recognized that society had a more pressing immediate need, one made clear from the vistas he gained atop nearly every Adirondack mountain peak. The United States needed to learn to manage its forests better and more sustainably in order to generate a renewable supply of a precious commodity and to protect the associated resources of water, wildlife, and human well-being. Marshall also understood keenly that society would only be willing and able to carve large wilderness areas out for protection if the bulk of the country’s forestland was productive and well managed. Finally, he was his father’s son, deeply earnest in his pursuits, faithful to an ethic of work, responsibility, and caring for the neglected, and already emerging as an insightful, strategic, and politically minded thinker. He recognized that his greatest hope of working the system for good and of advancing the cause of wilderness and all that was abused by society would come from a powerful platform within that system. He wasn’t about to veer off headstrong into a wilderness campaign. And he wasn’t going to fritter away his hours alone in the woods simply investigating how nature operated. Bob Marshall never delved into casual pursuits. When he hiked at his blistering pace, he always chased a destination, a distance goal, and a personal milestone. He was no sauntering Thoreau who mused over plants, scenery, or natural history. No, Marshall’s decision to attend graduate school was shrewdly calculated to advance his life goals; his thesis project would fit that grand scheme and position him for a job in the world’s preeminent forestry institution—the U.S. Forest Service. It would focus on forestry and advance his expertise in silviculture and systems of tree harvesting, areas in which his background was thin and Fisher’s expertise was already legendary. It would also base him in Petersham, where he could benefit from association with other students, faculty, and the frequent visitors from other universities, state agencies, the Forest Service, and Europe who appeared on the

LESSONS: BOB MARSHALL’S PLOT 81

doorstep of the Harvard Forest headquarters and shared their stories and camaraderie in the small Forest community. There the stream of conversations would advance his knowledge and broaden the web of contacts that he could draw upon for the rest of his life. Through this path, Marshall sought to build on the foundation he had established at home in Manhattan, at Syracuse, and in the Adirondacks to reach a prominent rank in national forestry. Once there he could achieve the multiple goals that inspired his daily and lifelong efforts: promote good forestry and resource use, advance the cause of wilderness, make a difference to society and nature, and live up to his father’s expectations. Louis Marshall’s reputation and the family’s place in society ensured Bob’s access to strategic and powerful connections; adding Harvard to this arsenal was a significant step, and the younger Marshall clearly planned to use his year in Petersham and Cambridge effectively. The four previous years of classes at Syracuse, summers at the university’s Cranberry Lake Ranger Camp, and a top score on the civil service exam guaranteed him a posting with the U.S. Forest Service that might be improved through strategic graduate work. A detour through Harvard would be likely to produce more options and connections and ultimately lead to a more rapid trajectory through the agency. As he wrote to his father in one of their nearly daily exchanges: “It is generally considered that advanced training will lead to greater opportunities and a faster pace through the forest agency. Harvard is the best place in the country to get that training along Silvicultural and Management Lines . . . and Fisher . . . is generally recognized as the foremost silviculturalist [sic] in the United States.” Once at Petersham, Bob believed he needed to capitalize fully on the assets represented by the university, Richard Fisher, and the Harvard Forest. With a solid forestry project, perhaps crowned by a publication, he could get on with his life’s ambition. In recent months he had mulled over these topics endlessly with friends, Syracuse faculty, and even the dean. But, judging from the correspondence that they had maintained since the death of Bob’s mother, it was clear that for both father and son the objective at Harvard was much more than undertaking a single study. Great as Pisgah and the old-growth topic may have been, the project was not the goal. Rather it was one more deliberate building block toward future success. All this is not to say that Marshall’s year at the Harvard Forest was spent in single-minded and somber pursuits, or that the young forester did not develop a love for the place, the people, and his experience. Quite the contrary. As he put it in a letter to his father: “You will note that everyone except the Director is a Syracuse man. Six of us live together in a very large old farmhouse which also contains the office. We have more room than we know what to do with. For once I think there will be plenty of room even for my junk. I have a tremendous writing table 6′ × 3′, LESSONS: BOB MARSHALL’S PLOT 82

a typewriting table, dresser, closet and limitless floor space all to myself. We cook our own meals which are therefore uniformly excellent, far better than the regular restaurant hash.” In Petersham Marshall bunked and enjoyed countless exploits with his closest college chum, Neil Hosley. He forged lifelong friendships with instructors Al Cline and Rupe Gast and with his fellow students. He also emerged as both the prankster of Community House and the grand chronicler of all of their exploits and accomplishments, all the while ratcheting his career forward and maintaining a daily stream of letters home. The experience also left an enduring appreciation for the “sagacious wisdom” and kindness of R. T. Fisher. His extracurricular writings and the impish grin adorning his mug in nearly every photo confirm that Marshall thrived in the close Harvard Forest community. This mixture of companionship and humor was even reflected in his thesis’s acknowledgments: “I am also much obliged to Messrs. Arthur Davis, Fred Goulet, and Otis Goulet for their cooperation while felling timber. It was necessary to study the stumps on several sample plots while the timber was being cut. The three choppers went to considerable trouble to avoid dropping trees on me while I was engaged in this work.” Marshall’s curious personal habits of self-evaluation and documentation further attest to the importance of his year with hemlock and Harvard. Beginning in 1927 when he was twenty-six and continuing until his sudden death in 1939, Marshall systematically reevaluated his place in life each year through a series of lists of “favorites” that he maintained in almost every conceivable category: friends, books, months, places, professional men, girls, authors, and more. His favorite “causes” at his death as enumerated in this life list (as well as in his will) were the Wilderness Society, union labor, the advancement of American Indians, and the Harvard Forest. Indeed, Marshall regularly ranked the Harvard Forest, Professor Fisher, and his summer months of fieldwork alongside students, faculty, and the Woods Crew at the top of his life list. This is all the more remarkable given the breadth of other experiences that this determined man was able to cram into his short life. He grew up in Manhattan, thoroughly explored the Brooks Range in Alaska, lived with northern Indians and Eskimos, and vigorously hiked the wild mountains of the West. He made a point of searching out and meeting Supreme Court justices, Civil War generals, great scientists, and scholars; and he worked alongside the likes of Hart Merriam, Gifford Pinchot, Olaus Murie, and Aldo Leopold. But every January, when Bob reworked his hand-scrawled life lists, updating them with the most recent year’s experiences, his mind faithfully returned to Petersham. There, in a project focused on logging and hemlock, he had bonded with a team of men from many different walks of life and was particularly inspired by a bespectacled gentleman who shared his love for the wild and its lessons for conservation. LESSONS: BOB MARSHALL’S PLOT 83

Marshall’s project was part of the grandest of the long-term experiments that Richard Fisher established, a magnificent precursor to the decades-long manipulations that constitute the Harvard Forest Long Term Ecological Research program today. Following Fisher, we now undertake fifty-year projects, such as pulling down two acres of forest to simulate a hurricane, warming the forest with miles of heating cable in the soils to mimic climate change, spraying nitrogen onto acres of pines and hardwoods to simulate the effects of increasing acid rain, and alternately girdling or harvesting hemlock to contrast the effects of an insect infestation with salvage logging. In the design of his long-term forestry study, Fisher sought to contrast the ecology of hemlock and white pine and to evaluate the effectiveness of different ways of promoting each of these species through logging. Although these two dominant conifers are similar in their longevity and abundance in old-growth forests, they contrast strongly in their growth rates, shade tolerance, and timber value. The study sought to investigate whether it was possible to purposefully manipulate their growth and relative abundance by harvesting the stands in very different ways. By coincidence, the site the Harvard group selected for its experiment in the summer of 1924 was owned by the New England Box Company, whose owners— the Dickinson brothers—were already thick in negotiations with Fisher over the sale of the Pisgah tract. The so-called Adams Fay parcel, named for previous owners, adjoined the Tom Swamp tract of the Harvard Forest and occupies an extraordinary site: a flat outwash plain that was thickly and rather uniformly covered with hemlock and pine. The remarkable homogeneity of the sandy site was ideal for experimentation, because it allowed nearly identical plots to be assigned to different harvesting treatments for comparison with each other and with additional plots that would be left intact and unharvested as controls. The treatments covered the range of common commercial logging practices along with some experimental approaches. This project fully engaged Fisher and senior scientist Rupe Gast, a brilliant though eccentric quantitative ecophysiologist. The tree felling, hauling, and associated work were all undertaken by the Harvard Forest Woods Crew, assisted by the graduate students and supervised by the faculty. The straw boss was newly hired lecturer Al Cline, who had just received his own graduate degree from Syracuse—a convenient decision, given that the incoming graduate students were all from his former department. Marshall dove in with the group, contributing to diverse aspects of the experiment from laying out plots, measuring timber volumes, and marking trees to hauling cordwood and burning brush. But his separate research project also played a key LESSONS: BOB MARSHALL’S PLOT 84

The Adams Fay lot of the Harvard Forest, showing the layout of the large experiment designed by Richard Fisher, with its many types of harvests. The plot that Bob Marshall dissected in detail is blackened in the center. (Brian Hall)

role in framing the larger study. Marshall sought to document hemlock’s growth patterns and its unique abilities to hunker down for decades in the deep shade, eking out an existence and barely growing, and then to capitalize on the death of surrounding trees with a burst of new growth. Though ultimately focused on hemlock, Marshall began by comprehensively dissecting all the trees on his 80-by-200-foot plot to shed light on the history of the entire forest. Fisher and his colleagues used these initial investigations to expand their understanding of the differences between pine and hemlock and to sharpen their hypotheses concerning how the forest would develop following each of the different treatments. Although Fisher was accustomed to applying his natural history skills to interpreting the long-term history of the forests that he studied, Marshall took the art of forest reconstruction and refined and formalized it to the level of science. In his efforts, Marshall was guided closely through regular meetings with Fisher, daily exchanges with Cline in the woods and the dorm where he lived with the boys, and intense strategy discussions with Rupe Gast. Gast devoted extensive time to providing his student with the background in the physics, meteorology, and electronics he needed to evaluate the environment and growth of plants. (Gast later exerted a major influence on Marshall’s decision to attend Johns Hopkins University for his doctoral work and on his thesis research on the growth and physiology of spruce at tree line in alpine Alaska.) Marshall also benefited from daily exchanges with the Harvard Forest Woods Crew of veteran loggers and farmers, who shared great depth of local knowledge. Through the process of working out the history of the Adams Fay area, Marshall developed, refined, and unified all the major steps to forest reconstruction that were subsequently applied by generations of Harvard Forest students. The scientific approach to forest history that emerged that summer is remarkably straightforward: scour the landscape for every scrap of information from living and dead plant material, the soil, human artifacts, and the local topography, and then integrate this with information from more typical historical sources such as interviews, newspapers, census data, correspondence, deeds, and other records. Marshall systematically dissected the forest. He cored trees and sectioned decaying logs to establish ages and growth records; he scoured the ground to locate uproot mounds and moss-covered stumps, and attempted to reconcile these clues on past disturbances with his tree records; and he ferreted through archives, libraries, and notes from interviews with previous owners and loggers to provide context and fill gaps in the story emerging from the woods. Every evening, save the few that they spent at movies in the adjoining mill town of Athol or on Bob’s infrequent trips back to Manhattan, he sat alongside the others at two lengthy tables in the large living room of the Forest headquarters and dorm, compiling notes, computing figures, or jawboning about work, life, and their futures. LESSONS: BOB MARSHALL’S PLOT 86

The Harvard Forest Woods Crew during the experimental harvest on the Adams Fay lot (1924). From left: Harry Upham, Rodney Stevens, and Bert Upham. (Harvard Forest Archives)

The publication of his project in the Harvard Forest Bulletin as “The Growth of Hemlock before and after the Release from Suppression” earned Marshall a footnote in the history of science. But his failure and that of his mentors to document the approach he developed and its value to ecology and conservation was a stunning lapse by someone so focused on fame and career. It would be a half-century before two publications in the journal Ecology (one by David Henry and Mark Swan on Pisgah in 1974, and another by Chad Oliver and Earl Stephens in 1977) exposed the science world to the field and laboratory methods developed by Marshall and employed by Harvard researchers ever since to reconstruct nature’s history. Those papers brought historical ecology into the limelight and earned it a solid place in the discipline, but even these authors appear to have been unaware that the techniques were first forged in the Petersham woods on what we now call the Bob Marshall Plot. LESSONS: BOB MARSHALL’S PLOT 87

Marshall’s efforts resulted in a comprehensive chronology of tree growth and response to wind and repeated harvesting. As revealed in the opening of his 1927 Harvard Forest Bulletin monograph, Bob relished this trip back in forest history and his newfound ability to extend the record back before the area’s colonial settlement: The history here considered commenced 272 years ago, at the time of the inception of the oldest element in the stand of 1924. In 1652, a year before Cromwell became Lord Protector of England, and thirty years before William Penn crossed the Atlantic, a hemlock seed germinated in the dense shade of the virgin forest and a tree commenced its long life of suppression. The history of the stand previous to 1822 can only be conjectured. The forest probably consisted principally of white pine, with considerable hemlock, and a sprinkling of chestnut, beech, yellow birch, and red oak. It was no doubt autochthonous in character. When one element dropped out, either the surrounding trees seeded in the spot or advance growth reproduction replaced the dead tree. But only the most shade-tolerant species could possibly survive with the slight amount of light which penetrated the canopy. Therefore, the understory consisted chiefly of that extremely shade-enduring species, hemlock, which, though it grew on the average about an inch in a century, was nevertheless able to maintain life. It was only when some natural catastrophe made a small opening in the forest that the trees had an opportunity to grow to a large size. No doubt the majority died after years of stunted existence. Frequently in larger openings the less tolerant white pine would seed in and overtake the slower growing hemlock. Then another period of suppression would ensue. One consistent and abundant element through time was hemlock, a fact Marshall attributed to a combination of the tree’s remarkable physiology and the complete absence of fire. He noted that, while white pine was favored whenever big holes were created in the canopy and the soils were disturbed, hemlock prevailed under most other conditions. What he and the Harvard group learned in 1924 took them a long way toward explaining hemlock’s great abundance on the site and its success in the region. From the tree-ring records, he documented that hemlock persisted under heavy shade, displaying minute rates of growth under conditions that the sensors built and deployed by Rupe Gast showed as supporting less than three-tenths of 1 percent of ambient sunlight. Marshall’s data also revealed that hemlock was able to increase its growth rate tenfold or more whenever more light became available. In contrast to most species, which lose or never display this flexibility, hemlock could bounce back repeatedly until it either became a dominant tree or was taken down

LESSONS: BOB MARSHALL’S PLOT 88

by a violent wind gust or a two-man saw. Whereas white pine had a boom-and-bust behavior in which it dominated after major disturbances through prolific seed production, long-distance dispersal, and rapid growth, hemlock employed a strategy of stealth and persistence. It invaded slowly, hunkering down, biding its time, and continually leveraging its position in the woods. From the growing appreciation of hemlock’s ecology and the site’s history emerging from Marshall’s plot, Fisher developed some guiding hypotheses for his big experiment. He proposed that, in the absence of fire, both hemlock and pine would persist on the site indefinitely, as it appears that they had for thousands of years. But he expected that the relative amount of each would depend strongly on the scale of disturbance. White pine would secure a greater foothold when intense windstorms or clear-cuts opened the canopy broadly. Hemlock would establish in the understory and be favored by lengthy periods with few large disturbances. Then, with the death of each pine from lightning, selective harvesting, or senescence, hemlock would increase toward a dominant position. The larger context of this experiment as it pertained to Marshall’s specific focus on the release of hemlock from suppression is laid out nicely in Marshall’s publication: In the autumn of 1924 the Harvard Forest marked for cutting a lot owned by the New England Box Company which contained a stand unusual in northern Massachusetts. It was composed of dense, almost pure, white pine and hemlock with very little ground cover or advance growth hardwood. The composition ranged from pure hemlock to nearly pure pine. But of special interest was the fact that the entire area was thickly sprinkled with old pine stumps which clearly testified that years before a heavy softwood cut had been made on the same area. Now as a general rule the forests which have followed nineteenth century softwood logging operations have resulted both in a conversion to hardwoods and a marked deterioration of type. But here softwood had followed softwood, and furthermore the new stand had both a large volume and excellent form. What was the history which had caused this anomaly? It was in answer to this question that the present study was undertaken. Almost as soon as the first hemlocks had been felled, it was noticed that at the center of every stump there was a core of wood from one to five inches in diameter which frequently had taken one hundred or more years to grow. At the outside of this core there was a very abrupt change in growth rate, and for a period of years rings from one-eighth to one-fourth of an inch thick were found. Coinciding in point of time with this acceleration in growth were old scars, evidently caused by logging. The obvious explana-

LESSONS: BOB MARSHALL’S PLOT 89

tion was that a previous stand had been cut, and the consequent infusion of light had released the long stunted hemlocks from suppression. The hypotheses laid out by Fisher and supported by Marshall’s work were tested directly in the large experiment, which sought to guide harvesting in the real world of commercial forestry. The experiment put these ideas to the test by establishing a gradient of disturbance size and intensity through different patterns of harvesting. The specific harvesting approaches included selection cutting, in which the canopy was thinned of one-quarter to one-half of its stems presumptively to increase hemlock; shelterwood harvests, in which the initial thinning was followed in a few years by removal of the remaining overstory to allow the release of hemlock and the establishment of many pines; strip cuts, which removed alternating sections of forest, producing strong gradients of shade to full sun that favored both species; and sizable clear-cuts, which exposed large areas to direct sunlight and overwhelmingly favored the establishment of white pine. Working alongside the Woods Crew, the students measured the trees in each area before and after every harvest. The entire group then stacked and hauled the wood and carted and burned the branches. At the end of the day, the scientists and woodsmen parted ways. The boys and younger mentors such as Cline retreated back to the old headquarters for dinner, their skull sessions, data analysis, and evening pranks. Gast lived offsite with his family, while Fisher maintained homes in both Petersham and Weston, a wealthy Boston suburb, and so was an episodic visitor. Judging from the evidence in photographs, journals, and letters (accompanied by a noticeable reduction in the frequency of Marshall’s correspondence home), the summer presented a thoroughly exhausting, stimulating, and engaging experience for the close-knit group in Petersham. Through the fall they conducted fieldwork, wrapping up the slash burning and wood hauling in midwinter while the men and horses worked the mill and reduced the logs to large and well-ordered stacks of lumber. As the winter turned into a muddy spring and summer approached, the boys completed their studies. Many stayed on for a second summer or more as they sought jobs, tied up loose ends, or established careers at the Harvard Forest, as Neil Hosley and Al Cline did. Bob Marshall followed a unique path, however. He pushed off immediately following the semester’s end, having completed his work on schedule and successfully converted his original Forest Service offer into a posting in Missoula. If he had glanced back on his way out west he would have realized that, in launching his career, he had contributed to an experimental legacy for future generations, established a fundamental historical approach for ecology, and advanced the knowledge of a key forest species.

LESSONS: BOB MARSHALL’S PLOT 90

The future did not play out for the Harvard Forest or for 1924’s group of students as they may have envisioned in their year together. But many of their lives remained intertwined, and nearly a century later their insights, methods, and approach to science-based silviculture have been fully vindicated. Their legacies certainly persist. Bob Marshall worked from a distance to publish his master’s thesis four years later, and his career advanced on a meteoric trajectory grounded in ambition, rare talent, boundless energy, and important connections. Just fourteen years after departing Petersham, he died in a train heading to Washington, D.C., likely of heart failure. At thirty-nine, he was chief of the U.S. Division of Indian Forestry and founding board member of the Wilderness Society. He had looked back annually to that grand summer as he updated his life lists. He also maintained contact with Professor Fisher, his friends Hosley and Cline, and figured strongly in key aspects of both friends’ lives. The grand experiment on the Adams Fay lot was resampled three times while Fisher was alive, keeping Cline and a regular stream of new graduate students busy. Yet, like so much of the scientific infrastructure established in the woods during the first quarter century at the Harvard Forest, the big experiment was abandoned following the 1938 hurricane. The neat experimental design of harvests was shredded—initially by the winds that flattened the remaining hemlock and pines and then by the salvage logging that left the landscape covered with stumps, skid trails, charcoal mounds, and residue from a portable sawmill. Al Cline captured the scene and response to the hurricane in the publication (Lutz and Cline 1956) that brought the big silvicultural study to a close: The stand left after the cutting of 1925 was completely blown down by the hurricane of September 1938. The stumpage was sold to a private operator. Because of the tangled condition of the trees, oftentimes piled in criss-cross fashion to a depth of twenty feet, no attempt was made to control the cutting or the extraction; the logging was done at the discretion of the operator. The logs were hauled on scoots by tractors. Although the logging was done in the winter, there was very little snow on the ground; consequently much of the remaining organic layer was broken up and mixed with the mineral soil, particularly along the many skid trails. After logging was completed, the slash was ricked and partly burned. The hurricane and attendant logging operation caused heavy damage to the reproduction; much of it was broken by falling trees or knocked down in the course of logging. Fortunately there was a good crop of pine seed on the trees when the hurricane struck. With the improved seedbed conditions brought about by the sec-

LESSONS: BOB MARSHALL’S PLOT 91

ond shelterwood cutting and the further scarification of the soil caused by cleaning up after the hurricane, a fairly abundant new reproduction started in 1939. It was nearly three-quarters of a century later before the next group of faculty and students refocused on the work that Marshall, Fisher, and the crew had initiated. Though it was challenging for us to relocate Marshall’s original plot given the intervening damage and regrowth, a group that included students Alex Ireland and Ben Mew eventually succeeded, using the original maps and locating persistent landmarks and features that Marshall had surveyed so very carefully. We also revisited and reimagined the original experiment, though the well-conceived cutting patterns can’t be discerned on the landscape today. Eight decades after the hurricane, the forest is now inspiring in many ways. And thanks to financial contributions by Marshall and other alumni, the Adams Fay lot was purchased from the Box Company and now belongs to the Harvard Forest. Its forest conditions repeat the historical pattern that Marshall reconstructed and Fisher predicted. Despite the vicissitudes wrought by wind and ax, both hemlock and pine continue to dominate the site. The hypothesis rooted in forest reconstruction has been supported, and Bob Marshall’s method has been in active use ever since.

LESSONS: BOB MARSHALL’S PLOT 92

FIVE

HEMLOCK AS A F O U N D AT I O N S P E C I E S All trees are equal, but some trees are more equal than others. —with apologies to George Orwell

W

e share the earth with approximately ten million other species. Many of these are widespread but infrequently seen: bacteria, fungi, and many insects, along with organisms that live in places that we are unlikely to visit, such as the deep oceans or the canopy of tropical rainforests. Others are superabundant and visible all around us: the common reed, ragweed, kudzu, and many aggressively spreading “invasive species.” Other species are less threatening but no less abundant, such as corn, cows, and potatoes, which have been transported worldwide by people and have established large populations outside their native ranges. But the vast majority of species are rare. Of the rarities, those large enough (tigers, for example), flashy enough (multihued tree frogs and some orchids), or of immediate use (Pacific yew, the original source of the anticancer drug Taxol) garner the majority of attention from conservation organizations and research scientists alike. Most of us would probably not miss kudzu if it were eradicated, and now that Taxol has been synthesized, the loss of a few yews probably wouldn’t make headlines. People are not likely to miss some treetop fungus that they never even knew existed. But there is a whole additional class of species—those we call foundation species—that are critical to the habitat that they help create. Foundation species generally are so common that we take them for granted, and they receive scant attention from most environmental or conservation professionals. Yet foundation species are so important to their local situation and the larger world that their disappearance would immediately diminish our daily lives. We depend upon the ecosystems that they build and maintain for a wide range of tangible and intangible services: clean air and water, wood products and wildlife, and solitude and inspiration. Eastern

93

hemlock is one such species, so let’s be clear on what it means to be a foundation species. Any foundation species has four key characteristics. The first three are ecological: foundation species are numerically abundant and account for much of the biomass or material in the ecosystem; they occupy the base of a network of interacting species—what ecologists call a food web; and they are well connected in that food web, with linkages to many other species through a large number of associated interactions. The fourth characteristic is perceptual and often deeply personal: without any prompting from scientists, we readily recognize a foundation species as an inseparable part of its ecosystem—we celebrate and enjoy its presence, and we mourn its loss. Marine ecologist, oceanographer, conservation biologist, and all-around mensch Paul Dayton of the Scripps Institution of Oceanography was the first person to define and discuss foundation species. He did so in a 1972 paper, which was buried in an almost entirely unknown and rarely cited proceedings on a conference focused on conservation issues in Antarctica. Outside of a small group of marine ecologists, the foundation species concept remained virtually unknown until a group of ecologists from the National Science Foundation–sponsored Long Term Ecological Research Network introduced it to a broader audience in 2005. Dayton outlined this concept by dramatically expanding on the idea of a keystone species that he and his doctoral dissertation advisor Robert Paine had developed while studying starfish and their prey in Washington State’s Puget Sound. Strongly rooted in the marine realm, the first foundation species that Dayton identified were marine sponges that grow in dense carpets on the seafloor under the Ross Sea. For most of his career, however, Dayton has studied kelp forests in the shallow subtidal waters of the western United States. Many species of kelp are also foundation species; they form dense underwater forests, are at the base of highly productive marine food webs, and tightly link the components of the food web in networks of species interactions through which energy and nutrients flow. Swimming through eerily silent, dimly lit kelp forests evokes mixed feelings of majesty and awe, just as walking into a quiet, shaded hemlock forest in New England reminds one of entering a cathedral. Foundation species create, define, and maintain entire ecological systems. They also determine in large measure how we perceive those systems, both intuitively and scientifically. We may walk through a “deciduous” forest without noticing the relative abundance of individual species such as red oaks, sugar maples, or paper birches. We may characterize “boreal conifer” forests without regard to the relative abundance of spruce, fir, or larch. But we are unlikely to walk into a hemlock forest or a redwood grove, or swim through a kelp forest, without knowing we have entered a special place, a sacred grove in which the vast majority of “trees” are a single species: H E M L O C K A S A F O U N D AT I O N S PE C I E S 94

dark, brooding hemlocks in hemlock forests; massive, towering redwoods in redwood forests; giant kelp gently swaying in rhythm with the waves. Thus our perception of its dominance in terms of size or abundance, which leads us to name an ecosystem for that individual species, is the first essential characteristic of a foundation species. When someone at the Harvard Forest mentions at lunchtime that they will be spending the rest of the day in “the Prospect Hill hemlock forest” there is no question about where in our 3,750-acre outdoor laboratory they can be found. But it is not enough for a species to be overwhelming in stature or numerically dominant for us to recognize it as a foundation species. Ecologists have identified a number of different kinds of “important” species in ecological systems besides foundation species—not only the aforementioned keystone species but also dominant species, core species, and ecosystem engineers. These terms are often used interchangeably with foundation species, depending on the fashion of the time. But scientists seek clarity in such classifications. To more plainly distinguish a foundation species and highlight the differences among different types of species that are crucial to well-functioning forests, we can use an analogy with the architectural features of a New England farmhouse or a massive Gothic cathedral made of stone. Dominant species, such as the pitch pines spreading out across abandoned fields on Cape Cod or Martha’s Vineyard, resemble the individual, nearly identical rows of wooden clapboards nailed in repeated courses up the sides of a farmhouse or the large stone blocks stacked up to make a cathedral’s walls. Unlike old-field pitch pines, yellow birch stands out in the woods but is rarely abundant. This core species among the northern hardwoods—tall in stature and with its distinctive peeling golden bark—is like one of the columns that form the towering arches that give a cathedral its character or the massive central chimney that caps the large and solid house. Core species are usually associated with satellite species, which are analogous to spandrels—the small triangular spaces between the outer edges of the arch and the adjacent wall that provide opportunities for visual interest but which do not provide any critical functionality in a cathedral. In our northern forests, white ash may be a good example of a satellite species; while pretty to look at (and valuable at the sawmill), our forests will not collapse without it. This last statement is unfortunately now being put to the test as the emerald ash borer, an insect that specializes on and rapidly kills ash trees, is spreading from the Midwest through the range of ash, including New England. In contrast, keystone species occupy the top and center of the cathedral’s arch; pull it out, and the arch collapses. As defined by Paine and Dayton, keystone species are the predators at the top of the food web; remove them, and the prey they eat might increase dramatically in abundance, outcompeting other species in the sysH E M L O C K A S A F O U N D AT I O N S PE C I E S 95

Pitch pines invading an abandoned field on Martha’s Vineyard. (David Foster)

tem and ultimately forcing major changes in many other components of the system. Here we can think of the wolves in Yellowstone, whose restored presence has brought elk populations back in check and allowed the heavily browsed groves of aspen and willow to expand. The end result of removing a keystone is an arch collapsed into a pile of unconsolidated rubble. Remove a keystone species, and we are left with a still-functioning ecosystem, but one that has far fewer species and less regulation of its component species than the original. Then there are the ecosystem engineers. These are the builders of houses and cathedrals, and the organisms—usually good-sized animals—that actively structure and shape an ecosystem. The defining example of an ecosystem engineer is the familiar beaver, which alters stream flow with its dams that lead to the creation of a landscape consisting of a diverse array of wetlands and ponds. The more elusive pileated woodpecker provides another example of an ecosystem engineer that, through its creation of distinctive immense and jagged holes in dead trees, provides cavities that afford nesting sites and habitat to many other organisms. All these different types of species are important, but foundation species are more fundamental and even more critical. Foundation species are the stones that support the ecosystem: there are a lot of them, they form the base of a well-supported structure, and they support the entire magnificent edifice. Without the foundation stones (even if they are underappreciated), none of the other parts are even relevant. Remove them, and the entire edifice is in jeopardy. Foundation species need not be showy or distinctive, though they may be. Often, however, they can be individually unobtrusive despite their abundance, so their importance goes unappreciated. In New England forests, eastern hemlock is a foundation species: with no hemlock, there is no distinctive and unique hemlock forest. The combination of the deep shade that hemlock casts, its year-round growth (whenever temperatures are above freezing), its effective interception of snow and rain, and the thick layer of needles fallen from its branches and underlain by poor, acidic soils, creates a unique environment for many other species of plants, animals, fungi, and bacteria. Hemlock canopies support diverse assemblages of spiders (including families absent or rare in nearby deciduous forests), and far more web-building spider species than those that actively hunt. In contrast, mammal, bird, and ant species tend to be less diverse in hemlock forests than in nearby deciduous ones, though many of these species are hemlock specialists or find refuge from predators within dense hemlock stands. It is important to keep in mind that our perception of a foundation species is related to our own size and position in ecosystems: humans are relatively largebodied animals, approaching six feet in height and often exceeding 150 pounds or (much) more in weight at maturity. We can live for over 100 years, far longer than any individual bacterium, most animals and fungi, and the majority of plants. We H E M L O C K A S A F O U N D AT I O N S PE C I E S 97

Beaver lodge and cut trees on the Tom Swamp tract. Many of the flooded dead trees are hemlocks. (David Foster)

Fisher House at the Harvard Forest, a magnificent colonial structure dating to the early nineteenth century and supported by a solid rock foundation. (David Foster)

are omnivorous predators, feeding at the top of the food chain and living high on the hog (or its closest vegetable equivalent). For us, a foundation species needs to be bigger, live longer, support many of the other species that we eat or on which we otherwise depend, and provide important resources and ecosystem services that we cannot live without. A large and long-lived tree such as eastern hemlock clearly shapes the ecosystems that it dominates and thus clearly fits the bill. But can we imagine foundation species at a scale smaller than our own? We also study the carnivorous northern pitcher plant, a common species that grows in bogs throughout Canada and the eastern United States. Ten-inch-tall pitcher plants host entire small ecosystems within their water-filled leaves, and many of the minute animals that live in these tiny pools can be found nowhere else in the world. Pitcher plants can live for decades, create novel habitats, and have distinctive cycles of nutrients and energy. From the perspective of the pitcher-plant mosquito, the H E M L O C K A S A F O U N D AT I O N S PE C I E S 99

pitcher-plant midge, the pitcher-plant fly, and three different pitcher-plant moths, this unique plant is a foundation species. From our perspective as large organisms that can crush a dozen pitcher plants underfoot without a passing glance, they are hardly worth noticing, much less considering as a foundation species. But for a one-quarter-inch-long pitcher-plant mosquito that lives its entire short life within a single pitcher, a pitcher plant is as large as a cathedral and its adjoining monastery. The organisms sharing an ecological system—what ecologists call a community or an assemblage—interact with one another in a variety of ways. Plants are the producers, using photosynthesis to turn sunlight, carbon dioxide, and water into the sugars and starches that make up the roots, stems, twigs, and leaves that are subsequently consumed by a variety of herbivores—literally “plant eaters.” Herbivores, in turn, are eaten by carnivores (meat eaters). After the plants, their herbivores, and the latter’s predators die, their dead leaves, broken branches, lifeless fur, scales, or bones are broken down into their fundamental atoms and molecules by a progressive range of smaller and smaller decomposers. The nutrients released by decomposition are absorbed by nature’s recycling system—complexes of bacteria, fungi, and plant roots (collectively known as mycorrhizae)—which takes up these nutrients along with water so that they can be used once again in growing plants, feeding herbivores, and supporting carnivores A building, whether a New England colonial farmhouse or a Gothic cathedral, also is an assemblage of distinct spaces and rooms. Each is a self-contained unit with its own function and parts, but each is linked to the others by doors, corridors, and hallways. Similarly, in ecological communities we can often find distinct smaller networks, or subwebs, in which species interact strongly with each other but only weakly with species in other subwebs. Just as the blocks of a foundation support all the different parts of a building and provide a platform that keeps the entire structure intact on its firm footings, foundation species link together the many different subwebs in an ecosystem. In the thick, organic soils of hemlock forests, for example, the subweb consists of bacteria and fungi that break down and decompose leaf litter; springtails, worms, and mites that feed on the bacteria and fungi; subterranean ants that feed on the springtails; and predatory beetles and spiders that feed on the ants, springtails, worms, and mites. Up in the canopy, it is a completely different system. Hidden among the evergreen foliage, grubs of wood-boring beetles burrow through bark and sapwood, riddling the trees with burrows and depositing frass, the combination of partially chewed and excreted material generated by insects feasting in tree canopies. These grubs, in turn, are eaten by woodpeckers. The organisms in these

H E M L O C K A S A F O U N D AT I O N S PE C I E S 100

two subwebs rarely interact—perhaps only when the woodpecker flicks slivers of bark and wood toward the forest floor, drops a hard-won grub, or sends its excreta hurtling onto the soil surface. But these spatially discrete, relatively small food webs are linked together by hemlock itself, whose fallen needles create the soil and forest floor, and whose trunks and branches provide a home and sustenance for the beetle larvae. But foundation species do more than provide the support for networks of interacting species that feed on each other. They also facilitate many other interactions that do not involve eating or being eaten. These interactions involve the by-products of life and metabolic activity (excretion or capture of nutrients, for example), as well as the provision of space for nests, reproduction, and socialization, and refuges from predators. In sum, foundation species connect a variety of subsidiary networks responsible for many different activities and functions in an ecosystem. Ecosystems are complex and include thousands of species, whereas the networks of interacting species studied by ecologists usually consist of fewer than a hundred, and often fewer than ten species. Although a knowledgeable and experienced naturalist might be able to identify the various roles that different species play within a given interaction web, it is surprisingly difficult to reliably identify the foundation species that links and supports them. In most instances, we become confident in our identification of a foundation species only after it has declined to such an extent that it is no longer functioning in its former role, or after it has disappeared completely from the landscape. If we can understand the ways that a single critical species links together other species through a variety of interactions, we may be able to identify it before it is too late. As a foundation species declines, the environment itself changes, leading to many direct and indirect effects on individual species and their interactions, and on the cycling of energy and nutrients. It can disrupt feedbacks between organisms, ecosystem processes, and the environment. Past a certain point, the connections between interacting species are severed, and the pathways of nutrients and energy flow are rearranged. For example, as hemlocks decline, resulting in changes to the local microclimate—including soil temperatures and moisture—associated changes in rates of decomposition and availability of carbon and nitrogen may benefit fungi at the expense of bacteria. The altered flow of resources through an ecosystem can itself lead to major shifts in the numbers of associated species. Some species may go locally extinct, while others increase substantially in abundance. Some of these disappear because they have direct and important interactions with the foundation species that cease. Others surge because of indirect effects, in an ecological example of the well-worn adage that “the enemy of my enemy is my friend.”

H E M L O C K A S A F O U N D AT I O N S PE C I E S 101

Once foundation species have disappeared, either functionally or entirely from a system, they will be replaced by other species as the entire system is reorganized. New interaction networks will be established, but none will be linked together in the same way that they were in the system grounded by the foundation species. In some cases, such as in young, early successional hardwood forests filled with red maple, gray birch, paper birch, or aspen, the singular roles that the lost foundation species played are taken over by a range of species, none of which by itself would be considered a foundation species. Ironically, this functional redundancy may derive from an increase in local species diversity and stability of the new stand, which is generally considered to be a positive thing by most people, including ecologists. Indeed, foundations can be rebuilt. House sills can be jacked up, floor joists can be replaced, and new frost walls can be poured. In forests, old foundation species may eventually recover, or new dominant species may replace them. But just as new foundations of poured concrete and pressure-treated timbers are different from rock-rubble walls and massive beams of oak, differences remain apparent. In northeastern North America, American beech provides a useful example. Beech has many characteristics of a foundation species, but its effects on forest ecosystems are decidedly different from those of hemlock. Like eastern hemlock, American beech is shade-tolerant, can become locally very abundant, and creates a distinctive dark environment in which few other species of tree can effectively compete. Beech leaves are quite different from other deciduous species and certainly distinct from hemlock, so the soils and the associated ground conditions in a beech grove are unlike those in many other forests. On the other hand, the decomposition rate of beech leaf litter and the recycling of the nutrients in these leaves in beech-dominated forests do not differ broadly from these same processes in many other hardwood stands. Beech provides plentiful food (in the form of beech nuts) for a variety of animals. But beech mast is episodic, and the bears and small mammals that depend on beech nuts one year may switch to acorns in another. Yet beech grows much farther north than oak, and there is substantial evidence that in oak-free regions, the decline of beech from beech bark disease is having at least subtle, and possibly significant, long-term effects on populations of bear and small mammals. Beech may not be a foundation species, but it clearly plays a major role in forested ecosystems. As illustrated by the complexity of this quick look at the characteristics of beech, it is a major challenge to ascertain whether a species is foundational in any ecosystem. Although we do not think that beech is a foundation species, we do believe that hemlock is. Only through detailed observation and experiments can we H E M L O C K A S A F O U N D AT I O N S PE C I E S 102

Ecologist Glenn Motzkin in a 200-year-old beech forest on Naushon Island, where the dense shade, deer browsing, and intense competition have eliminated nearly all other species. (David Foster)

make that determination. In the following chapters we describe four different but complementary approaches to understanding eastern hemlock’s role in the ecosystem. Through historical reconstruction, a natural experiment, a manipulative experiment, and modeling, we are developing a body of knowledge about the remarkable nature of a hemlock forest. This kind of research program is not a one-way process. The more we study the past and observe the unfolding present, the richer the contextual canvas becomes. Observations suggest experiments, but experiments generate new ideas that require new observations and further development of historical and contemporary contexts. Models force us to clarify our assumptions, and this clarity often forces us to reconsider them, reinterpret our existing results, or send us back into the field to start afresh. But whether we are playing in the lake mud, wrapping diameter tapes around the midsection of a tree in the forester’s embrace, logging out trees to learn how the forest regrows, or eying tons of virtual carbon flashing across a computer screen, we approach our work with the same sense of wonder and awe for hemlock that has inspired generations of poets. I robbed the Woods— The trusting Woods— The unsuspecting Trees Brought out their Burs and mosses My fantasy to please— I scanned their trinkets curious— I grasped—I bore away— What will the solemn Hemlock— What will the Oak tree say? Emily Dickinson, 1859 (Fr 57a) And this may tell us more about foundation species than any of the data we now present.

H E M L O C K A S A F O U N D AT I O N S PE C I E S 104

SIX

A R A N G E -W I D E H E M L O C K D E C L I N E

I

f you were to look at lake-sediment pollen records from places as far apart as Rogers Lake in southern Connecticut and Gould Pond in central Maine, you’d see similar patterns. Indeed, the records from most New England lakes feature a common sequence of shifts in vegetation that accompanied the changes in regional climate since the last glacier receded. Most of the transitions from one type of vegetation to another happened quite slowly, but a few were abrupt and involved dramatic reorganizations of plant and animal communities and their associated ecological processes. Using pollen analysis, paleoecologists have documented these broad trends in climate and vegetation since the melting of the great continental ice sheet. Tundra and then boreal spruce forest dominated during the late glacial interval (15,000– 11,600 years ago), giving way to white pine as the climate warmed during the early Holocene (11,600–10,000 years ago). It was during the next two millennia that hemlock, oak, beech, and other temperate species quite familiar to us in New England today expanded across the region, and they have largely flourished since then with only modest changes in forest composition. Though the period after 8,000 years ago was characterized by gradual shifts in vegetation, one anomalous event stands out: hemlock declined enigmatically and abruptly around 5,500 years ago and then recovered almost completely, following a period of regional scarcity that lasted nearly 1,500 years. With hemlock’s decline across the region today coming amid major concerns for the potential of abrupt species shifts with climate change in the twenty-first century, we look to this ancient dynamic to explore the patterns, causes, and consequences of an earlier catastrophic decline of this important tree species. Because the sudden decline of hemlock was so rapid and widespread, paleoecologists have long puzzled over it. Massive mortality occurred across the entire range of the species, from southeastern Canada and the Great Lakes states to the southern Appalachians. Although hemlock’s pollen disappeared nearly completely

105

at many lake sites, our work with the local records from Hemlock Hollow and the Black Gum Swamp at the Harvard Forest suggest that hemlock persisted in small populations throughout this lengthy decline. It’s likely that similar remnant populations were scattered across the landscape and indeed across the entire range, often far from the lakes that were sampled. The presence of these small, persistent populations enabled the species to eventually reestablish its original regional extent in a broadly synchronous recovery. Paleoecologists have interpreted the hemlock decline in a variety of different ways, and over the past few decades, various explanations have captured the imaginations of the entire ecological community. The major reduction in hemlock abundance was first identified in the 1930s in Connecticut lake and bog sediments by the Yale ecologist Edward Deevey, a colleague of George Nichols who was busily engaged in documenting the modern forests of the region. Given the known moisture sensitivity of hemlock and the emerging consensus among contemporary European paleoclimatic circles that the middle part of the Holocene (around 6,000 years ago) represented the warmest conditions of the past 12,000 years, Deevey suggested that the hemlock decline was the consequence of an interval of warm, dry climate. By the late 1970s and early 1980s, however, North American paleoecologists had begun to reassess many long-held interpretations of climate and vegetation history. This questioning was part of a general movement that emphasized regional variation in the nature of many ecological phenomena and the often different behaviors of the climate and vegetation in Europe and the Americas. Whereas Deevey and his contemporaries had trained scientists to interpret American pollen diagrams in a climatic framework developed from Scandinavian and British peat bogs, independent American data and interpretations began to emerge. This led to many shifts in thinking: the abandonment of established nomenclature and climate designations from Europe; emphasis on the role of the continental glaciers in northeastern Canada, which persisted much longer than those in other parts of the globe; the role of early humans in the extinction of the great megafauna (such as mastodons and large cats, sloths, and beavers); recognition of the ecological importance of natural and human-set fires; an emphasis on certain ecological phenomena leading to lags between climate change and vegetation change; and the potential for other major drivers of ecological change beyond climate. A central leader in the movement to extract ecological information from the records of past dynamics was Margaret Davis, who had completed her graduate studies in the late 1950s with Hugh Raup at the Harvard Forest and is now concluding her career as one of the world’s preeminent paleoecologists and a member of the National Academy of Sciences. Having broadened her perspective from New England to include pollen evidence from across all of eastern North America, Davis A RANGE-WIDE HEMLOCK DECLINE 106

proposed in the late 1970s an intriguing new explanation for the hemlock decline. At an international conference in India, she presented a paper arguing that three aspects of the decline suggested that it must have been caused by some kind of biological agent, such as a pathogen or insect, instead of climate. These attributes included its abrupt and catastrophic impact, its species-specific nature, and its comprehensive occurrence across the entire range of the species. Davis also made the intriguing suggestion that hemlock’s eventual recovery might have resulted from an evolution of resistance to the pest or pathogen and that the original culprit might therefore still be around. Davis waited nearly five years for the meeting’s much-anticipated proceedings volume to appear and reach broader ecological audiences, anguishing through an unfortunate lesson in the inadvisability of publishing breaking news in conference proceedings. Meanwhile, photocopies of her paper circulated widely among colleagues and graduate students excited by the innovative thinking. The importance of this novel interpretation to ecological theory and understanding cannot be overstated. Davis’s paleoecological predecessors would have assumed that a widespread ecological shift could have been driven only by climate, but she suggested instead that the system was being controlled from within. She presented the very real possibility that an everyday process—the local outbreak of an insect pest or disease— could expand to a scale broad enough to eliminate the host species across its range and transform ecological communities from the Appalachians to eastern Canada. Earlier scientists had tended to downplay the role of evolution in the broadscale vegetation changes of recent millennia, but Davis reasoned that natural selection must be at work and consequential in this seemingly insignificant interval on an evolutionary timescale. This work was characteristic of the way in which Davis advanced science through an innovative melding of ecology with the climatic and geological thinking that had dominated the world of paleoecology. With this single talk and paper on biological agents and the hemlock decline, Davis overturned fifty years of established interpretation and riveted the world of ecological research. Her stature growing, Davis was soon elected president of the Ecological Society of America. She encouraged her colleagues to revisit their own records to examine whether this hypothesis and her other interpretations withstood close scrutiny. They did. Paleoecologists soon were presenting a more complex understanding of the dynamics in nature that went beyond making the past more interesting; suddenly, understanding the past became essential to our understanding of the ecosystems that surround us today. In an era when the demise of American chestnut stood out as the sole example of the catastrophic loss of a dominant forest species, many researchers seized upon the long-ago decline of hemlock as a means of investigating the history, pattern, and consequences of the sudden removal and eventual A RANGE-WIDE HEMLOCK DECLINE 107

recovery of a dominant species. Meanwhile, Davis continued with her research and developed her own elegant approach to exploring the past dynamics of both hemlock and chestnut. At the University of Minnesota, where she had been appointed chair of the Department of Ecology and Evolutionary Biology, Davis worked with graduate student Taber Allison and postdoc Bob Moeller on some unusual New Hampshire lake sediments, comparing the characteristics of the ancient hemlock decline with the more recent disease-caused mortality of American chestnut. Their study site, Pout Pond in Belmont, New Hampshire, is one of only a few New England lakes in which it is possible to visibly distinguish the annual layers of sediment. Being able to pick through pollen and other materials year by year enabled the researchers to examine the rate and trajectory of decline in both species with great resolution. The chestnut blight provided the perfect event for comparison, because the fungus that kills the tree arrived on Long Island in 1904, and its spread was closely documented due to the quick and dramatic way it killed all trees of this valuable and beloved species within three to four years of its local arrival. The records that Davis and Allison developed for the two species revealed that the rate of decline for hemlock closely resembled that of chestnut; it had taken less than ten years for most of the hemlocks to die around Pout Pond. These results, which they published in Ecology, refined Davis’s initial hypothesis and suggested that a disease, such as a fungus, may have spread rapidly and quickly eliminated most of the hemlocks across New England. Lacking any way to identify the specific agent involved, Davis subsequently expanded the lineup of potential suspects to include insects, which she suggested might spread at a rate that would be difficult to separate from “instantaneous” in the paleoecological record. She began examining the many native insects for a plausible perpetrator. Among the various species that infest hemlock, the eastern hemlock looper, a voracious gray-green inchworm, caught Davis’s attention. The looper generally occurs in low abundance, so its chronic feeding on hemlock needles goes largely unnoticed. Yet it occasionally proliferates locally and then defoliates and kills patches of hemlocks. One notable infestation occurred on the edge of the Harvard Forest in the 1990s, and a much larger event was observed on the MassachusettsVermont border just a few years later. In both cases, defoliation progressed through the summer from the upper, sunny parts of the canopies downward. Hundreds to thousands of trees were killed in these local areas, yet surrounding regions were unscathed. Each of the outbreaks halted as rapidly as it began, leaving the surviving hemlock to recover. Davis recognized that a closely related species, the western hemlock looper, feeds on western hemlock in the Pacific Northwest. This observation led her to suggest that, through some ill-defined episode of extreme long-distance dispersal, perA RANGE-WIDE HEMLOCK DECLINE 108

haps aided by the jet stream, this western species may have jumped east about 5,500 years ago. Once in new, hemlock-rich territory, the insect could have feasted on the defenseless eastern hemlock and then gradually evolved in isolation into a varied counterpart of its western parent. The disastrous arrival of a novel pest for hemlock shares elements of the story for chestnut. In both cases, an eastern North American tree species succumbed to a novel foe that operated in the absence of any significant predators or any evolved defenses from the host. Davis postulated that, after living many centuries in dispersed and extremely small populations, hemlock would have evolved resistance to whatever organism— pathogen or pest—had decimated it, and then would have been able to rebound across parts of its former range. In yet another insightful contribution, Davis recognized that, at most sites, hemlock did not recover fully after its collapse to reach its predecline abundance, a fact she attributed to a change in climate during the ensuing period. Finally, after years of growing support for Davis’s general proposal, two Canadian scientists added hard evidence and mechanistic backing for her ideas. Working on peat bog sediments from Québec that dated to the hemlock decline, Najat Bhiry and Louise Filion from the Université Laval found numerous insect remains intermingled with tiny hemlock leaf fragments that appeared to have been chewed. When these fragments of mandibles and other body parts formed from decayresistant chitin were compared to modern-day insect collections, they matched those of the eastern hemlock looper. The case appeared solid and thoroughly convincing and has since been published in numerous textbooks as a classic example of ecology, biogeography, and evolution in prehistoric action. The idea that a major ecological shift could be generated by a biological agent captivated paleoecologists and other scientists, and this interest prompted numerous follow-on studies and questions concerning the consequences of the loss of a dominant species. For example, Janice Fuller, a researcher at Cambridge University who subsequently came to work at the Harvard Forest, carried out detailed analyses of two lake-sediment pollen records from southern Ontario to explore how the removal of hemlock affected the dynamics of the subsequent forest. Fuller found that a diverse group of trees took advantage of hemlock’s demise, including birches, oaks, elms, sugar maple, white pine, and American beech. While it wasn’t particularly surprising that different types of trees benefitted from the loss of hemlock, it was interesting that the compositional changes differed somewhat between the two study sites, suggesting that the consequences of hemlock’s demise could vary in important ways across the landscape. In related studies at the Harvard Forest, graduate student Tad Zebryk and David Foster analyzed pollen in the sediments of the Black Gum Swamp and HemA RANGE-WIDE HEMLOCK DECLINE 109

lock Hollow and also counted tiny pieces of charcoal to reconstruct fire history. Although the shady and moist conditions within hemlock stands typically make it difficult for fire to spread through them, a major charcoal peak associated with the hemlock decline in both records suggests that, as the hemlocks expired, wildfires burned the stems, branches, and foliage that had dropped to the forest floor. In another study, University of Toronto scientists explored the consequences of the hemlock decline on aquatic ecosystems by analyzing diatoms—unicellular aquatic algae—in a lake-sediment core from Ontario. They interpreted shifts in diatom communities as indicating a rise in lake enrichment and productivity as nutrients from soils in the watershed leached into the lake. The ecological community viewed these case studies as clever and opportunistic uses of this peculiar interval of the fossil record to glean new insights into the ways that ecosystems work and the role of foundation species. Davis’s hypothesis was consistent with local and regional patterns of pollen and was backed by direct evidence for damage to hemlock by a known defoliator. Yet despite its compelling logic, the Canadian evidence that supported it was geographically restricted, and there really was no way to test whether either a pathogen or insect had actually killed millions of hemlocks across the entire range of the species. Some scientists remained skeptical. For example, in our conversations with entomologist Joe Elkinton at the University of Massachusetts, he dismissed the suggestion that an insect could have been the principal culprit for such an encompassing regional decline. According to his work examining and modeling the outbreaks of many different kinds of insects on an equally diverse range of trees, the damage and resulting mortality patterns inflicted by insects are never as comprehensive, simultaneous, or geographically expansive as that which was experienced by hemlock 5,500 years ago or chestnut in the past century. Elkinton believed that, if hemlock’s decline had been triggered by a biological event, the perpetrator must have been a disease, as indeed it was with chestnut. At the same time, as questions emerged concerning the involvement of insects, a few scientists revisited the possible role of the long-held explanation: climate. By the 1980s and 1990s, researchers began to see that many phenomena related to the climate system were indeed global in nature. This perspective emerged in part from recognition that many changes, including abrupt ones, were being observed simultaneously in areas as far-flung as the Greenland ice sheet, northwestern Europe, and the monsoon region of South Asia. Perhaps the best example of this type of widespread climate change is what happened during the Younger Dryas interval. This fanciful name comes from Dryas octopetala or the mountain avens, a low-growing arctic and alpine plant whose pollen made major appearances only in two cold periods: the Older and Younger Dryas A RANGE-WIDE HEMLOCK DECLINE 110

events. Paleoecologists first found this indicator in European pollen records in sediments dating to the transition from the Pleistocene to the Holocene. Hundreds of paleoenvironmental records from around the world, including New England, have documented the Younger Dryas, beginning abruptly at 12,900 years ago and ending just as abruptly 11,600 years ago. The Younger Dryas and other short-term periods of cold conditions, such as the more blandly named “8.2K event” (8,200 years ago), when hemlock again abruptly declined and recovered, were brought on by changes in the circulation of the Atlantic Ocean caused by the delivery of huge volumes of fresh water from the waning ice sheet into the North Atlantic. (In and of itself, the notion that regional and even global climate could be shaped by the runoff of immense amounts of frigid water released by the collapse of an ice dam in the middle of North America is yet another extraordinary tale of scientific discovery, insight, and controversy.) The accumulating data on these past climatic events, along with extensive global scientific collaboration to understand the mechanisms underlying them, led to a pendulum swing of interpretation back toward a middle position, where the peculiarities of the North American scene were couched within a global framework in which climate remains the major driver of change. Through this work, the very nature of climate and climate dynamics also emerged as much more complex than previously conceived, and, even more important, they are often tied to many subtle and great shifts in vegetation. We have come to recognize that climate change does not just consist of simple broad shifts in mean annual temperature and precipitation. Rather, it is characterized by complicated variability in the extremes and seasonality of temperature and precipitation. These are brought about by changes in the position and strength of major atmospheric and oceanic currents, which, in turn, are driven by variations in solar radiation and atmospheric chemistry. The result is an interacting and often baffling array of potential combinations that include climate regimes and extremes unknown on earth today. With this recognition, it has become increasingly clear that the period broadly associated with the hemlock decline was characterized by shifts in climate, as first proposed by Deevey, and that similar changes occurred on both sides of the Atlantic. As a result, new efforts emerged that have reinjected climate into the story of hemlock decline. Quite serendipitously, our Harvard Forest research group came upon one key piece of evidence in this story as we collected and analyzed new lake-sediment pollen records from several sites on Cape Cod and the adjacent islands. We saw, as Margaret Davis had forty years earlier, that a dominant tree had experienced a pronounced and abrupt decline exactly 5,500 years ago. But in southeastern Massachusetts, in sites lying outside the range of hemlock, it was oak that underwent a major A RANGE-WIDE HEMLOCK DECLINE 111

regional collapse. And not only did oak decline at the same time as hemlock, but it also exhibited a nearly parallel recovery, rebounding to its predecline abundance around 4,000 years ago. This striking correspondence undermined a number of the major arguments advanced by Margaret Davis in her efforts to explain the hemlock decline. Most important, it refuted the assertion that the hemlock decline could be attributed solely to a species-specific insect herbivore or pathogen, independent of any broader environmental drivers. Because two distinctly different tree species were involved—the evergreen hemlock and deciduous oaks—if insects or pathogens were responsible for the declines, then there would have had to have been two quite different organisms simultaneously affecting two ecologically different groups of trees. It appears highly unlikely that two independently operating species would produce essentially simultaneous changes in different species in the absence of some other controlling factor. A simpler explanation is that both declines occurred when two of the most abundant trees—hemlock in inland areas and oak along the coast—were severely stressed by a broadscale driver such as climate change, operating across the entire region. Indeed, analyses of New England sediment cores by our University of Wyoming collaborator Bryan Shuman and other scientists have found a sequence of sand layers occurring in lake sediments at the depths of the hemlock and oak declines. We believe these mineral layers came during intervals of dry climate when the lake became shallower and much smaller in surface area, allowing the sandy, shoreline sediments to be washed toward the deeper, central part of the lake, where the cores are collected. New England can experience major droughts when cold sea-surface temperatures occur in the northwestern Atlantic, drawing in cold, dry, arctic air masses for extended periods. These exact conditions developed in the early 1960s, resulting in the widespread drying out of lakes, reservoirs, and streams and causing severe shortages in public drinking water across the region. These findings suggest that, under the broadly warm and dry climates persisting in the period from about 6,000 to 4,000 years ago, the hemlock and oak declines were initiated and sustained by a series of strong, extended droughts. Our ongoing research focuses on these ecological shifts, and we seek an even better understanding of their variability across space and time, as well as their rates, causes, and consequences. Although climate change now appears to have been the primary driver of the hemlock and oak declines, it certainly remains a possibility that insects or disease were the actual killers of drought-stressed trees. Indeed, it seems unlikely that climate change, even with severe droughts or temperature extremes, could kill trees across the full range of sizes and ages necessary to eliminate a species abruptly from a substantial area that would have varied greatly in local climate conditions. In A RANGE-WIDE HEMLOCK DECLINE 112

most cases, climate change results in gradual changes in vegetation by altering the likelihood of ongoing regeneration. As adult trees grow and eventually die, they are then replaced by seedlings and saplings of other species. This process plays out over many generations, and it is what we expect will occur here in the eastern United States under a warming climate. Over time, the species present in the understory will change, leading to a progressive shift in forest composition to more southern, warmth-loving, and drought-tolerant species. The way in which one set of dominant trees could rapidly replace another set is under a scenario of climate change coupled with a disturbance that kills the larger trees and rapidly opens the forest to the establishment and growth by replacement species. In our current and future landscape, such disturbances will likely be humaninduced—especially logging and development activity that leads to forest clearance—aided locally by agents such as ice damage and windstorms. The fact that a range of native and introduced insect species and diseases has evolved and caused significant mortality to oaks, maples, and hemlock makes it more likely that this type of disturbance will become an increasingly important cause of mortality as the climate changes. In the distant past, before intensive human land use was prevalent, it appears likely that the disturbances to forests were natural physical or biological agents. An abrupt-change scenario appears to be playing out today at a landscape scale in different parts of New England. For example, the island of Martha’s Vineyard experienced a widespread death of oaks over the last few years. Although mortality was scattered across much of the island, entire swaths have abruptly succumbed in some large, concentrated areas. The cause is the combined effects of multiyear defoliations by an array of insects—especially the native fall cankerworm along with the nonnative winter moth and gypsy moth—interacting with a brief drought. Under stress from two years of defoliation, entire hillslopes of oak trees died when a lengthy rainless stretch occurred as insects were attacking for a third year. Oak mortality approached 100 percent across many hundreds of acres, leaving behind all the other species that the insects ignored, including beech, sassafras, red maple, and black gum. Elsewhere across the island, the mortality ranged from 50 percent to practically zero, depending on the type of site and the number of infestations that had occurred. Since that initial event, the weakened oaks have suffered from other insects, including a previously unrecognized native wasp that went into an outbreak in 2012, resulting in widespread leaf and branch dieback and scattered tree mortality. A different but equally important interaction is playing out between the hemlock woolly adelgid and climate in the New England landscape today. While hemlocks are dying rapidly in the southeastern United States from the adelgid, farther north they linger on for many years before succumbing. In central New England, A RANGE-WIDE HEMLOCK DECLINE 113

A wide swath of oak forest on the island of Martha’s Vineyard was killed in 2008 by repeated insect defoliation, coinciding with drought. The remaining trees are largely beech. (David Foster)

Dead oaks surrounded by surviving beech on the Polly Hill Arboretum, Martha’s Vineyard. (David Foster)

the spread of the insect has clearly been slowed by extreme cold winter temperatures. As winter temperatures continue to rise, however, the adelgid’s range will expand, and the rate of hemlock decline in northern states will increase. A powerful interaction among climate change, insects, and trees is occurring at even larger scales across much of western North America, rivaling or even exceeding the spatial extent and magnitude of the long-ago hemlock decline. Widespread outbreaks of the mountain pine beetle are decimating lodgepole pine across tens of millions of acres, apparently facilitated by a long stretch of hot, dry summers and mild winters that allow beetle populations to grow and spread. This massive die-off is progressively affecting expansive areas of the Intermountain West, with increased wildfire risk having severe consequences for water quality and carbon sequestration in the region’s forests. Climate change is generating similarly abrupt ecological changes in Alaska, where warmer temperatures are inducing an increasing frequency of fire in the spruce forests, allowing them to be replaced by broad-leaved deciduous species, especially paper birch. The result is an abrupt shift in many key ecological processes and in habitat for important wildlife species such as moose. Although each of the various scenarios in different parts of the continent is highly complex and region-specific, a general ecological pattern of climate interacting with other physical and biological stresses is leading to major changes in individual species and entire ecological systems. The hemlock and oak declines of the past both involved the loss of dominant trees that played major roles in structuring their ecosystems; these widespread events thus present us with opportunities to examine some specific consequences of the elimination of foundation species. The loss of hemlock and oak was abrupt. It certainly took little more than a century and possibly just a decade or so. In inland areas of New England, the resulting compositional changes were what we might expect when a late successional species such as eastern hemlock experiences high mortality. Hemlock was replaced by early to mid-successional trees including white pine and birch, not unlike many of the changes associated with the present-day loss of hemlock. As described through Dave Orwig’s extensive work (summarized in chapter 7), pioneering but moderately shade-tolerant species such as black birch, and in some cases red maple and white pine, have taken full advantage of the recent hemlock mortality. In many ways, the modern dynamics mirror what happened thousands of years ago. Along the coast, it was a different story. There the prehistoric loss of oak produced somewhat counterintuitive changes in forest composition. At nearly all sites where oak declined during the middle Holocene, beech immediately increased in abundance. It is surprising that beech, a species especially common in the moister, A RANGE-WIDE HEMLOCK DECLINE 116

Wyatt Oswald, Lindsay Day, Elaine Doughty, and Ross Heinemann finishing a sediment core at Hemlock Hollow. (David Foster)

northern areas of New England, would be favored by drought or would even come to dominate under dry conditions, especially on Cape Cod or Martha’s Vineyard, where sandy, well-drained soils are so prevalent. These results have led us to conduct many additional studies and to offer a tentative hypothesis that the coastal area experienced cool rather than warm conditions, in concert with drought. This opposing pattern to the warming that occurred in inland locations would be consistent with the declining ocean temperatures in the northwestern Atlantic (initiated by the tremendous influx of cold runoff from land) that may have driven the region-wide shift in climate. Fewer fires during the oak decline, as evidenced by low charcoal abundance in lake-sediment cores from coastal sites, also are consistent with that interpretation. Although we continue to explore the dynamics, causes, and consequences A RANGE-WIDE HEMLOCK DECLINE 117

of the broadscale collapse of hemlock and the localized shift from oak to beech, we have emerged with one important lesson from these ancient ecological events: natural systems are extraordinarily complex and spatially variable. As a consequence, we should expect that their dynamics will likely surprise us in many ways as human activities continue to perturb the earth’s ecological and climate systems in major ways. As this chapter was being written, we became aware of a new publication by colleagues Bob Booth and Steve Jackson from Lehigh University and the University of Wyoming that explores the hemlock decline in the upper Midwest and discusses it in the context of our work on hemlock and oak. They identify major episodes of drought through their effects on the water table and the growth of raised bogs, which are highly sensitive to shifts in moisture balance. In a head-turning new twist, their data suggest that the last of the hemlock declines appears to precede the most significant drought by some two centuries. In light of these results, the article evaluates the many studies and hypotheses advanced over the years to interpret and explain the hemlock decline. The authors found subtle inconsistencies and problems with each (including both Davis’s and ours). They concluded that no one study or explanation was satisfactory, yet they too were unable to propose a comprehensive interpretation that withstood their own exacting criteria and level of scrutiny. This is the nature of science; it is seldom a pursuit of one ultimate answer, but rather a continual exploration of multiple avenues and complex interactions through which more knowledge and insights are gained and more wide-ranging and thought-provoking questions are raised. We now know much more about hemlock, the global climate system, and the nature of ecological systems than we did when Margaret Davis initially focused the attention of ecologists on this event four decades ago. We have clear evidence that climate was involved, and yet we fail to understand how climate alone could reduce a species rather uniformly across such a great range as that of hemlock. We have uncovered evidence for at least one insect pest that we know can decimate hemlock locally, yet we lack the empirical data or the theory to extrapolate the role of such a species more broadly. And now we see that more than one tree species was involved in the dynamics of the middle Holocene. Oak declined precipitously at the coast, but within the region occupied by hemlock, many other species exhibited only slight reductions. The resulting pattern yields a complex surface of responses across the region. Of note is how ecological knowledge has progressed through our collective understanding of the hemlock decline. This knowledge is emblematic of most of the work by Margaret Davis and of most innovative scientific research and theory. A fellow National Academy contemporary of Davis’s once acknowledged that, with the passage of time, nearly all of Davis’s heralded breakthrough ideas, in which ecological thinking was applied to the paleoecological stage, were ultimately shown to A RANGE-WIDE HEMLOCK DECLINE 118

be incomplete or incorrect. But this colleague concluded his appraisal with heartfelt admiration for her as a scientist and a person. He proposed that the genius of her work lay in its thought-provoking nature, which challenged the best established and emerging minds to engage in testing, refining, and, in many cases, rejecting her hypotheses. Davis’s bold and creative thinking inspired great intellectual effort, through which new techniques were developed and many additional breakthroughs were made in ecology, climate science, and the way we understand the role and future of humans in a changing world. While the specific arguments that she advanced have been refined, transformed, and, in some cases overturned, science has progressed through the pulses of discovery that they triggered. Thus our work goes on, for there are many dimensions of questions that emerge when a major species such as hemlock is abruptly replaced by others that are so different. Among the trees that took advantage of the hemlock decline were oaks, hickories, and beech; these nut-bearing species not only produce abundant and easily harvested food, they also create forests that offer substantially different environments for plants and animals. Beyond recognizing the decline as an abrupt but short-lived disturbance, we must think more broadly about the long-term consequences of hemlock’s replacement for plant and wildlife communities. How did mammals, ranging from small ones such as mice and squirrels to large species such as deer and bear, respond to what must have been a bonanza of available resources? And what about humans? Although people were widespread across the range of hemlock, there is no suggestion that they played any role in the decline itself— as they might have, for example, through the use of fire. There is much evidence, however, that this interval, from around 5,500 to 4,000 years ago, witnessed a great growth in human populations and significant shifts in culture across the eastern United States. Despite the droughts, climate was broadly favorable across the region, as warmer temperatures and longer growing seasons reduced the hardships of winter, presumably encouraged greater survival rates in humans, and enabled many plants to extend farther north. Under these conditions, the loss of hemlock likely proved a boon to humans, supplying nuts from the new trees, fruits from the many other plants, and meat from the diverse birds and mammals that capitalized on the protein- and carbohydraterich nuts that flourished in the absence of hemlock. This area of study is essentially untapped, and we are now pursuing it with archaeologists Elizabeth Chilton and Dianna Doucette, in conjunction with the climate insights provided by Bryan Shuman. As we explore these new areas of interest, we continue to seek the fundamental causes and characteristics of the decline itself.

A RANGE-WIDE HEMLOCK DECLINE 119

SEVEN

I N VA S I O N O F A N E XOT I C P E S T

AN ACT. No. 233. Declaring and adopting the hemlock as the State tree of Pennsylvania. Whereas, The hemlock (Tsuga Canadensis Linnaeus) is still today, as it was of old, the tree most typical of the forests of Pennsylvania; and Whereas, The hemlock yielded to our pioneers the wood from which they wrought their cabin homes; and Whereas, The hemlock gave its bark to found a mighty industry; and Whereas, The hemlock everywhere lends kindly shelter and sure haven to the wild things of the forests; and Whereas, The lighted hemlock at Christmas time dazzles the bright eyes of the child with an unguessed hope, and bears to the aged, in its leaves of evergreen, a sign and symbol of faith in immortality; now therefore, Section 1. Be it enacted, &c., That the hemlock tree (Tsuga Canadensis Linnaeus) be adopted as the State tree of Pennsylvania. —Gifford Pinchot, Governor, June 22, 1931

I

n today’s world of global commerce, organisms move freely to new continents and novel ecological settings. At last count, more than 400 forest insects had arrived in North America during the last couple of centuries, a few with devastating consequences. The eastern United States lays unfortunate claim to having experienced many of the textbook collapses of tree species from introduced insects and diseases. The majestic American chestnut, which provided timber, firewood, and nuts to communities from Alabama to Maine, became little more than an understory shrub because of a fungal pathogen from Asia in the early 1900s. Oak trees have been repeatedly defoliated by the gypsy moth caterpillar, which was purposefully imported to Boston in 1868 by a local entrepreneur, Étienne Léopold Trouvelot, in a fruitless attempt to create a domestic silk industry. In the years since, gypsy moth populations have episodically exploded, chewing up millions of oak canopies and

120

killing thousands of trees. American elm fell in one more pattern of destruction to another Asian fungus, the Dutch elm disease. It arrived in Ohio in the 1930s and rapidly removed the distinctively vase-shaped elm from moist forests, main streets, and city parks across the Northeast. These range-wide declines were possible because the trees lacked effective resistance and the exotic attackers encountered few predators. Currently, a litany of exotic invaders and impacts lies at our regional doorstep—so many that public agency staff refer to them in a dizzying “alphabet soup” of acronyms: beech bark disease (BBD), emerald ash borer (EAB), Asian long-horned beetle (ALB), and hemlock woolly adelgid (HWA). Tree declines from introduced pests generate a variety of short- and long-term effects on forests along with cascading impacts on other ecosystems and associated vegetation and wildlife. Remarkably, however, we have little solid science documenting these collective impacts. As a result, we are only beginning to develop answers to a fundamental question posed by this book: what difference does one species play? There are several reasons for this knowledge gap. The single greatest loss of a single tree species—chestnut—occurred before researchers were equipped with the tools and ethic of regional scientific collaboration necessary to document and interpret this tragedy. For other targeted species, their low abundance and the subtle ecological consequences of their declines have led to relatively little research. For example, the loss of American elm was felt more in city parks and streets than in forests. As for oak, the gypsy moth has reduced its growth and local abundance but seldom kills all the trees. Finally, scientific funders are not well positioned to support the sustained efforts required to evaluate the consequences of decades-long and regional-scale species declines. Grants from federal agencies generally range from two to five years and require a focus on detailed hypotheses. Stringing together a series of these awards into a consistent long-term study on a single species is daunting, especially today, when the average success rate on grant applications hovers in the neighborhood of 6 percent. Meanwhile, agencies that focus on forests, insects, and pathogens might be expected to seek such information so that they can anticipate how our forests may change in the future and thereby plan responses to species-specific declines. But there are many forest stresses and organisms to evaluate, and most resources are funneled to immediate, pressing problems. Consequently, the basic but critical questions of how forests respond to major invasions and how landowners and society should view and manage these changes go largely unanswered. It was this information gap that we sought to close when we geared up to study hemlock and the adelgid in the late 1980s. We seized on the expanding outbreak

I N VA S I O N O F A N E XOT I C P E S T 121

A foot-diameter beech tree with heavily fissured bark, indicative of beech bark disease. (David Foster)

A dead elm tree on Main Street in Groton, Vermont (2009). (David Foster)

of this new threat to our forests to make a concerted effort to evaluate the consequences of a decline in one tree species as it unfolds. In the face of the funding challenges, this work has required a bit of ingenuity and employed a range of resources from national agencies, foundations, and state sources along with a long-term commitment by the Harvard Forest. In 1951 the hemlock woolly adelgid was unintentionally shipped into North America on a Japanese hemlock meant to add variety and beauty to a backyard garden in Virginia. Over the next thirty years, the adelgid moved into and beyond nearby woods to infest millions of trees across eighteen eastern states. Just half a millimeter in size (roughly the size of a poppy seed), this insect attacks any eastern hemlock, regardless of size, age, or condition, and can kill entire groves in five years. Unlike many damaging insects, the adelgid does not eat leaves but rather competes with the plant for resources by extracting sugars and other food from it. The adelgid accomplishes this by settling along hemlock twigs at the base of the needle, inserting a long feeding tube or stylet into the twig, and feeding contentedly for months. Amazingly, we still don’t understand how it kills the trees. Some argue that it is purely through a depletion of resources. Others speculate that the adelgid exudes a toxic saliva that exacerbates the effects of resource loss. New studies reveal that the insect initiates a defensive reaction by the plant including local cell death, perhaps as a last-gasp effort to starve the insect. Over time this reaction may kill an entire branch. The biology of the adelgid explains its rapid increase and devastating impact. The population of tiny insects consists entirely of females, each of which can lay as many as 150 unfertilized eggs that are viable through parthenogenesis, a form of asexual reproduction. Since the insect completes two generations of young each year, staggering insect densities can grow on trees within a short period. A backof-the-envelope calculation reveals that a branch with only ten adults can produce more than thirty million individuals in two years. Extrapolate that to a fifty-acre stand of hemlock for an idea of the proliferation that occurs anywhere adelgids have entered a forest. Each infestation represents a tremendous source for dispersal and population growth. Because the adelgid is small and abundant, it is easily transported. The eggs and nymphs can easily hitch a ride on birds’ feathers, on deer and other mammal fur, and on humans’ clothing and vehicles, logs and firewood, and nursery plants. High winds during late spring and early summer, when the adelgid is mobile, blow large numbers of the insects to adjacent trees or much farther. Migrating birds assist in this long-distance transport. One study revealed that nineteen of twenty-two bird species captured leaving hemlock forests carried the adelgid. Several of the responsible species spend little time in tree canopies, suggesting that some adelgids must I N VA S I O N O F A N E XOT I C P E S T 124

View upward into a canopy of dead and dying hemlock in different stages of disintegration. (David Foster)

attach to birds on the ground. For an insect that almost never flies, the adelgid spreads rapidly over remarkably great distances. Just as hemlock is able to photosynthesize in winter, the adelgid feeds and develops on warmer days in winter and early spring, when most insects are dormant. Extreme cold, however, is limiting its spread northward. Temperatures below −13°F begin to kill the adelgid and have kept it from reaching the northern end of hemlock’s range. This cold sensitivity increases as winter progresses; the adelgid suffers mortality in March when exposed to temperatures that were not lethal in early winter. Winter temperatures are warming, however, and genetic analyses indicate that the insect is slowly adapting to severe cold. Thus, over time we expect that the adelgid will spread throughout hemlock’s range. Given scientists’ mixed success in documenting the consequences of the chestnut blight and other outbreaks, there were no blueprints to follow when we set out to investigate hemlock’s response to the adelgid. When asked how I would approach that task during my initial interview at the Harvard Forest in 1995, I underscored that we would have to start from scratch. I then outlined an approach that treated the adelgid’s arrival as a natural experiment in which the dynamics of the insect, individual trees, and entire forests could be investigated at multiple spatial scales over multiple decades. We initiated that effort by setting up permanent study plots in hemlock forests across the region that varied in infestation level, from heavy to light to none at all. Since then these plots have helped us document the rate of adelgid infestation, hemlock decline, and environmental and vegetation changes. As our study and interests expanded, we used aerial photographs to map every hemlock stand in a 3,000-square-mile region, from the shores of Long Island Sound in southern Connecticut to the Green Mountains of Vermont. From this detailed map we selected more than 100 forests to study intensively every few years. We also turned to our own land, which at that time was still free from the adelgid, to turn the hemlock-rich Prospect Hill and Simes tracts into comprehensive laboratories of hemlock and its demise. Working with many colleagues and teams of undergraduate and graduate students, we established a network of permanent measurement plots and affixed dendrometer bands on hundreds of trees to measure slight changes in their growth. We embarked on long-term sampling regimes of the soil, vegetation, forest canopy, and streams through intensive field campaigns and the deployment of a wide array of digital sensors, and then set about to initiate large-scale experiments. Through this expansive local and regional effort, we came to appreciate the challenges that had impeded previous efforts to document the declines of other tree species. This kind of research requires an unusual combination of ingredients: anticipation of the magnitude of the unfolding event long before it emerges broadly; I N VA S I O N O F A N E XOT I C P E S T 126

Dave Orwig describes the adelgid and hemlock to a group of Harvard College freshmen. (David Foster)

sustained commitment of researchers and their institutions, including leadership at multiple levels; substantial funding from many sources; flexibility in research execution as surprises occur; and strong collaborations with landowners, state and federal agencies, and other academic institutions. The single biggest surprise was the challenge in funding. Conducting our work has required an overlapping series of grants. Over three decades we have cobbled together support from the Long Term Ecological Research project and other programs at the National Science Foundation, the U.S. Forest Service, and many small foundations. Rejections have vastly outnumbered successful awards and were often based on frustrating rationales: the adelgid and hemlock were viewed as local problems and not of national significance, or the work was seen as too applied or, conversely, of little interest to the forest industry. The second greatest challenge was responding to unexpected events. Surprises came from the adelgid itself and landowners of the properties we studied, such as (respectively) declining abruptly after a cold winter or clear-cutting their forests despite prior arrangements to leave them intact. The very nature of academic science also posed challenges, since research is never a regimented business but rather a complex collaboration in which graduate students, postdoctoral researchers, and senior colleagues have great freedom to design their individual studies, which then need to be stitched together into a coherent framework. The activities and resulting data accumulated over nearly twenty years bear testament to both the challenges and the substantial scientific legacy that such efforts can achieve: • 1,100 plots established in the field and more than 30,000 hemlock trees measured • 4,000 soil samples dug, analyzed chemically, and permanently archived for future studies • More than sixty papers published • Forty undergraduate students and five graduate students trained, plus the involvement of 90 senior collaborators • A secure digital archive of all data and publications that is freely accessible through the Internet The result from this effort is a rich picture of a major ecological dynamic—the loss of dominant species from invasive organisms—that will unfortunately become more important as global commerce and environmental changes increase. For hemlock the picture is grim. Within a year or two after the adelgid reaches a forest, tree health begins to decline noticeably. Initially, new growth declines, the foliage thins throughout the canopy, and needles fall prematurely from interior I N VA S I O N O F A N E XOT I C P E S T 128

branches. The latter is especially noticeable on a wind-free day, when falling needles patter onto the forest floor like a gentle shower. As the infestation proceeds, needles fail to develop on most branches, and foliage becomes clustered toward the tips of branches and the uppermost part of the tree. Eventually the tree dies. Across our south central Connecticut plots, where the adelgid has been present for two decades, mortality has been as high as 98 percent, but it varies considerably. More than 80 percent of trees are dead at seven sites, and most others are in sharp decline. Yet at three northern sites, mortality rates remain below half of this. In all cases, seedlings and saplings generally succumb first, followed by larger trees. Once a tree is seriously in decline, a native insect, the hemlock borer, speeds mortality by laying eggs that develop into juveniles that feed beneath the tree’s bark. The presence of the borer is signaled by piles of purple-red bark at the base of dying trees, torn off by woodpeckers feeding on the borer larvae. An attractive mushroom—the hemlock varnish shelf fungus—often appears along the roots and on trunks of the tree, helping to decompose the wood during this process. Despite the bark’s decay resistance, hemlock wood deteriorates rapidly. Dead trees typically stand eight to ten years: fine twigs and branches are shed first, then the large branches and upper crowns, and finally the trunk topples to the ground. The rapid disintegration of hemlock contrasts with American chestnut, whose resilient dead trunks remain plentiful in our woods nearly 100 years after their death. The progressive loss of the hemlock canopy triggers dramatic changes in forest and stream environments, rippling out to affect many organisms and ecosystem processes. As light levels increase threefold or more, the abundance of other plants often explodes. Black birch, a deciduous hardwood tree, most commonly replaces hemlock. This fast-growing species occurs sporadically in hemlock stands, but it produces abundant seed annually and germinates readily in the thick organic soil. As light levels rise, carpets of birch seedlings appear and grow rapidly into dense sapling thickets. In just twenty years, a new birch forest is solidly established. Red maple, red oak, and white pine also frequently replace hemlock. In the resulting forests, dark understories are transformed into lush green displays of sedges, pokeweed, pilewort, and ferns, along with the low shrubs partridgeberry and wintergreen, and the exotic Japanese barberry, tree of heaven, and Japanese stilt grass. Changes in forest function—decomposition, nutrient cycling, soil respiration, and biomass production—are tightly linked to species composition, which strongly controls the woodland environment and the chemistry of leaves and other material that form the soil. When American elm, chestnut, and various oak species declined in the past, these hardwoods were replaced by species that were similar in seasonal leaf fall and chemistry. In contrast, hemlock—a needle-leafed shady evergreen— I N VA S I O N O F A N E XOT I C P E S T 129

A hemlock tree adjacent to the Hemlock Tower, equipped for the sampling of bark respiration. (David Foster)

shapes the soils, understory environment, and flows of energy and nutrients in fundamentally different ways. With increased sunlight penetrating the forest, air and soil temperatures increase by several degrees. Soil moisture also increases as hemlocks die and stop taking up moisture, but then declines as rapidly growing hardwoods suck up water at rates much greater than the hemlocks ever did. Streams within these forests consequently experience increases in light, water temperature, plant growth, and moisture when hemlocks die, but then they decline in water flow as hardwoods come to dominate the scene. Nutrient cycling changes profoundly as soil moisture and temperature increase, enhancing the decomposition of leaves and the thick organic material beneath the hemlocks. Decomposition releases nutrients such as nitrogen, freeing previously tied-up resources to the remaining plants. This nutrient pulse often persists just a few years, however, due to fierce competition among newly established hardwoods, understory plants, and billions of soil microbes. Longer-term changes in nitrogen cycling result from the strikingly different leaf chemistry of the hardwoods. For example, black birch leaves have more nitrogen and are less acidic than hemlock needles, both of which lead to much faster rates of nutrient cycling. Our collaboration with Bernhard Stadler from Germany has revealed that the mere presence of the adelgid alters forest function long before any signs of decline are observed. The hemlock’s defensive response to the insect results in higher nitrogen concentrations in the needles, stimulating the growth of bacteria, fungi, and yeasts. Rain and snow fall through the enriched branches, carrying carbon, nitrogen, and potassium to the soil, where they enhance nutrient cycling and productivity. Short- and long-term changes in carbon cycling also accompany the hemlock decline. Like all trees, hemlocks absorb carbon dioxide during photosynthesis and store large quantities of carbon in leaves, wood, and roots. As the trees die, carbon uptake ceases. Simultaneously, the release of carbon dioxide from the forest to the atmosphere accelerates as leaves, wood, and roots decompose and soils warm and decompose faster. These great fluctuations in carbon dynamics stabilize as young hardwood forests become established. The rate of carbon storage in forests of black birch and red oak is higher than in hemlock forests. Even so, it may take a century or more for a new hardwood forest to store as much carbon as the old hemlock forest, because of the significant pulse of carbon release and interruption in carbon uptake inflicted by the adelgid. To assess the past and future impact of the adelgid, we need to understand its rate and mode of spread. But how can we detect and track something that we can barely see? In the canopy the insect is invisible even with binoculars. And unlike most flying insects that can be lured and trapped, the adelgid moves only passively. I N VA S I O N O F A N E XOT I C P E S T 131

One approach for detecting the adelgid in forests emerged from the efforts of undergraduate Joe Brown, who spent a summer at the Harvard Forest working with entomologist Scott Costa from the University of Vermont. Painstakingly sampling more than 1,700 hemlock trees across many sites, Brown and Costa developed an effective but admittedly tedious method for locating and quantifying the severity of adelgid infestation. This approach, which involves inspecting branches for the presence or absence of the insect on trees at semi-random directions in a stand, provides an objective means of assessing its local distribution and movement. Needless to say, sampling across a large forest remains exceedingly taxing and time-demanding. We used Brown’s approach in forests across southern New England to examine the factors influencing the adelgid’s spread. Generally, we discovered that large stands often had higher mortality than small stands, as did those on ridge tops and those facing west—sites that are typically warmer and drier. We also confirmed that mixed forests of hardwood and hemlock are just as susceptible as pure hemlock stands. Region-wide, the intensity of infestation and hemlock mortality diminished to the north, parallel to the insect’s migration history. The trajectory of the decline is also progressive: the longer the stand is infested, the greater the mortality of hemlock. We did document major declines in the insect triggered by severe winters, confirming that temperature controls the severity of infestations and rate of hemlock mortality. But then a series of mild winters in central Massachusetts as we were preparing this book (2012–13) enabled small populations of the adelgid to explode and spread broadly across the Harvard Forest. Overall, the threat to hemlock is expanding relentlessly. Through this regional sampling, we uncovered a major indirect impact of invasive insects and pathogens: widespread harvesting of host trees. When landowners see their trees begin to decline, many bring in loggers to remove deteriorating trees from their landscape or to derive income from the timber. Other landowners, in anticipation of future infestation, cut host trees well before the insect arrives, something we are also witnessing with the emerald ash borer. This human response is important, because the extent and intensity of harvesting can easily rival the impact of the insect. For example, we observed hemlock harvests in 35 percent of the forests in Connecticut, ranging in intensity from thinning out of hemlocks to large clearcuts of hemlock stands along with more commercially valuable species. In Massachusetts nearly half of all stands sampled had been harvested, although only half of these were actually infested with the adelgid. Conversations with landowners, land managers, and consulting foresters confirmed these observations and highlighted several motivations behind harvesting: income, public safety, aesthetics, and perceived fire hazards. We seized on this widespread harvesting to expand our natural experiment I N VA S I O N O F A N E XOT I C P E S T 132

Harvesting logs and pulpwood with large equipment on the Harvard Forest Prospect Hill tract. (David Foster)

study to compare the consequences of logging versus the insect. One major objective was to develop guidelines for landowners regarding management responses. Black birch replaced hemlock on both cut and infested forests, but logged sites supported many more weedy and light-demanding understory species, including a much greater abundance of raspberry and sedges. Nitrogen cycling and availability was much greater after harvesting, especially on clear-cuts. Logging damage to residual trees and soil disturbance in removing logs was much more severe than the simple gradual death of trees from the adelgid. Our advice to landowners is to take these greater impacts into consideration when devising management strategies on their land. I N VA S I O N O F A N E XOT I C P E S T 133

Hemlock mortality also triggers dramatic changes in wildlife habitats. The variable nature of these changes is particularly highlighted with white-tailed deer: the rapid regeneration of hardwoods provides them with abundant browse, but the loss of the hemlock canopy eliminates cover that can reduce survival rates in severe winters. Snowshoe hare, porcupine, fisher, and bobcat are all likely to feel the loss of winter hemlock cover. Many birds that spend part of their life cycle in tree canopies are responding to the widespread loss of hemlock. A southern New England study led by Harvard undergraduate student (and current Princeton postdoc) Morgan Tingley showed that hemlock specialists such as the black-throated green warbler, Acadian flycatcher, blackburnian warbler, and hermit thrush were much less common or completely absent in former hemlock woods. As their favored habitat declines, these species are expected to suffer population declines and even contractions in their range. In contrast, many woodpeckers have benefited from the widespread tree mortality, and various songbirds, including the eastern wood pewee, red-eyed vireo, and veery, prefer black birch to hemlock. In other hardwood forests, the white-breasted nuthatch, tufted titmouse, brown-headed cowbird, and scarlet tanager have also increased. The hooded warbler, a locally rare bird generally restricted to coastal areas, has spread north beyond its historical range limit in forests succeeding hemlock. Terrestrial invertebrates are likely to increase in richness and overall abundance as hemlocks are replaced with hardwoods. Some species of ants, beetles, and grounddwelling spiders benefit following hemlock loss, while web-building spiders, several mites, centipedes, and millipedes are predicted to decline. We have not yet studied impacts on stream invertebrates, so our best predictions come from contrasting hemlock- and hardwood-dominated watersheds. Studies on our Prospect Hill tract suggest that the loss of hemlock will lead to an increase in the diversity of aquatic invertebrates. In turn, these modifications in food-web structure and stream habitat will likely trigger shifts in fish assemblages. In a study of headwater streams in the mid-Atlantic region, brown trout and brook trout were two to three times more abundant in hemlock streams than in hardwood streams. One concern for brook trout in particular is that the loss of shade will warm streams to temperatures above their survival limit. Many factors will figure into these dynamics, including the rate of hemlock mortality, the type of replacement species, and the role of groundwater in stabilizing the temperature of a particular stream. For fish, insects, birds, or mammals, in any alteration of habitat, there are both winners and losers.

I N VA S I O N O F A N E XOT I C P E S T 134

The already complex story of the adelgid developed another twist when a second invasive insect, the elongate hemlock scale, began to spread widely and interact with the adelgid. The scale was already in place (it had arrived in New York City from Asia in 1908), but it had caused little damage and therefore had gone largely unnoticed. Although it has many conifer hosts, in New England the scale prefers eastern hemlock. It’s mobile only in the spring, when its juvenile crawlers settle on the underside of hemlock needles and insert their feeding tubes through pores in the needles to extract nutrients. Once it’s actively feeding, the scale maintains a protected environment for itself and its eggs by producing a waxy outer shell or scale. Its life cycle differs from the adelgid in two distinct ways: adults reproduce sexually, and females lay only about twenty eggs once a year. Like the adelgid, however, the tiny (one-millimeter-long) scale is easily dispersed by wind and birds. Most scale infestations are overlooked until a large population develops. After multiple years of heavy infestation, trees begin to turn gray and suffer reduced growth and early needle drop. The scale is typically not as harmful as the adelgid, but it can weaken or kill already stressed trees. After remaining in the New York City area for decades, the scale evolved increasing cold tolerance and began to expand rapidly in the 1970s. It now occupies at least fourteen eastern states, where it often interacts with the adelgid. In our study region, the scale’s presence increased from less than 30 percent to nearly 90 percent of stands in the decade preceding 2009. We confirmed that scale infestations intensify in the presence of the adelgid, likely because of the enhanced nitrogen in needles triggered by adelgid feeding. Conventional wisdom suggests that two pests would be worse than one, but that appears not always to hold in this situation. In our collaborator Evan Preisser’s studies, the adelgid and the scale were 30 percent less dense when they co-occurred on a branch, and other research suggests that the scale even reduces the impacts of the adelgid. The scale’s recent invasion may be contributing to the observed slowing of hemlock mortality in parts of southern New England in the last decade. How long will this benefit from scale endure? The story continues to unfold, but it highlights the complexity of invasions and the challenges of making accurate predictions. Much has been learned about adelgid and its impacts on forests since we established our first long-term plots back in the 1990s. Scores of researchers and dozens of studies have provided a broad understanding of the many changes that a small invasive insect can trigger in an important forest ecosystem. Still, we have much more to learn about how the adelgid affects tree mortality, interacts with other insects, and indirectly affects hemlock forests. I N VA S I O N O F A N E XOT I C P E S T 135

EIGHT

CUT OR GIRDLE

The way a crow Shook down on me The dust of snow From a hemlock tree Has given my heart A change of mood And saved some part Of a day I had rued. —Robert Frost, “Dust of Snow”

T

he long-term observational studies of forest change described in the preceding chapter have enabled us to document in detail the natural dynamics and decline of hemlock forests throughout New England. These observations continue to be critical for understanding possible future conditions of our forests. They also provide a wealth of insights that can guide land managers and policy makers, and that we regularly use in discussions with owners, foresters, and conservation organizations. By their nature, we have had little control over the timing and logistics of these studies other than making the commitment to undertake them. We cannot decide when the adelgid colonizes a stand or when and where a landowner brings in a logging crew. It can be difficult to compare these “natural” events to various types of conservation management on equivalent sites. Nor can we readily determine from observational studies whether the changes we see in our forests as hemlock declines are simply the result of trees dying and falling apart, or whether the adelgid itself modifies the environment, leading to novel conditions that alter ecosystem processes and redirect successional pathways. Faced with these obstacles and in situations

136

where our powers of observation and analysis cannot unambiguously answer certain questions, we turn to field experiments. Large experiments, including simulation of clear-cuts, hurricanes, acid rain, and global warming conducted at the Harvard Forest, Hubbard Brook, and other well-known research institutions enable us and other scientists to pin down the underlying causes of the environmental changes we record in observational studies. For ten years now, we have been engaged in the Harvard Forest Hemlock Removal Experiment, known informally as Cut or Girdle. This experiment is a cornerstone of our Long Term Ecological Research program: a research initiative begun in 1988 and funded by the National Science Foundation. It uses a range of multidecade observational and experimental studies to understand the New England landscape and compare it with regions across the United States. As a site-based experiment, this study represents one of the four complementary approaches (along with paleoecological and historical reconstruction, landscape-level observations, and regionwide modeling) that we are employing to study how New England forests respond to the loss of a foundation species such as hemlock. The essential objective of Cut or Girdle—like that of any ecological experiment—is to identify a small number of critical processes (usually one or two) that drive the dynamics of an ecosystem, and then to simulate these processes so that we can determine how large an effect they really have. The areas subjected to the simulations are paired with adjoining identical plots that are left untouched to serve as controls or references for comparison. Unlike experiments carried out in a laboratory or in greenhouses, experiments that are done in the field allow the natural ecosystem to react to the experimental manipulation under the changing environmental conditions of the real world and in response to idiosyncrasies of the particular experimental sites. An effective field experiment thus comes quite close to simulating real-world processes and conditions. Nevertheless, the hallmark of a successful field experiment is that the “signal” (the effects of the manipulation) should far exceed the “noise” caused by small-scale differences in site characteristics or ongoing environmental changes. Such experiments provide a degree of control over when, where, and how a forest is altered, and they enable repeatability in both time and space that can never be achieved with observational studies. As a consequence of all these attributes, experimental results represent the gold standard of scientific evidence. At the same time, no experiment, including Cut or Girdle, can simulate every possible factor driving the ecosystem or control for every contingency or random event that might occur, especially when the experiment is designed to run for many years. With Cut or Girdle, we are interested in determining how a representative New England forest changes, depending on how the hemlocks are killed and how quickly they disappear from the forest. A process taking five to twenty years as adelCUT OR GIRDLE 137

gids colonize a stand and slowly kill the trees, leaving a ghostlike stand of decaying trees, differs in more than duration from a one- to ten-week period in which a landowner may contract with a logger to preemptively salvage the hemlock before the adelgids arrive. In establishing the study in 2003, we took advantage of two opportunities. First, the extensive land base of the Harvard Forest enabled us to locate this nearly twentyacre experiment in a section of mature hemlock forest that is broadly representative of hemlock stands found from Pennsylvania to Québec. Second, the adelgid had not yet arrived in our woods, so we were able to design the study with its presumed future arrival in mind. We could then compare the effects on the forest of two different ways that hemlock could die slowly—by scientists simulating adelgids through girdling the trees or by the action of the adelgids themselves. The study includes two kinds of experimental treatments and two kinds of controls. Each of the plots—experimental or control—is a ninety- by ninety-meter (295by 295-foot) square, covering just over two acres, or about the size of three football fields laid side by side. There are two of eight plots for each treatment plus two matching plots for each control, for a total of eight plots. For the first experimental treatment, we initiated the slow death of all the hemlocks by girdling them: cutting a continuous ring around the trunk with chainsaws or knives, which severs the sugartransporting phloem. For the other, “logged” treatment, we mimicked an intensive (but realistic) preemptive salvage harvest by cutting and removing all the hemlocks larger than eight inches in diameter, along with some additional merchantable trees. We paired each girdled and logged plot with a “hemlock control” plot in which no girdling or cutting occurred and a “hardwood control” plot dominated by black birch and red maple, which was similarly left alone. The hemlock control plots provide the formal controls for the girdled and logged treatments. Day-to-day and year-to-year environmental changes will influence all the plots more or less equally, but we expect the forest dynamics in the girdled and logged treatments to differ in many ways beyond the background growth that will occur naturally in the hemlock control. The hardwood control provides a bit of a different perspective. This forest, with its abundance of black birch and other hardwood species, represents what we expect most hemlock stands in northeastern North America to look like in fifty years or so, based on our long-term observational studies across central and southern New England. At the start of the experiment, all the hemlock stands studied were free from the adelgid. We sampled the forests and collected data in 2003 and 2004. After tallying all the information on the intact forests and their environment, we then applied the girdling and logging treatments in the early months of 2005. For the next five years we were able to compare how changes in forest processes caused by the slow CUT OR GIRDLE 138

A recently girdled hemlock with an insect sampler. (David Foster)

physical disintegration of hemlock or its rapid cutting and removal differed from natural changes from growth and events such as an ice storm in 2008. But, as we had expected from the beginning, we found adelgid on a few trees in the plots in 2009. Thus, beginning in 2010, the focus of the experiment changed. Now, instead of comparing girdled or logged stands to determine the effects of the adelgid on intact forests, we are comparing the effects of physical loss of hemlock (from girdling) to the effects of mortality in the hemlock “controls” due to the adelgid. The logging treatment continues to represent the alternative trajectory when landowners choose to harvest their hemlocks before the adelgid arrives. The site we selected is on the 325-acre Simes tract of the Harvard Forest in southeast Petersham. This outlying tract has an unusual institutional history, because it was donated to Harvard more than half a century after the Forest’s founding in 1907 and has been rather quietly ignored until now. In the center of the tract, where we situated the experiment, large hemlock-dominated stands are interspersed with hardwood forests, rocky ridges, and low-lying wetlands. The hemlock stands are second-growth; most trees are 80–100 years old, with a few hemlocks that approach 150 years in the sheltered valley where one set of the experimental plots is located. Records from the last century, barbed wire, and multiple-stemmed hardwoods attest to the site’s history of active use for agriculture and as a woodlot. Twentieth-century timber removals began in the early 1930s, and, soon after, many of the large trees across the tract were blown down by the 1938 hurricane. What we know about the history of the Simes hemlock stands fits nicely within the range of histories reconstructed in a detailed exploration of hemlock over the past four centuries by Jason McLachlan and David Foster. Hemlock dominated the six hemlock plots, along with scattered red maple, white pine, red oak, and black birch trees. Tree seedlings and understory plants were sparse. In contrast, in the hardwood control plots hemlock constituted less than a tenth of the initial basal area—a useful measure of forest composition defined by the cross-sectional area of the tree stems. Both hardwood plots had many black birches and red maples in the canopy and subcanopy, and supported more understory plant species and cover than did the hemlock plots. The species makeup of the two hardwood plots, however, presents two different scenarios for future trajectories of hemlock stands after the adelgid has moved through. The valley hardwood control plot has a much higher percentage of sugar maple and red oak, whereas white pine accounts for about one-third of the basal area in the hardwood control plot on the ridge. Understanding how a forest ecosystem changes through time and in response to environmental and anthropogenic factors requires diverse talents and a major team effort. A group of plant, animal, and community ecologists, plant physioloCUT OR GIRDLE 140

Aaron Ellison prepares field equipment on a logged plot on the Simes tract. (David Foster)

gists, ecosystem scientists, and wildlife biologists all have worked together on this large experiment. Audrey Barker Plotkin (Harvard Forest) coordinates the activities of the different researchers, and, with Kelley Sullivan (Harvard Forest) and Elizabeth Farnsworth (New England Wild Flower Society), keeps tabs on the plants— their germination, growth, death, and decomposition—and the arrival and spread of the adelgid. Aaron Ellison and Liza Nicoll (Harvard Forest) monitor environmental conditions (air and soil temperature, light availability) and keep the data up-to-date for everyone. Aaron and Tara Sackett (McGill University); Brooks Mathewson (Harvard Forest); Ally Degrassi (University of Vermont); and Ed Faison (Highstead) watch the animals: ants, spiders, and beetles; salamanders; rodents and shrews; and deer and moose, respectively. Kathleen Savage and Eric Davidson (Woods Hole Research Center) measure carbon dioxide as it moves from the soil into the atmosphere, while Adrien Finzi (Boston University) and his students assess longer-term changes in storage of carbon in the soil. Dave Orwig, with the help of Heidi Lux, keeps track of nitrogen as it moves in and out of the soil and helps to compare these results to the changes occurring in the larger landscape. We use two lenses—our experience observing hemlock forests and the adelgid, and our detailed knowledge of the site’s land use history—through which to view and interpret the experimental results. As ecologists studying the effects of disturbances such as insect outbreaks, disease, timber cutting, and wind on the forest, we sometimes find ourselves in the ethically murky position of planning the death of trees in order to better understand the forests we love. We have pulled over trees to simulate hurricane damage and cut trees to simulate windthrow gaps. As we homed in on the plot locations on the Simes tract, regret crept into our conversations. All the Harvard Forest studies—all the way back to Bob Marshall’s work—have shown repeatedly how resilient the forests in our region are. Yet being responsible for the loss of even a bit more hemlock galled us. We expected the logging treatment to be more extreme than girdling and to have a more jarring effect on all of us. But the prospect of watching the girdled trees gradually die and disintegrate in place was more painful to many of us than cutting the trees and using their wood for some practical purpose, such as siding the new machine shop and garage that our Woods Crew was building. Our own discomfort helped us understand some of the real emotions that have led landowners to cut their hemlock forests preemptively as the adelgid has moved north into New England. Now, five years after we initiated the experiment and cut and girdled those hemlock stems, our regret is tempered by the dynamism and growth we have witnessed in the two treatments and their rapid transformation into vibrant, albeit very different, types of forest. CUT OR GIRDLE 142

Adelgid-infested hemlocks die slowly, and the dead trees gradually fall apart. Although in principle we could have infested two experimental plots with the adelgid, the combination of the even more ethically challenging introduction of harmful species without a means to contain them, and our goal of contrasting the physical effects of hemlock decline with additional effects due to the adelgid, led us to the approach of girdling. In each of the two experimental plots we girdled every individual hemlock from the smallest seedling to the largest tree. For the trees we used chainsaws to make two circular cuts a few inches apart and each about two inches deep. Smaller saplings and seedlings were girdled with a knife. Shortly after girdling, sap flow was reduced by over 50 percent, but this did not immediately kill the trees. Rather, they struggled on for several months; some even lasted a few years, drawing sustenance from the sugars and starches stored in their leaves and branches, and obtaining water from the few vessels that escaped the girdling. But eventually they succumbed; over 80 percent of the trees died within two years, and the remainder had lost all their needles within three years. Although this rate of death was faster than the five to twenty years it has taken hemlocks to die from adelgid infestation across New England, it is comparable to the rate at which hemlocks have been succumbing to adelgid infestation under the warmer climate in Virginia and the Carolinas. At the same time that branches from the girdled trees were falling down, the boles were also susceptible to snapping where the girdling cuts were made. Because of these hazards, we always wear hard hats when working in the plots; in comparable adelgid-infested forests, authorities close off trails because of safety concerns from falling tree limbs. The specifics of the logging treatment were informed by our studies of hemlock harvests in Connecticut and Massachusetts, which removed 40 to 95 percent of the overstory hemlock basal area, along with a smaller percentage of other species. Typically, the slash of branches and leaves is left behind. A landowner’s goal in such preemptive salvage logging is often to capture the forest’s economic value before the adelgid arrives and renders the trees economically worthless. Because hemlock is not an especially valuable timber tree, harvests usually also remove white pines or oaks that have higher value as sawlogs, along with other hardwoods to yield cordwood for woodstoves. Loggers usually also cut smaller trees to ease access to the site by workers and their heavy equipment and to improve future stand quality. This extensive disturbance generally initiates a large cohort of seedlings and stump sprouts and the growth of a new forest. Although many might see this as close to a clear-cut, it is not. Rather, the best-formed oaks, pines, and other trees are retained with an eye to the future—to provide a seed source for these commercially valuable species, and to allow these trees to increase in size and value. Because such operations use large four-wheeled drive skidders or other machines to remove the logs, the soil is CUT OR GIRDLE 143

Complete mortality of a hemlock forest following girdling. (David Foster)

The bases of collapsed hemlock trees on the girdled plot. (David Foster)

Harvard Forest Woods Crew member John Wisnewski cuts logs hitched to a rubber-tired skidder. (David Foster)

scraped, scarified, and otherwise disturbed; smaller trees can be damaged; and slash is left scattered on the site. The seasoned loggers of the Harvard Forest Woods Crew harvested the two experimental plots in early 2005, using standard equipment—chainsaws and a medium-sized rubber-tired skidder—to haul the logs out of the woods. The winter operation used frozen trails to ease equipment access and reduce disturbance to the CUT OR GIRDLE 146

soil. In both plots, about two-thirds of the total amount of wood was removed, including all hemlock larger than eight inches in diameter, much of the low-quality maple and birch, and half of the white pine, red oak, and white oak. The remaining good-quality oaks and pines were left in place. The total harvest in these two two-acre patches yielded forty-seven cords of firewood, 39,000 board feet (equivalent to about seventy-eight cords) of logs for lumber, and 113 tons (equivalent to about forty-five cords) of hemlock pulpwood. About one-third of the logs and the hemlock for pulp were sold to saw and pulp mills in the region. The remaining white pine (8,000 board feet) and hemlock (6,000 board feet) logs were trucked to the Harvard Forest sawmill, where they were turned into lumber for the construction of our new garage. The firewood and waste from milling was used for heating Shaler Hall, which consumes between fifty and seventy-five cords every winter. In total, the operation netted (products sold less trucking and transport costs) just under $8,000, not counting the approximately 400 man-hours of labor and fuel costs. Roughly speaking, we broke even setting up this experiment, kept warm that winter burning the cordwood, and got a good start on our garage in the bargain. Ecologists are sometimes accused of quantifying the obvious, and many of the early responses of the forest to our treatments were consistent with what we expected and not especially surprising. But that really is the nature of science— insights rarely result from a series of eureka moments. Instead, the data accumulate, analyses are conducted, prior results are confirmed or modified, new findings build incrementally on the old, and the occasional surprises lead to new directions for study. Cut or Girdle has provided us with excellent data in a wide range of disciplines documenting the effects of hemlock’s removal from an ecosystem. The fast death of the hemlock trees from logging brought immediate changes. Opening up the canopy meant warmer days and cooler nights; without the moderating effect of a hemlock cover, winter was measurably colder and summers correspondingly warmer. These temperature differences are particularly striking on a hot summer day when walking from a heavily forested control plot out into an open and sunlit logged plot. The soil environment underwent major changes as well. Exposed to direct sunlight, the surface dried readily and fluctuated greatly in temperature. A few inches down, however, the soils remained wetter, as there were fewer trees sucking up water and transpiring it back into the atmosphere. Under these conditions of full sunlight, and with abundant moisture and nutrients, the vegetation rebounded rapidly. It took little time for the number of plant species to double, with many of these new species—ferns, sedges, brambles—appearing because they disperse easily to open CUT OR GIRDLE 147

The “green” Harvard Forest maintenance garage, constructed from hemlock logged at the Simes tract and sawn at the Forest. The garage is equipped with a solar array, wood furnace, and composting toilet. (David Foster)

sites and thrive in high light. The understory cover of these plants increased rapidly, quickly shading the soil exposed by the skidders. For the first five years after logging, we returned from the cut plots with many scratches as we waded through blackberries and raspberries to make our measurements. By 2010 the rapidly growing birches had overtopped, shaded, and killed the brambles, making our fieldwork much easier. The litter of falling leaves and small twigs transfers carbon, nitrogen, and other elements from the trees onto the forest floor. Once on the ground, this litter is broken up and shredded by small-to-microscopic animals; as bacteria and fungi complete this decomposition process, the carbon and nutrients stored within are released to the environment. We compared the litter falling in the different treatments by drying and weighing the leaves, twigs, acorns, and seeds falling from trees and captured in low-tech litter traps—laundry baskets lined with insect-screening mesh. Notably, because most of the large trees had been removed, litterfall in the logged treatment was dramatically reduced right after logging. Even with vigorous tree regeneration, litterfall in the logged plots in 2009 was only half of that observed before we had cut the plots. The flux of carbon dioxide out of the soil, which we call “soil respiration,” is a combination of exhalation (“respiration”) by microbes and tiny animals as they break down litterfall and dying roots, as well as the loss of carbon as plant roots respire. An immediate reduction in soil respiration was measured in the logged plots after the trees were killed, their leaf and branch litter were lost, and their roots died. Nonetheless, respiration levels recovered to precut levels within five years, likely due to the rapid establishment of understory plants, especially rapidly growing brambles and birch saplings. Although we did see short-term pulses in nitrogen availability a few years after logging and girdling, we did not observe any major changes in nitrogen cycling among treatments over the first five years. As expected, change was more gradual in the girdled plots. The trees’ continued presence, even though they were dying, kept the plots in partial shade and moderated the shift in conditions and most processes. Soil temperatures fluctuated less and displayed fewer extremes than in the logged areas. Belowground, the reduction in soil respiration that quickly followed the logged hemlocks did not reach its nadir in the girdled plots until all the trees were dead in a couple of years. In the understory of the girdled plots, black birch and red maple established thick carpets of seedlings, and many other herbs filled in the gaps. Understory cover of these plants increased slowly, paralleling the pace of hemlock death. Although we had worried that nonnative plants would invade the open forests, only two species—Japanese barberry and Oriental bittersweet—have colonized any of our plots so far and only the girdled ones. CUT OR GIRDLE 149

Leaf litterfall was remarkably high in the girdled plots. Within one year of girdling, huge amounts of hemlock needles, representing 80 percent of all litter, were shed by the dying trees. As birch, oak, and pine leaves gradually replaced hemlock in the girdled treatment, litter increased again, and by 2009 it was about the same amount as it had been before we girdled the hemlocks. One major result that we expected was realized. The death of hemlock in this experiment, whether fast or slow, resulted in an increase in species diversity, and not just of plants. The importance of hemlock extends up, down, and through the food web of other organisms that live in the forest and depend on hemlock for food and shelter. As hemlock declines, some animals disappear while others move in. We may notice this first with large animals. Early successional woodlands with dog-hair thickets of birch are highly attractive to moose and deer, both of which feed heavily on hardwood seedlings and stump sprouts. The importance of moose has become especially apparent, as this species has been steadily expanding its range from Maine and northern New England into central Massachusetts; in 2011 we installed fencing to keep moose and deer out of part of our plots, so that we could learn more about the interactions between these large herbivores and the tiny adelgid. Populations of small animals change too, and their small size and short life spans make them especially susceptible to changes in the larger environment. Observing the adelgid-infested and logged stands across central Connecticut described earlier, we saw that several groups of ants—notably large ants, including wood processors such as carpenter ants (genus Camponotus), large omnivores, scavengers, and slavemakers in the widespread genus Formica, and medium-sized soil-dwelling fuzzy ants (genus Lasius)—were uncommon or absent from mature, intact hemlock stands but rapidly moved into areas where hemlock was killed by the adelgid or logged out by landowners. We observed the same changes in this experiment and also discovered that the increase in the kinds and numbers of ants in the girdled and logged treatment plots led to a parallel increase in their predators: tiger beetles and hunting spiders. Although it may unsettle conservation biologists who simultaneously revere hemlock and are also single-mindedly focused on enhancing biodiversity, we have not been surprised that the loss of this foundation tree species, which so effectively dominates its environment and excludes many species, led to a substantial increase in biodiversity. In short, what we have seen over nearly ten years of comparing our experimental plots to the controls is that the changes in microclimate and forest-tree cover, induced either by girdling or logging, have had major effects on forest dynamics. Changes that began with vegetation composition then cascaded through the cycling of carbon and throughout the forest food web. Despite the changes initiated by the loss of hemlock, plots in both treatments had converged with those in the control CUT OR GIRDLE 150

plots in ecosystem function within five years, due to the rapid regeneration and resilience following disturbance. This pattern of speedy recovery is a hallmark of these temperate forests that has been seen in many other manipulative studies at the Harvard Forest and elsewhere. A central component of the design of this experiment was the expectation that the hemlock woolly adelgid would arrive eventually and infest our sites. This prediction came true, and several years earlier than we had anticipated. The adelgid was first observed at the Simes tract at low densities in 2006. By early fall 2009 all plots were infested, but with quite variable numbers. By spring 2013, dense aggregations of adelgid could be found on most hemlock branches. As a result of the arrival of the adelgid, our hemlock control plots have now been transformed into a new treatment: adelgid-infested plots. The first six years of these plots provided important information on variability within these plots in an uninfested condition. Going forward, however, we will compare forest structure and function in hemlock control plots before versus after the adelgid arrived to document the insect’s large impacts on forest structure and dynamics. Although we lament the adelgid’s arrival, we now can monitor responses following the precise time of infestation. This important detail is usually not known when we examine other infested plots across the landscape. At the same time, we will compare forest recovery in the girdled plots to recovery in the adelgid-infested hemlock “control” plots so that we can separate the effects of physical death of hemlock by itself from the additional impacts that the adelgid may have on forest dynamics. Based on Bernhard Stadler’s work on adelgids and nutrient cycling, we expect to see that the adelgid will rapidly move more nitrogen from the dying hemlock canopy to the forest floor. This excess nitrogen could cause some plants on the forest floor to grow better and others to die, and these effects will move up the food web, altering types of animals and their relative abundance in the forest. In the first two years of this new phase, the adelgid infestation has escalated. In 2009 we had to search carefully to find signs of adelgids. After the mild winter of 2011–12, the white woolly evidence of these tiny insects was obvious on most of the hemlock seedlings, saplings, and twigs fallen from canopy trees in all the treatment areas. The plots on the ridge are more heavily infested, so the two plots in the hemlock control treatment are likely to offer variations on the pace of deterioration. The hemlock trees themselves have not yet visibly declined, and we are curious to discover which of our harbingers of change will respond first: light, litterfall, tree growth, or nutrient cycling? Meanwhile, the trends we saw in the first years of the experiment continue. By 2011, the black birch seedlings in the girdled plots had grown into saplings topping CUT OR GIRDLE 151

out well above our heads. These plots look remarkably similar to forests decimated by the adelgid in southern New England. Dead hemlocks are dropping larger limbs and many stems are now falling, so walking through these plots is akin to climbing through a jungle gym. In the logged plots, the awkward brambly stage has passed as the birch saplings have grown, overtopped the raspberries and blackberries, and now shade the forest floor. Over the next decade we will discover whether the girdled and logging treatment plots continue to converge in structure and function, or whether they will take off on independent pathways. These comparisons will help us refine management recommendations to landowners facing loss of their hemlock forests. Cut or Girdle represents a large commitment of resources and time. How does this experimental approach clarify, corroborate, or contradict our long-term studies of the region’s “natural experiments,” both past and present? The vegetation response matches what we have seen in New England forests that have lost hemlock to the adelgid or to the logger’s chainsaw, but the experimental approach also allows us to construct a much more detailed story of how the flora fails or succeeds in passing from seed to seedling, from seedling to sapling, and from sapling to tree. Similarly, the coordinated observations of light, nutrients, flora, and fauna enable us to infer stronger connections between the forest structure and function. Finally, the experiment has become a hub of energy and scientific creativity. The boldness of this large, replicated real-world experiment has had the power to attract a range of additional researchers who bring their diverse perspectives to the study. Like other long-term experiments and observations, the value of this hemlock removal experiment will continue to grow. As the adelgid moves in, eats its way through the forest, and then continues its relentless move north, we will continue to explore the dynamics of these ever-changing forests.

CUT OR GIRDLE 152

NINE

MODELING THE DYNAMICS OF A FOREST GIANT One day that Tom Walker had been to a distant part of the neighborhood, he took what he considered a short cut homeward, through the swamp. Like most short cuts, it was an ill chosen route. The swamp was thickly grown with great gloomy pines and hemlocks, some of them ninety feet high; which made it dark at noonday, and a retreat for all the owls of the neighborhood. It was full of pits and quagmires, partly covered with weeds and mosses, where the green surface often betrayed the traveller into a gulf of black smothering mud: there were also dark and stagnant pools, the abodes of the tadpole, the bull-frog, and the water snake, where trunks of pines and hemlocks lay half-drowned, half-rotting, looking like alligators sleeping in the mire. —Washington Irving, The Devil and Tom Walker

I

n 1997 University of New Hampshire professor and longtime Harvard Forest collaborator John Aber wrote an editorial in the Bulletin of the Ecological Society of America with the title “Why Don’t We Believe the Models?” He lamented that insights gleaned from ecological models rarely interested his colleagues, who were mostly field researchers. When he initiated discussions “from a modeling perspective,” they often deteriorated into what he termed “the glazed-model-gaze.” The ambivalence to and distrust of models differed from attitudes in most other scientific fields, where quantitative model predictions and verification of those predictions are central to the whole research endeavor. The crux of the problem, Aber surmised, was that ecological modeling hadn’t been held to high enough standards. There was an impression among traditional ecologists that modelers could manufacture any output they wanted. Models were a “black box,” which data went into and results emerged after the model was tweaked in various ways. He suggested that if a consistent set of standards dictated how ecological models were presented, so that a model’s structure, along with its inputs and outputs, were clearly documented,

153

modeling would fast become a serious tool. Then ecologists would use models as a regular part of their work, just like statistics, economics, or climatology. It’s been sixteen years since Aber’s lament, and the chasm that separates modelers from other types of ecologists has indeed narrowed considerably. This is not to say that all field ecologists are now modelers, or vice versa. In fact, those of us who pursue modeling can safely say that, ever since we began seriously using simulation models to address ecological questions, our field boots have seen less and less action. In our discussions, however, we modelers rarely receive the glazed-model-gaze. Instead, most field ecologists now either dabble in modeling themselves or collaborate with full-time modelers. This arrangement has many rewards. By “modeling,” we are referring to computer simulations of ecological processes that are represented and linked together using mathematical and statistical functions. Within this broad framework, models take countless forms and operate over wide scales of time and space. In all cases, the process of developing a model causes us to formalize what we think we have learned from fieldwork—how an ecosystem functions, for instance, and how its component parts interact. In turn, models enable us to articulate and generate new hypotheses about the way ecosystems operate, to synthesize disparate studies, and to make predictions that extrapolate outside of the study to systems around us and into the future. All of these aspects can be tested further in the field. When done well, ecological modeling draws from observations collected in the field over lengthy periods. Models need to be calibrated, which is the process of programming the proper equations and variables—also called parameters—based on extensive field studies of various processes. Then models must be validated to evaluate their predictions against data sets that were collected independently from the calibration data. It follows, then, that many of the pioneering ecological models were developed as part of the International Biological Programme and its successor, the National Science Foundation’s Long Term Ecological Research program, in which Aber and Harvard Forest researchers participate actively. These efforts represent some of the largest investments in sustained ecological data collection and synthesis. Indeed, we would argue that the rise of modeling in ecology has been one of the hallmark successes of both programs. One such example is the JABOWA forest model, its name derived from the first two letters of the surnames of the three individuals who developed it: James Janak, Daniel Botkin, and James Wallis, who participated in the Hubbard Brook Ecosystem Study in central New Hampshire. JABOWA gave rise to a host of other forest models that are used to simulate the establishment, growth, and successional dynamics of individual trees in mixed species stands. These are called gap models, and they compute interactions among trees as they compete for light and other resources. The models have been used with MODELING THE DYNAMICS OF A FOREST GIANT 154

View from atop the Hemlock Tower to the fire tower on Prospect Hill, with intervening slopes of hardwoods, white pine, and hemlock. (David Foster)

great success to explore changes in the structure and composition of forests resulting both from climate change and from disturbances such as hurricanes and timber harvests. Harvard researchers have developed a forest model (whose lineage can be traced back to JABOWA) that has been used to help anticipate the effects of adelgidinduced mortality of hemlocks on the forest carbon cycle in eastern North America. The absorption of carbon dioxide out of the atmosphere and into the terrestrial ecosystem is one of the most important services that forests provide. When free in the atmosphere, carbon dioxide traps solar energy, increasing the earth’s average temperature. This process is termed the greenhouse effect, and without it the planet would be a chilly −4°F. But it is always possible to have too much of a good thing. Since the Industrial Revolution we’ve burned so many fossil fuels (coal and oil are, in fact, fossil forests) and released so much carbon dioxide back into the atmosphere that the greenhouse effect is now warming the planet more than we’d like. Lowering the concentration of carbon dioxide is central to the effort to slow the pace of warming. Forests can help us do that. Forests sequester and store massive amounts of carbon dioxide. Globally, forests are second only to the oceans as storehouses of carbon. In New England, where our relatively young forests are still recovering from the seventeenth- to nineteenthcentury agrarian past, along with heavy logging into the early twentieth century, forests continue to sock away carbon at a rate of about one ton per acre per year. To put these numbers into context, consider that forests in Massachusetts alone absorb enough carbon dioxide annually to offset the emissions released by one million U.S. homes. Of course, trees release their stored carbon back to the atmosphere when they die and decompose. Decomposition is effectively photosynthesis in reverse. So when an invasive insect such as the hemlock woolly adelgid threatens to kill off a major tree species, it can disrupt the carbon cycle. Given what we know about hemlock’s future in light of the adelgid, it is worth asking what impact we can expect the adelgid to have on the region’s long-term forest carbon dynamics. The question evokes scales of space (the eastern United States from Canada to the southern Appalachians) and time (decades) that are simply too large to address through field studies. What’s more, the adelgid has been killing trees in large numbers only since the 1980s, so we have no conclusive evidence regarding long-term future changes in the region’s forests, let alone the carbon implications. It is therefore a challenge that requires a simulation approach. Recently we teamed up with other Harvard ecologists, led by postdoctoral fellow Marco Albani and professor Paul Moorcroft, to use the Ecosystem Demography model (or ED, as it is called by everyone in the business) to simulate future changes in forest carbon across eastern North America with and without the effects of the adelgid. Though MODELING THE DYNAMICS OF A FOREST GIANT 156

View from the walk-up tower into the interior of the Prospect Hill hemlock forest. (David Foster)

the connection is not immediately apparent, ED is a distant relative to the forest gap models designed in the tradition of JABOWA. Ecological modeling is simply a way to formalize what we think we know about a complex system. Calibrating a model’s parameters involves integrating what we know—in terms of relevant rules and relationships—into computer code. It is the hardest part of the modeling process. Indeed, to talk to a forest modeler is mostly to hear complaints about “model parameterization.” ED was originally developed for the Amazon forest, so the processes of creating an eastern North American variant of ED required a completely new set of parameters, informed in part by the long-term data sets collected at Harvard Forest. To investigate the fate of hemlock, Albani and his colleagues also needed to develop a method to simulate the pattern of the adelgid spread. We know that most adelgid crawlers don’t disperse. We also know that a tiny fraction of crawlers spread much farther by wind, birds, and other means, and that these are the factors that dictate regional spread rates. But no one knows exactly what fraction of the adelgid population spreads long distances or exactly how far they might go. So our team built another model—a submodel that plugged into ED—to simulate adelgid spread. Fortunately, the U.S. Forest Service maintains a database that includes the year in which the adelgid was first observed in each county in the United States, and we were able to use this data to estimate the probability of infestation as a function of distance from the closest infested county. It was a simple approach, but one that produced a reasonable map of future adelgid infestation dates across our region. Albani simulated 100 years of forest and carbon dynamics beginning in the year 2000 using ED, first establishing a baseline of conditions by excluding the effects of the adelgid. Then he conducted a second round of simulations that introduced adelgid infestations in the time periods and locations that the spread model had estimated. Within the model, hemlocks succumb within a few years of the adelgid’s arrival, and the trees’ resources (light, water, and nutrients) are reallocated to other trees that might replace them, such as pines and hardwoods including black birch, red maple, and oaks. Within the simulations, the effects of the adelgid on carbon dynamics depended most strongly on how abundant hemlock was locally. For example, in the mid-Atlantic states hemlock constitutes only a small fraction of the total tree biomass, so the effect of the adelgid on forest carbon dynamics was expected to be trivial. In contrast, hemlock comprises as much as one-third of the tree biomass in central New Hampshire, so the adelgid is predicted to have a much larger impact there. Midway through the simulated century, the adelgid arrived in New Hampshire, and the modeled forests released a quick pulse of carbon dioxide as the hemlocks died and decomposed. Soon thereafter, carbon storage levels jumped as white pine and fast-growing hardwoods took over the sites once occupied by hemlock. MODELING THE DYNAMICS OF A FOREST GIANT 158

Integrated over the entire region, the difference in the forest ecosystem’s productivity (the carbon stored minus the carbon emitted) between the simulations with and without the adelgid yields our best estimate of the aggregate effects of the adelgid on forest carbon storage. For the first forty years of the simulation, ED predicted an 8 percent reduction in the uptake of carbon from eastern U.S. forests as a result of adelgid-caused mortality. Then, around the year 2040, the adelgid’s effect on carbon uptake turned from negative to positive as rapidly growing pines and hardwoods begin to dominate sites on which they replaced hemlock. For the following sixty years, ED predicted a 12 percent increase over the original conditions with hemlock as the forests recovered. This notable increase in the rate of carbon uptake is consistent with the slower rate of growth of hemlock compared to its major competitors. In addition to its carbon implications, there are several reasons that understanding the geography of adelgid range expansion and hemlock decline is a pressing concern. The most obvious is that forest managers would like to halt or slow the adelgid’s spread. And even if they can’t stop it, managers, large landowners, and conservationists would like to know if and when the adelgid might arrive, so that they can prepare in whatever way is possible. In addition, scientists are desperately trying to understand the range dynamics of all kinds of species, of which the adelgid is but one, albeit a very significant, example. This urgency stems from our ushering of countless species into new environments as by-products of trade, travel, and transportation. As we unwittingly help to scatter species across the globe, there is a vital need to know which are likely to become invasive and exactly where they will invade. To complicate matters, climate change is expanding the range of many native species, while contracting the range of others. Taken together, the forces of global change are promoting a wholesale rewriting of thousands of species’ range maps that once seemed relatively static. An interesting side effect has been a renaissance in the science of biogeography—the study of the distribution of species in space and through time—driven largely by the need for better approaches to species distribution modeling. Because the spread of the adelgid has been documented for more than sixty years, and because its host tree is so narrowly defined, it serves as an ideal study system and opportunity to advance the science of species distribution modeling. The classic approach relies on calculating correlations between species presence or abundance and a suite of abiotic factors such as climate, topography, and soil. Once these relationships are known, a modeler can map the potential species distributions and even update the range maps sequentially into the future based on climate change scenarios. This approach is popular for its ready answers, but it ignores the biology of species, including the central role of population and dispersal dynamics. MODELING THE DYNAMICS OF A FOREST GIANT 159

Range maps developed through this process serve primarily to describe a species’ potential habitat but cannot tell us the probability or timing of establishment in a specific location. In 2009 postdoctoral fellow Matthew Fitzpatrick arrived at the Harvard Forest as part of a project seeking to develop a better approach to species distribution modeling. He had worked extensively with the classic correlative approaches to mapping species ranges and hoped to develop an improved framework for species distribution modeling using the hemlock woolly adelgid as a study system. His approach would incorporate dispersal over time and population cycles, as well as habitat suitability and climatic controls. Fitzpatrick worked with our existing data to develop a model that incorporates the adelgid’s unique characteristics of dispersal and population growth. (As discussed, the real dynamics of adelgid spread are governed by long-distance dispersal events, which are rare and difficult to predict. Because the adelgid reproduces asexually, even a single individual colonizing a single hemlock needle can launch a new infestation. These characteristics can make the adelgid subject to sudden and dramatic range expansions.) The model includes an explicit representation of hemlock abundance throughout eastern North America and takes into account the lethal effects of cold weather on the insect. Each run of the model portrays the period from the 1951 discovery of the adelgid in Richmond, Virginia, up to 2008. After introducing a small number of adelgids into the Richmond area, in each sequential year the model simulates the development stages of the adelgid, including births, deaths, and dispersal. When the adelgid infests a new stand, the population grows in proportion to the abundance of hemlocks found there. This pattern continues until the hemlocks have all succumbed. If an individual adelgid crawler is to lead a charge into a distant forest, it must beat the one-in-100-million odds of being selected as a long-distance disperser, and it must also land in a location occupied by hemlock. For any adelgid to survive and reproduce another year, the winter temperature must not fall below a critical threshold (-13°F). Each of these processes occurs simultaneously all over eastern North America as gridded into one-square-kilometer cells. Each event in the model—establishment, mortality, local and long-distance dispersal—is drawn from a probability distribution (just like rolling a many-sided die), which was built using the best available empirical data. Given that a simulation includes millions of draws, each run of the model yields a different manifestation of modeled adelgid range dynamics. To interpret the results, Fitzpatrick ran the model 1,000 times and reported the average year that the adelgid arrived in each county within hemlock’s range. Qualitatively, the model predictions matched many attributes of the observed spread of the adelgid, showing the most rapid spread to the south and lower rates of

MODELING THE DYNAMICS OF A FOREST GIANT 160

spread elsewhere. Accordingly, the Southern Appalachians, with their warm winters and abundant hemlock, were infested quickly. The model also correctly predicted a slow spread across New England, highlighting the importance of winter temperatures in regulating the rate of spread. Indeed, the model suggests that variation in winter temperatures exerts a stronger control over adelgid spread than do the distribution and abundance of hemlock. The simulations also generally match observations regarding the current extent of the adelgid’s spread. Aside from some forests in northwestern Pennsylvania and southern New York, it appears as if the adelgid has reached the extent of its potential range and is circumscribed to the north by lethal winter cold snaps. Of course, given climate change, the range will likely extend farther north over time. We modelers often tell ourselves that we learn more when our models fail than when they succeed. While this serves to heal bruised egos, it also happens to be true. Fitzpatrick’s model failed to predict several key features of the spread of the adelgid, and this was arguably the most interesting part of his research. For example, the model did a poor job of predicting the timing of adelgid arrival into any specific area. This largely reflects the inherent difficulty in predicting rare long-distance dispersal events. In particular, the direction of long-distance dispersal appears not to be random, as the model’s parameters had specified. In general, the modeled dispersal was later than observed into the north and earlier than observed into the south. Put another way, the observed spread of the adelgid suggests that there is a greater longdistance dispersal of the adelgid to the north. This failure of the model forces the question: what unconsidered factor might lead to greater spread to the north? One potential answer is birds. Dispersing adelgids hatch during the spring, exactly the time when birds are migrating through the region and moving largely northward. Unfortunately, we don’t have ready data on the activity and effect of birds to plug into the model; nonetheless, the model results point toward an interesting factor that deserves more research. Another aspect to remember, and one that Fitzpatrick stresses, is that the observed pattern of spread across the Northeast is but one realization of many possible patterns. Were we to somehow turn the clock back to 1951 and do it over again, the dispersal pattern would be different, as it would be affected by a different set of unknowable events, such as a freak cold snap, or an individual bird or wind gust that happened to carry a crawler to a particularly faraway hemlock. By estimating the pattern of spread based on the average of a thousand simulations, we can reduce the influence of such idiosyncratic events. But averaging simulation results is only good from the perspective of reducing model uncertainty; it does nothing to reduce the influence of uncommon but highly consequential events in the real world.

MODELING THE DYNAMICS OF A FOREST GIANT 161

Fitzpatrick’s simulations show the myriad ways that the adelgid infestation could have played out. Seeing these, one wonders how things might have been different. What if the insect had been eradicated early, before it spread outside any plausible scenario for containment? Indeed, resource managers ask this type of “what if ” question every time a new pest is discovered. Consider that, in 2010, officials in Boston were alerted to the discovery of Asian long-horned beetles—one of the most potentially destructive nonnative insects in the United States—in a hospital parking lot directly across from Harvard University’s Arnold Arboretum. Thanks to fortuitous early detection and quick action, only six trees had to be felled and chipped. Thousands of nearby trees have since been examined, but to this day only those six ever showed evidence of the beetle. What if the agency had not responded with such intensity? One only needs to look west from Boston to the second largest city in New England, Worcester, where Asian long-horned beetles were discovered in 2008. In this case, the insect had already been in the city for at least a decade. Because the infestation was not detected early, officials have needed to examine over two million trees, 30,000 of which showed evidence of the insect and were cut down and chipped. In the effort to contain the beetle’s spread, many woodlots, suburban neighborhoods, and tree-lined streets in Worcester have been decimated. No doubt USDA officials had the adelgid spread pattern in mind as they commissioned TV and billboard advertisements encouraging citizens to report any sightings of the beetle, and as they imposed a 110-square-mile quarantine zone restricting the transport of untreated wood material. It is difficult to know whether this response will ultimately be effective in containing the spread of this invasive insect. The Future Scenarios project led by Kathy Lambert, Jonathan Thompson, and David Foster at the Harvard Forest is a nascent research initiative designed to help researchers and resource managers think through a myriad of “what ifs,” structured as alternative plausible scenarios of land use, insect infestation, and climate change. The scenarios are articulated by groups of informed stakeholders including policy makers, land planners, and conservation professionals. These narrative scenarios are then used to create landscape simulations of ecosystem change and the societal responses to these changes. Together the scenarios and simulations describe a range of plausible futures, without any pretense of actually predicting the future. Instead, comparisons among a set of scenarios can be used to prepare for a range of possibilities and to understand important areas of uncertainty. One scenario project has focused on exploring a range of land use futures for the forests in Massachusetts as defined by a group of ten stakeholders, mostly representing state agencies and environmental organizations. For example, one scenario envisions a future where the price of oil soars, and citizens of the commonwealth must look to their forests as a MODELING THE DYNAMICS OF A FOREST GIANT 162

source of biomass energy and as potential agricultural land. The simulations have helped the group understand the capacity of the forest to meet these needs, and the costs and benefits of doing so in terms of biodiversity and other ecosystem services. The group is now working toward a range of insect scenarios that will explore the potential impacts of and responses to the adelgid and other less-established invaders. Although the studies described in this chapter all focus on simulating the future or recent past, there is also great potential to use these models to help untangle lingering questions from the paleoecological record. There is ample precedent for this approach. Almost thirty years ago, Margaret Davis and Daniel Botkin joined forces to explore assumptions regarding tree migration in response to climate change, as inferred from pollen records. They used the JABOWA model to simulate the impacts of an abrupt four-degree Fahrenheit decrease in average temperature on forest composition in New England’s northern hardwood forests. In their simulations, it took as long as two centuries after the change in climate before cold-hardy spruce replaced sugar maple. The delay in species turnover reflects the life span and life history characteristics of the dominant tree species in question. Davis and Botkin showed that fossil pollen deposits, even if they faithfully record the composition and abundance of forest communities, may miss brief climatic events altogether, and that changes in the pollen record may lag behind the onset of more sustained climatic changes. This and other insights from their experiment changed how paleoecologists have interpreted their cores ever since. Of course, simulation models and our understanding of paleoecology have come a long way in the past three decades. Nonetheless, the potential for applying forest models to the paleo record remains as large as ever. One scientist heeding this call is friend of the Harvard Forest Steve Jackson, professor at the University of Wyoming. He and his students have started using dynamic global vegetation models (DGVMs) to explore the reciprocal relationships between climate, the atmosphere, and vegetation, all inferred from the expanding network of lake cores from across the United States. This class of simulation model has seen tremendous advances in the past decade (in fact, this has become more of a model synthesis than an individual model, but that is beyond this book’s scope of discussion). They were born out of a need to better link vegetation models to the global circulation models used by climate scientists to understand and forecast future climate change. The models used by Jackson are distinctive in that they explicitly couple terrestrial processes, such as photosynthesis, to atmospheric and physical processes, such as air currents and the reflectivity of the sun of the surface of the earth (known as albedo). They are the first class of model to quantify and exploit the reciprocal relationship between the atmosphere and biosphere. Once they are parameterized from pollen records, MODELING THE DYNAMICS OF A FOREST GIANT 163

these models are particularly suited for exploring and testing lingering hypotheses for the mid-Holocene hemlock decline. As noted throughout this book, hemlock’s story has unfurled at millennial scales of time and continental scales of geography. Accordingly, the arduous process of deciphering and telling hemlock’s story includes an important role for simulation modeling. These models enable us to efficiently explore the aggregate consequences of our assumptions as they play out over scales that are logistically impossible for field ecologists. Fortunately, ecologists of all stripes now embrace the “modeling perspective” that John Aber championed.

MODELING THE DYNAMICS OF A FOREST GIANT 164

TEN

REPRISE: EASTERN HEMLOCK AS A F O U N D AT I O N S P E C I E S I shall not here attempt further to define a foundation species, and perhaps I could never successfully do so. But I know one when I see one. —with apologies to Justice Potter Stewart

T

hroughout this book, we have referred to eastern hemlock as a foundation species—a species that creates its own ecosystem literally from the ground up; dominates the system in terms of numbers and mass; and is intimately linked to the majority of other species in the system. These characteristics of a foundation species lead us to perceive and recognize it as an inseparable part of the system—we know it when we see it—but for scientists there is an important difference between thinking (or hypothesizing) that something is so and demonstrating that it is indeed so. Strictly speaking, in fact, we do not demonstrate the truth of our idea or hypothesis of interest, such as “Eastern hemlock is a foundation species.” Rather, we aim to demonstrate that a so-called null hypothesis—for example, “Eastern hemlock is a species like any other”—is unlikely. If we can do that, then it is as Sherlock Holmes once said: “When you have eliminated the impossible, whatever remains, however improbable, must be the truth.” The historical, observational, experimental, and modeling work done by Harvard Forest researchers that has been focused on hemlock provides in aggregate one of the most thorough, detailed, and in-depth studies of a foundation species that has ever been conducted. Paleoecological reconstructions have shown that hemlockdominated systems throughout the region were dramatically reorganized following the first decline of hemlock, approximately 5,500 years ago. This reorganization differed in different places. In some forests, birches and oaks came to dominate,

165

whereas beech and sugar maple replaced it in others. When hemlock eventually recovered, however, the new hemlock forests everywhere appear to have regained many of the characteristics of a “hemlock forest.” This geographically widespread recovery is the first piece of evidence refuting the hypothesis that hemlock is a species just like any other. Long-term observations in southern and central New England also have revealed that hemlock forests have characteristics distinct from the broader range of features found in deciduous, mixed-conifer, and hardwood forests throughout the region. The evergreen, dark, cool, moist, cathedral-like understory environment of a hemlock forest differs from the understory environment of deciduous and mixed forests, regardless of the species mix in the latter. Furthermore, our experiments show that the loss of hemlock, whether from the adelgid or from preemptive salvage logging, results in immediate changes, both short-term and long-term, to the associated plant and animal species, along with shifts in storage and movement of energy, nutrients, and water through the ecosystem. The changes in the types and numbers of plants and animals as hemlock declines and then disappears can be quite dramatic and much more pronounced than when hardwood forests are thinned, logged, or undergoing succession—another nail in the coffin of the null hypothesis. On the other hand, shifts in carbon and nitrogen dynamics following hemlock decline are much more subtle and, at least in the few decades of “human time” over which we have observed these changes, provide less evidence for the idea that hemlock is a foundation species. The differences in water use between hemlocks and hardwoods have been solidly quantified by our studies on individual trees and measurements of atmospheric and hydrological flows from entire forests. Until we invent a time machine, models are our best way to shift our frame of reference from the decades of human time to the centuries of forest time. Our understanding of hemlock forests and our assertion that hemlock is a foundation species are tempered by the lessons learned from models of carbon storage in, and its movement into and out of, northeastern forests. These models suggest that one key service that forests provide—carbon storage—will return to prior rates within fifty years after hemlock disappears from our landscape. Beyond that, the rapidly growing hardwood forests will sequester carbon even more rapidly than the hemlock forests that they replaced. However, if we account for the carbon storage that never occurred because hemlock died, it actually will take more than a century for us to return to pre-adelgid conditions. By way of analogy, if you had $10,000 in the bank and you took it all out and then started putting $200 a year back into the bank, in fifty years you would be back where you started—at $10,000. But over the same time, you have lost fifty years of intervening interest on that initial $10,000, so it takes much longer than fifty years to recover your initial balance plus the forR E P R I S E : A F O U N D AT I O N S PE C I E S 166

gone interest. Think of hemlock as that initial balance. The potential carbon gain that could have occurred over the time it takes for the hardwood forest to recover that initial balance is interest forgone. So perhaps looking at carbon and nutrient dynamics for years or even a few decades provides insufficient evidence, either in favor of or opposed to, the hypothesis that hemlock is a foundation species. At the same time, we should also ask whether a forest built by a single foundation species always provides “more” or “better” ecosystem services than a forest of many species, none of them foundational. Those of us who live in old colonial houses are constantly bemoaning the faults of new construction as we watch, through our 200-year-old drafty single-pane windows, our neighbors’ thirty-year-old (or newer) houses fall rapidly into disrepair. Despite the lack of straight lines atop our wainscoting or level floorboards under our feet, we can convince ourselves that older is better. Certainly the rock-rubble foundations make for porous basements. Mice and chipmunks move through the chinks between the rocks and scamper around our walls in the autumn, the basement is too cold for anything except storing cabbage in the winter, water seeps in as the snow melts in spring, and the summer damp yields bumper crops of mushrooms on the ceiling. On the other hand, the walls do not crack in earthquakes, and twenty-firstcentury off-the-gridders would rather eat stored beets from their root cellars than fresh ones shipped to the grocery store from across the country. Similarly, we may implicitly assume that forests structured by foundation species must somehow be better than those that are not. But what does it mean to be a better forest? This question hit home in the summer of 2012 as we geared up for the annual sampling campaign in the hemlock removal experiment. On successive days, a graduate student (Ally Degrassi) and an undergraduate student (Yvan Delgado de la Flor Arana) working on the project each asked Aaron Ellison why there were so few other species to be seen in the hemlock forests. Ally and Yvan were new to the project: they had not been at the experimental site before, sampled along our New England transect, studied lake cores full of pollen, or run forest dynamic models. They both had read our scientific papers on hemlock forests and foundation species, however, and they came to the Harvard Forest and the hemlock removal experiment expecting something different from what they found. In particular, they both expected that, because we had asserted that hemlock forests were structured by a foundation species, they would be better forests than the nearby mixed deciduous woodlands. Their concept of “better” was that these forests would have more—and more different kinds—of the animals they planned to study: Ally’s small mammals and Yvan’s ants and spiders. Now that they were in the forest, however, they looked around and remembered some of the details of what they had read. There were fewer R E P R I S E : A F O U N D AT I O N S PE C I E S 167

kinds of ants, fewer kinds of small mammals, fewer kinds of amphibians, and fewer kinds of birds in hemlock forests than in mixed deciduous forests. Their recollections, together with the evidence in front of their eyes, led immediately to their next question: if hemlock forests are not as diverse as either the hardwood control forests in the same experimental block or as the young hardwood forests that replaced dying hemlock forests, why were we so worried about the loss of hemlock? Ecologists distinguish between foundation species and keystone species. Predators such as wolves are keystone species that sit atop the food web. They are not especially common, but they feed preferentially on dominant competitors, such as deer or elk, in the middle of the food web. By reducing populations of species in the web’s middle, weaker species (and more of them) can flourish. In other words, keystone species always enhance biodiversity from the top down. Foundation species, on the other hand, work from the bottom up. They can enhance, diminish, or have little effect on biodiversity. Their effects depend not only on their presence but also on all the interactions among all the other species in the system. When we walk through a forest stand, we see (and can measure) the biodiversity right around us. This local or within-site diversity is what Ally and Yvan were looking for. It is also the kind of biodiversity that a keystone species increases and that a foundation species may or may not enhance. But there is another kind of diversity— landscape diversity—that foundation species almost always enhance. This is the kind of diversity that you would see from the window of an airplane flying on a clear day. Flying over New England, we are accustomed to seeing a patchwork quilt of forest and fields, cities and towns, and a variety of houses clustered around town centers and scattered throughout the woodlands. This landscape diversity sets New England apart from, for example, the seemingly endless prairies of the Midwest (which now have been replaced over almost their entire extent by similarly endless fields of corn and soybeans) or the nearly unbroken urban corridor running along the Atlantic coast from New Haven, Connecticut, south to Washington, D.C. In a diverse forested landscape, we see a similar kind of patchiness. Over here are a few dozen acres of old beech and sugar maple, over there a patch of white pine, a mixed deciduous forest of young oaks and red maples, or a stand of dense hemlock. This patchiness, this landscape diversity, comes from a combination of the local diversity that we see within a site and the variability that occurs between the different sites. It is not as visible from the ground when we are standing within a single forest patch as it is from above when we can see all the different patches. But landscapelevel diversity is no less important than local diversity. Because foundation species define local ecosystems, their presence in some places and absence in others help create and maintain landscape diversity. But imagine the same landscape without hemlock forests. Looking down from the air, we R E P R I S E : A F O U N D AT I O N S PE C I E S 168

Beaver-cut hemlock adjacent a swamp near the Earl Stephens plot. (David Foster)

would see fewer different kinds of forest stands. If we were on the ground, we would be less likely to move from one kind of forest into another as we walked the same distance through the woods. The adelgid and salvage logging remove entire stands of eastern hemlock and replace them with hardwoods, resulting in forest patches that are more similar to one another and to existing mixed-hardwood stands on the landscape. Other irrupting insects such as the emerald ash borer, pathogens such as chestnut blight, and ongoing changes in land use as people move from cities to exurban areas also tend to homogenize the landscape. Soon we will be left with a landscape where one patch of forest look just like any other one. The end result of this biotic homogenization is a forested landscape consisting predominantly of what we only half-jokingly call New England’s “blah woods”—mixed stands of young red maple, birches, and oaks on cut-over, degrading soils. Most of the hemlock stands, and virtually all of the ones in which we have studied the effects of the adelgid and logging, are relatively young (100–200-year-old), second-growth stands. The few remaining old-growth hemlock stands—themselves the result of the hemlock resurgence after the first die-off 5,500 years ago—have notably different characteristics from second-growth stands. The loss of the majority of these old-growth stands following European settlement also reduced landscape diversity in ways that are difficult to measure. If we were able to go back to the 1500s or 1600s and study these old-growth stands as they developed, or if we could have protected them from the colonists’ axes, would we change our view of hemlock as a foundation species? This is a question we leave for the future. The model of adelgid dispersal discussed previously provides an important lesson related to such historical reconstructions, however. Matt Fitzpatrick’s model of adelgid dispersal predicted well the average pattern of its spread—quickly toward the south, more slowly toward the north—but failed to get the exact timing right. The model predicted that the adelgid should have gone south first, and quickly, and then turned its attention slowly northward. But the adelgid invasion actually unfolded with a first move north from its original site of introduction in Virginia, and it only later moved south. Although this was one possible outcome of Fitzpatrick’s model, such an outcome occurred only rarely in multiple model runs, and the average pattern looked substantially different. But reality happens only once, and it rarely hits the average. If a time machine enabled us to restart the pattern of adelgid spread, we would be unlikely to see the same path of dispersal, and maybe we would not yet be lamenting the loss of hemlock in New England. The measurements of hemlock’s change on the landscape, especially over the last 400 years; the results of our experiments on forest response to its removal; and our perception of it as a foundation species—all of these are conditioned by the unR E P R I S E : A F O U N D AT I O N S PE C I E S 170

folding of a particular set of historical occurrences. Some of these—long-term and seemingly abrupt climatic changes, hurricanes, ice storms, and outbreaks of native insects—we categorize as “natural,” whereas others—European colonization, logging and land clearing, stripping of bark for tannin, introduction of nonnative insects—we categorize as “anthropogenic.” But together, all of them are facets of the prism through which we now view the importance of hemlock to the landscape of northeastern North America. Now, in 2014, in light of the loss of chestnut in the early twentieth century, the ongoing decline of beech and sugar maple, the wholesale logging of white pine, white oak, red oak, and hickories from our forests throughout the last 400 years, and the extirpation of wolves and other keystone species from our region, hemlock appears to be a foundation species. It is not a species like any other. But despite the decades of observations, data, and models, we are not yet able to definitively decide whether, in the hemlock forests we revere, we are looking at the everlasting foundation of a restored colonial house, the ephemeral frost wall of new construction, or a battered and collapsing cellar hole returning to the earth from which it came.

R E P R I S E : A F O U N D AT I O N S PE C I E S 171

Lessons from Harvard Forests and Ecologists III. The Earl Stephens Plot It would appear that up to this time we have neglected to dissect our major subject of study, the forest. —E. P. Stephens, 1955

T

hese days, hikers tromp right by the spot on the heavily shaded path. With eyes drawn not to the dark depths of the old hemlock woods but toward the bright expanse of Harvard Pond just a few feet off the trail, most people miss the thirtyfive-foot-wide strip of paper birch and red maple that heads upslope through the woods. Only in the winter does this narrow stand jump out. Then, on sunny days when the snow catches in the canopy of the ancient hemlocks and leaves the ground beneath them a dull gray brown, the narrow band of young hardwoods lights up in brilliant white, broken only by the deep green of scattered hemlock saplings. This is the plot where Earl Stephens worked for nearly a decade to perfect Bob Marshall’s historical approach—what Stephens prosaically called “the historical reconstruction approach for determining forest trends” in his 1955 thesis. In the early 1950s, the 10- by 120-meter (33- by 394-foot) rectangle was a clear-cut gash in as old a forest as Earl’s advisor and Harvard Forest director Hugh Raup was willing to sacrifice in the name of science. The wooded slope perches just above the valley bottom where a stream once meandered through an open marsh known for centuries as “the Meadow Water” and which in its dammed state forms Harvard Pond. And it’s well below Petersham Common and the town’s best agricultural soils. Consequently, in contrast to the three-quarters of the region cleared for farmland, the long, rocky slope came through the colonial period as a woodlot or primary forest. From it and from other dwindling islands of trees, the landowners withdrew fuel, tan bark, small wood for fencing and tools, and logs sawn at Lincoln’s Mill at the valley’s south end, where the small stream drops abruptly toward the large Quabbin

172

Valley. In the old days the marsh surface of the Meadow Water had annually turned into a circus of men, oxen, wagons, and gathering families as dozens of landowners worked collectively to mow the native hay nourished by spring freshets. Then, as agriculture declined throughout the land and a larger dam was constructed, trees reclaimed every inch of dry land, the woods grew and matured, and one of the region’s loveliest scenes of forest and open water emerged. The peaceful quiet of this scene was destroyed in late 1948, when Stephens and Raup directed the Harvard Forest Woods Crew to fell and dismember the narrow rectangle of ancient woodland. Stephens then dissected the woods in a forensic effort to decipher each event and every spurt of growth in the forest’s history. Despite the countless lessons that had emerged from decades of Harvard Forest studies, the research group remained motivated by many questions that Stephens and Raup hoped to resolve in this woodlot vestige of the region’s original forest cover. How had the natural processes observed at Pisgah—wind, ice, insects, and fire—operated in this more protected site? What of the lengthy history of human activity? It was known that a long series of owners had used these woods in various ways over time, but what were their practices and how had they changed the makeup of the forest through the colonial period? And what about the mixture of hemlock, oak, pine, and birches that grew on these slopes and in many old woods? The trees were of various sizes, but did they differ in age? What characteristic of the forest history and these different tree species enabled them to coexist and each to thrive? Finally, Stephens was curious as to just how far back one could extend such a forest chronology. These lines of inquiry certainly built on the thinking and work initiated by Richard Fisher and his students at Pisgah, at the Adams Fay lot, and elsewhere. From the detailed field notes and correspondence that Hugh Raup left, it is clear that they began to emerge in his mind when he made his first forays to Petersham in the early 1930s in his position as professor and geographer at Harvard’s Arnold Arboretum. On each trip out from Cambridge, he stopped to visit and pick Fisher’s brain, sharing university news, collecting insights into the landscape’s history and vegetation, and seeking advice on sites to explore and botanize. Certain common themes and interests emerged from these excursions and exchanges. For example, while Raup roamed widely from the Slab City valley and the Simes tract to the top of Prospect Hill, the plot that he and Stephens selected to dissect lay just a half mile south of and along the same slope as Fisher’s favorite old-growth stand. A direct line connected those sites with the Bob Marshall plot northward up the Tom Swamp Valley. Raup raised many of the same questions that had long intrigued the first director. But he broadened the scope of inquiry with a geomorphological perspective on LESSONS: THE EARL STEPHENS PLOT 173

the landforms in Petersham that was strengthened from his years of work in the high Arctic and his close friendship with Harvard geologist (and Margaret Bryan Davis’s father) Kirk Bryan. Raup also could lend the insights that an accomplished botanist brings to understanding the relationships between plants and site conditions. In many ways, Raup brought an awareness of physical, biological, and cultural processes that would have been familiar to the original instigator of the Harvard Forest, Nathaniel Southgate Shaler. Like that broadly based scientist, who integrated the study of biophysical and social processes, Raup was keenly interested in deep time and the processes that shaped the soils and earth, as well as the plants upon them. Having framed a set of broad questions and selected a location that afforded a mixture of human and natural history, Stephens began his project much like Marshall did, with the demarcation of a permanent plot, many forest measurements, and a clear felling of the entire stand, this time by the next generation of the Woods Crew. Every tree and sapling was sawn as close to its base as possible. Then, as each tree bole was cut into cordwood sections at four-foot intervals, a thin disk of wood was sliced from the lower end and set aside for analysis the following winter. From hundreds of tapering stacks of these disks, each representing a single tree, would emerge records of the age and pattern of growth in diameter and height of every stem in the entire forest. Once the wood was hauled from the site, the brush was cut and burned, and dozens of boxes of labeled disks were carted back to Shaler Hall, Earl then set about disassembling the landscape itself. He first raked the entire surface of twigs, leaves, and decaying litter to expose the mineral soil, details of the microtopography, and the network of finely branching roots. Then he gridded the area with stakes and taut level lines of twine, from which he mapped the entire surface down to a six-inch contour. The mapping helped to pinpoint the location of downed logs, decaying stumps, and boulders with remarkable spatial precision. It also enabled him to characterize the many mounds and pits of earth formed by windthrown trees, which have persisted in our woods for centuries as barely recognizable lines of moldering moss-covered wood. Stephens aged every surface and feature in a relative way by judging its state of decay or erosion. This chronology was then improved by assigning a minimum date based on the age of the trees that grew upon it. If a decaying tree had sound wood, it was then sawn to determine the species, historical rate of growth, and age. Meanwhile, using a shovel and carting the rocks and soil from the area, he spent weeks bisecting mounds and adjoining pits in the plot and the surrounding woods with trenches four feet deep and five to twenty feet in length in order to expose the underlying soil horizons. With practice, care, and long hours in the trenches, he was able to map the complex cycles of upturn, overturn, burial, and erosion that had accompanied and followed the uprooting of the trees. Patterns began to emerge from LESSONS: THE EARL STEPHENS PLOT 174

A stilt-rooted hemlock that established a century ago on a stump, which is now largely decomposed. (David Foster)

this effort that helped him differentiate when a tree fell. Recent uproots were tall with an angular, sharp profile and distinct soil layers showing where the upturned soils had buried the former surface. In contrast, subtle low mounds marked uprooting events from ancient windstorms. From the ages of associated trees, stumps, and logs, Stephens could align this sequence with dates ranging from the 1938 hurricane and the chestnut blight in 1915 to ancient events that predated European settlement by many centuries. At the end of the initial two years of painstaking fieldwork, the site lay bare and peculiarly defiled. Every feature was numbered, named, probed, and dissected and each had been photographed, mapped, and then filed away as a print, sample, data point, or description that was available for further analysis. Gradually and methodically, Stephens revisited each bit of evidence in the lab and field; over the years the LESSONS: THE EARL STEPHENS PLOT 175

data and insights accumulated. New dates began appearing from the analyses, and the story line assumed a firmer chronological framework, based on the tree rings counted beneath a dissecting scope, property deeds ferreted out of the town hall and county courthouse, and the matching of field evidence with meteorological, historical, and agricultural records. The oldest hemlocks extended back more than two hundred years to the town’s founding in 1754. Their ages began to establish a minimum date for the stumps, mounds, pits, and logs that they grew upon. The very oldest materials came from the soil trenches. Over time, Stephens came to recognize and associate the distinctive shape and decay features of mound and pit sequences to a group of known hurricanes: 1938, 1815, 1788, 1635, and earlier. Extrapolating from this established series, he estimated that the most ancient features dated to the 1400s. He concluded that intense windstorms had periodically ravaged the New England landscape for thousands of years. Stephens’s idiosyncratic effort produced an unparalleled site-specific chronology of forest and landscape history, in which he developed a story line for every stem and each feature in what was clearly a dynamic forest. When he took stock of his immense accomplishments, he could envision a three-dimensional history of the old forest. In an era before the computer technology existed to display it, he produced a series of maps and sketches that captured the establishment and growth of every tree from a seedling stage into its established position in the canopy. He also chronicled the demise of many forest giants from wind and the ax. Stephens summarized in his thesis the techniques he had developed and integrated: The method is based upon the most conspicuous evidence in the forest: that furnished by the trees. Trees have the inherent capacity of providing directly and indirectly two general kinds of evidence that reveal forest trends: concrete and implied. Examples of concrete evidence are the trees themselves alive or dead; stumps; fragments of wood; charcoal; and mounds and pits resulting from uprooting. The implied evidence is founded upon known silvical characteristics and recognizable reactions of trees to the complex of environmental factors. Such evidence includes the general ability of hardwood trees to sprout from cut stumps, the influence of foliage-eating insect epidemics upon radial growth, and the occurrence of trees growing in perched positions above the general level of the forest floor. In later life Earl spoke about it matter-of-factly, but it was a monumental task that consumed him for nearly ten years. All the while he was supporting himself, and then a wife and children, as a junior scientist working on many other Forest projects. As might be expected from any great effort of science that took a decade

LESSONS: THE EARL STEPHENS PLOT 176

A large uproot mound at Pisgah, with sixty-year-old birches that are beginning to tilt as the mound erodes. (David Foster)

to unfold, it is the human elements in Stephens’s project that make it especially intriguing. Like the early Pisgah work and Bob Marshall’s thesis, Stephens’s project was a quintessential community effort that involved nearly everyone at the Harvard Forest. The study tapped senior scientists for inspiration, knowledge, and mentorship; fellow students for endless advice, second-guessing, and assistance in juggling the data and analyses; and the Woods Crew for all the heavy labor of sawing, hauling, digging, and contriving just the right tools for jobs that no one had ever imagined tackling before. We don’t know exactly how the project originated or when its comprehensive scope was established. But it is clear that, as the data accumulated and a focused picture of the site’s history and dynamics emerged, Hugh Raup became the study’s greatest advocate. Stephens had neither the appetite nor the podium from which to advance his work broadly, but Raup had both, as a senior professor at Harvard and an established voice in ecology and geography. As a result of Raup’s boosterism, Stephens’s study was widely heralded in forestry, ecology, and soils circles. It eventually became a metaphor for the Harvard Forest approach to ecology, wherein historical awareness lies at the heart of the way we conduct the science of ecology and advance conservation. Every possible clue to the past of a site or process is employed to interpret its current condition and anticipate its future. The approach emerged with Richard Fisher, and its field application was first realized in Bob Marshall’s reconstruction of the hemlock and pine forest north of Harvard Pond. But it gained its widespread foothold from the tireless efforts of Stephens and Raup. The similarities between the work of Marshall and Stephens are overwhelming and intriguing, but the connection between the two remains elusive. Nowhere in Stephens’s notes, thesis, or publication is there any reference to or acknowledgment of Marshall’s groundbreaking and distinctively similar efforts. Nor, in fact, is there any direct evidence that Marshall’s field methods were ever explicitly studied or followed by any other student or scientist. Indeed Marshall, like Stephens three decades later, did little to promote his work beyond ensuring that chunks of it were eventually published. In June 1925 he simply wrapped up his study and headed to Montana and the next step in his illustrious career. Though he turned back in his thoughts and writings to Petersham, Richard Fisher, and his close friends, he never laid any claim to the integrated methods that he had assembled to dissect a forest. This silence is intriguing, for Marshall became a tireless self-promoter of his writing and work. But on the subject of his forest history accomplishments, he was silent save for a single dismissive notation in 1927 that he had finally received a copy of his printed publication. In similar fashion, Stephens also left his legacy largely in the hands of others.

LESSONS: THE EARL STEPHENS PLOT 178

Though he retained an enduring passion for Petersham to his last days, he initially struggled with the task of turning his work into effective papers. He sought Raup’s help but then ultimately abandoned the effort. A highly independent thinker with big ideas that spanned the fields of ecology, forestry, and geography, Stephens was not a born writer despite his earnest attempts. The first draft of his thesis was clearly intended to be an opus, as by Raup’s estimate it exceeded 500 hand-scrawled and typed pages. A handful of sections from this initial effort survive, and they confirm Raup’s recollections. The work was so thoroughly edited and marked up with Raup’s exasperated marginal notes that it evidently sent Stephens into a scholarly tailspin from which he struggled to recover. He had plenty of motivation to finish: his mentor’s admiration for the study, his personal desire to inject ecological thinking into forestry, and his supportive wife’s growing encouragement to pull his lengthy student career to a close. Yet he was stymied by the massive accumulation of data and multitude of possible avenues to explore. Pressured and frustrated, Stephens abruptly changed tack. The consequence was eye-opening. In just a few weeks following his discouraged reading of Raup’s comments, Stephens’s committee approved one of the most distinctive dissertations in the history of Harvard biology. The final volume was a true antithesis to the thoughtful yet rambling first draft. Indeed, the text was a mere thirty-eight bound pages, while the appendices of maps, tables, and figures occupied roughly ten times that volume. The writing consists of a detailed summary of the field methodology; a series of short accounts of exemplary mounds, pits, stumps, and large downed logs; a brief section on the value of the historical approach to ecology; and a bibliography. But behind this thin tome lay five file drawers of field notes, photographs, and data, along with large-scale maps of the site and its features, their implications nearly unexplored. In the successful dissertation, he tapped a small fraction of the immense wealth of knowledge he’d gleaned from that clear-cut strip in the woods. He also wrote a short separate paper on the chronology of mounds and pits and the lessons this conveyed about the role of wind disturbance in forests. Published in a soils journal that was easily overlooked by ecologists and foresters, the paper nonetheless clearly demonstrated the utility of his reconstruction method to various sciences beyond ecology. That paper, Raup’s frequent use of Stephens’s work in his own talks, and word-of-mouth admiration for the heroic elegance of this graduate student’s research, did place Stephens in high regard. It also greatly enhanced the reputation of the Forest as a center for historical ecology. But in fairly short order Stephens pushed off, migrating with his wife, Blanche, and the kids to Ohio and then Alaska and the U.S. Forest Service, without ever writing up his ecological insights.

LESSONS: THE EARL STEPHENS PLOT 179

After his departure, Stephens and his plot eventually assumed a substantial role in Harvard Forest history and ecological lore. In the early 1970s a postdoctoral researcher at the Forest named Chad Oliver opened the five file drawers and recognized a gold mine. With Stephens’s blessing, he mined the data on hurricane history, harvesting, tree establishment, and growth releases, and published a truly classic study in Ecology on forest development, disturbance, and history. Applying updated concepts, concise analyses, and clear writing, Oliver achieved many of the goals that Stephens had hoped to accomplish during his decade in the woods. From his base in Anchorage, Earl served as coauthor on the paper and was extremely pleased with the result, as was the retired Raup. Meanwhile, the paper advanced Oliver’s career. In his work on his Yale Ph.D. thesis at the Forest, he applied aspects of the historical developmental approach to reconstruct the dynamics of hardwood stands just up the slope from the Stephens plot. That effort, augmented by his insightful adaptation of Stephens’s study, launched a brilliant career focused on understanding forest development and applying it to diverse forests across North America and beyond. Today Oliver is the Pinchot Professor of Forestry and Environmental Studies and director of Yale’s Global Institute of Sustainable Forestry. In separate conversations decades later, both men clearly relished the role that this forensic escapade played in their lives and in the development of new insights in ecology. Stephens retained an abiding love for Petersham. One summer morning as I walked over to the Forest from my home, which had been home to the Stephens family a few decades earlier, I spotted a well-traveled pickup and camper with Alaska plates parked in the shade of a towering maple in front of Shaler Hall. It was Earl. Like other graduates over the years, he was completing a nostalgic pilgrimage back to his intellectual roots. Through the course of the morning, we shared stories of his work, of Hugh and Lucy Raup, Ernie Gould, and other Harvard Forest denizens. That conversation led to a series of lengthy handwritten letters that we exchanged for decades until his death in 2009. In those letters he expressed his unending gratitude to Raup and the Forest for a life-changing experience. He gave much in return. Stephens’s memories of camaraderie, joy in the work and woods, and love of the place meshed well with my own experiences. It was clear that the decade he had spent in the woods just above Harvard Pond lived on for him. Today that productive decade lives on in a file cabinet, a broad strip of paper birch running through a hemlock wood, and ecological insights that, sixty years later, we all take for granted.

LESSONS: THE EARL STEPHENS PLOT 180

ELEVEN

W H E N D O I N G N OT H I N G I S A V I A B L E A LT E R N AT I V E : I N S I G H T S I N TO C O N S E RVAT I O N AND MANAGEMENT A third peculiarity about the forest is that it exhibits a dynamic beauty. A Beethoven symphony or a poem of Shelley, a landscape by Corot or a Gothic cathedral, once it is finished becomes virtually static. But the wilderness is in constant flux. A seed germinates, and a stunted seedling battles for decades against the dense shade of the virgin forest. Then some ancient tree blows down and the long-suppressed plant suddenly enters into the full vigor of delayed youth, grows rapidly from sapling to maturity, declines into the conky senility of many centuries, dropping millions of seeds to start a new forest upon the rotting debris of its own ancestors, and eventually topples over to admit the sunlight which ripens another woodland generation. —Bob Marshall, The People’s Forests, 1933

I

t is a poignant image. The date is late September 1938, just days after the Great Hurricane swept across New England, laying forests flat, ripping roofs, cupolas, and steeples from buildings, and turning thousands of boats into flotsam. Most of the timber visible on the Harvard Forest was uprooted or damaged, and the fate of three decades of research and experiments lay in uncertain condition, deep and inaccessible in the tangle of ravaged woods. The photograph captures Al Cline, the Forest’s acting director, leaning against an immense prostrate white pine, gazing across a scene of unimaginable chaos. While he clearly seeks to project an image of resolute authority and determination in his rumpled suit and stiff posture, the inescapable impression is that of a man overwhelmed, pressured, and in search of next steps. Cline’s dilemma was an extreme manifestation of the challenge that faces anyone who manages or is truly concerned about forests. Whether one is a landowner,

181

Al Cline, acting director of the Harvard Forest, amid uprooted trees following the 1938 hurricane. (Harvard Forest Archives)

forester, policymaker, or conservationist, the stewardship of forests eventually involves some level of anguish associated with the need to make difficult decisions about the fate of trees and woods. Today is an acutely worrisome time for our forests. The conversion of many broad swaths of woodland to residences, commercial buildings, and pavement continues unabated and remains the single greatest threat to natural areas, from New England to the tropics. Meanwhile, natural processes such as windstorms, ice and snowstorms, and insect outbreaks that have shaped our woods for thousands of years seem to be changing, increasing in their frequency and intensity. Add to that a growing progression of exotic organisms that damage many individual species and riddle our forests with weakened trees and the stark hulks of those that have died. The hemlock woolly adelgid is just one recent addition to an impressive biotic lineup of arboreal foes that includes chestnut blight, Dutch elm disease, gypsy moth, white pine blister rust, beech bark disease, winter moth, emerald ash borer, and—perhaps most worrisome of all given its appetite for so many types of trees—the Asian long-horned beetle. Collectively, this rogue’s gallery of what are derisively lumped together as “pests and pathogens” has made the northeastern United States an international showcase for tree species decline. The one silver lining is that through this history, including its grand recovery from the earlier onslaught of agricultural deforestation and industrial logging, our region has earned an even greater and well-deserved reputation for ecological resilience. Added to these physical threats from human, physical, and biotic forces, we now have climate change—shifts in temperature, precipitation, and their seasonal distribution and extremes—that will stress our forests further and lead to new and quite unanticipated changes. Indeed, it is the interaction among these various forces—climate change, human impacts, natural disturbances, and the outbreak of insects and disease—that is likely to exert the greatest impact on our woods. Shifts in climate per se, even extreme heat or extended drought, will seldom kill mature trees outright. In the face of even the extreme changes in climate projected for our region, many trees rooted in our forests will continue to live for decades or even centuries more; as a consequence, climate change will not directly cause much immediate change in the overall composition of our woods. But these shifts in environmental conditions will stress plants, including the largest and healthiest trees, as well as animals, in ways that make them more susceptible to other factors that can lead to death. Together these forces will create conditions that encourage novel behaviors, including outbreaks of native insects and disease that are currently not problematic and levels of damage that we have not experienced previously. In concert, the combination of chronic stress and acute impacts can kill even the largest and most robust trees. As trees die and create openings in our forests, it presents opportunities for the establishment and growth of new trees. Under the new climatic regime, southern I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 183

Panorama of the Harvard Forest following the 1938 hurricane. (Harvard Forest Archives; photograph by E. G. Stillman)

species will have an advantage as our native plants become somewhat less competitive. This is the key to climate-change impacts on nature. The death of mature individuals creates an opening for new individuals and species. Through this process of more rapid death and replacement, and after centuries of dramatic dynamics in our forests, we will begin to experience even greater rates of change. How do we conserve forests and other natural areas in the face of these many diverse foes? How should we respond when new threats loom and real impacts occur? In discussing these intertwined issues of conservation and management, hemlock serves as an effective focal point. For thousands of years before European arrival, hemlock was one of the most abundant species across broad stretches of the northeastern United States. Following decimation by intensive logging, land clearance, fire, and industrial logging for tanbark, it has recovered broadly under a regional regime of reforestation, reduced logging, and forestland protection. Its remarkable regional resurgence and its development into awe-striking stands of ancient trees make hemlock a magnificent metaphor for New England’s history of resilience and recovery from environmental damage. Clearly, it attests, nature can rebound from the most severe impacts when vast forest areas are conserved. I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 184

Yet hemlock’s upward trajectory is being arrested. The hemlock woolly adelgid is rampant; as it spreads across the region, an inevitable decline of hemlock is ensuing. With this new chapter in the saga of New England forests, landowners and land managers are experiencing some of the dread, loss, and helplessness that Al Cline lived through during the aftermath of the 1938 storm. Here we hearken back to his solutions to his dilemma and to the decades of ensuing Harvard Forest research reported throughout this volume to evaluate the many possible avenues for response. We argue that, in the face of the adelgid and other threats and impacts that will come to our woods, landowners have the ability to work actively with their forests and direct their future condition in specific positive directions. Yet we also underscore our conviction that, from many ecological perspectives, the soundest approach to land management in the face of uncertainty and change is to conserve the greatest expanse of forest possible and then forgo the human impulse to meddle unnecessarily. As Cline decided at Pisgah, the fastest route to recovery is simply to let the impacts play out and allow nature to recover as it has for millennia before. The challenge for New England is how to best ensure that our forests and other natural landscapes are able to cope with the broad range of impacts and stresses that I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 185

are likely to befall them. Our answer to that question is expressed comprehensively in the Wildlands and Woodlands vision that emerged from lengthy discussions at the Harvard Forest. Informed by history and science, it argues that the best way forward is through the permanent protection of the majority of forest in the region and flexible management that couples active harvesting with the establishment of large reserves that will remain significantly free from direct human impact. Wildlands and Woodlands draws its inspiration from the history of resiliency and recovery of New England’s forests. Specifically, following a lengthy period of agricultural deforestation and abusive harvesting, the region has reforested remarkably, with a wide array of woods now covering more than 80 percent of the land. Nonetheless, with a new wave of forest conversion for housing and infrastructure now perforating and fragmenting forests in every state in the region, the many benefits that forests provide—recreation, habitat, natural resources, and clean water and air—are being compromised. Wildlands and Woodlands views history as offering New England, and indeed most of the eastern United States, a fortuitous second chance to determine the fate of its forests. When confronted with a vast wilderness landscape four centuries ago, the European colonists and their successors cut much of the forest down. Faced with this new second forest expanse, we argue that conserving at least 70 percent of the landscape in forest is both possible and eminently sensible. Conservation of our forests will not only maintain all these values but also is the most likely means of ensuring the continuity of our natural areas and forest resources in the face of current and future disturbances and stresses. Large intact forest areas will always be more resistant and resilient than many fragmented pieces. Big blocks of forests are similarly best at providing most of the values that humans enjoy. Extensive interconnected forest will also be critical to ensure that species are able to move and adapt to environmental changes at local and regional scales. At a continental scale, our intact New England forests—continuous blocks on the local scale and expansive forest cover across the region—will continue to provide a fundamental link in the woodland habitat that extends up the Appalachian Mountains from Georgia to Canada. Retaining large blocks of land, many examples of each forest type, and a wide diversity of forests on many different sites provides the maximum safeguard in the face of multiple, uncertain, and interacting pressures and disturbances. Large intact blocks resist easy infiltration by invasive organisms and other external threats, while natural levels of diversity and redundancy provide a measure of insurance against individual disturbances exerting devastating effects. With regard to management, Wildlands and Woodlands recognizes that many thousands of private forest landowners control the New England landscape in relatively small parcels. Although this presents challenges to coordinated management I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 186

An increasingly common sight: houses perforating the New England forest. (David Foster)

and conservation, it also ensures that a diversity of approaches will be taken. Extensive private ownership makes it highly likely that the majority of land will be actively managed over time, which is a central recommendation of Wildlands and Woodlands. Yet we also advise that a substantial fraction of the region, embracing 10 percent or more of the forest, be set aside from harvesting and other manipulation in large reserves spanning many thousands of acres to a million acres or more. Although smaller reserves are extremely useful in protecting iconic sites, specific fragile habitats, or populations of individual species, the large reserves support intact ecosystems and landscape-scale patterns and processes. Expansive reserves ensure that, when large disturbances such as tornadoes or ice storms occur, the entire area is not affected and instead a mosaic pattern of damaged and undamaged areas result. These natural disturbance mosaics add variation and diversity to the more prevalent management mosaics that dominate small private landholdings. Large reserves of naturally varying forest can also serve as critical “controls,” by which we can gauge and improve the management practices that we undertake on the bulk of our forest landscape. Both objectives—that of preserving natural landscape patterns and of providing natural areas from which to develop new approaches to management— are highly consistent with the ecological and silvicultural thinking advanced by our predecessors Richard Fisher, Bob Marshall, Al Cline, Steve Spurr, Earl Stephens, and Hugh Raup. Regardless of the size and continuity of the area and despite all attempts at mitigation, every forest tract will eventually be affected by climate change and various intensities of disturbance and tree death. In the case of a hurricane, ice storm, or sudden outbreak of a native organism, such abrupt events leave dazed landowners to grapple with the aftermath. In contrast, the arrival of the adelgid and the subsequent gradual decline toward nearly complete hemlock mortality takes years or even decades. As the inevitable draws nearer, landowners may search widely and anxiously for answers. How are we to think about these looming disturbances and plan for the most appropriate management strategies to deal with them? Al Cline’s thoughtful and varied responses to the 1938 hurricane bracket the possible spectrum of action. In Petersham and across central Massachusetts, he persuaded his staff, local landowners, town officials, and state agencies to cooperate with the federal timber salvage program, harvesting as much of the downed timber as possible. The resulting effort was extraordinarily efficient and thorough—an example that truly makes one marvel at the efficiency of a federally coordinated regional effort undertaken with local community engagement. Nearly every available stick was harvested, and the resulting slash was burned. Today most areas of the Harvard Forest and New England I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 188

that were severely affected by the storm retain the tell-tale sign of salvage—uprooted and cleanly cut pine stumps leaning from the associated mound and pit of earth in a northwesterly direction—and remarkably few of the decaying downed trees that characterize the former old-growth forest at Pisgah. At Pisgah Cline charted a course of preservation for the uprooted ancient forest that was in keeping with its history, the science that had emerged from these woods, and the actions of his mentor, Richard Fisher. In those remote New Hampshire hills, Cline rejected the preponderance of professional advice and public sentiment to salvage the lumber. Instead he mandated that the hurricane-blasted forest should lay untouched by human hands. State and federal officials had strongly pressured Cline to follow the regional trend and actions of surrounding landowners by salvage logging the majestic logs, to reduce what they perceived as the great hazard of fire. Meanwhile, local mills and loggers clambered for access to the immense and valuable old-growth wood. Letters and notes in the archives include offers to purchase the timber and suggest that Cline perceived repercussions if he disappointed these various interests. Nonetheless, he held steadfast to his mentor’s vision; the saws stopped at the Harvard boundary, beyond which today lie the moldering remains of cut stumps and the tops of harvested trees. As Cline wrote in 1939 in response to an urgent appeal from logger Thomas Hanifin from the town of Belchertown, Massachusetts, “After thorough consideration of the purposes for which this old growth forest was originally acquired it has been decided to let nature continue to take her course free from interference by man. I must inform you that there is no possibility of making you a stumpage sale.” In the decades since, Pisgah has remained “virgin” and has yielded insights to the mind along with conviction to the soul that some early foresters embraced a range of values that resonate deeply with us today. Looking across that landscape, we see woody structures and contorted features that are structurally distinctive in our New England landscape, though once common in the land before European history. From that scene of vibrant new growth and in the legacy and remains of a once magnificent old-growth forest, we also recognize that nature needs no help from us. It can recover perfectly well on its own. The advice that we share with landowners concerning the adelgid spans a similar breadth of activity. In those cases where there is a tradition of intense management, where there is heavy public use of trails, roads, picnic areas, and campsites, or where timber revenues are important, extensive harvesting and clearing of the dead and dying trees matches landowner interests and objectives. For these landowners, getting into their woods and cutting the hemlocks (and often other trees) helps them feel engaged with their land and satisfied that they are exerting some measure of control over its future. Bent, broken, and poorly formed trees can be removed; valuI N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 189

Dave Kittredge gazes up one of the remaining large white pines on the Pisgah tract. (David Foster)

able or otherwise preferred species can be retained; and the appearance of the woods can be “improved” in whatever ways please the landowner. But for most areas, including the broad expanse of Harvard Forest lands and all of our most majestic older groves, we promote a strategy of letting nature take its course. In offering this advice, we are sharing what we’ve learned from 1938 and subsequent storms, outbreaks, infestations, and environmental extremes at the Harvard Forest regarding the nature of these events, their ecological consequences, and avenues for response. Any discussion of management approaches must begin with motivation. What are the guiding objectives of the landowner, land managers, or other decision makers? Is it to derive income from high-quality timber and firewood, or are these secondary interests to other values such as aesthetics, recreation, and personal enjoyment? Are there ecological interests such as the value of the land as habitat or for I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 190

protecting streams, wetlands, and ponds, and yielding clean water and other benefits to local communities? Articulating their central long-term interests and motivations helps landowner-managers understand not only what they want from their land but also how they should proceed. One important lesson that has emerged from watching countless responses to storms and tree mortality is that “catastrophic” events such as major insect infestations or windstorms trigger a strong human response to respond and to act. Regardless of their stated objectives or even long-held policy, most people who are faced with what they view as devastation, loss, and chaos in nature react with heartfelt dismay, empathy, and a sincere desire to repair the damage, clean the mess, and “stabilize” the situation. Much of our language about natural disturbances (including the word “disturbance,” which we employ regularly) reveals the attitude that society takes toward these events, including the underlying assumption that our actions can help. Individuals hired to manage in such cases often act precipitously, in a desire to demonstrate authority, control, and responsiveness that will reassure supervisors, shareholders, and the public. “Decisive” actions serve as outward signals that the uncontrollable is being subdued, the calamity rectified, and order reinstated. Land managers, for instance, have been quoted as stating that their salvage or preemptive cutting of hemlock stands is guided by a desire to minimize the impact of the adelgid and to mitigate the negative consequences of the biotic disturbance. This explanation is usually proffered with no acknowledgment of the irony inherent in the incorrect implication that the cutting of trees and driving of heavy equipment across the land will have less impact than the trees’ gradual decline due to the insect. One somewhat unexpected extension of the argument for an active response to disturbance is the suggestion that we use forest harvesting in a proactive or preventive manner to develop forest conditions that are more resilient to future disturbance events. This type of “protection forest” approach is being employed on various state water-supply lands in southern New England, including the Quabbin watershed, which produces over 250 million gallons of water daily for nearly three million people, or 40 percent of Massachusetts’s population. These state lands are managed with a mandate to “protect the drinking water supply watersheds,” and watershed forest managers under the application of the protection forest hypothesis have made these forests among the most intensively harvested lands in southern New England. The ostensible ecological rationale for this intensive harvesting is the belief that forests that are low stature (having short trees), uneven-aged, and varied in physical structure and composition are more resistant to physical disturbance, pests and pathogens, and chemical stresses such as acid rain than naturally developing woodlands. With less damage from disturbance and stress, the argument goes, there should be better water filtration and higher water quality. I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 191

A three-foot-diameter dead snag at Pisgah, surrounded by trees that grew up and established after the 1938 hurricane. (David Foster)

While intuitively compelling—the greater the variation, the more likely that elements of the forest will survive the next suite of stresses—actual evidence supporting this assertion is essentially nonexistent. Meanwhile, the effort to achieve the purported state of resilience requires harvesting trees, developing and maintaining roads, and running heavy equipment through the woods to remove the material, each activity having its own impacts on the forests, soils, and water quality. It is naive to expect predictable responses from tree species and ecosystems to these interventions, particularly in light of the highly uncertain future in which human disturbances may provide greater opportunities for the spread of invasive plant species and shifts in tree species distributions. An additional problem with the protection forest concept is its underlying assumption that water quality is harmed when forests are damaged by ice, wind, or insects. Again, although it’s intuitively compelling, this assertion is simply untrue. The protection forest idea remains a highly questionable and untested hypothesis, due to its lack of rigorous testing and experiments. Meanwhile, results from two efforts to implement harvesting to reduce the severity of impact from insect pests have had mixed success. The preemptive harvesting undertaken in an effort to diversify spruce fir forests (to reduce the susceptibility to spruce budworm defoliation) and the thinning of oak stands (to reduce gypsy moth defoliation) has actually led to higher defoliation levels than in unthinned stands. Ecological theory and practice show that the human responses to natural disturbance almost always exert a more severe impact on forest ecosystems than the event that they were seeking to mitigate and rectify. This makes great sense, since the harvesting and removal of trees is simply another layer of disturbance placed upon the existing disturbed forest. Therefore, when considering a management response to the adelgid infestation, the slow pace of hemlock decline should be seen as an opportunity for deliberation and a chance to avoid reactionary management, where action is a substitute for thought. Ideally, prior forest planning will determine the response to any disturbance, and salvage operations will be focused on areas that were previously slated for tree harvesting. Even in these cases, the meager economic returns from salvaging dying hemlock should be weighed against the innumerable ecological benefits these dead and dying legacies will provide for the future functioning of these areas. A historical perspective reminds us that many diverse disturbances have occurred in the past and have helped to shape our modern forests. Mortality occurs regularly at every scale: branches, individual trees, small groves, and entire landscapes. In many ways, therefore, the physical and biotic upheaval of ecosystems is a

I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 193

perfectly natural condition, albeit an episodic and comparatively rare one. From this ecological perspective, a wind-flattened forest, scorched hillslope, or beaver-flooded valley filled with dead trees is still a perfectly well-functioning ecosystem. It is no less intact and representative of the natural condition than an old-growth forest or a rapidly growing young stand of paper birch. The same is largely true of the impacts of most of our invasive organisms. While adelgid and other exotic species are not “natural” elements in New England forests, their impact in an individual forest is often quite analogous to and well within the range of damage from many natural disturbances, including extreme outbreaks of native pests. Looking back 5,000 years, we can see that hemlock and oak once experienced the level of impact associated today with the hemlock woolly adelgid. Tellingly, the language that we generally employ in situations where wind, fire, insects, or ice alter our forests reflects human values, sensibilities, or aesthetics but often has little relevance to ecological reality. We variously call these events “catastrophes” and “disasters” and label the forests in the resulting scenes as “damaged,” “destroyed,” or “lost.” While the imposition of human infrastructure—such as roads, bridges, and houses—into pristine natural areas may also be correctly described in such terms, ecosystems and landscapes that are transformed by extreme natural events remain fully functioning. Although the trees in a hemlock stand infested by adelgids are dead, dying, or in decline, the ecosystem remains intact and healthy. In our eyes, values may be lost. Yet as sad as it may be to lose the majestic hemlocks and gorgeous panoramas that we cherish, many resources are conserved, nutrients continue to be cycled, energy flows are maintained, and many organisms, established as well as new, thrive in these altered ecosystems. A corollary to our interpretation of disturbance as a natural process and of physical disruption as a natural condition is our recognition that ecosystems have immense capacity for recovery. Indeed, studies of processes as disparate as hurricanes and fires and the hemlock woolly adelgid demonstrate that, although the ecosystems may seem broken to our eyes and many trees may be literally quite broken, the forests remain intact in terms of their core biological processes. The dying trees leave roots in the ground and immense volumes of physical structures in the forest in the form of wood, branches, and leaves. These remnants mitigate erosion and additional soil disruption, and they shelter remaining plants, soil, and streams from major environmental shifts. Even as the affected trees and overstory succumb, they are replaced in many of their roles by remaining plants and new arrivals. The openings in the forests that the dead trees create and the light, moisture, and nutrients that they liberate aid their former competitors, fuel an explosion of new growth from herbs, shrubs, and saplings, and provide new habitat for a variety of wildlife species to use for dens, nests, and foraging. They also trigger the germination of I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 194

seeds and sprouting of new shoots. Although crippled in appearance, the ecosystem is vibrant and thriving; the unfolding changes border on the miraculous to behold. Our studies of windstorms and hemlock mortality document that the relatively minor changes in major ecosystem processes following such disturbances are largely the combined result of the retention of physical structure and the surprisingly rapid expansion of new leaf cover. Both help to shade the ground and soil surface from direct sunlight, thereby reducing major fluctuations in temperature. The great expansion of leaf cover and roots from the existing and establishing new plants allows rapid uptake of water and nutrients, continuing and reestablishing the forest’s control over water flow and evapotranspiration, and retaining nutrients in the ecosystem. This discovery came as quite a surprising result of a large experiment to simulate the impacts of an intense hurricane that we initiated in 1990 as part of our Long Term Ecological Research program. We pulled down mature forest trees on more than two acres of the Harvard Forest to mimic as closely as possible the type, extent, and patterns of damage experienced in similar forests in 1938. We took this rather extreme approach to improve our understanding of natural disturbance processes and forest recovery. The experiment enabled us to study the forest in detail before manipulation and then to comprehensively measure all physical characteristics and processes in an ongoing manner all the way to the present. The results were stunning. Though physical disruption was rampant and the word “catastrophe” emerged from every visitor’s mouth on first glance, forest processes and even major environmental conditions remained intact. The downed trees, remaining stems, and resurgent growth shaded the ground and maintained consistent soil temperatures and even light regimes at ground level. There was no appreciable nutrient loss, thanks to the efficient uptake of nitrogen and other limiting chemicals by the highly productive plants. In turn, the great availability of resources led to a rapid recovery of the new forest’s overall productivity. Over more than twenty years we have also seen remarkably little change in forest composition, as survivors and sprouters outnumber newly established plants. While these results have demonstrated the resilience of our forests, they have also highlighted the fact that many of our conclusions emerging from the 1938 hurricane were strongly colored by the extent of salvage harvesting after that event. Erosion, the great profusion of weedy and early successional tree species, and the sustained increase in flow of our streams and rivers were more likely the consequences of the efficient removal of dead, damaged, and surviving stems by the army of loggers than the direct consequences of the storm. Some additional general lessons have emerged from our research. After both wind and insect damage comes a period of transition in which the control by the previous forest dominants wanes and a set of new forest trees emerge in imporI N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 195

Snow- and ice-covered weir, measuring the flow of the stream that drains the Black Gum Swamp and Prospect Hill hemlock forest. (David Foster)

tance. Though wind, ice, and snow events appear to kill the vast majority of trees, the surprising fact is that most remain alive, at least for some lengthy period after the event. Hardwood trees stripped of their branches often sprout new and bushy epicormic branches up the length of their trunks, while snapped stems often sprout at the base. Most trees blown flat in a windstorm are uprooted rather than broken, a process that can leave up to half the roots in the ground and oddly enable the tree to function for many years even in a prostrate condition. In the aftermath of our hurricane experiment, more than 60 percent of the flattened trees releafed the following year, with the majority dying gradually over the next decade. The sprouting and releafing trees combine with and gradually trade off with the newly establishing trees in conferring continuity in leaf cover and important ecosystem processes. Though physically transformed, the forest continues to function. A similar gradual shift occurs in hemlock forests as they decline from the adelgid and in most other forests affected by insects and pathogens. Mortality comes relatively slowly—in four years in extreme cases and a dozen years or more in many New England forests. On the trap rock ridges in southern Connecticut, scattered large and thinly leaved hemlocks remain in forests that were invaded by the adelgid in the late 1980s. The slow loss of foliage and gradual decline of the hemlock cause a progressive shift in the understory environment, letting in light and delivering nutrients that allow other plants to establish and thrive, replacing many qualities that hemlock imparted to the stand. By the time that the last hemlock dies and the forest includes a jumble of branches, tops, boles, and snags in various stages of decay, a new forest is nearly fully established, dense with rapidly growing saplings of birch, maple, pine, and other species. Although the appearance, structure, and composition is distinctly different, forest continuity has been achieved. Remarkably, most people walking through these stands do not even notice the ongoing changes. Any level of human manipulation exacerbates these impacts and reduces the effectiveness of the damaged forest to effect this ecological transition in species, structure, and function. As demonstrated in the Cut or Girdle experiment, our studies across southern New England, and evaluations of the 1938 hurricane, intensive intervention including broadscale salvage logging or the preemptive harvesting of the declining trees can interrupt the process of forest continuity and recovery in significant ways. In the well-intentioned effort to improve the situation, clean up the mess, or recover value from an apparently damaged ecosystem, our management and restoration activities frequently exert a much greater impact on the ecosystem than the disturbance itself. Harvesting of the dying or damaged trees kills them instantly and removes their structure. This compromises the ongoing capacity of the forest to take up nutrients and moisture and often damages the many surviving plants, leading to abrupt shifts in ecosystem conditions. In such operations, seldom I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 197

are only dead and dying trees removed. After wind and ice damage, bent or leaning but thriving trees may be taken because of their lower potential for producing highquality timber. Healthy hemlock may also be removed in the expectation that these may die, but doing so eliminates any chance at detecting variation in resistance to the adelgid. In any salvage or restoration activity, it’s likely that healthy individuals of valuable timber species will be harvested to make the operation more viable financially or to conduct other desired work. Of equal importance, any type of operation requires access to the forest to transport equipment into and logs out of the site. At a minimum, this necessitates the cutting of additional trees to create work areas and skid trails. At an extreme it can result in the creation of an additional road network. Once human intervention proceeds, what started as a troubling meteorological or biological disturbance rapidly becomes a major harvesting activity. The ultimate impact often easily surpasses that of a more typical harvesting operation. Whenever a widespread disturbance strikes, concern about wildfire usually follows, perceived as a threat both to the forest and to surrounding property. Lots of speculation and strong opinions have been formed on this topic, but few studies collecting real data have been conducted in temperate forests in the northeastern United States. Ice storms and major windstorms are frequently cited as possible catalysts for forest fires in the New England landscape. Indeed, this was a major motivation for the massive salvage program following the 1938 hurricane. Still, no studies have been made of the changes in fuel loading after windstorms in New England forests, and there is little historical research that effectively links fires with windstorm events in the region. Most cutting follows on the intuitive notion that downed woody debris enhances long-term fire hazard. Although blowdowns in New England conifer stands may produce an increase in fire hazard, data from the hurricane experiment suggest that this condition is extremely short-lived. That’s because fine fuels (small branches, twigs, and needles), which are the main fire concern, were unevenly distributed and decayed rapidly. As a consequence, overall fire hazard was only slightly increased in the experimental study and for a brief period. The same is true for many sites with high hemlock mortality, as the fine branches decompose quickly, and the potential for fire declines rapidly with the establishment of understory vegetation. The final ecological consideration surrounding management is that of legacy— anything that happens in our forest today will have an enduring impact for a century or more. The uproot mounds and decaying boles from trees flattened in 1938 remain vividly evident at Pisgah, while tipped stumps characterize many stands at the Harvard Forest. Chestnut poles that died in 1915—nearly a century ago—lean against large hemlocks in many of our old woodlots. And the woods roads, stone I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 198

walls, lilacs, Japanese barberry, and ancient apple trees in New England woods provide evidence of land management decisions made by our predecessors centuries ago. In similar fashion, the impact of the hemlock decline and our response to it will play out for decades and centuries to come. Every action that we take in the woods today alters and adds to that legacy. We need to be thoughtful in our decisions and careful in our actions, because their consequences will ripple through the woods long after we are gone. The reality is that ecological consequences are rarely the sole or even the major considerations driving land management decisions. Other factors may be extremely important, including aesthetics, resources, wildlife, safety concerns, recreation, and the future condition of the stand. Moreover, given that many landowners look at their forestland as a bank account from which to draw in emergencies, we can’t ignore the financial considerations. There is no right or wrong approach except as evaluated against specific criteria. The critical issue for a landowner or manager is to define those criteria in advance of the actual need. Because of the great influence that management can have, the desire of landowners to learn about their management options, and the lack of existing guidelines, we developed a series of forestry options for landowners and managers to consider when dealing with or anticipating hemlock woolly adelgid on their land. If the objective is to manage in harmony with natural process and to minimize environmental impacts, then in most cases we recommend doing nothing, meaning no active management. Infested hemlocks will gradually die without much direct soil disturbance and with minimal environmental change. The forest will recover surprisingly rapidly but to a different condition, due to the replacement of hemlock primarily by a range of hardwood species with substantially different attributes. The remnants of the previous stand will persist for many decades as important structural elements that support a diversity of organisms. It is useful to underscore that, even in their death and their declining state, the hemlocks will continue to play an important role as many insects, invertebrates, birds, and mammals use the dead wood along with the habitats and organisms that this generates. On the other hand, intervention and active management enable the landowner to deflect the trajectory of the forest in particular desired directions and to obtain a variety of wood resources in the process. We use the word “deflect” deliberately, because there are constraints to what management can achieve: the history and current conditions of the stand, the nature of the site, the density of herbivores such as deer, and the qualities of the surrounding landscape all impose strong direction to the forest’s future development. That said, landowners have choices to remove or retain particular trees or species based on their appearance, value, or condition I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 199

relative to specific goals such as timber production or wildlife habitat. Management can also lessen specific hazards by removing dead, dying, or damaged trees and may work on problems with existing populations of invasive species such as Japanese barberry, honeysuckle, or buckthorn. In developing access to a property, it is also possible to create or improve networks of trails and wood roads in thoughtful and deliberate ways that enhance the use of the property. Management can exert its strongest impact on the forest’s future through decisions regarding what trees should be left, the size of the openings generated in the forest canopy, and the manner in which the soil and ground environment are treated. Removal of dead and especially living trees increases light penetration and can shift the balance in the capability of different tree species to establish and compete in the developing woods. Thus, the size and configuration of the openings created through harvesting can shape the relative proportion of shade-intolerant trees versus the generally slower-growing shade-tolerant species. In hemlock-dominated forests, we outline two approaches that contrast in cutting intensity, to highlight some of the options available to landowners and managers. Intermediate-intensity harvesting is also a viable approach that would be achieved through a combination of the two courses of harvesting. (1) Selection or shelterwood cuts remove 20–50 percent of the trees in a forest stand either in relatively small (1/8- to 1-acre) openings (selection), or in a distributed pattern (shelterwood). This removal generally leads to a modest increase in light and soil disturbance that encourages more shade-tolerant species to establish and grow. Natural disturbances of this severity are commonly documented in old-growth hemlock forests, and they provide opportunities for species such as yellow and black birch, red oak, and white pine. (2) High-intensity cuts remove over half of the trees and create larger-sized openings that increase the amount of light reaching the ground and generate more intense and widespread soil disturbance. These conditions lead to extensive sprouting of hardwoods and rapid establishment of shade-intolerant and early successional species such as paper birch that are less common following less extensive natural disturbances. If the infested forest is a mixture of hemlock with hardwoods or conifers, we recommend cutting hemlock in groups or throughout the stand to recover economic value and to speed the conversion to a hardwood stand or to encourage other conifer regeneration. It is important to recognize that both approaches offer other opportunities to vary the nature of the impacts and resulting conditions and future growth. One major determinant of soil conditions is the season of harvest. Operations occurring in the winter can take advantage of frozen soil or snow cover to lessen the impacts on the ground. On the other hand, the scarification or disturbance of the soil surface

I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 200

helps to regenerate light-seeded species such as birch and may be achieved through a warm-season harvest or operation of equipment in ways that enhance this effect. Harvesting has traditionally involved hand felling with chain saws, skidders to remove cut logs, and the development of a network of skid roads to access the forest, but the increasing use of mechanized equipment such as feller bunchers and forwarders provides the opportunity to control both the impact to the site and future access to the stand. Planting trees in New England woods is seldom feasible and almost never necessary, because the trees that are left postharvest will become the future forest and will dominate the seed rain within the stand for many decades to come. Nonetheless, owners who would like to retain an evergreen component after hemlock loss may want to directly plant seedlings of native evergreens such as white pine, red pine, and red spruce. Some have even planted the nonnative Norway spruce, because its full crown of needles resembles hemlock for habitat and shade characteristics. Others who want even more diverse hardwood mixtures can complement what fills in naturally with plantings of oak seedlings. Because of the financial cost, any planting operations should be planned carefully to ensure that planted species will not be overgrown by natural regeneration or eaten by deer. Seedling shelters, fencing, and thinning of competing vegetation may be required in some locations to help planted seedlings succeed. There is no single correct approach to forest management in New England, and certainly none that can be applied to all situations. Landowners should identify their needs and desires and then work creatively with the specific situation at hand. Most important is a thorough consideration of all factors to make informed decisions about the future of their forests. By the 1990s, the research group at the Harvard Forest had become convinced that the hemlock woolly adelgid presented a major force in the landscape, so we began to fashion a plan for our own forest lands. Our discussions and debates were extensive, animated, and ongoing. They also involved all members of our staff—from scientists, managers, and our Woods Crew to collaborating researchers in our Long Term Ecological Research project and visiting international and American researchers in our Bullard Fellowship program, as well as many outside forestry professionals. We quickly concluded that widespread control of the adelgid was impractical and would have potentially negative effects. Though local control can be achieved on individual trees and small stands through insecticide spraying or soil injections, these treatments are expensive and essentially unending. The cost at any large scale is prohibi-

I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 201

tive, and the potential for undesirable impacts on other species or soils, wetlands, and streams dissuaded us from considering this as a broad approach. In the end, we chose not to apply chemicals at all. From the outset, we were skeptical of biological controls such as predatory insects. On one hand, we were and remain concerned regarding the efficacy and application of the screening protocols and safeguards of the review process used in the selection and release of such organisms. The potential for damage to other organisms seemed great. Indeed, there are many global examples of biological control programs that have done major environmental harm through unanticipated impacts on nontarget organisms. One regional example of such unintended consequences is the dramatic decline of our large native silk moths such as the luna, cecropia, and Polyphemus moths, which resulted from the release of a parasitic fly to control the invasive brown-tail and gypsy moths. In this case, the parasite was ultimately effective in its intended role, but that achievement is often not realized, and the collateral damage caused by a generalist biocontrol agent gone wrong can be huge. On the other hand, we have also been skeptical that the native predators from Asia selected for biocontrol programs and release in New England have substantial potential to control the adelgid. In their native setting, these organisms are strongly aided by the fact that the Asian hemlocks all exhibit significant resistance to the adelgid, something that is completely lacking in eastern hemlock. This skepticism has been reinforced by the poor performance of each of the biocontrol prospects in field trials in New England. Despite millions of dollars spent on biocontrol efforts to combat the insect, and contrary to many claims in the press and the gray literature of breakthroughs and successes in laboratory trials and field releases, there is no evidence that any of the released Asian predators have established enduring reproducing populations or have sustained any measurable reduction of adelgid infestation in New England. Indeed, as we write this overview, we are not aware of any evidence that field releases of Asian biocontrol agents have saved a single hemlock tree from the adelgid. The sobering fact is that ongoing attempts at biological control show no greater promise than any over the last two decades. Nonetheless, there is one extremely bright side to this story: to date, there is no evidence of collateral damage by any of the released organisms to any native insects or organisms. In thinking of our own management options, our decision making was strongly influenced by our review of history and past responses to major forest disturbances. We have been especially focused on the fact that, in the wake of the thorough salvage operations following the chestnut blight and the 1938 hurricane, it is nearly impossible to locate areas where it is possible to study the long-term consequences of these major events in the absence of associated harvesting. This history and our studies of the adelgid and landowner behavior have convinced us that the majority I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 202

of hemlock forests across New England will be harvested in the aftermath of the adelgid, as after most disturbances. Therefore, one major role for the Harvard Forest lands is to provide contrasting conditions, including large areas in which the insect disturbance alone can be witnessed and studied in great detail. Consequently, we have chosen to study the impacts of salvage and harvest on outside lands while in our own backyard we are following a completely different course of action. While we will use some of our hemlock forests for various experiments, such as the large girdling and harvesting manipulation at the Simes tract, we will leave the bulk of our forests untouched as the adelgid infestation plays out. Our woods will therefore continue to serve a largely scientific and educational purpose. Today and over the next decade, they will provide an unparalleled record of the decline of a foundation species on sites where the history is known in remarkable detail. The dynamics will be recorded by countless collaborators using a diverse array of ground, tree, canopy and airborne devices that have not been previously available in locations during preinfestation periods, and also during and following an actual pest outbreak. In the future this new record will complement the historical records to provide the basis for studies not yet conceived on sites where the investigators will not need to unravel the confounding efforts of harvest and salvage operations. More important, this record will provide a critical contrast as future storms and introduced insects continue to shape the dynamic of our forests, and we contemplate the best course forward.

I N S I G H T S I N TO C O N S E RVAT I O N A N D M A N A G E M E N T 203

Lessons from Harvard Forests and Ecologists IV. Three Views from John Sanderson’s Woodlot The hemlock, which was abundant up to 1875, furnished not only lumber but tanbark. The great trees were cut down in June, and the bark was taken off and piled up to dry. It was used by local tanneries; there was one on West Street, where there is now a store, and another, a very prosperous one, conducted by Deacon Sanderson. —A. F. Johnson, 1922, “Some Reminiscences and Recollections of School District No. 10 [Petersham, Massachusetts],” manuscript, Harvard Forest Archives This farm formerly of more than 400 acres, is situated in the Bennett Hill district of the north part of Petersham. Sixty years ago it was extensively cultivated by John Sanderson, one of the wealthiest farmers in northern Worcester County. He was killed in his barn in 1831 in the act of taking a pair of unruly oxen off the cart tongue. —George Sumner Mann, n.d., Mann family genealogy, Harvard Forest Archives

I

t is likely the most famous farm in the scholarly fields of ecology and environmental history. When Hugh Raup transformed the slide show–based lecture that he had presented to enthusiastic audiences for years and published it in 1966 as “The View from John Sanderson’s Farm: A Perspective for the Use of the Land,” he produced an instant success for the journal Forest History and opened the eyes of people in many disciplines. The article used the Harvard Forest dioramas and the history of colonial management of the farm that became the Harvard Forest to insist on the critical need for understanding the role of humans and social forces in determining the fate of the land. The fundamental qualities of the land remained the same, he asserted, but the people and forces behind their actions changed over time, informed by incomplete knowledge of their circumstances: I suggest that the principal role of the land and the forests has been that of stage and scenery. The significant figures have always been the people,

204

and the ideas they have had about what they might do at specific points in time with the stage properties at hand. At each such point in time an actor could play his role only by the rules he knew—in terms of his own conception of his relation to the play of which he was a part. He was always hampered by lack of precise knowledge of the stage and its properties, the land and the forests. Perhaps more important than this, he had severely limited knowledge of the changing rules by which he and other actors of his time were playing. Both of these failings are perennial and no doubt will continue to be. Raup’s influential article drew from his decades of inquiry into land use history in Petersham; the extensive insights into the relationships among soils, vegetation, and human activity advanced by Steve Spurr, Earl Stephens, Walter Lyford and others; and Ernie Gould’s economic analysis of Harvard Forest’s history as a financially unsuccessful forest manager. “The View from John Sanderson’s Farm” presaged by forty years the injection of social science into the field of ecology and the development of such programs as the “Dynamics of Coupled Natural and Human Systems” by the Directorate for Biological Sciences at the National Science Foundation. Along with Raup’s other papers—“The History of Land Use at the Harvard Forest” and “Some Problems in Ecological Theory and Their Relation to Conservation”—it extended the legacy of ecological studies grounded in history and seeking to advance conservation that were the hallmark of Richard Fisher and have become the focus of our Long Term Ecological Research program. “John Sanderson’s Farm,” as the article became known, was immensely popular and quickly came to hold the distinction as the single most cited paper in the journal’s history. The themes that Raup explored have also stayed alive and contentious, as revealed by a recent challenge to his assertions regarding the failure of natural resource conservation by our Harvard Forest collaborator and Brandeis professor Brian Donahue in a 2007 article in Environmental History, aptly titled “Another Look from Sanderson’s Farm.” In all the illuminating attention on Sanderson’s farm and fields, one crucial fact largely escaped notice. Sanderson was more than a simple early nineteenth-century farmer, though he may have been typical in many regards. He was also a tanner, whose most important building was a small water-powered tannery, and his most treasured piece of woodland was a hemlock forest. The tannery was a critical economic engine in the Sanderson enterprise, and one that helped John earn a reputation as a shrewd businessman and his distinction as head of one of the region’s richest families. Meanwhile, the hemlock woodlot and the adjoining Black Gum Swamp (which despite its name is dominated by hemlock and red spruce) remained the only continuously forested parts of the expansive Sanderson farm through the

LESSONS: JOHN SANDERSON’S WOODLOT 205

eighteenth and nineteenth centuries. This “ancient woodland,” as the British would call such a heavily used but enduringly intact section of forest, provides a great deal of history and many lessons. The resulting tales have greatly informed our understanding of hemlock and its place and dynamics in the New England countryside. They also have expanded on Raup’s insights regarding the relationship between humans and the land.

Around Shaler Hall, it is known simply as “the hemlock forest.” Any reference to John Sanderson’s woodlot today would likely draw blank expressions from the undergraduates who spend their summers measuring trees there or from the scientists who have dedicated their lives to understanding it. The area is thoroughly dominated by hemlocks, including a few that approach 250 years in age. One of our oldest and most intensively studied forests, it is just a short walk from our main building, in the heart of our thousand-acre Prospect Hill tract. Age, history, and access have made it a premier site for research, but the allure of the hemlocks and their peaceful environment, with its quiet yet distinctive background sounds, are an unspoken motivation for every study conducted there. The stand does have remarkable scientific assets to draw researchers, most important of which is an unrivaled catalog of data. These records range from a century of detailed maps to 10,000 years of fossil records on climate, disturbance, and forest dynamics. We have a decadal record of the fluxes of carbon dioxide, energy, moisture, and other substances into and out of the woods along with minute-by-minute graphs of water flow in small streams that drain the area. The atmospheric measurements come from a set of instruments positioned just above the forest, atop a tower capped by a twenty-foot pole that once served as the mast of a small sailing vessel. This is just part of the bizarre array of ecological equipment and paraphernalia that distinguishes (and occasionally clutters) the forest. The tower sports another item that garners considerable attention from afar: a digital camera that yields captivating panoramas posted hourly on a website. The view scans across the tops of the hemlocks, pines, and hardwoods to the fire tower atop Prospect Hill. Whether in brilliant sunlight, heavy mist, or skies filled with fat flakes of snow, the images convey information about the state of the foliage and plants that can be related to the streams of environmental data flowing in tandem from many other devices toward a central digital archive, where all are curated for use by scientists across the globe. John Sanderson’s hemlocks shade the headwaters of two small watersheds that drain to separate coastal waterways. One flows to the vast Quabbin Reservoir down Bigelow Brook and the majestically wooded course of the East Branch of the Swift River. From there it is diverted via a buried aqueduct to metropolitan Boston, where LESSONS: JOHN SANDERSON’S WOODLOT 206

John Sanderson’s farm today, looking across cow pastures to the old Marsh place, known today as Fisher House. (David Foster)

it provides crystal clear and unfiltered drinking water to that urban population and is eventually released into Boston Harbor, Massachusetts Bay, and the Atlantic Ocean. The other watershed contains the ages-old course of water flowing north to the Millers River, then west to the Connecticut Valley and south by way of New England’s largest river through the heart of Connecticut to Long Island Sound. Thus, beneath the hemlocks it is possible to examine the intimate interactions between living and decaying plants, diverse organisms, and layers of soil that condition waters which eventually end up in kitchen sinks or seas miles and years apart. Since its appearance on the first map of the Harvard Forest—a crude survey that was hand-inked and -colored on thick vellum by the class of 1907—the SanderLESSONS: JOHN SANDERSON’S WOODLOT 207

John Sanderson Jr., the third generation head of the farm, sold the property in Petersham, acquired a farm in the fertile Connecticut Valley, established a highly successful bank, and became a state senator. (Harvard Forest Archives)

son woodlot has been studied by every generation of scientists to reside in Petersham. Over the years the collections of peculiar equipment and studies in those woods have yielded data that fill many archive drawers and digital storage devices. The work has advanced the missions and tapped the funding of nearly every federal agency that cares about the environment—the National Aeronautics and Space LESSONS: JOHN SANDERSON’S WOODLOT 208

Administration, the Department of Energy, the Environmental Protection Agency, the U.S. Forest Service, and the National Science Foundation. Today these interests and many others converge on a huge eighty-acre plot in the heart of the woodlot, where, in an effort funded by the Smithsonian Institution and Harvard University, every stem down to a thumb’s width is being measured, mapped, and recorded. This effort to link the intimate dynamics of every plant in the woods to its environmental drivers and the forest processes that control the movement of gases, material, and water is part of a global network of plots that extends from Malaysia to Australia, Brazil to Panama, and Yosemite on to Ontario and finally New England. For decades the Sanderson woodlot has contributed to all that we know about hemlock. These days we are converging on the woodlot in earnest, for it has become our ground zero in the documentation and analysis of the latest chapter in the species’ tumultuous history: its response to the hemlock woolly adelgid. We strive to follow the tiny insect to understand its movement and behaviors, and we seek to know, long before the first overt signs of collapse occur, when the trees first feel this foe’s impact. We will document the ripples of cause and effect as the gradual changes in the health, function, and form of this grand foundation species reverberate through the environment, affecting other plants and the habitats of these woods. All the while, the spectacle of slowly dissolving hemlocks will wrench the hearts of those objective scientists tending the instruments, counting the insects, and archiving those seemingly endless streams of data. Hugh Raup was right. The relationship of people to this woodlot and the tree that dominates it is a story in itself.

My first foray into John Sanderson’s woodlot was on a steamy July day in 1983. The ancient hemlock forest immediately captivated me with its contrast to the rest of the second-growth landscape that I had thus far seen in Petersham. I’d recently been studying wilderness landscapes in the vast forest and wetland expanses of Labrador, so I was drawn to the chance to study this oldest and least disturbed part of the Harvard Forest landscape. From my brief reading of Steve Spurr’s 1950s research, I knew that a few trees in this stand dated back a couple of centuries to the town’s founding, but beyond that the story faded. How long had hemlock been here, and how had the scene changed through the millennia? When did these trees get established, and what did the colonial landowners do with its original ancient growth? Had it changed much, or were Spurr and Raup correct in asserting that this and other ancient woodlands at Slab City and Tom Swamp were reasonable analogs for the forests that had prevailed in this landscape for thousands of years? The literature I had uncovered thus far left all these questions unanswered, and even my cursory reading of this landscape sugLESSONS: JOHN SANDERSON’S WOODLOT 209

gested that it would be impossible to derive any more information from the forest itself unless we applied tools and approaches unavailable to my predecessors. As I hiked along pondering these things, the hemlock’s deep shade offered me some modest respite from the July heat, but my golden retriever soon found a more refreshing niche—a small woodland pool filled with mud and just enough water to reach her belly. As she sank down into the water, I was intrigued by the thought that the mud layers below might reveal the deep history of these woods and offer insights into its even wilder past. That the idea came from my dog Åsa was apt, since she was named for the daughter of a Swedish friend and fellow paleoecologist who would have relished the scene and the direction of my thoughts. If we carefully analyzed the mud for the remnants of the plants, bugs, and other materials that accumulate over time, the pool might reveal answers dating back hundreds or even thousands of years. We might get lucky and be able to complement the early studies of Spurr and others with some innovative paleoecology. But a dose of reality dashed that flight of scientific elation. Åsa had beelined to the center of the tiny pool. Wouldn’t every moose, wolf, and deer that passed this way through prehistory have sought similar refreshment? Wouldn’t the destruction have escalated during recent centuries when most woodlots were grazed by cattle, sheep, and hogs? The scene of large trees and thick mossy groundcover had a pristine appearance, but the mud in this small pool was likely a churned-up soup. Still, it seemed worth exploring, even though the effort turned out to be both challenging and comical. The pool was only about twenty feet across and held just a few inches of water, but the only way we could establish a solid and stable platform over the mud surface was to deploy our standard and quite unwieldy equipment— an immense pontoon boat fashioned out of two seventeen-foot-long canoes bridged by an eight-foot sheet of marine plywood lashed across the thwarts. The scene of this unwieldy nautical vessel sitting in the deep woods was made more ludicrous by the fact that our craft nearly spanned the entire depression. Defying the self-conscious feelings of pursuing a blatantly foolish task, we completed our work with spirits buoyed by the discovery and retrieval of nearly a meter of dark oozy sediment. Back in the lab, however, disappointment settled in as graduate student Tad Zebryk processed the mud and began to examine the resulting microscope slides for pollen. His scans revealed no variation between the various samples: each level had more or less a similar plant composition. Even more troubling was the nature of the vegetation that the pollen counts revealed. Through much of the length of the core he found ragweed, a common weed of agricultural fields whose appearance in large numbers we conveniently use to identify the onset of colonial clearing of the New England forest. My fears based on Åsa’s romp in the pool appeared to be well justified: the core was clearly one homogenous mess. LESSONS: JOHN SANDERSON’S WOODLOT 210

A sediment-filled coring tube and other gear adjacent to Hemlock Hollow in the old Sanderson woodlot. (David Foster)

In a fairly desperate appeal for solutions and alternative explanations as we examined these discouraging results, I asked Tad what he thought of these clear indications that the mud had been thoroughly mixed. As if jolted into a new mode of concentration, his brow creased, and then his finger extended toward the wrinkled and mud-stained map of the coring site that was sitting on the table. Brown-streaked fingerprints obscured parts of the outline of the basin and the concentric lines that marked each contour at ten centimeters (nearly four inches) of depth. I had been impressed when Tad produced this bathymetric map and wondered how he’d done it. Now he sheepishly revealed that, since there was no easy way to probe the pool, he had simply put on hip boots and waded across the depression in a series of straight parallel lines, noting the depth of water and mud every meter. It hadn’t dawned on him that that we might end up coring literally in one of his footsteps. In no time, we had loaded the pickup with canoes, plywood, rope, and coring gear and were on our way back toward Åsa’s hollow. By following our detailed notes and positioning our platform just a few feet west of the original hole, we were able to retrieve a new meter of mud from a location that we hoped had escaped Tad’s well-intentioned but destructive transect. Back in the lab we were heartened by the appearance of discrete fine black layers of charcoal in the green matrix as we sliced the core lengthwise. Some weeks later, Tad’s speculation was confirmed by detailed graphs from one section of the core that depicted highly resolved fluctuations in mineral matter, charcoal, and the pollen of many different plant species. When radiocarbon dates arrived back from the lab, each of the dates fell into nice chronological order, and the truly ancient nature of the record was revealed. We were euphoric. At nearly 10,000 years old and comprising minute particles of plants, soil, and charcoal from within fifty feet of the tiny pool, this core and its record eventually became recognized as one of the most detailed and spatially resolved histories of the New England landscape. Quite remarkably, it portrays the dynamics of a single forest in which hemlock thrived and was challenged for more than 8,000 years. Our perseverance in coring brought us important discoveries, some major publications, and the need for a more proper designation of the site—as Hemlock Hollow. The story from that mud is told elsewhere in this volume, so the details need not be dwelled upon. The distinctive features emerging from this Sanderson Woodlot site, however, are worth highlighting. The long, continuous record has a highly unusual quality: its local scale. This results from the pool’s being surrounded by trees whose pollen has rained onto its surface, overwhelming any inputs from the much more distant sources that are represented in most ponds, lakes, and wetlands. By matching this local record with the somewhat broader picture obtained at the Black Gum Swamp 100 yards away, we were able to place the hemlock forest within a larger context of regional-scale vegetation and climate change. LESSONS: JOHN SANDERSON’S WOODLOT 212

The prominent message from this record was that hemlock has been the predominant tree in John Sanderson’s woods since it first migrated north to New England 10,000 years ago under a warming climate. A second notable fact was that whenever major disturbances hit this site, altering the forest in major ways, hemlock always recovered and reassumed its dominant role. In each case, regardless of whether fire, drought, insects, or people were involved, hemlock relentlessly returned to form what must have been an impressive scene of towering ancient trees— hemlock, pine, birch, oak, and magnificent spruce and black gum that would have dwarfed and darkened the tiny woodland hollow. This process of recovery was excruciatingly slow in human terms, each instance requiring 500 years or more before hemlock assumed its greatest abundance and settled into its dominant role for the next thousand years or more. This record provides a new perspective on the forest that we walk through today. In many ways, the Sanderson woodlot has the appearance of an ancient forest. Yet the larger trees just barely exceed 200 years in age, and the pollen record tells us the forest is still less than midway in recovery from the colonial disturbances wrought by John Sanderson and his kin. Given the widespread emergence of the hemlock woolly adelgid throughout the stand in 2012, the forest will never reach an old-growth condition this time around before this new dynamic of hemlock death begins. We can look to the midHolocene hemlock decline 5,500 years ago for perspective on how this adelgid episode may play out. The optimistic message emerging from the ancient script is that hemlock has always recovered from past devastating blows, so there is strong likelihood that the species will recover from this new one. The sobering news is that, following that great prehistoric decline, it took hemlock nearly 2,000 years to regain its former abundance. Finally, there is a truly wonderful message that emerges from this lengthy story of Åsa’s hollow. It is a message about the role of serendipity in science. The studies that Richard Fisher, Bob Marshall, Earl Stephens, and Tad Zebryk pursued were all based on collecting every last scrap of historical evidence that can be gleaned from any available source about a tree or site or landscape and its changes over time. Each such episode is always a novel pursuit. The different forms of evidence and the nature of the resulting information and its messages are never known at the outset. In this kind of historical and ecological research, any source is fair game, and the boundaries are only limited by one’s imagination and the quirks of history. In some cases such as the Pisgah Forest or the Sanderson woodlot, the site itself reveals most of the story, told by peculiar and often unanticipated sources such as cut stumps, downed wood, buried soils, and the mud in a woodland depression. In other cases, someone comes across the notes, samples, and charts left by Harvard Forest predecessors in the official archive or an attic or bookshelf and realLESSONS: JOHN SANDERSON’S WOODLOT 213

izes that they contain unmined gems of information. Who would have guessed that three graduate students in the 1920s would have preserved a record of tree growth from clear-cut old-growth forests at Pisgah, in the form of a series of paper strips stuffed in a yellowed envelope? Sixty years after these penciled marks captured every decade of growth in those stems, they told a compelling story of age-old growth and release after disturbances. In another instance, a grandson of Richard Fisher single-mindedly teamed with his mother to collect all her father’s writings from more than twenty far-flung relatives. Together they sifted through the voluminous haul and published a comprehensive volume of Fisher’s exchanges and thoughts on everything from the influence of Abbott Thayer on his work, to his insights into redwood forests, to the challenges of working with different Harvard administrations. Another example: a heavy oversized volume of handwritten pages that has sat for decades on a bookshelf in our archives was known generally to contain daily notes made by one of the early classes at the Harvard Forest, back when our residential “Community House,” the old Sanderson farmhouse, was the heart and center of the entire enterprise. The entries are dated, and the volume has been leafed through by dozens of bored or aimless scholars looking for a bit of diversion over the years, but it took a historically oriented and ecologically aware scholar who knew more than a little bit about Harvard Forest alums to recognize the writing in the volume and the insights that it holds. The tight script is Bob Marshall’s, and the journal turns out to capture a single and singular year in Harvard Forest history—the initiation of the greatest experiment in the institution’s history and the year when Marshall forged the Harvard Forest approach to forensic ecology. That record stimulated a new focus on Marshall that led to the recovery of his plot and the discovery of his role in developing historical approaches in ecology. So there is a lot of serendipity in our science, as we find unusual records that no one ever thought of seeking or even guessed were there. After all, who would have expected an unbroken 10,000-year record of forest growth and death to emerge from a small pool in the woods that had been ignored for decades and was likely disturbed by animals, humans, and forest processes? And how have so many books been written about the founder of the Wilderness Society without a single writer stumbling upon the treasure trove of personal writings, data, and unpublished photographs that have sat for decades in Petersham and that capture a joyous, inspirational period in the young man’s personal and professional growth? Therein lies the other side of our science coin. To seize on serendipity requires insight as well as a bit of perseverance. In their perennial quest to use colonial records to understand the nature of the early New England landscape, historians and scientists overlooked the fact that the notations of which trees served as boundaries and corner markers provide LESSONS: JOHN SANDERSON’S WOODLOT 214

The view from Sanderson’s farm over Shaler Hall and west toward the Quabbin Valley. (David Foster)

an unbiased record of the composition of the forests before they were disturbed by the new immigrants. It took imaginative ecologists to recognize this—Tom Siccama at Yale was one of the first, followed by Gordon Whitney, who built greatly on that effort during a year at the Harvard Forest, where he compiled the most exhaustive volume on the history of the northeastern forest. It has now taken Charlie Cogbill nearly three decades to track down and compile the comprehensive set of records from every nook and cranny in town offices across the Northeast, which are referenced in this volume. And it has taken Jonathan Thompson’s quantitative abilities, computer skills, and ecological expertise to fashion the 325,000-plus tree record into a revealing story of four centuries of change. Serendipity is key, but seldom do historical gems fall right into your lap, or announce themselves on the other end of a telephone line. Although that too can happen.

It was the autumn of 1999. The voice on the phone introduced himself as a dealer in rare books. He quickly followed by stating that the Petersham Historical Society had suggested that the Harvard Forest might be more interested in the book he was selling than they were. He explained that he often picked up boxes of miscellaneous books and old journals at auctions, and that a recent haul had included the account book for a farm in Petersham. Failing to feign indifference, I asked the name of the family. “Sanderson,” he replied. As I sat silent, he continued: “The volume begins in 1775 and runs well into the 1800s. It must cover multiple generations, for it contains a couple of different handwritings. It is organized in the typical style of a farm or business account book with pages listing expenses and others tallying income against the names of various individuals.” Then he played his trump card. “I gather that the Sandersons once owned your land and that you all have some interest in its history.” With this pronouncement I dropped all pretense of indifference and got right down to business. In a matter of minutes we had arranged for him to bring the leather-bound volume to Petersham for an assessment of its contents. Then we broached but did not resolve the challenge of establishing a fair value for something that was worthless to the world at large and yet priceless to the Harvard Forest, at least under its current leadership. When he arrived and I began flipping through its well-handled pages filled with notations, the value of “priceless” increased tenfold in my mind. The elongated volume had a soft leather cover that was worn through years of handling by rough hands and felt strangely smooth and cool in my hands. When I recognized LESSONS: JOHN SANDERSON’S WOODLOT 216

the details of the actions and transactions that it contained, the deal was sealed. In blue ink and flowing hand, the inside cover read “Jonathan Sanderson 1775.” Inside, most pages were dense with columns written in a range of cursive scripts by different hands. The entries shared a similar organizational pattern, but at first glance each was indecipherable. I could make out the pattern—names, dates, items, and columns of numbers. With a bit of effort, the words became recognizable. First were names that were familiar from the cemetery just fifty yards down Main Street from where I sat. The nineteenth-century neighborhood was all there: Mann, French, Wheeler, Sanderson, and more. Then there were the items bought, sold, or exchanged: cows, cattle, cheese, hay, butter, and bark. Finally, as others joined to share in this impromptu exploration, we began to decipher the activities: the laying of stone walls, mowing fields, and driving cattle to Boston. The language and currency gave pause, because the early entries were strangely British, including the use of pounds, shillings, and pence. But every page was revealing. As I read on, a vanished world began to take form in my mind. This was a world of concrete items, actions, and daily decisions by real people in a distant landscape that I now walk and study every day. It was a world that was previously known only through crude artifacts such as stone walls in the woods, census figures, ancient newspaper accounts, and trees growing in fields now abandoned for human use. In my hands, the account book, composed by the actors themselves, was speaking directly in words and numbers. The deal was done: $750. In no time, Linda Hampson dropped all her other tasks and was poring over the records and struggling with the handwriting. She transcribed and formatted each page on her computer screen so that it mirrored the original, allowing us to concentrate on the content rather than the laborious challenge of deciphering the unfamiliar scripts. We discovered that Linda brought two personal advantages to this effort. Her passion for New England history and antiquities made many terms, actions, and scenes familiar. It also turned out that, when completely stumped by a word, she could take a copy of the page home to her Yorkshire-born husband, who frequently recognized it as slang from the old sod. We learned much about the Sandersons and their farm as the transcribed volume took shape; over time it transformed our understanding of the Harvard Forest land, its history, and the hemlock woodland. Some insights were quotidian, such as the routine purchase and bartering of animals, labor, and farm produce. Others opened our eyes to the unsuspected breadth of the world of eighteenth- and nineteenth-century farm families. Most remarkable was the frequency with which the Sandersons engaged the commercial world in Boston and beyond: selling cheese and butter, purchasing calico and other cloth, tools, and goods such as sugar, and dealing with drovers and neighbors who would share in the task of driving the livestock on the long walk to the Brighton slaughterhouses or on to Boston markets. LESSONS: JOHN SANDERSON’S WOODLOT 217

Some entries opened new perspectives on daily life, such as regular notations of the purchase of draughts at the French homestead. Today, all that remains of this grand inn just over the hill is a sprawling cellar hole at the crossing of two woods roads, largely hidden in low growth beneath a few immense sugar maples. Yet from the descriptions and scattered artifacts around the site, it must have been a magnificent colonial house with ells, outbuildings, and barns that provided both a home for a large family and respite to passing travelers and animals. On the subject of liquor, we were bemused by frequent notations for gallons of rum purchased “for mother.” This raised the speculation that such a notation either served as a general euphemism for any liquor purchase, or it provided a solid indication that the aged mistress of the house spent much of her time curled up in a rocker with her knitting and a bottomless cup. There were other major revelations. Many of the nineteenth-century stone walls that bisect our woods and line the ancient roads were not built by the strapping men of the Sanderson family, as we had always imagined. Rather, the journal’s tallies of work days note weeks spent assembling these walls by a family who rented a separate farm in the town of Shutesbury from John Sanderson. Many other debts were worked off in part through labor in Sanderson’s fields and improving his land. Learning about previously unknown real estate assets, the great number of cattle, oxen, and horses that passed through the farm, and the scale of business that he conducted, we came to recognize that landowners in rural New England such as John Sanderson were not poor dirt farmers scratching a living out of barren and rocky soil. Rather, they were successful and calculating businessmen who thrived in a world where land could be improved through hard labor, and materials were regularly exchanged with neighbors, distant cities, and far-flung parts of the world. Going back to the census, newspaper accounts, and gazetteers, we confirmed that Petersham was one of the most prosperous towns in northern Worcester County. A town leader, Sanderson was considered among the most successful of its citizens. From the journal it also became clear that John was a diversified producer and broker of foods, goods, and services. Beyond farming, he was a small industrialist, as indicated by one gazetteer that stated “much of [Sanderson’s] wealth came from his tannery.” The leather business was inextricably linked to the land, for it depended on hides produced on local farms that were soaked for months in vats of a tanninrich solution derived from the bark residue of hemlocks in local woodlots. John’s wife, Lydia, carried the family and farm enterprise forward with son John after her husband was crushed at age sixty-two in the barn by an ox. They all seem to have been shrewd business operators. From this small tanning industry, they turned a regular profit in hard cash that was then invested in the farm or other land holdings. Hugh Raup concludes his history of John Sanderson’s farm by noting that the LESSONS: JOHN SANDERSON’S WOODLOT 218

Archaeologist Dianna Doucette (lower left) and students excavate John Sanderson’s tannery. (David Foster)

family wisely sold out at the height of land values in Petersham and then purchased a magnificent farm in the agriculturally rich Connecticut River Valley town of Bernardston, Massachusetts, where they started one of the region’s major banks, raised prize oxen, and launched additional careers that included politics for state senator John Sanderson Jr. This success was originally grounded in no small measure in the Petersham hemlock woodlot, which provided one critical raw material and the small stream that powered the tannery. Every town in New England had one or more tanners along with the operations needed to produce enough leather to meet the local demands for work, home, and pleasure. Despite their ubiquity, such small-scale tanneries are poorly understood by historians and archaeologists, and their ecological implications are completely unexplored. When we launched our archaeological excavation of John Sanderson’s tannery in collaboration with Elizabeth Chilton at the University of Massachusetts and Dianna Doucette at Harvard’s Peabody Museum, we were surprised to learn that it was the first attempt in the northeastern United States to excavate a tannery and to learn exactly how such critical operations were constructed and run. The stream provided two key resources–water to wash the gore off the hides and to soften them through soaking, and the energy to drive the immense grindstone that pulverized the hemlock bark so that its rich tannins could be leached out. Brought by the wagonload, hides and hemlock bark converged at the tannery. The skins headed to a separate building for flensing and beaming, a sequence of processes that hand-stripped the fat from the hide and then limbered it up by working it over a wooden beam. The bark was delivered to the mill, where it was crushed by the ponderous millstone that rotated around on its flat and crudely serrated edge to shred the bark, which was then soaked in vats of water with the hides. After we understood these details, the history and use of the hemlock woodlot became much clearer. Jonathan Sanderson, the family patriarch, began to carve his farm from the Petersham wilderness in the 1770s, just decades after the town was founded. Over time his son John followed his lead and expanded the arable and pasture lands out from the Main Street homestead. Because the hemlock woodlot and the Black Gum Swamp were a large distance from the barns and had poor drainage and low fertility, they were among the family’s last acquisitions. In the extensively cleared landscape of the early nineteenth century, however, timber, firewood, and bark became increasingly scarce and valuable commodities. The Sandersons cut the woodlot heavily the first time and then harvested it repeatedly, presumably husbanding the many resources and favoring species that served specific needs for the farm businesses. The chestnut, which sprouted prolifically, was cherished for posts, beams, and other building materials due to its rapid growth, straight grain, and resistance to decay. Oak and pine timber could be readily sold or used as needed. But LESSONS: JOHN SANDERSON’S WOODLOT 220

A large stone used to grind hemlock bark sits on a stone pier adjacent to the stream that supplied the tannery operation. (David Foster)

for John and his major industry, the hemlock was key. Its sustained abundance in the woods must have been a result of deliberate management. While hemlock wood was much less valuable than that of many other species, the bark would have kept the tannery crew busy, including brother Joel, who ran the operation and supervised another three to five men and boys. Our tannery excavation remains in its early stages, progressing slowly as it also serves as one focal activity in our summer research and educational program, which teaches integrated historical approaches to ecology and conservation. But through the efforts of Doucette, archaeologist Tim Binzen, and some superb students, the general layout of the buildings and operation of the site have become clear. A small Cape Cod–style house with numerous ells comprising barns, sheds, a well house, and outhouse sat atop the steeply rising banks of a brook that reaches ten feet across during a spring freshet. Today it is a lovely scene of trees and stone and water, but in the tannery’s heyday the same view would have been bleak and likely nauseating. The land was undoubtedly treeless and bare with cartways and trails eroding the slopes and stream banks. The stench of putrid hides and their scrapings would have filled the air and accompanied the fetid odors emanating from dozens of vats filled with hides soaking in hemlock-infused water. The stream, with its many duties, would have wandered from the marsh above through an open pasture filled with cattle, flowed into the millpond and mill, past all the working men and processing areas, and then left the site as an infusion of silt, manure, and trimmed fat, all stained dark brown with tannins. On a steamy August day, travelers on the Athol–Petersham road would have hurried by the site. The grueling effort of running the tannery was very much a seasonal boom-orbust affair. Today the stream seldom runs in the summer once the trees leaf out and the forested watershed begins to evaporate vast quantities of water. In the deforested landscape that John Sanderson knew, this atmospheric diversion of water by trees would have been substantially reduced, yielding more stream flow, but the tanner would still have had to rely on favorable weather and ingenuity to run his mill for a few months each year. Water was stored in two locations. A half-acre millpond sits behind a massive rock dam that spans the valley within view of both the tannery and the miller’s house. A quarter mile upstream, nearly ten acres of marsh is dammed today by beavers that have capitalized on the long, low rock dam that the Sandersons erected and used to manage the large volume of water. By controlling the flow from the marsh, they could have kept the millpond full. The census records from Petersham list three tanneries, with Sanderson’s regularly noted as the most productive. Joel Sanderson and his crew processed about a thousand hides a year, an extraordinary yield given that it could take a year or more LESSONS: JOHN SANDERSON’S WOODLOT 222

of soaking to process a single hide. The operation would have required an extensive complex of vats and processing capability, for which our archaeological foray is just beginning to account. At the same time, this level of production would have consumed an immense amount of hemlock bark and hides, likely far more than could have been produced from the farm alone. Consequently, in this business, as in most of their other enterprises, the Sandersons engaged in a constant stream of transactions with many people. This commercial venture brought considerable cash to the farm and helped keep it prosperous. At its heart lay the fields that produced the cattle, swine, goats, sheep, and calves and the woodlot that harbored a grove of hemlock that endured throughout the tumultuous New England colonial period. From the farm journal, we see the labor of cutting the woods that yielded bark, a bit of timber, and cordwood. In the records from Hemlock Hollow, we see the consequence as repeated harvests turned a diverse old-growth forest into a woodland of sprout chestnut and hemlock. It was the resilience of hemlock and the care of the woodlot owners that enabled hemlock to persist through those years of use and to reemerge as the dominant species long after the farm was sold and chestnut declined. And it was hemlock that made the whole tannery operation work and ultimately enabled John Sanderson, his family, and farm to thrive.

LESSONS: JOHN SANDERSON’S WOODLOT 223

T W E LV E

LAMENT

Welcome are all earth’s lands, each for its kind; . . . The solid forest gives fluid utterances; They tumble forth, they rise and form . . . Shapes bracing the earth, and braced with the whole earth. —Walt Whitman, “Song of the Broad-Axe,” Leaves of Grass, 1900

E

very hike through the hemlocks at the Harvard Forest these days lends to an ongoing voyage in self-exploration. Like all groves of this great species, the environment is still and inspiring: cool in the heat of summer; eerie with a low mist hanging above the snow on a warm winter morning; and thick with darkness cut by shafts of bright light whenever the sun shines brilliant above the dense foliage. The magic of these woods persists, and the awareness of their past as primeval forest, colonial woodlot, and inspiration for creative people of all types adds a deep temporal dimension to one’s walks. In most groves, this overriding sense of history is underscored by the many old and decaying stems of chestnut that still lean awkwardly against the towering hemlock. Each of these resistant stems rests where it died in 1915, when the Asian chestnut blight swept through these stands. But our hemlock woods can also be disorienting and, frankly, bizarre, for they are increasingly filled with scientific apparatus and ecological debris that is often inexplicable, even to those of us who should have a good command of all the studies transpiring across this property. Affixed throughout the landscape are bits of luminescent flagging and scattered posts of wood, metal, and painted PVC. In broad swaths of forest, each tree is marked and identified, often with numbered aluminum tags hanging loosely from nails driven partway into the wood. Across the eightyacre “big grid” of permanently mapped and regularly censused trees, every stem is also anonymously blazed with a horizontal splash of yellow paint at “breast height.”

224

These marks signify that the stem’s vital details have been recorded and are duly stored in a digital database and available to anyone, anywhere, with access to the Internet. They also ensure that when the next census of the tract is made, each tree’s girth will be measured at exactly the same place. High above these blazed hemlocks, the energy, moisture, and gases that enter or leave the woods are being measured through a baffling array of devices. The vantage point atop our eighty-foot-high observation tower provides the most unusual and engaging view in the region (recorded via hourly digital images) of varying treetops with their leaves, flowers, and fruit, the arched leaders of hemlock, and the distant sight of the fire tower atop the tallest hill in the region. Back on the ground, low square boxes, eighteen inches across and fashioned of stainless steel, automatically tilt open from a hinged side at regular intervals to sample all the air that emerges from below as the soil breathes. The forest, the soils, and the streams are all gauged and monitored; snow depths are measured on a ridiculous-looking eight-foot-square inflatable mattress appropriately called a “snow pillow”; and thick electric cables snake through the woods distributing power to the various studies. Naturally, all the data are streamed back to base seamlessly via radio waves and a wireless Ethernet. This system connects every device and carries photographs from the “canopy cams” as well as the motion-sensitive cameras that capture unsuspecting animals (and people) moving through their unblinking fields of view. Our woods are now filled with such apparatus, deployed to monitor the unfolding of an event akin to the one that killed those decaying chestnut trees a century ago. Back then, a single student designed a presumably comprehensive study of plots arrayed across the landscape to capture the expected subtle variation in chestnut’s response to an unknown fungal blight. This well-conceived effort was rendered useless when the disease blanketed the region and killed all trees in all settings, quite oblivious to any potential mitigating factor. This time, with a mixture of historical perspective and twenty-first-century hubris, we are poised and anxious. We are watching an ecological drama unfold, albeit at an agonizingly slow pace. We are documenting the latest chapter in the epic story of a great species, and we do so with mixed feelings of scientific engagement, selfconsciousness, and lament. We balance emotion and a sense of inquiry, aware of lessons learned and the expected future to come. In our explorations of hemlock we have come to realize much, and in the process we have also lost much. But we are reconciled to our forest’s fate, as we also seize on glimmers of hope for a bit longer association with hemlock as it fades from the New England landscape. While the great stands in the Smoky Mountains fell within a seeming instant, the adelgid is struggling—in vain so far—to move north much beyond Massachu-

LAMENT 225

The arching leader of a hemlock, photographed from the Hemlock Tower seventy-five feet above John Sanderson’s woodlot. (David Foster)

setts. Yet given a warming climate and the capabilities of an enigmatically evolving insect, it is only a matter of time before the insect spreads and northern stands begin to decline. Entomologists hope to locate a predator in Asia that can control the insect here, but so far there has been no sustained success, and there is growing and well-deserved fear concerning the unintended consequences of adding even more exotic organisms to our landscape. Some signs of resistance are being found in scattered trees, and it may be possible to breed or engineer even more, yet it is staggering to contemplate how we might begin to replace the forests that already have died across thousands of square miles. In the face of such loss and despair, some enthusiasts advocate protecting at least a few stands of hemlock through radical campaigns of insecticide application. At the same time, many agencies, organizations, and landowners already have cut their forests preemptively and intensively, entailing a more drastic impact, all in the name of management and an attempt to minimize the consequences of the adelgid. What will this loss mean? In the late 1980s, shortly after the adelgid arrived in southern New England, I received a call from a Nature Conservancy staff member in Connecticut who was helping the organization formulate a campaign to deal with the threat looming in the region. It was a notable conversation, as my responses simultaneously surprised and disillusioned the young woman. “If hemlock dies, what will happen to biodiversity on these sites?” “Well, I expect that it will go up, probably greatly.” “Oh, OK. How about streams? What will happen to the number and abundance of species in these pristine habitats?” “I’d expect those to show major increases as well, what with greater light, nutrients, and more readily decomposed foliage.” “Well, then,” she said, now clearly probing for anticipated tales of calamity, “what about the productivity of the woods and water?” “That should certainly increase in each of these habitats.” “Hmmm,” she paused. “So . . . is this a good thing? What do we actually think is the problem?” “Ah,” I responded, smiling as I recognized the dilemma that this situation posed. Traditional conservation values diversity in natural systems and seizes on compelling story lines in its public campaigns. How does one argue the value of a slow-growing tree of marginal commercial value that excludes the bulk of species in the region from the dense, dark habitats it creates? In the face of that challenge, I said: “This is most definitely a tragedy. We will lose hemlock. We will lose our flagship for old-growth and primeval forest in the LAMENT 227

Students disappear down a trail in the Prospect Hill hemlock forest. (David Foster)

Northeast. We’ll lose distinctive variation in our landscape. And we will lose the history and experiences embodied in these woods.” There you have it. In many ways—and especially with regard to many of the values that many conservationists, scientists, landowners, and wood producers bring to their work—the loss of hemlock can be easily seen as a good thing. More plants, insects, and animals will occupy these sites. Productivity will go up. More valuable tree species may become established. The woods and waters will experience conditions unimaginable when hemlock prevailed. And yet we will have lost a major force of variation in our landscape. Five thousand years ago a similar event occurred, and it appears that the human population flourished in a manner not seen before or after. The combination of warm conditions with the absence of one of the landscape’s most dominant trees encouraged nut-bearing species to flourish, which in turn supported more wildlife, plant life, and people than ever before. The range-wide destruction of hemlock may well have been a positive force in the lives of many in that ancient world. Nonetheless, the coinage of conservation and human value systems is clearly inadequate if the loss of this majestic species translates into a positive. Hemlock imparts variation to the land and life. It lends distinctive qualities to environments in which only a few species thrive. These areas stand out in our landscape. In its solemn shadows, hemlock provides specialized ecological niches, creating opportunities for just a handful of plants and animals. One of the species that responds positively to this dark, solemn tree, at least in an emotional and intellectual manner, is us. As we walk among their stems we relish these massive trees and their dark woods. As with so many species, we use these woods but do not take root there. Like our Indian predecessors in New England and the deer that abound here today, we delve into but do not live among these trees. Hemlock woods are special, and they stay with us even though we do not stay with them. Daily, I walk out to that old, solemn, and gaudily bedecked hemlock wood to savor it one more time. Nowadays the adelgid is also there, everywhere, and in massively unavoidable numbers. While the forests remain grand and inspiring, I am beginning to sense that the familiar and fateful shadow of gray is appearing in the needles, and the canopy is a bit less firm and dense. I may be imagining it, but we are definitely coming to the end. Yet there remains more to explore and much to experience. We can relish the time we have spent with the tree, and we can hearken back with full-hearted appreciation to all those early Harvard Foresters’ exploits with hemlock. We can also envision a life to come in which these deep, dark, and depauperate hemlock forests are gone, trusting that they may return again. LAMENT 229

This page intentionally left blank

B I B L I O G R A P H I C E S S AY S

P R E FAC E David R. Foster

The writing in this volume was inspired and informed by research conducted by many people over more than a century at the Harvard Forest. All the published literature from these studies is available online in the Harvard Forest bibliography on the Harvard Forest website (http://harvardforest.fas.harvard.edu/research-publi cations). Documents about hemlock can be accessed through a keyword search. The relevant archival documents, photographs, and data supporting all our research are increasingly accessible in various online catalogues from the Harvard Forest archive through the same website. The Harvard Forest dioramas in the Fisher Museum convey the broad elements of the story of forest history, management, and conservation in New England in a readily accessible manner. For readers who cannot make the trip to Petersham or desire high-quality images of the sequence of changes in our land, the dioramas and their story are accessible online (http://harvardforest.fas.harvard.edu/fisher-mu seum) and in the short book New England Forests through Time (Foster and O’Keefe 2000). C H A P T E R 1 . H E M LO C K ’ S F U T U R E I N T H E CO N T E X T O F I TS PA S T David R. Foster

Delving into the literature on the history, ecology, and dynamics of hemlock is an enjoyable pursuit that can lead in many directions. Overviews of the historical and postglacial changes of the New England landscape are available in numerous formats, ranging from specialized peer-reviewed journal articles and more accessible syntheses to highly illustrated overviews intended for general audiences and visitors to the Fisher Museum. The most complete overview, which also shows the importance of these dynamics for the broad range of ecological studies conducted

231

at the Harvard Forest, is presented in Forests in Time (Foster and Aber 2004). In this volume synthesizing the results of our Long Term Ecological Research (LTER) program, we address environmental and human history, broadscale forest response to land use and climate change, landscape dynamics, and associated changes in many species of wildlife. An even more detailed series of overviews are provided in the 2002 special issue of the Journal of Biogeography, which was devoted to studies from the Harvard Forest, including articles addressing the nature of the pre-European forest (Cogbill et al. 2002); land use history and change (Hall et al. 2002); the role of fire in the context of vegetation and human dynamics before and following European arrival (Parshall and Foster 2002); wildlife dynamics that accompanied shifts in land use and vegetation over the past 400 years (Foster et al. 2002b); and the need for a strong understanding in history, ecology, and human activity in developing effective means of managing and conserving nature (Foster 2002). Chestnut, hemlock, and oak have undergone notable dynamics through the postglacial and historical periods in New England, which have been examined by numerous authors in species-specific and comparative studies. Fred Paillet has evaluated the life history, ecology, role, and dynamics of chestnut throughout its range and through comparative studies of chinquapin in the southeastern United States as well as European chestnut in that continent (Paillet 1982, 1984, 1988, 1993, 2002; Paillet and Rutter 1989). Fred’s work has included a specific focus on New England and has involved detailed field studies and many discussions at the Harvard Forest. The longer-term dynamics of chestnut in relationship to fire, human activity, and environmental change over the past 10,000 years in the upland oak-dominated forests of central Massachusetts have also been investigated (Foster et al. 2002a). As described in chapter 6, the longer-term dynamics of hemlock have been a focus of Harvard Forest researchers since Harvard alumna Margaret Davis identified the range-wide decline in the species 5,500 years ago and attributed it variously to a pathogen or pest (Davis 1981a). Janice Fuller examined the detailed features of this decline and the responses of associated species in her thesis (working with Keith Bennett) at Cambridge University, focused on forests in Ontario, Canada (Fuller 1997, 1998). At the Harvard Forest, Janice worked with us to explore both these ancient dynamics and the shifts in hemlock abundance over the last 1,000 years that were driven by the combination of Little Ice Age changes in climate and European settlement in New England (Fuller et al. 1998, 2004). Expansion of detailed paleoecological studies by Harvard Forest researchers to the southeastern coast and islands off of Massachusetts uncovered the first evidence for the great decline in oak populations 5,500 years ago, which paralleled what was by then known as the mid-Holocene hemlock decline (Foster et al. 2002b). The implications of the similarities in the dynamics of these very different species inB I B L I O G R A PH I C E S S AY S 232

cluded recognition of a strong climate signal as a causative factor in the abrupt decline of hemlock. This discovery and new research direction stimulated a synthesis and comparison of the dynamics in the two trees (Foster et al. 2006), along with much more detailed evaluations of coastal climate and vegetation dynamics (Oswald et al. 2007a, b; 2009, 2010; Oswald and Foster 2011a, b). The most detailed record of hemlock’s change in New England emerged from the study of the small Hemlock Hollow at Harvard Forest (Foster and Zebryk 1993), which was an outgrowth of Tad Zebryk’s thesis. Field-based ecological studies by our predecessors at the Harvard Forest have been especially notable in advancing the understanding of hemlock. These include work on hemlock in the context of broader forest associations at the Harvard Forest and following the 1938 hurricane (Spurr and Cline 1942; Cline and Spurr 1942; Spurr 1950, 1956a, b); close study of the morphology of hemlock and detailed examination of its growth into the canopy (Hibbs 1981a, b; 1982); and a series of papers on the development of mixed hardwood and hemlock forests by Oliver (1975, 1978; Oliver and Stephens 1977) and Kelty (1984, 1986, 1989; Kelty and Gould 1987) that included evaluations of the influence on forest productivity of hemlock’s inclusion in these stands. The details of hemlock’s interactions with the hemlock woolly adelgid are largely based on the extensive work over the last fifteen years by Dave Orwig and his many students. C H AP T E R 2 . AN I CO N I C S PE C I E S David A. Orwig

Much has been written about hemlock over the years outside of the world of science, especially in the field of literature. The tree’s distinctive characteristics and habitat produce an ethereal quality that makes them a common focal point for both prose and poetry. Nowhere are the qualities of hemlock more stunning than in old-growth forest settings. Abrams and Orwig (1996) describe the cathedral-like canopy of towering hemlocks and white pine in northwestern Pennsylvania, while Orwig et al. (2001) and D’Amato et al. (2008) characterize old-growth hemlock in southern New England. The general branching, tree form, life history characteristics, and wood products derived from hemlock are well summarized in a series of early twentieth-century publications, complete with beautiful black-and-white photographs, including Frothingham (1915), Merrill and Hawley (1924), and Clepper (1944). Our understanding of the wood qualities and defects associated with hemlock (e.g., ring shake) are well described in a Harvard Forest paper by our colleague and former Harvard Forest scientist Bill Wilson (1962) and in McWilliams and Schmidt (2000) and Brown and Sendak (2006). B I B L I O G R A PH I C E S S AY S 233

One interesting early description of the deep shade and acidic litter cast by hemlock was written by the eminent ecologist Rexford Daubenmire (1930) while an undergraduate student at Butler University. Slow decomposition of hemlock litter compared to hardwood species (Kizlinski et al. 2002; Cobb et al. 2006; Cobb 2010) and its characteristic needle chemistry (McClure and Hare 1984; Ingwell et al. 2009) also help create the characteristic spongy organic layer found in many hemlock forests. With the advent of new technology, more sophisticated instruments are available to monitor light levels and to assess how trees capture carbon, respire, and use water. A series of articles led by Harvard Forest ecophysiologist Julian Hadley (Hadley 2000a, b; Hadley and Schedlbauer 2002; Hadley et al. 2008; Daley et al. 2007) and graduate student Sebastian Catovsky (Catovsky and Bazzaz 2000; Catovsky et al. 2002) provide detailed information and the bulk of our current understanding of the physiological characteristics of hemlock, including seasonal patterns of carbon storage and water use. Earlier work on the water-conducting anatomy of hemlock (Ewers and Zimmermann 1984) yields insight into the limitation of water movement in hemlock relative to that of hardwood trees. Hemlock’s amazing ability to remain alive for decades in the dark understory can be seen clearly when examining increment cores or cut stumps in hemlock forests. Early work by Bob Marshall, described in this volume and in the Harvard Forest Bulletin (Marshall 1927), highlighted the extreme slow growth in forest-grown hemlocks compared to those growing in the open. Later work by Abrams and Orwig (1996), McLachlan et al. (2000), and D’Amato and Orwig (2008) corroborated these findings in increment cores from mature and old-growth hemlock stands. One intriguing aspect of hemlock’s life cycle that has not been previously documented is its ability to layer—or vegetatively regenerate—from its branches. A few years ago, while hiking and admiring the old-growth forests on Wachusett Mountain with Bullard Fellow Peter Del Tredici from Harvard’s Arnold Arboretum, we made a fascinating discovery. On an upper slope of the mountain, about 100 yards from an old-growth hemlock stand, we found a central hemlock stem surrounded by a ring of smaller hemlock stems, each of them having a distinct sharp bend upward from the soil surface. Upon carefully removing the upper surface of the soil and organic matter, we discovered that former lower branches extending from the central stem had been covered with needles and soil over time; these branches eventually had formed roots where the soil accumulation was deepest. At this location the lateral branches, extended upward, formed the unusual bend to the stems we encountered. We documented that the central stem was 200 years old, and that the nine surrounding stems (ranging from 61 to 160 years old) had originated from branches of this old hemlock. We surmised that the open, high-light conditions present when the hemlock was young encouraged the extension of lower branches B I B L I O G R A PH I C E S S AY S 234

that would normally self-shade and eventually prune off. After further inspection, we recorded and measured eight additional locations on Wachusett with evidence of hemlock layering. From the earliest descriptions, one often noted quality of hemlock stands is that the abundance and cover of understory vegetation is much lower than in adjacent hardwood stands. Studies examining a wide variety of hemlock forests consistently show low species richness and abundance of understory vegetation (Rogers 1980; D’Amato et al. 2009; Orwig et al. 2012). Despite the low species occurrence, however, studies incubating soils from hemlock forests in a greenhouse show that the soils are often rich repositories for buried seed that can respond to disturbance by producing an abundance of seedlings from species not seen in closed-canopy conditions (Catovsky and Bazzaz 2000; Yorks et al. 2000; Sullivan and Ellison 2006; Farnsworth et al. 2012). Dugan (2004) and Orwig (2008) provide fairly complete overviews of the abundance of animals that take advantage of the unique habitat that hemlock creates on land and in nearby aquatic ecosystems. Several studies highlight the abundance of salamanders (Brooks 2001; Mathewson 2009) and other soil invertebrates in these sites (Dilling et al. 2007; Mallis and Rieske 2011; Sackett et al. 2011). Benzinger (1994a, b) provides an excellent overview of hemlock as bird habitat, and Garrett (2002), Tingley et al. (2002), Becker et al. (2008), and Allen et al. (2009) offer useful field studies providing data on individual bird species abundance. Similarly, Yamasaki et al. (2000) provides a good summary of the various small mammals and carnivores that inhabit hemlock forests. There are only a few studies that have examined invertebrates in streams flowing through hemlock forests (Snyder et al. 2002; Collins et al. 2007; Willacker et al. 2009), and only one study has examined fish communities associated with hemlock ecosystems (Ross et al. 2003). Early descriptions of brook trout as “hemlock trout,” however, suggest the close association between this species and hemlock habitat (Lose 1931). While many advances have been made in remote sensing technology, it is still beyond the capacity of most sensors to yield accurate maps of particular species distributions across the landscape. Painstaking analyses of aerial photographs remain the best approach for mapping hemlock. Although there are no broad, speciesspecific maps of hemlock in the eastern United States, several latitudinal or longitudinal studies focus on hemlock distribution across this region, including New England (Rogers 1978; Orwig et al. 2002, 2012; Evans et al. 2011). In addition, by using remotely sensed images and examining the relationship between hemlock presence and landscape terrain variables, several studies have mapped hemlock within particular sections of the landscape (Royle and Lathrop 1997; Bonneau et al. 1999; Pontius et al. 2005; Koch et al. 2006; Narayanaraj et al. 2010; Clark et al. 2012). B I B L I O G R A PH I C E S S AY S 235

L E SSO N S F RO M H A RVA R D F O R E S TS A N D E CO LO G I S TS : I . T H E PI SG AH F O R E S T David R. Foster

Henry Thoreau’s personal explorations of old forests and wild landscapes were extensive and ranged from his local forays in Inches Woods and detours through north central Massachusetts to multiple trips up Mount Monadnock and through northern Maine (Foster 1999). Of course, Thoreau’s greatest contribution to the understanding of wild nature came through his recognition that wildness lies within and around us all. The full history of Richard Fisher’s days in Dublin, New Hampshire, and the insights that he acquired into nature, art, and conservation at the side of his uncle Abbott Thayer remains to be told. What is clear is that Thayer was an influential mentor and friend who strongly shaped his nephew’s thinking and natural history abilities. Thayer, always eccentric, mercurial, and prone to mood swings, pursued his passions for wildlife (especially birds) and art throughout a productive career that included extended residence in Paris, close friendships and collaborations with Daniel Chester French and George de Forest Brush, and eventual contributions to the permanent collections at the Freer Gallery of Art, Metropolitan Museum of Art, National Academy of Design, Smithsonian American Art Museum, and Art Institute of Chicago. In a telling commentary on modern artistic sensibilities, it has been Thayer’s angelic works, often using his own children as models, that have found their way most consistently into the nation’s great museums. One notable accomplishment emerging from Thayer’s twin passions of art and nature study was his work on camouflage and its potential adaptation for military purposes (Meryman 1999). This effort culminated in a volume with his son Gerald, Concealing Coloration in the Animal Kingdom, and references to Thayer as the “father of camouflage.” Richard Fisher’s mother, Ellen, a gifted artist who, according to family depictions, rivaled her famous brother in skill and sensitivity, contributed to the family support by selling decorative greeting cards. Fisher’s family, from whom many stories of his upbringing and Thayer’s influence come, maintained a family home adjacent the Harvard Forest in Petersham until the 1990s, when they donated the residence and hilltop land to the Forest for housing of undergraduates and visiting groups. Photographs, field notebooks, and correspondence from the Harvard Forest archive enable us to reconstruct the early trips by Fisher and others visiting groves of large trees and old-growth stands in the area around Richmond and Winchester, New Hampshire. The Pisgah Mountain region ultimately attracted his most focused attention. When the best stands were threatened by logging, he succeeded B I B L I O G R A PH I C E S S AY S 236

in protecting the most magnificent twenty-two acres of old growth, supported strongly by the Society for the Protection of New Hampshire Forests and its president (“Forester”) Philip Ayres. Photographs of marked trees and the presence of ancient stumps ringing that tract reveal how real the threat of logging was in the 1920s. Astutely, Fisher reached out nationally in his effort to protect the Pisgah wilderness and was rewarded with broad and deep support. The New York Times and newspapers as far as California covered the story of “The Virgin Forest Our Pilgrim Fathers First Saw,” as one newspaper headline hyperbolically described it. Fisher wrote personally to many friends, including Harry James, whose letter in response is one of many in our archives. Along with one check for a thousand dollars from a passionate female supporter came dozens for ten or twenty dollars, each accompanied by grateful letters of support and encouragement. One envelope arrived with a single dime from a boy in Illinois, accompanied by a note that brought a heartfelt personal response from Fisher. Fisher’s love for and appreciation of the scientific lessons emerging from oldgrowth forests is apparent in his correspondence, writings, and photographs. Two of the Harvard Forest dioramas draw directly from this passion and those lessons: the 1700 Pre-Settlement scene and the Old-Growth diorama that depicts Fisher, his dog Johnny, and Nathaniel Shaler in the old hemlock–white pine forest alongside Harvard Pond. Published writings by Fisher that draw from his work at Pisgah are found in the introductions to some Harvard Forest Bulletins and his essays on New England forests, such as “A Museum of Forest Antiquity” (Fisher 1927), “Soil Changes and Silviculture of the Harvard Forest” (Fisher 1928), and “New England Forests: Biological Factors” (Fisher 1933). His love for the old forests and the scenes and wildlife visible along Harvard Pond led him to designate this area as the first wildlife sanctuary at the Harvard Forest, and led his students to place the Fisher memorial—a bronze plaque in an immense boulder—at this site that inspired the Old-Growth diorama. Not all have appreciated Fisher’s vision, however. A lapse occurred a generation after his death, when the Harvard Forest leadership altered the layout and structure of the Fisher Museum. To increase efficiency and economize on space in the early 1970s, they arranged to drop the ceiling by thirty feet, add a second floor above the dioramas, and carve the cavernous space into a lecture area and two smaller alcoves. While the overall utility of the space was increased, the grandeur of the museum was greatly diminished. There was also a symbolic negative consequence. The riveting view of the large Old-Growth diorama that was originally received upon walking in the museum’s main entrance is now blocked by two large columns that support the second floor. The central focus of the Fisher Museum—the lesson of learning from the history of the land and natural forests—is largely obscured. B I B L I O G R A PH I C E S S AY S 237

Steve Spurr completed his thesis, “Stand Composition in the Harvard Forest as Influenced by Site and Forest Management,” at Yale in 1950. Publications that derived from that effort include “Forest Associations in the Harvard Forest” (Spurr 1956a) and “Natural Restocking of Forests Following the 1938 Hurricane in Central New England” (Spurr 1956b). After a few years as assistant director of the Harvard Forest, he moved to the University of Michigan, where he compiled his classic text Forest Ecology. This volume has been used by generations of students in forestry and ecology and has been expanded in subsequent decades by Burton Barnes, who had become Spurr’s coauthor, and then further developed by Barnes’s own students. Spurr eventually left Michigan to become the president of the University of Texas, where he presided over what has been described as a tumultuous period in the institution’s history. The first exposure that most ecologists had to the Harvard Forest approach to applying historical technique to ecology came with Henry and Swan’s (1974) publication “Reconstructing Forest History from Live and Dead Plant Material: An Approach to the Study of Forest Succession in Southwest New Hampshire.” While the approach to forest reconstruction that Henry and Swan describe did originate with Bob Marshall and was subsequently perfected by Earl Stephens, neither of these two early scientists did much to make their innovative work and methods known. The Henry and Swan study is also notable for two peculiarities. Given that it focused on a forest that was already known to have established after a major disturbance in the 1600s and lay prostrate when it was studied just three decades after the 1938 hurricane, its central conclusion that natural disturbance is important in the life history of forests seems quite apparent. The selection of the specific site for study within the heterogeneous landscape also strongly conditioned the results. The plot that was reconstructed sits on the exposed ridge at Pisgah that had supported the largest trees in the former old-growth forest and then was entirely blown down. A sample elsewhere would have yielded much less downed wood and fewer tip-up mounds, and might have led to the conclusion that the old-growth forests were less majestic and suffered less disturbance. A second enigma is that, despite the paper’s robust conclusions, the results section contains a number of striking inconsistencies. One example: The majority of the trees in the plot and across the Pisgah landscape blew down in a westerly or northwesterly direction coinciding to the peak hurricane winds (Foster 1988b; Foster and Boose 1992). Nonetheless, the paper has a convincing but incorrect diagram of eighteen stems blown down from the 1938 storm lying in a southwesterly direction. There is one inexplicable enigma in the Pisgah landscape itself. Although we have had excellent success using pollen to reconstruct ancient forest history from the soils of hemlock forests and small vernal pools such as those that abound at PisB I B L I O G R A PH I C E S S AY S 238

gah, Peter Schoonmaker’s (1991) attempts at replicating this effort at Pisgah itself proved fruitless. Despite years of effort (including the subsequent involvement of Fraser Mitchell from Ireland) and work at many promising hollows across Pisgah, we were never able to interpret any of the resulting data. In general, the records exhibited little variation over time and did not display any of the major dynamics that we know occurred at the site: the original development of the great forest of pine and hemlock, its cataclysmic demise with the hurricane, and its replacement by the newly emerging forest of hemlock and hardwoods. This lack of success at Pisgah did not prevent Peter and others from using the same techniques profitably at the Harvard Forest (see Foster et al. 1992) or generating some valuable papers describing the insights to ecology and conservation that come from paleoecology (see Schoonmaker and Foster 1991 and Foster et al. 1990). Thus, we have gained much from Pisgah, but many mysteries remain. C H AP T E R 3. P R E H I STO RY TO P R E S E N T W. Wyatt Oswald, David R. Foster, and Jonathan R. Thompson

For more than a century, backward-looking scientists have analyzed deposits of bog and pond sediments to reconstruct past environmental changes. Von Post (1917) published the first formal study of pollen preserved in peat, launching a field that would be developed in subsequent decades by researchers working on pollen records from peat bogs across northern Europe (Godwin 1935, 1946; Iversen 1944). The relationship between pollen in sediments and vegetation on the surrounding landscape has been explored in various ways, including calibration with empirical data and simulation modeling (Prentice 1985; Sugita 1993, 1994), yielding the understanding that most of the pollen grains accumulating in peat bogs come from nearby vegetation, whereas in the case of bigger lakes there is a larger source area for the pollen. Ever the keen observer of natural processes, Thoreau described lakes as “pollenometers” and explained in his journal that “lakes detect for us thus the presence of pine pollen in the atmosphere” (Foster 1999). However, collecting sediments from lakes presents a greater logistical challenge than sampling peats, and thus it wasn’t until the development of lake-sediment coring devices by Livingstone (1955) and Wright (1967) that it was possible to take advantage of lakes as archives of landscape-scale environmental and ecological processes. Dozens of lake-sediment pollen records have been analyzed for sites in southern New England, beginning with the work of Deevey (1939, 1951) and Davis (1969) and including recent work at the Harvard Forest (Faison et al. 2006; Foster et al. 2006; Lindbladh et al. 2007; Oswald et al. 2007a,b, 2009, 2010; Oswald and Foster 2011a). Tom Webb and his students from Brown University were among the first researchers to recognize the value of assembling and analyzing pollen data in networks involvB I B L I O G R A PH I C E S S AY S 239

ing many study sites (Bernabo and Webb 1977; Overpeck et al. 1992; Williams et al. 2001). Such efforts have enabled the reconstruction of the postglacial migration of the tree species of eastern North America (Huntley and Webb 1989; Bennett 1985), and the unexpectedly rapid migration rates of trees as revealed by the pollen data that has been termed “Reid’s paradox” (Clark et al. 1998). In addition to these insights into past vegetation composition, lake sediments can also be used to study past climate. Reduced water depth during periods of drought is recorded in the sediment as a layer of sand (Shuman et al. 2001; Newby et al. 2009), whereas temperature shifts can be reconstructed via chemical analyses of the sediment (Huang et al. 2002; Hou et al. 2011). These analyses have revealed that New England has experienced both gradual and abrupt environmental changes over the last 15,000 years. Climatic conditions were generally good for hemlock between 10,000 and 6,000 years ago, although occasional, short-lived, cold events appear to have had deleterious effects on hemlock populations (Fuller 1998; Oswald and Foster 2011b). Numerous studies have examined the better-known major decline of hemlock 5,500 years ago (see chapter 8), including the analyses of the sediments of Hemlock Hollow at the Harvard Forest by Foster and Zebryk (1993). Pollen records sampled in detail for the last couple of millennia show the dynamics of New England forests during the Little Ice Age (Oswald and Foster 2011a) and changes in the composition of the vegetation associated with European settlement (Fuller et al. 1998). We have a particularly detailed snapshot of New England forest types just prior to the onset of widespread deforestation and European agricultural practices, thanks to the monumental efforts of Charlie Cogbill and colleagues (2002 and Thompson et al. 2013) to assemble and analyze the settlement-era witness-tree data for the New England states. Tree-ring data and soil pollen records have been used to reconstruct the impacts of historic human activities and other disturbances on hemlock stands in central Massachusetts (McLachlan et al. 2000). C H A P T E R 4 . T R E E - FA L L S A N D TA N B A R K Anthony W. D’Amato

A broad overview of the importance and estimated extent of old-growth hemlock forests in southern New England can be found in Eastern Old-Growth Forests, edited by Mary Byrd Davis (1996). The text includes a chapter by several forest ecologists and old-growth enthusiasts from southern New England and New York that describes early old-growth studies in the region, including an excellent summary of Nichols’s work at the Colebrook tract and the early Harvard Forest studies at Pisgah. In addition, more recent efforts to “discover” old-growth remnants on the highly impacted southern New England landscapes are presented. These initial studies were expanded over the last two decades, ultimately resulting in new estiB I B L I O G R A PH I C E S S AY S 240

mates of the extent of old-growth forests in the region (cf. Dunwiddie et al. 1996 with D’Amato et al. 2006 for further details). The factors protecting these areas from centuries of land use, including rugged slopes and gnarled trees, certainly restrict the degree of extrapolation of findings from these areas to the rest of the New England landscape; however, the opportunity these areas have afforded for direct observations of natural hemlock dynamics and structures has been invaluable. Examples of this work can be found in Orwig et al. (2001) and D’Amato et al. (2008). One of the better works describing the conservation and scientific value of oldgrowth forests is George Peterken’s Natural Woodland (1996), which offers a fascinating, though often overlooked, synthesis and comparison of research on important forest areas and concepts in the United States and Europe. A large portion of this work was inspired and penned during Peterken’s 1989 tenure as a Bullard Fellow at the Harvard Forest. His text focuses on several remnant stands of old growth on the two continents and highlights the lessons learned from research conducted in these areas. The Pisgah tract is included among the old-growth stands covered and, although not containing the depth of detail found in the works of Cline and Spurr (1942) and Foster (1988a), the volume offers a concise overview of the broad lessons on hemlock forest dynamics stored within that ancient grove. There are numerous individual studies documenting the dynamics of oldgrowth hemlock systems, which Craig Lorimer (1995) summarizes in a broad overview in Proceedings of the Conference on Hemlock Ecology and Management. Although largely focused on hemlock systems in the Lake States region, Lorimer’s report captures the primary models of disturbance and development for old-growth hemlock forests across its range and includes a discussion of approaches to restoring oldgrowth conditions to second-growth forests. In the past two decades, managing for old-growth conditions has become a major focus for many conservation and management agencies, reflecting the core visions for ecological forestry put forth by Steve Spurr and Al Cline in the late 1940s. L E SSO N S F RO M H A RVA R D F O R E S TS A N D E CO LO G I S TS : I I . B O B M A R SH AL L ’ S P LOT David R. Foster

Given the prevalence of second-growth forests throughout southern New England, there is no shortage of literature describing the history, structure, and dynamics of these stands, some of which has been based on detailed forensic analyses of these forests. Much of this work is linked with the Harvard Forest, including Bob Marshall’s formative work in 1927. The reconstructive approach introduced by Marshall, which employs the growth patterns revealed on cut stumps to infer past forest history and development, has been a common thread throughout many of these B I B L I O G R A PH I C E S S AY S 241

studies, including the seminal papers on second-growth dynamics by Oliver and Stephens (1977), Hibbs (1982), Kelty (1986), and McLachlan et al. (2000). These reconstructive approaches to forest development were formalized by Lorimer (1985) and have led to several decades of research examining the dynamics of old-growth and second-growth hemlock systems throughout the species’ native range. See the articles by Frelich and Lorimer (1991), Ziegler (2002), and D’Amato and Orwig (2008) for examples of these methods being applied to old-growth hemlock forests. Despite the absence of any publication that clearly laid out the techniques Marshall developed, it is clear from the record of research that the use of these combined techniques was novel, attributed to Marshall, and then subsequently employed by many other researchers at the Forest and beyond. The importance of the insights garnered by this approach and Marshall’s study were immediately apparent, as indicated by this acknowledgment by Russ Lutz and Al Cline (1956) in the Harvard Forest Bulletin no. 27, which summarized thirty years of research on white pine– hemlock forests: “The history of the stands on the four case areas here described was worked out in considerable detail by Marshall (1927) from stump analyses made in 1924, at the time of the first cuttings under the direction of the Harvard Forest staff. His observations contributed much toward an understanding of existing stand conditions, particularly of the successive changes in stand form and composition that resulted from previous logging operations and other disturbances.” Similarly, in a letter of November 21, 1927, Cline relayed to Bob: “we used your technique to detect 100-year-old releases in old-growth forests at Hinsdale [New Hampshire].” Bob Marshall and his father, as well as other family members, were prolific letter writers and retained most of their correspondence. As part of a larger project examining Bob’s history at Harvard and his relationship to the development of scientific and conservation thinking at the Harvard Forest, we have developed a considerable archive of these materials, including copies of all the papers cited and quoted in this volume. From these records and documents that Bob developed during his one-year stay in Petersham, we can glean much about his admiration for the place and the people, as well as his development as a scientist and conservationist. These documents include substantial 1924 correspondence when Bob was in residence in Petersham; Bob’s paper written for his Silviculture VI course at Syracuse University, “The Past History, Present Status and Probable Future Development of Forestry in the Adirondacks with Certain Observations on the Recreational Limitations to Empirical Silviculture within the New York State Forest Preserve”; and Marshall’s high school essay, “Why I Want to Be a Forester.” In an October 1924 letter to his father and brother George, Bob captured some of the diversity of the Harvard Forest life:

B I B L I O G R A PH I C E S S AY S 242

The past week has been another which has past [sic] variedly, pleasantly and interestingly. The six days were divided about as follows: 2 1/2 days Historical hemlock study 1 ″ Mixed Stand Study 1/2 ″ Skull practice 1/2 ″ Lab work on thermocouples with Rupe 1/2 ″ Returning, shoveling snow, etc. 1/2 ″ Surveying 1/2 ″ Visiting world’s largest chair factory lumber yard in Gardner The latter was very interesting. Upon viewing the hundreds of thousands of chair rockers stored away in sheds awaiting the assembling process I was nearly dumbfounded, never having conceived of so many rockers in the whole world. No wonder they same [sic] the present is a comfort loving generation! The snow has been steadily melting from the ground. However, until Friday it remained quite fresh and deep, and the novel sensation of working under such conditions gave me, to borrow a trite and colloquial expression, a tremendous kick. Professor Fisher returned from Weston Thursday, where he had taken his family to their winter home. He is now living in the house and eating with us. He is a very fine type of man with whom to be associated, and you would never think by his manner that he is easily one of the few outstanding foresters in the country. Most of to-day was spent driving around with Professor Fisher to interview old timers in regard to the history of the Adams-Fay Lot in connection with my historical study. We got scant dope of value but much of interest. In the evening I went over to Rupe’s for supper, and then studied with him till 10:40. It is now 11:30. Neil and Art went up to the latter’s home for the day to hunt, but were unsuccessful. This past week has been the entire open season for deer in Massachussets [sic]. The number killed has been amazing when one considers that forty years ago the deer were supposed to be extinct in Mass, and also that there is probably not a single spot in this entire section which is over 2 miles from some road. This evening at five thirty, while Hank, Grant and I were reading and listening to the radio, Al dashed in a[nd] told us that an auto had turned

B I B L I O G R A PH I C E S S AY S 243

turtle along the State Road just below the college. In a jiffy we were outside where Bert Upham and the Prof were already in the formers car. We hopped aboard and in less than a minute were at the scene of the accident. We found a Ford completely turned over on the side of the road and two men pinned under it. The six of us and another man who had spread the alarm put our shoulders to it and soon lifted it enough so that one of the men could crawl out. The other who had been nearly strangled was also dragged out. Miraculously neither received any injury worth mentioning. The accident had occurred when the Ford tried to pass the man who gave the alarm. They skidded on the slippery road and shot off. They were certainly quadruply lucky: first, that they were not killed outright; second that there was snow which had more give than the frozen ground; third, that the man they tried to pass noticed the accident and give the alarm so quickly; and forth [sic] that they tipped over at the only place on the entire nine miles from Athol to Petersham where they could have summoned so much man power so quickly. The humorous part of the incident came when the two men got into the shattered car, turned the self-starter, and away she went, the shattered roof and windshield being the only indications that anything extraordinary had ever happened. Love from Bob In December he was still working on hemlock history and thoroughly embracing the fieldwork, as evidenced by the account in a letter to his father: Life has been mostly boring for the past week. By this I do not imply the slang term but the real meaning of the word. Have been boring hemlock to find out information concerning their growth. Tuesday, Thursday, Friday, Saturday and to-day were all spent that way, or on office work connected with the results of my boring. Last Tuesday was one of the most pleasant days I ever spent working in the field. I reached the compartment of the Forest on which I was going to work at 8 a.m., driving down with Neil. Then I spent about an hour on reconnaissance, picking out an area to bore. Incidently [sic] I climbed a bare hill and got a dandy view of the surrounding country, which was about 80 per cent wooded and 20 per cent pastoral. I was just about to commence work when there was a noise in the bushes nearby and four deer dashed away. That made a total of eleven seen this winter in Massachusetts, in which 30 years ago the wild deer were practically extinct. Just a couple of weeks ago while

B I B L I O G R A PH I C E S S AY S 244

counting rings I heard steps near by me, but did not look up because I did not want to lose my place. When I was finished and glanced about, there stood two does within 100 feet of me. I remained perfectly quiet for about five minutes and so did they, until finally something startled one of them and with the characteristic whistle, they bounded away. It was not until then that I was able to see a buck which had been with them and even closer to me, but hidden behind a ledge just below me. But to return to last Tuesday, after the pleasant deer incident I began to bore. I was way up on top of some high ledges and after finishing each tree I would look off and enjoy the pleasant view. The air was perfectly clear; also mild and balmy and time passed very pleasantly. About 3:30 I finished the job and took another reconnaissance walk, this time through the very fine woods on Mr. Choate’s property, boring a few trees on the way. This scenery reminded me strongly of the Adirondacks, so you can well imagine how much I enjoyed myself. It was 5:30 before I came out on the State Road, with six miles to home. I was awfully glad to had told the fellows not to call for me, because the walk back just around sunset was worth far more to me than anything I could possibly have done in the hour I might have saved. I reached the house at 7, after 10 1/2 hours in the field and 11 1/2 since the last meal, feeling more satisfied than ever that I had chosen forestry for a career. In 1934 Bob joined Al Cline, other staff, and former students of Fisher in designing the memorial plaque for the deceased director, employing the term “sagacious wisdom” to characterize their former mentor. Bob’s regard for Fisher was long-standing, as revealed in a 1928 letter to his father concerning the resignation of Horace Greeley as the chief of the U.S. Forest Service. In that correspondence Bob placed his Harvard Forest mentor in the pantheon of leaders in his field by stating that “Graves, Fisher and Zon alone are abler foresters. . . .” Bob’s deliberate nature and keen sense of responsibility, along with his father’s sobering influence, is illustrated by the fact that he penned his first will shortly after taking his first position with the U.S. Forest Service. He revised this document a number of times as family circumstances changed and in order to reflect his varying passions for major causes in his life. Like Richard Fisher and other Harvard Forest directors, Bob Marshall’s life list indicates that his favorite author was Thoreau. In a June 23, 1928, letter to his family from his research position with the U.S. Forest Service in Missoula, Montana, Bob wrote, “The past winter ranked with the one at Petersham as one of the two most enjoyable winters of the 28 that I have so spent.”

B I B L I O G R A PH I C E S S AY S 245

After his initial lengthy stint with the U.S. Forest Service, Marshall advanced his educational background by completing his Ph.D. thesis at Johns Hopkins. The project he chose—“Climatology and Ecophysiology of Tree Growth at Tree-line in the Arctic”—linked one of Marshall’s favorite landscapes with a set of skills and research questions that would have been quite familiar to Rupe Gast. The two maintained an active correspondence after Marshall left Petersham, one that was filled with tales of exploits from Bob and many growing complaints from Gast about his rather precarious soft-money position at Harvard. Gast also provided Bob with considerable advice regarding options for graduate studies, including a strong recommendation that he consider the University of Massachusetts, where the two could continue to work together. Although there is little record of exchanges regarding the specific topic of Marshall’s thesis, its strong link with many of Gast’s interests suggest that the old Harvard Forest mentor had a lasting effect on Bob. Although Marshall did not follow Gast’s recommendation for graduate studies in Amherst, he relied heavily on Gast for letters of recommendation. The Adams Fay tract, where all of Marshall’s thesis work was advanced, was purchased by the Harvard Forest a few years after Marshall completed his degree. In a 1932 letter, Richard Fisher wrote to Bob thanking him for his contribution toward the successful purchase of the northern parcel of the Tom Swamp tract from the New England Box Company. Today the tract lies peacefully at the north end of Tom Swamp and Harvard Pond, where a solitary staked plot memorializes Bob Marshall’s efforts and insight. The history of this lot and its acquisition by the Forest is described well by Hugh Raup in his introduction to the Harvard Forest Bulletin by Lutz and Cline (1956) on white pine and hemlock: The land involved in Cases 15–18 was not a part of the original Harvard Forest tracts that were acquired by the University in 1907. In 1924 it was in the possession of the New England Box Company of Greenfield, Massachusetts, held by that company for the timber it contained. The late Richard T. Fisher, first director of the Harvard Forest, realizing the significance of the area for additions to the research program of the Forest, was largely responsible for an arrangement with the Box Company whereby experimental cuttings could be made. The Forest staff was permitted to design and manage the cuttings, beginning in 1924, while the Company took the lumber. At that time the property was known as the “Adams-Fay Lot,” and has continued to be called by that name in the Forest records. It adjoins Compartment VIII of the Tom Swamp Block on the north, and after the Box Company transferred it to the Forest in 1932 it became Compartment IX of that block.

B I B L I O G R A PH I C E S S AY S 246

Like so many efforts in the first quarter century of the Harvard Forest, the big experiment that Fisher initiated and Bob devoted his graduate career to was abandoned following the 1938 Great Hurricane and the subsequent salvage that leveled most of the area. Nonetheless, a successful effort to relocate the plot and resample the forest eighty years after Bob laid it out supported the hypotheses that Fisher sought to test (see Ireland et al. 2008). C H AP T E R 5. H E M LO C K A S A FO U N D AT I O N S PE C I E S Aaron M. Ellison and Benjamin H. Baiser

Ecologists identify foundation species based on three defining characteristics (Ellison et al. 2005a): they are numerically abundant and account for most of the biomass in an ecosystem; they occupy the base of the food web (i.e., they are the plant producers); and they are connected with many other species in the food web. More broadly, Mark Harmon (personal communication, September 2012) notes that we perceive foundation species differently from other species: foundation species define a system and are inseparable from it. Measuring all these characteristics to help us identify foundation species before they decline or disappear is the primary goal of both defining and studying them. Not surprisingly, we tend to think about, care for, and manage rare species much more than common ones. But given that foundation species shape ecosystems, we should be paying much more attention to identifying and protecting such common species before their populations decline below functional levels (Gaston and Fuller 2007, 2008). All ecological systems consist of a mix of both rare and common species. The general pattern is that biotic assemblages are dominated numerically by a few species with high abundances, but the majority of species are found at relatively low abundances (Fisher et al. 1943; Preston 1948; MacArthur 1960; Ulrich et al. 2010). There are different ways of explaining these patterns of abundance. Niche models explain the relative abundance as a consequence of dividing up resources (Sugihara 1980), whereas neutral models predict species abundance distributions based on immigration, emigration, and extinction (Hubbell 2001). George Orwell, whose reduction of the “Seven Commandments of Animalism” to the maxim “All animals are equal, but some animals are more equal than others” could be considered the conceptual parent of the foundation species concept (Orwell 1945). But the marine ecologist Paul Dayton was the first scientist to articulate it, describing species that “have roles in the maintenance of the community disproportionate to the abundance or biomass of the species” (Dayton 1972). At the time, he was discussing the effects of pollution on benthic marine assemblages at McMurdo Sound, Antarctica, that were dominated by sponges. This was published in a relatively obscure conference proceedings volume, yet another warning B I B L I O G R A PH I C E S S AY S 247

to the junior scientist about limiting the reach of one’s research by publishing in a conference proceedings. Nevertheless, Dayton’s paper has continued to have a great influence on how marine ecologists characterize the organisms in the systems that they study. Dayton provided examples of foundation species from both terrestrial and aquatic ecosystems, but many ecologists working in those habitats appear to have been unaware of his work. In the decades following, terms were independently proposed with some or all of the attributes of foundation species. For example, Hanski (1982) defined core species as locally abundant and regionally common and contrasted them with satellite species that are sparse or rare. Grime (1987) characterized dominant species that competitively exclude subordinate species by garnering a disproportionate share of resources and contributing most to productivity. Holling (1992) hypothesized that all terrestrial ecosystems are controlled and organized by a small set of keystone species that share the characteristics of core species, dominant species, and keystone predators. In the mid-1990s the idea that particular species can create physical structures in the environment, limit or amplify variation in environmental conditions, or provide resources or habitats for other species was independently proposed by Huston (1994), who termed such species structural species, and Jones ( Jones et al. 1994), who called them ecosystem engineers. As virtually all species modify their environment to some degree (through the process of niche construction; see OdlingSmee et al. 2003), we are clear in our characterization of foundation species that their activities are disproportionate to their abundance or biomass (Ellison et al. 2005a). This is analogous to cornerstone species (Bracken and Low 2012), which are uncommon or rare and exert strong “bottom-up” effects on higher trophic levels. Foundation species may play extremely important roles in food webs, which are a specific class of ecological networks that depict feeding interactions of species in a given location. The arrangement, number, and distribution of species and links are measured by a number of different metrics that have been used to compare different food webs (Vermaat et al. 2009), understand the dynamic stability of food webs (May 1973; Allesina and Tang 2012), and explore how species extinctions cascade through food webs (Dunne et al. 2002). The most important roles in an ecosystem of a foundation species are most likely not related to its being eaten. Nontrophic interactions such as facilitation (Bertness and Callaway 1994), parasitism (Dobson and Hudson 1986), and competition (Schoener 1983) all play key roles in structuring ecological communities but are only recently being integrated into studies and analyses of ecological networks. For example, Kéfi et al. (2012) place nontrophic interactions into three categories based on how they influence focal species. The first category contains interactions that directly modify feeding parameters (e.g., consumption rate) such as B I B L I O G R A PH I C E S S AY S 248

creating hunting perches for predators. The second category includes interactions that modify nonfeeding parameters (e.g., metabolism and reproduction) such as pollination. The final category is interactions that modify flows across ecosystem boundaries such as seed dispersal and the transfer of nutrients (e.g., bird guano). Incorporating nontrophic interactions into dynamic food web models is especially important when considering foundation species, which exert most of their influence through nontrophic interactions. In sum, foundation species are the most abundant species in the overall ecological community, where they provide the support for networks of interacting species that feed on each other and that interact in ways that do not involve eating or being eaten, and they also are the hubs that connect many subsidiary networks. As a foundation species declines past a certain point, models predict that the species and interaction networks that depend on it are simplified as individual species are lost. Some of these disappear because they have direct and important interactions with the foundation species, while others disappear because of indirect effects: they depend on species that themselves depend on the foundation species. At the same time, the decline or loss of a foundation species can reduce the flow of resources through the system, resulting in an overall reduction in numbers of associated species. Once populations of producers or herbivores and intermediate carnivores shrink below a critical point, predators that depend on them no longer have enough to eat, and they disappear. Finally, once foundation species disappear either functionally or entirely from a system, they may be replaced by other species. If, as is the case in New England’s hemlock forests, these replacements are a mixture of species such as birches and maples, none of which are foundation species, the entire system may be rearranged. New interaction networks are established, but none are as well-linked as in a system grounded by a foundation species. These illustrations of the consequences of foundation species loss are derived from mathematical models of ecosystems. We emphasize, therefore, that they are hypotheses about what could happen when foundation species are lost. As in many fields of science, theories and models continue to outpace the data available to test them; the results described in chapters 6–9 are a first step in both testing these models and in giving us the insights needed to identify foundation species before they are lost. These approaches to combining observations of past and current conditions, experiments, and models mirror the methods proposed for “modern” and successful forestry (Stephens 1955).

B I B L I O G R A PH I C E S S AY S 249

C H AP T E R 6. A R A N G E - W I D E H E M LO C K D E C L I N E W. Wyatt Oswald and David R. Foster

The postglacial history of vegetation in New England is known from a network of dozens of pollen records from ponds and bogs, with study sites ranging from Connecticut (Davis 1969) to Maine (Anderson et al. 1992). The abrupt decline of hemlock 5,500 years ago, first observed by Deevey (1939), is a striking feature that appears in all paleoecological records from the region, including Hemlock Hollow and Black Gum Swamp at the Harvard Forest (Foster and Zebryk 1993). Deevey assumed that the hemlock decline was caused by warm, dry climate, an interpretation informed by and consistent with paleoenvironmental evidence for mid-Holocene warmth in Europe (Sernander 1908). However, the application of the European framework for understanding past changes in the environment and vegetation was called into question in the 1970s and 1980s by North American scientists working on questions as varied as the ecology and demise of the Pleistocene megafauna (Martin 1973) and the possibility that vegetation lagged behind climate as hospitable conditions returned to the northeastern United States after the retreat of the glaciers (Davis 1981b, 1983). Margaret Davis was foremost among those interested in the peculiarities of the North American paleoecological record and the role of ecological interactions across the millennia. She was particularly interested in the hemlock decline, and she introduced her biological agent hypothesis at a conference in Lucknow, India, in 1978 (Davis 1981a). That interpretation was strengthened by the details of the decline that appeared in the pollen data from Pout Pond, New Hampshire (Allison et al. 1986), and a likely culprit was identified when Bhiry and Filion (1996) encountered hemlock looper remains in sediments of mid-Holocene age. Given the assumption that the hemlock decline was attributable to a pathogen or insect pest, various studies proceeded to use the event as a paleoecological experiment in foundation species removal (Foster and Zebryk 1993; Fuller 1998; St. Jacques et al. 2000). In the meantime, research on the patterns, causes, and consequences of past changes in climate began to identify events that were global in scale. The Younger Dryas, a cold interval at the Pleistocene-Holocene transition, is the best-known climatic event with an apparently worldwide reach (Alley et al. 2003; Rodbell 2000), but other shorter-term events are emerging as records are developed with increasingly higher temporal resolution, including the so-called “8.2K event,” in which a sharp cold event triggered responses globally (Alley et al. 1997; Alley and Ágústsdóttir 2005). Cold and dry climatic intervals occurred in New England during the mid-Holocene (Shuman et al. 2009), and we have proposed that they played a key

B I B L I O G R A PH I C E S S AY S 250

role in the hemlock decline and the coincident decline of oak along the coast (Foster et al. 2006). The drought that affected New England in the 1960s may represent a modern analogue for the climatic conditions of the mid-Holocene (Namias 1966). The current landscape-scale mortality of oak in parts of Martha’s Vineyard gives us insights into the causes and ecological consequences of the mid-Holocene declines of hemlock and oak, but to visualize the scope of tree mortality in New England 5,500 years ago, we must look to western North America. The onset of warmer and drier conditions has led to widespread outbreaks of mountain pine beetle in the Rockies (Kurz et al. 2008) and to tree mortality and a fire-mediated shift from black spruce to broadleaf deciduous trees in boreal Alaska ( Johnstone et al. 2010). It appears that human populations in New England were influenced by a combination of environmental and biological changes in the mid-Holocene (Munoz et al. 2010), and it would serve us well to anticipate the societal impacts of future climatic variability. C H A P T E R 7. I N VASI O N O F A N E XOT I C PE S T David A. Orwig

New England forests encompass just over thirty-three million acres, and the vast majority of them have been exposed to at least one or more invasive insect pests. Many of these forests have experienced moderate to substantial impacts on their structure and composition by these outbreaks. Liebhold et al. (1995), Orwig (2002), and Dukes et al. (2009) provide details of the historical introductions, impacts, and also potential threats from other insects migrating toward the region. In addition, a few recent books have documented the social and ecological impacts of the loss of dominant trees in the past: Bolgiano (2007) and Freinkel (2007) describe the loss of chestnut, while Campanella (2003) documents the loss of American elm. Much of what we know about the life cycle of the hemlock woolly adelgid comes from the extensive work of entomologist Mark McClure. During the late 1980s and 1990s, he documented all aspects of the adelgid, from its parthenogenetic life cycle and two generations a year (McClure 1987, 1989a), to the many vectors that disperse it (McClure 1990), to its density-dependent population trends (McClure 1991) and patterns of early spread in New England (McClure 1989b). As part of these formative studies in the late 1980s, he also discovered an interesting wrinkle in this insect’s complex life cycle. As hemlocks start to decline, a percentage of the spring adelgid generation develop wings and fly away in search of a spruce tree that is needed to complete their life cycle (McClure 1987). Fortunately, there are no suitable spruce hosts in North America, so at least part of the population dies without further damage or long-distance dispersal. Recent studies have documented that the

B I B L I O G R A PH I C E S S AY S 251

adelgid in eastern North America resulted from a single introduction from southern Japan (Havill et al. 2006), and that the spruce species that it can migrate to, form galls, and sexually reproduce on, is tigertail spruce in Japan (Havill et al. 2011). One of the surprising facts of woolly adelgid dynamics is the remaining uncertainty regarding how they actually kill hemlock trees. In their investigations of adelgid feeding on hemlock twigs, Young et al. (1995) speculated that the insect may inject toxic saliva into the twigs. Field studies support this hypothesis by documenting that even low adelgid densities result in reduced branch growth (Miller-Pierce et al. 2010; Miller-Pierce and Preisser 2012). Recent work by our collaborator Evan Preisser at the University of Rhode Island has further investigated the mechanisms of adelgid damage and highlighted that feeding by adelgids may induce systemic responses within the plant, such as localized cell death as part of a hypersensitive defensive response (Radville et al. 2011) and increased water stress and amino acid concentrations (Gomez et al. 2012). Evidence for altered tracheid formation and reduced water transport associated with adelgid feeding further highlights the association of feeding with water stress (Gonda-King et al. 2012). One of the few factors that controls adelgid population densities is cold temperature. Following initial observations that colder winters lead to reduced population numbers the following spring, researchers at the University of Vermont collected adelgid from branches in southern New England and exposed them to increasing colder temperatures in growth chambers to examine what conditions adelgids could tolerate (Parker et al. 1998, 1999; Skinner et al. 2003). Harvard Forest aided this effort on several occasions by leading investigators to stands with high adelgid densities. Field examinations supported the initial laboratory findings by showing a strong relationship with adelgid mortality and latitude and, by proxy, cold temperature) (Paradis et al. 2008; Trotter and Shields 2009; Orwig et al. 2012). These results, along with genetic evidence that adelgid had gained cold tolerance over time (Butin et al. 2005), have helped shape predictions on how rapidly and far to the north the adelgid will be able to spread and survive (Paradis et al. 2008; Dukes et al. 2009; Albani et al. 2010). One of the early challenges in documenting adelgid abundance and spread was that a standardized quantitative protocol for assessing presence and abundance did not exist, so descriptions were largely subjective and not comparable across sites (i.e., “high, moderate, low, and absent”). Such a protocol was developed at Harvard Forest one summer by undergraduate student Joe Brown and entomologist Scott Costa from the University of Vermont (Costa and Onken 2006). Scott was a great collaborator and a true joy to work with, and he spent a large portion of his career working on adelgid-related studies, including developing a method to mass apply

B I B L I O G R A PH I C E S S AY S 252

adelgid-killing fungi onto trees Unfortunately, he died in 2012 before this work could be fully developed. The magnitude and duration of adelgid impacts within a forest vary tremendously from location to location and are controlled by a variety of site factors (Orwig et al. 2002, 2012), weather factors (McClure 1989b; Paradis et al. 2008; Dukes et al. 2009), vectors (McClure 1990), and random events (Fitzpatrick et al. 2012). Shortly after the adelgid arrived in New England in the mid-1980s, it was believed that trees would die within four years (McClure 1991), but over time it became clear that many trees could survive ten to fifteen years or more with adelgid, at least in the north (Orwig and Foster 1998; Orwig et al. 2012). Examining the dynamics of an invasive pest as it spreads is an extremely challenging but interesting endeavor. After speaking with dozens of scientists, managers, and homeowners, we were able to find a variety of hemlock locations and settings with adelgid present to establish long-term plots and follow how the forests changed over time (Orwig and Foster 1998). Revisiting these stands over time has been invaluable in documenting the structural and compositional changes associated with the decline and loss of this magnificent tree. By mapping hemlock from aerial photographs across large portions of Connecticut and Massachusetts, we were able to examine even larger variation in hemlock distribution across the region (Orwig et al. 2002, 2012). This base map of hemlock has provided the platform for a broad suite of studies examining various consequences of hemlock decline and loss, including impacts on wildlife (Tingley et al. 2002; Mathewson 2006, 2009); microenvironment and ecosystem function (Cobb et al. 2006; Orwig et al. 2008); carbon storage (Raymer et al. 2013); logging impacts (Kizlinski et al. 2002; Orwig and Kittredge 2005); hemlock regeneration (Preisser et al. 2011); hemlock physiology (Smith and Orwig 2008); subsequent spread of adelgid (Preisser et al. 2008); and even property value declines (Holmes et al. 2010). Knowledge of a range of hemlock stands with varying levels of adelgid infestation led to a rich and thoroughly enjoyable collaboration with Bernhard Stadler of the University of Bayreuth in Germany. Using detailed studies of microbial abundance and the chemical composition of precipitation falling through hemlock trees with and without adelgid, we were able to document how the presence of adelgid enriched the flow of energy and nutrients from the canopy to the forest floor and increased the abundance of bacteria, yeast, and fungi on hemlock needles (Stadler et al. 2005, 2006). One of the unexpected observations of visiting stands throughout southern New England was the amount of hemlock cutting that had occurred. When ground truthing the first aerial photograph–derived hemlock map for Connecticut, it be-

B I B L I O G R A PH I C E S S AY S 253

came clear that hemlock cutting had increased due to adelgid damage (Orwig et al. 2002). These observations led us to explicitly examine cut-over areas and compare them with stands that were left to decline due to adelgid (Kizlinski 2002; Kizlinski et al. 2002). Additional observations in Massachusetts confirmed that hemlock cutting was frequent, but not always due to adelgid (Orwig et al. 2012). These experiences helped formulate our management ideas (see chapter 11; Orwig and Kittredge 2005; Foster and Orwig 2006) and led to additional studies at the Harvard Forest’s Simes tract (Ellison et al. 2010) and studies of logging at the Arnold Arboretum and in central Massachusetts. As hemlock declines and is lost, terrestrial and aquatic habitats will change as well, and some species will decline while others will increase following the loss of hemlock and its replacement with hardwoods. The sources cited above to describe hemlock’s habitat and physiological traits are also helpful in predicting changes following hemlock’s loss. There are no physiological or wildlife studies that have followed hemlock function or species abundances from a pre-adelgid condition up to and including replacement by other species. The Tingley et al. (2002) investigation that documented major changes in bird communities once hemlock was removed did not include pre-adelgid data. Other studies either predict what will happen or are in the process of monitoring changes as hemlocks decline. One of the biggest challenges associated with invasive pest and pathogens—the question of how to control them—also leads to the most questions. Early on, when adelgid was first being described, Mark McClure outlined the chemical options that were available (McClure 1987, 1995) and helped locate and describe the first mass-produced biocontrol agent used in the fight against the adelgid (Sasaji and McClure 1997). Periodic syntheses on biological control are produced by the USDA Forest Service to help inform landowners and managers (McClure 2001; Reardon et al. 2004; Onken and Reardon 2011). Dave Mausel, whose doctoral and postdoctoral research focused on introducing the tooth-necked fungus beetle throughout the eastern states to fight adelgid, worked for two summers at the Harvard Forest, assisting in our first landscape sampling of adelgid in Connecticut. Witnessing the adelgid’s impacts firsthand certainly helped to inspire Dave to pursue potential control mechanisms. Information on chemical control can be found in Cowles (2009) and also in the proceedings from the USDA symposia on hemlock woolly adelgid in the eastern United States, held every three years. Finally, early records of when and where we first saw the elongate hemlock scale during our adelgid sampling have helped to document an interesting wrinkle in the adelgid-hemlock story. In the mid-1990s, elongate hemlock scale was barely mentioned as a problem for hemlock, despite detailed work on its life cycle and biology in the 1980s by McClure (1979, 1989b). While sampling foliage for adelgid, B I B L I O G R A PH I C E S S AY S 254

we also recorded the presence of any other pests. It was only after resampling many of these stands years later that we realized how rapidly scale had dispersed through the landscape (Preisser et al. 2008, 2011). The presence of these two co-occurring pests intrigued us and led to several studies and ongoing research examining these interactions and how they influence hemlock health and foliar chemistry (Smith and Orwig 2008; Miller-Pierce et al. 2010; Miller-Pierce and Preisser 2012). C H A P T E R 8. C U T O R G I R D L E Aaron M. Ellison, David A. Orwig, and Audrey A. Barker Plotkin

There are many classic examples of large field experiments in ecology: the forest stream studies at Hubbard Brook, New Hampshire, where entire small watersheds have been studied exhaustively and then treated by logging, liming, clear-cutting, and herbicides (Likens and Bormann 1995); studies in Ontario and Wisconsin, in which entire lakes or large sections of lakes have experienced the selective removal of different fish species or changes in pH to simulate acid rain (Carpenter 2003; Schindler 2009); and studies across the Konza tallgrass prairie in Kansas, where the role of bison grazing has been contrasted with the effects of fire, fire and bison grazing together, or grazing by cattle alone (Knapp et al. 1998). A large series of experiments initiated around 1990 has formed a centerpiece of the Harvard Forest LTER program: a “hurricane” experiment, in which two acres of forest were pulled down to simulate the effect of a storm like that of 1938 (CooperEllis et al. 1999; Barker Plotkin et al. 2013); various warming experiments, in which the soil or forest understory were heated to evaluate the impact of climate change on forest ecosystems, plants, and insects (Melillo et al. 2011); and a nitrogen saturation experiment, in which nitrogen was sprayed onto forest soils to simulate enhanced levels of acid deposition (Magill et al. 2004). In each of these studies, the researchers investigated the conditions in manipulated and control areas at similar sites using the same methods. The experimental results illustrate the differences for forest dynamics between specific manipulations and natural variation and ongoing environmental changes in intact forests. The experiments themselves usually had more than one objective: to simulate existing, anticipated, or potential extremes in a natural disturbance or environmental stress; to compare a variety of potential management treatments; or to disrupt ecosystems in fundamental ways in order to evaluate their response and basic organization. Cut or Girdle was developed as a complement to these earlier experiments. It took nearly a year to design and develop this large experiment with its coordinated set of manipulations and controls. Finding the proper location for a twenty-acre study turned out to be a significant challenge, despite the fact that our 3,500 acres of forests include considerable amounts of hemlock. But northeastern forests are B I B L I O G R A PH I C E S S AY S 255

typically mosaics of different forest types, and what we needed was a large, relatively homogenous set of adjoining hemlock stands. While we do have some beautiful, large hemlock-dominated stands, including the magnificent woods on the Prospect Hill (Sanderson woodlot), Slab City, and Tom Swamp (Earl Stephens plot and R. T. Fisher Memorial) tracts, most of these are relatively old forests that have been set aside as reserves for long-term observational studies, rather than for harvesting or manipulative experiments. Old maps, recollections, and contemporary collaborators frequently directed us to forests that turned out to be dominated by oak or pine, with only an understory of hemlock. In itself, this confusion was informative because it provided an important indicator of the strong influence that hemlock exerts on our perceptions; when standing in a forest understory, hemlock’s thick foliage restricts any sense of the abundance of the other trees that reach into the upper canopy. In almost every situation, the actual abundance of hemlock is greatly overstated by most observers and incorrectly perceived in our memories. Finally, by combining data in Harvard Forest inventory records with new analyses of aerial photos, and following many days of tromping through the forests, we located two large blocks of hemlock stands suitable for the experiment. Cut or Girdle—known in our publications as the Harvard Forest Hemlock Removal Experiment (Ellison et al. 2010)—is the most recent long-term experiment set up at Harvard Forest. Here we were interested in understanding how forest ecosystems would change following the removal of a single foundation species. We also sought to test hypotheses derived from large-scale observational and modeling studies concerning the changes in forest dynamics as hemlocks were killed by the hemlock woolly adelgid or by preemptive salvage logging (Orwig and Foster 1998; Albani et al. 2010; Orwig et al. 2012) and to test general ecological theories about the role of foundation species (Ellison et al. 2005a). Cut or Girdle is an example of a “Before-After-Control-Impact” experiment (Gotelli and Ellison 2012), a common design used to study the effects of environmental change, especially due to human intervention or large-scale construction (such as dams, power plants, and so on). The “Before” and “After” mean that we take many of the same measurements both before and after the treatments are applied. This enables us to track the normal variation and ongoing changes in the experimental plots due to succession and environmental variability. The “Control” and “Impact” mean that we have both control plots and treated plots. Our experiment was set up in two blocks; each block contains four plots, each about two acres (90 by 90 meters, or 0.81 hectares). Each plot within a block was assigned to a treatment—girdled, logged, hemlock control, or hardwood control— and so replication was between, not within, the blocks. Each of the blocks occupies

B I B L I O G R A PH I C E S S AY S 256

a different habitat: one is in a moist, sheltered valley, whereas the other is on a dry, exposed ridge. The two blocks enable us to examine the effects of the broader environment—aspect, exposure, soil type, land use history—whereas the replication of treatments allows us to evaluate how the different canopy manipulations affect forest dynamics (Ellison et al. 2010; Gotelli and Ellison 2012). However, because there is only one of each treatment within each block, we cannot determine whether the effects we observe on forest dynamics of logging on the ridge are statistically different from the effects of logging in the valley. In many environmental impact studies, the “Before” is assumed to be “natural,” “pristine,” or “stable.” The reality is that, in most locations, there is a long history of human activity and use overlain on background variability. Perhaps nowhere in North America is the history of land use as pronounced or as well documented as it is in New England (Foster and Aber 2004). This land use history can temper or constrain the range of responses to our experimental manipulations and must inform how we interpret the results. The Cut or Girdle experiment is located in the Simes tract, named for Olive Simes, who willed the property to Harvard in 1970. It has a long history of use. At its southern end are the remains of an old house foundation, a well, extensive stone walls, large maple trees, old apple trees, scattered barberry, and daylilies. Farther into the tract, the majority of the land has been cleared at one time or another. The trees are mostly 80 to 100 years old; among the oldest, at nearly 150 years old, are the hemlocks in the valley block. In the 1930s Harvard Forest assistant director Al Cline described the southern part of the Simes tract as “covered with an inferior growth in which gray birch predominates . . . [along with] a mixed growth of hardwood (both cordwood and prospective sawtimber), pine and hemlock of decidedly better quality, and of older age than the growth on the central and eastern portions.” Most of the larger trees on the site were destroyed by the 1938 hurricane, and in February 1939 Cline wrote to Olive Simes that “the hurricane resulted in the complete blowdown of all of the larger timber, including at least nine-tenths of the area supporting stands which a lumberman would consider merchantable and a good logging chance. No stand completely escaped.” In July 1980 a detailed inventory of the entire tract documented that “hemlockhardwoods” dominated just over eighty-four acres. By 2003, when we began scouting for good study locations, there were many well-developed stands of mature hemlock. Before we logged or girdled the canopy, hemlock composed approximately 60 percent of the initial basal area and initial stem density in the six hemlock plots. In the hardwood control plots, hemlock was 10 percent or less of the initial

B I B L I O G R A PH I C E S S AY S 257

basal area and stem density; other species included red maple and sugar maple, black birch, red oak, and white pine. Both hardwood control plots started with many black birch and red maple individuals in the canopy and subcanopy. Trees in the logged treatment were killed immediately and removed from the site, whereas trees in the girdled treatment died slowly over time (Ellison et al. 2010). Five years after canopy manipulations, the overstory in each of the two treatments had lost between 60 and 70 percent of the living stems and basal area. As the overstory died and the canopy opened up, the average temperature increased by about 2°C in the girdled plots and 4°C in the logged plots; diurnal and seasonal variability increased in parallel (Lustenhouwer et al. 2012). By 2012, however, the microclimates in the two canopy manipulation treatments were similar, recovering to levels seen in the hardwood control treatment. The forest understory, with its seeds, seedlings, and saplings waiting in the wings (i.e., advance regeneration), responded rapidly and predictably to the removal of the hemlock overstory. These responses have been summarized in a set of reports focused on seed bank dynamics (Sullivan and Ellison 2006; Farnsworth et al. 2012), understory regrowth (Farnsworth et al. 2012; Orwig et al. 2013), and advance regeneration (Orwig et al. 2013). Both before and after the canopy manipulation treatments were applied, the seed bank in all plots was dominated by birch and brambles (raspberries and blackberries). Birch and bramble seedlings rapidly formed a dense mass in the understory of the logged plots. Brambles were thick in those plots through 2011, but by 2012 they had virtually all disappeared after being overtopped and outcompeted by the now two- to three-meter (six- to ten-foot) tall birch saplings. Brambles were scarcer in the girdled plots, but birch was abundant and grew almost as fast as it did in the logged plots. By 2012 birch saplings were denser in the girdled plots than in the logged plots; self-thinning was already apparent in the logged plots. Although some of these differences in vegetation structure and dynamics that were observed between the logged and girdled treatment plots can be attributed to the more rapid opening of the canopy in the logged plots (Ellison et al. 2010; Lustenhouwer et al. 2012), others are better attributed to the effects of soil disturbance by logging machinery and the one- to two-foot tall piles of slowly decomposing slash left by the logging operation (Smith 1986; Foster and Orwig 2006). Similar contrasts were apparent when comparing vegetation changes following salvage logging after the 1938 hurricane to those seen in the Harvard Forest’s simulated hurricane experiment, in which no salvage logging was done (Foster and Aber 2004; Barker Plotkin et al. 2013). Logging and girdling resulted in small, short-term changes in ecosystem dynamics due to rapid pulses of litter and death of tree roots, followed and compenB I B L I O G R A PH I C E S S AY S 258

sated for by rapid regrowth of vegetation. These results were summarized by Orwig et al. (2013). In general, interannual variability exceeded differences among treatments. Soil carbon flux in girdled plots showed the strongest response: 35 percent lower than controls after three years, and slowly increasing thereafter. Ammonium availability increased immediately after logging and two years after girdling, due to increased light and soil temperatures, nutrient pulses from leaf fall, and reduced uptake following tree death. The net effect of these short-term perturbations on long-term dynamics, however, is still not well understood. Modeling suggests that the loss of hemlock will reduce overall net primary production and carbon stocks for twenty to fifty years, but, after that, the rapid growth of young, early successional hardwoods will lead to stands with productivity and standing stocks of carbon that will be higher than the mid-successional hemlock stands that the hardwoods replaced (Albani et al. 2010). However, the “opportunity costs” of fifty years of lowered productivity may take more than a century to make up. Finally, the food webs of animals living in the soil, the forest floor, and the forest canopy are reorganizing as hemlock disappears. In a study along Orwig’s observational transect across Connecticut (Orwig and Foster 1998), we found that several genera of ants, notably Lasius and Formica, were absent in hemlock-dominated stands but would colonize stands as hemlock cover declined to below 50 percent of the basal area (Ellison et al. 2005b). We predicted a similar compositional change in the Cut or Girdle plots, and indeed we have observed it (Sackett et al. 2011). Spider and beetle assemblages are changing in parallel (Sackett et al. 2011), deep forest birds likely will move elsewhere (Tingley et al. 2002), and moose are moving in. By late 2009 the hemlock woolly adelgid could be found on hemlocks in all of the experimental plots. The warm and dry winter of 2011–12 provided perfect conditions for the adelgid populations to expand, and the prevalence and abundance of this insect was noticeably greater in the summer of 2012. The patterns of infestation within plots is following predicted patterns (Turner et al. 2011), as is the regional spread of the adelgid as the climate warms (Fitzpatrick et al. 2012). Over the next decade, we have the unfortunate opportunity to be able to compare plots newly infested by the adelgid with the girdled plots in which the trees were killed standing in place. In the future, the state of these adelgid plots will be the new “normal.” C H A P T E R 9. M O D E L I NG T H E D Y N A M I C S O F A F O R E S T G I A N T Jonathan R. Thompson

Virginia Dale and her colleagues (1998) at Oak Ridge National Laboratory wrote a reply to Aber’s (1997) lament in the Bulletin of the Ecological Society of America suggesting that he inappropriately emphasized “belief ” of models rather than focusB I B L I O G R A PH I C E S S AY S 259

ing on the understanding they could provide. They also suggested that the failings Aber identified were based on unrealistic expectations of models and modelers. Aber’s (1998) reply to their comment pointed out their many points of agreement and stated that any remaining disagreements were based on semantics and not true difference. Together the three short papers provided a concise review of the merits and shortcomings of ecological modeling at a time when ecological modeling was poised to move into the mainstream of ecological research. Aber was among the original cohort of scientists involved with the establishment of both the Hubbard Brook and Harvard Forest LTER sites. Thanks in no small part to Aber, both sites have since made major contributions to the science of ecosystem modeling. As the home of the original JABOWA forest simulator (Botkin et al. 1972), Hubbard Brook will always be associated with forest gap models. In truth, though, both Hubbard Brook and Harvard Forest have made larger contributions to biogeochemical and whole ecosystem models. For example, the widely used PnET family of models are used to simulate the carbon, water, and nitrogen dynamics of forest ecosystems and have been developed, calibrated, and validated based on long-term data and experiments at the Harvard Forest (Aber and Federer 1992; Aber et al. 1995; Aber et al. 1996). Similarly, the Terrestrial Ecosystem Model (TEM) is a continental- to global-scale biogeochemical model developed by Harvard Forest scientists that is used by researchers worldwide (Melillo et al. 1993). Because these models have not been used to model hemlock dynamics explicitly, we do not discuss them in detail. After JABOWA, the advancement of forest stand modeling followed two paths. The first was growth and yield simulators used to aid in the prediction of timber harvest volume (Weiskittel et al. 2011). The second was individual-based forest simulators designed to study demographics and neighborhood processes (Pacala et al. 1993). The Ecosystem Demography model (ED) is a distant relative to JABOWA whose ancestry can be traced through this latter path. ED’s major advancement over the original individual tree models is a series of complex equations that approximate the behavior of running many gap models over a large area simultaneously, in terms of numbers of trees, their functional types, and size. Through this approach, ED formally scales leaf-level physiological processes to the ecosystem scale. One of the team members, Harvard professor Paul Moorcroft, is a Harvard Forest LTER collaborator who helped pioneer this approach to regional-scale forest modeling while working in the Amazon Basin in the 1990s. His lab has since produced a version of ED that is specifically designed to simulate the climate, land use history, and vegetation dynamics of eastern North America. ED’s estimates of carbon and water flux throughout the ecosystem were calibrated using data from the eddy flux tower located at Harvard Forest and validated using U.S. Forest Service field data B I B L I O G R A PH I C E S S AY S 260

and other flux towers in the Northeast. These flux towers measure the exchange of gases between the atmosphere and the forest canopy and enable accurate estimation of forest growth and decomposition. The simulations of forest ecosystem processes produced by this version of ED are impressive in their ability to reproduce the field data and the flux tower estimates. It is this version of ED that has been used to forecast future changes in forest carbon dynamics with and without the presence of hemlock (Albani et al. 2010). C H A P T E R 1 0. R E PR I SE : E AST E R N H E M LO C K A S A F O U N D AT I O N S PE C I E S Aaron M. Ellison

For a practicing scientist, it is insufficient to assert that a particular item has a set of properties (e.g., all swans are white, have wings, and float on the water), identify an object that appears to have those conditions (e.g., look, there’s a white, winged, floating object), and then conclude that the object is the particular item (the white, winged, floating object is a swan). We could be looking at an egret, a wooden statue of an angel that was painted white, or any number of other objects. Similarly, just because we can define a foundation species and see that eastern hemlock has many of the characteristics of a foundation species, we cannot simply say that it is a foundation species. Rather, a scientist employs a particular set of rules—the so-called scientific method—to determine if the white, winged floating object is a swan, or if the dominant, massive tree with lots of connections to other things in the forest is, in fact, a foundation species. The scientific method differs substantively from the evidence required to prove something beyond a reasonable doubt in the courtroom, as the U.S. Supreme Court has demonstrated repeatedly: the epigraph for chapter 10 is based on the infamous 1964 decision in Jacobellis v. Ohio (378 U.S. 184) establishing a threshold test for obscenity and censorship (“I know it when I see it”). Science proceeds by the process of rejection and elimination of alternatives: Karl Popper’s hypothetico-deductive method (Popper 1935). We start with observations (such as those described in chapter 7), construct a suite of plausible hypotheses to explain the observations, collect additional data (observations, results from experiments, output from models), and use these data to eliminate one or more of the hypotheses. Eventually, as in a Sherlock Holmes story (the quotation in chapter 10’s opening paragraph is from p. 111 of Sir Arthur Conan Doyle’s 1890 novel The Sign of Four), only one of our initial hypotheses is left standing, but all we can assert is that we have failed to reject it. But as we learn more, we will continue to confront our hypotheses with new data (Hilborn and Mangel 1997), and it is always possible that some new datum will lead us to reject the hypothesis that up until then had withstood all challenges. When we have a set of hypotheses that both has held up in spite of repeated attempts to falB I B L I O G R A PH I C E S S AY S 261

sify it, and makes predictions—which themselves are falsifiable—about other, previously unobserved phenomena, then we may feel secure enough to describe it as a scientific theory. Although scientific theories, including the theory of evolution and the theory of relativity, are, in principle, falsifiable, they are as close to true facts as anything we know about in science. Scientific theories decidedly are not, as popular use and parlance would suggest, unproven speculation. At the time of this writing, the assertion that “eastern hemlock is a foundation species” is a scientific hypothesis in need of testing. Some of the data we have, especially on tree physiology and forest-atmosphere carbon exchange (Hadley 2000a; Hadley and Schedlbauer 2002) and on the fauna associated with hemlock forests (Tingley et al. 2002; Ellison et al. 2005b; Dilling et al. 2007; Mathewson 2009; Rohr et al. 2009; Mallis and Rieske 2011; Sackett et al. 2011) are consistent with this hypothesis (i.e., they would fail to reject it). Other data, especially on nutrient and carbon dynamics in hemlock soils (Cobb and Orwig 2002, 2008; Cobb et al. 2006; Orwig et al. 2008, 2013; Cobb 2010; Templer and McCann 2010) neither strongly support nor strongly reject the hypothesis. But soil dynamics play out not only over years and decades but also over centuries, and it may take equally long to collect sufficient data to mount a convincing challenge to the hypothesis that eastern hemlock is a species like any other. The definition of a foundation species does not say anything about how it should affect biological diversity in the system that it undergirds. Some (hypothesized) foundation species, such as red mangroves in the tropics or poplars in the southwestern United States, create local hotspots of species diversity (Ellison et al. 2005a; Barbour et al. 2009; Ferrier et al. 2012). Both the flora and fauna of eastern hemlock forests, on the other hand, are generally species poor (e.g., Ellison et al. 2005b; Rohr et al. 2009; Sackett et al. 2011). Hemlock forests do support an unusually high diversity of canopy-dwelling, orb-weaving spiders (Mallis and Rieske 2011), and a unique suite of deep-forest birds (Tingley et al. 2002). Biological diversity can be measured at many levels besides species richness in a particular plot or forest stand, however. Biodiversity itself is most often defined by ecologists as the sum total of genetic, species, and ecosystem diversity within a particular region (Hawksworth 1996), but virtually all the information we have for biodiversity in hemlock forests is species diversity at the stand level. Two recent studies have found relatively low genetic variation within populations of eastern hemlock, possibly as a result of a population bottleneck associated with the prehistoric (5,500 years before present; see chapters 3 and 6) hemlock decline (Potter et al. 2008, 2012). At the same time, these genetic studies found high genetic differentiation among stands. This pairing—low within-stand diversity, high between-stand

B I B L I O G R A PH I C E S S AY S 262

diversity—leads to relatively high landscape-level genetic diversity, similar to the landscape-level diversity of forest types that we see across New England. Landscape-level or gamma (γ) diversity equals the product of local or alpha (α) diversity and the variability or beta ( β) diversity between localities in which we measure α diversity (e.g., stands) (see the contributions in Ellison 2010 for detailed discussion). With hemlock forests mixed in with hardwood forests on the landscape, β diversity is relatively high, and so γ diversity is high too. But as hemlock is lost and the forested landscape is homogenized, β diversity declines. If species unique to hemlock forests are lost, average α diversity will decline too. But whether average α diversity declines or remains relatively constant among similar (mixed) hardwood stands, γ diversity is likely to decline as eastern hemlock (or other foundation species) are lost. Because β diversity is harder to measure than α diversity, however, we have much less data on the former than the latter for northeastern U.S. forests, and so there is as yet insufficient data with which to convincingly test the hypothesis that the presence of eastern hemlock significantly increases either β or γ diversity. In summary, our sense remains that eastern hemlock is a foundation species. It passes Potter Stewart’s “I know it” test, and we could probably convince a jury of our (nonscientist) peers. But convincing the scientific community will require yet more data, which we will continue to collect at the Harvard Forest and throughout southern and central New England even as eastern hemlock disappears from our landscape. The lessons learned from our studies of eastern hemlock may yet lead to a general theory of foundation species. L E SSO N S F RO M H A RVA R D F O R E S TS A N D E CO LO G I S TS : I I I . T H E E AR L ST E PH E N S P LOT David R. Foster

Earl Stephens wrote extensively, but little of this made its way into peerreviewed scientific journals. Consequently, the majority of the material presented here was derived from the Harvard Forest archives and various files, journals, correspondence, photographs, and unpublished materials generated by Stephens and Hugh Raup. These were supplemented by extensive conversations that David Foster had with Raup and Ernie Gould and through a series of delightful exchanges of letters with Earl. The most informative paper regarding Earl’s methods and insights is his 1955 “Research in the Biological Aspects of Forest Production.” Here he accomplishes three tasks that are highly relevant to this volume: he describes in fairly good detail the methods that he applied in his forest reconstruction study; he outlines three major approaches through which scientists may develop the kinds of ecological in-

B I B L I O G R A PH I C E S S AY S 263

sights that inform timber production and forest management—empirical observation, experimentation, and reconstruction; and he effectively links a discussion of forest ecology with that of forest management. This paper is as close as Stephens came to realizing his thesis goal of developing a new paradigm in forestry based on the ecological insights emerging form an understanding of forest history. The topic of the uprooting of forest trees is fascinating, both for the role that the uproot mounds and pits may play creating important microhabitats (see Carlton 1993; Carlton and Bazzaz 1998a, b), and for the insights that documenting these features shed on forest dynamics and disturbance history. In addition to Earl’s comprehensive study (Stephens 1956), there was subsequent research on the roles, dynamics, and characteristics of uproot mounds conducted by Harvard Forest scientists Walter Lyford (Lyford and MacLean 1966; Lyford et al. 1963), John Goodlett (Hack and Goodlett 1960), David Foster (1988a), and Peter Schoonmaker (1991). The analysis of Willett Rowland’s data after the 1938 hurricane revealed the surprising insight that the majority of trees uproot rather than break in intense windstorms (Foster 1988b; see also Cooper-Ellis et al. 1999). Correspondence with Earl’s family yielded a marvelous photograph and a comprehensive biographical sketch of all his remarkable activities and achievements following his departure from Petersham. C H A P T E R 1 1 . W H E N D O I NG NOT H I N G I S A V I A B L E A LT E R N AT I V E : I N SI G H TS I NTO CO NSE RVAT I O N A N D M A N AG E M E N T David R. Foster and David A. Orwig

The Harvard Forest approach to management in the face of disturbances such as windstorms and insects has been evolving since the organization was founded in the early 1900s, and it will continue to be refined in the face of new experiences, studies, and discussions with colleagues. Many of the thoughts and conclusions expressed in this chapter emerged from a paper that we wrote for Conservation Biology (Foster and Orwig 2006) as part of a collection of perspectives on salvage logging edited by David Lindenmayer and Jerry Franklin (see also Lindenmayer et al. 2004). Our previous papers and this chapter also draw strongly from Harvard Forest research and real-world experience—especially from the studies by Al Cline, Steve Spurr, David Henry, and others on Pisgah, all the work by Willett Rowlands, Cline, Lutz and others after the 1938 hurricane, and research in the Harvard Forest LTER program. That latter effort sought to contrast the effects of the physical impacts of the hurricane to those chemical and climatic stresses associated with global change and acid rain. The surprising result of our hurricane emulation (Cooper-Ellis et al. 1999; Barker Plotkin et al. 2013) was that, although the forest was physically disrupted

B I B L I O G R A PH I C E S S AY S 264

and had all the appearance of being greatly impaired, its ecosystem processes were much more intact than in the seemingly normal-looking stands altered by climate and chemistry. That work, inspired in many ways by the great watershed manipulation at Hubbard Brook, led to a reevaluation of the 1938 hurricane (Foster et al. 1997, 2004). One of our greatest sources of insights into forest ecology and management has been through interactions with colleagues who visit and share their experiences with us through the Bullard Fellowship program at Harvard University. Hallway conversations, lengthy hikes, and discussions in Petersham, on visits to field sites, or at various LTER venues and conferences can challenge our thinking on many fronts. The greatest insights have come from a remarkable group of collaborators from the Pacific Northwest ( Jerry Franklin, Tom Spies, Fred Swanson, Mark Harmon, and Julia Jones) and a series of visitors from Europe and its islands (George Peterken, Fraser Mitchell, Keith Kirby, Peter Thomas, Matts Lindbladh, Liz Almgren, and Simon Smart) and Australia/New Zealand ( John Ogden, David Bowman, David Lindenmayer, Neal Enright, George Perry, John Herbohn, and Andrew Bennett). Jack Putz from the University of Florida challenged us to reexamine many of our fundamental beliefs with his proposal that, in the face of the threat from the hemlock woolly adelgid, we and other landowners should work collectively to protect a large swath of forest through chemical treatment. Meanwhile, Duncan Stone raised the question that he has been asking in Scotland: if foundation species are lost and have no native substitutes, perhaps there are some useful exotic species that might be imported and planted. In the case of his native Scots pine and our hemlock, Duncan has suggested Norway spruce. Indeed, some Connecticut land management agencies have planted understory spruce in forests where only hardwood species remain after the death of the hemlock. Bernhard Stadler initiated an entirely new focus on biogeochemical changes and processes initiated by the insect itself and returned to help guide our efforts (Stadler et al. 2005, 2006). Meanwhile, ever since his days as a Bullard Fellow from neighboring University of Massachusetts, Dave Kittredge has served as the forest policy analyst at the Harvard Forest, in which capacity he has continually refined our understanding of landowners, their management options, and their responses. Among other activities, Dave collaborated with Dave Orwig in the production of a successful outreach pamphlet on managing hemlock forests, which was distributed to thousands of forest practitioners throughout New England (Orwig and Kittredge 2005). Peter Del Tredici introduced us to the resistant Chinese hemlock and engaged us in the thought-provoking management questions and forest dynamics of the magnificent woods on Hemlock Hill at Boston’s Arnold Arboretum. Debby

B I B L I O G R A PH I C E S S AY S 265

Kaspari brought a new way of viewing our landscape and its rich mixture of natural and cultural elements. Each of these interactions was initiated or advanced through Bullard Fellowships. A considerable amount of our understanding of hemlock management and its consequences has come from concerted efforts to study what was actually happening across the New England landscape. Early work focused on Connecticut illustrated the striking contrasts in the role, magnitude, and trajectory of forest dynamics on sites declining from the adelgid and sites that were harvested (Kizlinski et al. 2002; Cobb et al. 2006; Orwig et al. 2002). Those efforts also led us to discover how differently these processes can play out in adjoining states (Orwig et al. 2012). These studies have been profitably compared and contrasted with the silvicultural approach taken at the Quabbin Watershed, where state managers and scientists have sought to develop a “protection” forest that, through active management, is resilient to natural disturbance, novel threats and climate change. Bruce Spencer, Thom Kyker-Snowman, and Paul Barten have shared their experience and thinking on these forests and that approach through many field trips and enjoyable conversations. Finally, much has been learned in our own backyard, through experiments in the LTER program, through our active management of many forests orchestrated by John Wisnewski and Audrey Barker Plotkin, and through discussions with colleagues Jonathan Thompson, Tony D’Amato, Emery Boose, Betsy Colburn, Kristina Stinson, and Bill Sobczak. L E SSO N S F RO M H A RVA R D F O R E S TS A N D E CO LO G I S TS : I V. T H R E E V I EW S F RO M J O H N S A N D E R S O N ’ S WO O D LOT David R. Foster

In many ways, the story of the Prospect Hill hemlock forest is very much one of people and their use of the land, both as a resource and as an object of study. In colonial times this old forest served as a woodlot, supplying timber, cordwood, and hemlock bark to three generations of Sandersons: Jonathan, son John, and then his wife Lydia along with their son John. The life of this family and the landscape containing the hemlock forest came to represent New England farm and woodland history through Hugh Raup’s classic work “The View from John Sanderson’s Farm” (Raup 1966). A reevaluation of this long-standing perspective was offered by our Brandeis colleague, environmental historian Brian Donahue (2007) in “Another Look from Sanderson’s Farm: A Perspective on New England Environmental History and Conservation.” Raup did not conduct extensive work in New England history, forestry, and ecology himself, but he did develop a fairly comprehensive overview of the ownerB I B L I O G R A PH I C E S S AY S 266

ship and history of the Harvard Forest with Reynold Carlson (1941) and produced many papers based from the original scholarship of his students such as Earl Stephens and other staff, including Ernie Gould, John Goodlett, and Walter Lyford. The role of the tannery operation and other investments including real estate in the financial success of John and the other Sandersons has not been adequately investigated but awaits a good treatment by a student of environmental history or economics. What we do know about the tannery operation was motivated by David Foster’s living in the Lyford House adjacent the tannery site, and his subsequent pursuit of its historical, archaeological, and paleoecological context. Indeed, in some early Harvard Forest ecological studies, such as Steve Spurr’s thesis, the site is incorrectly referred to as a grist mill. The archaeological research is ongoing. An initial survey was completed by the Archaeological Services Group at the University of Massachusetts under Tim Binzen and Mitch Mulholland; this has been continued by a group including Dianna Doucette, Elizabeth Chilton, and Foster, largely through summer field projects. Much more remains to be learned through that effort and further work with the Sanderson Farm account book. The story of the Hemlock Hollow and all the science emerging from the paleoecological, historical, and ecological studies there and in the adjoining hemlock forest is well described in other chapters and essays. L AM E N T David R. Foster

In the fall of 2012 the Harvard Forest LTER program received its fifth round of six-year funding from the National Science Foundation. This work is engaging colleagues and students from around the world to build on the rich legacy of scholarship and policy-relevant research that was initiated by Richard Fisher. A central thrust of the new research is the development of scenarios for New England’s landscape fifty years from now as a consequence of land use and environmental change. That work will draw on the talents of scientists, historians, artists, conservationists, and diverse types of land managers, policy makers, and stakeholders, and it will certainly be informed by the lessons that we have learned from hemlock over the years. As our scientific predecessors have inspired us, we will continue to draw from the past as we envision the future.

B I B L I O G R A PH I C E S S AY S 267

This page intentionally left blank

REFERENCES

Aber, J. D. 1997. Why don’t we believe in models? Bulletin of the Ecological Society of America 78: 232–233. Aber, J. D. 1998. Mostly a misunderstanding, I believe. Bulletin of the Ecological Society of America 78: 256–258. Aber, J. D., and C. A. Federer. 1992. A generalized, lumped-parameter model of photosynthesis, evapotranspiration and net primary production in temperate and boreal forest ecosystems. Oecologia 92: 463–474. Aber, J. D., S. V. Ollinger, C. A. Federer, P. B. Reich, M. L. Goulden, D. W. Kicklighter, J. M. Melillo, and R. G. Lathrop Jr. 1995. Predicting the effects of climate change on water yield and forest production in the Northeastern U.S. Climate Research 5: 207–222. Aber, J. D., P. B. Reich, and M. L. Goulden. 1996. Extrapolating leaf CO₂ exchange to the canopy: a generalized model of forest photosynthesis validated by eddy correlation. Oecologia 106: 257–265. Abrams, M. D., and D. A. Orwig. 1996. A 300-year history of disturbance and canopy recruitment for co-occurring white pine and hemlock on the Allegheny Plateau, USA. Journal of Ecology 84: 353–363. Albani, M., P. R. Moorcroft, A. M. Ellison, D. A. Orwig, and D. R. Foster. 2010. Predicting the impact of hemlock woolly adelgid on carbon dynamics of Eastern U.S. forests. Canadian Journal of Forest Research 40: 119–133. Allen, M. C., J. Sheehan Jr., T. L. Master, and R. S. Mulvihill. 2009. Responses of Acadian flycatchers (Empidonax virescens) to hemlock woolly adelgid (Adelges tsugae) infestation in Appalachian riparian forests. Auk 126: 543–553. Allesina, S., and S. Tang. 2012. Stability criteria for complex ecosystems. Nature 483: 205–208. Alley, R. B., and A. M. Ágústsdóttir. 2005. The 8k event: cause and consequences

269

of a major Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–1149. Alley, R. B., J. Marotzke, W. D. Nordhaus, J. T. Overpeck, D. M. Peteet, R. A. Pielke, R. T. Pierrehumbert, P. B. Rhines, T. F. Stocker, L. D. Talley, and J. M. Wallace. 2003. Abrupt climate change. Science 299: 2005–2010. Alley, R. B., P. A. Mayewski, T. Sowers, M. Stuiver, K. C. Taylor, and P. U. Clark. 1997. Holocene climatic instability: a prominent, widespread event 8200 years ago. Geology 25: 483–486. Allison, T. D., R. E. Moeller, and M. B. Davis. 1986. Pollen in laminated sediments provides evidence for a mid-Holocene forest pathogen outbreak. Ecology 67: 1101–1105. Anderson, R. S., G. L. Jacobson Jr., R. B. Davis, and R. Stuckenrath Jr. 1992. Gould Pond, Maine: late-glacial transitions from marine to upland environments. Boreas 21: 359–371. Barbour, R. C., J. M. O’Reilly-Wapstra, D. W. De Little, G. J. Jordan, D. A. Steane, J. R. Humphreys, J. K. Bailey, T. G. Whitham, and B. M. Potts. 2009. A geographic mosaic of genetic variation within a foundation tree species and its community-level consequences. Ecology 90: 1762–1772. Barker Plotkin, A. A., D. R. Foster, J. Carlson, and A. Magill. 2013. Survivors, not invaders, control forest development following simulated hurricane. Ecology 94: 414–423. Becker, D. A., M. C. Brittingham, and C. B. Goguen. 2008. Effects of hemlock woolly adelgid on breeding birds at Fort Indiantown Gap, Pennsylvania. Northeastern Naturalist 15: 227–240. Bennett, K. D. 1985. The spread of Fagus grandifolia across eastern North America during the last 18 000 years. Journal of Biogeography 12: 147–164. Benzinger, J. 1994a. Hemlock decline and breeding birds. I. Hemlock decline. Records of New Jersey Birds 20: 2–12. ———. 1994b. Hemlock decline and breeding birds. II. Effects of habitat change. Records of New Jersey Birds 20: 34–51. Bernabo, J. C., and T. Webb III. 1977. Changing patterns in the Holocene pollen record from northeastern North America: a mapped summary. Quaternary Research 8: 64–96. Bertness, M. D., and R. M. Callaway. 1994. Positive interactions in communities. Trends in Ecology and Evolution 9: 191–193. Bhiry, N., and L. Filion. 1996. Mid-Holocene hemlock decline in eastern North America linked with phytophagous insect activity. Quaternary Research 45: 312– 320.

REFERENCES 270

Bolgiano, C. 2007. Mighty Giants: An American Chestnut Anthology. American Chestnut Foundation, Bennington, Vermont. Bonneau, L. R., K. S. Shields, and D. L. Civco. 1999. Using satellite images to classify and analyze the health of hemlock forests infested by the hemlock woolly adelgid. Biological Invasions 1: 255–267. Botkin, D. B., J. Janak, and J. Wallis. 1972. Rationale, limitations, and assumptions of a northeastern forest growth simulator. IBM Journal of Research and Development 16: 101–116. Bracken, M. E. S., and N. H. N. Low. 2012. Realistic losses of rare species disproportionately impact higher trophic levels. Ecology Letters 15: 461–467. Brooks, R. T. 2001. Effects of the removal of overstory hemlock from hemlockdominated forests on eastern redback salamanders. Forest Ecology and Management 149: 197–204. Brown, J. P., and P. E. Sendak. 2006. Association of ring shake in eastern hemlock with tree attributes. Forest Products Journal 56: 31–36. Butin, E., A. H. Porter, and J. Elkinton. 2005. Adaptation during biological invasions and the case of Adelges tsugae. Evolutionary Ecology Research 7: 887–900. Campanella, T. J. 2003. Republic of Shade: New England and the American Elm. Yale University Press, New Haven, Connecticut. Carlton, G. C. 1993. Effects of microsite environment on tree regeneration following disturbance. Ph.D. thesis, Harvard University, Cambridge, Massachusetts. Carlton, G. C., and F. A. Bazzaz. 1998a. Regeneration of three sympatric birch species on experimental hurricane blowdown microsites. Ecological Monographs 68: 99–120. Carlton, G. C., and F. A. Bazzaz. 1998b. Resource congruence and forest regeneration following an experimental hurricane blowdown. Ecology 79: 1305–1319. Carpenter, S. R. 2003. Regime Shifts in Lake Ecosystems: Pattern and Variation. Vol. 15, Excellence in Ecology series. International Ecology Institute, Oldendorf/ Luhe, Germany. Catovsky, S., and F. A. Bazzaz. 2000. The role of resource interactions and seedling regeneration in maintaining a positive feedback in hemlock stands. Journal of Ecology 88: 100–112. Catovsky, S., N. M. Holbrook, and F. A. Bazzaz. 2002. Coupling whole-tree transpiration and canopy photosynthesis in coniferous and broad-leaved tree species. Canadian Journal of Forest Research 32: 295–309. Clark, J., C. Fastie, G. C. Hurtt, S. J. Jackson, W. C. Johnson, G. King, M. Lewis, J. Lynch, S. W. Pacala, I. C. Prentice, E. Schupp, T. Webb III, and P. Wyckoff. 1998. Dispersal theory offers solutions to Reid’s Paradox of rapid plant migration. Bioscience 48: 13–24. REFERENCES 271

Clark, J. T., S. Fei, L. Liang, and L. K. Rieske. 2012. Mapping eastern hemlock: comparing classification techniques to evaluate susceptibility of a fragmented and valued resource to an exotic invader, the hemlock woolly adelgid. Forest Ecology and Management 266: 216–222. Clepper, H. E. 1944. Hemlock, the state tree of Pennsylvania. Bulletin no. 52. Pennsylvania Department of Forests and Waters. Harrisburg, Pennsylvania. Cline, A. C., and S. H. Spurr. 1942. The virgin upland forest of central New England: a study of old growth stands in the Pisgah mountain section of southwestern New Hampshire. Harvard Forest Bulletin no. 21: 1–58. Petersham, Massachusetts. Cobb, R. C. 2010. Species shift drives decomposition rates following invasion by hemlock woolly adelgid. Oikos 119: 1291–1298. Cobb, R. C., and D. A. Orwig. 2002. Impacts of hemlock woolly adelgid on decomposition: an overview. Pages 317–323 in R. C. Reardon, B. P. Onken, and J. Lashomb (eds.), Proceedings of the Symposium on the Hemlock Woolly Adelgid in Eastern North America. New Jersey Agricultural Experiment Station, New Brunswick, New Jersey. Cobb, R. C., and D. A. Orwig. 2008. Changes in decomposition dynamics in hemlock forests impacted by hemlock woolly adelgid: restoration and conservation of hemlock ecosystem function. Pages 157–167 in B. Onken and R. Reardon (eds.), Proceedings of the Fourth Symposium on Hemlock Woolly Adelgid in the Eastern United States. U.S. Forest Service, Morgantown, West Virginia. Cobb, R. C., D. A. Orwig, and S. J. Currie. 2006. Decomposition of green foliage in eastern hemlock forests of southern New England impacted by hemlock woolly adelgid infestations. Canadian Journal of Forest Research 36: 1–11. Cogbill, C. V., J. Burk, and G. Motzkin. 2002. The forests of presettlement New England, USA: spatial and compositional patterns based on town proprietor surveys. Journal of Biogeography 29: 1279–1304. Collins, B., W. V. Sobczak, and E. A. Colburn. 2007. Subsurface flowpaths in a forested headwater stream harbor a diverse macroinvertebrate community. Wetlands 27: 319–325. Cooper-Ellis, S., D. R. Foster, G. Carlton, and A. Lezberg. 1999. Forest response to catastrophic wind: results from an experimental hurricane. Ecology 80: 2683– 2696. Costa, S., and B. Onken. 2006. Standardizing sampling for detection and monitoring of hemlock woolly adelgid in eastern hemlock forests. U.S. Forest Service, Forest Health Technology Enterprise Team, FHTET-2006-16, Morgantown, West Virginia. Cowles, R. S. 2009. Optimizing dosage and preventing leaching of imidacloprid for REFERENCES 272

management of hemlock woolly adelgid in forests. Forest Ecology and Management 257: 1026–1033. Dale, V., and W. Van Winkle. 1998. Models provide understanding, not belief. Bulletin of the Ecological Society of America 79: 169–170. Daley, M. J., N. G. Phillips, J. C. Pettijohn, and J. L. Hadley. 2007. Water use by eastern hemlock (Tsuga canadensis) and black birch (Betula lenta): implications of effects of the hemlock woolly adelgid. Canadian Journal of Forest Research 37: 2031–2040. Daley, R. K., W. C. Branch, and T. Lotti. 1930. Life history of the climax forest on the Pisgah Tract, Winchester, New Hampshire. M.F.S. thesis, Harvard University, Cambridge, Massachusetts. D’Amato, A. W., and D. A. Orwig. 2008. Stand and landscape-level disturbance dynamics in old-growth forests in western Massachusetts. Ecological Monographs 78: 507–522. D’Amato, A. W., D. A. Orwig, and D. R. Foster. 2006. New estimates of Massachusetts old-growth forests: useful data for regional conservation and forest planning. Northeastern Naturalist 13: 495–506. ———. 2008. The influence of successional processes and disturbance on the structure of Tsuga canadensis forests. Ecological Applications 18: 1182–1199. ———. 2009. Understory vegetation in old-growth and second-growth Tsuga canadensis forests in western Massachusetts. Forest Ecology and Management 257: 1043–1052. Daubenmire, R. F. 1930. The relation of certain ecological factors to the inhibition of forest floor herbs under hemlock. Butler University Botanical Studies 1: 61–76. Davis, Margaret B. 1969. Climatic changes in southern Connecticut recorded by pollen deposition at Rogers Lake. Ecology 50: 409–422. Davis, Margaret B. 1981a. Outbreaks of forest pathogens in Quaternary history. Pages 216–227 in D. Bharadwaj, Vishnu-Mittre, and H. Maheshwari (eds.), Proceedings of the Fourth International Palynological Conference, Birbal Sahni Institute of Paleobotany, Lucknow, India. Davis, Margaret B. 1981b. Quaternary history and the stability of forest communities. Pages 132–153 in D. C. West, H. H. Shugart, and D. B. Botkin (eds.), Forest Succession Concepts and Application. Springer, New York. Davis, Margaret B. 1983. Holocene vegetational history of the eastern United States. Pages 166–181 in H. E. Wright Jr. (ed.), Late-Quaternary Environments of the United States, vol. 2: The Holocene. University of Minnesota Press, Minneapolis. Davis, Margaret B., and J. C. Goodlett. 1960. Comparison of the present vegetation with pollen-spectra in surface samples from Brownington Pond, Vermont. Ecology 41: 346–357. REFERENCES 273

Davis, Mary B. 1996. Eastern Old-Growth Forests: Prospects for Rediscovery and Recovery. Island Press, Washington, D.C. Dayton, P. K. 1972. Toward an understanding of community resilience and the potential effects of enrichment to the benthos at McMurdo Sound, Antarctica. Pages 81–95 in B. C. Parker (ed.), Proceedings of the Colloquium on Conservation Problems in Antarctica. Allen Press, Lawrence, Kansas. Deevey, E. S., Jr. 1939. Studies of Connecticut lake sediments. I. A postglacial climatic chronology for southern New England. American Journal of Science 237: 691–724. ———. 1951. Late-glacial and postglacial pollen diagrams from Maine. American Journal of Science 249: 117–207. Dilling, C., P. Lambdin, J. Grant, and L. Buck. 2007. Insect guild structure associated with eastern hemlock in the southern Appalachians. Environmental Entomology 36: 1408–1414. Dobson, A. P., and P. J. Hudson. 1986. Parasites, disease and the structure of ecological communities. Trends in Ecology and Evolution 1: 11–15. Donahue, B. 2007. Another look from Sanderson’s Farm: A perspective on New England environmental history and conservation. Environmental History 12: 9–34. Doyle, Sir A. C. 1890. The Sign of Four. Spencer Blackett, London. Dugan, C. 2004. Saving our hemlock forests. Massachusetts Wildlife 54: 24–30. Dukes, J. S., J. Pontius, D. A. Orwig, J. R. Garnes, V. L. Rodgers, N. Brazee, B. Cooke, K. A. Theoharides, E. E. Stange, R. Harrington, J. Ehrenfeld, J. Gurevitch, M. T. Lerdau, K. A. Stinson, R. Wick, and M. Ayres. 2009. Responses of insect pests, pathogens, and invasive plant species to climate change in the forests of northeastern North America: what can we predict? Canadian Journal of Forest Research 39: 231–248. Dunne, J. A., R. J. Williams, and N. D. Martinez. 2002. Network structure and biodiversity loss in food webs: robustness increases with connectance. Ecology Letters 5: 558–567. Dunwiddie, P., D. R. Foster, D. J. Leopold, and R. Leverett. 1996. Old-growth forests of southern New England, New York, and Pennsylvania. Pages 126–143 in Mary B. Davis (ed.), Eastern Old-Growth Forests: Prospects for Rediscovery and Recovery, Island Press, Washington, D.C. Ellison, A. M. 2010. Partitioning diversity [introduction to the Forum]. Ecology 91: 1962–1963. Ellison, A. M., M. S. Bank, B. D. Clinton, E. A. Colburn, K. Elliott, C. R. Ford, D. R. Foster, B. D. Kloeppel, J. D. Knoepp, G. M. Lovett, J. Mohan, D. A. Orwig, N. L. Rodenhouse, W. V. Sobczak, K. A. Stinson, J. K. Stone, C. M.

REFERENCES 274

Swan, J. Thompson, B. Von Holle, and J. R. Webster. 2005a. Loss of foundation species: consequences for the structure and dynamics of forested ecosystems. Frontiers in Ecology and the Environment 9: 479–486. Ellison, A. M., A. Barker Plotkin, D. R. Foster, and D. A. Orwig. 2010. Experimentally testing the role of foundation species in forests: the Harvard Forest hemlock removal experiment. Methods in Ecology and Evolution 1: 168–179. Ellison, A. M., J. Chen, D. Díaz, C. Kammerer-Burnham, and M. Lau. 2005b. Changes in ant community structure and composition associated with hemlock decline in New England. Pages 280–289 in B. Onken and R. Reardon (eds.), Third Symposium on Hemlock Woolly Adelgid in the Eastern United States. U.S. Forest Service, Morgantown, West Virginia. Evans, D. M., W. M. Aust, C. A. Dolloff, B. S. Templeton, and J. A. Peterson. 2011. Eastern hemlock decline in riparian areas from Maine to Alabama. Northern Journal of Applied Forestry 28: 97–104. Ewers, F. W., and M. H. Zimmermann. 1984. The hydraulic architecture of eastern hemlock. Canadian Journal of Botany 62: 940–946. Faison, E. K., D. R. Foster, W. W. Oswald, E. D. Doughty, and B. C. S. Hansen. 2006. Early Holocene openlands in southern New England. Ecology 87: 2537–2547. Farnsworth, E. J., A. Barker-Plotkin, and A. M. Ellison. 2012. Seed bank, seed rain, and vegetation dynamics in Tsuga canadensis stands: reorganization of forests after foundation species loss due to logging or simulated attack of Adelges tsugae. Canadian Journal of Forest Research 42: 2090–2105. Ferrier, S. M., R. K. Bangert, E. I. Hersch-Green, J. K. Bailey, G. J. Allan, and T. G. Whitham. 2012. Unique arthropod communities on different host-plant genotypes results in greater arthropod diversity. Arthropod-Plant Interactions 6: 187–195. Fisher, R. A., A. S. Corbet, and C. B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample from an animal population. Journal of Animal Ecology 12: 42–58. Fisher, R. T. 1927. A museum of forest antiquity. American Forestry and Forest Life 33: 348. ———. 1928. Soil changes and silviculture of the Harvard Forest. Ecology 9: 6–11. ———. 1933. New England forests: biological factors. New England’s Prospect: 1933. American Geographical Society special publication 16: 213–223. Fitzpatrick, M. C., E. L. Preisser, A. Porter, J. S. Elkinton, and A. M. Ellison. 2012. Modeling range dynamics in heterogeneous landscapes: invasion of the hemlock woolly adelgid in eastern North America. Ecological Applications 22: 472– 486.

REFERENCES 275

Foster, D. R. 1988a. Disturbance history, community organization and vegetation dynamics of the old-growth Pisgah Forest, southwestern New Hampshire, U.S.A. Journal of Ecology 76: 105–134. ———. 1988b. Species and stand response to catastrophic wind in central New England, U.S.A. Journal of Ecology 76: 135–151. ———. 1999. Thoreau’s Country: Journey through a Transformed Landscape. Harvard University Press, Cambridge, Massachusetts. ———. 2000. Hemlock’s future in the context of its history: an ecological perspective. Pages 1–4 in K. A. McManus, K. S. Shields, and D. R. Souto (eds.), Proceedings of the Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America. U.S. Forest Service, General Technical Report NE-267, Newtown Square, Pennsylvania. ———. 2002. Thoreau’s country: a historical-ecological perspective to conservation in the New England landscape. Journal of Biogeography 29: 1537–1555. Foster, D. R., and J. D. Aber, eds. 2004. Forests in Time: The Environmental Consequences of 1,000 Years of Change in New England. Yale University Press, New Haven, Connecticut. Foster, D. R., J. D. Aber, J. M. Melillo, R. Bowden, and F. Bazzaz. 1997. Forest response to disturbance and anthropogenic stress: rethinking the 1938 hurricane and the impact of physical disturbance versus chemical and climate stress on forest ecosystems. BioScience 47: 437–445. Foster, D. R., and E. R. Boose. 1992. Patterns of forest damage resulting from catastrophic wind in central New England, USA. Journal of Ecology 80: 79–99. Foster, D. R., S. Clayden, D. A. Orwig, B. Hall, and S. Barry. 2002a. Oak, chestnut and fire: climatic and cultural controls of long-term forest dynamics in New England, USA. Journal of Biogeography 29: 1359–1379. Foster, D. R., B. Hall, S. Barry, S. Clayden, and T. Parshall. 2002b. Cultural, environmental, and historical controls of vegetation patterns and the modern conservation setting on the island of Martha’s Vineyard, USA. Journal of Biogeography 29: 1381–1400. Foster, D. R., G. Motzkin, J. O’Keefe, E. R. Boose, D. A. Orwig, J. L. Fuller, and B. Hall. 2004. The environmental and human history of New England. Pages 43–100 in D. R. Foster and J. D. Aber (eds.), Forests in Time: The Environmental Consequences of 1,000 Years of Change in New England. Yale University Press, New Haven, Connecticut. Foster, D. R., and J. F. O’Keefe. 2000. New England Forests through Time. Harvard University Press, Cambridge, Massachusetts. Foster, D. R., and D. A. Orwig. 2006. Pre-emptive and salvage harvesting of New

REFERENCES 276

England forests: when doing nothing is a viable alternative. Conservation Biology 20: 959–970. Foster, D. R., W. W. Oswald, E. K. Faison, E. D. Doughty, and B. C. S. Hansen. 2006. A climatic driver for abrupt mid-Holocene vegetation dynamics and the hemlock decline in New England. Ecology 87: 2959–2966. Foster, D. R., P. K. Schoonmaker, and S. T. A. Pickett. 1990. Insights from paleoecology to community ecology. Trends in Ecology and Evolution 5: 119–122. Foster, D. R., and T. M. Zebryk. 1993. Long-term vegetation dynamics and disturbance history of a Tsuga-dominated forest in New England. Ecology 74: 982– 998. Foster, D. R., T. M. Zebryk, P. K. Schoonmaker, and A. L. Lezberg. 1992. Postsettlement history of human land-use and vegetation dynamics of a hemlock woodlot in central New England. Journal of Ecology 80: 773–786. Freinkel, S. 2007. American Chestnut: The Life, Death, and Rebirth of a Perfect Tree. University of California Press, Berkeley, California. Frelich, L. E., and C. G. Lorimer. 1991. Natural disturbance regimes in hemlockhardwood forests of the upper Great Lakes region. Ecological Monographs 61: 145–164. Frothingham, E. H. 1915. The eastern hemlock. Bulletin of the U.S. Department of Agriculture 152: 1–43. Fuller, J. L. 1997. Holocene forest dynamics in southern Ontario, Canada: fineresolution pollen data. Canadian Journal of Botany 75: 1714–1727. ———. 1998. Ecological impact of the mid-Holocene hemlock decline in southern Ontario, Canada. Ecology 79: 2337–2351. Fuller, J. L., D. R. Foster, J. S. McLachlan, and N. Drake. 1998. Impact of human activity on regional forest composition and dynamics in central New England. Ecosystems 1: 76–95. Fuller, J. L., D. R. Foster, G. Motzkin, J. McLachlan, and S. Barry. 2004. Broadscale forest response to land-use and climate change. Pages 101–124 in D. R. Foster and J. D. Aber (eds.), Forests in Time: The Environmental Consequences of 1,000 Years of Change in New England. Yale University Press, New Haven, Connecticut. Garrett, J. D. 2002. Avifauna of hemlock-dominated natural communities of the Quabbin Watershed in Massachusetts. M.S. thesis, University of Massachusetts, Amherst. Gaston, K. J., and R. A. Fuller. 2007. Biodiversity and extinction: losing the common and the widespread. Progress in Physical Geography 31: 213–225. ———. 2008. Commonness, population depletion and conservation biology. Trends in Ecology and Evolution 23: 14–19. REFERENCES 277

Godwin, H. 1935. Post-Glacial floras: vegetation phases reconstructed from pollenanalysis of peat. Discussion on the origin and relationship of the British flora. Proceedings of the Royal Society of London B 118: 210–215. ———. 1946. The relationship of bog stratigraphy to climatic change and archaeology. Proceedings of the Prehistorical Society 12: 1–11. Gomez, S., C. M. Orians, and E. L. Preisser. 2012. Exotic herbivores on a shared native host: tissue quality after individual, simultaneous, and sequential attack. Oecologia 169: 1015–1024. Gonda-King, L., L. Radville, and E. L. Preisser. 2012. False ring formation in eastern hemlock branches: impacts of hemlock woolly adelgid and elongate hemlock scale. Environmental Entomology 41: 523–531. Gotelli, N. J., and A. M. Ellison. 2012. A Primer of Ecological Statistics, 2nd ed. Sinauer Associates, Sunderland, Massachusetts. Grime, J. P. 1987. Dominant and subordinate components of plant communities: implications for succession, stability and diversity. Pages 413–428 in A. J. Gray, M. J. Crawley, and P. J. Edwards (eds.), Colonization, Succession and Stability. Blackwell Press, Oxford, U.K. Hack, J. T., and J. Goodlett. 1960. Geomorphology and forest ecology of a mountain region in the central Appalachians. USGS Professional Paper 347: 66. Hadley, J. L. 2000a. Effect of daily minimum temperature on photosynthesis in eastern hemlock (Tsuga canadensis L.) in autumn and winter. Arctic, Antarctic, and Alpine Research 32: 368–374. ———. 2000b. Understory microclimate and photosynthetic response of saplings in an old-growth eastern hemlock (Tsuga canadensis L.) forest. Ecoscience 7: 66–72. Hadley, J. L., P. S. Kuzeja, M. J. Daley, N. G. Phillips, T. Mulcahy, and S. Singh. 2008. Water use and carbon exchange of red oak– and eastern hemlock–dominated forests in the northeastern U.S.: implications for ecosystem-level effects of the hemlock woolly adelgid. Tree Physiology 28: 614–627. Hadley, J. L., and J. L. Schedlbauer. 2002. Carbon exchange of an old-growth eastern hemlock (Tsuga canadensis) forest in central New England. Tree Physiology 22: 1079–1092. Hall, B., G. Motzkin, D. R. Foster, M. Syfert, and J. Burk. 2002. Three hundred years of forest and land-use change in Massachusetts, USA. Journal of Biogeography 29: 1319–1335. Hanski, I. 1982. Dynamics of regional distribution: the core and satellite species hypothesis. Oikos 38: 210–221. Havill, N., M. E. Montgomery, and M. Keena. 2011. Hemlock woolly adelgid and

REFERENCES 278

its hemlock hosts: a global perspective. Pages 3–14 in B. Onken and R. Reardon (eds.), Implementation and Status of Biological Control of the Hemlock Woolly Adelgid. U.S. Forest Service, Forest Health Technology Enterprise Team (FHTET)2011-04, Morgantown, West Virginia. Havill, N. P., M. E. Montgomery, G. Yu, S. Shiyake and A. Caccone. 2006. Mitochondrial DNA from hemlock woolly adelgid (Hemiptera: Adelgidae) suggests cryptic speciation and pinpoints the source of the introduction to eastern North America. Annals of the Entomological Society of America 99: 195–203. Hawksworth, D. L. 1996. Biodiversity: Measurement and Estimation. Chapman & Hall, London. Henry, J. D., and J. M. A. Swan. 1974. Reconstructing forest history from live and dead plant material: an approach to the study of forest succession in southwest New Hampshire. Ecology 55: 772–783. Hibbs, D. E. 1981a. Development of the forest canopy: a key to silviculture. Northern Logger and Timber Processor 29: 62–78. ———. 1981b. Leader growth and the architecture of three North American hemlocks. Canadian Journal of Botany 59: 476–480. ———. 1982. Gap dynamics in a hemlock-hardwood forest. Canadian Journal of Forest Research 12: 522–527. Hilborn, R., and M. Mangel. 1997. The Ecological Detective: Confronting Models with Data. Princeton University Press, Princeton, New Jersey. Holling, C. S. 1992. Cross-scale morphology, geometry, and dynamics of ecosystems. Ecological Monographs 62: 447–502. Holmes, T., A. M. Liebhold, K. F. Kovacs, and B. Von Holle. 2010. A spatialdynamic value transfer model of economic losses from a biological invasion. Ecological Economics 70: 86–95. Hou, J., Y. Huang, B. N. Shuman, W. W. Oswald, and D. R. Foster. 2011. Abrupt cooling repeatedly punctuated early-Holocene climate in eastern North America. The Holocene 22: 525–529. Huang, Y., B. Shuman, Y. Wang, and T. Webb III. 2002. Hydrogen isotope ratios of palmitic acid in lacustrine sediments record late Quaternary climate conditions. Geology 30: 1103–1106. Hubbell, C. S. 2001. A Unified Theory of Biodiversity and Biogeography. Princeton University Press, Princeton, New Jersey. Huntley, B., and T. Webb III. 1989. Migration: species’ response to climatic variations caused by changes in the earth’s orbit. Journal of Biogeography 16: 5–19. Huston, M. A. 1994. Biological Diversity: The Coexistence of Species on Changing Landscapes. Cambridge University Press, New York.

REFERENCES 279

Ingwell, L., J. Brady, M. Fitzpatrick, B. Maynard, R. Casagrande, and E. Preisser. 2009. Intraspecific variation in Tsuga canadensis foliar chemistry. Northeastern Naturalist 16: 585–594. Ireland, A. W., B. J. Mew, and D. R. Foster. 2008. Bob Marshall’s forest reconstruction study: three centuries of ecological resilience to disturbance. Journal of the Torrey Botanical Society 135: 411–422. Iversen, J. 1944. Viscum, Hedera and Ilex as climate indicators: a contribution to the study of the post-glacial temperature climate. Geologiska Föreningens i Stockholm Förhandlingar 66:463–483. Johnstone, J. F., T. N. Hollingsworth III, F. S. Chapin, and M. C. Mack. 2010. Changes in fire regime break the legacy lock on successional trajectories in Alaskan boreal forest. Global Change Biology 16: 1281–1295. Jones, C. G., J. H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers. Oikos 69: 373–386. Kéfi, S., E. L. Berlow, E. A. Wieters, S. A. Navarrete, O. L. Petchey, S. A. Wood, A. Boit, L. N. Joppa, K. D. Lafferty, R. J. Williams, N. D. Martinez, B. A. Menge, C. A. Blanchette, A. C. Iles, and U. Brose. 2012. More than a meal . . . integrating non-feeding interactions into food webs. Ecology Letters 15: 291–300. Kelty, M. J. 1984. The development and productivity of hemlock-hardwood forests in southern New England. Ph.D. thesis, Yale University, New Haven, Connecticut. ———. 1986. Development patterns in two hemlock-hardwood stands in southern New England. Canadian Journal of Forest Research 16: 885–891. ———. 1989. Productivity of New England hemlock/hardwood stands as affected by species composition and canopy structure. Forest Ecology and Management 28: 237–257. Kelty, M. J., and E. M. Gould Jr. 1987. Effects of understory removal in hardwood stands. Northern Journal of Applied Forestry 4: 162–164. Kizlinski, M. L. 2002. Vegetation and ecosystem response to eastern hemlock decline and logging: direct and indirect consequences of the hemlock woolly adelgid. M.F.S. thesis, Harvard University, Cambridge, Massachusetts. Kizlinski, M. L., D. A. Orwig, R. C. Cobb, and D. R. Foster. 2002. Direct and indirect ecosystem consequences of an invasive pest on forests dominated by eastern hemlock. Journal of Biogeography 29: 1489–1503. Knapp, A. K., J. M. Briggs, D. C. Hartnett, and S. L. Collins. 1998. Grassland Dynamics: Long-Term Ecological Research in Tallgrass Prairie. Oxford University Press, New York. Koch, F. H., H. M. Cheshire, and H. A. Devine. 2006. Landscape-scale prediction of hemlock woolly adelgid, Adelges tsugae (Homoptera: Adelgidae), infesREFERENCES 280

tation in the southern Appalachian Mountains. Environmental Entomology 35: 1313–1323. Kurz, W. A., C. C. Dymond, G. Stinson, G. J. Rampley, E. T. Nielson, A. L. Carroll, T. Ebata, and L. Safranyik. 2008. Mountain pine beetle and forest carbon feedback to climate change. Nature 452: 987–990. Liebhold, A. M., W. L. MacDonald, D. Bergdahl, and V. C. Mastro. 1995. Invasion by exotic forest pests: a threat to forest ecosystems. Forest Science Monograph 30: 1–49. Likens, G. E., and F. H. Bormann. 1995. Biogeochemistry of a Forested Ecosystem, 2nd ed. Springer, New York. Lindbladh, M., W. W. Oswald, D. R. Foster, E. K. Faison, J. Hou, and Y. Huang. 2007. A late-glacial transition from Picea glauca to Picea mariana in southern New England. Quaternary Research 67: 502–508. Lindenmayer, D. B., D. R. Foster, J. F. Franklin, M. L. Hunter, R. F. Noss, F. A. Schmiegelow, and D. Perry. 2004. Salvage harvesting policies after natural disturbance. Science 303: 1303. Livingstone, D. A. 1955. A lightweight piston corer for sampling lake sediments. Ecology 36: 137–139. Lorimer, C. G. 1985. Methodological considerations in the analysis of forest disturbance history. Canadian Journal of Forest Research 15: 200–213. ———. 1995. Dynamics and structural characteristics of eastern hemlock stands. Pages 43–59 in G. Mroz and A. J. Martin (eds.), Proceedings of the Conference on Hemlock Ecology and Management. Department of Forestry, University of Wisconsin-Madison. Lose, C. 1931. The Vanishing Trout: A Study of Trout and Trout Fishing in the Waters of Central Pennsylvania. Altoona Tribune Press, Altoona, Pennsylvania. Lustenhouwer, M. N., L. Nicoll, and A. M. Ellison. 2012. Microclimatic effects of the loss of a foundation species from New England forests. Ecosphere 3: 26. Lutz, H. J., and A. C. Cline. 1956. Results of the first thirty years of experimentation in silviculture in the Harvard Forest, 1908–1938. Part II. Natural reproduction methods in white pine–hemlock stands on light, sandy soils. Harvard Forest Bulletin no. 27: 5–69. Petersham, Massachusetts. Lyford, W. H., J. C. Goodlett, and W. H. Coates. 1963. Landforms, soils with fragipans, and forest on a slope in the Harvard Forest. Harvard Forest Bulletin no. 30: 1–29. Petersham, Massachusetts. Lyford, W. H., and D. W. MacLean. 1966. Mound and pit microrelief in relation to soil disturbance and tree distribution in New Brunswick, Canada. Harvard Forest Paper no. 15: 1–17. Petersham, Massachusetts.

REFERENCES 281

MacArthur, R. 1960. On the relative abundance of species. American Naturalist 94: 25–36. Magill, A. H., J. D. Aber, W. S. Currie, K. J. Nadelhoffer, M. E. Martin, W. H. McDowell, J. M. Melillo, and P. A. Steudler. 2004. Ecosystem response to 15 years of chronic nitrogen additions at the Harvard Forest LTER, Massachusetts, USA. Forest Ecology and Management 196: 7–28. Mallis, R. E., and L. K. Rieske. 2011. Arboreal spiders in eastern hemlock. Environmental Entomology 40: 1378–1387. Marshall, R. 1927. The growth of hemlock before and after the release from suppression. Harvard Forest Bulletin no. 11: 1–43. Petersham, Massachusetts. Martin, P. S. 1973. The discovery of America. Science 179: 969–974. Mathewson, B. 2006. Differences in eastern red-back salamander populations in eastern hemlock–dominated and mixed deciduous forest in north central Massachusetts. M.F.S. thesis, Harvard University, Cambridge, Massachusetts. ———. 2009. The relative abundance of eastern red-backed salamanders in eastern hemlock–dominated and mixed deciduous forests at Harvard Forest. Northeastern Naturalist 16: 1–12. May, R. M. 1973. Stability and Complexity in Model Ecosystems. Princeton University Press, Princeton, New Jersey. McClure, M. S. 1979. Self-regulation in populations of the elongate hemlock scale, Fiorinia externa (Homoptera: Diaspididae). Oecologia 39: 25–36. ———. 1987. Biology and control of hemlock woolly adelgid. Connecticut Agricultural Experiment Station Bulletin no. 851, New Haven, Connecticut. ———. 1989a. Evidence of a polymorphic life cycle in the hemlock woolly adelgid, Adelges tsugae (Homoptera: Adelgidae). Annals of the Entomological Society of America 82: 52–54. ———. 1989b. Importance of weather to the distribution and abundance of introduced adelgid and scale insects. Agricultural and Forest Meteorology 47: 291–302. ———. 1990. Role of wind, birds, deer, and humans in the dispersal of hemlock woolly adelgid (Homoptera: Adelgidae). Environmental Entomology 19: 36–43. ———. 1991. Density-dependent feedback and population cycles in Adelges tsugae (Homoptera: Adelgidae) on Tsuga canadensis. Environmental Entomology 20: 258–264. ———. 1995. Managing hemlock woolly adelgid in ornamental landscapes. Connecticut Agricultural Experiment Station Bulletin no. 925, New Haven, Connecticut. ———. 2001. Biological control of hemlock woolly adelgid in the eastern United States. U.S. Forest Service, Forest Health Technology Enterprise Team FHTET2000-08, Morgantown, West Virginia. REFERENCES 282

McClure, M. S., and J. D. Hare. 1984. Foliar terpenoids in Tsuga species and the fecundity of scale insects. Oecologia 63: 185–193. McLachlan, J., D. R. Foster, and F. Menalled. 2000. Anthropogenic ties to latesuccessional structure and composition in four New England hemlock stands. Ecology 81: 717–733. McWilliams, W. H., and T. L. Schmidt. 2000. Composition, structure, and sustainability of hemlock ecosystems in eastern North America. Pages 5–10 in K. McManus, K. Shields, and D. Souto (eds.), Proceedings of the Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America. U.S. Forest Service, General Technical Report NE-267, Newtown Square, Pennsylvania. Melillo, J. M., S. Butler, J. Johnson, J. Mohan, P. A. Steudler, H. Lux, E. Burrows, F. P. Bowles, R. Smith, L. Scott, C. Vario, T. Hill, A. J. Burton, Y. Zhou, and J. Tang. 2011. Soil-warming carbon–nitrogen interactions and forest carbon budgets. Proceedings of the National Academy of Sciences 108: 9508–9512. Melillo, J. M., A. D. McGuire, D. W. Kicklighter, B. Moore III, C. J. Vorosmarty, and A. L. Schloss. 1993. Global climate change and terrestrial net primary production. Nature 363: 234–240. Merrill, P. H., and R. C. Hawley. 1924. Hemlock: its place in the silviculture of the southern New England forest. Yale University School of Forestry Bulletin no. 12, New Haven, Connecticut. Meryman, R. S. 1999. A painter of angels became the father of camouflage. Smithsonian Magazine, April, 116–128. Miller-Pierce, M. R., D. A. Orwig, and E. L. Preisser. 2010. Effects of hemlock woolly adelgid and elongate hemlock scale on eastern hemlock growth and foliar chemistry. Environmental Entomology 39: 513–519. Miller-Pierce, M. R., and E. L. Preisser. 2012. Asymmetric priority effects influence the success of invasive forest insects. Ecological Entomology 37: 350–358. Munoz, S. E., K. Gajewski, and M. C. Peros. 2010. Synchronous environmental and cultural change in the prehistory of the northeastern United States. Proceedings of the National Academy of Sciences 107: 22008–22013. Namias, J. 1966. Nature and possible causes of the northeastern United States drought during 1962–65. Monthly Weather Reviews 94: 543–554. Narayanaraj, G., P. V. Bolstad, K. J. Elliott, and J. M. Vose. 2010. Terrain and landform influence on Tsuga canadensis (L.) Carrière (eastern hemlock) distribution in the southern Appalachian mountains. Castanea 75: 1–18. Newby, P., J. P. Donnelly, and B. Shuman. 2009. Evidence of centennial-scale drought from southeastern Massachusetts during the Pleistocene/Holocene transition. Quaternary Science Reviews 28: 1675–1692. REFERENCES 283

Odling-Smee, F. J., K. N. Laland, and M. W. Feldman. 2003. Niche Construction: The Neglected Process in Evolution. Princeton University Press, Princeton, New Jersey. Oliver, C. D. 1975. The development of northern red oak (Quercus rubra L.) in mixed species, even-age stands in central New England. Ph.D. thesis, Yale University, New Haven, Connecticut. ———. 1978. The development of northern red oak in mixed stands in central New England. Yale School of Forestry and Environmental Studies Bulletin no. 91, New Haven, Connecticut. Oliver, C. D., and E. P. Stephens. 1977. Reconstruction of a mixed-species forest in central New England. Ecology 58: 562–572. Onken, B., and R. C. Reardon. 2011. Implementation and status of biological control of the hemlock woolly adelgid. U.S. Forest Service, Forest Health Technology Enterprise Team FHTET-2011-04, Morgantown, West Virginia. Orwell, G. 1945. Animal Farm. Secker and Warburg, London. Orwig, D. A. 2002. Stand dynamics associated with chronic hemlock woolly adelgid infestations in southern New England. Pages 36–46 in R. C. Reardon, B. P. Onken, and J. Lashomb (eds.), Proceedings of the Symposium on the Hemlock Woolly Adelgid in Eastern North America, New Jersey Agricultural Experiment Station, New Brunswick, New Jersey. ———. 2008. Eastern hemlock: irreplaceable habitat. Massachusetts Sierran, Summer, 3–5. Orwig, D. A., A. A. Barker Plotkin, E. A. Davidson, H. Lux, K. A. Savage, and A. M. Ellison. 2013. Foundation species loss affects vegetation structure more than ecosystem function in a northeastern USA forest. PeerJ 1: e41. Orwig, D. A., R. C. Cobb, A. W. D’Amato, M. L. Kizlinski, and D. R. Foster. 2008. Multi-year ecosystem response to hemlock woolly adelgid infestation in southern New England forests. Canadian Journal of Forest Research 38: 834–843. Orwig, D. A., C. V. Cogbill, D. R. Foster, and J. F. O’Keefe. 2001. Variations in old-growth structure and definitions: forest dynamics on Wachusett Mountain, Massachusetts. Ecological Applications 11: 437–452. Orwig, D. A., and D. R. Foster. 1998. Forest response to the introduced hemlock woolly adelgid in southern New England, USA. Journal of the Torrey Botanical Society 125: 59–72. Orwig, D. A., D. R. Foster, and D. L. Mausel. 2002. Landscape patterns of hemlock decline in New England due to the introduced hemlock woolly adelgid. Journal of Biogeography 29: 1475–1487. Orwig, D. A., and D. B. Kittredge. 2005. Silvicultural options for managing hemlock forests threatened by hemlock woolly adelgid. Pages 212–217 in B. Onken REFERENCES 284

and R. Reardon (eds.), Third Symposium on Hemlock Woolly Adelgid in the Eastern United States, U.S. Forest Service, Morgantown, West Virginia. Orwig, D. A., J. R. Thompson, N. A. Povak, M. Manner, D. Niebyl, and D. R. Foster. 2012. A foundation tree at the precipice: Tsuga canadensis health after the arrival of Adelges tsugae in central New England. Ecosphere 3: 10. Oswald, W. W., E. K. Faison, D. R. Foster, E. D. Doughty, B. Hall, and B. C. S. Hansen. 2007a. Post-glacial changes in spatial patterns of vegetation across southern New England. Journal of Biogeography 34: 900–913. Oswald, W. W., and D. R. Foster. 2011a. A record of late-Holocene environmental change from southern New England, USA. Quaternary Research 76: 314–318. Oswald, W. W., and D. R. Foster. 2011b. Middle-Holocene dynamics of Tsuga canadensis (eastern hemlock) in northern New England, USA. The Holocene 22: 71–78. Oswald, W. W., D. R. Foster, E. D. Doughty, and E. K. Faison. 2009. A record of Lateglacial and early Holocene environmental and ecological change from southwestern Connecticut, USA. Journal of Quaternary Science 24: 553–556. Oswald, W. W., D. R. Foster, E. D. Doughty, and D. MacDonald. 2010. A record of Holocene environmental and ecological changes from Wildwood Lake, Long Island, New York. Journal of Quaternary Science 25: 967–974. Oswald, W. W., B. C. S. Hansen, and D. R. Foster. 2007b. Comparison of pollen and stomata in late-glacial and early-Holocene lake sediments from eastern Massachusetts. Rhodora 109: 225–229. Overpeck, J. T., R. S. Webb, and T. Webb III. 1992. Mapping eastern North American vegetation change of the past 18 ka: no-analogs and the future. Geology 20: 1071–1074. Pacala, S. W., C. D. Canham, and J. A. Silander. 1993. Forest models defined by field measurements. I. The design of a northeastern forest simulator. Canadian Journal of Forest Research 23: 1980–1988. Paillet, F. L. 1982. The ecological significance of American chestnut in the Holocene forests of Connecticut. Bulletin of the Torrey Botanical Club no. 109: 457–473. ———. 1984. Growth-form and ecology of American chestnut sprout clones in northeastern Massachusetts. Bulletin of the Torrey Botanical Club no. 111: 316– 328. ———. 1988. Character and distribution of American chestnut sprouts in southern New England woodlands. Bulletin of the Torrey Botanical Club no. 115: 32–44. ———. 1993. Growth form and life histories of American chestnut and Allegheny and Ozark chinquapin at various North American sites. Bulletin of the Torrey Botanical Club no. 120: 257–268.

REFERENCES 285

———. 2002. American chestnut: the history and ecology of a transformed species. Journal of Biogeography 29: 1517–1530. Paillet, F. L., and P. A. Rutter. 1989. Replacement of native oak and hickory tree species by the introduced American chestnut (Castanea dentata) in southwestern Wisconsin. Canadian Journal of Botany 67: 3457–3469. Paradis, A., J. Elkinton, K. Hayhoe, and J. Buonaccorsi. 2008. Role of winter temperature and climate change on the survival and future range expansion of the hemlock woolly adelgid (Adelges tsugae) in eastern North America. Mitigation and Adaptation Strategies for Global Change 13: 541–554. Parker, B. L, M. Skinner, S. Gouli, T. Ahikaga, and H. B. Teillon. 1998. Survival of hemlock woolly adelgid (Homoptera: Adelgidae) at low temperatures. Forest Science 44: 414–420. ———. 1999. Low lethal temperature for hemlock woolly adelgid (Homoptera: Adelgidae). Environmental Entomology 28: 1085–1091. Parshall, T., and D. R. Foster. 2002. Fire on the New England landscape: regional and temporal variation, cultural and environmental controls. Journal of Biogeography 29: 1305–1317. Parshall, T., D. R. Foster, E. K. Faison, D. MacDonald, and B. C. S. Hansen. 2003. Long-term history of vegetation and fire in pitch pine–oak forests on Cape Cod, Massachusetts. Ecology 84: 736–738. Peterken, G. F. 1996. Natural Woodland: Ecology and Conservation in Northern Temperate Regions. Cambridge University Press, Cambridge. Pontius, J., R. Hallett, and M. Martin. 2005. Using AVIRIS to assess hemlock abundance and early decline in the Catskills, New York. Remote Sensing of Environment 97: 163–173. Popper, K. R. 1935. Logik der Forschung: Zur Erkenntnistheorie der Modernen Naturwissenschaft. Springer, Vienna. Potter, K. M., W. S. Dvorak, B. S. Crane, V. D. Hipkins, R. M. Jetton, W. A. Whittier, and R. Rhea. 2008. Allozyme variation and recent evolutionary history of eastern hemlock (Tsuga canadensis) in the southeastern United States. New Forests 35: 131–145. Potter, K. M., R. M. Jetton, W. S. Dvorak, V. D. Hipkins, R. Rhea, and W. A. Whittier. 2012. Widespread inbreeding and unexpected geographic patterns of genetic variation in eastern hemlock (Tsuga canadensis), an imperiled North American conifer. Conservation Genetics 13: 475–498. Preisser, E. L., A. Lodge, D. A. Orwig, and J. S. Elkinton. 2008. Range expansion and population dynamics of co-occurring invasive herbivores. Journal of Biological Invasions 10: 201–213.

REFERENCES 286

Preisser, E. L., M. R. Miller-Pierce, J. Vansant, and D. A. Orwig. 2011. Eastern hemlock (Tsuga canadensis) regeneration in the presence of hemlock woolly adelgid (Adelges tsugae) and elongate hemlock scale (Fiorinia externa). Canadian Journal of Forest Research 41: 2433–2439. Prentice, I. C. 1985. Pollen representation, source area, and basin size: toward a unified theory of pollen analysis. Quaternary Research 23: 76–86. Preston, F. W. 1948. The commonness and rarity of species. Ecology 29: 254–283. Radville, L., A. Chaves, and E. L. Preisser. 2011. Variation in plant defense against invasive herbivores: evidence for a hypersensitive response in eastern hemlocks (Tsuga canadensis). Journal of Chemical Ecology 37: 592–597. Raup, H. M. 1966. The view from John Sanderson’s farm: a perspective for the use of the land. Forest History 10: 2–11. Raup, H. M., and R. E. Carlson. 1941. The history of land use in the Harvard Forest. Harvard Forest Bulletin no. 20: 4–62. Petersham, Massachusetts. Raymer, P. C. L., D. A. Orwig, and A. C. Finzi. 2013. Hemlock loss due to the hemlock woolly adelgid does not affect ecosystem C storage but alters its distribution. Ecosphere 4: art63. Reardon, R. C., B. Onken, C. A. S.-J. Cheah, S. Salom, B. L. Parker, S. D. Costa, and M. Skinner. 2004. Biological control of hemlock woolly adelgid. U.S. Forest Service, Forest Health Technology Enterprise Team FHTET-2004-04, Burlington, Vermont. Rodbell, D. T. 2000. The Younger Dryas: cold, cold everywhere? Science 290: 285– 286. Rogers, R. S. 1978. Forests dominated by hemlock (Tsuga canadensis): distribution as related to site and postsettlement history. Canadian Journal of Forest Research 56: 843–854. ———. 1980. Hemlock stands from Wisconsin to Nova Scotia: transitions in understory composition along a floristic gradient. Ecology 61: 178–193. Rohr, J. R., C. G. Mahan, and K. C. Kim. 2009. Response of arthropod biodiversity to foundation species declines: the case of the eastern hemlock. Forest Ecology and Management 258: 1503–1510. Ross, R. M., R. M. Bennett, C. D. Snyder, J. A. Young, D. R. Smith, and D. P. Lamarie. 2003. Influence of eastern hemlock on fish community structure and function in headwater streams of the Delaware River basin. Ecology of Freshwater Fish 12: 60–65. Royle, D. D., and R. G. Lathrop. 1997. Monitoring hemlock forest health in New Jersey using Landsat TM data and change detection techniques. Forest Science 43: 327–335.

REFERENCES 287

Sackett, T. S., S. Record, S. Bewick, B. Baiser, N. J. Sanders, and A. M. Ellison. 2011. Response of macroarthropod assemblages to the loss of hemlock (Tsuga canadensis), a foundation species. Ecosphere 2: art74. Sasaji, R., and M. S. McClure. 1997. Description and distribution of Pseudoscymnus tsugae sp. nov. (Coleoptera: Coccinellidae), an important predator of hemlock woolly adelgid in Japan. Annals of the Entomological Society of America 90: 563– 568. Schindler, D. W. 2009. Lakes as sentinels and integrators for the effects of climate change on watersheds, airsheds and landscapes. Limnology and Oceanography 54: 2349–2358. Schoener, T. W. 1983. Field experiments on interspecific competition. American Naturalist 122: 240–285. Schoonmaker, P. K. 1991. Long-term vegetation dynamics in southwestern New Hampshire. Ph.D. thesis, Harvard University, Cambridge, Massachusetts. Schoonmaker, P. K., and D. R. Foster. 1991. Some implications of paleoecology for contemporary ecology. Botanical Review 57: 204–245. Sernander, R. 1908. On the evidence of post-glacial changes of climate furnished by the peat mosses of northern Europe. Geologiska Föreningens i Stockholm Förhandlingar 31: 423–448. Shuman, B., J. Bravo, J. Kaye, J. A. Lynch, P. Newby, and T. Webb III. 2001. Late Quaternary water-level variations and vegetation history at Crooked Pond, southeastern Massachusetts. Quaternary Research 56: 401–410. Shuman, B. N., P. Newby, and J. P. Donnelly. 2009. Abrupt climate change as an important agent of ecological change in the Northeast U.S. throughout the past 15,000 years. Quaternary Science Reviews 28: 1693–1709. Skinner, M., B. L. Parker, S. Gouli, and T. Ashikaga. 2003. Regional responses of hemlock woolly adelgid (Homoptera: Adelgidae) to low temperatures. Environmental Entomology 32: 523–528. Smith, D. M. 1986. The Practice of Silviculture, 8th ed. John Wiley & Sons, New York. Smith, H. S., and D. A. Orwig. 2008. Influence of hemlock woolly adelgid and elongate hemlock scale on leaf-level physiological performance of eastern hemlock. Pages 308–316 in B. Onken and R. Reardon (eds.), Proceedings of the Fourth Symposium on Hemlock Woolly Adelgid in the Eastern United States, U.S. Forest Service, Morgantown, West Virginia. Snyder, C. D., J. A. Young, D. P. Lemarie, and D. R. Smith. 2002. Influence of eastern hemlock forests on aquatic invertebrate assemblages in headwater streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 262–275.

REFERENCES 288

Spurr, S. H. 1950. Stand composition in the Harvard Forest as influenced by site and forest management. Ph.D. thesis, Yale University, New Haven, Connecticut. ———. 1956a. Forest associations in the Harvard Forest. Ecological Monographs 26: 245–262. ———. 1956b. Natural restocking of forests following the 1938 hurricane in central New England. Ecology 37: 443–451. Spurr, S. H., and A. C. Cline. 1942. Ecological forestry in central New England. Journal of Forestry 40: 418–420. Stadler, B., T. Müller, and D. A. Orwig. 2006. The ecology of energy and nutrient fluxes in hemlock forests invaded by the hemlock woolly adelgid. Ecology 87: 1792–1804. Stadler, B., T. Müller, D. A. Orwig, and R. C. Cobb. 2005. Hemlock wooly adelgid in New England forests: canopy impacts transforming ecosystem processes and landscapes. Ecosystems 8: 233–247. St. Jacques, J.-M., M. S. V. Douglas, and J. H. McAndrews. 2000. Mid-Holocene hemlock decline and diatom communities in van Nostrand Lake, Ontario, Canada. Journal of Paleolimnology 23: 385–397. Stephens, E. P. 1955. Research in the biological aspects of forest production. Journal of Forestry 53: 183–186. ———. 1956. The uprooting of trees: a forest process. Soil Science Society of America Proceedings 20: 113–116. Sugihara, G. 1980. Minimal community structure: an explanation of speciesabundance patterns. American Naturalist 116: 770–787. Sugita, S. 1993. A model of pollen source area for an entire lake surface. Quaternary Research 39: 239–244. ———. 1994. Pollen representation of vegetation in Quaternary sediments: theory and method in patchy vegetation. Journal of Ecology 82: 881–897. Sullivan, K. A., and A. M. Ellison. 2006. The seed bank of hemlock forests: implications for forest regeneration following hemlock decline. Journal of the Torrey Botanical Society 133: 393–402. Templer, P. H., and T. M. McCann. 2010. Effects of the hemlock woolly adelgid on nitrogen losses from urban and rural northern forest ecosystems. Ecosystems 13: 1215–1226. Thompson, J. R., D. N. Carpenter, C. V. Cogbill, and D. R. Foster. 2013. Four centuries of change in northeastern U.S. forests. PLOS ONE [end of August 2013; details TK] Thompson, J. R., A. Wiek, F. J. Swanson, S. R. Carpenter, N. Fresco, T. Hollingsworth, T. A. Spies, and D. R. Foster. 2012. Scenario studies as a synthetic and

REFERENCES 289

integrative research activity for long-term ecological research. BioScience 62: 367–376. Tingley, M. W., D. A. Orwig, R. Field, and G. Motzkin. 2002. Avian response to removal of a forest dominant: consequences of hemlock woolly adelgid infestations. Journal of Biogeography 29: 1505–1516. Trotter, T. R., and K. S. Shields. 2009. Variation in winter survival of the invasive hemlock woolly adelgid (Hemiptera: Adelgidae) across the Eastern United States. Environmental Entomology 38: 577–587. Turner, J. L., M. C. Fitzpatrick, and E. L. Preisser. 2011. Simulating the dispersal of hemlock woolly adelgid in the temperate forest understory. Entomologia Experimentalis et Applicata 141: 216–223. Ulrich, W., M. Ollik, and K. I. Ugland. 2010. A meta-analysis of species–abundance distributions. Oikos 119: 1149–1155. Vermaat, J. E., J. A. Dunne, and A. J. Gilbert. 2009. Major dimensions in food-web structure properties. Ecology 90: 278–282. von Post, L. 1917. Om skogsträdpollen i sydsvenska torfmosselagerföljder. Geologiska Föreningens i Stockholm Förhandlingar 38: 384–390. Weiskittel, A. R., D. W. Hann, J. A. Kershaw, and J. K. Vanclay. 2011. Forest Growth and Yield Modeling. John Wiley & Sons, New York. Willacker, J. J., W. V. Sobczak, and E. A. Colburn. 2009. Stream macroinvertebrate communities in paired hemlock and deciduous watersheds. Northeastern Naturalist 16: 101–112. Williams, J. W., B. N. Shuman, and T. Webb III. 2001. Dissimilarity analyses of late-Quaternary vegetation and climate in eastern North America. Ecology 82: 3346–3362. Wilson, B. F. 1962. A survey of the incidence of ring shake in eastern hemlock. Harvard Forest Paper No 5: 1–11. Petersham, Massachusetts. Wright, H. E., Jr. 1967. A square-rod piston sampler for lake sediments. Journal of Sedimentary Petrology 37: 975–976. Yamasaki, M., R. M. DeGraaf, and J. W. Lanier. 2000. Wildlife habitat associations in eastern hemlock—birds, smaller mammals, and forest carnivores. Pages 135–143 in K. A. McManus, K. S. Shields, and D. R. Souto (eds.), Proceedings of the Symposium on Sustainable Management of Hemlock Ecosystems in Eastern North America, U.S. Forest Service, General Technical Report NE-267, Newtown Square, Pennsylvania. Yorks, T. E., D. J. Leopold, and D. J. Raynal. 2000. Vascular plant propagule banks of six eastern hemlock stands in the Catskill Mountains of New York. Journal of the Torrey Botanical Society 127: 87–93.

REFERENCES 290

Young, R. F., K. S. Shields, and G. P. Berlyn. 1995. Hemlock woolly adelgid (Homoptera: Adelgidae): stylet bundle insertion and feeding sites. Annals of the Entomological Society 88: 827–835. Ziegler, S. S. 2002. Disturbance regimes of hemlock-dominated old-growth forests in northern New York, U.S.A. Canadian Journal of Forest Research 32: 2106–2115.

REFERENCES 291

This page intentionally left blank

C O N T R I B U TO R S

BENJAMIN H. B AI S ER , a recent postdoctoral researcher at the Harvard Forest, is assistant professor of ecology and conservation at the University of Florida, where he studies how human activity influences the structure and dynamics of ecosystems and how this translates to changes in large-scale biodiversity patterns. He develops computational models to explore the structure and dynamics of ecological networks and has worked with bird, plant, invertebrate, and protozoan communities in diverse field locations, including the Florida Everglades, Hawk Mountain Sanctuary, and bogs across Vermont and Massachusetts. AUDREY A. BARKER PLOTKIN coordinates the Long-Term Ecological Research program and site use at the Harvard Forest. Her research delves into long-term forest development, especially how forests respond to disturbances such as wind and insect outbreaks. She is involved in land management as a Massachusetts licensed forester and part-time farmer. ANTHONY W. D’AMATO, a former graduate student at the Harvard Forest, is associate professor of silviculture and applied forest ecology at the University of Minnesota. His research examines the impacts of natural and human-generated disturbances on the growth and development of forest ecosystems. In the spirit of Bob Marshall and the generations of Harvard Forest scientists who followed, this work relies heavily on the records of change and dynamics contained within tree rings and long-term field observations. AARON M. EL L I SON is a senior research fellow in ecology at the Harvard Forest, adjunct research professor in biology and environmental conservation at the University of Massachusetts, and lead author of A Field Guide to the Ants of New England. He has studied how various ecological systems—including salt marshes and mangrove swamps, pitcher-plant symbiota and freshwater sponge endobiota, forest-floor

293

ant assemblages, tropical rainforests, and temperate hemlock forests—disintegrate following disturbances and then reassemble through time. DAVID R . FOST ER is director of the Harvard Forest, senior lecturer on biology at Harvard University, adjunct research professor in the Department of Environmental Conservation at the University of Massachusetts, and author of Thoreau’s Country: Journey through a Transformed Landscape. His research combines paleoecology, environmental history, and ecology to explore the past and current condition of landscapes and to consider how best to conserve them for the benefit of nature and society. He collaborates with the Highstead Foundation and other partners to advance the Wildlands and Woodlands vision of conserving more than 70 percent of the New England landscape in forests and farms. D AV I D A . O RW I G is the forest ecologist at the Harvard Forest and adjunct research professor at the University of Massachusetts. His research encompasses broad areas of ecosystem response to natural and human disturbance and environmental change. Using a combination of community ecology and tree-ring studies, he has spent the last two decades studying the ecology of old-growth forests and helping to conserve these ecosystems across the eastern United States. His work focuses especially on the complex interactions between invasive pests and pathogens and dominant host tree species, and the implications of these dynamics for ecological theory and forest practice. W. W YAT T O S WA L D is an associate professor of science at Emerson College, Boston, research fellow at the Harvard Forest, and associate editor for the journal Quaternary Research. His research analyzes lake-sediment records to reconstruct past changes in climate, vegetation, and fire. He has worked in New England, Alaska, and the western United States. At Emerson he teaches courses on climate change, earth science, and ecology, and in 2010 he received the school’s Helaine and Stanley Miller Award for Outstanding Teaching. JONAT HAN R . T HOMPSON is a senior ecologist at the Harvard Forest. Relying heavily on ecosystem and landscape simulation models, he studies the aggregate consequences of land use, species invasion, conservation, and climate change on the composition, configuration, and carbon stores of forest ecosystems in the United States. STEPHEN J. LONG is a freelance writer and editor, the founder of Northern Woodlands magazine, and the author of More Than a Woodlot.

CONTRIBUTORS 294

INDEX

Page numbers in bold refer to photographs and illustrations. 51; postglacial migration, 57, 58; precolonial decline, 62; shade tolerance, 13–15, 19 Bennett, Keith, 232 Berkshire old-growth groves, xiii, 65–67, 66, 70 Bhiry, Najat, 109, 250 Binzen, Tim, 222, 267 biocontrol (predatory insects), 6, 202, 227, 254 biodiversity: abundance patterns in ecosystems, 247; after Cut or Girdle experiment, 150; genetic diversity, 262–63; harvesting trees to maintain, 191–93; increased after hemlock loss, 227–29 (see also forest composition); less in hemlock forests, 167–68, 262; maximized by intermediate-intensity disturbances, 70 birch trees: black birch, xvi, 17, 19, 116, 129, 131, 134; and fauna, 134, 150; and hemlock decline, xvi, 109, 116, 129, 149, 151–52, 258; hemlock more vulnerable than, 2; leaf litter, 13, 131; lower branches shed, 11; pollen, 54; postglacial migration, 55; yellow birch, 95 birds: adelgid transported, 124–26, 161; in hemlock forests, 19–20, 100–101, 235; hemlock loss and, xvi, 134, 259. See also specific species black birch: growth rate, 17; and hemlock decline, xvi, 116, 129; leaf litter, 131; preferred by some birds, 134; seeds in hemlock soils, 19. See also birch trees

Aber, John, 259–60 Abrams, M. D., 233, 234 Adams Fay lot: 1924 experiment, 73, 79, 80, 84–90, 85, 243; after the 1938 hurricane, 91–92; purchased, 92, 246. See also Marshall, Robert adelgid. See woolly adelgid agriculture. See farmland Alaska, 116, 251 Albani, Marco, 156–58 Allison, Taber, 108 American beech. See beech trees (American beech) “Another Look from Sanderson’s Farm” (Donahue), 205, 266 ants, 100, 134, 150, 259 Arana, Yvan Delgado de la Flor, 167 Ashuelot region, New Hampshire, 28–30 Asian long-horned beetles, 121, 162, 183 bark (hemlock): eaten by porcupines, 20; logging for, 68; respiration sampling, 130; uses, x, xi, 23–24, 218–22 (see also tanning and tanneries) Barnes, Burton, 238 BBD. See disease: birch bark disease (BBD) beaver, 97, 98, 169 beech trees (American beech): abundance, 61; beech bark disease, 121, 122, 183; beech forest, 103; compared to hemlock, 1, 2, 102; ecosystem role, 102; and hemlock decline, 109, 119; and oak collapse, 116–17; pollen,

295

stewardship of Harvard Forest, 36–38, 181, 185, 188–89 Cogbill, Charlie, 61, 216, 240 Cole, Thomas, xxi Colebrook Forest, 25, 27, 65, 69–70, 74, 240 collapse of hemlock (5,500 years ago): about, 1, 58–59, 105; biological agent hypothesis, 58–59, 107–9, 110, 112–13, 118, 250; climate change as driver, 58–59, 106, 110–13, 118, 223, 250–51; first identified, 106; forest compositional changes, 116, 119, 165–66; and forest dynamics, 109–10; and hemlock genetic variation, 262; humans benefited, 119, 229; pollen and lake sediment records, 240, 250 (see also lake and pond sediments); recovery, 213; simultaneous oak collapse, 111–12, 251 cones (hemlock), 2, 12, 55 Connecticut, 22, 23, 61, 129, 132. See also Colebrook Forest; New England conservation: agenda, 9; challenges today, 183–86; of foundation species, 247; hemlock loss as lens for, 9–10; managing for old-growth conditions, 241; natural areas and their champions, 25–26; Pisgah Forest, 33–34, 36–38, 65, 236–37; Wildlands and Woodlands vision, 186–88. See also forest management construction, hemlock’s uses in, xi, 22–23 core species, 95, 248 Costa, Scott, 132, 252–53 Cut or Girdle experiment, 137–52, 139, 255–59 cutting trees. See logging

Black Gum Swamp: about, 205–6, 220; sediment study, 106, 109–10, 212, 250; weir, 196 Bob Marshall Plot, 87. See also Marshall, Robert Booth, Bob, 118 Boston, xiii–xiv, 120, 162, 206–7 Botkin, Daniel, 154, 163 Branch, W. C., 33, 35, 38, 41 brook trout, 20, 134, 235 Brown, Joe, 132, 252 Burroughs, John, 64 Cape Cod, 54, 95, 111–12, 117 carbon storage and cycling, 17, 131, 149, 156– 59, 166–67, 259 Carrière, Élie-Abel, x Catovsky, Sebastian, 234 chestnut (American chestnut): dead trunks’ persistence, 129; fungal blight, 7, 60, 63, 70, 108, 120, 183, 224, 225, 251; leaves, 8; mortality compared to hemlock collapse, 107–8, 109; once common, 61; Paillet’s work, 232; persistence (resilience), 6–9; pollen, 51, 54; postglacial migration, 59, 62; relationship with hemlock, 59, 70; uses, 220 Chilton, Elizabeth, 119, 220, 267 climate. See climate change; drought; temperature climate change: and forest composition, 113, 183–84; global changes, 110–11, 250; importance of dispersal studies to, 55–57; interaction with other stresses, 183; lake sediments used to study, 240 (see also lake and pond sediments); postglacial warming, 54–55, 105; and prehistoric hemlock collapse, 58–59, 106, 110–13, 118, 233, 250–51; and prehistoric oak collapse, 117, 233; species’ ranges expanded, 159; temperature drop modeled, 163; warming experiments, 255; western North American impacts, 116, 251 climax forest concept, 68–69 Cline, Al, 182; and the Adams Fay lot experiment, 86, 90, 91–92; on the history of the Pisgah forest studies, 30–32; and Marshall, 77, 83, 86, 91, 242, 243–44, 245; 1942 paper with Spurr, 34, 38–39, 241; on the role of disturbance, 69; on the Simes tract, 257;

Dale, Virginia, 259–60 Daley, Robert, 33, 38, 41 D’Amato, Tony, 66, 233, 234, 241, 242, 266 Davidson, Eric, 142 Davis, Arthur, 83 Davis, Margaret: biological agent hypothesis, 106–9, 110, 112, 118–19, 232, 250; tree migration work, 163, 239 Davis, Mary Byrd, 240 Day, Lindsay, 117 Dayton, Paul, 94, 247–48 death of hemlock trees: from adelgid infestation, 5, 6, 9, 124, 125, 128–29, 143, 197, 253; beaver-killed trees, 98, 169; dead trunks

INDEX 296

183, 224, 225, 251; Dutch elm disease, 121, 123; increasing role in tree mortality, 112; and prehistoric hemlock collapse, 107, 108, 112–13, 250 (see also insects) dissection of a forest, 174–80. See also Earl Stephens plot project distribution of hemlock: as example of ecological legacy, 2–4; Fisher on, 33; mapping, 20–22, 23, 126, 235, 253; postcolonial period, 61–62; prehistoric distribution, 55–57, 58–61; range, x, 7, 20–22, 55–57, 58–59, 240 diversity of species. See biodiversity dominant species, 95, 248 Donahue, Brian, 205, 266 Doucette, Dianna, 119, 219, 220, 222, 267 Doughty, Elaine, 117 drought: drought shadow, beneath hemlocks, 13; and hemlock decline, 58–59, 118, 119 (see also collapse of hemlock); and hemlock photosynthesis, 17; mature trees resistant, 183; 1960s drought, 112, 251; and oak declines, 113, 114, 117; registered in lake sediments, 45, 52, 112, 240 Dunwiddie, Peter, 67 “Dust of Snow” (Frost), xviii–xix Dutch elm disease, 121, 123, 183, 251 dynamic global vegetation models, 163–64

and logs, 5, 42, 74–76, 125, 192, 199; and fire risk, 198; and forest composition, 129– 31, 147–49, 150, 166, 195, 197, 258–59; from girdling, 138, 143, 144, 145, 149–50; impact on ecosystem, 129–31, 134, 149–51, 258–59; inability to sprout, 9; mortality rates, 129, 135; process, 128–29; rapid decay, 129; scale infestations and, 135, 254–55; sound of dying trees, ix–x. See also collapse of hemlock; decline of hemlock; logging; woolly adelgid decline of hemlock: chestnut and, 59; after colonization, xxi, 1–2 (see also farmland; logging); conservation implications, 9–10 (see also conservation; forest management); Cut or Girdle experiment, 137–52, 139, 255–59; demand for tannins, 23–24 (see also tanning and tanneries); ecosystem changes from, xvi, 101, 109, 156–59, 227–29, 254; emotional impact, xvii, xxii–xxiii, 142, 227; and forest composition, xvi, 129–31, 147–51, 166, 197, 227; landscape diversity lost, 168–70, 229; management approaches (see forest management); mortality rates, 129, 135, 197; in the past 400 years, 61–62; pre-colonial distribution and decline, 61–62; preemptive harvesting, xvi, 132–33, 138, 142, 143–47, 146, 197–98, 227, 253–54; scale infestations and, 135, 254–55; and species diversity, 150; wildlife habitat affected, 134. See also collapse of hemlock; death of hemlock trees; logging; woolly adelgid decomposition: of beech litter, 102; carbon released, 149, 156, 158; function, 100; of hemlock litter, 13, 234; of hemlock wood, 129, 198; nutrient cycling, 100, 131, 149; old-growth pine resistant to, 43 Deep Pond coring, 44–48, 46. See also lake and pond sediments deer: adelgid transported by, 124; hemlock death and, xvi, 134, 150; hemlock useful to in winter, 20; in the 1920s, 243, 244–45 Deevey, Edward, 106, 111, 239, 250 Degrassi, Ally, 142, 167 Del Tredici, Peter, 234–35, 265 Dickinson, Emily, 104, xix Dickinson brothers, 33, 84 disease: beech bark disease (BBD), 121, 122; chestnut blight, 7, 60, 63, 70, 108, 120,

Earl Stephens plot project, 172–80, 263–64. See also Stephens, Earl P. eastern forests: “better” forest concept, 167– 68; carbon storage, 156, 166–67 (see also carbon storage and cycling); cleared for agriculture, 1–4, 62 (see also farmland); dominant, core, satellite, and keystone species, 95–97; fluctuations between softand hardwood dominance, 33; growth and development patterns, 70; hurricane damage, salvage, and recovery, 36–38, 37, 40, 41–43, 42, 91–92, 181, 182, 184–85, 188–89, 195, 238, 257; landscape diversity, 168–70; local abundance of hemlock, 158; modern threats, 183 (see also climate change; logging; pests; and specific pests); mortality and recovery as natural process, 193–97; planting trees, 201; postcolonial forest composition, 61–62; prehistoric forestation, 54–55,

INDEX 297

Elkinton, Joe, 110 Ellison, Aaron, 141, 142, 167 elm trees, 109, 121, 123, 183, 251 elongate hemlock scale, 135, 254–55

eastern forests (continued) 57, 58–61 (see also collapse of hemlock); primary vs. secondary forests defined, 72; recovery and resiliency, 2–4, 147–49, 186, 194–97 (see also resilience of hemlock); residential areas within, 187; 17th- and 18thcentury records, 61, 215–16. See also conservation; forest composition; Harvard Forest; hemlock forests; New England; old-growth forests; Pisgah Forest eastern hemlock. See collapse of hemlock; death of hemlock trees; decline of hemlock; distribution of hemlock; hemlock forests; hemlock trees eastern hemlock looper, 108–9, 250 Eastern Old-Growth Forests (Davis), 240 “Ecological Forestry in Central New England” (Cline and Spurr), 38–39 ecological models: about, 154, 259–60; adelgid dispersal model, 160–62, 170, 252–53; dynamic global vegetation models, 163–64; Ecosystem Demography model (ED), 156–58, 260–61; of foundation species loss, 249; Future Scenarios project, 162–63; gap models, 154–56, 260; hemlock loss and carbon storage, 156–59, 166–67, 259; JABOWA forest model, 154–56, 163, 260; models distrusted, 153–54; PnET models, 260 Ecology (journal): Davis-Allison study, 108; Henry and Swan paper, 39–41, 87 (see also Henry, David; Swan, Mark); OliverStephens paper, 87, 180, 242 (see also Stephens, Earl P.); Spurr paper, 39 (see also Spurr, Steve) Ecosystem Demography model (ED), 156–58, 260–61 ecosystem engineers, 97, 98, 248 ecosystems: dominant, core, satellite, and keystone species, 95–97, 168; impact of foundation species decline, 101–2, 249; impact of hemlock loss, xvi, 101, 109, 129–31, 149–51, 156–59, 227–29, 254, 258–59; interspecies relationships within, 100–101, 248–49; mortality and recovery as natural process, 193–97. See also ecological models; foundation species efts, 19 Egler, Frank, 65

Faison, Ed, 142 farmland: cleared for agriculture, 1–2, 62, 64–65; forest recovery in former farmland, 2–4, 186; second-growth forests affected by, 75 Farnsworth, Elizabeth, 142 Filion, Louise, 109, 250 Finzi, Adrien, 142 fire: in Alaska and the West, 116; concern about, 36–38, 198; hemlock decline following, 59; during/after the oak decline, 117; during/after the prehistoric hemlock collapse, 110; rare in New England, 61 firewood, 143, 147 fish, 20, 134, 235 Fisher, Richard T., 31; Ashuelot region’s forests explored, 28–30; background, 28, 236; longterm logging experiment, 84–86, 85, 89–90 (see also Adams Fay lot: 1924 experiment); and Marshall, 78–81, 79, 82, 83–86, 89–90, 91, 243–44, 245, 246–47; memorial plaque, 237, 245; Mount Wachusett survey, 67; and the Pisgah Forest, 26, 29–34, 65, 236–37; Raup and, 173; on the role of disturbance, 69; silvicultural approach, 30–32, 41, 178, 189; writings collected, 214 Fisher House, 99, 207, 236 Fisher Museum, xv, 32, 231, 237. See also Harvard Forest dioramas fishers, 20, 134 Fitzpatrick, Matthew, 160–62, 170 flux towers, 260–61 food webs, 100–101, 248, 259 forest composition: climate change and, 113, 183–84; Cut or Girdle experiment, 140; hemlock loss and, xvi, 6, 129–31, 147–51, 166, 195, 197, 227, 258–59; historically, 58–62; after logging, 62, 133 (see also logging: overall impact); New England generally, 41–43; postcolonial period, 61–62 Forest Ecology (Spurr), 238 forest ecosystems. See also hemlock forests forest history. See historical ecology

INDEX 298

fungi, 129. See also disease: chestnut blight Future Scenarios project, 162–63

Forest History, Raup’s article in, 204–5 forest management: action vs. inaction, 185, 190–93, 197–98, 199–201 (see also Pisgah Forest: 1938 hurricane and recovery); and adelgid spread, 132–33, 138, 142–47, 159, 189–91, 193, 199–201, 201–3, 265–66; challenge of, 181–86; Cline’s approach, 36–39; Fisher’s approach, 30–32, 41, 178, 189; legacy, 198–99; mortality as natural process, 193–97; Norway spruce as hemlock substitute, 265; for old-growth conditions, 241; planting trees, 201; post-hurricane timber salvage, 36–38, 37, 91–92, 188–89, 195; protection forest approach, 191–93. See also conservation forest reconstruction: Marshall and, 86–90, 92 (see also Marshall, Robert); Stephens and, 172–80, 263–64. See also historical ecology forests. See eastern forests; forest composition; hemlock forests; old-growth forests; and specific forests Forests in Time (Foster and Aber), 232 Foster, David: on the adelgid and hemlock loss, xvi, xvii; and the Future Scenarios project, 162; hemlock forest history study, with McLachlan, 62–63; and the Hemlock Hollow sediment study, 210–12, 240; oldgrowth vs. second-growth understory documented, 74; pollen analysis studies, 109–10; and Sanderson’s woodlot, 209–10, 267; stand-replacing disturbance event synthesis, 71; and Stephens, 180; Thoreau’s Country, xiv foundation species: Dayton’s work, 94, 247– 48; definition and importance, 4, 93–94, 97, 101, 247, 248, 249, 262; difficult to identify, 101, 261–62; and ecosystem interactions, 100–101, 248–49; environmental impact of decline, 101–2, 249; hemlock as, 4, 93–95, 97–104, 150, 165–71, 262–63; humans in relation to, 97–99; key characteristics, 94–95, 247; vs. keystone species, 168, 248; landscape diversity enhanced, 168–70; smaller species as, 99–100 Fraver, Shawn, 76 Frost, Robert, xvii–xix, 136 Fuller, Janice, 109, 232 Fuller, Margaret, xiii funding for research, 121, 128, 137, 208–9, 267

gap-phase replacement, 70 Gast, Rupe: and Marshall, 83, 86, 243, 246; and the 1924 Adams Fay lot experiment, 79, 84, 86, 90 genetic diversity, 262–63 girdling, 138, 139, 143, 144, 145, 149–50. See also Cut or Girdle experiment glacial period (most recent), 54–55 Gleason, Henry, 57–58 Goodlett, John, 264, 267 Gould, Ernie, 180, 205, 267 Goulet, Fred and Otis, 83 greenhouse effect, 156. See also climate change “The Growth of Hemlock before and after the Release from Suppression” (Marshall), 87–88 growth rate of hemlock, 2, 15–17, 70, 88–89, 234 gypsy moth, 113, 120–21, 183, 193, 202 Hadley, Julian, 234 Hampson, Linda, 217 Hanifin, Thomas, 189 hardwoods, 15, 17, 33, 197. See also specific trees Harmon, Mark, 247 Hart, Clarisse, xvii Harvard Forest: about and history, xv, xxiii– xxiv, 26; adelgid’s arrival, 151, 259; archives, 206, 213–16, 242–46; atmosphere, 224; Cline’s stewardship, 36–38, 181; endowment, 79; first map, 207; maintenance garage, 147, 148; management approach, 30–32, 36–39, 41, 178, 188, 201–3, 265–66 (see also forest management); 1938 hurricane and recovery, 40, 42, 182, 184–85, 188–89, 195, 257 (see also hurricanes); tracts and plots, 26, 27, 34, 41, 85, 140 (see also Adams Fay lot; Prospect Hill tract; Sanderson farm; Simes tract); Wildlands and Woodlands vision, xv, 186–88. See also Fisher House; Fisher Museum; Harvard Forest studies; Hemlock Hollow; Pisgah Forest Harvard Forest Bulletin: Branch-Daley-Lotti study, 38; Marshall’s project, 87–88; Raup’s introduction (1956), 246

INDEX 299

Hemingway, Ernest, 1 hemlock borer, 129 hemlock forests: atmosphere, ii, 11, 13, 94–95, 224, 229, 233; Berkshire old-growth groves, xiii, 65–67, 240–41; compared to salt marshes, xii–xiii; dead trees’ environmental impact, 74–76; deep shade, ii, 11, 13, 234; dissection and chronology of a forest, 174–80 (see also Earl Stephens plot project); disturbance’s role in shaping, 68–72, 88–89, 241; diversity of sites and environments, 20–22; eulogy for, ix–xix; fauna, 19–20, 97, 100–101, 134, 235, 262; food webs, 100–101, 259; forest floor, 13, 14, 19, 49, 234; hemlock looper and, 108–9, 250; Marshall on typical history, 88–89 (see also historical ecology); McLachlan-Foster study, 62–63; percentage of hemlock, 256, 257–58; Pisgah old-growth forest, 69; rain and snowfall, and drought shadow, 13; scent, 13; second-growth stands, 72–76, 140, 170; soil, 19, 147, 235, 258 (see also soil); temperature, xiii, 11, 13, 147; trails closed in adelgid-infested areas, 143; understory, 3, 11, 16, 166; understory vegetation, 19, 74, 88, 235; unique environment, 97; uproot mounds, 43, 174–75, 177, 179, 264. See also collapse of hemlock; death of hemlock trees; decline of hemlock; distribution of hemlock; Harvard Forest; hemlock trees; old-growth forests; Pisgah Forest; resilience of hemlock Hemlock Hill (Arnold Arboretum, Boston), xiii–xiv Hemlock Hollow: 220-plus-year-old tree, 50; pond sediment study, 53, 59, 106, 109–10, 117, 210–13, 211, 233, 240, 250 hemlock looper, 108–9, 250 Hemlock Removal Experiment. See Cut or Girdle experiment hemlock trees (eastern hemlock): bark, x, xi, 20, 23–24, 68, 130, 218–22; carbon storage, 17, 131, 166–67, 234, 259 (see also carbon storage and cycling); cones, 2, 12, 55; durability of old-growth logs, 76; established over stump, 175; etymology, x; as foundation species, 4, 93–95, 97–104, 150, 165–71, 262–63; genetic variation, 262; growth rate, 2, 15–17, 70, 88–89, 234; layering, 234–35;

Harvard Forest dioramas, xv, 30, 32, 204, 231, 237 Harvard Forest studies: about (generally), 128, 136–37; Adams Fay lot experiment, 73, 79, 80, 84–90, 85, 91–92; adelgid carbon impact model, 156–58; adelgid coldtolerance study, 252; adelgid dispersal model, 160–62, 252–53; adelgid infestation impact studies, xv, 121–24, 126–29, 151–52, 202–3, 209, 224–25, 253; Cape Cod lake-sediment study, 111–12; Cline and Spurr’s work, 34, 38–39, 241 (see also Cline, Al; Spurr, Steve); collaborations, 265 (see also specific studies and individuals); Cut or Girdle experiment, 137–52, 139, 255–59; downed tree core study, 76; and ecological models generally, 260; equipment and tree markings, xv, 206, 224–25; Fuller’s work, 109, 232; funding, 128, 137, 208–9, 267; Future Scenarios project, 162–63; Hemlock Hollow study, 53, 59, 106, 109–10, 117, 210– 13, 211, 233, 250; Henry-Swan study and paper, 39–41, 71, 87; hurricane impact simulation, 195, 255, 264–65; “John Sanderson’s Farm” article, 204–5, 209, 218–20, 266–67 (see also Sanderson farm); Long Term Ecological Research program, 84, 154, 195, 205, 232, 255, 264–65, 267; Marshall’s work, 79, 80, 81–90, 85, 174, 178, 234, 241–47; McLachlan-Foster study, 62–63; Pisgah tree growth records, 214; pollen analysis studies, 109–10; ring shake study, 22; Sanderson woodlot, 206–10, 211; sediment core samples, 44–47, 46, 53, 54, 109–10, 210, 211, 233 (see also lake and pond sediments); seed bank studies, 19; Stephens’ forest dissection/ reconstruction, 172–80, 263–64; tannery excavation, 219, 220–22, 267; water movement measurement, 17; and the writing of this book, xxiv–xxvi. See also Pisgah Forest; woolly adelgid; and specific individuals Harvard Forest Woods Crew: and the Adams Fay lot experiment, 84, 86, 87, 90; and the Cut or Girdle experiment, 146; and Stephens’s research, 173, 174, 178 harvesting trees. See logging Heaney, Seamus, xviii Heinemann, Ross, 117

INDEX 300

individualistic concept of ecology, 57–58 insects: as biocontrol agents, 6, 202, 227, 254; in healthy hemlock forests, 19, 20, 97, 100–101; hemlock death and, 134, 150, 259; hemlock pests, 108–9, 112, 129, 135, 250, 254–55 (see also woolly adelgid); insecticides, 6, 201–2, 227, 265; introduced species, 120–21, 194, 202, 251–52 (see also introduced species); and oak mortality, 113, 114, 115, 120–21, 193 (see also oak trees); pitcher plant and, 99–100; and prehistoric hemlock collapse, 107, 108–9, 110, 250; remains in lake sediments, 52. See also ants; spiders; and other species intermediate-disturbance hypothesis, 70 introduced species (invasive species), 120–21, 159, 194, 202, 251–52. See also Asian longhorned beetles; gypsy moth; woolly adelgid invertebrates, 20, 134, 235. See also insects Ireland, Alex, 92 Irving, Washington, 153

leader (top), 226; old trees (photos), 50, 66; old-growth diameters, 74; as Pennsylvania state tree, 120; pests (see eastern hemlock looper; elongate hemlock scale; hemlock borer; woolly adelgid); ring shake, 22, 73, 233; seedlings (photo), 56; shade tolerance, 11–15, 88–89, 234; sound of dying trees, xix; structure, 11–13, 12; as timber source, xi, 22–23, 233 (see also logging); water use/ loss, 17–19, 166, 234; white pine’s relationship with, 56, 88–89. See also collapse of hemlock; death of hemlock trees; decline of hemlock; distribution of hemlock; hemlock forests; needles (hemlock); resilience of hemlock hemlock varnish shelf fungus, 129 hemlock woolly adelgid. See woolly adelgid Henry, David, 39–41, 71, 87, 238 historical ecology: Earl Stephens’ forest dissection, 174–80; Henry-Swan study, 39–41, 71, 87, 238; “John Sanderson’s Farm” article, 204–5, 209, 218–20, 266–67 (see also Sanderson farm); Marshall’s contributions, 86– 90, 92, 214, 238, 241–42; serendipity’s role, 213–14. See also paleoecology “The History of Land Use at the Harvard Forest” (Raup), 205 hobblebush, 16, 19, 74 Hosier, Paul, 65 Hosley, Neil, 83, 90, 91, 243 housing, forest lost to, 186, 187 Hubbard Brook, 260. See also JABOWA forest model human beings: adelgid transported by, 124; prehistoric hemlock collapse and, 119, 229, 251; in relation to foundation species, 97–99; response to hemlock loss, xvii, xxii– xxiii, 142, 227 (see also forest management) hurricanes: emulation study, 195, 255, 264–65; identifying effects from land features, 176; 1938 hurricane and recovery, 36–39, 37, 40, 41–43, 42, 69, 70–72, 91–92, 181, 182, 184– 85, 188–89, 195, 238, 257; trees uprooted rather than broken, 181, 264 HWA. See woolly adelgid

JABOWA forest model, 154–56, 163, 260 Jackson, Steve, 118, 163–64 James, Harry, 28, 29, 237 Janak, James, 154 “John Sanderson’s Farm” article. See “The View from John Sanderson’s Farm: A Perspective for the Use of the Land” (Raup) Johnson, A. F., 204 Journal of Biogeography (2002 edition), 232 kelp, 94 keystone species, 95–97, 168, 248. See also foundation species Kittredge, Dave, 190, 265 lake and pond sediments: analyzing samples, 52–54, 239–40; charcoal in, 61, 117; coring, 44–48, 46, 53, 117, 210, 211, 239; depth, 48; and dynamic global vegetation models, 163–64; hemlock collapse documented, 108, 109–10, 250; Hemlock Hollow study, 53, 59, 106, 109–10, 117, 210–13, 211, 233, 250; occasional cold, dry centuries indicated, 58; from Pisgah, 238–39; pollen grains in, 51–52 (see also pollen); sand layers, 45, 52, 59, 112

Ice Age (most recent), 54–55 ice storms, 70

INDEX 301

maple trees, 2, 13, 17, 51, 54, 61. See also red maple; sugar maple Marshall, Louis, 77, 82, 242 Marshall, Robert (Bob): about, xiv, 78–82, 83, 242; Adirondack Reserve paper, 77, 78; archive, 242–46; and Fisher, 77–79, 82, 83, 243, 245; on the forest in flux, 181; on forests’ dynamic beauty, xvi; Harvard Forest work, 72–74, 73, 79, 80, 81–90, 85, 174, 178, 234, 241–47; journal, 214; later career, 25, 90, 91, 245–46 Martha’s Vineyard, 95, 113, 114, 115, 117, 251 Massachusetts: Asian long-horned beetles, 162; carbon storage, 156; forests and natural areas, 27; hemlock distribution map, 22, 23; land use future scenarios, 162–63; oldgrowth groves, 65–68, 66, 70, 75; postcolonial forest composition, 61; prehistoric oak collapse, 111–12. See also New England; Pisgah Forest; and specific locations Mathewson, Brooks, 142 Mausel, Dave, 254 McClure, Mark, 251, 254 McLachlan, Jason, 62–63, 234, 241 Melville, Herman, 44 Mew, Ben, 92 migration rate of hemlock, 55–57 models. See ecological models Moeller, Bob, 108 Mohawk Trail State Forest, 75 Moorcroft, Paul, 156–58, 260 moose, 21, 150, 259 Motzkin, Glenn, 103 mountain pine beetle, 116, 251 Mount Monadnock, xiv, 28, 29, 236 Mount Wachusett, 67–68, 234–35 Muir, John, 25

Lambert, Kathy, 162 land records, 61, 215–16. See also witness trees landscape diversity, 168–70 leaf litter, 13, 131, 149, 150. See also needles (hemlock) Leopold, Aldo, xiv, 83 Leverett, Robert, 67 Little Ice Age, 62 Livingstone, Dan, 47 Livingstone, Robert, 65, 239 lodgepole pine, 116, 251 logging: Fisher’s long-term experiments, 84–86, 85, 89–90; forest recovery after, 62, 72–74, 133, 147–49, 186; hemlock benefitted, 63, 89–90; hemlock disproportionately affected, 2; hemlock’s commercial value, x, xi, 22–23, 68; hemlock’s resilience, 72–74; intensity (two approaches), 200– 201; overall impact, 64–65, 147–49, 197–98, 200–201, 258; Pisgah forest, 236–37 (see also Pisgah Forest); post-hurricane salvage, 36–38, 37, 91, 188–89, 195; preemptive harvesting, xvi, 132–33, 138, 142, 189–90, 191, 193, 197–98, 227, 253–54; protecting large reserves from, 188; protection forest approach, 191–93; on the Simes tract, 140, 143–47, 146 (see also Cut or Girdle experiment) Long, Steve, xvi Longfellow, Henry Wadsworth, 11 Long Term Ecological Research program: ecological modeling, 260–61 (see also ecological models); financial support, 128, 267; foundation species concept introduced, 94; Harvard Forest studies, 84, 154, 195, 205, 232, 255, 264–65 (see also Cut or Girdle experiment; Harvard Forest studies) looper (hemlock looper), 108–9, 250 Lorimer, Craig, 241, 242 Lotti, Thomas, 33, 38, 41 Lutz, Russ, 242 Lux, Heidi, 142 Lyford, Walter, 205, 264, 267

Natural Woodland (Peterken), 241 needles (hemlock): adelgid’s impact, ix–x, xix, 5, 9, 128–29 (see also woolly adelgid); density and alignment, 11–13, 12; as food source, 20; light absorption, 11–15; litter, 13, 14, 234 (see also hemlock forests: forest floor); other pests and, 108–9, 135 New England: adelgid’s spread, 132, 161 (see also woolly adelgid); carbon storage, 156 (see also carbon storage and cycling); colonial

mammals: adelgid transported by, 124; in hemlock forests, 20, 97, 235; hemlock mortality and, 119, 134, 150. See also deer; moose Mann, George, 204

INDEX 302

ment of, 57; rare, 64–65; vs. second-growth stands, 170; Sullivan on, xiii; variability, 34. See also hemlock forests; Pisgah Forest Oliver, Chad, 87, 180, 241, 242 “On a Tree Fallen Across the Road” (Frost), xvii–xviii Orwell, George, 247 Orwig, Dave, 127; publications, 233, 234, 235, 241, 242, 251, 259, 265; studies, 67–68, 116, 142, 259 Oswald, Wyatt, 117

and early American landowners, 216–20 (see also Sanderson farm); droughts, 112, 113, 117, 118, 251 (see also drought); few oldgrowth forests left, 64–65; forest composition, 41–43 (see also forest composition); forests and natural areas, 25–30, 27, 251 (see also hemlock forests; old-growth forests); lake sediments, 44–45, 48, 52 (see also lake and pond sediments); land clearances, 1–2, 62; landscape diversity, 168–70; last glacial period, 54–55; 1938 hurricane (see hurricanes); planting trees, 201; postcolonial forest composition, 61–62; prehistoric forestation, 55, 57, 58–61, 105; protection forest approach, 191–93. See also specific states, locations, and topics New England Box Company, 33–34, 84, 92, 246 New Hampshire, 27, 28–30, 61–62, 158. See also New England New York (state), 27, 161 Nichols, George, 25, 65, 70, 74, 240 Nicoll, Liza, 142 nitrogen, 131, 133, 149, 166, 195, 255 nutrient cycling, 100, 131, 149. See also carbon storage and cycling; nitrogen

Paillet, Fred, 232 Paine, Robert, 94 paleoecology: change vs. stability, xvii; Davis’s impact, 118–19 (see also Davis, Margaret); and ecological modeling, 163–64; and global climatic events, 110–11; importance, 48–49; individualistic concept of ecology supported, 57–58; oak decline documented, 111–12, 116–18, 232–33; prehistoric forestation documented, 55, 57, 58–61, 105–6, 212– 13, 239–40; and the prehistoric hemlock collapse, 106–8, 250–51. See also collapse of hemlock; lake and pond sediments; pollen; and specific individuals pesticides, 6, 201–2, 227, 265 pests. See disease; insects; woolly adelgid; and other pests Peterken, George, 241 Petersham, 218. See also Harvard Forest; Prospect Hill tract; Sanderson farm Phillips, John, 79 photosynthesis. See needles (hemlock): light absorption Pinchot, Gifford, xiv, 28, 83, 120 pine trees: disappearance from Pisgah study area, 41–43; growth rate, 15; logging, and hemlock growth, 89–90; lower branches shed, 11; mountain pine beetle, 116, 251; after 1938 hurricane, 91–92; pitch pines, 95, 96; planting, 201; pollen, 54. See also white pine Pisgah Forest: Branch and, 34, 35; chestnut blight, 70; Cline’s publications, 30–32, 34, 38 (see also Cline, Al); conservation, 33–34, 65; described, 32; downed tree core study (2011), 76; Fisher’s explorations and studies,

oak trees: abundance, 61; carbon storage, 131; growth rate, 15, 17; and hemlock decline, 109, 119, 129; hemlock more vulnerable than, 2; leaf litter, 13; logging, 143; lower branches shed, 11; mortality, xxii, 113, 114, 115, 120–21, 193, 251 (see also gypsy moth); pollen, 51, 54; prehistoric decline, 111–12, 116–18, 232–33, 251 old-growth forests: in the Berkshires, 65–67, 66; catastrophe and recovery, 38, 39, 40, 41–43, 42, 58–59, 69, 70–72; climax forest concept discredited, 68–69; complex structure, 74–76, 75; conservation efforts (see conservation); disturbances and change more common than thought, 62–63; disturbance’s role in shaping, 68–72, 75–76, 241; downed trees’ persistence, 42, 43, 74; durability of old-growth logs, 76; fluctuations between soft- and hardwood dominance, 33; hemlock often dominant, 4, 68; layering seen in, 234–35; postglacial establish-

INDEX 303

pulp (for paper), 23 Putz, Jack, 265

Pisgah Forest (continued) 26, 28–33; Harvard parcel, 33–34; HenrySwan study and paper, 39–41, 71, 87, 238; history, 26–28; Marshall invited to study, 77–78; 1938 hurricane and recovery, 36–39, 40, 41–43, 42, 69, 70–72, 185, 188–89, 192, 195, 238; old-growth forest (ca. 1920), 69; Peterken’s discussion of, 241; pollen data unhelpful, 238–39; role of disturbance in shaping, 68–72; second-growth stands, 72–76; Spurr’s publications, 34, 38–39, 238; study plots, 41; tree growth records, 214; in 2012, 42; uproot mound, 177; white pine, 190 piston corer, 47–48 pitcher plants, 99–100 planting trees, 201 Plotkin, Audrey Barker, 142, 266 PnET models, 260 pollen: analyzing data, 52–54, 239–40; climate change not accurately recorded, 163; Dryas octopetala, 110–11; in Hemlock Hollow sediments, 210 (see also Hemlock Hollow); hemlock pollen, 51; Pisgah data unhelpful, 238–39; precolonial forest record, 62; prehistoric forest record, 105, 106, 109; prehistoric hemlock record, 54–55, 58–59, 105–6, 109–10, 212–13; prehistoric oak collapse, 111–12; useful characteristics for paleoecology, 49–52. See also collapse of hemlock; lake and pond sediments porcupines, 20, 134 Pout Pond (New Hampshire), 108, 250 Pratt, Zadock, and Prattsville, xi–xii prehistory. See collapse of hemlock; paleoecology Preisser, Evan, 135, 252 primary forests, 72–73 Prospect Hill tract: environmental and fossil data, 206; harvesting logs and pulpwood, 133; hemlock forest, 3, 157, 224–25, 228, 229; history of forest changes, 59; long-term study begun, 126; Sanderson woodlot, 205– 10, 212–13, 220–23; streams and watersheds, 196, 206–7, 222; tower, 18, 206; views from tower, 155, 157, 206, 225. See also Harvard Forest; Hemlock Hollow; Sanderson farm protection forest approach, 191–93, 266

Quabbin Reservoir and watershed, 173–74, 191, 206, 266 range maps (generally), 159–60 range of eastern hemlock, x, 7, 20–22, 55–57, 58–59, 240. See also climate change Raup, Hugh: article on “John Sanderson’s Farm,” 204–5, 209, 218–20, 266–67 (see also Sanderson farm); Davis as student of, 106; on the history of the Adams Fay lot, 246; and the Stephens plot study, 172–74, 178–80 “Reconstructing Forest History from Live and Dead Plant Material” (Henry and Swan), 39–41, 87, 238. See also Henry, David; Swan, Mark red efts, 19 red maple, 17, 116, 129 Reid’s paradox, 55, 240 “Research in the Biological Aspects of Forest Production” (Stephens), 263–64 resilience of hemlock: demonstrated in Hemlock Hollow, 213; generally, xxi, 2–4, 55, 184; after the 1938 hurricane, 41–43 (see also hurricanes); in primary second-growth stands, 72–74 ring shake, 22, 73, 233 Rowland, Willett, 264 Sacket, Tara, 142 salvage. See logging Sanderson, Joel, 222 Sanderson, John, Jr., 208, 218–20, 222–23 Sanderson, Jonathan ( John), 205, 216–18, 220 Sanderson farm, 207; Hemlock Hollow sediment study, 53, 56, 59, 106, 109–10, 117, 210–13, 211; history, 217–23, 266–67; Sanderson’s journal, 216–20, 223; scientific assets, 206–8; tannery, 205, 218–23, 219, 221, 267; view from, 215; woodlot, 205–10, 212–13, 220–23 Sargent, Charles, vi, x satellite species, 95, 248 Savage, Kathleen, 142 scale (elongate hemlock scale), 135, 254–55

INDEX 304

tanning and tanneries, xi–xii, 23–24, 218–23, 219, 221, 267 temperature: adelgid and, 6, 113–16, 126, 132, 151, 160, 161, 252; in hemlock forests, xiii, 11, 13, 147; of soil, 147, 149, 195 Terrestrial Ecosystem Model (TEM), 260 Thayer, Abbott, 28, 214, 236 Thompson, Jonathan, 61, 162, 216, 266 Thoreau, Henry David: admired by Fisher and Marshall, 245; musings on pollen grains, 49, 239; woods explored, xiv, 25, 26, 28, 236 Thoreau’s Country: Journey through a Transformed Landscape (Foster), xiv timber salvage operation. See logging Tingley, Morgan, 134, 254 Tom Swamp tract, 21, 98, 246. See also Earl Stephens plot project tree rings, 15, 70, 72, 73 Trouvelot, Étienne Léopold, 113 tulip tree, 15

scent of hemlock, 13 Schoonmaker, Peter, 41, 71, 264 Schuman, Bryan, 46 secondary forests (defined), 72 second-growth forests, 72–76, 140, 170. See also hemlock forests seeds, 2, 19, 20, 55, 235, 258. See also cones; soil serendipity, 213–16 Shaler, Nathaniel, 30 Shaler Hall, 206, 215 Shepard, Ward, 36 Shuman, Bryan, 46, 112, 119 Siccama, Tom, 216 Simes, Olive, 257 Simes tract, 126, 139, 140, 141, 257–58. See also Cut or Girdle experiment soil: hemlock death and, 131, 147, 149, 195; in hemlock forests, 19, 147, 235, 258; logging’s impact, 147, 200–201, 258 Southern states, x, 6, 124, 143, 161 species distribution modeling, 159–62 spiders: hemlock death and, 134, 150, 259; in hemlock forests, 19, 97, 100, 262 spruce trees: adelgid and, 251; dominant in preindustrial Maine, 61; planting, 201; as possible hemlock substitute, 265; postglacial migration, 55; preemptive harvesting, 193; red spruce, 19, 61 Spurr, Steve, 34, 38–39, 205, 238, 241 Stadler, Bernhard, 131, 151, 253, 265 stand-replacing disturbance events, 71–72. See also hurricanes: 1938 hurricane and recovery Stephens, Earl P., 87, 172–80, 205, 242, 263– 64 Stevens, Rodney, 87 Stone, Duncan, 265 stone walls, 218 streams: hemlock and, 17–19, 20, 129, 134; Prospect Hill watersheds, 206–7; and tanneries, 220, 222; water-flow measurement, 196, 206. See also fish; water sugar maple, 13–15, 19, 109 Sullivan, Kelly, 142 survey records, 61, 215–16 Swan, Mark, 39–41, 71, 87, 238 Syracuse School of Forestry, 77

understory. See forest composition; hemlock forests Upham, Burt, 87, 244 Upham, Harry, 87 uproot mounds, 43, 174–75, 177, 179, 264 Vermont, 27, 61. See also New England “The View from John Sanderson’s Farm: A Perspective for the Use of the Land” (Raup), 204–5, 209, 218–20, 266–67. See also Raup, Hugh; Sanderson farm Virginia, 124, 160 von Post, Lennart, 49, 239 Wallis, James, 154 warblers, xvi, 20, 134 water: drinking water, 191–93; flow measurement, 196, 206; hemlock use/loss, 17–19, 166, 234; Prospect Hill watersheds, 206–7. See also drought; fish; streams Webb, Tom, 239–40 Wells, Diana, x wetlands, 17–19, 20, 51 white ash, 95 white pine: downed trees’ persistence, 42, 43; growth rate, 2; and hemlock decline, 109, 116, 129; logging, 89–90, 143, 147; on the

INDEX 305

forest function altered, 129–31, 253; forests already lost to, xiii–xiv; hemlock mortality rates, 124, 129, 135; impact on hemlocks, 4–6, 5, 9, 124, 125, 128–29, 131, 143, 160–61; introduced, 124, 160, 251–52; kill mechanism, 124, 252; long-term consequences, 4–6, 156–59, 213; management approaches, 132–33, 138, 142, 143–47, 146, 159, 189–91, 193, 199–203, 265–66 (see also forest management); map of infestation, 7; migration and range expansion, 4, 124–26, 131–32, 158, 160–62, 170, 225–27; one of many arboreal pests, 183; postinfestation survival period, 143, 253; preemptive harvesting due to, xvi, 132–33, 138, 143–47, 146, 189–90, 191, 197– 98, 227, 253–54; study methodology, 126, 224–25; study purpose, xv; study started, 121–24, 126; trails closed in infested areas, 143; wildlife habitat affected, 134. See also death of hemlock trees; decline of hemlock Wright, Herb, 47, 239

white pine (continued) Pisgah tract, 41–43, 190, 242; planting, 201; postglacial migration, 55; relationship with hemlock, 56, 88–89. See also pine trees Whitman, Walt, 224 Whitney, Gordon, 74, 216 “Why Don’t We Believe the Models?” (Aber), 153–54 Wildlands and Woodlands, xv, 186–88 Wilson, Bill, 234 wind damage, recovery from, 195–97. See also hurricanes Wisnewski, John, 146, 266 witness trees, 61, 240 woodpeckers, 97, 100–101, 129, 134 woolly adelgid (HWA): arrival in Harvard Forest, 151, 259; biology and life cycle, 124, 160, 251; carbon dynamics impact model, 156–59; challenges of studying, 136–37; climate (temperature) and, 6, 113–16, 132, 151, 160, 161, 252; control methods, 6, 201–2, 227, 254, 265; and the Cut or Girdle experiment, 151–52; dispersal model, 160–62, 170;

Zebryk, Tad, 109–10, 210–12, 233, 240

INDEX 306