AGGREGATING FOR WINTER - Winter World: The Ingenuity of Animal Survival - Bernd Heinrich

Winter World: The Ingenuity of Animal Survival - Bernd Heinrich (2003)


All kinds of creatures form tight-knit societies in winter, even those that don’t crash a cozy cabin and even those that don’t need to seek warmth. One November day some years ago I squatted down on the freshly fallen but already matted leaves in the woods. Within seconds I detected the unmistakable odor of stinkbug. Digging under the leaves, I found dozens of them massed together, presumably settling in for hibernation. I’d disturbed them in their bivouac, and they were giving off their foul-smelling defense secretions. I did not need to taste them—I knew they tasted as bad as monarchs (but not nearly as bad, according to this gourmet, as a mass of overwintering spider eggs!).

It is not only the stinkbug that smells or tastes foul. Almost any insect that is brightly colored except some mimics of them) is sure to do the same. The ladybird (or ladybug) beetles that aggregate by the thousands both in my Maine cabin and my Vermont home in the fall, when hoping to stay the winter and when unduly disturbed, put out a foul burnt-rubber smell that is overpowering. Some species of these beetles of the family Coccinellidae aggregate by the millions, and in California and other areas of the western United States where they mass up under rocks or at the base of trees up in the mountains, they are scooped up in buckets and sold to gardeners for aphid control. The reason for aggregating in both stinkbugs and ladybird beetles is likely for the purpose of massing stink power. That is, if you want to be associated with stink, make your own, look like a stinker, and better still go where others stink like you.

Aggregating in winter brings animals, at least some snakes, other advantages. One of the most amazing snake aggregations are those of the Manitoba red-sided garter snakes (Thamnophis sirtalis parietalis). Like stinkbugs, garter snakes give off a foul smell when you tread on or otherwise molest them. Each fall writhing masses of these snakes pack themselves like live spaghetti into specific crevices—ten thousand in a single depression of a few cubic feet—in the rocks of a barren region near Winnipeg. The snakes spending the winter at these spots avoid freezing and gain protection. While they’re all together, they perform another primal function just before dispersing in the spring.

Males emerge from the rocky crevices before females and then wait at the periphery to intercept the females. As soon as a female emerges from the den, she is enveloped in a ball of dozens of suitors. Curiously, some of these males mimic females, and are mistaken as such by other males (Shine and Mason 2001). Their behavior just doesn’t make sense (yet), but with more information I trust that it eventually will.

Aggregation behavior has its share of such mysteries. Some years ago I received a letter from a man in Alaska who wrote of seeing a communal crow roost in winter woods, where “the ground was littered with fighting crows who were murderously hacking one another to death.” He had never seen anything like it. Neither have I. Nor could I make any sense out it, no matter how hard I tried to twist the scenario into a logical possibility.

Communal bird roosts are not closed societies. In the Maine woods, my colleague John Marzluff (who worked with me on ravens for three years) routinely introduced long-captive ravens into established communal raven roosts, and these new birds showed no hesitation in joining and were immediately accepted by the group. The next morning they then followed the roost occupants to the crowd’s feeding place, such as a deer or cow carcass. There was never any exclusion of strangers. Communally roosting birds even tolerate other species. Crow roosts in the Old World sometimes contain magpies, jackdaws, and ravens. Although on the whole, roosts of corvids in North America tend to be species-specific, large roosts of icterids (blackbirds) near Burlington, Vermont, sometimes contain red-winged blackbirds, common grackles, and cowbirds. There seems to be almost no limit to the numbers of birds tolerated. There are reports of crow roosts in the western United States containing several million individuals.

Nothing in the scientific literature about roost behavior would allow for group fighting, and birds that fight and are territorial during the breeding season relinquish their antagonisms in order to be in roosts. What could there be to fight over, since the occupants join up because they ultimately need each other? Also, when social birds do fight, they don’t hack themselves to death. (So how do I explain that letter from the man in Alaska? Bear with me a bit and I will try.) Crows sometimes act aggressively, but from what I have seen it has always been a group of them attacking another individual; I have never seen more than one intended victim at a time. It seemed bizarre that crows would wage the equivalent of a war at a communal roost, and that they would stay and fight there, without escaping, until bloody mutual destruction had set in. In short, I did not believe what I read because it did not fit into my previous knowledge, preconceptions, or experience. It just seemed to be another one of those bizarre reports that I hear all the time that are almost invariably the result of false identification or faulty observations. I would therefore have banished the report of the murderous crows from my mind had it not been for some observations of crows I made one evening in the city of Burlington, Vermont.

I have watched crows every winter for the past twenty years as they come to the Burlington area each winter. At dusk they start to form their communal sleeping roosts, which number in the thousands. They come flying into their roosts in endless long strings or queues flying high, forming diffuse, gray cloudlike aggregations against the snowcap of distant Mount Mansfield. Just when I think the last of them have arrived, I see many more behind, in what seems like an endless stream. All of them converge on one darkening spot near or in the city.

As spectacular as these flights to the evening roost are, to me the most notable thing about them is where the birds settle for the night. The roosts are never in the forests where crows commonly nest in the summer, nor are they in the pines along the fields that crows like to inhabit. Curiously, the city center itself now seems to be the preferred winter roost site, as the birds settle not far from the hustle, the bustle, the traffic, and the lights. Once I saw them roost in a patch of pines flush up against the I-89 Interstate highway, where they were surrounded by the busy Burlington exit ramp.

One evening I watched them again as they came into the heart of the commercial district on Church Street. Round and round they flew in swirling clouds above the evening town crowd going to restaurants and theaters. It seemed as though they were looking for a place to land. I watched them fly over patches of trees at the edge of town that looked to me like ideal roosting sites, yet the birds still kept coming back into the center of town. Eventually they settled in several young cottonwoods next to Bove’s restaurant. The birds were soon closely packed upon the branches, as ever more continued to stream in. That is when it suddenly occurred to me why the roost was here rather than elsewhere.

For decades there has been a heated debate about why birds join flocks. In the 1950s the British biologist V.C. Wynne-Edwards speculated that birds form communal roosts in order to assess their population size, so that they could then decide whether it was appropriate to reproduce, in order to maintain a stable population. To biologist’s ears now this idea sounds about as plausible as that of the sun orbiting the earth to an astronomer. Wynne-Edwards’s theory on bird flocking for the common good at the expense of the individual was termed “group selection.” Animals can act cooperatively, but the specific example of Wynne-Edwards’s theory sounded so ridiculous that his baby was soon summarily tossed out. But other ideas were spawned. William D. Hamilton came along and proposed the “selfish herd” hypothesis, which posited that animals form groups for their individual safety, using one another as shields. By being in a communal roost, the individual birds also profit from many eyes with which to see approaching danger. Hamilton’s hypothesis made sense, and it also fit empirical observations.

The Israeli biologist Amotz Zahavi then vigorously promoted a third hypothesis that seemed plausible to account for other group behavior as well. Zahavi proposed that when birds roost communally they gain information on where to find food. His “information center,” or IC, hypothesis provoked a good deal of controversy and generated hundreds of papers in the learned journals, perhaps because it looked suspiciously like a form of “group selection.” Our data on raven communal roosts in Maine, principally the demonstration of naive birds being led to cattle carcasses after joining a roost (Marzluff, Heinrich, and Marzluff 1996), gave perhaps the first empirically tested proof of Zahavi’s IC hypothesis. (Of course, there are still those who, quibbling about mechanism, claim that the raven’s behavior is “group selection,” unless it can be defined in terms of tit-for-tat reciprocation among individual birds, not exploitation of individuals for information held within the group.) Neither hypothesis would explain the golden-crowned kinglets’ tendency to congregate at night, because being in hiding and possibly also being in a torpid state, they could not instantly respond and escape a predator, even if one of the kinglets gave an alarm. Nor would they benefit from the group vis-à-vis finding food, as their food is highly scattered. What they do group for is warmth; unlike corvids, they seek body contact.

As I watched the crows in the city and wondered which if any of the existing aggregation hypotheses might apply, I was reminded of seeing other huge crow roosting aggregations elsewhere in both North America and Europe. Crows used to be thought of as strictly rural birds, but in the last fifty years they have started roosting in cities all over the world. There was clearly something significant in the way that the birds avoided the woods to be in the town lights and the bustle. It was not just a stray observation. The birds were free to roost in forests available within a half mile of the city, yet they had flown miles just to come here where they had to search long for a suitable landing site, eventually choosing the few available trees downtown.

Crows have greatly increased in numbers all over the North American continent over the last century, and they have been moving ever increasingly into cities. As one indication of both of these suppositions, I refer to a 1946 report in the Oklahoma Game and Fish News (Vol. 2: 4-7, 18), written by H. Gordon Hanson, a biologist of the Oklahoma Fish and Game Commission. His report gives the map locations of 47 “major” winter crow roosts in Oklahoma, those with 200, 000 or more crows per roost. “It all started back in 1933,” Hanson writes, “when it was first brought to the Oklahoma Game and Fish Commission’s attention that the numbers of crows wintering in the state had begun to increase alarmingly and to spread over a much larger area.” The crows fed on crops and came from the northern nesting areas in the prairie provinces of Canada where they “molest the nests of waterfowl in the duck factories.” Although the winter crow roosts in Oklahoma had increased considerably in size and number, the Oklahoma Commission perfected metal cylinder bombs filled with steel shot and dynamite and then carried out an annual crow-bombing campaign. In eleven years government bombers bragged up a tally of approximately 3, 763, 000 crows killed. Given these data—and the probability that people living in cities would find the dynamiting objectionable—I speculate those bombed, and perhaps other, crow roosts were out in the country, instead of in cities as is so often the case now.

Crows are now federally protected birds, safe from the likes of state game commission dynamiters and others who took their lead to kill crows at every opportunity. Nevertheless, they also have their natural enemies, primarily great horned owls. Great horned owls nest in the winter, when they require much food to feed their fast-growing young. I have found two of their nests and heard them sing at night within two miles of downtown Burlington. The owls are perceived by crows to be their greatest enemies, and if they find one, the alarm goes out on the roost and dozens of crows converge quickly to harass the owl relentlessly, with the goal of driving it as far as possible away from where the crows will sleep at night. As my detailed observations of Bubo, a tame but free great horned owl, in One Man’s Owl revealed, crows easily outmaneuver these great but somewhat clumsy predators in the daytime. At night it’s a different story. It pays the crows to be in Hamilton’s selfish herd at night, even as it pays them to be in Zahavi’s information center at dawn. The two are not mutually exclusive. A communal roost can serve more than one function. The more benefits it confers, the more it is likely to evolve and to be maintained.

What I thought I was seeing in Burlington—crows trying to get into the city center—was no aberration. In one study of crow roosts in the Sacramento Valley city of Woodland, California, biologists Paul W. Gorenzel and Terrell P. Salmon document that the common crow’s winter communal roosts are preferentially located in commercial (rather than residential) areas of cities, characterized by high nighttime light levels, and paved parking lots and commercial areas that have high noise and disturbance levels from vehicles and people. There are no great horned owls prowling downtown Woodland or Burlington or most other well-lit, noisy city centers.

That finally brings us back to the inexplicable anecdote about the crows “fighting” and “hacking” one another to death in a roost. What we see is not necessarily what is. The story is almost always in the details. The observer had mentioned that the episode occurred in the woods, hence it might have been in a location where one or a pair of great horned owls had access to the roost. It also could have been on a heavily overcast night in which much confusion ensued when the predators struck. With perhaps thousands of birds fluttering around in confusion, the predators would have maimed many crows, which would have then fallen, injured, to the ground. I have myself once seen an owl-killed crow, from which only a small portion of flesh was removed. When the observer arrived on the morning after an owl attack, he might have heard a bedlam of alarm from the injured crows. With many crows all around in the dark, the owls would not need to hold back. The carnage of crows that was described at the crow roost reminded me of a scene related to me by ornithologist Jeremy Hatch where a great horned owl raided a tern colony: “A dry pond-bed was strewn with about forty corpses: most headless, wings are generally torn off, or at least broken. Occasionally legs are gone. No evisceration. No plucking.”

Crows in pain and in fear and not knowing what had hit them would blame and may strike out in frustration and anger at others near them. It is impossible to reconstruct what actually happened in the case of the warring crows, but there is one thing of which we can be absolutely sure. There are many advantages in different animals to aggregate in the winter besides keeping warm, but doing so in order to fight is not one of them.

A currently more public and more scary myth than that of the presumed warring crows is that the species might become extinct due to the much publicized West Nile virus. Crows occasionally die from this virus (and other causes), and there are cases of humans killed by it as well. After a dead virus-infected crow is found, “all" the crows may disappear in that area. But there is no reason to conclude that “the first thing that happens when the West Nile virus invades an area is that all the crows die—" (“The Silence of the Crows,” in the Washington Post, 30 August, 2002), much less that we are “witnessing the disappearance of the crow from the American landscape." I see no evidence that these birds are disappearing from “the American landscape"; there is no evidence that they are being depleted either by mutual warring or by the virus. If anyone finds a virus-killed crow, it is most likely to be simply where there are a lot of crows—a communal roost, and since roosts are strictly temporary, lasting only through the winter, it is therefore a given that the crows disappear after some die. We don't need another round of government bombing to kill 200, 000 crows in a roost to get rid of a few sick ones.