experiment before. First, long-held captive birds, who could not have had knowledge of any food bonanzas in the field, joined crowds at roosts without hesitation after we released them near the roosts in the evening. The next dawn, they showed up at the baits where that particular roost crowd was feeding. Control birds released on the same evening without access to a roost did not show up at the bait. This was definitive proof that the communal roost served as an information center.
This year we would do the reverse experiment of releasing potential recruiters at baits rather than potential recruitees at roost. A huge complication to be expected was that the roost birds were already feeding at other carcasses, and hence uninterested in changing feeding sites. We would try to restrict the number of potentially competing carcasses by removing those we knew about.
To set up this new experiment, we captured twenty wild birds and maintained them for months in our large aviary in the Maine forest. John watched these birds daily to tabulate who made submissivegestures to whom at food. He found that they were all aligned in a dominance hierarchy, starting with the most dominant bird who challenged all and yielded to none, and to whom all others yielded. We could then pick birds from either end of the dominance spectrum for our recruitment trials in the field. We had twenty radios for this experiment, which we would use to track the birds.
The result of the first release of the year, at a food pile near the inlet to Lake Webb, was spectacular! Freed at dusk, our radioed bird did not feed, although she had not eaten for two days. Unlike most of the others, she did not bolt away either. Instead, after we opened the door of the cage, she calmly walked to a puddle near the bait and drank. Then she flew onto a tree above the bait and preened vigorously for half an hour. Next, cawing loudly, she flew north in the direction of a roost. Roosts are noisy at night, and perhaps she heard the din. We knew that the birds at that roost had just finished feeding at another bait. All the conditions that one can never control for in the field were miraculously just right. Better still, our radio signals indicated that the bird entered the roost that night. The result the next dawn was stunning: At first light they came—a string of more than thirty ravens, all flying directly from the roost to the bait that they could have known about only from our radioed bird, who was in or near the lead! Results like that convinced us that ravens can and do recruit from the roost.
For the particular trial I’m about to describe, John picked a bird of low status. Our studies had already shown that the reason only vagrant juveniles recruit to a food bonanza is that it allows them to overpower the territorial adults to gain access to food, although there could be other reasons as well, such as wanting company during feeding at food. We now expected such subordinate juveniles as our present subject to recruit more than dominant juveniles, because subordinates should have the most to fear from the adults. On the other hand, in a crowd of juveniles, they experience less aggression at a carcass or bait, since the dominants fight with each other and are also attacked by the defending adults, thus giving the subordinates more chance to get past the socially distracted dominants to get at the meat.
John attached a radio transmitter to the bird’s tail. He also attached a large red plastic tag to each wing. Each tag had a large number written on it, enabling us to identify the bird not only by her radio signal but also by sight. He then put her into a pet travel-cage and did not feed her for two days so that she would appreciate the super-bonanza of food we had in store for her. Meanwhile, I drove the two hundred miles from Vermont, in a snowstorm as it happened, to help conduct the experiment. My job was to release this bird at a fresh food bonanza, about 150