Mind of the Raven: Investigations and Adventures With Wolf-Birds

by Bernd Heinrich

  This anecdote reinforced my impression that the ability of these birds to anticipate the actions of others, coupled with their good memory, are traits that can compensate in competition with larger and more dominant associates. My observations were possible only because I was so closely in their midst. My rearing them from nestlings, and daily association with them for ten months, had won me their trust, which made the expression of their fine-grained unfiltered and hence complex behavior possible in my presence. The aviary also compensated for my inability to fly. I could follow them here, while at the same time provide an experimentally crowded situation that elicited flexible and innovative behaviors that otherwise might occur only rarely in the field where the birds can more easily avoid each other if they choose.

  The ravens routinely recovered food they saw me bury in the snow. They also seemed to anticipate the actions of not only other ravens, but other animals as well. Gerald Fitz, a ravenphile from Lowell, Vermont, is convinced that his tame raven quickly figured out his buddy, the Fitz’s beagle. Gerald had rescued the raven as a nestling from a local granite quarry, where a raven’s nest had been destroyed three years in a row by blasting. The raven had grown up with the beagle. The two played with each other, and the raven followed the dog around, apparently enjoying its company. In the winter, the beagle followed the raven when it cached food in the snow, predictably taking the hidden food. The raven responded by caching in high places that the beagle could not reach. He also cached food in vertical bolt holes that had been drilled into a sill. After caching meat in the drill holes, he capped them with pebbles that fit snugly so that the dog could not lick them or the meat out.

  Terry McEneaney, who has been watching ravens for a long time, has many stories to tell. “I saw the pair of magpies near the house here {in Gardine, Montana} become terribly excited, vehemently scolding a raven. I knew something was up so I continued to watch. Soon the raven approached the magpie nest in a juniper tree. Since a magpie’s nest has a nearly impenetrable dome of thick limbs, the raven couldn’t just tear into it. The raven proceeded to methodically pull out one stick at a time, until it had finally made a hole. Then he reached in and pulled out a pinfeathered young, flew by the house, and cached it nearby under a bush in the open sagebrush. He flew right back, got another young, and cached it in a different direction under another sagebrush. He then repeated this with a third and a fourth young. He kept the fifth one in his bill, and flew off with it and went right by his nest in the nearby cliff, as if to show his mate, who was brooding their own young. She left the nest and joined him as he cached this fifth one, then she followed him back to the magpie nest. She now took the sixth young magpie and brought it back to their nest to feed their three raven young. He took the seventh and cached it under another sagebrush, then came back to get the eighth, the final one. This one he did not cache. Instead, he took it into a dense juniper where he had some protection from the screaming magpies, who had in the meantime recruited helpers to harass the offending ravens. There, in the safety of the branches, he dismembered the young magpie and ate the whole thing.”

  Lesly Woodroffe of Alna, Maine, told me a similar anecdote of a raven that raided a nestful of four young cottontail rabbits. It decapitated them, cached them one by one, and only then stopped to eat.

  Ravens’ policy is: “Cache while it lasts and eat later.” At the calf, deer, and moose carcasses where I have watched ravens, most of the participants exhibit almost frenetic activity as they haul off meat, especially if the meat can be torn off in large chunks. On December 24, 1991, I stretched out under a spruce blind covered with snow to watch a crowd of about forty ravens disassembling a calf carcass. In one 135-minute stretch, my four individually marked birds made twenty-three, twenty-one, eleven, and nine trips, respectively. The birds worked steadily all day long, so the forty birds may have made more than four thousand caches that day. Knowing how much they can eat per day and how much they carried, I presume they could each only have eaten the equivalent of one to two caches in that time.

  The average time per caching trip in the incident I have just described was ten minutes. This seems long. When caching in soft snow, ravens merely thrust their bill with the meat into the snow, then release it with a push of the tongue. Loose snow from around the hole falls in and covers the meat as the bird retracts its bill. Usually, the raven makes a few more shoveling motions with its bill to draw more snow in from the sides. When the snow is crusted, the raven will lay the meat down onto the crust, peck a hole through the crust, then pick up the meat and push it down into the hole. It will cover the hole with pieces of crust or shove snow over it, as appropriate. Similarly, on bare ground, a raven will thrust food into crevices or holes, or dig holes with its bill if none exist, and then will pick up such debris as sticks, leaves, or grass from nearby to cover the food. Ravens also cache up in trees. Making the cache itself rarely requires more than half a minute, but it may take many minutes to tear off enough meat for any one cache. Just before flying off with the throat bulging full of meat, and with additional meat sometimes dangling from its bill, a raven usually stops a second or two, looking in all directions and blinking its eyes. Then it flies far off in one direction with vigor and resolve, as if having decided where to go before departing. Almost every trip is in a different direction. No two caches are ever in the same place. The big question posed by these observations is: Why do they go to all of this trouble to fly long distances, and to disperse all of their caches, when they could potentially cache nearby without flying, or flying only short distances? The key is, individuals or lone pairs cache much closer to the source than do birds from crowds.

  Unlike the pair that McEneaney watched and the many pairs I also have watched, ravens in crowds in the field almost always cache food very far from where they get it. I have seen ravens from these crowds in the wild make many thousands of caching trips, but I’ve seldom seen them actually make a cache. Most likely, the ravens near the food pile do not get to see others making caches, either. That’s probably no accident. Ravens may have to make their caches far enough away so that other ravens don’t find them. Watching from the top of a tree, as a raven might, I sometimes see the birds carrying meat fly off out of sight, over the next ridge, down the valleys. When I’ve found cache sites—tracks and scrapes in the snow—it has been only by walking at random in the woods. Caching close to a carcass not only invites thieving by other ravens, it also poses another danger. Such keen-scented carnivore scavengers as coyotes, fox, fisher, raccoon, and weasel, as well as other animals from bumblebees to squirrels, concentrate their search for food near where they previously have been successful. That is, if dozens of ravens put their caches in the same area, all the caches would be in greater jeopardy than if they were spread out.

  All of my observations of raven-caching behavior indicated more flexibility than had ever been observed in any other animal. I wanted to get publishable results, and turned to more experimental conditions for a tighter focus on specific aspects of the caching behavior, such as memory, using fifteen individually identified wild-caught birds who were kept in one of the side areas of the aviary off the main complex. From this pool of birds, I would capture from two to four birds at a time and place them into a second side area of the aviary. Continuously well fed birds don’t cache food (Gwinner, 1965), so I kept them without food for two days to induce the caching response. After the two days, they were allowed into the large experimental enclosure where a pile of chopped meat chunks had been placed on the snow. Up to an hour commonly elapsed before the birds ventured to contact the food, but after they did, they usually began caching immediately. I allowed them another half hour of feeding and caching, then chased them back to their side aviary. I waited for different durations, from one day to one month, before I would let them back into the experimental area to see if they could recover their caches. If the memory interval to be tested was more than two days, the birds had to be fed in the meantime, and then again allowed to go w
ithout food for two days before the retrieval part of the experiment, so that they would be motivated to recover caches.

  I found that the birds easily remembered caches made a day or two before. They had poor ability to remember caches made two weeks earlier, and they were virtually unable to recover month-old caches. As usual, some of the details of these experiments, which I began in February 1992, were more interesting than the general results.

  In one experiment, White Slash and Blue Diamond (names referring to symbols painted on wing tags, which I’ll here shorten to Slash and Diamond) were allowed on February 4 to feed and to make caches. For the first two and a half hours, Slash spent most of her time flying about. She fearfully approached the bait—chopped pork lung—dozens of times, doing jumping jacks with her bill open in fright. Finally, she edged close enough to grab a piece. Diamond had perched nonchalantly during that time, but after Slash got a piece of meat, he suddenly became active. He followed her as she fed briefly. She eventually cached her piece of meat, but not until Diamond was almost but not quite out of sight at the opposite end of the aviary. Immediately after Slash cached, Diamond flew over and dug in the snow at her cache site for three minutes. The snow was deep, and he apparently had not watched closely enough to succeed in recovering Slash’s cache. A few minutes later, Slash herself showed no problem recovering her own cache. She then continued feeding.

  I left them without food for another day, hoping they would become more motivated to make multiple caches. In the morning session, they both acted uninterested in the pile of chopped pork lung. In the afternoon, Slash approached the meat within one minute after I put it down. She fed, and then again made a cache in the snow. As on the previous day, Diamond immediately flew there to dig. Strangely, Slash did not intervene. Instead, as Blue was preoccupied in trying to dig up her first cache, she erupted into a virtual caching frenzy, making seven caches in seven minutes while Diamond remained thoroughly preoccupied with his digging. Finally, he again gave up, then came back to watch Slash, who abruptly stopped caching, instead flying round and round the aviary with a chunk of meat in her bill, as if suddenly not knowing where to put it. She finally cached it only when Diamond once again resumed trying to dig up her initial cache, the only cache he watched her make. This time, Diamond was finally successful in finding the meat. After eating it, he came to the bait pile himself. When Slash was preoccupied in feeding there also, Diamond suddenly made four caches in three minutes. Afterwards, I chased both of them out of this portion of the aviary into the adjoining aviary in order to learn if they could recover their caches the next day.

  On February 6, I allowed both birds back into the enclosure to find out who would recover which cache, and how fast. Although Slash had acted frenziedly the previous day while caching, she was calm and composed today. Nothing seemed to disturb her, not even me. She calmly went to one of her caches and ate the piece of meat retrieved from it. Then she flew to the next one, and so on.

  Diamond, who had the day before made only four caches of his own, first went to where Slash had made three caches and probed briefly in the snow. He did not seem to know the precise locations and recovered no meat, although for Slash he was probing too close for comfort. She flew over. Diamond apparently knew Slash’s intent to punish his intended trespass, because he made a submissive display before she even got there. He had never made submissive displays to her before. Slash did not attack. Instead, she walked by him and simply reached down into the snow and pulled out her cached meat to hide it elsewhere.

  The observation done, I opened the gate leading to the main aviary where the raven crowd was feeding on a food pile. Slash instantly flew in to join them at their feast, grabbed a piece of meat, and came right back into the part of the aviary from which I had just chased her! Although she was a wild-caught bird, she perched with her back to me. She held the food in her bill and all the time watched Diamond instead of me, probably waiting for an opportunity to cache when her competitor was not looking. Neither paid much attention to me. Their attentions were on each other and where to cache their food so that the other could not see them.

  In another experiment on February 10, I observed recoveries of caches that Yellow O and Diamond had made the day before. Diamond had made thirty-three caches, and Yellow O only two. In one hour, Diamond recovered eight of his but made ten mistakes—six unsuccessful recoveries of his own caches, and four apparent recovery attempts where there was no cache. Yellow O, in contrast, recovered both of his caches plus two more of Diamond’s, and made no mistakes. Although the birds probably were capable of recovering many more caches, they had no need because they were soon satiated, so I chased them out to continue the experiment the next day.

  Later in the month, on February 27, Number 106 Blue and White Blank were tested for their ability to recover the caches they made two weeks ago. White Blank had made five caches, and today went to all five. Number 106, a new bird from the same group, had made thirteen and tried to recover only three. He dug for a full eight minutes at one site, before he got his piece of meat.

  I later allowed four birds to cache simultaneously, removed them from the aviary, and let them back in one month later. Only two of these birds had cached, and these two walked continually all over the enclosure giving the impression of searching. Perhaps they knew the food was in this enclosure somewhere, but they didn’t seem to know exactly where. Meanwhile, the two birds who had done no caching at all sat quietly, hardly moving. They acted like birds in the control experiment.

  The control consisted of forty food caches that I made in the birds’ absence (while they were out of sight in a side aviary). I marked the locations of those false caches in snow with branch sprigs and twigs. Then I let the birds in. None of the forty false caches was recovered by them. I recovered them all myself after one day.

  My experiments were the first to prove that any bird could remember others’ cache locations, and my anecdotes consistently demonstrated that ravens also have counter-strategies that foil the considerable cache-parasitism capabilities of their fellows in a crowd. In experiments to test this three years later, in 1995, I provided Fuzz, Goliath, Houdi, and Lefty with piles of food when they were either all four together as a group, or when they were each provided food in isolation, one at a time. In the first situation, when I put down a pile of food chunks, all four instantly descended onto it, grabbed all they could, and started caching immediately. Their caching trips were short. Each bird attempted to return immediately to the food pile to get another load of meat, as if to carry it away before the others might get it. Only the bird who eventually got the very last piece of meat from the pile invariably did not cache that piece immediately. Instead, it carried the last piece about, leisurely fed from it, and eventually cached what remained of it with great deliberation. Similarly, when only one bird at a time was given access to the same amount of meat, caching proceeded at a leisurely pace. The birds therefore appeared to monopolize as much of the temporarily available food as possible, and caching was their way of choice. If that was all, they should have cached closer to the source when they were in a crowd than when alone, to increase caching speed. Instead, with others present, they went farther to cache, going into the side aviaries where they could cache in private. The results made sense if the birds also attempt to take the uncached food away from others and to hide their caches from their sight.

  Ravens’ caching behavior relies on their capacity to “track” objects. If they see a piece of food put into the snow, they “know” it is there and may dig six or more minutes for it. As soon as I or another bird removes it, they stop. Their minds seem to be like a chalkboard. They have the capacity to register many locations on it, but with each retrieval they “erase” that site from further consideration. I saw something even more sophisticated in one probe. In four different trials, I showed three of my ravens two halves of a cashew nut, and as they watched me from a perch, I buried the halves next to each other in a bare patch of snow. White t
wice found one of the nuts in less than a minute, then left. Orange showed faith that there was more to be had by continuing to dig after the one piece had been removed. The pieces of nuts were difficult to find in the loose, powdery snow, and in two trials he dug for 3:15 and 2:40 minutes, respectively, without finding them. In the third trial, Orange got one piece in 75 seconds, and both Orange and White continued to dig for another 2:50 minutes before giving up. At the fourth trial, Orange got a piece of nut in five seconds, then Red joined in and they both dug for 70 more seconds.

  Although it may seem logical to suppose that the ravens have awareness of others’ actions because they anticipate others’ responses before they occur, not everyone agrees with that conclusion. The alternative supposition is that the behaviors are stimulus-response phenomena in an infinitely complex chain of unconscious reflex responses to stimuli, ad infinitum. Most behaviorists would be convinced of awareness if, as has been shown with some primates, ravens lie. What is a lie? I would argue that pretending to make a cache and then actually hiding the food someplace else amounts to a lie.

  In deception, a signaler behaves in such a way that a receiver registers something that is, in fact, not occurring. As a result, the signaler benefits and the receiver pays a cost. That is a behavioral ecologist’s definition. To a psychologist, however, lying is not only the giving of false information. Lying, from the psychological perspective, implies that the giver of information is aware of the fact that he is giving false information and that he knows this information will be interpreted by an intended receiver. I have evidence that ravens lie when I wear my hat as a behavioral ecologist. However, as I have indicated, I attempt to go beyond that perspective in this book.