A new paper in Current Biology by Désirée Brucks and Auguste van Bayern shows what looks like altruistic food sharing in grey parrots (click on screenshot below; pdf here; reference at bottom).
I use the term “altruistic” a bit loosely, as true biological altruism occurs when an individual sacrifices its reproductive potential to enhance that of another and unrelated individual, while what we see in the parrots may be an evolved “reciprocal altruism” that is a byproduct of an evolved social behavior in the wild. If you help another individual with the expectation of a return, you might actually be enhancing your fitness by sharing, while true biological altruism reduces your fitness. That’s why we don’t expect “true” altruism to evolve in nature, and we see it, as far as I know, only in humans, where it’s a cultural phenomenon. And we expect reciprocal altruism, in which you scratch someone else’s back who then scratches yours, in species having small social groups in which individuals can recognize and remember each other. (This is one explanation for reciprocal altruism in our species, who for millions of years lived in small bands.)
The two European researchers studied both African grey parrots (Psittacus erithacus) and blue-headed macaws (Primolius couloni): the experiment was designed to test whether parrots trained to exchange tokens (metal washers) for food (walnuts), would give a conspecific individual a token so that it could get a walnut when the token-giver couldn’t. That is, were the parrots acting selflessly?
The gray parrot, famous for its intelligence, is shown in videos and pictures below, and here’s a blue-headed macaw:
Here’s the paper:
Individuals of both species were trained to exchange metal tokens for walnuts by putting the token into the hand of a researcher that was stuck through the hole of a two-compartment plexiglas cage. There was also a hole between the compartments themselves, each of which housed a bird. The object was to see whether, when a bird that had tokens but no access to walnuts, it would pass the token to a bird in the next compartment—a bird that could exchange it for walnuts. There were four conditions of the experiment: a test and three controls. Sample size was small: 4 pair of grey parrots and 3 pairs of macaws.
There were four conditions of the experiment, as shown in the figure below:
(A) was the Test condition (T), in which a bird had tokens but couldn’t exchange them for walnuts (exit hole blocked as shown by blue bar). It could, however, pass the tokens through two holes in the partition to the bird in the left compartment, who could then exchange them for the treat (hole open in that side, experimenter’s hand shown ready to accept tokens and hand out nuts).
(B) is the “social control” (SC), which tested whether the availability of nuts to the recipient was essential in passing the tokens between birds. As you see, neither bird can get nuts because the holes to the experimenter are blocked.
(C) is the “non-social control” (NSC), which tests whether birds who can’t get nuts will nevertheless pass tokens through the hole even though there’s no bird in that compartment (and that side has the hole open to the experimenter).
Finally, (D) is the “motivatonal control” (MC), in which only the bird that already has token can get nuts, but the other bird has no access to nuts. This tests whether the bird will pass tokens to its partner when the partner stands to gain nothing. In several cases the partner bird without tokens would beg for the token by making soft noises, but this occurred only when that bird knew it could exchange the tokens for nuts.
Before I show you a video of what the experiment looked like, here’s a graph of the results, showing how many tokens were passed under all four conditions, and in each species. Grey parrots are light gray, while the macaws are dark gray.
It’s clear that nearly all the tokens were passed by grey parrots, and nearly all of those under the test condition (T)—when the recipient could get nuts but the original token-owner couldn’t. Some tokens were passed by the grey parrots under three conditions (but not when the proposed recipient couldn’t get nuts). The bars show that the test condition resulted in significantly more tokens passed than the other conditions (stars indicate significance), and this difference holds for grey parrots (starred bars) but not for the macaws (non-starred bars; the interpretation of these lines is a bit confusing, so I may be wrong). But seven of the eight pairs of grey parrots showed token passing when the partner was in need.
As the authors note:
The majority of instances in which the AGPs [African grey parrots] directly gave tokens to their neighbor occurred when their partner was in need (i.e., in the T where it could exchange tokens for food; Figure 3; 82.6% of all direct giving transfers), whereas only 15.5% of direct giving instances occurred when it had no exchange opportunity (i.e., in SC when its exchange hole was blocked). Thus, the results of these two control conditions (non-social and social) confirm that the motivation underlying the subjects’ token transfers most likely was to help their neighbor to obtain food.
This video gives you a clear picture of how the experiment worked:
Such a cool study! Those macaws are nasty pieces of work!
Another observation is that pairs of greys who had previous experience of each other—for example birds that had fed each other when together or preened each other—tended to pass more tokens in the Test condition than birds who didn’t “know” each other in this way.
What we see here, apparently, is the first instance of a non-mammal showing “prosocial” behavior under test conditions: helping other members of your species when they are in need. This has previously only been seen in a few species of any type. The authors summarize:
What puzzles me about the results is the authors’ explanation of why grey parrots are prosocial and macaws are not (indeed, macaws sharing a space would often grab bowls of food away from each other rather than sharing). One would think that if the behavior reflects evolution in the wild, a species that lives in small groups, in which individuals get to know each other, would be more likely to evolve reciprocal altruism than species forming large flocks in which individuals don’t get to know the others, thus having little chance for reciprocity (you have to recognize individuals to evolve that behavior). Yet it is the altruistic grey parrot that forms large flocks in nature, while the selfish macaws form small bands: exactly the opposite of what I’d expect. In fact (see below), gray parrot flocks are not stable, and divide up during foraging: another reason to expect that reciprocal altruism wouldn’t evolve.
But the authors somehow interpret the results the other way around, and I don’t see why. The New York Times story on the experiment, for example, says this:
The researchers think different social systems in the wild may help explain the different results. African greys live in huge, constantly shifting flocks. It might be important for the birds to immediately build good reputations, so that if they need help in the future — such as extra food, or help chasing off a predator — they’ll get it. Blue-headed macaws live in smaller, unchanging groups. So quickly building up a reputation might not be as important.
That is “a thought-provoking potential explanation,” said Katherine Cronin, an animal welfare scientist at Lincoln Park Zoo in Chicago, who has studied helpful behavior in animals.
“Immediately building good reputations” is not a convincing explanation for me. Cronin adds it’s counterintuitive, like I noted, but the paper doesn’t point that out, and in fact suggests that the difference in social group size explains the difference in behavior. But the authors don’t say how, even while pointing out the difference:
The limited available observations of wild populations suggest that AGPs form huge flocks for roosting of up to 1,200 individuals with fission-fusion dynamics during foraging, whereas BHMs, albeit clearly very social, seem to form considerably smaller and probably more cohesive flocks.
That would suggest that the evolution of prosocial behavior is easier in BHM (blue-headed macaws] than in the grey parrots. But what we see is the opposite of these expectations. Well, what we have is a mystery, and there may be other aspects of these birds’ ecology that could explain the difference. Still, the experiment is really nice, and the results, despite the small samples, pretty telling.
Here’s a picture of prosocial token sharing under the test condition. You can see that the grey parrots dispensing the washers can’t get food, but the food hole is open to the recipient:
______________
Brucks, D. and A. M. P. von Bayern. Parrots voluntarily help each other to obtain food rewards. Current Biology, online.
Sub
sub
Always risky to do behavior studies on domesticated animals when the subjects can see (and worse, interact with) the experimenter. Subtle and non-intentional cues by the experimenter can influence their behavior. I am surprised that this was allowed and was published; are standards so low in this field?
Remember “Clever Hans”:
https://en.wikipedia.org/wiki/Clever_Hans
Yes, I suppose they should have blanked out the experimenter except for a view of her hand, or maybe use a robotic hand. Good point. But there are videos that one could look at to see if there’s anything odd.
But the videos were probably aimed at the subjects, not the experimenters. The cues could be very subtle, and probably completely unintentional, as in the Clever Hans case.
I suggest you write a letter to the editor (and/or to the authors) raising this issue.
Good idea.
I agree – and parrots are very bright, clever enough to pick up cues and clues. And how were the parrots brought up? Casual observations on cockatoos we kept 50 years ago suggested great differences in behaviour between individuals reared in different ways.
That’s another huge point. People who own Gray Parrots usuaully are very social with them; macaws are often kept in aviaries with fewer human interactions.
A serious oversight, I should think. I just quickly browsed the article and I didn’t see any mention of how they controlled for this very well known effect. I wonder how that (no mention of the possible confounding variable) got past review.
Grey parrots are Democrats and the macaws are Republicans.
Yes, or the experimenters working with each.
For the record, I read and appreciated this science post!
Regarding the relative prosocial behavior characteristics of larger vs. smaller “communities” — I’m sure this has already been studied in that more familiar social animal, the human. Is there a difference in the behavior of residents of small villages vs. big cities?
Remember that whatever evolved behavior we show now evolved over millions of years when we lived in small groups. We’ve had big cities for less than ten thousand years, a tiny fraction of the time since our lineage began evolving as social hominins.
Fascinating!
I haven’t yet read the study or the NYT article, but I wonder, in the test case T trials were there any instances of the parrot receiving tokens sharing their nut(s) with the parrot that gave them the token?
Regarding the social group sizes correlating with the difference in results between the AGPs and the macaws, perhaps AGPs form smaller long term groups similar to the macaws and those smaller groups join together in shorter term much larger groups and group size doesn’t have anything to do with the differences in results.
I didn’t see any mention of nut-sharing in the paper, but if you’re curious you can see it from the links above. It’s just a small piece of nut, as far as I can see, and so would seem hard to share.
Although they measured pre-existing social relations among members of a tested pair (“affiliation”), and said it had no effect, it’s possible that the macaws, despite living in the enclosure with each other, did not regard one another as members of the same band. A species that lives in small bands may have a more restricted notion of who is ‘deserving’ of a token. A species that lives in large flocks of fluctuating membership may have a ‘low bar’ for admission into the cooperating group: gray parrots have a more “expanded circle”. The keepers at Loro Parque would probably be aware of the pre-existing social structures in the birds, and perhaps they shared that knowledge with the experimenters.
The best test, but of course much more difficult, would be to compare the results of pairs of macaws from the same wild group with pairs from different wild groups.
Also, it’s important to remember that the sample size for the underlying question asked here is the number of birds (14), not the number of repetitions of a test or number of behaviors observed. Of course, the same bird can be observed repeatedly, giving a better estimate of that bird’s behavioral inclinations, but it’s still just one bird’s inclinations. I haven’t gone through their statistics to be sure they did them right; I have no reason to think they didn’t.
Yes – I would share my sandwiches with anyone, provided they did not support Manchester Utd!
I have a theory. The particular type of nuts used in the experiment were not considered very desirable by Blue Headed Macaw society. Good enough to eat oneself but not considered polite to share with neighbors.
This reminds me a bit in the observations I have seen often with some other species on the water (Pelicans & Cormorant). The Pelicans work together as a group to feed although you do not see them sharing any food. The cormorant does his fishing alone and stays separated from others. However you do see cormorant hanging around the pelicans for some reason. However, if the cormorant comes up with a fish, the pelicans close by go after the cormorant in fast assault and will steal the fish if the cormorant does not quickly swallow the fish. The pelicans are very nasty in this way.
They are ornery birds. A word to the wise; when snorkeling do not try to sneak up under them and grab their feet. Hoo-boy, they do NOT like that.
And who could blame them? 😉
Yes, I don’t think you want to get too close to any pelican. And, how you you like it if they grabbed your feet…
I hear people don’t like it either when sharks do it.
Even if the shark is only joking.
This is a really cool experiment, and easy to get results. My suggestion is to try it on other intelligent species of parrots and corvids to try and pinpoint a situation in the wild that points more clearly as to why certain species share/don’t share. In a way, as others have pointed out, I can see the major flaw with the experiment is that the participants are not wild but rather pre-trained and tame, so it may be difficult to reveal any evolutionary influence.
That is “a thought-provoking potential explanation,” said Katherine Cronin.
She should be a teacher; she has a very kind way of saying “you are not quite correct.”
I started Googling questions related to this question out of curiosity. It seems that grey parrots are known, by their owners at least, for being particularly empathetic pets. One article mentioned that this is because they are prey animals (Google also tells me that blue-headed macaws have few to no predators past the nestling stage) who have to be especially sensitive to their environment. Whether there is any truth to this idea I don’t know. Perhaps a group of animals in danger of being preyed upon would be more sensitive to distress signals from other group members?
Macaws, including the Blue-headed Macaw, are attacked by eagles and hawks in Peru. I witnessed and photographed such an attack, and Jerry posted the photos here:
https://whyevolutionistrue.wordpress.com/2016/09/30/readers-wildlife-photos-and-a-video-3/
Interesting. If you google macaw and “known predators”, you’ll get multiple sites saying they have no known predators. I wonder where the disconnect is, or if this is one of those things that gets published in one place and then many others use that site as a reference. The top result is from the National Aquarium so maybe other sites are using that as a reference.
Oh wait, I see my mistake… that gives results for the hyacinth macaw, specifically, for whatever reason. Doh!
Hyacinth macaws are the biggest macaws, with the biggest beaks, so they might be too powerful for most eagles and hawks. I don’t know anything about that species in the wild.
Regardng sensitivity to danger signals, that’s not true; see my video of macaws all (apart from one straggler) reacting in unison to an alarm call:
Do macaws band together to react defensively as a group? (I believe grey parrots do, which made me wonder if this sort of ‘pack like’ behavior led to increased empathy. I believe you see increased theory of mind skills in animals that hunt in packs, so it seemed the same might be true of animals that do the opposite, by defending in groups.)
Roo, in the encounter with the eagle, all the macaws stuck together, which I iamgine is a defensive maneuver. But macaws always stick together. Where they are not trapped or killed by humans, they travel in pairs that hang out with other pairs in large flocks.
I’m skeptical of the “defending in groups” hypothesis. Pretty much all herding animals from wildebeest to schooling fish and murmuring starlings are doing so for protection. Little empathy would seem to be involved.
To both comments above… I wonder if a more coordinated defense would require more communication (grey parrots, or so I just read, become quiet as a flock and then burst out screeching unison, for example.) Or maybe not. Just a muse for the sake of musing.
Lou, thank you for the videos – what beautiful creatures!
I was thinking about why the 2 species differ in this behavior and came up with a decent theory. Wanted to write about it, but not sure you’d receive my email. I tried to contact you before, but no response. Maybe I have a bad email address, but not sure