In evolutionary biology, we often draw a line between “altruism” and other cooperative traits. Altruistic traits are special in that they lead to a net cost to one’s survival and reproduction. Some traits are clear cases: when a bee stings you it dies, so the suicidal bee sting is an altruistic trait.
But nature abhors clear categories. Most seemingly “altruistic” behaviors are ambiguous in whether they pose net average fitness benefits or costs. This leads to some confusion for several reasons. First, most acts that most people call altruistic are not altruism in the evolutionary sense. Any form of “reciprocal altruism” is not really altruism as defined above. This is why people who study animal cognition invented the term “prosocial” behavior to avoid the word “altruism” and all the inevitable semantic arguments with evolutionary biologists. Second, the kinds of costs we can easily measure (like time and energy) are not the same as fitness costs which can only be measured after whole lifetimes have gone by, so this means we can’t categorize most of the cooperative behaviors that we are studying. Third, many people equate helping kin with “altruism” but much helping between kin might actually be mutually beneficial. For example, natural selection might have shaped me to care about the survival of my family, not just because that helps them survive, but also because it helps my own survival.
For traits like this that pose both costs and benefits to the helper (food sharing in vampire bats is an example), it’s better to think of there being a spectrum where the exact cost/benefit ratios of a cooperative trait can slide around from positive to negative depending on the circumstances. It’s not a completely different behavior just because you move from -0.1 to +0.1 direct fitness effects.
Thankfully, to better understand a cooperative trait, we don’t always need to try to unambiguously classify traits or exactly measure the change in lifetime reproductive success that comes from performing the behavior. Instead, we can just change the factor we think is important and see if the animal’s helping decisions also changes as one would expect from theory. Rather than measuring lifetime fitness consequences, we can test the design of the trait. In this case, what information is involved the decision-making process?
For example, many animal parents will go to extreme lengths to protect their babies (e.g. below is footage of a mother moose attacking a truck) and various theories (inclusive fitness and parent-offspring conflict) makes predictions about the design of this behavior.
A mother’s brain should be designed by natural selection to put herself at some degree of risk to save her offspring, but not too much risk. Unlike the sterile worker bee, she is not a genetic dead-end, so she should not carelessly cast away her own life for any potential benefit to her genetic kin. Theory predicts that at some risk to her own survival and reproduction, she should give up on trying to save her offspring. There’s a risk factor that can be tuned up and down that should have an effect on the probability of A helping B, and this factor should interact with genetic relatedness. One could also tune up the kinship factor. As the famous quote by Haldane goes: I would give up my life to save 2 brothers or 8 cousins.
Risk should decrease my willingness to help and increase the degree to which I care about someone’s relatedness to me. As you dial down the risk, I am more willing to help: I would not run into a burning building, suffering certain third degree burns, to help a total stranger, but I would do it if I had protective gear. As you dial up the risk to me, the circle of people I would be willing to help in that situation should shrink: I would jump in the ocean to save a stranger, but I would only jump in shark-infested waters to save my child.
Like other animals, we don’t weigh the costs and benefits consciously, but the emotional urgency we feel to help or not help depends on situation-based cues that have, in our evolutionary past, acted as reliable indicators of inclusive fitness benefits of helping in situations similar to the one we are facing.
So how can we test this in food-sharing vampire bats?
A few years ago, I was trying to record contact calls from hungry vampire bats. So put a caged hungry bat in a larger flight room with the other bats and put a microphone on it. I discovered to my surprise that other bats would feed these trapped individuals through the cage bars. At some point, this gave me the idea to do a small side experiment to test the idea that idea that risk increases nepotism.
To manipulate the perceived risks of helping, I created a novel “rescue” condition, where any donor vampire bat had to leave her warm, safe, dark, and comfy roosting location alongside her groupmates, then descend to an illuminated spot (vampire bats are very light phobic) where the trapped bat was stuck, and then to feed the trapped bat, she has to press her face to the cage bars (which sometimes makes her surroundings invisible) and regurgitate across cage bars. Compared to normal food-sharing, they don’t seem to like doing this. But they do it.
Sixteen of 29 bats were fed by others when trapped. They were fed by both kin and nonkin, but the degree of nonkin sharing declined quite obviously. All 15 starved bats that were tested in both trapped and free conditions received less food when trapped, and they received a consistently greater proportion of this food from closer relatives when trapped than when free. The vampires were more willing to feed mothers, daughters, and sons in the rescue condition. This is what we should expect if the bats’ nepotistic biases are exaggerated under dangerous conditions.
This paper was published in the journal Behavioral Ecology. Or if you lack institutional access you can get it from me here.