Social Living - Cooperation; Costs and Benefits

  1. Cooperation among non-relatives. Kin selection (or structured- population) selection may explain the evolution of altruism among related animals, but we often see beneficial behaviors performed by one animals to another that is definitely NOT closely related. Can such behaviors be explained by natural selection?
    1. Reciprocal altruism. "You scratch my back, and I'll scratch yours". The donor of an act pays an immediate cost to benefit the recipient, but the recipient then returns the favor at some future time. A problem with this explanation is that NS could favor recipient that do not return the favor. If the two individuals alternate or work together to achieve a mutual benefit, this is called cooperation (not really altruistic, as both benefit). But even the possibility of achieving a mutual benefit does not guarantee that cooperation will evolve.
    2. The Prisoner's Dilemma. W.D. Hamilton (again!) and Axelrod (a social scientist) applied a famous 'game' called the Prisoner's Dilemma to the problem of cooperative behavior in animals. The PD game is based on this story: two suspected (in fact actual) criminals are caught by the police. The police have some independent evidence, enough to convict each of the criminals for R years. If both prisoners provide evidence against the other, they will both be convicted and go to jail for P years, with P > R. If this were all, clearly neither prisoner should talk. However, the police offer each of the criminals (separately, without the knowledge of the other) the chance to have a much shorter sentence (T < R) if he will provide evidence agaisnt the other, who will then get a much harsher sentence S (> P). This game has the following payoff matrix:

      3. Payoffs to A Player B cooperates Player B defects
      Player A cooperates R S
      Player A defects T P

    3. If the payoffs are now thought of as positive fitness increments instead of jail time, the game can be applied to animal cooperation (with T>R>P>S). The ESS of this game, if played only once, is Defect always (allD) for both animals. However, if the same players were to play many times, it is clear that they COULD reap much higher rewards if they did cooperate. Is there any behavioral 'rule' that allows cooperation to be the ESS when Defect is also present in the population? Axelrod and Hamilton staged a contest, played by computer, in which scientists sent in 'rules' which competed against other rules. The winner in that contest was a rule called Tit-for-Tat (TFT), in which a player always plays C on the first move, and then always copies what his partner did on the last move. Thus, in a match up against D, TFT loses only on the first move, and then always plays D, whereas with another TFT, TFT always cooperates. However, for TFT to be the ESS, the number of repetitions of the contest has to be quite large and the end should NOT be predictable. Furthermore, TFT cannot increase when rare in a population of allD (as there won't be any TFTs around to cooperate with, and TFT always does marginally worse than allD). There have been lots of recent changes to these general rules, including ingenious ways that TFT can increase when rare (e.g., by kin-selection or structured-population evolution). In any case, this theory provides the basis for understanding cooperative behaviors between animals that are not close relatives.
    4. We have already seen a payoff matrix of this structure in the discussion of parental care. Recall that I mentioned that genetic monogamy (i.e., no defecting) was characteristic only of species in which individuals mated for life - that is, the number of repeated breedings with the same individual was likely to be large, and the end would be unpredictable (usually only via the death of the partner). Similar cooperative behavior is seen among primates when the partners are likely to stay in the same group for long periods of time. This can happen even when the partners are highly competitive, such as adult males.
  1. Why are there so many kinds of social behaviors among animal species? There are two kinds of explanations – general and particular. General explanations help to explain why many species show SIMILAR social traits, such as a tendency to help relatives rather than non-relatives. Particular explanations can help to explain why one species DIFFERS in social behavior from other species, especially closely-related species. One class of particular explanations relies on species differences in ecology – the relationships to their physical and biotic environment. Among the most important of these ecological relationships is that of predator and prey, so the first things that you might want to know is what a given social species eats and by what it is eaten. Knowledge of these two aspects of trophic ecology can reveal much about the costs and benefits of being social.


  2. Foraging costs and benefits.
    1. The major general cost of being social is having to share food with members of the group. This cost is called within-group competition. The importance of this cost depends on how easy it is to obtain food. When food is locally superabundant, there is little cost, whereas when food is locally scarce, sharing it can reduce fitness markedly. However, it also depends on how easy it is AVOID sharing. If food is evenly distributed, then it is relatively easy to avoid sharing, because it is just as easy to find your own food as to join a (possibly reluctant or defensive) group member at their food source. In contrast, when food is clumped, then if you not in a clump, it may be hard to find an alternative food source. In this case, you are better off sharing someone else’s food clump than trying to find one of your own, and you may even be willing to fight to get access to such a food clump.
    2. There are possible foraging benefits of being social.
      1. faster discovery of food. With more eyes to find food, the group as a whole can detect food sources much more quickly than can a solitary animal. However, this benefit is unlikely to yield a net increase in feeding rate, after sharing of the food among the group members, unless the food sources found contain so much food that several individuals can feed without affecting each other’s intakes.
      2. better protection of food sources against competitors. If food is clumped and the clumps are scarce, groups of individuals may compete over access to food clumps. Such competition is called between-group competition. Nearly always, larger groups win in such encounters. If the food source contains enough food for long enough, the advantage to a large group of being able to defend the food clump against smaller competing groups may be large enough to offset the disadvantage of having to share the food among more individuals (in the larger group).

  3.    Predation costs and benefits.
    1. Costs. Larger groups are usually spread out across larger areas and this makes it easier for a predator to find and detect larger groups. In addition, a larger group may be a preferred target for a predator because of the greater overall chance that at least one group member is injured, weak, old, or otherwise easy to catch and kill.
    2. Benefits. A few weeks ago, we listed the various ways in which being social can reduce an individual’s predation rate, including the dilution effect, the selfish herd, the confusion effect, and increased group vigilance (early warning). We will not discuss them here (see previous lecture notes).

  4.   Putting costs and benefits together. Two main models.
    1. Within-group competition (cost for larger groups) vs. between-group competition (benefit for larger groups). In this model, first proposed by Richard Wrangham in 1980, individual foraging success would at first increase with group size as larger groups won in contests over favored food clumps. However, once group size was greater than the defended food clump could support, feeding rate would decline with group size. The best or optimal group size (for feeding) would be an intermediate number. This model does not take into account possible costs or benefits of grouping for reducing predation risk.

  5. Within-group food competition (cost for larger group) vs. predation reduction (benefit for larger group). In this theory, larger groups gain no feeding benefit at all, so that feeding rate always declines with increasing group size. However, survival in larger groups is better than in small groups, so that individuals end up with the best fitness in groups of intermediate size.
  1.   Testing the cost-benefit models – an example with brown capuchin monkeys (Prof. Janson's research).
    1. Measuring food intake. I followed every individual in four groups of capuchin monkeys ranging in size from 3 to 13 individuals (2 to 12 adults). I followed each individual for up to 6 hours at a time, recording systematically what they ate, how rapidly they ate it, and how long they ate it. From these data I could estimate the total energy intake of each individual. I found that an individual’s food intake per visit to a feeding tree declined rapidly with increasing group size, and there was no benefit of more rapid encounter of the group with food trees in larger groups, nor was the effect of between-group competition large enough to change the strong decline in food intake per food source.
    2. How much individual food intake declined in larger groups did depend, however, on the size (and fruit production) of the feeding tree. In smaller trees (less than 10m in diameter), individuals food intake declined strongly with group size. In large trees (more than 20m in diameter), food intake did not depend at all on group size – there was so much food available that food competition did not matter. Averaging out over all the trees used by these capuchins, however, individual food intake did decrease, with members of modal sized groups (of 10-12 animals) eating only one quarter of the fruit per visit to a tree of that a solitary animal would expect to eat.
    3. Does food competition help to explain primate group sizes? There are two ways of addressing this question. First, one can try to see how increasing food competition might limit the size of groups in one particular population. Second, one can derive an index of food competition for many different primate species and see if across species, group size and food competition are correlated. I performed both approaches.
      1. For the first, I looked at the consequences of reduced food intake in fruit trees on other aspects of capuchin behavior. I found that as food competition increased, individuals foraged for more hours per day to compensate for reduced food intake, leading to increased distances moved per day and less resting time. In fact, in the largest group I studied, individuals hardly rested at all, spending 11hours and 50 minutes of every 12 hours looking for food. It is reasonable to expect that if groups were to get much larger, there would simply not be enough hours in the day for individuals to compensate for increased food competition through increased foraging effort. Rather than accept reduced total food intake per day in groups of 14 or more individuals, group members would leave the group instead.
      2. For the second, across-species approach, I used the observation that foraging effort in capuchins increased with group size to compensate for reduced food intake = food competition. I used the AMOUNT of this increase in foraging effort (which I called Relative Ranging Cost or RRC) as an index of the severity of food competition for different primate species. Those that did not increase foraging effort with group size were assumed to have little or no food competition, while rapid increases in foraging effort with group size (high RRC) probably reflect high levels of competition. If food competition limits group size, species with high RRC show have small average group sizes and those with low RRC should have large group sizes. This is exactly what I found, but only for primates that eat fruit (why leaf-eating species do not seem to follow this rule is an interesting question that is still not fully understood).
    4. The preceding analysis still leaves unresolved what the benefits of group living are for capuchins. I could rule out the importance of between-group competition for food clumps (as this would have been reflected in the individual food intake values that I measured), so the only remaining plausible alternative was reduction of predation risk.
      1. However, although I saw a modest number of attacks by large eagles on my study groups of capuchins, I never saw one killed. This lack of direct evidence is not as serious as it might seem, when you consider that these are very long-lived animals (up to 45 years in captivity and well beyond 20 years in the wild), which means that death rates MUST be low. If less than 10% of the adults die per year, then one should expect only one death per modal group of 10 animals per year. If I am only with any one study group about 15% of its time, I have only a 1/7 chance of seeing a death happen. I did not get ‘lucky’.
      2. What other kinds of evidence would support the anti-predator argument? One line of evidence is WHERE individuals ‘hung out’ in the group. Recall from the selfish herd argument that the best place to avoid predators is the center of a group. If predation is an important influence on capuchin behavior, we should see two predictions fulfilled: 1) individuals that are especially vulnerable to predation should strongly prefer to stay in the center of the group, and 2) individuals should be more vigilant at the edge or periphery of the group (where predation risk is highest) than in the center. In fact, young capuchins (infants and young juveniles), which are easier to kill than adults, and slower to recognize and flee from a predator attack, strongly preferred to stay in the center of the group. The second prediction was also supported – individuals were in fact more watchful at the edge of the group than in the center, where they were safer. One would imagine that all individuals would want to be in or near the center of the group, but because of food competition, some adults (subordinate, aggressively unsuccessful ones) 'preferred' to forage near the edge of the group to avoid food competition, even at the expense of greater risk of predation. The only adult that could maintain a high food intake in the center of the group AND be safe from surprise attack by predators was the group’s dominant male. Thus, it was common to see the group’s infants near the dominant male, who actually tolerated them and even defended them against food competition by other adults.