Abstract: One specific prediction emerging from trait-group models of natural selection is that when individuals possess traits that benefit other group members, natural selection will favor “cheating” (i.e. not possessing the group-beneficial trait) within groups. Cheating is selected within groups because it allows individuals to avoid bearing the relative costs typically associated with group-beneficial traits, but to still reap the benefits associated with the acts of other group members. Selection between groups favors traits that benefit other group members. The relative strength of within- and between-group selection then determines the equilibrium frequency of those who produce group-beneficial traits and those that do not. Here we demonstrate that individual-level selection, that is selection within groups can also produce an intermediate frequency of such group-beneficial traits by frequency-dependent selection. The models we develop are general in nature, but were inspired by the evolution of antibiotic resistance in bacteria. The theory developed here is distinct from prior work that relies on reciprocity or kinship per'se to achieve cooperation and altruism among group members.

Abstract: In vertebrates, the mating preferences of individual females can be flexible and the probability of a female mating with a particular male can be significantly increased by her having previously observed another conspecific female affiliate and mate with that same male. In theory, such mate-choice-copying behaviour has potentially important consequences for both the genetic and social (`cultural') transmission of female mating preferences. For copying to result in the `cultural inheritance' of mating preferences, individual females must not only copy the mate choice decisions of other females but they also should tend to repeat this type of behaviour (i.e. make similar mating decisions) subsequently and to generalize their socially induced preference for a particular male to other males that share his distinctive characteristics. Here, we show experimentally that individual female guppies, Poecilia reticulata, not only copy the observed mating preferences of other females for particular males, but that the preference now assumed via copying is subsequently repeated and generalized to other males of a similar colour phenotype. These results provide empirical evidence for social enhancement of female preference for particular phenotypic traits of chosen males rather than for the particular males possessing those traits, and thus have important implications for our understanding of the role of social learning in the evolution of female mating preferences and of male epigamic traits.

Abstract: Abstract. Considerable evidence across many taxa demonstrates that prior social experience affects the outcome of subsequent aggressive interactions. Although the 'loser effect', in which an individual losing one encounter is likely to lose the next, is relatively well understood, studies of the 'winner effect', in which winning one encounter increases the probability of winning the next, have produced mixed results. Earlier studies differ concerning whether a winner effect exists, and if it does, how long it lasts. The variation in results, however, may arise from different inter-contest intervals and procedures for selecting contestants employed across previous studies. These methodological differences are addressed through a series of experiments using randomly selected winners and three different inter-contest intervals in the pumpkinseed sunfish, Lepomis gibbosus. The results indicate that a winner effect does in fact exist in pumpkinseed sunfish, but that it only lasts between 15 and 60 min. Based on these results, predictions about the behavioural dynamics of hierarchy formation are discussed, and it is suggested that it may be impossible, in principle, to predict the outcome of dominance interactions between some individuals before they are actually assembled to form a group. Finally, the possible mechanisms underlying the winner effect are explored.

Abstract: For the purpose of distinguishing between mutualism and reciprocity in nature, recent work on the evolution of cooperation has both oversimplifed and undersimplified the distinction between these two categories of cooperation. This article addresses the resulting issues of model testability, clarifies the role of time and argues that the category of `pseudo-reciprocity' is an unnecessary complication.