CONCEPTS OF SOCIAL STRUCTURE FROM A COMPARATIVE VIEWPOINT Comparative sociology challenges us to think in truly comparative terms about social structures as alien to humans (or to each other) as, for example, the societies of coelenterates and cooperative arachnids. Recent empirical work, particularly in pheromone chemistry, impels fresh theoretical analysis of this problem of general sociology (compare Merton 1967, pp. 162-65). First, a basic tool for comparing social structures is investigation of the means by which conspecifics single out recipients of aid. In the insect societies, recognition proceeds primarily by chemical (pheromone) signaling, and the last fifteen years of research have laid the foundation for a chemical theory of invertebrate sociality including chemical badges of colony membership and pheromone-based recruitment, attraction, foraging, and communal defense (see Shorey 1976, who cites 726 papers and monographs on these topics). Sophisticated as these mechanisms are in some respects, however, they do not enable the kind of individual recognition and formation of durable alliances we take for granted in all human and most other vertebrate societies. Insect social structures are not based on the ability of individuals to identify individual nestmates and to form individual cooperative bonds with certain of them. Because the precursors of modern social insects did not explore such evolutionary paths, social insect evolution has run in narrow channels, lacking the individual This content downloaded from 157.55.39.17 on Fri, 02 Sep 2016 05:29:57 UTC All use subject to http://about.jstor.org/terms EVOLUTIONARY BIOLOGY OF SOCIAL BEHAVIOR 229 components to branch out in the open-ended social structures characteristic of vertebrate cooperative hunting (which, as just suggested, may be a prototype for human social structures). An important corollary of these basic vertebrate/invertebrate differences in ability to recognize individuals concerns the heritability of the social structures that emerge. Since recognition of individuals is generally absent, social insect evolution is probably a product primarily of kin selection, with the chemical ability to identify kin serving as the substitute for vertebrate networks of reciprocal altruism. In many cases, kin selection implies a considerable continuity in the social structure across generations; witness the manner in which advanced insect societies have typically a long-lived queen who rears many generations of workers in the same established colony (see Wilson 1971, p. 428). By contrast, a cooperative hunting network such as that depicted in the I3crit model developed earlier in this paper may in principle be totally ahistorical, being independently regenerated in each new generation. Of course, most vertebrate social structures do not exhibit such extremes (e.g. see the well-known phenomenon of inherited dominance rank in some nonhuman primate societies, with inherited rank normally passing from mother to daughter); certainly kin selection has important applications here as well as to invertebrates. Nevertheless, both models and data suggest that an exceedingly influential factor in higher vertebrate social evolution has been the flexibility afforded by the potential for starting fresh in each generation, with all the possibilities for exploitation of genetic and environmental diversity (and ultimately for cultural evolution) thus afforded (cf Eisenstadt 1964 for related analyses in human macrosociology). An important illustration of this theme is the crucial role of "nomads" that is being discovered in various vertebrate societies, notably those of lions among carnivores (Schaller 1972) and of langurs among primates (Sugiyama 1967). Quite possibly the "takeoff' potential of vertebrate, as opposed to invertebrate, sociality is mainly a consequence of this partial absence of historical constraint, which allows the pieces to be put together in quite new ways from one generation to the