Favored Places in the Selfish Herd: Trading OS Food and Security

Abstract

A key element in John Holland's approach to the study of complex adaptive systems has been the emergence of interesting macro-level phenomena from a number of simple rules at the micro level. Hamilton, an early member of the BACH brainstorming group that Holland thanks in his introduction to Emergence, used this approach to show that group formation could emerge naturally from individuals' "concerns" over predation in a one-dimensional environment. In this chapter, we use Holland's approach to extend Hamilton's study to include food competition, along with predation, in a two-dimensional world. In the spirit of Holland's approach, we combine empirical data, mathematical modeling, and computer simulation to understand more completely the behaviors that emerge at the group level from individual activities. Group living over extended time entails automatic and substantial costs: disease and parasite transmission, and constant competition for resources. It will, thus, evolve only when specific benefits outweigh the automatic costs [1, 4, 14, 16, 24, 26, 29, 35]. Yet many animals group together, under a variety of conditions. Determining the functionally important costs and benefits in any case is important, but often difficult. Hamilton's seminal paper "Geometry for the Selfish Herd" [4] used a functionally one-dimensional model in which extremely simple rules generated grouping: in a one-dimensional system with a random predator, a prey's domain of danger (the arc in which it was the individual closest to the predator) was reduced by having neighbors. Empirical work since has suggested a variety of particular ways in which heightened safety might arise; these are neither true alternatives, nor easy to distinguish empirically. Other benefits proposed have been: increased foraging efficiency under specified circumstances (i.e., patchily distributed food that is neither consumable nor economically defensible), and for fish, possibly efficiency of movement generated by use of vortices (see Parrish and Hamner [25] for a recent review). Here we examine groups of fish to focus our discussion of the costs and benefits of grouping in fish "congregations" [27]. Congregations typically (1) have distinct edges and relatively uniform densities, and (2) can show alignment and coordinated movement of individuals ("schooling" in fish).