Signal evolution in prey recognition systems.

In this paper a graphical model first developed in the context of kin recognition is adapted to the study of signalling in predator-prey systems. Antipredation strategies are envisioned as points along a signal-to-noise (S/N) axis, with concealing (low S/N) and conspicuous (high S/N) strategies being placed at opposite sides of this axis. Optimal prey recognition systems should find a trade-off between acceptance errors (going after a background cue as if it were a prey) and rejection errors (not going after a prey as if it were background noise). The model also predicts the types of cues the predator should use in opposite sides of the S/N axis.