egy is found in spiders, whose eyes and nervous systems are structurally quite different from those of vertebrates, suggests that this kind of visual organization reflects a more general ecological constraint on predators.
The jumping spiders also use their acute vision to regulate the actual pouncing in a way that converges with the procedures of cats. Spiders learn to see the distance to their prey in terms of whether it is a "jumpable" distance, in terms of a "body-scaled metric" ( Warren, 1988). The general affordance here is that predators who pounce must regulate their predation in terms of a body-scaled distance to their prey. If the prey is too far (which is to say, if it is out of reach of the predator's pounce), then the predator is not in adequate position. A predator that cannot reliably attain such an adequate position in a significant percentage of its predatory bouts will lose--not gain-- food energy by predation.
Once again, it is worth emphasizing that careful observation and experiment, even on invertebrates like jumping spiders, reveals not only flexibility but functionally specific adaptation of behavior to environmental variation. Learning to catch prey is very likely a behavioral pattern common to many predators, even fairly simple ones. Given the variability in size and capacity (not to mention other properties) of most animals, predator and prey alike, and given the variable nature of most habitats, it is difficult to see how evolution could bypass such flexibility.
Affordances serve as the foci of selection pressures affecting the populations of behavioral units of a given animal. These foci favor certain behavioral units and select against others, but only in the context of aggregations of units resulting in complete action sequences. This selection pressure occurs at all time scales, from the short-lived behavioral, through the ontogenetic, up to the phylogenetic time scale. Some affordances are so short-lived as to have little effect on the animal's subsequent behavior. However, other affordances persist throughout the behavioral time scale, and still others persist throughout the ontogenetic time scale or even longer. In these cases, there may be sufficiently persistent selection to create conditions that in an evolutionary sense favor certain action sequences over other sequences. If these selective regimes are combined with developmental processes within the organism that will tend to increase the reproduction of favored behavioral units or action sequences--and, as I show in chapter 5, there is good reason to believe that nervous systems embody a variety of such selective retention processes--then the favored activities may be properly said to be selected for.
The factors underlying the tendency for favored action units or sequences to be reproduced are multiple. Differences in the behavioral units across a population are in part due to differences in reaction norms with the species gene pool, and if the affordances selecting for that behavior are sufficiently persistent, then there may be competition in the population for those affordances and some corresponding shift in the gene pool (Note