Introduction
Many researchers are actively engaged in cross-species comparisons of animal intelligence. These researchers make inferences about relative intelligence by analyzing animals' overt learning behaviors in a particular experimental paradigm (e.g., numerical competence, concepts of similarity and difference, foraging and memory), as reviewed by Kamil (1988a) and Macphail (1987). Such inferences are adequate provided that the behavioral analyses are complete in terms of cause, function, evolution, and development (Burghardt, 1984; Kamil, 1984; Snowdon, 1983). Unfortunately, the study of behavior still is often partitioned between those who, as a group, study cause and development and those who study evolution and function. These two groups can be characterized, respectively, by the differing perspectives of comparative psychology and ethology (Marler & Terrace, 1984; Parker, P&G1). Comparative psychologists generally examine the cause and development of learning by comparing the mechanisms and processes in just a few animal species that are trained and tested under rigorously controlled laboratory conditions. Ethologists, in contrast, examine the function and evolution of learning by comparing natural behaviors in a variety of animals, from ants to elephants, often using closely related species to see if each species' performance depends on the particular aspects of the situation in which it is observed or tested (Kroodsma et al., 1984). Incomplete knowledge of each other's domains may cause both groups of researchers to perform experiments that are inadequate for the purpose of comparing intelligence.