INTRODUCTIONThe sociobiology debates of the 1970s increased interest in the biology of behavior. At the same time, the growth of cladistics increased interest in how to do systematics and phylogenetic reconstruction. Yet, there are surprisingly few recent papers dealing explicitly with behavior from a phylogenetic perspective. Lack of communication between students of behavior and students of systematics is partly to blame. If one says to a behavioral ecologist, "Isn't it curious that there are white bears in the arctic?" he may say that there is nothing curious about it because they are white like all the other arctic mammals, and the fact that they are bears is irrelevant to the broad patterns of evolution. If one asks the same of a systematist he may reply that there is nothing curious about it because they are still bears like all the others, and the fact that they are white is irrelevant to the broad patterns of evolution. Both perspectives are partly right, and both are less than the whole story. Systematists tend to look for constraints of history, while behaviorists usually prefer to work with a warm ball of clay that lies ready to take on any shape the outside forces push upon it.Some of what follows is review and some is more philosophical, but the point of the paper is simple. Determining homology among behaviors is no different than determining homology among morphological structures. Behav ior is not special, it is only more difficult to characterize. Ethology (the study of behavior) is a relatively young science and does not yet have the benefit of centuries of debate and consensus, but that provides more reason for us to take up the challenge now. Ethology has made almost no advance with respect to a phylogenetic understanding of behavior since the late 1950s, and most Further ANNUAL REVIEWS 362 WENZEL modern ethologists simply do not work toward that goal. To honor the proud . heritage of Lorenz and Tinbergen we need only to be brave and begin.There is an immense literature dealing with "evolution of behavior," but only a tiny fraction of ethological efforts are relevant to the question of how one postulates homology among specific elements of an animal's behavioral repertoire. The majority of studies on behavioral evolution are related to theories of the process of evolution, and they therefore compare grades to illuminate the way in which analogous transitions occur in different groups (33). The focus of these studies is the transition itself, and homology of the steps is n01: an issue. Also the taxonomic literature is skewed toward finding species-spc:cific behaviors that allow identification more easily than morpho logical varii ation allows (1). Such unique traits do not assist in reconstructing phylogeny; only shared traits are useful for finding a nested hierarchy of order.I have tried! to include here both classical and more recent works that illustrate explicit postulates of homology, or cases where behavioral characters were critical for defining or supporting a phylogenetic scheme, but ...