I develop the argument that for a true kin recognition system to evolve, selection must act on both parties: not only must recognition be favored in the donor of care, but reliable identification must be favored in the potential recipient of the care. This perspective suggests two complementary hypotheses, which I discuss and evaluate with data drawn from studies of birds. According to thesignature adaptation hypothesis, when the sender benefits by reliably identifying itself, selection will act directly on phenotypic characters so as to enhance their signature properties. I summarize our studies on parent-offspring recognition in four species of swallows which are consistent with this hypothesis. In particular, acoustical and perceptual analyses of chick calls show that the calls of colonial swallows are more individually distinctive than are the chick calls of noncolonial swallows. According to theantirecognition hypothesis, when the sender doesnot benefit by reliably identifying itself, selection will act so as to minimize signature characteristics. I suggest two contexts for research on this hypothesis. The first context occurs when parentage is uncertain due to extrapair copulations and/or egg-dumping, and the second context occurs when there is a long period between the onset of chick mobility and chick fledging, as occurs typically in gulls. In both instances, parents will be favored to recognize genetic relatedness of offspring but offspring will be favored to conceal it. To date, data from bird studies are consistent with the prediction that the interests of chicks win out in this situation.