Signature Systems and Kin Recognition

Synopsis. This paper examines mechanisms of individual identification in the kin rec? ognition process. In type 1 kin recognition, the individual {e.g., a parent) learns the identifying cues (the "signature") of particular kin {e.g., offspring) at some stage when there are reliable contextual cues as to kinship {e.g., the parent finds the young in its nest). These kin are subsequently recognized via their signatures in situations in which contex? tual evidence of kinship is lacking. The primary selection pressure in this case is for a mechanism that can generate a large number of distinctive signatures, and thus permit recognition in large groups. I develop a model that predicts the necessary information capacity of a signature system, and I test this prediction for the signature call of the bank swallow {Riparia riparia). The measured information capacity of 17 bits corresponds well to the prediction and, moreover, is substantially greater than that of the homologous call of the similar but non-colonial rough-winged swallow {Stelgidopteryx ruficollis). In type 2 kin recognition, reliable contextual evidence as to kinship is absent at all times, and kin must be recognized via a signature-matching process. For example, the individual might compare its own signature to that of the unknown individual, inferring kinship if their signatures are sufficiently similar. I describe a multi-locus genetic signature mechanism that would permit discrimination of kin from non-kin, and suggest how this model might be tested.