Kin Recognition Using Innate Labels: A Central Role for Piggybacking?

Abstract

Kin recognition enables incest avoidance and kin selection, and it is this last point that has probably fueled most interest in it among evolutionists. Kin recognition can rest on context (such as a shared nest-site), but there is good evidence that innate components are important, leading to much attention to the allele-frequency dynamics of such systems.

The earlier theoretical framework distinguished between green-beard alleles, recognition alleles, and phenotype matching. Useful definitions of these categories can be framed, but consideration of the molecular mechanisms involved erodes these as absolutely distinct. The most operational approach is to concentrate on phenotype matching. Kin-recognition involves distinction between classes of individuals: it does not necessarily entail individual recognition, nor does individual recognition necessarily facilitate kin recognition.

Social insect colonies, unlike those of marine invertebrates, are genetically heterogeneous and are made up of one or more family groups. This make-up has led to a concentration on the problem of distinguishing nestmates from non-nestmates when the nestmates often differ from oneself in the labels that they carry. Under current models an individual compares the arrays of labels carried by others with a template and rejects them as nest-mates if the number of matches between their labels and that of the template falls below a critical value. The template may have various origins, being derived from one or more referents (e.g., the individual itself, the queen(s), or the whole assemblage of nestmates). The models indicate that a few highly variable loci would enable the system to work. The amounts of genetic variation envisaged are large for most loci, but not for histocompatability loci. The possible systems vary in the stringency with which the matches are assessed with regard to the template.

The work on social insects has concentrated on determining which individuals are referents involved in template formation, and on the ontogeny of this formation. Significant variation between species is apparent.

Two models have been proposed for kin recognition in marine invertebrates, differing as to whether genetic identity is required for a match at a locus, or whether possession of at least one allele in common suffices. Because this recognition leads to agonistic interactions, it is likely that common genotypes have higher fitness than rare ones because they are involved in fewer agonistic interactions. Although both models lead to heterozygote advantage, each also leads to fixation of the initially commonest allele. The observed polymorphism is therefore not maintained by this kin recognition, but rather persists in spite of it. It seems that polymorphism at the label loci is maintained by selection for some other function; their recognition role is truly an “effect,” with the recognition function piggy-backing on one or more other systems. The immune system may be the most important such system.

Variation in the genetical sources of the labels is likely to be an important source of variation in kin-recognition systems. In higher vertebrates and in social insects, the likely use of many labels suggests the possibility of evolutionary switching between them in order to use those with the highest current levels of polymorphism. Unless the cost of adding further receptors is significant, however, the increasing incorporation of labels into the recognition system would result in the reduction of recognition selection on individual component loci. In higher vertebrates and in social insects, the likely use of many labels suggests the possibility of evolutionary switching between labels, or the steady incorporation of additional labels into the recognition system. Evolutionary switching, which allows use of those labels with the highest current levels of polymorphism, would be favored if the cost of adding additional receptors is significant. The addition of further labels, reducing the strength of recognition selection on individual loci, would be favored if it is cheap to add necessary receptors.