Polymorphism and selection inCochlicella acuta

Abstract

Cochlicella acutais a small, pointed sand-dune snail, which is polymorphic for shell banding, shell opacity and ground colour. From a survey of about eighty colonies between latitudes 58° N and 43° N in west and northwest Europe and a more extensive study of one of them, morph frequencies have been determined and corresponding habitats classified, according to their appearance. It has been shown previously that the principal alleles responsible for the shell characteristics ofC. acutabehave as a supergene, that the known variations occur in certain preferred associations and that only a few morphs occur commonly. Now, the frequencies of certain morphs are shown to be associated with particular habitat types. Three such relationships are identified. Discontinuously opaque unbanded forms, which usually possess a buff/ amber ground colour and are consequently of sandy appearance, predominate on backgrounds of exposed sand, where they are cryptic. Heavily-banded forms, which are dark in colour occur most commonly on dark coloured substrates while continuously opaque morphs with few bands which are of black and white appearance, occur principally and numerously in turfy habitats, where in contrast with the cryptic relationship of the other two associations, they are quite conspicuous. Moreheavily-banded forms of the sympatric speciesHelicellahave also been found to occur most commonly where the substrate is dark. Predation by rooks, Corvusfrugilegus, of both species is established. It is suggested that predation of this kind is likely to be selective and that, largely though not entirely, it determines the cryptic habitat/morph frequency relationships, such visual selection in turfy habitats being less influential. By using Fisher’s maximum likelihood method, the frequencies of the supergenes and genes that are responsible for the principal features of the shell polymorphism ofC. acutaare computed from the sample data already recorded. An analysis of these frequencies in relation to habitat and to latitude shows certain supergenes and genes frequencies to be related to habitat, irrespective of latitude; others to latitude irrespective of habitat type. A third category is independent of both. These relationships are attributed to selection of at least three kinds, identified respectively by being related to habitat type, to latitude and to neither. Visual predation, which consistently favours an absence of banding in sandy habitats, irrespective of latitude is probably the major influence of selection related to habitat. Latitude-related selection, which acts very strongly against the supergeneCO 00in all types of habitat in northern latitudes, though progressively less in southern ones, is probably due to certain aspects of climate. These two kinds of selection appear to underlie most, but not all, of the supergene and gene frequency distributions within the range investigated. The exceptions are those determining the banded forms, which are discontinuously opaque. Gene interaction of sufficient strength to surpass the effects of other selective factors is suggested.