An analysis has been made of the variation in shell shape and shell characteristics of 889 Australian and New Zealand specimens of the genus Dicathais, using multivariate techniques. Shell measurements taken were: the overall length, length of spire, length of aperture, and width of aperture. Weight of the shell plus the preserved animal was also recorded. The sculpture of the shell, thickness of the lip, and the presence or absence of a reddish or purplish colouration or banding on the inside of the lip, were assessed qualitatively. Principal component analyses of the size measurements for each site showed that the first principal component, which accounted for greater than 95% of the variation at each site, was associated with variation in the ‘size’ of the animal. Canonical analysis of the size measurements showed a cline in shell shape from the animals on the western side of Australia to those on the eastern side of Australia and New Zealand. The resulting canonical variates were associated with variation in the ‘shape’ of the shell. Principal component analyses of the between-group matrix and of the within-group matrix of the size measurements showed that the site means exhibited a similar pattern of dispersion to that of the animals within each site. Canonical analysis of the shell characteristics showed that variation along the first canonical axis was largely produced by shell sculpture, while variation along the second resulted from differences in colouration/banding. The generalized variances of the correlation matrices for the size measurements showed that groups with similar shell shape were associated with the presence of granite substrata and/or mussel beds or, alternatively, with limestone substrata, but canonical correlation analysis of the relationship between the size measurements and shell characteristics showed that no consistent trend was evident over all sites. A subjective examination of the structure of the radula of 84 animals showed that two distinct morphological forms were present, but that they were not correlated either with sex or any of the named shell forms or site groupings. An analysis of the growth curves of 27 animals of the two forms from the eastern and western coasts of Australia, held in the laboratory, was carried out. The eastern coast form showed a loss of sculpturing and a change in shell shape when kept under west coast conditions and on a mussel diet. Water temperature, diet, substratum, and degree of exposure to wave action were all found to show associations with variations in either shell shape or shell characteristics. It is suggested that the selective force of the habitat which produces changes in shell shape and shell characteristics of the animals at any site is a complex of factors, many of which are interrelated. The genetic basis for the development of shell shape and production of the shell characteristics in Dicathais may be similar to that found in Nucella lapillus (L.) in the Northern Hemisphere. These data suggest that the Dicathais found at the sites studied in this investigation are all part of the same ‘population’, the shell shape and shell characteristics of the adult populations being determined both by selection and phenotypic expressions caused by the selective force of the habitat at each site. It is concluded that the genus consists of a single highly variable species. The value of the application of multivariate analyses to this type of study is shown to lie in the way in which the techniques provide an overall picture of the variation within sites and of the variation between sites.