principal component analysis
Springer Online Journal Archives 1860-2000
Abstract The morphological variation of Littorina saxatilis and L. neglecta on a microscale (vertical transects down a shore) and on a geographical scale (Northeastern Atlantic) was examined to see if the sampled snails could be consistently separated into two groups on morphological criteria. Size, shell form and shell banding pattern, subopercular pattern and size at maturity were recorded for 21 samples from different shore levels (barnacle zone — usual habitat of L. neglecta, low littoral fringe and high littoral fringe — usual habitat of L. saxatilis) in Iceland, Norway, Sweden, Isle of Man and Wales. A multivariate approach (principal component analysis) was applied to analyse 14 quantitative shell characters. Between-shore variation in shell shape of samples from similar levels was generally larger than within-shore variation of samples from different levels, while size of snails was consistently smaller in barnacle zone compared to littoral fringe samples. The frequency of snails with qualitative characters, used in earlier studies to define L. neglecta, differed between samples from the littoral fringe and the barnacle zone at each site (except at the Swedish site), but the differences between geographical areas were generally larger. The results indicate that snails fitting earlier descriptions of L. neglecta are present in the barnacle zone in Iceland, Norway, Isle of Man and Wales, but that this form is not clearly distinguishable from L. saxatilis. Snails with intermediate shapes were common and, furthermore, the qualitative and the quantitative characters used to define L. neglecta were not closely coupled. No snails of L. neglecta form were found at the Swedish site, yet shape differences over the intertidal gradient were greatest at this site. The extensive morphological variation found in L. saxatilis, between shores and between microhabitats on the same shore, is likely to be due to low gene-flow within the species and a heterogeneous environment. The area over which mating can be random is probably small, thus a subpopulation will behave as a semi-isolate with more or less unique ecological and evolutionary factors influencing its detailed morphology.
Type of Medium: