CHARACTER VARIATION IN 2 LAND SNAILS FROM DUTCH LEEWARD ISLANDS - GEOGRAPHY ENVIRONMENT AND EVOLUTION

Cerion uva (L.) and Tudora megacheilos (Potiez and Michaud) are the dominant land snails of Aruba, Bonaire and Curaao; both display the high degree of intraspecific character variation so often noted in terrestrial gastropods. This study correlates patterns of morphologic diversity with geographic and environmental distribution. Projected upon factor axes, the samples of each species fall into four groups corresponding to ce:degrees of geographic isolation: Aruba, Bonaire, eastern and western Curaao. Factor axes represent the major trends of variation in shell size and shape. Subgrouping within geographic clusters is influenced by environmental parameters. In Cerion, for example, shells collected from volcanic rocks tend to be larger than those from limestone. In all geographic regions, these volcanic samples are distinguished by their higher projection on an axis representing increased size and concomitant changes of shape. Pulmonates usually favor limestone, but the anomaly of Cerion's smaller size is related to the unfavorable climatic conditions associated with limestone areas in the Dutch West Indies (dry, windy, and poorly vegetated). If the morphologic differences among geographic regions are similar to ecologicallyinduced variation within regions, then interregional differences are probably adaptive, and not a fortuitous result of geographic isolation, as many have claimed. Cerion is small and Tudora large in eastern Curaao; Tudora is small and Cerion large in western Curaao. This interregional relationship is parallelled by significant negative correlations between sizes of the two species for samples within regions. Diversity patterns did not change significantly during 16 years separating the studies of Baker and Hummelinck. Baker's data show the same distinction of four geographic areas and similar patterns of distribution within areas; significant positive correlation exists between measures made by Baker and Hummelinck on samples collected from the same localities. Fossil samples from cave phosphates do not conform to modern distributions. The basic character diversity pattern-good (presumably genetic) distinction among regions and ecologically induced variation within regions-requires a revision of Baker's classification. Separate names for the four regions are warranted, but the cumbersome nomenclature for intraregional diversity should be dropped. © Oxford University Press.