SELECTIVE PREDATION OF THREESPINE STICKLEBACKS (GASTEROSTEUS ACULEATUS) BY GARTER SNAKES.

Abstract

Natural selection operates through the differential survival and reproduction of alternative phenotypes within a population. When a population is composed of two or more phenotypes that are specified by different genotypes, natural selection may operate upon these phenotypes. Natural selection may result from both differential fertility and survival, and both aspects of selection must be investigated to appreciate the relative impact of selective agents. The threespine stickleback, Gasterosteus aculeatus (Gasterosteidae), is renowned for intraand interpopulation variation. The most obviously varying morphological feature in freshwater threespine sticklebacks is the lateral plate series (Fig. 1) which varies in arrangement (morphs of Hagen and Gilbertson, 1972) and the number of plates. Other morphological features vary, but they are more difficult to measure and have received less attention. Hypotheses to explain lateral plate variation within G. aculeatus include NeoLamarckian environmental induction (Bertin, 1925), introgression and intergradation of genetically divergent subspecies (Miller and Hubbs, 1969) and adaptation to local conditions through natural selection (Hagen and McPhail, 1970; Hagen and Gilbertson, 1972; Moodie et al., 1973; and others; see Bell, 1976a). Recently, G. aculeatus on the Pacific Coast of North America has been the subject of numerous investigations to identify the mechanisms responsible for observed variability. These investigations have involved a variety of approaches: analysis of geographical variation (Miller and Hubbs, 1969; McPhail, 1969, 1977; Hagen and McPhail, 1970; Hagen and Gilbertson, 1972; Moodie and Reimchen, 1976), direct measurement of changing phenotypic frequencies over time (Hagen and Gilbertson, 1973a), examination of the effects of selective agents under natural conditions (Hagen, 1967; Semler, 1971; Moodie, 1972; Kynard, 1972; Hagen and Gilbertson, 1973a), laboratory studies of selective predation (McPhail, 1969; Moodie et al., 1973) and determination of the genetic basis of morphological features subject to selection (McPhail, 1969, 1977; Hagen, 1973; Hagen and Gilbertson, 1973b). These studies and others indicate that (1) variation of some morphological features has a strong genetic component, and (2) different classes of variants for these features have different probabilities of survival and reproduction under laboratory and natural conditions. These features thus are subject to natural selection. The effect of predatory fishes on morphological variation in threespine sticklebacks is comparatively well understood (Moodie et al., 1973). All of the studies indicating the importance of predatory fishes as a cause of morphological variability in threespine stickleback populations used specimens from Washington and British Columbia. However, populations primarily from southern California form the basis for claims that gene flow among divergent populations (i.e., introgression and intergradation) and not selection has caused variability (Miller and Hubbs, 1969). The virtual absence of samples with a modal frequency of seven lateral plates per side in southern California (Miller and Hubbs, 1969; Rutter, 1896), a mode characteristic of populations sub-