PHENOTYPIC PLASTICITY IN BODY SIZES AND SEXUAL SIZE DIMORPHISM IN EUROPEAN GRASS SNAKES.

Abstract

Many animal species show considerable geographic variation in average adult body sizes. Although this phenomenon offers a unique opportunity to tease apart the factors influencing body size, interpretation has proved to be difficult because several processes may contribute to the observed variation (e.g., Andrews, 1976; Case, 1978; Dunham et al., 1978). First, average body size is influenced by adult survival rates in species in which growth continues after maturity: animals may be smaller in one area than another simply because they are, on average, younger (e.g., King, 1989). Second, even if survivorships are similar among areas, so that the body-size differences are due to differences in growth patterns, there are two possible explanations for such variation in growth. Body sizes may differ either because of local genetic modifications (possibly due to adaptation) or because of a direct phenotypic effect of differing food availability on growth rates (e.g., Berry et al., 1987; Dobson and Murie, 1987; Ebenhard, 1 990). To distinguish among these alternative interpretations, we need two kinds of data on growth trajectories of animals: first from the field (to determine whether the populations differ in actual growth patterns, rather than simply survivorships) and second the response to experimental manipulation of food supply (to determine whether the observed differences in growth rates are due to genetic differences or phenotypic plasticity). This paper presents information on European grass snakes. We show that body sizes and the degree of sexual difference in body size are greatly reduced in an island population compared to the nearby mainland, that these differences are due to modified growth patterns and not just survivorship, and that the low growth rates and small asymptotic body sizes of the island snakes are a phenotypic response to local conditions (probably, low food availability). Insular populations of snakes offer some of the most dramatic examples of geographic variation in body size (e.g., Case, 1978; Schwaner, 1985; King, 1989; Shine, 1987; Schwaner and Sarre, 1988; Hasegawa and Moriguchi, 1989; Forsman, 1991). For example, Schwaner (1985) showed that body masses of adult Australian tigersnakes varied up to tenfold among adjacent islands. Correlational analyses suggest that predators attain larger sizes in areas where larger species of prey are available, and this may be true both for snakes (Schwaner, 1985; Hasegawa and Moriguchi, 1989) and for mammalian predators (Gittleman, 1985; Erlinge, 1987). Experimental studies on mammals have shown that geographic variation in body sizes may be due both to phenotypic plasticity (e.g., Dobson and Murie, 1987) and to local adaptation (e.g., Berry et al., 1987; Ebenhard, 1990). The only experimental study to address the determinants of such differences in snakes has been that of Barnett and Schwaner (1984), who raised juvenile tigersnakes. These authors documented rapid growth in captive snakes from a giant population, but obtained no comparable information on snakes from dwarf populations. We studied two populations of a nonvenomous natricine colubrid species, the grass snake (Natrix natrix), which is abundant over much of Europe (Arnold and Burton, 1978). The mainland study area was near Maryd, 15 km south of Lund in southern Sweden (55?40'N, 13?30'E). The area contains a mixture of arable land, grazed meadow, and mixed deciduous forest. Detailed data have already been published on body sizes, sexual size dimorphism, growth rates, diets and reproductive biology of grass snakes from this area (Madsen, 1983, 1987). Those papers also describe the methods used to capture, mark and measure snakes, and to obtain prey items by forced regurgitation. The same methods were used for the study of island snakes. Our island population was on Hallands Vadero (56?27'N, 12?44'E), a small (2.6 kM2) island approximately 3 km from the Swedish coast. One quarter of the island is forested, with the remainder consisting of meadows bordered by blackthorn, stony areas with juniper, and bare rock (Madsen and Stille, 1988). The two study areas are approximately 100 km apart. Geological evidence suggests that the two areas probably have been separated for several thousand years, since glaciation-induced reductions in sea level (Devoy, 1987). However, it is possible that the effective period of separation may have been longer than this (e.g., these oviparous snakes may not have existed in this region during glacial periods) or considerably briefer (due to fortuitous dispersal of snakes from one area to another). Male snakes from the island were similar to the mainland animals in mean SVL (Fig. 1; N = 22, 41; means = 52.2 versus 53.0 cm, t = 0.40, df= 61, P = 0.69) but females were much smaller on the island (N = 28, 44; means 59.7 versus 69.9 cm, t = 4.14, df= 70, P < 0.001). Indeed, some male snakes from the mainland actually attained larger sizes than did any of the females from the island population, although the mean values were lower (Fig. 1). Two-factor analysis