Geographic Variation of Multiple Paternity in the Common Garter Snake (Thamnophis sirtalis)
The common garter snake, Thamnophis sirtalis, is the most widely distributed reptile species in North America. Although multiple paternity has been documented in this species, variation in reproduction and ecology suggests that the frequency of occurrence of multiple paternity may vary. We investigated the occurrence of multiple paternity in snakes on Vancouver Island with the following aims: (1) to detect the occurrence of multiple paternity at this location; (2) to determine whether life-history variation and single versus multiple paternity were associated; and (3) to determine whether local rates of multiple paternity differ in comparison to a previous study of this species. Sixteen females and their offspring were analyzed using three highly polymorphic microsatellite loci. Only six of 16 litters showed direct evidence of multiple paternity. Results also showed evidence of a trade-off between offspring size and number of offspring per litter and that females that were multiply mated generally made a higher reproductive investment than females that were singly mated. Rates of multiple paternity in this study and in the previous study differed and were significantly different when litters of fewer than five offspring were eliminated from the analysis. Although we cannot determine the causes of variation in multiple paternity given our data, we suggest two possible mechanisms, one genetic and one ecological, that may lead to different frequencies of multiple fertilizations in this species.Abstract

Hypothesized causal relationships between life-history variables in the form of a “path” diagram, with direction of arrows indicating presumed direction of influence (“cause”). Relationships between SVL of mother, litter size, and average size of neonates as posited in Gregory and Skebo (1998), with presumed influence of proportion of live young added. Dashed lines show factors that might influence attractivity of females to males and hence the number of potential fathers of a litter

Plot of “adjusted” litter size versus average SVL of littermates. This graph shows the semipartial relationship between neonate SVL and litter size from the multiple linear regression equation, neonate SVL = constant + litter size + neonate mass + proportion live young, as dictated by the causal relationships in Figure 1. Litter size was adjusted by linearly regressing it on the other independent variables (neonate mass and proportion live young) in the original regression equation, thereby holding them constant statistically, and then expressing it as a residual from that relationship (for method of calculating semipartial relationships, see Nie et al., 1975). M = multiply sired litter, S = singly sired litter