HABITAT VARIATION WITHIN SYMPATRIC SNAKE POPULATIONS

Radiotelemetry was used to monitor the habitat use of 21 timber rattlesnakes (Crotalus horridus) and 20 northern copperheads (Agkistrodon contortix mokeson) during a 3-yr period beginning in 1979. Habitat was quantitatively defined on the basis of 14 structural environmental factors. To examine intraspecific habitat variation, species groups were subdivided by sex and the female subgroup was further separated into gravid and nongravid segments. The nongravid female A. c. mokeson subgroup was deleted from the analysis because of its small sample size. A random series of habitat samples was obtained from the study site for comparison with the remaining 5 subgroups. A multivariate analysis of variance of 6 groups with 14 variables indicated that group centroids differed significantly (P < 0.01). The first 3 discriminant functions accounted for 94% of the total variance and represented the following respective habitat gradients: rocky, open sites to heavily forested sites with few rocks; sites predominantly covered with leaf litter to sites with dense forest-floor vegetation; and sites with large-diameter overstory trees to sites with small-diameter trees. Gravid specimens of both species were most closely associated in discriminant space and clearly separated from other segments of their respective species groups. The utilization of rocky, sparsely forested sites and decreased variation in site selection were associated with gravidity. A 2-group discriminant-function analysis using melanistic and nonmelanistic male C. horridus indicated a gradient in habitat utilization from mature forest sites with numerous fallen logs to young forest sites with predominant leaf litter cover. Melanistic specimens preferred the former and nonmelanistic specimens the latter. The results suggested that A. c. mokeson was poorly adapted to the study site due to the preference of all population segments for open sites and young forest stands. C. horridus was better adapted, utilizing a broad range of the available habitat. The analyses serve to emphasize the importance of physiological factors in determining the spatial relationships of snakes in temperate environments.