Feeding Ecology of the Rattlesnake, Crotalus viridis oreganus, in Northern Idaho

Abstract

Prey remains were obtained from 104 Crotalus viridis oreganus from northern Idaho from 1978 to 1987. Rattlesnakes fed on eight species of mammals, one lizard, and one bird. Voles (Microtus montanus and M. longicaudus), deer mice (Peromyscus maniculatus), and cottontail rabbits (Sylvilagus nuttallii) comprised nearly 92% of the biomass ingested and 80% of the numbers of prey taken. Snakes less than one year of age fed exclusively on shrews (Sorex vagrans). The only ontogenetic shift in prey was related to an increase in mammalian prey size with snake body size. The regression of the proportion of snakes with prey to an index of small mammal abundance was not significant. Feeding occurred from mid-April to early October. Reproductive females fed in the spring until early August, and in the fall after parturition. Feeding during the reproductive year may explain how some female C. v. oreganus in our study area can reproduce in two consecutive years. The trophic relationships of populations of snakes are important aspects of their life history and ecology (Fitch, 1987; Mushinsky, 1987), because feeding is essential to growth and reproduction of any heterotrophic animal. Obtaining sufficient energy has been shown to be important in the reproductive ecology of several snake species (see Seigel and Ford, 1987). Detailed analyses, rather than lists of prey taken from dead snakes, are needed to better understand the foraging ecology of these predator animals (Mushinsky, 1987). The diet of the western rattlesnake, Crotalus viridis, has been well documented and includes mostly small mammals with occasional reptiles, birds, and amphibians (Fitch and Twining, 1946; Fitch, 1949; Klauber, 1956; Gannon and Secoy, 1984; Diller and Johnson, 1988; Macartney, 1989). Some populations of juvenile C. viridis rely primarily on lizards and occasionally amphibians for prey and then shift to mammalian prey as the snakes grow larger (Fitch and Twining, 1946; Fitch, 1949; Klauber, 1956; Mackessy, 1988). Females in most populations of C. viridis have a biennial or longer reproductive cycle (Rahn, 1942; Fitch, 1949; Klauber, 1956; Diller and Wallace, 1984; Macartney and Gregory, 1988) and pregnant female crotalines have a reduced food intake or do not feed (Fitch, 1960; Keenlyne, 1972; Keenlyne and Beer, 1973; Macartney and Gregory, 1988). The major objectives of this segment of our long term study of Crotalus viridis were to determine the food habits, degree of specialization, and ontogenetic shifts in the food of C. v. oreganus. These characteristics were considered in reference to an environment in which lizards are uncommon and some female rattlesnakes are able to reproduce annually, although most follow a biennial cycle (Diller and Wallace, 1984; and unpubl. data). STUDY AREA AND METHODS The study area was located in northern Idaho in the eastern portion of the Columbia Plateau Province. The climate is characterized as modified temperate continental, with most precipitation in the winter. Mean number of frost free days varies from 150-170 per year. Yearly precipitation varies from 38-60 cm. The mean maximum and minimum temperatures in January and July are 3 and -5 C, respectively and about 30 and 12 C, respectively. Since 1982, we have concentrated our studies at a hibernaculum complex on the upper breaks of the Clearwater River canyon at about 700 m in elevation. Two other species of snakes, the racer Coluber constrictor, and the gopher snake Pituophis melanoleucus, inhabit the area and utilize the same hibernaculum complex. The only lizard species we have found in the area is the western skink, Eumeces skiltonianus. Snakes were captured by hand, drift fences, and by driving roads. Drift fences (Diller and Johnson, 1988) typically were 30 m long barriers of metal 51 cm tall. Snake funnel traps constructed of hardware cloth were 1.2 x 0.6 x 0.3 m high and located at 15 m intervals along the drift fence. Traps were checked every three to seven days except during periods of cool, rainy weather, and never left unattended more than 13 days. Prey items were obtained from This content downloaded from 157.55.39.147 on Wed, 20 Apr 2016 05:24:54 UTC All use subject to http://about.jstor.org/terms FEEDING ECOLOGY OF CROTALUS VIRIDIS autopsied snakes (N = 509), road-killed individuals (N = 4), and hand-captures (N = 408). Prey remains were collected from hand-captured snakes by palpating a scat or holding snakes in the laboratory until a scat was produced. A hair key (modified from Moore et al., 1974) was used to identify most prey remains. Prey availability was estimated from drift fence captures from 1982 to 1987. Estimates of prey biomass were obtained from local specimens or the literature (Steenhof, 1983). Snout-vent length (SVL), body mass, sex, and life history category were recorded for each snake from which prey remains were taken. Life history categories used in this study were: young-first year of life; immature-second year of life until sexual maturity; mature male-a sexually mature male; reproductive female-a female that will give birth during the year of capture; nonreproductive female-a sexually mature female that will not give birth the year of capture; and postpartum female-a female captured in the fall that has recently given birth. Young and immatures were identified by known ages of marked snakes. Snakes were classified as mature based on SVL (greater than 52 cm for males and 55 cm for females, Diller and Wallace, 1984), by palpation, or by inspection of reproductive tracts following dissection. Reproductive and nonreproductive females were identified based on length-mass relationships (Diller and Wallace, unpubl. data), palpation, or dissection. Any snake that was captured in a drift fence trap with an obvious prey in the stomach was assumed to have fed in the trap and was not included in the analyses. The date of feeding was estimated by adjusting the date of capture depending on where in the snake the prey remains were found (in the stomach, small or large intestine) or by adjusting the date of finding a scat in the drift fence trap. Adjustments to estimate the date of feeding from the date of capture were obtained from current feeding trials with C. v. oreganus, feeding trials with C. v. lutosus (L. Diller, unpubl. data), and a summary of digestion data in the literature (Skoczylas, 1978). The following number of days were subtracted from the date of capture to estimate the date of feeding: prey in the stomach, 1 day; bolus in the small intestine, 3 days; scat palpated in the field, 7 days; bolus in the large intestine, 14 days; and scat in a drift fence trap, 10 days. We determined the relationship of snake size to prey size by calculating Kendall's tau, a nonparametric statistic, utilizing snake SVL and the maximum skull width of the mammalian prey. The X2 test of homogeneity was used to determine differences among prey availability (based on actual drift fence captures) and prey taken by the various life history categories of snakes. We assumed that actual drift fence captures that are biased toward animals with greater movements best approximate the availability of prey to a sit-and-wait predator. However, we tried adjusting actual drift fence captures based on relative movements of the major prey species. This analysis may yield more accurate estimates of density of prey, but we believe it produces erroneous conclusions concerning the availability of prey to a rattlesnake population. An index of small mammal density was calculated from the capture of small mammals in drift fence traps (numbers of mammals/30 m of drift fence/100 trap days). The proportion of snakes feeding each year was estimated from 1982-1987 by dividing the number of snakes with prey by the total number of snakes captured during the feeding season (approximately mid-May to very early October). Proportional data were treated with an arcsin squareroot transformation before calculating the relationship of mammal population density to frequency of feeding. The Shannon diversity index, H', was used to estimate food habits specialization (in biomass) of the life history categories of snakes (Schoener, 1968; Mushinsky and Hebrard, 1977). The formula used was H' = -2 pilogl0pi, where Pi = the proportion of the ith food item to the total prey recorded for each category. When using log,0, H' can vary from 0 (maximum specialization) to 1 (no specialization). We compared the diversity indices among the life history categories with a two-tailed Student's t-test and calculated the degrees of freedom following Zar (1984). We also calculated the Shannon index from the data on food habits of Crotalus viridis from southwestern Idaho (Diller and Johnson, 1988) and from central California (Fitch, 1949), and compared these indices to that of the present study.