Conservation Genetics of the Desert Massasauga Rattlesnake (Sistrurus catenatus edwardsii)
Populations of the Desert Massasauga Rattlesnake (Sistrurus catenatus edwardsii) have declined rapidly as desert grassland communities have become reduced and fragmented. To provide information useful for management of remaining populations, the genetic characteristics (based on microsatellite DNA loci) of the last demonstrably extant population in the state of Arizona were compared to a population in the Rio Grande Valley of New Mexico. Results indicated that genetic diversity was relatively high in both populations, with statistically significant heterozygote deficiencies detected at only one of six loci in each population. Contingency tests, Wright's F-statistics, and Bayesian clustering algorithms all indicated substantial subdivision between populations in Arizona and New Mexico, but only contingency tests supported differentiation within the Arizona population. A preliminary hierarchical analysis of variance (incorporating both our data and published microsatellite data for the Eastern Massasauga) indicated that 73% of the total molecular variance was explained by variation within populations, with variation between the two subspecies accounting for 15% of the total variance. Results support the high conservation value of individual populations, as well as the need for further population genetic studies of the Desert Massasauga Rattlesnake.Abstract
Location of the San Bernardino (SB) and Belen (B) populations within the historic distribution of Sistrurus catenatus edwardsi (black). The historic distribution of S. c. tergeminus and S. c. catenatus are collectively illustrated in gray. Current distribution in Arizona is substantially reduced; SB is the last demonstrably extant population.
Frequency of captures along a 17.1 km transect (US Highway 80) through Sistrurus catenatus edwardsii habitat in Cochise County, Arizona.
Geary's c correlogram for capture frequency data partitioned into one-mile distance lags. Significant autocorrelation for an interval was inferred if the test statistic exceeded the expected value when no autocorrelation is present (c = 1), as determined by 999 permutations of the data. For Geary's c, values below one indicate positive autocorrelation and values above when indicate negative autocorrelation.
Contributor Notes
Associate Editor: J. M. Quattro.