Telemetric study of winter habitat selection by the American bullfrog, Rana Catesbeiana, in east-central Kansas.

Abstract

The overwintering behavior and ecology of American Bullfrogs are poorly understood. A Jolly-Seber capture/recapture population estimation was conducted during fall 2001 and summer 2002 to estimate winter mortality of American Bullfrogs at a pond located on the property of Wolf Creek Nuclear Operating Corporation in Coffey County, KS. American Bullfrogs (12 adults ~ 69g) were collected and a radio transmitter was implanted into each frog. Frogs were located once a week and ten habitat variables were recorded at each frog's location (frog points), at one point, < 1 m, 1-4 m, and> 4 m from the frog's location (non-frog points), and in areas where frogs were not located (pond points). Locations were recorded on aerial photos and Spearman rank correlation was conducted to determine if movements were correlated among frogs. Principal Components Analysis was conducted on the habitat variables. ANOVA revealed that the only significant differences between frog, non-frog, and pond points were on PC-1 and PC-2, with a Tukey's studentized range test grouping frog and non-frog points as a significantly different group than the pond points on PC-l and PC-2. The high, positive loading score for temperature and the high, negative loading score for dissolved oxygen on PC-l show that as temperature increased at sample sites dissolved oxygen decreased. The high, positive loading scores for depth and % gravel substrate show a positive correlation between these variables. American bullfrogs were selecting shallow areas of the pond with lower temperatures and higher dissolved oxygen. The estimate of winter mortality could not be estimated due to an absence of ranids at the pond during summer sampling. Movements were uncorrelated among frogs during the winter sampling and the purpose for long-range movements is unclear.