14 J. Whitfield Gibbons and Raymond D. Semlitsch 



Table 5. The number of drift fence captures and recaptures of adult Ambystoma 

 talpoideum entering and exiting Rainbow Bay, South Carolina. The 

 percentages indicate the increasing effectiveness of the technique for this 

 species as a greater proportion of the population is collected. 



Sampling year 

 1978-79 1979-80 1980-81 



Total number entering (marked and unmarked) 459 2,133 450 



Total number exiting 



Already marked 193 836 264 



Unmarked 70 200 19 



Percent of already marked individuals 



of those entering 42% 39% 59% 



Percent of unmarked individuals exiting 



of those entering 15% 9% 4% 



Sampling error (based on assumption that 100% 



of specimens on inside of fence were marked) 27% 19% 7% 



SRP. However, the 14 specimens captured represent a major sample 

 compared to those previously reported from either South Carolina or 

 Georgia in the previous century (Golley 1962, 1966). The findings sug- 

 gest that the species is not necessarily rare or restricted in its geographic 

 range or habitat preference, but is merely difficult to capture by conven- 

 tional trapping methods. 



Another form of bias is that adults of certain species may not be 

 captured in pitfall traps, although the smaller juveniles may be suscepti- 

 ble and reveal an unexpected abundance. This phenomenon was wit- 

 nessed with the rainbow snake, Farancia erytrogramma, in which more 

 than 100 subadult animals were captured in an area where large adults 

 have never been seen (Gibbons et al. 1977). Such captures must be 

 interpreted cautiously, but their value for certain purposes is obvious. 



CONCLUSIONS 



Drift fences are capable of collecting large amounts of data on a 

 daily basis over long periods of time (> 5 years). For some species they 

 provide the only effective sampling technique, and for many it is highly 

 cost effective. However, variation in morphology, ecology, and behavior 

 of each species must be considered. If the limitations and biases of the 

 drift fence and pitfall trap technique are considered, population sizes, 

 seasonality, migration patterns, diversity, and distribution of many spe- 

 cies of animals can be effectively determined. 



ACKNOWLEDGMENTS.— We are grateful to the more than 400 

 individuals who have worked at or visited the Savannah River Ecology 



