Factors affecting pitfall capture rates of small ground vertebrates in arid South Australia. II. Optimum pitfall trapping effort

Pitfall trapping is widely used for sampling small terrestrial vertebrates. This study investigated the influence of trapping period, number of trap s ites and drift fences on the capture rates of small vertebrates in order to ascertain optimum pitfall trapping efficiency. Capture rates at fenced pit s were five times higher than at unfenced pits. Our study indicates that on e fenced site established in chenopod shrubland in the Olympic Dam region a nd trapped for four nights yielded less than one third of the local chenopo d shrubland reptile inventory. Capture rates of reptiles at Olympic Dam in arid South Australia were significantly higher on the first three nights of trapping than on the seven subsequent nights. The most efficient number of trap nights for maximising the number of reptile species captured was five nights. Trapping only a single pitfall site per habitat type for 5 nights captured approximately 30% of the reptile species in that habitat compared with 55%, 65% and 73% for three, five and ten sites respectively. Typically, approximately 70% of the reptile species in chenopod shrubland w ere captured by trapping 5 sites for 7 nights, a method recommended for tho rough surveys of species present in chenopod shrublands. Captures of mammal s were more variable than those of reptiles and, as a result, sampling of m ammal species benefited from more trapping nights than did sampling of rept iles. New mammal species continued to be recorded at a high rate for 8 nigh ts and even 10 sites trapped for 10 nights typically yielded only 60-80% of the species present in chenopod shrubland. Standard 4-night trapping perio ds adopted by most surveys will fail to detect many rare species, or those that are difficult to trap. Repeated 7-night trapping sessions for up to 21 nights are recommended for surveys where more accurate species inventories are required.