Effect of Moving Traps between Trapping Stations upon Rodent Retrapping Data

An experiment was conducted in which Sherman traps, on successive nights, were either left in situ or moved 10 m to the next adjacent trap station. Rodents were significantly more likely to be recaptured at a new location if the trap there was the same one they had occupied the previous night, and previously occupied traps were also significantly more likely than clean traps to capture rodents other than the original occupant. These results indicate that estimates of home range and spatial exclusivity based upon retrapping methods are likely to be influenced by hitherto ignored methodological details of trap emplacement. INTRODUCTION The ecology and behavior of cryptic small mammals have been studied extensively by live trapping. Many authors have noted potential biases in these methods. Traps at different stations may not be equivalent, due to residual odors (Daly et al., 1980; Wuensch, 1982); different classes of animals may not be equally trappable (Boonstra and Krebs, 1978); trapping may influence the very ranging phenomena under study (Sheppe, 1967). Despite these problems, the method is so convenient that it remains prevalent. Retrapping of marked individuals has been used to study home range, spatial dispersion, juvenile dispersal and other aspects of behavior and demography. In most grid-trapping studies, traps are left in situ throughout a trapping period, but for one reason or another, the investigators in several studies have picked up traps nightly and then returned them to trapping stations without attending to trap identity (e.g., Brant, 1962; Wallen, 1982). The fact that residual odors affect trap response suggests that the identity of the occupant of a trap on one night may influence the probability of the same or other individuals entering that trap on the next night, and thus may bias our perceptions of home range, spatial separation and other parameters. Whether a particular trap remains at a particular trapping station or is moved among them may thus affect results. The experiment reported here was undertaken to compare data collected when traps were moved between stations with data collected when traps were stationary. METHODS The experiment was conducted in eight replicates. Each replicate involved 2 consecutive nights of trapping in a previously untrapped area within the ecological reserve of the Boyd Deep Canyon D -sert Research Center, Palm Desert, California. All trapping was conducted in Colorado descrt habi-tat in which creosote bush (Larrea tridentata) was the dominant perennial shrub. The eight replicates were begun on 22 December 1981; 21 April, 28 April, 5 May, 18 May, 25 May, 12 December and 20 December 1982. On Night 1 of each replicate, 80 thoroughly washed Sherman live traps (8 x 9 x 23 cm), each baited with ca. 2 g of rolled oats, were set out in 20 groups of four traps each. Within each foursome, one trap was set at each corner of a square with 10-m sides. Foursomes were set at least 40 m apart. Ten foursomes were designated "Experimental" and 10 "Control," at random. Traps were left from dusk until dawn, at which time all captured rodents were identified by species and sex, weighed, examined for reproductive condition, marked and released. Traps were left closed in situ throughout the day. On Night 2 of each replicate, trapping was conducted at the same stations. Control traps remained at their original positions. Experimental traps were each moved 10 m, by rotating each foursome anticlockwise to the next station. All traps were again baited with ca. 2 g of rolled oats and left from dusk until dawn. Each captured rodent was released after its species, sex, identifying mark if any, reproductive condition, body weight and capture site were recorded.