An analysis of multiple captures in Permoyscus with a critique on methodology
Multiple captures obtained in Sherman traps are reported for Peromyscus maniculatus (66 intraspecific, 6 interspecific) and P. difficilis (12 intraspecific, 4 interspecific) in Colorado. Multiple capture frequency in P. maniculatus. (3.53%) was higher than in P. difficilis (1.94%), which probably resulted from the greater weights of P. difficilis. Only one case of injury or death was noted in 84 multiple captures. Multiple captures occurred in traps in which the treadles were significantly less sensitive to tripping than traps that captured single mice, which suggested that pairs of mice did not enter the traps simultaneoulsy. Adult mice were disproportionatley sampled in single captures because of the insensitivity of the tripping mechanism. For the same reason, immature mice were disproportionately sampled in multiple captures. These trap-related biases can create spurious age and sex associations within multiple captures. Trap bias should be tested explicitly before attempting to analyze multiple-capture data from single-capture traps. A "socially null" model of random encounter of traps by captured pairs fits many of the data presented here and is suggested as the appropriate null hypothesis for these analyses. If live-trapping data must be used to address questions of social behavior, the use of actual multiple-capture traps, preferably to study long-term associations between individuals, is suggested.