Differencial efficiency of two sampling methods in capturing non-volant small mammals in an area in eastern Amazonia

ABSTRACT This study was the first to evaluate the efficiency of trapping methods in the study of small mammals in the Carajás National Forest, southeastern Brazilian Amazon. It is an area with a unique vegetation type (metalofilic savannah or Canga). The aims of this study were to compare the efficiency of two trapping methods (i.e. live-traps and pitfalls), the bait types used, and evaluate if trapping success varied seasonally. We used four sampling grids, each with six parallel transects. The trap effort for live-traps and pitfalls was 51,840 trap*nights and 10,800 bucket*nights, respectively. We used three types of bait: a paste of peanut butter and sardines, bacon, and bananas. We placed one type of bait in each trap, alternating between points. We recorded 26 species of small mammals, 11 from the order Didelphimorphia and 15 from the order Rodentia. Pitfalls captured a higher number of species compared with live-traps. The capture rate, the mortality rate and the quantity of juveniles and adults did not differ significantly between methods. Capture rate for pitfalls differed significantly between seasons. The majority of species were captured by a single method. Species were equally attracted to the traps regardless of the type of bait used. Some of our results differed significantly from other studies in Amazonia and such variation should be taken into account when designing survey methods for Amazonian small mammals.