EN
Edge effect is an inherent problem when using trapping grids to estimate density of small mammals, resulting in a sampling area larger than the area of the grid. Distances between captures of individuals are used to estimate A(W), the effective sampling area of a trapping grid, but grid size sets a limit for the largest detectable distance. The spool-and-line technique is proposed here as a new method to estimate A(W). Movement distances based on the spool-and-line technique were compared to similar movement distances based on capture-recapture of three species of marsupials of the Atlantic Forest of Brazil. Distances based on the two methods were uncorrelated, and only ln-transformed distances based on the spool-and-line were normally distributed. The maximum distance moved (MaxD) estimated by the spool-and-line was chosen as the more accurate and practical distance to estimate edge effect. Estimates of the effective sampling area and densities for the common opossum Didelphis aurita (Wied-Neuwied, 1826), were compared using MaxD based on spool-and-line (MaxDspool), capture-recapture (MaxDcap), and also the distance between traps (DT). MaxDspool reflected more accurately density variation between seasons. Movement distances of small mammals based on the spool-and-line technique permit more accurate estimates of density and its dynamics.