| Resumo: | As part of one study of the population dynamics and habitat use of the white-tailed deer at the tropical deciduous forest of "Chamela" at Jalisco, we analyzed the uselfulness of tallying animals in transects as a census method. We compared the density estimates by nine models, between July 1989 and June 1991. Estimates varied beween 5 to 22 deer/km2, depending on the model and season. The negative-binomial, exponer>tial- polynomial, and Cox2 modeis, gave the highest densities; the exponential-series, strip transect with 20 m-width, and Coxl, the lowest; and the Fourier-series, half-normal, and strip transect with 1Om-width models, the intermediate. Frequency histograms of deer observations varied between seasons and between years, suggesting that the non parametric model based on the Fourier series might be the most appropriate for estimating deer density in this type of tropical habitat. Therefore, the mean density for the period of study was 11 ± 1 deer/km2. The main advantage of the direct count for this type of habitat is that, unlike the indirect methods, the transect line method does not require specific assumptions related to aspects of deer biology for a particular area. The main difficulties of this method in the study area were: low number of deer detectad, low visibility making difficult to detect exactly where the animal was befora it fled, and difficulty to set the transects rendomly in this tropical habitat.
|
|---|
