A test of the scent-station survey technique for bobcats

In autumn 1988 and 1989 the authors reintroduced Felis rufus (n = 31) to Cumberland Island, Georgia, and conducted 15 scent-station surveys during. September-February 1988, 1989, and 1990 to obtain scent-station indices (SSI) as bobcat density was increased. There was a positive relationship (r² = 0.45) between population size and SSI, but because SSI variance also was correlated positively with SSI, the authors transformed data to meet the assumption of homoscedasticity for the regression model (r² = 0.73). Predictions of population size using individual scent-station surveys had poor precision. Four replicate scent-station surveys had an 80% probability of detecting only large (≥25%) changes in populations of high density (0.5 bobcats/km²). It is recommended that: 1) multiple scent-station surveys be conducted each year to monitor changes in bobcat populations; 2) SSI values should be calculated as proportions and transformed to reduce heteroscedasticity; 3) each stratum in a sampling design should contain as many stations as possible to minimize the problem of discrete data (number of visits) analyzed as a continuous variable (proportion of stations visited); 4) scent stations should be placed as far apart as logistically feasible to minimize multiple visits by individual bobcats; and 5) results of the power analysis should be used as a minimum guideline for estimating sample-size requirements. -from Authors