Optimum effort to estimate habitat use when the individual animal is the sampling unit

• 1 Studies of habitat use in which the individual animal is the sampling unit should ideally sample each individual sufficiently to achieve a stable estimate of its habitat use. Data are typically obtained by radio-tracking, which can be labour-intensive. Hence, optimization of sampling effort is desirable. A method to determine optimum sampling effort is described with reference to an example from a study of Natterer's bat Myotis nattereri, in which data were collected by radio-tracking with individual bats followed continuously for entire nights.
• 2 Habitat use by Natterer's bat was assessed by compositional analysis, which compares the composition of habitats used with those potentially available. Therefore, we plotted running per cent foraging time spent over a range of habitat types against the cumulative foraging time recorded. We visually estimated the optimum sum of foraging time required to determine stable estimates of the composition of habitat use from the plots. Then, by reference to the full-time budget, the total tracking effort expended at the point when this optimum amount of foraging data had been recorded was determined and expressed in units of the number of nightly tracking sessions that had been undertaken to collect that amount of foraging data.
• 3 Stable estimates of habitat use were attained after a mean of 22 ± 7.7 h of foraging time, which were obtained in a mean of 4.6 ± 1.9 nights of radio-tracking effort. Thus, in this Natterer's bat study, where habitat preference was assessed by compositional analysis, it was appropriate to aim to collect foraging data during five nights of radio-tracking for each bat sampled.
• 4 The method presented is also applicable to studies where tracking data are discontinuous. A variation of the method can be applied in studies where a Euclidean distance method is to be used for the analysis of habitat use.