Abstract

Sampling methods to estimate acridid density per surface area unit in grassland habitats were compared using presence-absence data and count data. Sampling plans based on 6 yr of surveys were devised to estimate the density of Chorthippus spp., Euchorthippus spp., and Calliptamus italicus L. These acridids represented > 90% of species in the study area. Sampling plans based on count data provided a reasonable tool when densities were > 1/m(2) and when the level of precision was 0.20-0.30. A binomial sampling plan can be used to estimate C italicus density with a level of precision >= 0.28. Sampling characteristics, i.e., estimated mean, actual precision, and sample size, were established on validation data sets with bootstrapping analysis. Sampling costs were also calculated according to density-dependent functions. Comparison between binomial sampling and enumerative sampling of C. italicus showed that binomial sampling required less time than enumerative sampling when densities were <= 2/m(2) and when fixed precision was > 0.35. Plot area bad no significant effect on sample variances of counts.