A geostatistical description of the spatial distribution ofLimonius californicus(Mannerheim) wireworms is presented, as are unbiased estimates of spatial dependence for different sampling scales, patterns, and areas. The spatial structure was usually described by the spherical model, indicating an aggregated spatial arrangement. Localized discontinuity was>0.48 at scales of 5, 10, and 20 m and increased with scale. Spatial dependence occurred at a range of 66 m. Stratified hexagonal sample patterns generally gave more accurate summary statistics, parameter estimates, and r2values for theoretical spatial models than did random sampling. With few exceptions, summary statistics for the subsampling patterns reasonably estimated those for the complete 309-sample grid. However, perception of the spatial structure deteriorated with reduced samples. Frequency distributions estimated by Poisson and negative binomialkdistributions gave similar interpretations as the geostatistical analyses. However, geostatistical analysis is a true spatial analysis and yields information on spatial distribution, whereas traditional “spatial dispersion” methods only infer association from frequency distributions.