Can ground-based assessments of forest biodiversity reflect the biological condition of canopy assemblages?

We identified 3933 spiders from 109 species from the 18 forest patches sampled. Both types of sampling defined differences in community composition between forest types in a similar manner; hence, either method could be used to evaluate differences or test management regimes in well-replicated experiments of forest type. However, the association in community composition between ground and canopy assemblages at the individual site-based level was weak; we found low correlation between the two data sets indicating that surrogacy between methods was not supported at this level. Furthermore, disparities in spider habitat association, body size, hunting guild and vertical stratification of spider families indicates that where detailed species and family-based information is required, or if inventorying is necessary, then multiple targeted surveys are essential.