Two novel sampling techniques were used to survey the spider community of the leaf litter in a deciduous forest in Kentucky, USA. Using modified pitfall traps and litter-grab techniques, we sampled separately the top, middle and bottom litter layers from April–October. Our sampling program captured over 3,000 spiders encompassing 18 different families. Both techniques revealed that the web-spinning families were more abundant in the lower litter layers. In contrast, the non-web building cursorial spiders, which actively pursue their prey, were more abundant in the top litter layer. Cursorial spiders, on average, were larger than the web-building spiders found in the leaf litter. Web-building spiders from the top litter layer were also larger than the web-building spiders caught in the middle and the bottom litter layers.

Comparison between the two sampling techniques revealed that the spider community profile is greatly influenced by the sampling method employed. The stratified litter-grab technique revealed the numerical dominance of Dictynidae (38% of the spiders captured) and Linyphiidae (32%), families that are predominately minute web-building spiders. In contrast, the pitfall-trap technique suggested Lycosidae (24%), a family of active foragers, to be numerically dominant, with Dictynidae representing only 1% of the spiders captured. The results indicate that major groups of spiders differ in their vertical distribution within deciduous leaf litter, and that sampling method can dramatically affect inferences about spider community structure.