Random or non-random structure?: What species co-occurrence, spatial and temporal variation reveal about the psocoptera community associated with eastern hemlock in the southern Appalachians

Psocoptera associated with eastern hemlock in the southern Appalachians have been demonstrated to be impacted by insecticide treatments used to control the hemlock woolly adelgid. However, it is unknown if the temporal and spatial co-occurrence of these species prior to disturbances from insecticides is random or non-random. As such, the spatial and temporal co-occurrence of 31 species of canopy-dwelling Psocoptera were examined. Six untreated eastern hemlocks were sampled in August 2006, September 2007, and June 2008. Samples collected represented three different strata within the canopy (top, middle, and bottom). Presence/absence matrices were constructed for each temporal and spatial sample from each host tree and two co-occurrence metrics were calculated, the C-score and the variance ratio. These metrics were compared with the respective indices calculated for 5,000 null matrices that were randomly assembled using fixed-equiprobable algorithms. Psocoptera assemblages on eastern hemlock were found to be non-randomly structured and were segregated within the host tree spatially throughout all sampling times. Five of the species were found only in the top strata of the tree, four species were found only in the middle, and 18 were found only in the bottom strata of the tree. In addition, three species were found in the bottom and middle strata, and one species was found in the top and middle strata. The resulting temporally consistent stratification and non-random pattern of Psocoptera co-occurrence illustrates the importance of incorporating temporal and spatial factors when examining the expression of community structure and when designing sampling methodologies for canopy studies.