The objective of this study was to assess the effects of forest applic
ation of diflubenzuron (Dimilin(R)), a gypsy moth suppressant, on nont
arget invertebrates in litter and soil on the forest floor. Four exper
imental watersheds (two treatment and two reference) in the Fernow Exp
erimental Forest, West Virginia, were sampled for two years before and
one year after application of diflubenzuron (70 g ha(-1), Al). Soil a
nd leaf litter arthropods were monitored before and after application
using soil cores and litter bag colonization samplers. Frequency analy
sis, BACIP, and ANOVA were used to analyse the data for diflubenzuron
treatment effects. During the 36 month study, soil core samples were d
ominated by mites (49%) and springtails (28%). A total of 19 taxonomic
groups were suitable for statistical analysis. We detected no signifi
cant treatment effects based on total organism counts or counts by tro
phic categories (p < 0.05). There were no significant treatment effect
s for populations of major taxonomic groups, except for Araneae (spide
rs). Analysis of leaf litter bags also showed no significant differenc
es in total numbers of invertebrates or in trophic categories between
treated and reference watersheds during the 12 month post-treatment st
udy. Density of one species of springtail (Sminthurus purpurescens) an
d springtails as a group were significantly lower in treated watershed
s. High variation was associated with both soil core and leaf bag samp
les; because of this variation, a significance level of 0.05 should be
considered very conservative. Long-term trends in soil biota were evi
dent, demonstrating the need for well-established pre-treatment baseli
ne data for pesticide impacts on soil organisms.