We examined the ability of a large, purpose-built rhizotron (undergrou
nd observation chamber), to examine the effects of two pesticides on t
he interactions between soil invertebrates, root growth and above-grou
nd plant production within a grassland ecosystem. The pesticides, Durs
ban 5Gn(TM) and Tornado(TM) (Carbaryl), killed most of their target sp
ecies (soil arthropods and annelids, respectively), but some taxa thri
ved in pesticide-treated plots, probably due to reduced inter-specific
competition, decreased predation and/or increased food supply. In art
hropod-impoverished (Dursban-treated) plots, the total length of visib
le root in a selected portion of the soil profile was reduced by <40%
compared to that in control plots and the rate of root turnover was al
so reduced. Associated with this was a >30% increase in above-ground f
oliar production (by mass). In earthworm-impoverished (Carbaryl-treate
d) plots, the total length of visible root was not significantly diffe
rent from that in the control plots and the rate of above-ground folia
r production was also comparable. The usefulness of rhizotrons in exam
ining the effects of pesticides on the interactions between soil inver
tebrates and plant roots is discussed.