Organic pesticide modification of species interactions in annual plant communities.

A method is proposed and tested for assessing multispecies responses to three pesticides (atrazine, 2,4-D and malathion). Pesticides were applied at two concentrations, on model plant communities grown in raised beds using soil containing a natural seed bank. Cover by species was monitored over time in nested 10 and 20 cm diameter neighbourhoods around Poa annua and Calandrinia ciliata target plants. All tested compounds modified relative species abundance, altered dominance and simplified the treated communities. Community biomass decreased with atrazine and 2,4-D treatments, but not with malathion. Each chemical altered species interactions for all treated communities, including the identities of interacting species and the timing of interactions. Each target species had its own suite of interacting species that individually changed with chemical treatment. When cover was used as a predictor of neighbour influence, analysis of species interactions using 10 cm neighbourhoods indicated more interactions than using 20 cm neighbourhoods. When biomass was used as the predictor, use of the 20 cm neighbourhoods indicated more interactions. This method of using model plant communities for field toxicity testing is simple, economical and effective. It uses naturally occurring plants while reducing the environmental heterogeneity common in most field studies.