We investigated the acute (5 d) effects of particle-associated azinphosmethyl (AZP) in multispecies microcosms and assessed the results in the context to data obtained from a parallel field study undertaken in the Lourens River, South Africa. A runoff simulation was carried out in stream microcosms containing the macroinvertebrate fauna of an uncontaminated Lourens River site exposed to particle-associated AZP (control and 200, 1,000, 5,000, 20,000 μg/kg; three replicates each) for 1 h. Measured AZP concentrations in filtered microcosm water resulted in the following values: Not detectable (control) and 0.03, 0.2, 1.1, and 6.9 μg/L, respectively. The two highest treatments resulted in significantly (analysis of variance [ANOVA]) reduced total numbers of individuals, while the number of taxa was affected in the 20,000 μg/kg treatment only. A comparison with previous data suggests that observed effects partly resulted from particle-associated AZP. Particularly affected were six out of 14 macroinvertebrate taxa such as mayfly and stonefly taxa. In parallel, the distribution of macroinvertebrates at a pesticide-free and a contaminated stretch of the Lourens River was monitored five times during the spraying season in 2001 and 2002. Out of the 14 core taxa found in the microcosm study as well as in the field approach, 10 showed comparable reactions in the microcosm experiment and in their field distribution; they were either classified as affected or unaffected in both studies. Thus, we conclude that particle-associated AZP has the potential to affect the invertebrate community structure of the Lourens River and that microcosm studies employing field-relevant exposure scenarios may be valuable for a local risk assessment of pesticide-related community disruptions in the Lourens River.