Context: Risk assessments currently undertaken to approve pesticides for use against locusts in Australia are predominated by laboratory-generated data, often collected from non-endemic species. Very little information exists concerning the potential ecological effects of insecticides used for locust control. Aim: To determine the effect of aerial ultra-low volume (ULV) applications of the organophosphorous insecticide, fenitrothion (used for locust control), on avian assemblages in arid and semiarid agro-ecosystems. Methods: Avian abundance was surveyed in pesticide-exposed and unexposed locations during standard locust control operations, using fenitrothion, in western New South Wales and Queensland in the spring and summer of 1996, 1997 and 2000, using a transect-point method. Temporal changes in assemblages were analysed using multi-dimensional ordination (MDS) and permutational analysis of variance (PERMANOVA) with PRIMER (V7). Key results: Avian assemblages varied over time, inde