An image analysis system for measuring insect feeding effects caused by biopesticides

A video-imaging system was developed to quantify and analyse, in real time, the feeding behaviour of insect larvae exposed to peroral biopesticides. Image processing algorithms were developed to separate the leaf image from the background and to estimate the consumed leaf area in each image. The feeding rhythm was then characterized by: average rate of leaf defoliation, meal size, inter-meal time, number of meals and rate of leaf defoliation per meal. The feeding rhythm was measured for mature larvae of Spodoptera littoralis (Lepidoptera), which were pre-fed with the entomopathogenic nematode Steinernema riobrave in an alginate gel carrier. The leaf consumption h(-1) in the nematode-treated larvae was lower than in the control treatment and the number of meals h(-1) was smaller in the nematode treatment than in the control. This pre-mortal effect of the nematodes on larval feeding suggests that plant damage could be reduced significantly during the time between ingestion of the nematodes and mortality. The same system was used to test the effect of feeding Bacillus thuringiensis (Bt) transgenic cotton to Helicoverpa armigera larvae; it reduced leaf consumption h(-1) and increased the inter-meal time. This feeding profile suggests that the Bt toxic protein in the plant suppressed feeding, because the longer resting times than those of the controls were probably needed by the larva to recover from the Bt toxicity. The imaging system could easily be adapted and applied to feeding behaviour analysis in diverse insect pathology and entomology studies. (C) 2000 Silsoe Research Institute.