The impact of commonly used organophosphate (chlorpyrifos, methamidophos), carbamate (carbaryl), and pyrethroid (cypermethrin) insecticides on insect natural enemies was compared with that of a nucleopolyhedrovirus (Baculoviridae) of Spodoptera frugiperda (J. E. Smith) (Lepidoptera Noctuidae) in maize grown in southern Mexico. Analyses of the SELECTV and Koppert Side Effects (IOBC) databases on the impact of synthetic insecticides on arthropod natural enemies were used to predict approximately 75-90% natural enemy mortality after application, whereas the bioinsecticide was predicted to have no effect. Three field trails were performed in mid- and late-whorl stage maize planted during the growing season in Chiapas State, Mexico. Synthetic insecticides were applied at product label recommended rates using a manual knapsack sprayer fitted with a cone nozzle. The biological pesticide was applied at a rate of 3 x 10(12) occlusion bodies (OBs)/ha using identical equipment. Pesticide impacts on arthropods on maize plants were quantified at intervals between 1 and 22 d postapplication. The biological insecticide based on S. frugiperda nucleopolyhedrovirus had no adverse effect on insect natural enemies or other nontarget insect populations. Applications of the carbamate, pyrethroid, and organophosphate insecticides all resulted in reduced abundance of insect natural enemies, but for a relatively short period (8-15 d). Pesticide applications made to late-whorl stage maize resulted in lesser reductions in natural enemy populations than applications made at the mid-whorl stage, probably because of a greater abundance of physical refuges and reduced spray penetration of late-whorl maize.

The efficacy of bioinsecticides that act by ingestion can be improved by the use of phagostimulant formulations that increase the consumption of the pathogen. A granule recipe comprising of pregelatinized flour, starch, ground maize cob, maize oil, and water was identified that was highly palatable to Spodoptera frugiperda larvae. To compare the performance of a multiple nucleopolyhedrovirus (SfMNPV) in aqueous spray and phagostimulant granules, multiple linear regression procedures were employed to analyze the results of seven independent field trials involving spray applications and eight independent field trials involving application of phagostimulant granules for control of S. frugiperda in maize. Prevalence of infection was positively correlated with the quantity of inoculum applied but negatively correlated with the interval between application and sampling of larvae for both granular and spray formulations. Virus-induced mortality was greater when inoculum was applied in granular than in spray formulation. A field experiment involving the application of 1:5 Â 10 12 occlusion bodies (OBs)/ha in quantities of 6, 18, and 32 kg granules/ha indicated that the quantity of granules applied did not significantly affect viral mortality at any timepoint. Persistence of activity of viral inoculum applied in spray or granular formulations was significantly greater when applied in the granular formulation, with approximately 23% of the original activity remaining at 8 days postapplication compared to <1% in the spray-treated plants. We conclude that the efficacy of SfMNPV was improved by formulation; the granular formulation caused a higher prevalence of infection in S. frugiperda larvae and persisted longer on crop foliage compared to an aqueous spray application. The cost of the formulation ingredients alone was approximately US$0.50/kg. When combined with the virus, the cost of the basic bioinsecticide components would be about US$17/ha, although this estimate does not include additional costs involved in the commercialization of a potential bioinsecticide product. Ó

The behaviour of two abundant predators in Mesoamerican maize crops, Chrysoperla rufilabris larvae and Doru taeniatum adults, towards healthy and nucleopolyhedrovirus-infected Spodoptera frugiperda larvae was compared. C. rufilabris did not discriminate between healthy and virus-infected prey, although the mean search time was approximately two times longer towards virus-infected larvae. In contrast, D. taeniatum directed a greater proportion of their attacks towards virus-infected prey but there was no significant difference in the search time. Prey consumption time did not differ significantly for each type of prey by either predator, although prey consumption was much faster in D. taeniatum. Viable virus was detected in D. taeniatum faeces up to 3 d after feeding on infected S. frugiperda larvae, whereas virus was inactivated in the gut of C. rufilabris. Both predators were shown to have acidic guts. A field experiment demonstrated that D. taeniatum that had fed on infected prey could contaminate foliage resulting in the transmission of the disease at a low prevalence (4.7%) to S. frugiperda larvae in a field maize crop