Practical sampling procedures for panicle-infesting insect pests of sorghum

Within-field spatial patterns were investigated and optimal sample sizes were calculated for five panicle-feeding insect pests of sorghum including: sorghum midge (SM), Contarinia sorghicola (Coquillett), corn earworm (CEW), Heliothis zea (Boddie), fall armyworm (FAW), Spodoptera frugiperda (J. E. Smith), sorghum webworm (SWW), Celama sorghiella (Riley), and rice stink bug (RSB), Oebalus pugnax (F.). Spatial distributions of all species were significantly aggregated, based on tests for non-randomness and Iwao's and Taylor's regression coefficients. Regression analysis of sticky trap data revealed that SM numbers/trap were inversely related to the distance between the trap and the upwind field border, thus indicating the relationship between wind and SM dispersal. Frequency distributions of insect counts generally conformed to the negative binomial distribution; however calculation of common ;In situ visual estimates of SM densities on panicles were highly correlated with estimates provided by whole panicle samples (r;Mail survey questionnaires were used to assess current attitudes and practices regarding sorghum pest management and field scouting among farmers and pest management advisors in five Texas Lower Coastal Bend counties. Of those surveyed, 30-39% of farmers and 56-72% of pest management advisors did not generally follow state Extension Service guidelines for economic thresholds of SM, CEW, and RSB. The most frequently cited reason for non-use of economic thresholds among farmers was uncertainty about their ability to accurately sample these three insects, while pest control advisors most frequently cited distrust of published thresholds. The average time farmers and pest control advisors were willing to spend scouting a 16 ha sorghum field for insects was 20-34 minutes.;Sequential sampling plans were developed for each species studied. Truncation points for the plans were based on the maximum amount of time decision-makers were willing to spend sampling, rather than according to optimal precision formulae. A method, based on Monte Carlo procedures, for adjusting economic thresholds to minimize expected costs resulting from decision errors was proposed.