Development and validation of sequential sampling plans for Sitophilus species associated with pet specialty stores.

Abstract

AbstractEight pet stores in Kansas, USA were sampled between February and August 2002, using traps baited with food and pheromone lures for capturing multiple species of beetle adults. Thirty traps were arranged in a grid pattern in each store and were checked every 2-3 weeks. The most common and abundant beetles captured in traps in all eight stores were the Sitophilus spp. (rice, granary and maize weevils). The rice weevil, Sitophilus oryzae, was the most common and predominant of the three weevils. Trap capture data from each store were used to calculate mean numbers of Sitophilus spp. trapped per week and associated variance and weekly presence or absence of adults. About 60% of the weekly trap capture data from the 8 stores were used for developing fixed precision and binomial sequential sampling plans and the other 40% of the data were used for testing the performance of these plans through computer simulations using the "Resampling for Validation of Sampling Plans" software. Green's fixed precision sampling plan was used for estimating Sitophilus spp. density of 0.62 insects trapped per week that corresponded with 50% of infested traps. The actual precision and sample sizes needed for estimating density at the fixed precision levels of 0.25, 0.35 and 0.50 were determined. Wald's sequential probability ratio test plan and a fixed sample size binomial plan were developed to classify infestation level with respect to an infestation threshold (50% of infested traps). Operating characteristic and average sample number curves generated using the validation data sets were used to gauge performance of the binomial plan. In addition, the actual errors in classifying infestation levels were also determined. The development, performance and utility of these sampling plans in retail stores are discussed.