CLASSIFYING RESISTANCE SEVERITY IN FIELD POPULATIONS - SAMPLING INSPECTION PLANS FOR AN INSECTICIDE RESISTANCE MONITORING PROGRAM

Detecting insecticide resistance before it increase above a critical level has been a driving force in developing tests to estimate resistance severity in field populations. Here, estimates of resistance severity (r) are classified into one of three categories. The classification is based on comparing r with two preselected levels of resistance severity (theta) to obtain a three decision sampling plan. Using the example of fenvalerate resistance in beet armyworm, Spodoptera exigua (Hubner), each theta was selected by comparing resistance test responses of populations that differed in resistance and exposure to insecticides. Test response was corrected for survival of moths after exposure to a 1,00-mu-g/g concentration of fenvalerate. This concentration was selected by comparing probit lines of susceptible and fenvalerate-selected strains. The determination that r was greater or less than each theta was made using sampling inspection plans based on the binomial distribution. A sequential sampling plan using the sequential probability ratio test (SPRT) required less sampling effort than single sampling plans of equal strength and was suited to our operational needs. Using the same susceptibility test, sequential analysis of survival proportions from the 1,000-mu-g/g concentration was also compared with probit analysis of survival proportions from multiple concentrations. Sampling effort required for the SPRT was substantially lower than that required for probit analysis. Despite differences in probit analysis and SPRT assumptions, probit estimates of resistance severity at 1,000-mu-g/g agreed with sequential classification of r using a 1,000-mu-g/g concentration.