Viable spore count as an index of effective dose of Bacillus thuringiensis

Leaf samples from Cercis occidentalis trees, collected 5–75 hr after application of varying concentrations of a Bacillus thuringiensis formulation, Dipel, were used as dose sources for third instar Schizura concinna larvae. Dose, defined as logarithmically transformed, estimated viable spore counts, and duration of time the formulation was on the leaf prior to sampling were considered in combination as predictors of insect mortality. By use of the multiple logistic model both factors were found to be significant predictors of insect mortality (P < 0.0001). Available methods for assessing goodness of fit indicate that the multiple logistic model adequately fits these data.