Mortality of last instar western spruce bud worm, Choristoneura occidentalis Freeman, exposed to a series of increasing doses (topical application) or concentrations (diet incorporation) of 10 insecticides was analyzed over 7 d by time-dose-mortality regression. Regression based on the complementary log-log model was useful for analyzing time trends for all dose levels simultaneously. Each of the 10 insecticides acted more quickly when applied topically. In both kinds of experiments, speed of lethal action of the insecticides depended on dose or concentration. The statistical methods developed to estimate lethal doses as functions of time and median times of death as functions of dose are more efficient than methods published previously.