A warning note is sounded concerning dependence on log-probit regression lines for the detection of resistance to insecticides. By the use of only a few serial dosages, a straight-line relationship between dosage and mortality may result from testing a mixed population containing small but significant numbers of resistant individuals.
Further, the difference in LD30 between this population and a homogeneous susceptible one may be so small that the presence of resistance will not be suspected. The use of discriminating dosages and the rearing and testing of the offspring of survivors of these dosages is suggested as the only sure way of deterting resistance.
