Methods are described for discovering whether a mixture of two poisons is as toxic as predicted on the hypothesis of independent joint action. These include X2tests and a procedure for finding the maximum likelihood estimate of the coefficient of correlation in resistance to the poisons. The methods are illustrated using data from insecticidal tests.

In the insecticidal tests, flour beetles, Tribolium castaneum Herbst, were sprayed with, or exposed to films of, different insecticides in solution in Shell oil P 31. The insecticides were pyrethrins, D.D.T., and B.H.C., singly and in pairs. The statistical analysis of the results showed that pyrethrins and D.D.T. could have acted independently both as films and as direct sprays; D.D.T. and B.H.C. could have acted independently as films but not as direct sprays; and B.H.C. and pyrethrins could not have acted independently either as films or as direct sprays.

These findings are discussed, and it is concluded that independent action should be regarded as a special case of a more general type of joint action for which the term dissimilar is proposed. A general method of approach is suggested for the conception and development of hypotheses of the joint action of poisons.