The problem of whether a given amount of active ingredient is more effectively applied in concentrated or dilute form is discussed. Ifyis percentage mortality andx1andx2are respectively log‐concentration of active ingredient and log‐volume (or deposit) of insecticide applied,ymay be regarded as a uniform, continuous function ofx1andx2. The amount of active ingredient is constant, i.e.x1+x2=k, so thatdx/dx1=−dy/dx2=δy/δx1−δy/δx2Probit mortality,Y, can be substituted foryin (4). Thus, whether an active ingredient is better applied in concentrated or dilute form depends on the relative magnitudes of ∂y/∂x1and ∂y/∂x2, or of ∂Y/∂x1and ∂Y/∂x2. Equation (4) is true whenever an insecticide consists of an active ingredient in a diluent, whatever the dosage‐mortality relationship. Previous work is discussed in the light of (4) and its probit form, and it appears that the concentration at which an active ingredient is best applied can depend upon the nature and quantity of the active ingredient, and the method of application of the spray. There may be other facto*** The probit form of (4) is applied to the probit plane and confirmed experimentally. Flour beetles,Tribolium castaneumHerbst were sprayed with pyrethrins in Shell oil P 31, and it was found that ∂Y/∂x1>∂Y/∂x2, so that a given quantity of pyrethrins was more toxic in