REALIZED HERITABILITY ESTIMATION FOR PESTICIDE RESISTANCE TRAITS

Tanaka & Noppun (1989) have described a procedure for estimating realized heritabilities, h2, from artificial selection experiments for pesticide resistance. Here I describe three processes especially relevant to such selection experiments that may introduce important biases into resulting estimates of h2. First, imperfect bioassays, in which different individuals receive different doses, have diminished abilities to discriminate between individuals on the basis of their tolerances. This results in an overestimate of S, the selection differential, and a corresponding underestimate of h2. Second, sublethal effects of pesticide exposure may generate variation in reproductive output among survivors of a round of artificial selection, with more tolerant individuals frequently producing more progeny. This results in an underestimate of S and a corresponding overestimate of h2. Third, unequal selection applied to males and females, especially likely to occur when mass selection is applied to adults of both sexes after mating has occurred, may cause S to be overestimated and h2 to be underestimated. Biases introduced by these processes may in some cases be large, adding an unacceptable amount of uncertainty to h2 estimates. Only if these biases are avoided can heritability estimates be usefully incorporated into prospective models of resistance evolution.