Dr Kettlewell's proof that the dark form carbonaria of Biston betularia has replaced the type, at least in part as the result of selection by bird predators, gives me the right to bring my calculation (Haldane 1924) on this matter up to date. Assuming random mating, if the relative fitnesses of the genotypes CC, Cc and cc are as 1:1 - k:1 - K, and the gene frequencies are pC + qc, then the annual change in the value of p is Δp = (pq(kp - kq + Kq)) / (1 - 2kpq - Kq2) . But K and k are not constants, even in a constant environment. They change with the frequency of other genes (or possibly plasmons). At present in industrial areas, K is positive and k nearly zero, since C is apparently fully dominant. However, on the genetic background of unpolluted areas, Cc is intermediate and k may be about 1/2K (Ford 1955).