A COMMUNICATION1 from B. R. Speicher has directed attention to certain genetical data which do not seem to favour my suggestion2 that the pre-conjugation phenomena observed in the gametogenesis of the honey bee and of Cynips kollari may be interpreted in terms of derived tetraploidy. Now while I admit that the findings of the experiments of Dzierzon, Michailoff and Newell, as quoted in Mr. Speicher's letter, do suggest a simple condition of female diploidy and male haploidy, the facts are by no means fatal to a theory of derived tetraploidy–diploidy based on cytological observations. It has recently been shown that tetraploid hybrids do not necessarily exhibit independent segregation of chromosomes in gametogenesis and also that gametic viability in these forms is often differential, and only those gametes carrying one complete genom or the other are functional. It is thus possible for an organism to behave genetically as a diploid while in its gametogenesis traces of a different cytological condition may be shown. Genetical methods, therefore, may not be infallible criteria of evolution.