Variation in Scutellar Bristle Number an Alternative Hypothesis
Alan Robertson · The American Naturalist · 1965
In summary, I have put forward here an alternative viewpoint which I believe has some value in simplifying the basic approach but which is still not without difficulties. One obvious example is the occurrence in high scoring lines of flies with bristles at completely new sites. I feel that the use of the probit transformation has led to some conclusions about the basic mechanism of scutellar bristle variation which are incorrect. I have in mind the conclusions about the relative sizes of the intervals on the underlying scale corresponding to various scores on the bristle scale and also the emphasis on the canalization at four bristles. It is true that there are four sites but the essence of this phenomenon is to me the precise determination of the unit process at each. In other words, it is essentially an all-or-none or small number problem. Some necessary statistical consequences about the relationship of mean and variability derive from this view, which may play some part in explaining the well-known phenomenon of the high variability of mutant genotypes. Finally, I feel that in any future work on this character, a consideration of the actual location of missing or extra bristles is necessary.