environment directly on the germ cells. The most fundamental phenomena in the unit character theory are unquestionably the segregations of characters that appear in the offspring of hybrids in so-called Mendelian inheritance. In the most typical cases, grandparental characters reappear in definite proportions of the progeny of the hybrid generation. The interpretation, according to the theory of unit characters, is in the hypothesis of purity of the germ cells of the hybrid generation with respect to the segregated characters; which means that the germ cells of the hybrid generation are pure with reference to the contrasting characters united in the soma; in other words, that corresponding contrasted characters can not both remain in the same germ cell, but are segregated in different ones and may thus appear pure in the descendants of a hybrid generation.[1]
We may well doubt that absolute purity of grandparental characters in the offspring of the hybrid generation occurs, and the results unquestionably vary with the environment; but I believe that we have to admit the general principle of segregation. However, the theory of segregation of unit characters in the germ cells is in no way necessary to explain the results; it is in fact inconsistent with the highly variable result; if unit characters were segregated in the germ, we should expect very definite constant results.
If we take our stand on the epigenetic basis and regard the germ cells as no more complex than direct investigation would lead us to suppose, then we have to admit that segregation in the germ cells can involve only constituents of the germ cells themselves. But any variation thus induced in the germ cells would be a factor in each process of the development, and would hence tend to influence every character that appears. Such a hypothesis involves the conception that germ cells contain elements capable of segregation; and this is so. Even if the principle of segregation of characters in inheritance had never been discovered, the principle of segregation of germ-cell elements would still hold, for the two discoveries were made absolutely independently.
I refer to the work of Guyer and Montgomery on the chromosomes, which has been followed by a long series of very exact studies. These studies certainly suggest segregation of parental chromosomes in varying proportions in different germ cells. Indeed, I know of no other interpretation of chromosome behavior that is consistent with the facts. Whatever value we may attribute to the chromosomes in cellular physiology, the variable relations established by their differential segregations, even if only quantitative differences are concerned, must involve endless secondary effects in the long series of cell generations that make
- ↑ Recent modifications of the theory of purity of the germ-cells do not essentially modify the argument.
