(1) and (2) give females, both red eyed, since both contain a red-factored X chromo­some. (3) and (4) give males, (3) giving rise to red-eyed males, since it contains a red-factored X chromo­some, (4) producing males with white eyes since this X chromo­some is lacking the factor for red eyes. Since all four combina­tions must appear in equal numbers (provided the experi­mental material is ample enough, which was the case in these experi­ments), in the F1 genera­tion both males and females should have red eyes and in the F2 genera­tion all the females should have red eyes and half of the males should have red, half white eyes. These results were obtained.

The experiments were carried further. No white-eyed females had appeared thus far. On the same assump­tions of the rela­tion of the X, X, and Y chromo­somes to the heredity of sex as well as to eye colour it was possible to predict under what condi­tions and in which propor­tions white-eyed females should arise. Thus if a red-eyed female of the F1 genera­tion (a cross between white-eyed male and normal female) be mated with a white-eyed male the result should be an equal number of white-eyed males and white-eyed females if the chromo­some theory of sex determina­tion were correct. The reasoning would be as follows:

The red-eyed female, having the chromo­some constitu­tion XX should form two kinds of eggs in equal numbers with the constitu­tion X and X; the white-eyed male having the chromo­some constitu­tion XY should form two kinds of spermatozoa X and Y. The following four types of individuals must then be produced in equal numbers:

(1) XX, (2) XX, (3) XY, and (4) XY.

In this case (2) must give rise to white-eyed females and (4) to white-eyed males, while (1) must give rise to red-eyed females and (3) to red-eyed males. Hence white-eyed males and females and red-eyed males and females are to be expected in this case in equal numbers, and this was actually observed.

The numerical agreement in this and the other experi­ments between the expected and observed result cannot well be an accident. The fact that the inheritance of sex-linked characters in man follows the same laws as in Drosophila is a strong argument in favour of the assump­tion that in man, also, sex is determined by two kinds of spermatozoa.

Morgan and his students discovered no less than thirty-six sex-linked characters in Drosophila, and each behaved in a similar way to the red and white eye colour in regard to sex-linked inheritance, so that the chromo­some theory of sex determina­tion rests on a safe basis. That sex is merely determined by the number of X chromo­somes, not by the Y chromo­some, is proved by the facts that the Y chromo­some may be completely absent as in Protenor and that Bridges[186] has found a type of female Drosophila with a chromo­some formula XXY whose sex was not affected by the super­numerary Y.

3. On the basis of all these experi­ments and theories it is comparatively easy to explain a number of phenomena concerning sex ratios which before had been very puzzling. In bees it had been shown many years ago by Dzierzon that the males develop from unfertilized eggs while the females, queens and workers, develop from fertilized eggs. This is intelligible on the assump­tion that the unfertilized egg contains only one X chromo­some while the spermato­zoön carries into the egg the second X chromo­some. But if the male bee produces two types of spermatozoa we should expect that only one-half of the fertilized eggs should be females, the other half males. But it happens that of the two types of spermatozoa only one is formed since in one of the cell divisions which lead to the forma­tion of spermatozoa one viable spermato­zoön only is formed while the other one perishes. It is, therefore, quite possible that it is the female-producing spermato­zoön which survives while the male-producing spermato­zoön dies.

It is occasionally observed that an insect shows one sex on one side of its body and the opposite sex on the other side. Boveri suggested that this phenomenon of gynandro­morphism is due to the fact that the spermato­zoön for some unknown reason does not fuse with the egg nucleus until after the egg has undergone its first cell division. In this case it fuses with the nucleus of one of the two cells into which the egg divides (or in some cases even one of the later cells?). As a consequence the one-half of the embryo which arises from the cell which was not fertilized would have only one X chromo­some and in a case like the bee would develop partheno­genetically, while the other half of the body, developing from the cell into which a spermato­zoön has penetrated, would be fertilized. The latter half of the body would be female, the former male. In his last paper before his untimely death, Boveri has given proof for the correctness of this interpreta­tion as far as gynandro­morphism in the bee is concerned.[187]