The fact that, in such a case as that which we have considered, it is immaterial which parent exhibits the dominant character, suggests that it is very improbable that this mechanism lies in the cytoplasm, because the amount of extra nuclear material con tributed to the zygote by most Metazoan spermatozoa is extremely small in amount, although it seems to include a representation of most cell organs. Attention is thus directed to the nucleus. From the fact that the only permanent or more accurately, persistently recurring structures within the nucleus are the chromosomes, and that the number of these bodies is constant in all members of a species, it is clear that either the number of chromosomes in the nucleus of the gametes is half that in body cells, or that after fertilization the number in the zygote nucleus is in some way halved. In fact, in animals and plants both mechanisms may be found, but that which is universal in Metazoa is the reduction of the number in the gamete nucleus.
A full amount of the details of the process by which this result is reached will be found in the article CYTOLOGY ; the fundamental phenomenon is that during the first stage of a division of a germ cell the chromosomes which compose a pair come together and lie side by side, so accurately adjusted to one another that they present the appearance of a single longitudinally split chro mosome. In many cases the original chromosomes then split longi tudinally, so that the nucleus appears to contain a number of threads, each composed of four chromosomes lying side by side. The number of such threads is necessarily half that of the chro mosomes present during the division of an ordinary cell of the animal. (See CYTOLOGY.) By two divisions which follow rapidly on one another, the four elements which build up each of these threads (paired chromo somes) become distributed into four cells, which, in the case of the male, are all functional spermatozoa, whilst in the female three are functionless, whilst the fourth is a mature or ripe ovum. When fertilization takes place the nuclei of the ovum and spermatozoan swell up, become exactly similar in size and char acter, and chromosomes appear within them. From the nature of the process of maturation it is clear that each will contain only half the number present in the normal body cell. These nuclei then fuse, and the chromosomes present in them neither fuse, nor are they united in pairs, but each becomes attached to the spindle, splits longitudinally and is separated into two halves at the first cleavage division which immediately follows. Thus each cell of the new individual contains two complete sets of chromosomes, one derived from its female, the other from its male parent. The accuracy of this statement has been confirmed by observation of special cases in which one or more of the chromosomes of one of the parents is visibly different from the corresponding chromosome of the other.
When gametes come to be formed it is apparent that each will receive a single complete set of chromosomes, a haploid group, and that this group need not be purely maternal or paternal in nature, but will owe its constitution to chance, certain chromo somes coming from one and the rest from the other parent.
This process of gamete formation and fertilization obviously affords a mechanism which will completely account for the ob served facts of a simple case of Mendelian heredity, if the factors which determine the development of the alternative characters lie in a chromosome.
Furthermore, it should follow that, in a cross involving two pairs of Mendelian characters, one of two things must happen. If the factors involved lie in the same chromosome the two characters will stick together in heredity, if they lie in different chromosomes, each should be inherited independently of the other, and the phenotypic nature of the individuals of the f 2 second hybrid generation will be determined by chance. Innumerable cases of each kind are now known.
In the case of Drosophila it has been shown that there are four groups of mutations, which are linked together in their inheritance.
These linkage groups differ very greatly in size. One includes only three mutant genes, two include about 8o each, whilst the fourth includes more than 200. If the linkage be dependent on the situa tion of the factors which produce these mutations in the same chromosome, it should follow that Drosophila should have a hap loid number of four chromosomes. This is the case, and the indi vidual chromosomes differ in length much as do the linkage groups.
Thus there is a very great probability that the factors which de termine Mendelian heredity lie in the chromosomes. It remains to determine their distribution. It is clear from the fact that many mutations may co-exist in the same individual without inter ference, that the factors determining them are discrete entities; to them is given the name gene. It was observed first in Droso phila, and since in other animals and plants, that the linkage be tween genes which lie in the same chromosome is not absolute, that in a certain definite proportion, differing for every two muta tions considered, a process which is called "crossing over" occurs; that is, genes which, as they lie in the same chromosome, should enter a single gamete, do not do so. This has been explained by a fracture of the chromosbme involved between the two genes, and a reunion of its parts, not with each other, but with appropriate fragments of the homologous chromosome when the two lie side by side during the process of maturation. By arguments based on this hypothesis it has been shown that all the unexpected oc currences in the heredity of the mutations of Drosophila can be accounted for if the genes have a linear and fixed distribution along the chromosomes. Indeed, maps of the chromosomes of this ani mal, showing the location of each gene have been published.