March 9, 1922] 



NATURE 



313 



Mendelising pairs as there are chromosome pairs, 

 namely, seven), that this relation has as yet been 

 established ' only for species of Drosophila. But it is 

 also true that not a single animal or plant has yet been 

 found in which the number of known hereditary 

 groups of genes is greater than the number of chromo- 

 some pairs. It is to be anticipated that some one will 

 before long announce such a discovery, for it is very 

 probable that if two linked genes happen to be so far 

 apart as to give 50 per cent, of crossing-over, they 

 will appear to be in different groups. But such a 

 situation need cause no alarm when (or if) it arises, 

 and will not, of course, refute the correspondence of 

 linkage and chromosomes, unless it can be shown that 

 each such member belongs to a different linkage 

 system. P'urthermore, it is to be anticipated that 

 where compound groups of chromosomes exist, such 

 as have been described in some grasshoppers and 

 bugs, peculiar relations are likely to be found. 



The evidence from two species of Drosophila other 

 than D. melanogaster should also be taken into account. 

 In D. ohscura Lancefield has shown that there are five 

 pairs of independently assorting characters. There 

 are also five pairs of chromosomes. In D. virilis Metz 

 has found six pairs of chromosomes, and up to the 

 present at least five independent loci. The fact that 

 no crossing-over takes place in the male makes the 

 evidence for the independence of the pairs practically 

 certain. 



Origin of Mendelian Genes. 



Mendelian heredity is sometimes slightingly referred 

 to as a particular kind of heredity dealing with char- 

 acters that are due to losses of wild-type characters. 

 This view ignores some significant facts and considera- 

 tions. To argue that because a character is lost or 

 modified there must be a corresponding loss in the 

 germ-plasm is clearly a non sequitur. Each organ of 

 the body is the end result of a long series of stages in 

 embryonic development. Any change in any one of 

 the stages would be expected to alter the end product. 

 There are no grounds for assuming that such changes 

 must necessarily be losses, although losses also might 

 sometimes produce such effects. The argument has 

 all the earmarks of reasoning by analogy. 



However, the discussion need not rest any longer 

 on philosophical grounds, since we have crucial ex- 

 perimental data which show that loss of a character 

 is not necessarily due to loss of a gene. One case will 

 suffice. In addition to the white-eyed mutant of 

 I3rosophila there are ten other eye colours that lie 

 in the same locus. Obviously there cannot be ten 

 kinds of absences. The only other possible explana- 

 tion of ten absences would be that there were ten 

 genes here so close together that crossing-over does 

 not take place. Hence they appear to be in the same 

 locus. Now, fortunately, the origin of these ten 

 mutations is known, and shows — if they were really 

 a closely linked nest of genes— that when the last one 

 appeared there must have been at the same time 

 mutation in nine other genes in order to get the results. 



NO. 2732, VOL. 109] 



The rareness of mutation precludes such an interpreta- 

 tion. Attempting to save the interpretation of reces- 

 sive characters as due to absence of genes, it has been 

 argued that perhaps only a part of the wild-type gene 

 is lost when a new recessive character appears. It is, 

 however, not obvious why the hypothesis needs to 

 be saved. It is simpler and suffices to cover our 

 ignorance to say that a change has taken place. 



There is another question connected with these 

 multiple allelomorphs — changes in the same locus — 

 that is very important. Any given individual may 

 normally have at most any two of the genes (one 

 derived from the father, and one from the mother), 

 but never more than two. When there are two such 

 mutant genes present they behave towards each other 

 in the same way as does any mutant gene towards its 

 wild-type allelomorph. It follows that the Mendelian 

 behaviour is not a peculiar relation of a mutant gene 

 to a wild-type gene. It would seem, therefore, highly 

 probable that wild-type genes behave in this way 

 towards each other, and, in fact, where two wild 

 types exist in Nature that differ in a single allelomorph, 

 they are found to give a Mendelian segregation when 

 brought together. 



The discovery of a large number of mutants in the 

 same species may be expected in time to furnish some 

 idea of the number of hereditary genes that exist in 

 a species, or, in other words, to tell us how many 

 different kinds of genes plus the cytoplasm constitute 

 a species. At present, even in the case of Drosophila, 

 we are far from being able to make such a calculation. 

 There are, however, one or two rough estimates which 

 seem to indicate that the number of genes is more 

 than several thousands. The upper probable limit 

 cannot even be guessed. 



How the genes bring about their effects, which are 

 shown as modifications of the protoplasm (or by- 

 products of it), is entirely unknown. If it seems 

 desirable at present to limit the definition of heredity 

 to cover only the distribution of the genes in successive 

 generations, the result of their effects on the proto- 

 plasm becomes a problem of embryology. To many 

 geneticists, however, no such limitation seems desirable, 

 because it may appear that the ultimate constitution 

 of the genes themselves can be discovered only by 

 working backwards, through the effects produced, to 

 the nature of the material that furnishes the first 

 stage in the elaboration. With this pious hope I 

 heartily agree, but in the meantime I do not think 

 it desirable to let premature attempts in this direction 

 interfere with clear-cut methods of research that 

 Mendelian results supply. 



Finally, the question as to whether all hereditary 

 characters arise, or have arisen, through mutational 

 changes in the germ-plasm similar to those found 

 occurring to-day, can be settled only by future evidence. 

 Guessing is scarcely worth while. One point, however, 

 seems fairly well established— namely, that in several 

 cases where differences in wild species have been 

 subjected to the experimental analysis employed by 

 geneticists for variation arising by mutation, they give 

 the same kind of results. 



