CHROMOSOME THEORY OF HEREDITY 239 



chromatin are essential to life, and in that sense both parts of 

 the "constitution"; but in all organisms with chromosomes, 

 the chromatin has become specialised to discharge special 

 functions in regard to heredity. The factors ■ of heredity, I 

 would say, are those which are segregated in the gametes ; 

 they are so segregated by the activities of the chromosomes in 

 which they are lodged, and they constitute a mechanism 

 which has two distinct functions, both of prime importance. 

 The first function is to act as the self-regulating machinery of 

 heredity ; the second, in conjunction with sexual reproduc- 

 tion, is to allow the multiplication and more especially the 

 recombination of variations, and so to afford the possibility of 

 evolutionary change. It is noteworthy that the opinion of 

 biologists in the last three-quarters of a century upon the 

 function of sexual reproduction has oscillated between these 

 two extremes ; some have maintained that the effect of sexual 

 fusion of germ-plasms would help to preserve specific constancy, 

 others that it would encourage variability. It is the merit of 

 the chromosome-gene theory of heredity that it enables us 

 to reconcile these apparently incompatible ideas. 



The chief method of variation revealed by recent breeding 

 experiments is that known as mutation of single genes ; in 

 other words, the discontinuous and definite change of single 

 factors. The amount of the change, as revealed in the visible 

 alterations produced, may be very small or very large ; it may 

 apparently be due to a subtraction, an addition, or a qualitative 

 alteration which cannot be classified either as addition or as 

 subtraction. It is a curious irony that the occurrences in the 

 Evening Primrose upon which de Vries based his mutation 

 theory of evolution have finally been shown not to be mutations 

 in the true evolutionary sense at all, i.e. they are not due to 

 direct effects of changes in the hereditary constitution, but are 

 recombinations of already existing differences brought about 

 by conditions special to the genus (Renner, Ztschr. Ind. Abst. 

 Vererb., 18, 191 7). 



I remember discussing the general question with Morgan 

 in 191 2, and saying that, if we were to find a number of very 

 minute mutations of single factors, we should for all practical 

 purposes be back again in the old Darwinian position, with 

 the difference that we possessed information which he and his 

 immediate followers did not, as to the nature, extent, and loca- 

 tion of variations, and the mechanism of their transmission. 

 Events have moved in the direction which I thought ; the wide 

 occurrence of Mendelian behaviour having been demonstrated 

 on the large and striking variations, which, however, do not 

 appear to be the most important for evolution, attention is now 

 being focused upon smaller mutations, and we are coming 



