An Introduction to a Biology 



stability or unfixity of a character, on the other hand, is due 

 to the fact that though one of the germ-cells which pro- 

 duced the organism bearing the character in question con- 

 tained the factor for that character, the other germ-cell 

 contained the other factor of the pair. In Mendelian terms 

 the organism was heterozygous for that character. Fixa- 

 tion, that is to say, is due, according to Mendelian prin- 

 ciples, not to the continued breeding to a particular standard 

 over a great number of generations, but to the union of 

 germ-cells bearing the appropriate characters. In other 

 words, time plays no part in the fixation of a character. 



(2) Another idea essential to Mendelian principles is 

 that all that the breeder can do is to effect re-combinations 

 of existing characters. The creation of new characters has 

 no place in the Mendelian scheme. 



(3) Lastly, the Mendelian principles render possible for 

 the first time the prediction of the result of the union of a 

 given pair of individuals, in cases which have already been 

 subjected to Mendelian analysis ; and not only this, but the 

 numerical ratios in which the various characters will appear 

 amongst the offspring resulting from the union can also be 

 predicted. 



If these conclusions be compared with the first three 

 conclusions arrived at after a consideration of M. Bergson's 

 biology, it will be seen that they are flatly contradictory. 

 His main conclusions with regard to life are untrue of 

 Mendelian characters. This may mean either (1) that M. 

 Bergson's conclusions are ill-founded, or (2) that the Men- 

 delian characters are dead or, at any rate, appertain to the 

 least vital parts of the organism. I believe the latter alter- 

 native to be nearer the truth. If it is nearer the truth, we 

 have, I think, a clue which will enable us to relegate the 

 Mendelian characters to their true position among the 

 characters of living things ; and a suggestion which may 

 enable us to determine, without experimentation, which 

 characters are likely to behave in a Mendelian way in heredity 



99 



