30. INTRODUCTION. 



continuous variation was capable of a Mendelian interpretation. 

 The whole subject of the inheritance of characters with fluctu- 

 ating variability has been ver}^ abl}^ dealt with recently by both 

 Tammes^ and Nilsson-Ehle.- The latter was the first to show 

 that characters, which to the eye appear similar, may in reality 

 be due to different genes which are inherited independently. 

 The red colour of the pericarp of some wheats is composed of 

 three factors, each of which will independently produce a red 

 colour, although less intense in tone than that due to the com- 

 bination. Recent work everywhere endorses the complicated 

 nature of most characters and has resulted in a large increase 

 in the number of factors recognised, while at the same time the 

 visible effect due to each factor appears smaller. In the paper 

 quoted above, Nilsson-Ehle discusses fully the question of 

 fluctuating variability and variation in general, and points out 

 that fluctuating variability may only exist as an effect of 

 environmental influence. If the number of factors n is large, the 

 number of homozygotic combinations possible will be much 

 larger, i.e., 2' and the differences between these combina- 

 tions will be smaller than the differences between the factors 

 themselves. If the heterozygotic forms are intermediate in 

 value between the homozygotic combinations, we may obtain a 

 continuous series in the Fg generation and as the combinations of 

 middle values occur most frequently, the form of the curve, 

 obtained in a graphic representation of the E^ generation, will 

 resemble that of an ordinary frequenc}^ curve. The hetero- 

 zygotic combinations, which occur at different points on this 

 curve, will in the F^ and succeeding generations give a progeny 

 which varies within much smaller limits than those of the Fg. 

 Assuming a large number of small factors, this is sufficient to 

 explain all variation which is not induced by environment. 

 A plant which exhibits small fluctuations in any one character 

 may be heterozygotic as regards that character and it should be 



* Tnmmes, I.e. 



» Nilsson-Ehle, I.e. 



