73 



3. When forms are crossed which are distinguished by a greater or 

 lesser number of units, the combinations which are possible stand in exact 

 proportion to the number of units involved. 



4. While different independent units can produce practically the same 

 external effect and while different combinations can show practically the same 

 external results, cases occur where, from crossings between sorts which ex- 

 ternally resemble each other, there arise new combinations with more or 

 less sharply distinct external characters due to a special grouping of the units. 

 In this manner forms which appear as 'new' and which might be mistaken 

 for mutations, can arise. 



5. Through continued regrouping of certain factors there may take 

 place what is commonly known as an acclimitization. The stronger and more 

 hardy combinations survive while the weaker and less resistant perish 

 in the struggle.* 



That continuous hereditary variations do not originate as a result of a 

 relatively limited number of independent units arising through " mutation " as 

 claimed by DeVries, is pointed out by Nilsson-Ehle (1. c. p. 7), who emphasizes 

 the practical importance of a correct understanding of this question. Thus, 

 " in the light of the conception of the existence of a relatively limited number 

 of different Forms or Elementary Species, one would expect to find in old 

 races a relatively small number of characteristic forms, and thus would run 

 the danger of over-looking much valuable material." 



The striking constancy of those isolated forms which have been investi- 

 gated is held responsible in part at least, for the idea that these represent 

 the essential units themselves. The reason for this constancy in self -fertiliz- 

 ing plants has already been explained. 



When the hereditary differences between a number of different bio- 

 types are investigated and classified, the continuity of hereditary variation 

 is revealed. These differences group themselves around a mean in the same 

 manner as do modifications. 



In discussing the various terminologies which have been proposed by 

 different writers for the different classes of variation, Baur (3) of Germany, 

 points out the confusion which has followed the application of the term 

 fluctuation to both modifications and to hereditary variations. The views 

 advanced by this writer are quite in accord with those of Nilsson-Ehle, who 

 has accepted the following general classification, although the difficulty of 

 making sharp distinctions is fully recognized: 



1. Modifications (not hereditary). 



2. Hereditary variation. 



. A distinction is then made between 



(a) Individual, continuous modification and 



(b) Individual, continuous variation. 



*This question is discussed more in detail by Nilsson-Ehle in a paper given by him at the International 

 Genetics Conference held in Paris, September 18-23rd, 1911. This paper is entitled " Acclimatation par 

 reco mbinaison de facteurs mendeliens." 



