THEORIES IN PRACTICE 377 



\4duals in a single generation to put it to the 

 test. For the number of possible combinations 

 increases in geometrical ratio, as we have seen, 

 with the increased number of characters under 

 consideration. And a really penetrating view 

 of the situation reveals to us hereditary factors 

 in the germ plasm of each individual plant that 

 would be numbered, could we isolate them, not 

 merely by tens or scores; not merely by hun- 

 dreds or thousands; but rather by hundreds of 

 thousands or millions. 



To those experimenters who have been prone 

 to think of "unit characters" as few in number, 

 such a statement will perhaps seem anomalous. 

 Yet there can be no question that it is fully 

 justified. 



In point of fact, what the present day student 

 of heredity usually speaks of as a unit character 

 might better be referred to as a "unit complex," 

 or by some allied term that would suggest its 

 complicated character. The word "gene-com- 

 plex" has been suggested in a similar connection. 



It would appear that the real purpose of 

 selective breeding through many generations is 

 to remove one after another of the factors that 

 dominate or mask other factors, so that subor- 

 dinate or recessive factors may make themselves 

 manifest. 



