GENERAL RELATIONS IN INHERITANCE 303, 



of these, called Havana, averaged 20 leaves to the plant, 

 the leaves being large, with an average area of about 7 

 square decimeters. The other variety, called Sumatra, aver- 

 aged 26 leaves to the plant, the leaves being smaller in 

 size, with an average surface of about 3 square decimeters. 

 The first generation offspring (Fi) from this cross were 

 intermediate in number and size of leaves between the two 

 parent stocks. When the Fi individuals were interbred, an 

 F2 generation was produced which showed great variation 

 In the number and size of the leaves. Some had a mean num- 

 ber of leaves (26) about the same as In the variety Sumatra, 

 with the large leaf size of the variety Havana. These plants 

 were Interbred for several generations, selecting In each 

 case for further propagation the individuals with the largest 

 and most numerous leaves. After several years a stock was 

 thus produced that was superior both in leaf number and 

 leaf size to both the original parents. It averaged about 30 

 leaves to the plant, and these leaves were as large as, or 

 larger than, those of the larger variety Havana. 



Thus by crossing two stocks, a new stock is produced 

 showing a distinct advance on the original stocks, a stock 

 with more numerous and larger leaves than before occurred. 

 How was this result produced? 



Without doubt It was due to the gradual collecting to- 

 gether into a single Individual of genes that were before 

 separated in the two original stocks. Assume (as suggested 

 by East and Hayes) that each stock contained certain genes 

 that tended to increase the number of leaves. Represent 

 these genes by capital letters, while genes that do not tend 

 to cause Increased number of leaves are designated by 

 lower case letters. Then the genes of the variety Havana 

 could be represented by some such combination as AAbb- 

 ccdd, etc., while the variety Sumatra could be represented 

 as aaBBCCDD, etc. By crossing the two and selecting the 

 best combinations for further breeding, there Is finally ob- 



