542 WHITE— STUDIES OF INHERITANCE IN PISUM. 



of some of the F^ intermediates. On the four-factor interpretation, 

 the extreme scarcity in F2 of the large-seeded grandparental type is 

 accounted for by regarding its factorial composition as due to all 

 four factors in a homozygous condition (AABBCCDD). Combin- 

 ing the F2 classes I., II. and III., the F„ ratio of small and interme- 

 diate seed weights to large seed weight is 793 : 3 or 264 : i which is 

 somewhat close to the theoretically expected ratio (255:1). Like- 

 wise the relation of the small-seeded Fj's to the remainder of the 

 Fo progeny on a four-factor basis is theoretically 7:248:1, while 

 Tschermak's actual numbers were 22 : 771 : 3 or y.^i : 257 : i. Accord- 

 ing to his provisional hypothesis, the 22 small-seeded F^'s repre- 

 sent not only those which will breed true (aabbccdd) but also small- 

 seeded forms which will give intermediates (aabbccDd, aabbCcdd, 

 etc.). Tschermak finds no evidence in his experiments for believing 

 that sterility is in any way responsible for the small Yo numbers of 

 the large-seeded segregates. He also finds no reason for believing 

 in a differential relation of the environment which would be so much 

 more unfavorable to the large-seeded types. 



6. Height, Stem Diameter, Internode Length and Internode 



Number. 



As described by Mendel and most geneticists since 1900, the 

 heredity of height or length of stem in peas represents a very 

 simple problem, the presence and absence of a factor for tallness. 

 While Mendel's description and interpretation of results from cross- 

 ing tails and dwarfs accounts for most of the facts derived from 

 studying the genetics of two varieties differing in height, it fails to 

 account for all the facts when pea varieties in general or as a 

 whole are under consideration. Height in peas is generally arbi- 

 trarily divided into dwarfs (23-90 cm.), half drawfs (90-150 cm.) 

 and tails (150-300 cm.). As pointed out by Lock, Keeble and 

 others, height of a given variety in any given year is very much in- 

 fluenced by environmental conditions, so that in any detailed study 

 of the heredity of height, parents, F^ and subsequent generations 

 should be grown side by side, as this method insures a minimum 

 amount of variation in the environment. The environmental condi- 

 tions which modify height are numerous, including defective or 



