Size inheritance and the pure line theorv. 227 



publication of the Carnegie Institution of Washington. M.v colleague, 

 Dr. Joiix C. Phillips, has studied size inheritance in crosses between 

 Roueu and Mallard ducks. M,y botanical colleague, Professor East, and 

 his pupils. Hayes and Emerson, have meantime cairied out many size 

 crosses among plants, where large numbers of individuals are obtained 

 much more readily than among animals. Mention should also be made 

 of the investigations of Tammes, Wichler, Goldscioiidt, and others, 

 whose results have been either published or reviewed in this journal. 



As to the facts, all observers are in substantial agreement. 



They may be summarized thus: 1. When animals or plants 

 are crossed which have racial differences in size or other 

 characters, in respect to which each race shows continuous 

 variation about a different mean, the Fi progeny are of 

 intermediate size^). They may or may not be more variable 

 than the races crossed, but quite commonly are not. 2. The Fa 

 generation as a whole commonly varies about the same inter- 

 mediate mean as the Fi generation, but its variability as 

 measured by the standard deviation or the coefficient of varia- 

 tion is usually greater than that of the Fi generation. (The 

 greater variability of the Fa generation was not indicated in our earlier 

 observations on rabbits, but conies out in many of the observations 

 made later, as it does unmistakably in most of the observations made 

 on plants.) The increased variability of F2 as compared with Fi may 

 in extreme cases include forms as large as the larger parental race or 

 as small as the smaller race, and showing a tendency to vary in F3 

 about the same size as characterized the F2 parent. 



Some illustrative cases may be cited. Phillips (1912, 1913) 

 crossed two breeds of ducks which differed markedly in size, namely 

 Rouens and Mallards. The average adult weight of the Eouen race 

 used was, for males. 2321 grams, and for females 2237 grams. Corres- 

 ponding weights for the Mallard race were 1068 and 928 gi-ams respec- 

 tively. The Rouens accordingly were more than twice as large as the 

 Mallards. The two races did not overlap in weight, as appears from 

 Table 1, where the animals are classified by weight. In this table the 

 mean weight of the Mallards is taken as the center of class 2 and the 



*) I leave out of consideration here such differences as exist between tall and 

 dwarf plants, and between brachydactylous and normal men. In such cases a simple 

 Mendelizing difference exists, which shows both dominance and segregation in typical 

 fashion. Aside from this simple difference, however, ordinary size differences exist in 

 such cases, which I doubt not follow the ordinary rules of size inheritance. 



16* 



