228 Castle. 



mean weight of the Roueus as the center of class 10. The 70 F2 off- 

 spring have their mode in the intermediate class 6, though they range 

 all the wa3' from class 2 to class 9. The 63 F2 offspring likewise have 

 their mode in class 6, and are slightly more variable than Fi , though 

 only one aberrant individual falls beyond the range of Fi. 



The results of Mac Do well on the bone dimensions of rabbits are 

 similar to those just described for ducks. Fi is in all cases intermediate 

 and F2 usually but not always shows increased variability. The rabbit 

 material was, however, scarcely as favorable as the ducks because the 

 races employed were less pure. 



The extensive and carefully executed studies of Emerson and East 

 (1913) upon crosses of maize involving differences in size and other 

 quantitative characters afford excellent illustrations of the usual con- 

 sequences of size crosses. The simplest and clearest cut cases relate 

 to the size of the ear or of the seeds borne upon it. The behavior of 

 eai- diameter in crosses is shown in Table 2 (Emerson and East, p. 56). 



Both Fl and Fs are intermediate in character in comparison with 

 the parent races, but F2 is shghtly more variable. Different lots of Fi 

 progeny (combiued in Table 2) give coefficients of variability of 8'29 

 and 6'88 respectively, whereas Fo progeny have coefficients ranging 

 from 9'66 to 11-77. The extreme ranges of the parent races are not 

 attained in Fo. 



Table 3 shows the result of crossing two races of corn (A and B) 

 differing in seed width. In this cross also, Fi and F» were alike inter- 

 mediate, but the latter was slightly more variable. It was found that 

 the Fi plants differed in genetic character as to seed width. An Fs 

 with low seed width (143 mm.) produced an Fg likewise low (mean 

 141'3 mm.); an Fs with seed width above the average (178 mm.) produced 

 an F3 of like character (mean 172'9 mm.). In interpreting this case it 

 must be borne in mind that among the Fi individuals differences of 

 the same sort occurred as were thus shown to exist among the Fä 

 individuals. For a high Fi individual (198 mm.) produced F» offspring 

 likewise high (mean 178-7 mm.) but an Fi individual of seed width 

 178 mm. produced F2 offspring close to the general average (162'2 mm.). 

 Between the Fi parents in the two cases there was a difference of 

 20 mm. : between their progeny there was a corresi)onding difference of 

 16'5 mm. This .shows that genetic differences in seed width existed 

 among the Fi as well as among the Fs individuals. It seems i)robable 

 therefore that similar genetic differences might have been found in the 

 uncrossed races had they l)een looked for, and it becomes at least an 



