242 THE SECOND-CHROMOSOME GROUP 



in each sex the pure creams were separated from the others, and 

 finally the more difficult separation of heterozygous cream from eosin 

 was undertaken. The separation of the creams from the other colors 

 is accurate, but the final separation, that of the heterozygous creams 

 from the eosins, must be regarded as only a close approximation. 

 The sharp 1:2:1 ratio (160 : 318 : 148) which was obtained from 

 this separation probably represents among the eosins a small number 

 of the darkest heterozygotes, while the lightest of the pure eosins were 

 likewise classified among the heterozygotes. Probably 10 per cent 

 of the combined eosin and heterozygous cream class overlapped enough 

 so that the separations might or might not correspond to genetic 

 differences. One test of the correctness of the classification of inter- 

 mediates was made. From culture M 79 a heterozygous male and a 

 heterozygous female were selected, and the results (culture M 75) 

 showed that both individuals were of the supposed type. 



No attempt has been made to secure a stock homozygous for the 

 cream gene but without eosin. The cultures and experiments in 

 which such not-eosin creams must have constituted one-fourth of the 

 wild-type flies prove that such a stock could not be distinguished by 

 inspection from a wild stock. 



That the action of cream II is specific to eosin was suggested by 

 crosses of cream with vermilion (X chromosome) and with pink (third 

 chromosome). A careful examination of the F 2 flies from these 

 crosses showed no dilution of either vermilion or pink by the cream, that 

 is, the double recessives vermilion cream and pink cream (not-eosin) 

 are indistinguishable from vermilion and pink respectively. 



LINKAGE METHOD OF ANALYSIS FOR MULTIPLE-GENE CASES. 



The proper method of study for cases of multiple factors or of modi- 

 fiers is by means of linkage experiments, whereby all guesswork as 

 to the number and effect of modifiers can be eliminated. In Dro- 

 sophila such a study is rendered particularly easy by the fact that in 

 the male there is no crossing-over of any of the chromosomes. In con- 

 sequence, if two recessive genes which belong to the same chromosome, 

 e. g., black and vestigial of the second chromosome, enter the cross 

 from opposite parents ("repulsion"), the F 2 never shows flies which 

 have both these mutants at the same time. The double recessive 

 class is entirely unrepresented, and the 2 :1 :1 :0 ratio of "absolute 

 repulsion" results. This ratio holds, whatever may be the amount of 

 crossing-over in the female, for the lack of double-recessive sperm 

 prevents the double-recessive eggs from revealing themselves. This 

 ratio is in marked contrast to the 9:3:3:1 ratio, which obtains when 

 the two genes belong to different chromosomes, e. g., curved of the 

 second chromosome and ebony of the third chomosome. 



The light color cream was known to be eosin plus a recessive modifier 

 which belonged to an autosome group. To find whether this group 



