July 7, 1911] 



SCIENCE 



19 



pect such Fj whites to be in part hetero- 

 zygous, but in larger part homozygous, and 

 of these homozygotes some should be domi- 

 nant whites. Now a homozygous domi- 

 nant white individual, however mated, should 

 produce only white offspring, and only oc- 

 casionally would a pair of heterozygous domi- 

 nant whites be secured, the condition neces- 

 sary to allow yellow to recur in F,. But we 

 must remember that homozygous whites as 

 well as heterozygous ones occur in Fj. 

 Therefore matings of F, whites selected at 

 random with yellows all known to be hetero- 

 zygous should produce in a probable majority 

 of cases white F^ progeny which would be 

 homozygous. Two of these mated together 

 should, of course, breed true in F3. Accord- 

 ingly we expect pairs of F, whites in most 

 cases to produce only white F, oiispring. 

 Kellogg records five F^ matings of whites 

 from white X yellow F^ pairs (see Table II.). 

 Of these, four produced only white F, individ- 

 uals but one produced a mixed brood of 28 

 whites and 29 yellows. 



Let us next consider what is to be expected 

 from yellow X yellow F, pairs derived from 

 white X yellow F^ parents. The F, whites, as 

 we have seen, are often homozygous. There- 

 fore Fj yellows descended from them are 

 pretty sure to be heterozygous and should in 

 general produce mi-xed broods. Kellogg re- 

 cords five such matings (see Table II.), four 

 of which gave mixed broods aggregating 

 29 whites : 66 yellows, or approximately 

 Iw : 2.3y. The fifth brood produced 9 off- 

 spring, all yellow. From so small a number it 

 is impossible to decide whether one or both 

 parents were homozygous in yellow, as seems 

 probable. It is evident, however, that the 

 great majority of the F^ yellows of this origin 

 are, as expected, heterozygous. 



Finally mention should be made of a mating 

 of F^ yellows derived from white Fj parents, 

 which produced only yellow young in F3. 

 This seemingly contradictory result exactly 

 accords with Mendelian expectation, for in 

 this cross yellow was recessive in F„ there- 

 fore upon reappearing in F„ it should be 

 homozygous and so breed true. 



At every point, in this series of experiments, 

 the Mendelian expectation is realized, not 

 only in F., but also in F^, if we base that ex- 

 pectation on the behavior of the same races in 

 the single matings made by Kellogg. It is 

 therefore idle to seek for explanations of 

 " perturbed inheritance " due to double ma- 

 tings where no perturbation is discoverable. 



In a second series of double matings made 

 by Kellogg, the same white race (Bagdad) was 



TABLE IV. RESULTS OF MATING FRENCH YELLOW 



FEMALES DOUBLY, VIZ., WITH FRENCH YELLOW 



AND WITH BAGDAD WHITE MALES 



employed together with " French yellow." In 

 three matings (Table IV.) the French yellow 

 female was used, and in the same number of 

 cases (Table V.) the Bagdad white female 

 was used. In each case the female was mated 

 both with a male of her own race and with 



