18 



SCIENCE 



[N. S. Vol. XXXIV. No. 862 



TABLE III. RESULTS OF MATING ISTRIAN YELLOW 



FEMALES DOUBLY, VIZ., WITH ISTRIAN YELLOW 



AND WITH BAGDAD WHITE MALES 



Fifteen matings of F^ yellows inter se 

 (Tables II. and III.) gave in F, 109 

 whites : 321 yellows, or Iw : 3y, exactly as in 

 the corresponding generation following single 

 matings of the same races. 



Five matings of F^ white with Fj yellow 

 (Table II.) gave 93 whites : 98 yellows, close 

 equality of the two sorts, again as in the 

 single matings. 



Thus far we have no evidence of any de- 

 parture in the double matings from Mendelian 

 inheritance as strict as prevails in the single 

 matings. If such departures occur they are 

 to be found in the remainder of Kellogg's 

 statistics, which are based on F3 matings. No 

 F3 broods from single matings of the races 

 under consideration are reported by Kellogg. 

 Surely he has not overlooked the elementary 

 fact that F3 matings should on a Mendelian 

 expectation produce a different result from 

 the F., ones, yet on no other supposition can 

 I account for the fact that Kellogg speaks of 

 " irregularities " that occur with " non-Men- 

 delian regularity." These " irregularities " 

 of F3 as compared with F^ are not " non- 

 Mendelian " ; they are part of the Mendelian 

 expectation as a moment's consideration will 

 show. The F^ individuals, whether yellows or 

 whites, produced by a (single mating) cross 

 of yellow with white are all heterozygous and 

 so are expected to jsroduce mixed broods, but 

 of the F„ individuals only part are heterozy- 

 gous. The F, yellows from Fj yellow pairs 



should be two thirds of them heterozygous 

 like the F^ yellows, but one third of them 

 should be homozygous and so capable of pro- 

 ducing all yellow broods. The same would 

 be true also of the F^ whites. 



Kellogg reports seventeen matings inter se 

 of F, yellows derived from Fj ancestors, one 

 or both of which were yellows. Six of these 

 matings produced only yellow progeny indi- 

 cating that the pair contained at least one 

 homozygous yellow individual. One of the 

 matings produced only white progeny, a sur- 

 prising result if it is not a misprint. The re- 

 maining ten matings produced mixed broods, 

 like those of the F^ generation, aggregating 

 64 whites : lYl yellows, or Iw : 2.7y, a slight 

 deficiency of yellows but no greater than the 

 expected Mendelian deviations from the exact 

 1 : 3 ratio. 



Let us now consider what is to be expected 

 regarding F„ whites mated inter se. If the 

 egg of a Bagdad white female were fertilized 

 by sperm of a Bagdad white male, we should 

 expect Fj and all subsequent generations from 

 such fertilizations to contain only white indi- 

 viduals. Seven F^ broods (Table II., sum- 

 mary) fall in this category as previously 

 noted, being derived from seven different F^ 

 pairs coming from white Fj pairs produced 

 by a white mother doubly mated, but as re- 

 gards these eggs evidently purely fertilized. 

 These seven pairs produced only white off- 

 spring both in F. and in F3. 



In case the Bagdad white egg were fertil- 

 ized by yellow sperm and gave yellow F^ in- 

 dividuals which produced mixed broods, and 

 white individuals from these broods were 

 mated inter se, we should expect them to pro- 

 duce only white offspring, white being the 

 recessive character in such cases and so pre- 

 sumably pure. Six of the F, white pairs 

 (Table II.) fall in this category and produce 

 only white F3 offspring, except for one indi- 

 vidual possibly a yellow " straggler," as sug- 

 gested by Kellogg. 



In case the F„ white individuals were de- 

 rived from a crossed (and so dominant) white 

 Fj individual mated with a yellow one like- 

 wise a dominant heterozygote, we should ex- 



