Maech 5, 1909] 



SCIENCE 



397 



the same sex-tendency, are, first, the moth, 

 Airaxas grossulariata, and, secondly, the can- 

 ary-bird. The two cases appear to be similar, 

 but as the former has been more fully worked 

 out, we may confine our attention to that. 

 The case of Abraxas has already been pre- 

 sented in part to the readers of Science by 

 Bateson and Punnett." 



This moth has a rare variety, lacticolor, 

 known originally only in the female sex. Por 

 brevity in description we may call the typical 

 grossulariata condition G, and the lacticolor 

 condition L. The latter is a Mendelian reces- 

 sive to the former. 



Cross 1. — The cross L? X Gr<^ gives only G 

 offspring in both sexes, but of course all bear- 

 ing L as a recessive character. See Table I. 



Cross 2. — Heterozygotes (produced by cross 

 1), when bred inter se, produce GS, Gc? and 

 L?, but in no case Ix? offspring. 



Cross S. — ^A heterozygote Gc?, mated with 

 LS, produces all four possible combinations, 

 G5, Gc?, L? and L<?. " The Lc^s thus produced 

 were the first that had ever been seen." Now 

 comes the most remarkable part of the whole 

 story. 



Gross If — When the newly produced Lc7s 

 were mated either with heterozygous G2s pro- 

 duced by cross 1, or with wild G?s, the off- 

 spring were all G in the male sex, all L in the 

 female sex. 



Cross 1 establishes beyond question the re- 

 cessive nature of the color character L. Cross 

 4 shows that the G2, whether cross-bred or 

 wild in origin, is heterozygous in color-char- 

 acter, bearing L as a recessive character. No 

 homozygous G2s have been found. Crosses 1 

 and 3 show that the male may be, as regards 

 character G, either homozygous, GG, or hetero- 

 zygous, GL, and cross 3 shows that it may 

 also be homozygous in L, that is, LL. In 

 other words, there is no correlation between 

 the male sex-character and either color-char- 

 acter. There does, however, clearly exist re- 

 pulsion between the female sex-character and 

 the color-character G, so that, whenever an 

 alternative is offered, f emaleness and L go into 

 one gamete, maleness and G into another. 



"Vol. 27, p. 785, May 16, 1908. 



But such alternatives manifestly occur only in 

 oogenesis, not in spermatogenesis. In no other 

 way can we account satisfactorily for either 

 the difference in result between the reciprocal 

 crosses, 1 and 4, or the failure of cross 2 to 

 produce the group Lc?. 



Bateson completes the explanation by offer- 

 ing the further suggestion that there is no 

 disjunction of the sex-characters in spermato- 

 genesis because the male does not carry the 

 female sex-determiner at all, but is homozy- 

 gous, S3. Consequently, when the L character 

 once gets into a male individual, as by cross 1, 

 where heterozygous GLc?s are produced, then 

 in the spermatogenesis of such an individual 

 gametes are sure to be formed in which the 

 male character is associated indifferently 

 either with G or with L. This, however, per- 

 mits of the production of (homozygous) Lc?s 

 only in cases where the egg bears the c? char- 

 acter associated with L, a condition realized 

 in cross 3, but not in cross 2 or cross 4. Don- 

 caster summarizes the case in a table, which 

 is here reproduced as Table I. 



Table I 

 A&rcKBos crosses, Doncaster's interpretation 



Lact. 



female 

 Gross. 



male 



Hetero- 

 zygous 

 female 



Hetero- 

 zygous 

 male 



Lact. 

 female 



Hetero- 

 zygous 

 male 



Hetero- 

 zygous 

 female 



Lact. 

 male 



LL?,? 



GL?^ 

 GL^^ 



LL?^ 

 GL3>3' 



GL?^ 

 LL^'^ 





L9,G^ 

 GJ',L,? 



G^,L3> 



L9,G,? 

 LJ', L^' 



|GL9^ = 

 IGL^^^-. 



fGL9^ = 

 LLp,? = 

 GL^^. 



[gg^j'^ 



[GL9^ = 

 LL 9 J = 

 GL^,? = 



lLL^'^ = 



(LL9^ = 



-gross, female 

 :gross. male 



-gross, female 

 =lact. female 

 =gross. male 

 =gross. male 



=gross. female 

 =lact. female 

 =gross. male 

 =lact. male 



ilact. female 

 -gross, male 



If, in Table I, we substitute X for the 

 symbol 2, discarding the symbol c? altogether, 

 and consider all individuals bearing X to be 



