658 



SCIENCE 



[N. S. Vol. XXVI. No. 672 



an extreme demand upon the specific 

 powers of chemical substances. 



Polydactylism has thus far failed to give 

 clear indications. Sometimes the inherit- 

 ance is Mendelian, while in other strains 

 or individuals dominance is so irregular 

 that the descent becomes untraceable. 

 Such irregularities of dominance here, as 

 elsewhere, may be referred with some 

 probability to the disturbing influences of 

 other undetected factors. It is much to be 

 hoped that cases of difference in the 

 ground-plan numbers of some radial type 

 will be found amenable to experimental 

 tests. Here the problem may be found in 

 a somewhat simplified form on account of 

 the elimination of serial differentiation. 



One most interesting class of characters 

 remains untouched. I refer to right- and 

 left-handedness. I can form no surmise as 

 to the laws which will govern the descent 

 of these characters. From Mayer's obser- 

 vations on Partula we learn that parents 

 of either twist may bear young of either 

 twist. The numbers in the uteri were so 

 small that the absolute numbers were in- 

 significant, and it may be an accident that 

 no mixture of types was found in any one 

 uterus. Direction of twist is a funda- 

 mental meristic phenomenon, being, as 

 Crampton and Conklin have proved, de- 

 termined as early as the first cleavage 

 plane; and great light on the problem of 

 cell division might perhaps be obtained if 

 the inheritances of these differences could 

 be determined. The only case we have 

 studied, that of Medicago, in which the 

 fruits are right- or left-spirals according 

 to species, proved unworkable, perhaps on 

 account of the minute size of the flower 

 and the roughness of the manipulations. 



I must now refer to the one positive case 

 alluded to above, in which a chromosome 

 difference has been proved to be associated 

 with a somatic difference. McClung, 



studying the accessory chromosome first 

 observed by Henking was the first to insist 

 on its importance. He showed that in cer- 

 tain insects half the sperms have it and 

 half are without it. This fact led him to 

 make the natural suggestion that the struc- 

 ture might be concerned in the differentia- 

 tion of sex. This suggestion has been shown 

 by Wilson to be correct, but the accessory 

 body proves to be the peculiarity of the 

 sperms which are destined to form females, 

 not of those which will form males, as had 

 been previously supposed. It was with no 

 ordinary feelings of pleasure that in the 

 past week many of us in Woods HoU, and 

 again the large audience assembled in this 

 room, beheld the fine series of photographs 

 which so amply demonstrate Wilson's far- 

 reaching discovery. 



The definiteness of the facts is evident 

 beyond all question, and whether the ac- 

 cessory body is in these types the "cause" 

 of femaleness or only associated with that 

 cause, we have at last the long-expected 

 proof that sex is determined in the germ 

 cells, so far as these specific cases are con- 

 cerned. In those cases we may even go 

 farther and declare that the female is 

 homozygous in femaleness, while the male 

 is heterozygous in sex. Such a result ac- 

 cords well, I think, with the general con- 

 elusions to which breeding experiments, on 

 the whole, point. For though great dis- 

 parities between the proportion of the 

 sexes occur in certain matings, these dis- 

 parities seem to be obliterated in suc- 

 ceeding generations. If the one sex were 

 homozygous and the other heterozygous, 

 such impermanence of the divergences is 

 what we might naturally expect.^ 



' In these remarks I have of course in view the 

 case where the actual number of the two sexea 

 show strange departures from equality. The phe- 

 nomena recorded by Doncaster in Abraxas grossu- 

 lariata and by Standfuss in Aglia tail, where the 

 proportions of the sexes belonging to two varietal 



