Mabch 3, 1911] 



SCIENCE 



329 



is found in the Polish fowl. The essential 

 facts brought out are that the proportion of 

 the median element to the lateral elements of 

 the Y comb varies in the Fi progeny all the 

 way from to 90 per cent, of median element. 

 In the Fa and subsequent progeny the amount 

 of median element is distinctly correlated with 

 the amount present in the parents. The au- 

 thor's interpretation of this is that median 

 comb is imperfectly dominant over no-median 

 comb, and that the degree (or potency) of 

 dominance is inherited. Chapter II. deals with 

 the inheritance of the extra toe found in 

 various breeds of fowls. The facts here are 

 that while " extra toe " is sometimes dominant, 

 it sometimes fails to dominate. Lumping all 

 statistics together, it would appear to be the 

 case that the greater the degree of dominance 

 of extra toe in Fi the higher the proportion of 

 polydactylous young produced in subsequent 

 generations, thus apparently confirming 

 Castle's conclusions regarding the inheritance 

 of degrees of polydactylism in guinea-pigs. 

 Analysis of the data with reference to gametic 

 constitution of the parents, however, leads 

 plainly to the conclusion that " the average 

 condition of toes in the offspring of second or 

 later generation hybrids can not be used as 

 evidence of inheritance of the degree of pa- 

 rental development of the toes, since these are 

 dependent upon the same basal cause, namely, 

 the hidden gametic constitution of the pa- 

 rents." 



Chapter III. deals with syndactylism, or 

 " web foot " condition. This character appar- 

 ently segregates as though it were dominant to 

 normal foot, though the dominance itseK may 

 fail to appear in Fi. Chapter IV. deals with 

 rumplessness, a character of certainly puzzling 

 behavior in inheritance. Here the author's 

 earlier provisional conclusion that rumpless- 

 ness is recessive is reversed, and now this 

 condition is held to be dominant, the allelo- 

 morphic factors being " inhibitor of tail " and 

 its absence. The principle of imperfection of 

 dominance is adduced to explain the fact that 

 the Fi progeny are tailed. Some data regard- 

 ing the inheritance of presumably congenital 

 winglessness are given in the next chapter. 



but definite conclusions are not reached. The 

 next two chapters present some very valuable 

 evidence on the much-mooted problem of the 

 effectiveness of selection of fluctuating varia- 

 tions. The characters dealt with are booting 

 (feathering of legs) and nostril form. The 

 facts here are of great importance and almost 

 unique in the literature in that they give 

 for a bisexual organism data as to definite 

 grades of the character in individuals of 

 known gametic constitution with respect to 

 this character. The results show at once the 

 inherent fallacy of the basic assumption of the 

 biometric method of dealing with inheritance, 

 which tacitly assumes that all somatic varia- 

 tions are of equal hereditary significance,* 

 and at the same time the failure of selection 

 of fluctuating variations within a gametically 

 homogeneous population {= the nearest thing 

 to a " pure line " one can ever get in a bisex- 

 ual organism) to produce any result. If all 

 degrees of booting in parent and offspring of 

 all sorts of gametic constitution are lumped 

 together in one table there is a sensible corre- 

 lation between parent and offspring. One 

 might hence conclude that grades of booting 

 are inherited in a blending fashion, and 

 could be genetically modified by systematic 

 selection of slight " favorable variations." If, 

 however (and in showing this with such pre- 

 cision and clearness lies one of the most valu- 

 .able contributions of the work), a table is 

 made in which the individuals included are 

 homogeneoiis gametically there is no correla- 

 tion whatsoever between parent and offspring 

 in respect to grade of booting. The offspring 

 of parents with " much boot " have on the 

 average no more of the character than those 

 whose parents have " little boot," provided 

 both sets of parents are gametically alike as 

 regards " booting." Under these circum- 

 stances one would obviously not make progress 

 in selecting for increased booting. Nostril 

 form shows the same thing. 



The next two chapters deal with crest 



^ Cf. a more detailed discussion of this point in 

 a paper entitled " Biometric Ideas and Methods 

 in Biology: Their Significance and Limitations," 

 which is shortly to appear in " Scientia." 



