me\di:liax experiments with fowls. 245 



contain one cliai-acter and the other half its allelomorph, as it is 

 called. My object hei-e is merely to di'aw attention to certain 

 points in my results which indicate that the segregation is not 

 necessarily complete. If we consider single characters and take 

 only the eight chicks of the first brood, omitting the one that 

 died, and the two of the second brood, we see that the Mendelian 

 ])roportion of 3 dominants to 1 i-ecessive is, in many of the 

 characters, either exactly or approximately exhibited. Thus we 

 have in the colour of the plumage (i coloured to 2 white, in the 

 structure of the feathers 6 noi-mal to 2 silky, in the combs 5 rose 

 to 3 single, in the colour of the skin 6 black to 2 normal, in the 

 crest 6 with crest to 2 without. The double hallux, however, 

 occurs only once, instead of six times, and as the male F^ had 

 normal feet although the extra toe chaiucter was present in tlie 

 cross, it is by no means certain that this chai'acter was segregated 

 in his gametes. The featheiing of the legs and the ci'est on the 

 head exhibit a condition which does not harmonise exactly with 

 the Mendelian theory. The leg-feathering is present in all the 

 specimens except No. 10 in the second brood, but it exhibits very 

 different degrees of development. One explanation of this may 

 be that such a character is naturally more developed in the 

 dominant homozygote, where the chai'acter was present in both 

 gametes, than in the hefcerozygote, where it was present only in 

 one. and this is probably the case. But there is one specimen. 

 No. 5, in which the feathering is the merest trace, and Meudelism 

 affords no explanation of such variations within a character, 

 it merely takes account of its presence or absence. Similar 

 i-emarks may be made concerning the crest : of the 8 chicks of 

 the first brood there are 6 with crest and 2 without, the Mende- 

 lian proportion, and there are 4 with well-developed crest and 

 2 with a vei-y slight crest. The latter may be the hetero- 

 zygotes for this chai-acter, although there should be four of these 

 to two dominant homozygotes. 



I have noticed an interesting correlation between the crest 

 and the chai-acter of the comb. As Bateson points out, the rose 

 comb of the Silky has its posterior end divided into three 

 irregular points. He desci'ibes the comb as a rose plus a trifid 

 element which is the cause of this condition of the posterior end, 

 and states that in P\ fi'om Silky x Single, regular rose combs are 

 produced in those individuals which have the i-ose factor without 

 this trifid element. In otlier words, the comb of the Silky is re- 

 presented not by one factor but by two. The i-egular or ordinaiy 

 i-ose comb, as it occui's in Hamburg fowls, has a flat upper 

 surface which is triangular in shape, the apex of the triangle, 

 called the peak or pike, being free from the head and projecting 

 backwards. Bateson 's statements are for the most part confirmed 

 by my results, but I have a- somewhat different interpretation to 

 offer. The rose comb of the Silky is not only trifid postei-iorly 

 but abbreviated. This chaiacter occurs in botli the F/s in my 

 experiment ; it is dominant. In the F^'s the same character 



