K. Pearson 
73 
The odds against such a deviation from equality are enormous, i.e. about 
10,000,000 to 1. But it will be noted that the abnormal members are largely in 
defect. Mr Nettleship himself remarks that this deficiency may probably be 
explained by unavoidable imperfections in the record, especially in the earlier 
generations. The excess of normals, however, has been more than maintained in the 
recent generations ; and if, in defending the Mendelian ratio of equality, we impugn 
the record on the ground that members of the stock are ashamed to own their 
defect and screen it, what then becomes of the evidence it undoubtedly provides 
for Mendelian theory in the " invariable continuity of descent of the disease in the 
affected branches, and its permanent disappearance from all other divisions of the 
genealogy " ? The screening of the disease mixst clearly have become an hereditary 
character, and the above evidence of segregation would be worthless. This 
principle of segregation is so vital, is such a wonderful, and I am inclined to say, 
glorious addition to our knowledge of heredity, that I am inclined in these cases of 
family abnormalities to defend Mendel against Mendel, or to preserve Mendelian 
segregation at the expense of Mendelian ratios. We have seen that the odds 
against a ratio of 3 to 2, which is the combined result of our siDlit-foot families, are 
very slight, and this applies equally to the case of Night Blindness, where a 
random sample deviating as much as the present one from the ratio of 3 to 2 
would occur once on the average in 6 or 7 trials. There is an additional point 
also to be noted. The Cure of Vendemean, according to Mr Nettleship, states 
that if the original night-blind individual, "the Jean Nougaret of 1637 could 
reappear to-day, except for a few individuals recently established in the district, all 
could salute him as their ancestor" {loc. cit. p. 2SS). If this be true, then the 
Mendelian ratio to be expected is more than 1 to 1, for the husband or wife of the 
night-blind individual would, in certain cases, be screening their own defect, and 
the offspring ought to have been night-blind in the ratio of 3 to 1. On the whole, 
while these cases give very definite evidence of the segregation factor, they do 
not seem to me to favour the segregation in rigid Mendelian proportions. 
(4) When we turn to the actual nature of the deformity in the case I am 
dealing with, we find all the principal types observed by Messrs Lewis and 
Embleton recur, with one exception which may exist, but owing to the few 
individuals who have been radiographed, this cannot be asserted. I refer to the 
existence of cross bones {Biometrika, Vol. Vi. p. 41). On the other hand, indi- 
viduals exist showing syndactyly and Polydactyly; gross deformity, several " fingers " 
placed promiscuously on a large misshapen " hand " ; or branching fingers which, 
without careful radiographic examination, it would perhaps be hard to classify as 
polydactyle or misplaced. 
The usual type of hand is to outward appearance, the hook-like form of my 
Plate XI, but the degree of defectiveness in the bones of the hand, even as 
affecting the carpus, may vary widely. In one case (II. 3) the hands were perfect 
but the feet deformed. There seems to be some evidence that in the younger 
generations the defectiveness is increasing ; thus all the children of II. 6 had two- 
Biometrika vi 10 
