A. Barrington and K. Pearson 
453 
probably closely alike in the four cases, although the categories in the case of the 
shorthorn render the comparison harder. 
(10) On Collateral Inheritance. 
The work of tabling pairs of whole siblings is somewhat lengthy in the case of 
shorthorns, the breeding life of a cow is short, and there is a good deal of changing 
in the case of the bull used. Accordingly we have contented ourselves with the 
cases of whole brothers and whole sisters*. It is much easier to table cases of 
half siblings, and six tables of half siblings — brother-brother, sister-sister, brother- 
sister, the bull parent being common, and again the cow parent being common — 
have been constructed -f. These half-sibling tables are of peculiar interest, because 
there is a good deal of obscurity at present about the degrees of likeness we should 
expect to find between half siblings. A pi^iori we should expect the degree of 
likeness of half siblings to be half that of whole siblings, but this relationship has 
not been verified on any of the material hitherto considered : half siblings are 
relatively more alike than whole siblings, the addition of another common parent 
does not double the resemblance. 
In Table IX below we have put together the results with regard to whole and 
half siblings as far as pigmentation results have yet been worked out. 
Several conclusions may be drawn from this table. In the first place we see 
that half siblings through the bull are more alike than half siblings through the 
cow, and this is so for all three combinations. It would thus appear that the bull 
is more influential than the cow, but there is no prepotency of the bull over the 
cow in the parental resemblances. We should thus have the paradox that the bull 
caused the offspring to be more alike than the cow does, but not by causing them 
to be more like himself ! Notwithstanding that the whole-sibling resemblances 
in shorthorns are of the same general value as those for eye-colour in man or coat- 
colour in Basset Hounds or Greyhounds of different litters, our opinion is that 
they are too low. Our experience with the parental tables leads us to believe 
that the grade of the resemblance is nearer that for hair-colour in man or coat- 
colour in horses. But we should not expect the reduction to be as great in this 
as in the parental case, owing to the special emphasis breeders lay on certain classes 
of parents. It is quite clear that neither in horses nor in shorthorns is the degree 
of resemblance for half siblings half that between whole siblings. The ratio 
of resemblance of maternal half siblings (the only case worked out for horses) to 
whole siblings is for horses '57, and for shorthorns is '56. These results are closely 
in accord, and our impression that a constant ratio exists between half and whole 
sibling resemblance, and that this ratio is greater than '5, is not shaken by the 
values in the case of Basset Hounds. We think that a very profitable line of 
investigation would deal with the relation of half to whole brethren in a variety 
of characters, and, if possible, with measurable characters as well as pigments. 
So far as we can see in this, as in other features of collateral inheritance, short- 
horns range themselves well within the limits already found for other species. 
* Tables a and /3 of the Appendix. t Tables 7, 5, e, ri, 0, of the Appendix. 
