110 Cc. H. DANFORTH 
the other three characters do not, so that linking can only be 
tested by comparing the proportional distribution of one pair of 
contrasted characters with reference to other pairs. The table 
shows the observed values for all possible combinations and the 
expected values for each of these based on the behavior of the 
several characters considered separately. 
It will at once be apparent that the observed and expected 
values agree very closely in the first five sets of combinations. 
The slight apparent deviations in favor of a correlation between 
polydactyly and broad comb and between polydactyly and 
booting are too small to be of signficance. So far as this part 
of the data goes it is in full accord with Davenport’s (06) con- 
clusion that correlation of characteristics in poultry is very rare. 
But, on the other hand, between brachydactyly and booting 
there is found to be a close correlation. Here the values are 
clearly significant. Of thirty-one brachydactyl chicks, all had 
booted tarsi; and of thirty-six chicks with booted tarsi, thirty- 
one were brachydactyl. The distribution is shown graphically in 
figure 5 where it also appears that the length of the fourth toe is 
rather more variable in specimens with booted shanks than in 
those with smooth shanks. 
There is, then, a distinct relation, which becomes apparent in 
embryonic stages, between brachydactyly and booting, but 
there is no evidence of a significant connection between either 
of these characters and comb form, color or polydactyly. 
SIGNIFICANCE OF THE CORRELATION BETWEEN BRACHY- 
DACTYLY AND BOOTING 
The explanation for the relation between short toes and 
feathered tarsi is not obvious. Several possibilities suggest 
themselves. 
In the first place, it might be that we have here a case of true 
linkage, such, for example, as Morgan (715) and others have 
successfully demonstrated in fruit flies. Morgan believes that 
in such cases the determiners for both of the linked characters 
are located in the same chromosome. If this were the true 
explanation, ‘crossing over’ might possibly be expected to occur 
