class="tdc br">...

13 4 762 2011 444 ... ... 10 ... 2011 2621 ... ... 9 ... 2333 444 ... ^[A]5 11 ... 2333 2621 ... ^[A]1 2 ... 2618 444 ... ... 2 ... 2618 2621 ... ... 5 ... 3776 444 ... ... 2 ... 3776 2621 ... 1 14 ... 820 2016 4731 ... ... 10 ... 2255 4731 ... ... 16 ... 5143 4731 ... ... 45 ... 6479 4731 ... ... 31 ... 832 ^[B]2618 5119 ^[B]1 ... 23 ... 3776 5119 ... ... 28 ... 4404 5119 ... ... 9 ... 4732 5119 ... ... 3 ... 5803 5119 ... ... 21 2 6481 5119 ... ... 11 ... 834 2324 5090 ... ... 26 ... Total 2 11 367 7

The distribution of offspring in the 24 families of table 5 is in fair accord with any of the three hypotheses, but seems to favor the second, for that hypothesis calls for families with combless children, whereas such are not to be expected on the first hypothesis. Moreover, agreement with the second hypothesis is fairly close, for that calls for 3 families with combless children and there were actually 3 such families showing a total of 1.8 per cent combless, where expectation is 2.8 per cent. What is opposed to any hypothesis is the appearance of some Y-combed offspring; and to account for this the hypothesis is suggested that the germ-cells of some parents with oo comb contain traces of the I-comb determiner. The word "traces" is used because the median element in these Y-combed offspring is practically always very small. It is fair, consequently, to conclude that oo × oo gives oo-combed, and occasionally combless, offspring. This conclusion is further supported by the statistics derived from extracted oo comb of all generations bred inter se, which give: Y 11, oo 427, and no comb 8, where the 11 Y-combed birds are those just referred to as progeny of F₂ parents. The non-median comb, consequently, probably contains only non-median germ-cells.