INHERITANCE OF WINTER EGG PRODUCTION 105 



as a 'somatic' high, mediocre, or zero producer, save in a very 

 few instances, and these all clearly result from some peculiarity 

 of management, such as birds hatched late in May with records 

 of nearly 30 eggs. Such exceptions are apparent rather than 

 real. 



In one or two instances females, such as female no. 6067 mated 

 to male no, 5584, table 3, have been encountered where the 

 number of high producers does not meet expectation. In the 

 case just cited, the progeny were obviously subnormal in vitality, 

 but were kept because they came from a high-hatching line. 

 We can readily understand that cases may arise where a bird is 

 genetically high although its record is mediocre, but it is hard 

 to see how a bird genetically mediocre should markedly transcend 

 the division line between classes. Some records, especially 

 records not far above the division line are to be expected, but 

 since it is a universal experience with Mendelian ratios that 

 individual families frequently deviate markedly from expected 

 ratios, although the average fit closely, it has seemed wisest to 

 take this way out of such a difficulty, rather than assume that 

 the genetic constitution is not represented by the somatic 

 record. * • 



In compiling the tables, the genetic constitution of each male 

 is first determined by the ratios in which his offspring occur. 

 When this result is checked against the parents of the male in 

 question, it may happen that a male of the class indicated by his 

 progeny could not have arisen from such parents. It is possible, 

 however, to adjust all such discrepancies where a division point 

 of 30 eggs is used. Doubtless adjustments can be made in the 

 case of the other division points, but it does not seem necessary 

 to pursue the matter further. Inability to make such adjust- 

 ments would indicate that neither theory has any basis in fact. 



The same care has been taken in classifying the females. To 

 do so, however, is not as simple as it seems, for one change often 

 involves others, and a long chain of changes is often necessary 

 in order to reduce all the data to a harmonious whole. We have 

 gone over the data with this end in view, for both theories 

 with the division point at 30 eggs. With the single exception 



