INHERITANCE OF WINTER EGG PRODUCTION 113 



birds react variously. Some are apparently immune, others 

 take the disease in a mild form and without apparent detriment 

 to egg production, while others take it with various degrees of 

 severity (our Cornish, for example, are particularly susceptible). 

 Such individuals may cease laying entirely for a time or lay 

 spasmodically. The presence of roup, therefore, complicates 

 matters greatly. 



The degree to which these changes in management have af- 

 fected our records is a matter of surmise only. We have no 

 precise data on this point. The evidence from the records 

 themselves indicates that it may be of small moment as attested 

 by the presence of low records in clean years. Nevertheless, the 

 presence of these changes in management introduces an element 

 of doubt, especially because of the variability in susceptibility 

 of individuals and families to roup. 



Size of families 



There is a difficulty,^ common to Pearl's data and my own, 

 that renders it impossible at present to decide between the two 

 theories and their several modifications and which renders it 

 somewhat doubtful that either scheme has any foundation in 

 fact. This difficulty lies in the fact that the adult female off- 

 spring of each pair are so few in number that it is almost always 

 possible to refer any observed ratio to some theoretical ratio that 

 will bring each mother into line with the rest of her group. 

 Thus an observed 2 : 2 ratio will fit any of the following theo- 

 retical ratios, viz., 2 : 2, 1 : 3, 3 : 1, 3 : 5, or 9 : 7. Since the progeny 

 of a single male in numerous instances is fairly large, it may be 

 urged, by the law of errors, that the agreement between observed 

 and theoretical ratios is adequate proof that the scheme correctly 

 represents the actual mode of inheritance of winter egg produc- 

 tion. This would be true only under certain conditions, which 

 are: First, there must be no bias. Second, the sample of females 

 mated must be a representative sample of the population. 



^ This difficulty is not confined to egg-production statistics, but unfortunately 

 is encountered in much Mendelian work, particularly with mankind. 



