88 H. D. GOODALE AND GKACE MACMULLEN 



recall that the evidence favors the view that this class is wholly 

 artificial and depends for its existence upon an artificial division 

 point in the calendar year. However, no difficulties need be 

 expected in fitting the observed to expected ratios if the alterna- 

 tive theory be modified to consist of Pearl's three classes, for it 

 can be assumed that the double nulliplex class accounts for the 

 zero producers. We have applied this modification to several 

 of the observed ratios and found that it fitted well. 



THE MATERIAL AND ITS TREATMENT 



The kind and sources of data used 



The observations presented in this paper are the winter egg 

 records of our Rhode Island Red pullets for the years 1913 to 

 1917, inclusive, together with Pearl's data on Barred Plymouth 

 Rocks, pure Cornish, and the crosses between the two breeds. 

 (Pearl, '12). We have also a small amount of data on pure 

 Cornish and one cross of these with the Rhode Island Reds, the 

 latter, however, of prime importance. 



The arrangement of data 



The data are presented in the following order: 



1) A table (no. 4) showing Pearl's observed and expected 

 ratios compared with those of the alternative theory. 2) A table 

 (no. 5) showing the observed and expected ratios of our own 

 data arranged according to the two theories, together with 

 such comment as seems necessary. 



The primary object of this paper is to test the applicability 

 of Pearl's theory to our data. Consequently, we have reduced 

 our data to essentially the same form employed by Pearl. More- 

 over, we have included in our data only birds hatched in April 

 and May in order to conform to Pearl's practices in this 

 respect. 



In this study we have limited ourselves to a consideration of 

 one phase of the problem of the inheritance of winter egg pro- 

 duction. Later we purpose to present our data in a form that 



