84 MASS. EXPERIMENT STATION BULLETIN 264 



365 days. These studies were undertaken to discover whether the length of 

 the molt period was a criterion of second-year production. The following 

 constants came from the correlation table: 



Number of birds 541 



Mean length of molt period — days 78.45 



Molt period standard deviation — days ±31.81 



Mean second-year egg production 140.21 



Production standard deviation ±40.95 



Coefficient of correlation —.2960 ± .0265 



Correlation ratio .3572 



Mean second-year egg production ranged from 11 to 250 egjgs. There ex- 

 isted a wide variability here such as was observed in Section 10 when pullet- 

 year records were considered. 



The coefficient of correlation was tested for linearity of regression by the 

 X- method, which gave a probability of .4369 of linearity. The absolute mag- 

 nitude of this constant is sufficiently close to .30 to suggest a significant nega- 

 tive correlation between molt duration and second year production. This 

 correlation was in a sense false, however, because the time consumed in molt- 

 ing came out of the second-year production period more often than out of 

 the pullet-year production period. It seems inadvisable, therefore, to attach 

 very nmch weight to this constant. 



The correlation ratio of molt duration to second-year production was high- 

 er than the coefficient of correlation. The difference between the two cons- 

 tants by Blakeman's test is 3.57 times the probable error of difference. It 

 may be assimied, therefore, that the regression of molt duration upon second- 

 year production was not linear, and that the correlation ratio expresses a 

 significant negative relation between duration of molt and second-year egg 

 production. 



Summary 



1. This report includes studies of some of the factors possibly affecting 

 the duration of annual molt, also the relation of length of molt period to 

 vigor and to second-year egg production. In this study 936 pedigreed Rhode 

 Island Red hens that were trapnested for two complete laying years were 

 used. 



2. The simple correlation coefficient is used when regression is linear by 

 the X- test. When regression is non-linear, the correlation ratio is employed 

 and Blakeman's test for linearity applied. 



3. Hatching date did not affect duration of complete molt. 



4. Age at first egg was not significantly correlated with length of molt 

 period. 



5. Intensity showed a significant negative correlation with molt duration 

 of .3339. 



6. Length of winter pause was independent of duration of molt period. 



