BR()Oi:)INESS AND FECUNl^lTY IX FOWL 7.5 



ard deviation in times broody is greater than tlie mean because 951 birds (83 

 per cent) fell into classes 1-4., leaving only 17 per cent in classes 5-13. Tlie 

 modal class is 2. 



The mean length of liroody periods is 15.10 days with a standard deviation 

 of ± 3.78. There is, therefore, much greater uniformity in length of period 

 tlian is observed for number of periods. Evidently tlie number of periods 

 otfeTs a more fertile field for improvement than is offered by the length of 

 period. 



The negative coefficient of correlation indicates that an increase in number 

 of broody periods is accompanied by a decrease in their average length. .V 

 reduction in number of periods would therefore be accompanied by an increase 

 in their length. That this relationship is far from absolute is shown by the 

 magnitude of the correlation coefficient. Certainly the time lost in non-pro- 

 duction has been very significantly reduced by decreasing the number of 

 broody jieriods, as talile 1 shows. 



23. Correlation Betxceen Winter Rate and Annual Rate — Pullet Year. 



The records for 2212 individuals both broody and non-broody are available 

 for study. This relationship is important because both rates bear a rj.ther 

 intimate relation to egg production. The fact has previously been pointed 

 out that broody birds tend to be more intense winter layers than are non- 

 broody birds, but that the former are likely to carry a lower annual rate. An 

 intimate correlation between winter rate and annual rate would suggest that 

 rate of laying for the year may be predicted from the winter rate. Constants 

 calculated are as follows: — 



Number of liirds ....... 2242 



Mean winter rate ....... (i6.41 



Winter rate standard deviation .... ±9.38 



Mean annual rate ....... 56.46 



Anmuil rate standard deviation .... ±9.85 



Coefficient of correlation +.4900 ± .0108 



The above constants indicate a slightly greater relaLive standard deviation 

 in annual rate than exists for winter rate. Such a condition liiight be sur- 

 mised from the fact that broodiness and complete molt may lioth aifect annual 

 rate but for the most part are not concerned in winter rate. 



A rather intimate correlation exists between winter and annual rate. Evi- 

 dently those birds above the average in winter rate would be expected to be 

 above the average in annual rate. The practice of selecting for high winter 

 rate is without doubt sound from the standpoint of securing high annual rate. 



24. Correlation Between the Presence of Broodiness and Winter Production 

 above the Mean of Broodies and Non-broodies Combined — Pullet Year (Flocks 

 1916-1923). 



The absolute correlation between the presence of broodiness and winter pro- 

 duction above the mean of all birds is of much concern to poultrymen striving 

 for high winter records. Such information will show whether or not broody 

 birds tend to lay more eggs before March first than do broody-free birds. In 

 section 8 some evidence is presented to indicate that broody birds do actually 

 lay at a slightly higher rate than non-broodies when they are laying; but late 

 r— npl maturity, winter pause and the occasional winter broody period mav 



