lUiOODlNKSS AM) KKCrXDITV 1\ rowi, (I., 



Xiiiiilicr ol' liirds . . . . 



MiMii times l)n)(Kly 

 J'iiiies l)r()t)dy staiulaiil tU'\ iatioii 

 Mean winter rate . . . . 

 A\'iiitVr rate standard deviation 

 Cuettic'ient of correlation 



2221 

 l.VA 

 rtl.f)!) 



±9M7 

 +.0706 ± .0142 



Tlu' al)o\e I'onstants sliow tlie mean winter rate to be greater tlian tiie mean 

 December rate ])reviously calculated. The above winter rate really signifies 

 that, on the average, the birds laid 66.45 per cent of the maximum possii)le 

 number of eggs when they were laying, since all pauses of four or more days 

 have been deducted in calculating winter rate. The standard deviation in 

 winter rate is only ± 9.37 compared with a figure of ± 20.40 for I)eceml)er 

 rate. The winter pause and the fact that many of the birds actually lay their 

 first egg during December account for the wider variability in December rate. 



The coefficient of correlation between times broody and winter rate is almost 

 identical with that between times broody and December rate. This is a con- 

 stant of small magnitude, and is a fal.se correlation because the population is 

 made up of both broody and non-broody birds. 



7. Correlation Beticeen Times Broody and Winter Rate for Broodt/ Birda 

 .Hone — Pullet Year. 



In order to ascertain any possible relationship between winter rate and 

 degree of broodiness, the correlation between times broody and winter rate 

 has been calculated for broody birds alone. The constants obtained are as 

 follows: — 



Xum!)er of birds 1098 



Mean times broody 2.89 



Times Itroody standard deviation .... ±1.93 



Mean winter rate ....... 67.57 



Winter rate standard deviation .... ±9.63 



Coefficient of correlation —.0314 ± .0203 



Regression l)roodiness on rate .... — .0063 



Regression rate on broodiness .... — .1564 



The mean winter rate in those iiirds that actually went l)roody during their 

 pullet year is 67.57 compared with 66.45 for broodies and non-broodies com- 

 liined. Such a difference is of no significance. 



The coefficient of correlation is negative. Its small magnitude, together 

 with the size of its probable error, leads to the assumption that there is ab.so- 

 lute independence between winter rate and degree of broodiness as measured 

 liy times broody. 



.S'. Correlation Between the Presence of Broodiness and Winter Rate ahove 

 the Mean of Broodies and Non-Broodies Combined — Pullet Year. 



Tlie absolute correlation between the presence of broodiness and higli rate 

 is of importance to the breeder. Such a constant will indicate whether or not 

 the broody trait carries with it higher winter intensity than does the non- 

 l)roody trait. The coefficient of correlation is calculated below according to 

 Yule. 



