74 TECHNICAL BULLETIN 7 



mean annual rate are represented liy In-oken lines. Tlie groups of birds used 

 in making the two charts are not identical because winter rate records are 

 availal)le on a considerable number of individuals that did not survive to com- 

 plete annual records. However, the two groups are so nearly identical that 

 the mean percentage of broody birds closely agrees in the two charts. 



The increase in mean winter rate from 1916 to 1923 is 4.06, while the increase 

 for annual rate in the same period is 6.00. This fact indicates that annual 

 rate has increased more rapidly than winter rate as the percentage of broody 

 birds has been reduced from year to year. The greater degree of parallelism 

 in the two graphs on chart 1 suggests tliat a change in percentage of liroody 

 birds is usually accompanied by a change in winter rate. Chart 2 shows a 

 lesser relationsliip between percentage broody and annual rate. 



In general the two charts furnish evidence that lioth mean winter rate and 

 mean annual rate may be increased while the percentage of broody birds is 

 being reduced. The lowering of the percentage of broody birds to at least 30 

 per cent, as has been accomplished in the flock studied, appears to be advan- 

 tageous from the standpoint of annual jiroduction. 



Tiie next section is devoted to a study of the relation between the number 

 of broody periods and the mean length of broody periods. It seems desirable 

 to ascertain if the average length of broody period is affected by the numl)er 

 of periods. Does the frequency of onset of broodiness tend to shorten or 

 lengthen the period? The coefficient of correlation is again made use of and 

 the number of broody periods is tabulated against the mean length of period, 

 using 1135 birds that were broody in the pullet year. 



22. Correlation Beticeen Times Broothj and Mean Lenath of Brooihi 

 Periods — Pullet Year. 



Any attempt to decrease the intensity of liroodiness must be accomplished 

 either by reducing the number of periods or by reducing the length of these 

 periods. The coefficient of correlation is here calculated to discover a possible 

 relationship between number and length of broody periods. Constants calcu- 

 lated are as follows: 



Number of birds ....... 113.5 



Mean times broody 2.S9 



Times broody standard deviation .... ±3.67 



Mean length of periods 15.10 



Length of periods standard deviation . . . ±3.78 



Coefficient of correlation —.2338 ± .0189 



Regression times broody on length . . . — .4620 



Regression length on times broody . . . — .1183 



The total number* of birds showing one or more broody periods is slightly 

 greater than the number in sections 12, 21, 30 and 32, broody records being- 

 available on a few birds on which annual rate records are lacking. The stand- 



"The total number of hirds goin^ bi-oody was 1135. Of this group, 1017 individuals 

 were first broody after March first so that the actual length of the period of non- 

 production attributable to broodiness could be definitely recorded. There were 118 

 birds broody before March first. The mean lensjth of broody period for the 1017 

 birds is 1.5.95 days, while that for the grou]) of 1135 birds is 15.10 days. This slight 

 difference in mean length of period is not significant and may be attributed to our 

 inability to separate broody pause from winter pause in those 118 birds going broody 

 before March first. The method of allowing a bird but four days to begin laying after 

 her return to the laying house following broodiness during the winter season is faulty 

 in that it actually assigns a shorter broody period during winter than the mean of 

 summer broody periods. 



