104 
POUL TP J ’ SCIENCE 
While the correlations between the individual months produc¬ 
tion and the annual total is the constant of greatest economic 
importance, since it furnishes the foundation of the equations for 
the prediction of annual production from the record of minor 
periods, it is somewhat misleading physiologically. In corre¬ 
lating between the record of an individual month and the record 
of the year as a whole we are determining the relationship be¬ 
tween the record of a minor period and of a major period in 
which the record of the individual month is one of the compo¬ 
nents. We have, therefore, determined as in an earlier paper 
(Harris, Blakeslee and Kirkpatrick, 1918, p. 44-45, diag. 1) the 
correlation between the record of the individual months and the 
total record of the other 11 months of the year. 
The coefficients showing the relationship between the produc¬ 
tion of the individual months and the eleven other months of the 
year appear in the second correlation column of table 1 for the 
first year and table 2 for the second year. These coefficients are 
represented graphically by the circles bordering the shaded areas 
in the two figures of diagram 1. The differences between the 
two coefficients and the percentage reduction in the correlation 
resulting from the elimination of the individual month used as 
first variable from the annual total is shown in the two final col¬ 
umns of tables 1 and 2. The difference is also represented graph¬ 
ically by the breadth of the shaded area in diagram 1. The 
actual magnitude of the reduction may be seen from the differ¬ 
ences or from the relation of the shaded to the unshaded area in 
the figures of diagram 1. It may be noted that the correlations 
for the eleven mouth periods follow very closely indeed tho ;e for 
the twelve month periods. 
We now have to compare the correlations between the indi¬ 
vidual months and the annual totals and the individual months 
and the eleven month totals as deduced from the first year and 
the second year performance of the birds. The differences 
deduced from tables 1 and 2 appear in table 3. 
These differences show that the correlations between the indi¬ 
vidual months production and the total production of the remain¬ 
ing eleven months are higher for the first year in November, 
December, January and February, whereas they are generally 
lower for the first year in March to October. A number of these 
differences may be considered clearly significant in comparison 
