594 
Journal oj Agricultural Research 
Vol. XXIX, No. 12r 
an average length of 148 days, and in¬ 
clude for each trial, and for each cow, 
expressed as a daily average: (1) Milk 
yield; (2) per cent fat (3) nutri¬ 
ment daily; (4) nutriment for main¬ 
tenance; and (5) nutriment for pro¬ 
duct. The “ nutriment daily” (cor¬ 
responding to milk production require¬ 
ments as used here) was estimated from 
the weights and analyses of the feeds 
consumed by the cow. The “nutri¬ 
ment for maintenance” was estimated 
from the live weight of the cow, allow¬ 
ing 7.92 pounds of nutrients per day 
per 1,000 pounds live weight (4, p- 10). 
The “nutriment for product” (corre¬ 
sponding to lactation requirements as 
used here) was estimated as “nutri¬ 
ment daily” minus “nutriment for 
maintenance.” Haecker’s data may be 
reworked and rearranged to better serve 
the present purpose. From his data 
the writer has calculated the lactation 
requirements per pound of milk and 
quite diverse also in breeding—Ayrshire,. 
Brown Swiss, Guernsey, Holstein, 
Jersey, Shorthorn, crossbred, and na¬ 
tive. 
The mean nutrients requirement for 
lactation per pound of milk for each of 
the fat percentage classes has been 
derived from Table I and the data are 
shown graphically in Figure 1. It is 
evident at once from the figure that the 
nutrients for lactation may be ex¬ 
pressed as a linear function of the fat 
percentage. We have now to deter¬ 
mine a value for K L , in a curve of the 
of the equation Nl=Kl (2.66 + 0? f° r 
these particular data and see how its 
slope conforms to the observed results. 
The curve is simply fitted so that the 
algebraic sum of the deviations is zero, 
and this gives a value of 0.049 for Kl* 
The calculated values are given in 
Table II and graphically in Figure 1. 
The agreement with the observed 
values is very close indeed. 
Table I.— Correlation of the variables , percentage fat content of milk and nutrients 
for lactation per pound of milk a 
r=0.6529±0.0327 
a Data from Minn. Bui. No. 140. 
b These two records are from the same cow (Lou II), stated in the text (p. 20) to have been out of con¬ 
dition at the time. It has seemed advisable to exclude these records in the computations. 
correlated this result with the per cent 
of fat in the milk. Table I shows the 
correlation surface. 
The coefficient of correlation, 0.6529 
±0.0327, is perhaps not so high as one 
might expect to find in a relation of 
this kind. It must be borne in mind, 
however, that there are several other 
variables involved here, such as varia¬ 
bility in the estimated consumption of 
nutrients as compared with the actual 
(including loss of body tissue); and the 
estimated maintenance requirements 
as compared with the actual (including 
gain of body tissue). The 46 cows used 
in Haecker’s experiments ranged in age 
from 2 to 14 years, and in weight from 
593 to 1,347 pounds. They were 
Table II. — Lactation requirements per 
pound of milk as affected by percentage 
fat content of milk * 
Per cent of 
fat in milk (t) 
Number of 
observa- j 
tions i 
Nutrients per pound of 
milk 
Observed 
values 
Calculated 
values, Nl 
[0.049 (2.66 
+0] 
2.55. 
4 
0.253 
0.255 
3.05.. 
14 
.286 
.280 
3.55_ 
23 
.310 
.304 
4.05. 
14 1 
.333 
.329 
4.55.- 
31 ! 
.333 
.353 
5.05. 
28 1 
t .380 
.378 
5.55_ 
23 
. 409 
.402 
6.05.. 
3 
; .435 
.427 
« Data from Minn. Bulletin 140. 
