244 THE INDIAN JOURNAL OF VETERINARY SCIENCE AND ANIMAL HUSBANDRY [III, III.

A study of the above Table shows that both skewness and co-efficient of
variation are lower in the case of maximum yield than in the mean first 3 or 4
lactations yield, which means that the maximum yield being less variable is a more
reliable measure of a cow's milking capability than her mean yield for several
lactations.

That the mean yield and maximum yield are definitely correlated is shown
in Graph I, in which each point in the dotted curve represents the mean
yield for first four lactations for the different cows, and that in the continuous
curve the maximum yield corresponding to this mean yield, the mean yields being
arranged in the ascending order according to their magnitude from left to right.
It will be noticed that except for a few exceptions here and there, the maxi-
mum yield goes on increasing regularly with the increase in mean yield.

From the correlation surface drawn for these 2 variables, the following cons-
tants were obtained:—

                   r

+0.8763 .0112

                   n

0.8889 .0102

                   n2r2

0.0222 .0144

As will be noted the value of r is very high. The value of n is slightly higher
than that of r, as these values generally are in case of linear regressions, and the
value of n2r2, compared with its probable error is very low and quite insignificant.
It is only 1.54 times its probable error, and therefore permits the use of a linear
equation to fit the data. This has, therefore, been done, and the following equation
obtained—

            y = 1040 +1.0358 x
                where y = maximum " corrected " lactation yield ;
                          x = mean "corrected" yield for first four
                                       lactations.

The fit of the regression line to the actual means is shown in Graph II. It
will be noticed that the fit is good—the line passing through the observational
points remarkably well.

By using the equation given above, the maximum " corrected " yield of a cow
can easily be determined from her mean " corrected " yield for first 4 lactations.
Thus if a cow gives a mean " corrected " yield of 2,000 lbs. in her first 4 lactations,
according to this equation, she will give a maximum of 3,112 lbs. in any one lacta-
tion. The variation around this expected milk yield which would include 50 per
cent. of the cows of 2,000 lbs. mean " corrected " yield for 4 lactations may be found
from the relation of the standard deviation of such a record.