LACTATION YIELDS AS MEASURES OF A COW'S MILKING CAPACITY.         241

factor, i.e., selective breeding coupled with rigid selection. Those who have studied
the evolution of these herds are aware of the large number of cattle culled in order
to bring these herds to their present level of milk production. To quote a single
instance, out of a total of 208 female calves born in the Agricultural College Dairy,
Lyallpur, from the time it was started, and excluding the 23 calves that are being
reared, only 78 were reared to maturity, and of those 36 culled (46 per cent.)
on account of poor production.

However, the age at which cows have been culled on account of poor produc-
tion has varied considerably at different dairy farms, and no uniform standard has
been followed. Thus, though at the College Dairy, Lyallpur, out of a total of
36 cows culled on account of poor performance 24, or 66 per cent. were culled
after milking for one lactation only, i.e., were selected very early, this task of
weeding out the inferior animals was considerably deferred at other places, such
as the Pusa Dairy Farm, where animals were usually tried for two or three lactations
generally before they were discarded. Evidently these two Dairy Farms have
assigned different values to the various lactation yields of a cow as a measure of her
milk-giving capacity, for though at Lyallpur the first lactation yield has been
reckoned as quite reliable, at Pusa it has been more often than not ignored.

It costs a huge sum, however, to rear a calf from birth to productive age.
It is, therefore, a waste of labour and capital to cull cows without giving them a
fair trial. It is equally unsound, however, to maintain poor producers in a herd
for any unnecessarily lengthy period. The question of determining the best age
for selection—age up to which a cow should be milked to know her true value as a
yielder —and of comparing the relative reliability of different lactation yields of a
cow as measures of her milking capability is, therefore, of some practical import-
ance. But so far as the author is aware, no results of any study of this nature
under Indian conditions have yet been reported. The present attempt has been
made to fill this gap.

                                               MATERIAL.

The material for this study has been derived from the same sources as for the
previous one [Lall Chand Sikka, 1931], i.e., milk records of the Sahiwal cows of
Ferozepore and Pusa herds; though only those cows have been included which
have been kept for at least 4 lactations. 194 such cases were available, and their
records—776 in all—form the basis of this study.

To eliminate from these records the differences in yield as caused by varia-
tions in service period, and consequently to get more uniform data, all of them
were first standardised to 85 days service period basis, by using the correction