Changes in Milk Production with Age and Milking Frequency

Summary 1.Multiplicative age-correction factors for Holsteins were derived from 43,573 Herd Improvement Registry lactations of 5,374 cows milked twice daily and 5,627 cows milked thrice daily. These factors were based on comparisons of production by the same cows in consecutive lactations. The factors are shown in table 2. 2.The biases which culling of cows may cause in age correction factors are discussed. Several sets of factors for Holsteins are compared in figure 2. Figure 3 shows the actual HIR averages before and after correction of these factors. 3.The present findings agree well with the widely used Kendrick factors. The differences indicate that the following four points need attention whenever a revision of the Kendrick factors is made: (a)The Kendrick factors seem a bit too large at ages under 35 mo. Part of this might come merely from the higher persistency of heifers, since the present data included many lactations which went as long as 365 days. Part of it may have come from neglecting the difference between late first calvers and early second calvers in the smoothing when the Kendrick factors were prepared. (b)Some difference needs to be made between late first and early second calvers at those ages when records by both are present; i.e., at around 33 to 42 mo. This might be done rather simply by having a sharp discontinuity in the standard age-correction factors between 34 and 35 mo. and adding about 0.05 to the standard factor if the calving is a second one which occurred earlier than 35 mo., but subtracting about 0.04 from the standard factor if it is a first calving which occurred at ages later than 34 mo. (c)The Kendrick factors appear a bit too low at ages around 45 to 66 mo. The apparent difference is small but does have a little importance, because so many of the records on which cows are culled or are chosen to be the dams of herd sires are made at these ages. (d)The Kendrick factors seem decidedly too large at the older ages. 4.Because all of the commonly used age corrections remove most of the variance caused by age, their accuracy can be increased only a little by further changes. Since any changes involve much confusion and clerical labor, those which the present study indicates to be desirable should first be verified on other data, especially with respect to these points: (a)Is any material portion of them the peculiar result of the present data having included many lactations as long as 365 days? (b)Is the slope of the age curve really different enough in HIR and in DHIA data to need attention? (c)Is the explanation proposed here for the peculiar shape of the R curve at around 36 mo. really valid and is there any simpler and equally accurate device for adapting age corrections to it ? 5.Multiplying 3×records by 0.80 when they are first lactations and by 0.833 when they are later lactations is suggested for correcting them to the basis of twice-a-day milking. The data hint that the difference between 2×and 3×milking is even larger than 4: 5 at extremely young ages.