Estimating Milk Yield of Sows

Hourly milk yields were obtained for nine consecutive hours at d 14 of lactation from 92 crossbred (four breed) sows during their first and second lactations. Additionally, milk yields were obtained at d 14 and 20 of lactation from 175 mature Yorkshire X Landrace sows for nine consecutive hours. Each measurement (4,806 observations) was obtained by weighing the pigs before and after each hourly nursing. These data were examined with the objective of understanding the consequences of reducing the number of measurements obtained from each sow. Four data sets were formed from the measurements, namely, first parity crossbred sows, second parity crossbred sows, d 14 and d 20 milk yields from the Yorkshire X Landrace sows. Differences (P less than .01) among the hourly milk yield estimates found in all four data sets were largely accounted for by the low yield for h 1 and moderate reduction for h 2 compared with the means of nine measurements. There were no significant differences among the hourly yields for h 3 through 5 compared with h 3 through 9 in three of the four data sets. Mean hourly milk yield for first parity crossbred sows, however, increased as additional hourly measurements were taken. Using the mean hourly milk yield estimate of h 3 through 5 would give an unbiased estimate of h 3 through 9, except for first parity sows, and a standard error only 13% greater than the standard error for nine hourly measurements. The shorter procedure would save considerable time and labor.