Relative accuracy of various methods of estimating milk yield from sample day production in crossbred cows.

摘要

168 crossbred (Holstein-Friesian X Hariana) cows were maintained from 1969-85 at the Department of Animal Breeding, Chaudhary Charan Singh Haryana Agricultural University, Hisar, India. Cows were milked twice a day at approximately 12 h intervals from the 5th day of calving to the end of lactation (between 200 and 300 days), and samples were combined to give the actual lactation yield. Three methods of estimating milk yield: centring date (CDM), test interval (TIM) and sample day production (SPM) were assessed for accuracy. For CDM, samples were taken at regular intervals of 30, 40, 50 and 60 days. For TIM and SPM, samples were taken irregularly - the duration of each test interval being chosen randomly. There were 5 sampling plans: AM+PM (both morning and evening), all-AM (only mornings), all-PM (only evenings), AM-PM (mornings and evenings alternately) and PM-AM (evenings and mornings alternately). The average error in estimated milk yield for each method is tabulated without adjustment, with adjustment for the shape of the lactation curve (SA) and with adjustment for both SA and diurnal variation (SDA). Adjustments reduced the average error in all 3 methods, and more so when the shorter sampling intervals (30-40 days) were used. Varying the sampling interval did not affect the average error. The PM-AM sampling plan gave an average error closest to that of the AM+PM plan. There were no significant differences in relative efficiency between the 3 sampling schemes, but relative efficiency decreased as the sampling interval increased for all 3 schemes. Adjustments could improve the efficiency by no more than 1%.