Comparison of methods of estimation of lactation yield from monthly records
DOI:
https://doi.org/10.1590/S1678-3921.pab1979.v14.16663Keywords:
milk recording, selection, milk production, repeatabilityAbstract
Four methods of estimation of lactation yields were compared: Brazilian official method (Q); test interval method (T); T corrected for length of inicial interval (TI) and for initial and final intervals (TIF). Actual production (R =sum of daily yields) was used to measure errors of estimation. For 423 lactations (176 cows) corrected for environmental effects, correlation coefficients between actual and estimated production, by each of the Q, T, TI and TIF methods, were respectively: 0.98, 0.95, 0.95 and 0.94; and repeatabilities, 0.47, 0.46. 0.46 and 0.44 (standard error = 0.05). Cows were ranked by their MPPA calculated by each method. Rank correlation coefficients with actual rank were 0.98, 0.95, 0.96 and 0.95, for Q, T, TI and TIF, respectively. The ranking of six bulls on their predicted difference was not much affected by the method used for estimation.