Estimation of Some Variance Components of Bovine Semen Criteria and Their Use in the Design of Experiments

Open Access
Publisher Website
Google Scholar
Abstract
A year''s semen production records of 68 bulls from 5 dairy breeds in 4 studs were used to estimate the bull ejaculate, breed, season, and interaction variance components of 5 criteria used to measure semen production. The means and standard deviations of individual observations for the 5 criteria were: ml of semen per ejaculate, 8.0 [plus or minus] 2.4; per cent motile sperm, 66 [plus or minus] 9.0; billions of sperm per ml of semen, 1.71 [plus or minus] 0.62; billions of motile sperm per ejaculate 8.9 [plus or minus] 3.9; and billions of total sperm per ejaculate, 13.6 [plus or minus] 6.0. Component of variance analyses showed that the bull and ejaculate components accounted for more than 90% of the total variance in all criteria except per cent motile sperm. Consequently, these 2 variances were used in estimating the power of the test by Tang''s method. This estimate showed that there was little increase in sensitivity to be gained by taking more than 10 ejaculates per bull. It was calculated that to have a 75% chance of detecting a threatment difference of 50% of the mean of motile sperm per ejaculate at the 5% level of significance would require about 10 bulls per treatment and 5 ejaculates per bull. The final choice of power of the test, level of significance, treatment difference to be detected, and number of bulls and ejaculates to be used depends upon resources available and upon the value of the detectable difference to the experimenter.