Four methods of measuring daily sperm production (DSP) were evaluated using seventy-two boars. The DSP was determined; I, by unilaterally castrating boars 0 and 48 hr after depletion of the epididymal sperm reserves and then measuring changes in the reserves (fourteen boars) ; II, by measuring changes in the epididymal sperm reserves of four groups of boars killed 5, 24, 52 and 72 hr following depletion of the reserves (forty-six boars) ; III, by dividing the total number of spermatozoa in the right and left epididymis of sexually active boars by the epididymal transit time (twelve boars) ; IV, by quantitative testicular histology (same twelve boars as for Method III). Method I was unsuitable for estimating the DSP of individual boars because of normal variation in sperm numbers between epididymides within boars. Method II revealed that there was too much variation in epididymal sperm reserves among boars to obtain a reliable mean estimate of DSP for the forty-six boars. Method III did not give reliable results for individual boars because epididymal transit time varies among boars. It was concluded that the DSP of individual boars could best be measured by Method IV though Methods III and IV gave similar estimates for the mean DSP of the twelve boars (15·0 × 109 versus 16·2 × 109). The sperm output of these boars averaged 15×5 × 109/day when ejaculates were collected at 72-hr intervals, suggesting little or no absorption of spermatozoa from the epididymides or excretion of spermatozoa in the urine of sexually active boars. The mean DSP/g of net testis was 24·5 × 106 (S.E.±0·3) as determined by quantitative histology. Only spermatozoa from the cauda epididymidis were ejaculated during depletion.