Scrotal circumference in living Holstein bulls was found to be a better measure of testis size than various linear measurements. This variable was easy to measure, was highly repeatable (r=0.98) when measured by different investigators and was highly correlated with testis weight (r=0.92). The regression equation for estimating testis weight (Ŷ) from scrotal circumference (X) in mature bulls was Ŷ − 1298.5+50.2X. Scrotal circumference (Ŷ) and age of bulls in mo. (X) showed a curvilinear relationship described by the multiple regression equation Ŷ=−7.65+5O.1O log X−12.44 (log X)2. Semen was collected twice each Tuesday and Friday for 4 weeks from groups of Holstein bulls (10 bulls per group) which were 17 to 22, 34 to 42, 42 to 53 and 59 to 69 mo. of age. In addition 15 proven bulls over 72 months of age, varying greatly in physical condition, were placed on the same schedule after an initial period of depletion. Average sperm output per week in order of increasing age group was 28.4, 41.5, 38.0, 42.7 and 28.2 × 109. These values are not corrected for any loss in the collection equipment. Corresponding values for sperm output per week per gram of testis were 61.2, 58.5, 53.6, 52.8 and 34.4 X 106. These data strongly suggest that there is a reduction in spermatogenesis per unit of testis in old bulls, and indicate the desirability of capitalizing upon the large spermatogenic potential of sires 2 to 6 yr. old. Correlations between scrotal circumference and sperm output per week were 0.81,0.72, 0.64, 0.40 and −.22, respectively. The youngest group of bulls was ejaculated a year later and the correlation between scrotal circumference and sperm output per week was 0.69. The correlation between scrotal circumference measurements in consecutive years was 0.91. These results indicate that scrotal circumference is a useful indicator of potential current sperm output and subsequent testicular development in young bulls. However, it appears to be of little value in bulls more than 5 to 6 years of age, where other changes in the testes apparently mask the relationship of testis size to total spermatogenic potential.