Two experiments were conducted to investigate the level of fertilization failure that was attributed to oestrous ewes being mounted and marked by the harnessed ram but where service accompanied with ejaculation had not occurred. In ewes marked by harnessed rams, the occurrence of failure of service with ejaculation was determined by examining a 2 cm cube polyurethane sponge which was placed in the vagina before mating. The vaginal sponge was removed and examined at 0 to 24 hours after the onset of oestrus. In experiment 1, a second sponge was inserted at the time of removal of the first sponge and removed 24 hours later. The presence of spermatozoa in the vaginal sponge was taken as evidence of service. The service rate was determined by this method in one half of the flock. The fertilization rate (experiment 1) or the proportion of ewes that returned to oestrus following natural mating (experiment 2) was determined in the remainder of the ewes. In experiment 1 the service rate, as determined by examination of the sponges, was 86 per cent and the fertilization rate was 88 per cent. In experiment 2 the service rate was 94 per cent and the non-return to service rate was 94 per cent. The vaginal sponge reduced the number of 1 1/2 year old maiden ewes exhibiting oestrus but had no influence on the parous ewes. It was concluded that lack of service with ejaculation in oestrous ewes may be the major cause of reproductive failure in flocks where the levels of failure of fertilization are low.