Ab8tractThree methods of cannulating the uterus of the ewe are described and the patterns of secretion collected from the cannulated uteri are presented.The first procedure involved cannulation of the uterus near the uterotubal junction. When one horn was cannulated by this method and the cannulated horn ligated at the external bifurcation, no change in the length of the oestrous cycle or in the ovulation rate was observed. However, in some animals examined 60-80 days after cannulation, these treatments caused a slight increase in the height of the luminal and glandular epithelium of the uterus.There were large variations between ewes in the volume of fluid collected from the cannulated horn. In most ewes, a peak in flow was obtained close to oestrus and an average of 3 or 4 ml of fluid was collected during each oestrous cycle. However, larger volumes than average were obtained during the first oestrous period following ligation of the uterine horn at the external bifurcation. There was a considerable difference in the volume of fluid collected from each horn of the same animal. In animals with both the uterine horn and the fallopian tube cannulated, the volume of fluid collected from the uterus was less than that from the fallopian tube. In anoestrous ewes, secretion from one uterine horn averaged O· 12 mljday and showed little variation with time.In the second procedure, the uterus was cannulated by transection anterior to the cervix and insertion of a cannula into the transected end. Short oestrous cycles were frequently observed in these ewes after cannulation. The volume of uterine fluid collected varied greatly between animals but in most cases peaks in volume occurred at each oestrus.Uterine fluid was also collected from four multiparous ewes with a cannula inserted through the cervical canal. In these ewes, the oestrous cycles were generally shorter than normal and there were large variations between animals in the volume of fluid collected.Overall, cannulation via the uterine horn resulted in the collection of less fluid than if either of the other two methods were used. In all cases, the variation between animals was greater than the variation between cycles within the same animal. The physiological significance of secretion rates from cannulated uteri is discussed.