The ovaries of 59 pluriparous cows of unknown reproductive history were palpated, scanned and dissected on the day of slaughter to compare the accuracy of rectal palpation and transvaginal ultrasonography with a 5 MHz linear array for the detection of corpora lutea and follicles. The rectal palpation was first carried out to judge the presence of follicles of more than 5 mm diameter, and corpora lutea which were classified as young (days 1 to 4), mid-cycle (days 5 to 16) or old (days 17 to 21) according to morphological criteria. The cows were then examined for follicles and corpora lutea by ultrasonography and the corpora lutea were again classified directly as young, mid-cycle or old according to their appearance. The cows were then slaughtered, their ovaries dissected, and the follicles over 5 mm in diameter were counted and the corpora lutea were classified in the above mentioned age categories. For the detection of a mid-cycle corpus luteum the sensitivity and predictive value of rectal palpation were, respectively, 83.3 per cent and 73.2 per cent and for ultrasonography the sensitivity and predictive value were 80.6 per cent and 85.3 per cent, respectively. However, both techniques were inaccurate for the detection of young and old corpora lutea. For detecting follicles ultrasonography was a significantly better method than rectal palpation. Ultrasonography detected 95 per cent of follicles larger than 10 mm whereas rectal palpation detected only 71 per cent of these follicles. Both techniques failed with follicles 5 to 10 mm in diameter; only 21.5 per cent were detected by rectal palpation and 34.3 per cent by ultrasonography.