Ovarian ultrasonography and measurement of circulating concentrations of anti-Müllerian hormone (AMH) have been used to estimate follicle number and predict fertility in mammalian females, but no study has evaluated follicle number and circulating concentrations of AMH in ewes known to differ in fertility. We tested the hypothesis that ewes that had failed to lamb (BARREN) in four consecutive annual breeding seasons of 21–35 d have fewer follicles and diminished circulating concentrations of AMH compared to closely matched ewes that consistently produced lambs (FERTILE) under the same breeding protocols. Once identified, BARREN ewes (n = 19) were paired by breed and sire to a FERTILE control ewe (n = 19) and reproductive tracts were recovered at necropsy. Visible antral follicles in both ovaries were counted and a representative cross section of one ovary was embedded for histological evaluation of pre-antral follicle numbers. Paired t-tests indicated that BARREN ewes had fewer antral follicles, fewer primordial follicles, and decreased circulating concentrations of AMH compared to FERTILE ewes (P ≤ 0.01), but there were ewes in each fertility classification that had ovarian phenotypes like the opposite fertility classification. The best technologies we have currently for estimating follicle numbers are ultrasonography and measurement of circulating concentrations of AMH, but no single technique is perfect for predicting fertility. A better understanding of the under-lying biological mechanisms linking AMH, folliculogenesis, and fertility is required to improve the use of measurements of follicle number for predicting fertility in livestock.