Crossbred ewes (1/2 Dorset, 1/4 Rambouillet, and 1/4 Finnsheep) from a flock being selected for spring fertility, defined as ability to lamb following ram exposure in May and June in Virginia (37 degrees N latitude), were used to study the duration of the seasonal anestrus. In the first 3 yr of the study (1992, 1993, and 1995), mature ewes were divergently selected based on EBV for fertility, and the duration of anestrus was measured by continuously exposing the ewes to vasectomized rams equipped with marking harnesses from mid-January until approximately August 1. Only ewes that had lambed in the previous fall were used to ensure that ewes were in a comparable physiological state, and the same rams remained with the ewes in each year to avoid induction of estrus by introduction of novel rams. The duration of anestrus in high-fertility ewes (n = 26; mean fertility EBV of 12.6%) was 28.4 d, which was significantly less than the 70.2 d of anestrus observed for low-fertility ewes (n = 15; mean fertility EBV of .3%). Five high-fertility ewes did not exhibit a period of anestrus. The regression of number of days of anestrus on fertility EBV was -2.15 +/- .72 d/%. In yr 4 (1997), 11 high-fertility and two low-fertility ewes were evaluated. None of these ewes exhibited a clear seasonal anestrus; six unequivocally cycled continuously. Between January 23 and July 31, the mean duration of anestrus for these ewes was only approximately 11 d. The duration of anestrus for high-fertility ewes seems to be the shortest reported for temperate sheep breeds.