Plasma progesterone concentrations in mares were determined in two experiments during the time that the luteal glands were detectable by transrectal ultrasonography. In both experiments, corpora lutea were classified into two types of morphologies based on their ultrasonic appearance: centrally nonechogenic luteal glands (fluid-filled) and uniformly echogenic luteal glands (non-fluid-filled). In Experiment 1, daily blood samples were taken from horse mares during August through October and May through July. There were no significant effects of season or luteal morphology on progesterone concentration. There was a significant main effect of day, but no day-by-season or day-by-morphology interactions. Progesterone increased significantly between Days 1 and 3 (mean progesterone concentration, 2.5 vs 5.2 ng/ml, respectively), between Days 3 and 4 (5.2 vs 7.8 ng/ml), and between Days 4 and 5 (7.8 vs 11.0 ng/ml). Progesterone did not decrease significantly until between Days 11 and 15 (11.6 and 6.1 ng/ml). Subsequent decreases occurred between Days 15 and 16 (6.1 vs 3.9 ng/ml), and Days 16 and 17 (3.9 vs 2.5 ng/ml). In Experiment 2, blood samples were obtained from pony mares at 1 2-h intervals for 3 h before and 2 h after the defined onset of luteal development (end of evacuation of the ovulatory follicle). Additional blood samples were taken at 5, 8 and 12 h after the onset of luteal development, and thereafter at 12-h intervals for 5d. There were no significant differences between centrally nonechogenic luteal glands (n=7) and uniformly echogenic luteal glands (n=5) during the first 5 d of luteal development. There was no time-by-morphology interaction, but there was a significant time effect. The first significant increase in progesterone concentration occurred between Hours 12 and 24 (0.5 vs 1.1 ng/ml). Additional increases were detected between Hours 24 and 36 (1.1 vs 2.6 ng/ml), Hours 36 and 48 (2.6 vs 4.3 ng/ml), Hours 48 and 60 (4.3 vs 6.1 ng/ml), Hours 60 and 72 (6.1 vs 9.4 ng/ml), and Hours 72 and 96 (9.4 vs 13.8 ng/ml). The hypothesis was supported that fluid-filled corpora lutea do not differ from non-fluid-filled corpora lutea with regard to progesterone production.