The luteolytic effects of exogenous prostaglandin F2alpha (PGF) that did and did not simulate natural 13,14-dihydro-15-keto-PGF (PGFM) pulses were studied during mid-diestrus in 42 Holstein heifers. Plasma concentrations of PGF were assessed by assay of PGFM. In experiment 1, a single intrauterine injection of 4.0 mg of PGF into the uterine horn ipsilateral to the corpus luteum resulted in a precipitous progesterone decline, whereas sequential injections of 0.25 or 1.0 mg every 12 h resulted in a stepwise decrease (P < 0.05) following each injection. A progesterone increase occurred during the first 5 min before the luteolytic decrease but only for the 4.0-mg dose. From the results of experiment 2, a 2-h intrauterine infusion of a total of 0.5 mg of PGF was judged to best simulate a natural PGFM pulse. In experiment 3, simulation of sequential pulses at 12-h intervals resulted in a continuous precipitous decrease in progesterone to <1 ng/ml by the beginning of the fourth simulated pulse. In contrast, a single simulated pulse resulted in a 6-h progesterone decrease to a constant concentration for 3 days after treatment, followed by a return to control concentrations. The mean +/- SEM interval between the pretreatment and posttreatment ovulations was shorter (P < 0.05) in the group with sequential simulated pulses (14 +/- 1 day) than in the group with a single pulse (21 +/- 1 day). Results indicated that excessive PGF doses may stimulate nonphysiologic progesterone responses and supported the hypothesis that sequential PGF pulses are required to stimulate natural luteolysis in cattle.