Pregnancy failure during placentation in lactating dairy cows was associated with low concentrations of serum progesterone. Beef cows have greater serum progesterone and less pregnancy failure. Experiment 1 determined that reduction of serum progesterone affected late embryonic/early fetal loss in suckled beef cows. Cows (n=40) received progesterone from two new or used controlled internal drug releasing devices, replaced every 5d, beginning on Day 28 of gestation (mating=Day 0); CL were enucleated on Day 29. Retention of pregnancy was 77% in treated cows and 97% in 78 control cows (P<0.05). Experiment 2 determined how pregnant, lactating dairy cows with high or low progesterone concentrations during Days 28–34 differed in luteal function or in serum progesterone during replacement therapy. Luteal tissue from such cows was assayed for progesterone and expression of mRNA for genes of endothelin and prostaglandin (PG) systems. Secretion of progesterone and prostaglandins by dispersed luteal cells was determined during incubation with LH, endothelin-1, or arachidonic acid. Neither luteal progesterone nor mRNAs for endothelin or prostaglandin systems differed. Endothelin-1 inhibited secretion of progesterone more (P<0.05) in luteal cells from cows with low versus high serum progesterone, when incubated with arachidonic acid. Secretion of prostaglandin F 2 α was increased and that of 6-keto-PGF 1 α decreased by endothelin-1 in vitro. Serum progesterone during replacement was lower (P<0.05) for cows with low than high serum progesterone at lutectomy. Thus, clearance, more than luteal production, determined peripheral progesterone in pregnant, lactating dairy cows.