The aims of this study were to determine the plasma concentrations of follistatin in rams and to assess if the testis contributes to circulating follistatin and if there is uptake or production of follistatin by the head in rams. Catheters were inserted in the carotid artery, jugular vein and spermatic vein of intact rams during the non-breeding season (experiment 1; n = 5) and breeding season (experiment 2; n = 4). In experiment 1, blood samples were collected from 5 rams every 10 min for 4 h, commencing 20-60 min after surgery. After 2 h of sampling 1 microgram gonadotrophin-releasing hormone (GnRH) was injected intravenously. In experiment 2, blood samples were collected from 4 of the rams used in experiment 1 by venipuncture 30 and 15 min before surgery and every 15 min throughout surgery. Commencing 1 h after surgery, matched samples were taken from each of the vessels every 10 min for 4 h (1-4 h after surgery), then every hour for 20 h (4-24 h after surgery) and then every 10 min for 4 h (24-28 h after surgery). In both experiments, follistatin secretion was non-pulsatile and there were no significant differences between the concentrations of follistatin in any of the vessels. There was a significant (P < 0.05) increase in the concentrations of follistatin in each of the vessels throughout the 4 h of 10-min sampling in both experiments. In experiment 2 plasma concentrations of follistatin in the jugular vein were significantly (P < 0.05) lower before surgery than at other stages of the experiment. During the non-breeding season (experiment 1) the concentrations of follistatin in all vessels were about 2-fold higher (P < 0.001) than during the breeding season (experiment 2). Concentrations of follistatin were measured in the testicular tissue of the ram, bull, monkey and rat and were found to be 13.6, 2.1, 2.5, 0.8 ng/g testis respectively. In experiment 3, blood samples were collected every 15 min for 4 h from castrated rams (n = 6) in the absence of treatment with testosterone propionate (TP) and after 7 days of treatment with a physiological dose of TP during the breeding and non-breeding seasons. There was no effect of stage of breeding season or TP on the plasma concentrations of follistatin and these concentrations in the castrated rams were similar to the concentrations in the intact rams in experiment 2. In experiment 4, the function of Leydig cells was stimulated by administration of human chorionic gonadotrophin but this had no effect on plasma concentrations of follistatin. These experiments show that the concentrations of follistatin in the plasma of rams are measurable, that the testis is not the major contributor to circulating follistatin and that there is no significant uptake or production of follistatin by the head in rams. It appears that the contribution of the testis to circulating follistatin may vary with the stage of the breeding season, being greater during the non-breeding season than the breeding season. The gonadotrophins and testosterone do not appear to have a direct effect on the secretion of follistatin in rams. The increase in concentrations of circulating follistatin during surgery and more frequent blood sampling suggest a stress-related effect on the production of follistatin.