It is well accepted that the C cells of the thyroid contain somatostatin, but the role in local endocrine function has not yet been firmly established in this organ, and it has not been proved that thyroidal somatostatin is released into the circulation. We have measured the contents of somatostatin-like immunoreactivity in the effluent of canine thyroid glands perfused without recirculation with a synthetic buffer medium. During basal conditions a definite release was consistently found in the order of 10 pg/ml corresponding to 12 pg/min. The somatostatin-like immunoreactivity was studied in dilution experiments and by gel-filtration chromatography, and found to have properties identical to those of synthetic cyclic somatostatin, which was also recovered quantitatively when added to sampling tubes. Various compounds were infused in concentrations that are highly active in pancreas perfusion experiments. 14-min infusion of arginine, 5 and 11.5 mmol/liter; isoproterenol, 10 and 23.7 nmol/liter and 68.7 mumol/liter; acetylcholine, 5 mumol/liter, carbamylcholine, 10 and 100 mumol/liter; glucagon, 1 and 30 nmol/liter; and porcine calcitonin, 1 and 100 ng/ml did not affect the basal release of somatostatin-like immunoreactivity significantly. Neither did an increase from the control level of 4 mmol/liter glucose of 10 or 20 mmol/liter, nor an increase in the control level of 4.4 mmol/liter K+ to 7.5 or 14.4 mmol/liter. Each of these compounds were tested in three or four dogs. The effect of an increase in Ca++ from the control level of 1.5 mmol/liter to 2.25, 3.0, and 4.5 mmol/liter was tested in random order in five thyroid lobes. All three doses elicited an immediate increase in effluent somatostatin-like immunoreactivity. In most experiments the response was biphasic with an early spike, followed by a stable level that was maintained during prolonged Ca++ infusion. The secretory response was not diminished through a series of repeated short pulses of calcium infusion. The response to 3.0 mmol/liter Ca++ (control period 8.4 +/- 1.5, test period 337 +/- 110 pg/ml, mean +/- SE) and 4.5 mmol/liter Ca++ (control period 9.5 +/- 1.4, test period 386 +/- 125) were significantly higher than 2.25 mmol/liter Ca++ (control period 7.2 +/- 1.0 test period 140 +/- 39), while there was no significant difference between responses to the two high doses. Infusion of salmon calcitonin, 10 ng/ml and 1 microgram/ml; or porcine calcitonin, 1 microgram/ml during calcium stimulation (2.25 mmol/liter of Ca++) did not induce alterations in the release of somatostatin-like immunoreactivity. The results demonstrate that thyroidal somatostatin is mobilizable, and it appears to be selectively sensitive to calcium stimulation, indicating a possible role in calcitonin release control.