108 patients with endemic nontoxic goitre have been treated in the field with thyroxine (T4), triiodothyronine (T3), and potassium iodide (KI), singly or in combination, or with placebo. After 6 months of continuous treatment, goitre size decreased significantly in the 7 actively treated groups, but not in the one treated with placebo. The combination of 150 micrograms T4 + 150 micrograms KI daily seemed the most effective treatment, both clinically and by its suppression of the 131I uptake and the TSH response to TRH, followed by 100 micrograms T4 + 25 micrograms T3 or 200 micrograms T4, but the difference from the other groups was not statistically significant. The increase in the pulse rate (PR) and the shortening of the photomotogram of the Achilles tendon reflex (PMG) were taken as indices of thyrotoxicity and side-effects of the treatment. There was no significant difference in the side-effects between any two of the active groups if effectiveness was also taken into account. The decrease in goitre size was not correlated to either the final serum T3 value achieved at the end of the treatment, or the thyroidal 131I uptake. There was, however, a weak but significant correlation between the decrease in goitre size and the TSH response to TRH. This casts some doubt to the concept that thyroid hormones decrease goitre size solely by suppressing the pituitary TSH release. Of the 30 patients treated with KI singly or in combination and studied in this respect, 8 developed autoantibodies against thyroglobulin and/or the thyroidal microsomal antigen compared to 1 of 22 not receiving KI (P = 0.08).(ABSTRACT TRUNCATED AT 250 WORDS)