Although thyroxine (T4) 5'-deiodinase activity is diminished in liver homogenates of starved rats, no information is available regarding the effect of starvation on net T4 to triiodothyronine (T3) conversion in the intact rat. It appeared important to clarify this relationship since rat liver homogenates are widely used as a model for the study of the factors responsible for reduced circulating T3 in chronically ill and calorically deprived patients. In contrast to the expected selective decrease in circulating T3 levels in calorically restricted humans due to diminished T4 to T3 conversion, 5 d of starvation of two groups of male Sprague-Dawley rats resulted, paradoxically, in a greater decrease in serum T4 than in serum T3 levels. Kinetic studies show that starvation is associated with no change in the metabolic clearance rate (MCR) of T3, a 20% increase in the MCR of T4, a 67% reduction in turnover rate of T4, but only a 58% reduction in the turnover rate of T3. Moreover, in the first group of rats studied, direct chromatographic analysis of the isotopic composition of total body homogenates after the injection of 125I-T4 showed that 21.8% of T4 is converted to T3 in control rats and 28.8% in starved rats, suggesting that virtually all extrathyroidal T3 in starved and control rats is derived from the peripheral conversion of T4, and that there is little or no direct thyroidal secretion of T3. Our findings strongly point to a reduced thyroidal secretion of T4 as the primary cause of the observed reduction in circulating T3. Since the mechanisms leading to reduced levels of plasma T3 differ in humans and rats, it may be important to reexamine the use of liver homogenate preparations as models for study of the pathogenesis of the "low T3 syndrome" in humans.