To non-anaesthetized rats starved for 3 days, [U-14C]acetone, NaH14CO3, L-[U-14C]lactate, [2-14C]acetate or D-[U-14C]- plus D-[3-3H]-glucose was injected intravenously. From the change in the plasma concentration of labelled acetone versus time after the injection, the metabolic clearance rate of acetone was calculated as 2.25 ml/min per kg body wt., and its rate of turnover as 0.74 mumol/min per kg. The extent and time course of the labelling of plasma glucose, lactate, urea and acetoacetate were followed and compared with those observed after the injection of labelled lactate, acetate and NaHCO3. The labelling of plasma lactate was rapid and extensive. Some 1.37% of the 14C atoms of circulating glucose originated from plasma acetone, compared with 44% originating from lactate. By deconvolution of the Unit Impulse Response Function of glucose, it was shown that the flux of C atoms from acetone to glucose reached a peak at about 100 min after injection of labelled acetone. In comparable experiments the transfer from lactate reached a peak at 14 min after the injection of labelled lactate. It was concluded that acetone is converted into lactate to a degree sufficient to account for the labelling of plasma glucose and is thus a true, albeit minor, substrate of glucose synthesis in starved rats.