Barley (ffordeum vulgare L. var. Prior) leaves converted more '4C-glutamic acid to free proline when water-stressed than when turgid; neither decreased protein synthesis nor isotope trapping by the enlarged free proline pools found in wilted tissue seemed to account for the result. This apparent stimulation of proline biosynthesis in wilted leaves was not observed when radioactive ornithine or P5C (A1-pyrroline-5-carboxylate, an intermediate following glutamate in proline synthesis) were used as proline precursors unless proline levels were high as a result of previous water stress. We interpret this to mean that any stimulation of proline synthesis by water stress must act on P5C formation rather than its reduction to proline. Experiments showing greater apparent conversion of 14C-glutamate to proline do not unequivocally prove that proline synthesis is stimulated by water stress, as P5C feeding studies show that proline oxidation is inhibited under comparable conditions. This inhibition could account, at least in part, for increased proline labeling, and must be considered an alternate possibility.Although several plants have been reported to accumulate free proline during periods of water deficit (2,12,15,16,18) or when subjected to other stress (7, 8), the biochemical changes linking water stress and proline accumulation are not well understood. Barnett and Naylor (2) and Morris et al. (13) have observed increased incorporation of 14C-glutamic acid into proline in wilted leaves as compared to turgid leaves of Bermuda grass and turnip. This increase in radiotracer incorporation might reflect a stimulation of proline biosynthesis by water stress, but the data presented did not fully eliminate the possibility that inhibited protein synthesis might account for the results.The Bermuda grass experiment was done after a long period of drought, so that the large amount of free proline already present during "4C-glutamate feeding could have acted as a trapping pool for radioactive proline, accounting for the results in the absence of an actual increase in the rate of conversion of glutamate to proline. In this paper we describe 14C-glutamate feeding experiments designed to avoid these difficulties in interpretation, and to localize the point in the proline biosynthetic pathway at which a stress-induced stimulation would most likely occur.