Additions of methionine sulfoximine (MSX), an inhibitor of glutamine synthetase (GS), result in an increase in NH13 in seedling leaves of Cs ( (9) suggested that the major reaction leading to the release of CO2 in photorespiration was the conversion of glycine to serine. If this were true, the release of NH3 by photorespiration would be as significant as the release of CO2. In C3 plants, the rates of photorespiration can be considerable, up to 25% the rate of photosynthesis (9). Thus, in a system where NH3 accumulation can be measured, NH3 derived from photorespiration could be far in excess of the NH3 produced by the reduction of N03 . Under normal conditions, however, little NH3 is recovered from leaf tissue (1, 21). Miflin and Lea (1 1 NH3, and subsequently additions of MSX, an inhibitor of GS, were shown to result in accumulations of NH3 (4,5,10,12,18) and in the inhibition of photosynthesis (13,14). In 1978, Keys et al. (7) proposed a photorespiratory nitrogen cycle whereby NH3 released from glycine would be efficiently reassimilated by GS in the cytosol. The resultant glutamine (20) would be transferred to the chloroplast where it would serve as a substrate for GOGAT, and the resultant glutamate would serve as the N-donor in the transaminase reaction leading to glycine formation in the peroxisome. The overall reactions and potential sites of inhibition are illustrated in Figure 1. Sommerville and Ogren (16) showed that Arabidopsis mutants lacking GOGAT accumulated NH3 under conditions which permitted photorespiration. In mutants lacking serinetranshydroxymethylase activity, NH3 was necessary for the continued synthesis of glycine, again under conditions which permitted photorespiration (17). Thus, their results support the hypothesis of Keys et al (7).Recently, Platt and co-workers (13, 14, and personal communication) observed an accumulation of NH3 in spinach leaf discs and an inhibition of photosynthetic activity in the presence of MSX. Under conditions where photosynthesis was inhibited, they still saw an accumulation of NH3 in experiments where the 02 content was reduced to 2%. These conditions should have inhibited photorespiration (15,19) and hence NH3 accumulation. These observations agree with neither the Arabidopsis-mutant studies (16,17) nor with the original hypothesis of Keys et al (7,20) as it suggests alternate and significant sources of NH3 in leaf tissue. Explanations for this apparent contradiction could be the excessively high levels of MSX (8 mM) used in Platt's experiments and to the use of leaf discs rather than whole leaves.In view of the fact that corn leaf pieces showed an accumulation of NH3 in the presence of MSX (12) and that high levels of MSX (2.5 mM) caused a general inhibition of root metabolism (A. Oaks, unpublished), it seemed to us that the role of MSX in leaves needed to be reexamined. In this paper, we show that a maximum level of NH3 is released by relatively low concentrations of MSX