Heterozygous plants of barley (Hordeum vulgare L. cv. Maris Mink) with activities of chloroplastic glutamine synthetase (GS) between 47 and 97 of the wild-type and ferredoxin-dependent glutamate synthase (Fd-GOGAT) activities down to 63 of the wild-type have been used to study the control of photosynthetic fluxes. Rates of CO2 assimilation measured over a range of intercellular CO2 concentrations and photon flux densities (PFDs) were little different in the wild-type and a mutant with 47 GS, although total activities of ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) decreased by about 20 with a decrease in GS to 50 of the wild-type. The quantum efficiencies of photosystem II electron transport (PSII and CO2 assimilation CO2) were determined. PSII was lower than expected in mutants with 50 less GS under conditions which enhance the photorespiratory flux, but were identical to the wild-type under non-photorespiratory conditions, suggesting that at high rates of photorespiration the electron requirement for net CO2 assimilation declines in plants with decreased GS. This discrepancy in the electron requirement between the wild-type and the 47 GS mutant was enhanced at high temperatures and low CO2, conditions which favour oxygenation by Rubisco. Photochemical and non-photochemical chlorophyll a fluorescence quenching as well as the quantum efficiency of excitation-energy capture by open photosystem II reaction centres were differentially affected in mutants with less GS relative to the wild-type when CO2 was lowered or the PFD was varied. The quantum efficiencies of electron transport in photosystems I and II were closely correlated under a range of PFDs and CO2 concentrations, confirming that the rate of linear electron transport was much lower in plants with less GS. It is shown that GS exerts considerable control (flux control coefficients between 0.5 and 1.0) on the electron requirement for CO2 assimilation at high fluxes of photorespiration relative to CO2 assimilation. Apart from the control of GS on protein and Rubisco contents, GS in the wild-type has also some direct positive control on CO2 assimilation. However, negative control on CO2 assimilation was found in mutants with 50 less GS. These data, taken with the data on electron requirements for CO2 assimilation, suggest that CO2-fixing processes other than that catalysed by Rubisco, such as carboxylation of phosphoenolpyruvate, or an inhibition of photorespiration (e.g. glycine decarboxylation), may contribute to the observed CO2 exchange and photosystem II electron transport in plants with less GS. In the 63-Fd-GOGAT mutant, rates of CO2 assimilation were appreciably lower than in the wild-type under a range of PFDs and CO2 concentrations, which largely reflected lower contents of Rubisco in the Fd-GOGAT mutants. Assimilation of CO2 was inhibited appreciably at high CO2 concentrations. There was little difference in the electron requirement for CO2 assimilation between the wild-type and mutants with less Fd-GOGAT, although there were indications that a triose-phosphate/glycerate-3-phosphate shuttle or cyclic electron transport operates to balance ATP generation and NADP reduction. The latter was supported by a curvilinear relationship of photosystem I and II electron transport in the 63 Fd-GOGAT mutant. A positive control is exerted by Fd-GOGAT on the amounts of protein and Rubisco and on CO2 assimilation.