Nitrate reductase (NR) activity in spinach leaf extracts prepared in the presence of a protein phosphatase inhibitor (50 microM cantharidine) was measured in the presence of Mg2+ (NRact) or EDTA (NRmax), under substrate saturation. These in-vitro activities were compared with nitrate reduction rates in leaves from nitrate-sufficient plants. Spinach leaves containing up to 60 micromol nitrate per g fresh weight were illuminated in air with their petiole in water. Their nitrate content decreased with time, permitting an estimation of nitrate reduction in situ. The initial rates (1-2 h) of nitrate consumption were usually lower than NRact, and with longer illumination time (4 h) the discrepancy grew even larger. When leaves were fed through their petiole with 30 mM nitrate, initial in-situ reduction rates calculated from nitrate uptake and consumption were still lower than NRact. However, nitrate feeding through the petiole maintained the in situ-nitrate reduction rate for a longer time. Initial rates of nitrate reduction in situ only matched NRact when leaves were illuminated in 5% CO2. In CO2-free air or in the dark, both NRact and in-situ nitrate reduction decreased, but NRact still exceeded in-situ reduction. More extremely, under anoxia or after feeding 5-amino-4-imidazole carboxyamide ribonucleoside in the dark, NR was activated to the high light level; yet in spite of that, nitrate reduction in the leaf remained very low. It was examined whether the standard assay for NRact would overestimate the in-situ rates due to a dissociation of the inactive phospho-NR-14-3-3 complex after extraction and dilution, but no evidence for that was found. In-situ NR obviously operates below substrate saturation, except in the light at high ambient CO2. It is suggested that in the short term (2 h), nitrate reduction in situ is mainly limited by cytosolic NADH, and cytosolic nitrate becomes limiting only after the vacuolar nitrate pool has been partially emptied.