A comparison was made of stomatal behaviour, and related phenomena, between leaves of garden pea (Pisum sativum cv. Feltham First) inoculated with powdery mildew fungus (Erysiphe pisi) and uninfected leaves on healthy plants. Twenty four hours after inoculation, stomata opened more widely in the light in infected leaves than in healthy leaves. Thereafter, stomatal opening was progressively reduced by infection and stomata failed to close completely in the dark until, 7 d after inoculation, all movements ceased and stomata remained partly open. Transpiration in the light followed closely the pattem of stomatal opening and, after an early increase compared with healthy controls, was progressively reduced by infection. Evidence is presented that transpiration from the fungus was less than the reduction in transpiraation from the leaf which was caused when development of the mycelium increased the boundary layer resistance of the leaf. Seven days after inoculation, transpiration in the dark was greater from infected leaves than from healthy leaves because of partly open stomata in the dark. Net photosynthesis in infected leaves was reduced within 24 h of inoculation to a level below that found in healthy leaves and thereafter it declined progressively. The initial reduction was due to a transient increase in photorespiration, for when the glycolate pathway was inhibited by a 2% O2 concentration there was no difference between the (gross) photosynthetic rates of healthy and infected leaves. Changes in photorespiration rate were confirmed from the interpretation of the CO2 burst on darkening. Reduced stomatal opening was a contributory cause of the reduction in net photosynthesis in the later stages of infection. Since the rate of gross photosynthesis, but not the rate of photorespiration, of infected plants fell below that of healthy plants, and infected plants had a higher rate of release of CO2 in the dark than healthy plants from the third day after inoculation onwards, infected plants consume an increasingly greater proportion of their photosynthate in respiratory processes than do healthy plants. The CO2 compensation point of infected plants increased at every time of sampling after inoculation.