The appearance of transverse sections of maize leaves indicates the existence of two airspace systems serving the mesophyll, one connected to the stomata of the upper epidermis and the other to the stomata of the lower surface, with few or no connections between the two. This study tests the hypothesis that the air-space systems of the upper and lower mesophyll are separated by a defined barrier of measurable conductance. A mathematical procedure, based on this hypothesis, is developed for the quantitative separation of the contributions made by the upper and lower halves of the mesophyll to carbon assimilation using gasexchange data. Serial paradermal sections and three-dimensional scanning-electron-microscope images confirmed the hypothesis that there were few connections between the two air-systems. Simultaneous measurements of nitrous-oxide diffusion across the leaf and of transpiration from the two surfaces showed that the internal conductance was about 15% of the maximum observed stomatal conductance. This demonstrates that the poor air-space connections, indicated by microscopy, represent a substantial barrier to gas diffusion. By measuring the CO2 and water-vapour fluxes from each surface independently, the intercellular CO2 concentration (c i) of each internal air-space system was determined and the flux between them calculated. This allowed correction of the apparent CO2 uptake at each surface to derive the true CO2 uptake by the mesophyll cells of the upper and lower halves of the leaf. This approach was used to analyse the contribution of the upper and lower mesophyll to CO2 uptake by the leaf as a whole in response to varying light levels incident on the upper leaf surface. This showed that the upper mesophyll was light-saturated by a photon flux of approx. 1000 μmol·m(-2)·s(-1) (i.e. about one-half of full sunlight). The lower mesophyll was not fully saturated by photon fluxes of nearly double full sunlight. At low photon fluxes the c i of the upper mesophyll was significantly less than that of the lower mesophyll, generating a significant upward flux of CO2. At light levels equivalent to full sunlight, and above, c i did not differ significantly between the two air space systems. The physiological importance of the separation of the air-space systems of the upper and lower mesophyll to gas exchange is discussed.