Intercellular CO2 concentration of leaves (Ci) is a critical parameter in photosynthesis. Nevertheless, uncertainties in calculating Ci arise as stomata close. Here, by modifying the assimilation chamber of a commercial gas-exchange equipment to directly measure Ci, we demonstrate overestimation of calculated Ci (i.e. Ci(c)) without stimulating stomatal closure. Gas exchange was measured on one side of the leaf while measured Ci (Ci(m)) was acquired simultaneously on the other side of the leaf in hypostomatous passion fruit (Passiflora edulis Sims) and amphistomatous sunflower (Helianthus annuus L.) and common bean (Phaseolus vulgaris L.). The adaxial surface showed comparable Ci(c) and Ci(m) in sunflower, whereas in common bean, where the adaxial surface has a low stomatal density, Ci(c) markedly differed from Ci(m) when the stomata remained open. However, the latter discrepancy disappeared when measuring the leaf flipped upside down so that the gas exchange was measured (i.e. Ci was calculated) on the abaxial side, which has a much higher stomatal density. The passion fruit showed the largest discrepancy on the astomatous side, indicating that the cuticle has a large impact on the calculation. Direct measurement of Ci is recommended as a more accurate estimate than the calculation when stomatal gas transport is restricted. Occurrence of overestimation and prospects for direct measurement are discussed.