Abstract
A computer was used to carry out numerical iterations of the diffusion pattern of water vapour in models of sub-stomatal cavities. The cavities were simple hemispheres and cylinders having uniformly wet interior surfaces. In the simulations water was allowed to diffuse from the vapour saturated wet surfaces to a single hole which simulated a stoma. After many simulations we concluded that about ¾ of all water evaporation occurs in the region corresponding to the guard cells. We think that a large degree of peri-stomatal evaporation will occur in real sub-stomatal cavities. We have reviewed some of the literature on liquid pathways and resistances in leaves and conclude that the resistance to water flow from veins to the nearest guard cell is large enough to cause substantial localized dehydration of the guard cell resulting in hydroactive stomatal closure. We also suggest that: (1) the use of apoplastic dyes to trace the pathway of water movement in leaves cannot give the correct answer; (2) the reported conductivity of epidermal tissue to water is impossibly high; and (3) previous gravimetric measurements of peristomatal evaporation on large scale models of sub-stomatal cavities were subject to water loss by convection and therefore underestimated the degree of peristomatal evaporation. We also argue that all wet interfaces take up CO2 more or less uniformly even though evaporation of water is mostly peristomatal. The ramified internal structure of leaves therefore allows CO2 uptake with relatively little water loss.