The three most commonly used methods for estimating mesophyll conductance (g(m)) are described. They are based on gas exchange measurements either (i) by themselves; (ii) in combination with chlorophyll fluorescence quenching analysis; or (iii) in combination with discrimination against (13)CO(2). To obtain reliable estimates of g(m), the highest possible accuracy of gas exchange is required, particularly when using small leaf chambers. While there may be problems in achieving a high accuracy with leaf chambers that clamp onto a leaf with gaskets, guidelines are provided for making necessary corrections that increase reliability. All methods also rely on models for the calculation of g(m) and are sensitive to variation in the values of the model parameters. The sensitivity to these factors and to measurement error is analysed and ways to obtain the most reliable g(m) values are discussed. Small leaf areas can best be measured using one of the fluorescence methods. When larger leaf areas can be measured in larger chambers, the online isotopic methods are preferred. Using the large CO(2) draw-down provided by big chambers, and the isotopic method, is particularly important when measuring leaves with high g(m) that have a small difference in [CO(2)] between the substomatal cavity and the site of carboxylation in the chloroplast (C(i)-C(c) gradient). However, equipment for the fluorescence methods is more easily accessible. Carbon isotope discrimination can also be measured in recently synthesized carbohydrates, which has its advantages under field conditions when large number of samples must be processed. The curve-fitting method that uses gas exchange measurements only is not preferred and should only be used when no alternative is available. Since all methods have their weaknesses, the use of two methods for the estimation of g(m), which are as independent as possible, is recommended.