Abstract The objective of this study was to quantify the accuracy of transpiration estimated using heat balance gauges, porometry, and deuterium tracers methods. Measurements were made on one Eucalyptus tree and three Prunus trees during three measurement periods (MPs) in the summer of 1991 at the Institute of Hydrology, Wallingford, UK. Gravimetric measurements of transpiration ( T g ) of the potted trees were used as the standard for comparison. Continuous estimates of transpiration were made using constant-power heat balance gauges ( T h ). A stomatal conductance-based transpiration ( T s ) was calculated using the Penman-Monteith equation. Deuterium oxide was used as a tracer for calculating transpiration ( T d ) for 5-day periods. There were no systematic differences between daily T g and T h for the one Prunus tree for which daily T g could be accurately measured, or, for all trees, between T g and T h for the entire MP. The root mean square difference (RMSD) between daily T g and T h was 0.26 kg per tree day −1 for the one Prunus tree. There was a consistent underestimation of daily T g by T s , while T h on these days was closer and not consistently different. The RMSD between daily T g and T d was 1.0 kg per tree day −1, more than twice the error for T h . For daily and 5-day periods, the T h RMSD was lower than the RMSD from T s , and T d , respectively. Positive and negative aspects of each method are discussed.