A correct evaluation of coffee transpiration is required for many purposes, such as hydrological services, drought, sustainable irrigation, even in regions where water is not a limiting factor. Two sap flow methods are commonly used in order to assess transpiration of individual plants: i) the Granier¿s method, based on the thermal dissipation, yields sap flow density, but this requires a specific calibration; and ii) the energy balance method (e.g. Dynagage ®), which provides a calibration-free measurement of sapflow, but with inherent drawbacks (price, fragility, intolerance of the plant). It was hypothesized that an adequate combination of both methods would allow efficient and long-term evaluation of coffee sap flow in the field. We tested both methods on 2 m high adult coffee plants, transplanted in containers and acclimated in a greenhouse 6 months before the experiment. Nine plants were equipped with sap flow probes, 6 Granier (half were 2 cm long and half were 1 cm long), and 3 with Dynagage®. Diurnal transpiration of all plants was controlled gravimetrically using an electronic scale (daily error < 1% of transpired water), after covering the soil. We obtained good correlations between every sap flow method and gravimetry. R2 ranged between 0.71 and 0.94 and was higher for Granier¿s probes. The CV was less than 25%. Both methods were reliable. As expected, the Dynagage® probes yielded non-biased estimations of transpiration, whereas the home-made Granier¿s probes required a specific calibration: we recorded distinct calibration coefficients for the 2 cm and for the 1 cm probes. We confirmed that the natural thermal gradient in the stem of these short trees was not negligible. Taking into account in every method allowed a significant improvement in the estimations of transpiration. We concluded that Granier¿s probes were attractive for coffee sapflow in the field, considering their cost-effectiveness, accuracy and suitability for long-term monitoring, but