The heat pulse method for measuring sap flow in cotton (Gossypium hirsutum L. cv. SJ‐2) stems was developed and tested to provide a means for investigating transpiration of single plants under field conditions without disturbing the environment. The temperature difference between two radially inserted thermocouples, one 9 mm above and the other 4 mm below a heater piercing the stem, was measured every 0.3 s following emission of a heat pulse. These dimensions are a compromise solution between ideal configuration, ease of probe construction, and resolution of the signal detector. Two data were recorded: the time elapsed from pulse emission to the first re‐occurrence of the initial temperature difference, and the time at which the temperature difference was maximum. At low and moderate convective sap velocities in the stem (0‐0.17 mm s⁻¹), the first datum can be measured with precision to accurately estimate sap velocity. At velocities higher than 0.22 mm s⁻¹ the initial differential temperature change is too small and evolves too rapidly to detect when temperature returns to its initial value, but the temperature wave has a well defined maximum that is related to sap velocity. For convective velocities between 0.17 and 0.22 mm s⁻¹, both data can be used. Calibration against transpiration measurements in potted cotton plants, grown outdoors, showed that the apparent effective conducting stem area is 0.71 of the total cross‐section. This fraction was found to remain constant over the entire range of stem diameters reached by cotton grown as an annual crop. The transpiration rate of a cotton plant is determined by multiplying the effective conducting area by the sap velocity. The study shows how the heat pulse quantifies water uptake of cotton and can be carried out in field conditions.