An account of the various methods which have been used in an attempt to measure the rate of sap flow in plants has been given by Crafts, Currier and Stocking (2). Methods which require the measurement of the amount of material moving through a known cross-sectional area, or the timing of the movement of some material indicator such as salts, dyes or radioactive material, suffer from the disadvantage that the vessels conveying the sap stream must be severed, either to measure the quantity of sap or to introduce the indicator, causing an unknown disturbance to the initial state. Our experience confirms that on cutting the vessels of a living tree the rate of sap flow always changes greatly in magnitude, often instantaneously. Heat used as an indicator, however, does not have this disadvantage as it can enter the sap stream by thermal conduction through the walls of the vessels. Measurements using heat have been made by Huber (5) and his school (e.g., Huber and Schmidt (6)), and Dixon (3). Huberts first experiments were made on a tropical liana with such a high rate of sap flow that heat applied for one or two seconds was still recognizable as a pulse at the junctions of a thermocouple 30 cm downstream from the heater. The time till the first appearance of heat at the thermocouple was assumed to be the same as the time taken for the sap to move this distance. For slower sap speeds it became essential to distinguish between the effect of convection by the moving sap and the transport of heat by thermal conduction. Dixon achieved this by placing the thermoj unctions at distances of 1 cm and 2 cm from the heater and considering only cases when the more distant, junction at some stage indicated a higher temperature than the nearer one. Huber and Schmidt (6) devised a method for the same purpose in which one junction was 2.0 cm downstream and the other 1.6 cm upstream from the heater. In all these experiments the heater wires and the thermojunctions were laid either on the intact bark or just under the bark on the surface of the sap wood. There are two fundamental drawbacks to these heat transport experiments. First, it will be shown below that the assumption made in each case that the speed of the sap is identical with that of the heat pulse is not justified. In conifers the sap speed is actually about three times that of the heat pulse. Secondly, comparison between the curves presented for galvanometer deflections versus time and those calculated theoretically shows that, in spite of cotton wool lagging, there was considerable loss of heat; this reduces the usefulness of the results.