BY THE majority of clinicians, the therapeutic value of diathermy is attributed to the production of heat within the tissues, resulting in a rise in temperature. Hence it is of importance to know the actual rise in temperature during a treatment in order to prevent damage to the tissue by too high a temperature, and yet to have a sufficient rise in temperature to be of therapeutic value. Two types of diathermy electrodes are in general use. One is an electrode of pliable tin. The other consists of a felt pad soaked in electrolyte and applied to the skin, connection to the diathermy machine being made with copper gauze sewn to the outer surface of the pad. It has been reported that apparently greater heating is produced with a metallic electrode. Bordier (1) has observed greater perspiration and redness of the skin with a metallic electrode than with the saline pad. By inserting a thermometer through an incision in the skin over the pectoralis major, Simon (2) has observed a similar effect. In the present paper, we have investigated further the problem of the temperature distribution at the surface and in the tissues below these electrodes. The temperatures were measured by thermocouples within number 22 hypodermic needles, soldered to the point of the needle, a method used by Bazett and McGlone (3). The thermocouple circuit is similar to the circuit of Clark (4). Preliminary experiments were carried out on dogs and are reported elsewhere (5) with all the technical details. On human subjects, one type of electrode was placed on the skin over the vastus lateralis muscle, while an electrode of the other type was placed in the corresponding position on the other thigh. Both electrodes had the same area of 60 square centimeters. A large metallic electrode, 400 sq. cm. in area, was placed on the inside of each thigh and joined to the one opposite by a copper wire. The outer electrodes, beneath which the temperatures were measured, were connected to the diathermy machine. Thus the same diathermy current would traverse an identical tissue path, the only differences being due to the two types of electrodes. This manner of comparing electrodes also eliminated variations due to fluctuating currents, variations in room temperature, and errors in duration time of treatment, since the same errors would affect both electrodes simultaneously. Two thermocouple needles were inserted into the tissue beneath each electrode. One thermocouple needle was inserted subcutaneously while the other was placed within the muscle approximately 0.75 in. from the surface. A thermocouple was also placed on the surface between the skin and the electrode. The three thermocouples beneath each electrode are referred to in the curves in Charts I and II as surface, subcutaneous, and intramuscular thermocouples. Before being used, the pad electrode was soaked in a 10 per cent sodium chloride solution. With this concentration of saline, no appreciable surface heating occurs (6).