The numerous electrical, thermometric, and chemical measurements necessary in the full conduct of an experiment with the respiration calorimeter has often raised the question of the desirability of making at least a portion of these observations more or less automatic. This seems particularly feasible with the observations ordinarily recorded by the physical observer. These observations consist of the reading of the mercurial thermometers indicating the temperatures of the ingoing and outcoming water, records with the electric-resistance thermometers for the temperature of the air and the walls and the body temperatures, and the deflections of the thermo-electric elements. Numerous plans have been proposed for rendering automatic some of these observations, as well as the control of the heating and cooling of the air-circuits. Obviously, such a record of temperature measurements would have two distinct advantages: (1) in giving an accurate graphic record which would be permanent and in which the influence of the personal equation would be eliminated; (2) while the physical observer at present has much less to do than with the earlier form of apparatus, it would materially lighten his labors and thereby tend to minimize errors in the other observations. The development of the thread recorder and the photographic registration apparatus in recent years led to the belief that we could employ similar apparatus in connection with our investigations in this laboratory. To this end a number of accurate electrical measuring instruments were purchased, and after a number of tests it was considered feasible to record automatically the temperature differences of the ingoing and outcoming water from the calorimeter. Based upon our preliminary tests, the Leeds & Northrup Company of Philadelphia, whose experience with such problems is very extended, were commissioned to construct an apparatus to meet the requirements of the respiration calorimeter. The conditions to be met by this apparatus were such as to call for a registering recorder that would indicate the differences in temperature between the ingoing and outcoming water to within 0.5 per cent and to record these differences in a permanent ink line on coordinate paper. Furthermore, the apparatus must be installed in a fixed position in the laboratory, and connections should be such as to make it interchangeable with any one of five calorimeters.