In the numerous previously published reports which describe the construction of and experiments with the respiration calorimeter, but little attention has been devoted to a statement of the routine. Since, with the increasing interest in this form of apparatus and the possible construction of others of similar form, a detailed description of the routine would be of advantage, it is here included. PREPARATION OF SUBJECT. Prior to an experiment, the subject is usually given either a stipulated diet for a period of time varying with the nature of the experiment or, as in the case of some experiments, he is required to go without food for at least 12 hours preceding. Occasionally it has been deemed advisable to administer a cup of black coffee without sugar or cream, and by this means we have succeeded in studying the early stages of starvation without making it too uncomfortable for the subject. The stimulating effect of the small amount of black coffee on metabolism is hardly noticeable and for most experiments it does not introduce any error. The urine is collected usually for 24 hours before, in either 6 or 12 hour periods. During the experiment proper urine is voided if possible at the end of each period. This offers an opportunity for studying the periodic elimination of nitrogen and helps frequently to throw light upon any peculiarities of metabolism. Even with the use of a long-continued preceding diet of constant composition, it is impossible to rely upon any regular time for defecation or for any definite separation of feces. For many experiments it is impracticable and highly undesirable to have the subject attempt to defecate inside the chamber, and for experiments of short duration the desire to defecate is avoided by emptying the lower bowel with a warm-water enema just before the subject enters the chamber. Emphasis should be laid upon the fact that a moderate amount of water only should be used and only the lower bowel emptied, so as not to increase the desire for defecation. The clothing is usually that of a normal subject, although occasionally experiments have been made to study the influence of various amounts of clothing upon the person. There should be opportunity for a comfortable adjustment of the stethoscope and pneumograph, etc., and the clothing should be warm enough to enable the subject to remain comfortable and quiet during his sojourn inside the chamber. The rectal thermometer, which has previously been carefully calibrated, is removed from a vessel of lukewarm water, smeared with vaseline, and inserted while warm in the rectum to the depth of 10 to 12 centimeters. The lead wires are brought out through the clothing in a convenient position.