The respiration of a strain of Sarcina lutea, growing as individual cocci of uniform size and of relatively high viability under our experimental conditions, was studied by means of the Warburg technique. Suspensions of thoroughly washed bacteria in water or in glucose solution were enumerated by direct and plate counts, and pH determinations were made before and after manipulation in the manometer. A simple method is described for obtaining accurate gas mixtures with a low concentration of oxygen. The oxygen consumption of various suspensions of the washed cocci was found to be fairly constant in water suspension at about 2.6 cu. mm. O2 per mgm. dry weight per hour, or approximately 7µ2 O2 per single cell per hour: that is, over three times its volume of oxygen was consumed per hour by each micrococcus. This value is for the nearly constant level attained some hours after the start of the experiment. Oxygen consumption is considerably more rapid at first and falls along a roughly arithlogarithmic curve towards an asymptote. This early excess appears to represent an oxygen debt due to partial asphyxia produced in the course of preparation of the suspensions. In 0.2 to 0.5 per cent glucose solution, the rate of oxygen consumption falls along a similar curve, but the rates are 100 to 300 per cent greater than for suspensions of cocci in water. The respiratory quotient in water is constant at about 0.67; in glucose it falls from 0.95 to 0.71 in seven hours (one experiment each). The oxygen consumed by a suspension in water is independent of oxygen concentration when this is above one per cent. For glucose suspensions 2.5 per cent O2 or more is required for maximal consumption. Below the critical values of oxygen tension, oxygen consumption becomes less with diminishing oxygen concentration. When air is admitted to the suspension after a period of complete anoxia, there is observed a high initial rate of consumption followed by progressively declining rates during several hours until a relatively constant rate is approached. When small amounts of oxygen (0.5 per cent or less) are admitted after anoxia, consumption is greater at first and then falls as when air is admitted; but the maximal rate is low, usually less than the normal in air. The oxygen debt developed by these organisms is not apparently associated with the production of detectable acid metabolites (in water suspensions, at least), nor with the formation of carbon dioxide during asphyxia. It may result from the accumulation of non-acid metabolites, or of amounts of acid metabolites that do not equal or exceed the buffering capacities of the cells. The oxidation quotient, even in glucose solution, in contrast to the data for most cells that have been studied, is not over 0.6.