It is especially necessary to guard against the development of hypercapnia and hypoxia in the patient during thoracic surgery because so many factors may combine to interfere with pulmonary ventilation. It is especially difficult to judge the depth of anesthesia and adequacy of gas exchange when spontaneous respiration is replaced by controlled respiration. The possible advantages of deliberate hyperventilation were explored in 128 patients undergoing thoracic surgery. In most of them ventilation was achieved by manual compression of the reservoir bag, but in about 20 patients mechanical respirators were used. Samples of arterial blood were obtained for determinations of pH, pCO 2 , and oxygen saturation. Four case histories are given showing the possibility of maintaining normal acidbase balance or producing an alkalosis in this way, and illustrating the inadequacy of spontaneous respiration in some cases. There was no evidence that the methods of artificial respiration used hindered the return of venous blood to the heart. Excessive depth of anesthesia was avoided by using concentrations and quantities of anesthetics such that it was virtually impossible for deep planes of anesthesia to be attained. No ill-effects attributable to alkalosis were seen even in patients maintained in that state for several hours. The results indicate that an extensive evaluation of the relative merits of respiratory acidosis and alkalosis is desirable.