The role of anesthesia in contributing to surgical mortality has been studied in 33,224 patients given either spinal anesthesia or a general anesthetic to which muscle relaxants were added. There were no deaths atributable to anesthesia in the 16,000 physically fit patients anesthetized by either technique. As the patients' physical condition worsened, deaths related to anesthesia increased in incidence; in the moribund patients, 1 in 16 patients given spinal anesthesia died of causes related to the anesthetic, and in 1 in 10 patients, general anesthesia could not be excluded as contributing to death. Of 6,000 physically fit patients who received a muscle relaxant, none died. No evidence of an inherent toxicity of muscle relaxants could be found. When deaths were related to the use of muscle relaxants, errors of omission or commission were always apparent.