Studies on 38 children and adults who ingested large amounts of salicylate revealed a poor correlation of salicylate level in serum with clinical severity of salicylate intoxication. Serial measurements of serum salicylate levels in dogs and in 17 previously well patients revealed that the disappearance of salicylate from circulation is a first-order reaction in the post-absorptive period. It is possible, therefore, to calculate the half-life of circulating salicylate. In the present study the mean salicylate half-life in 17 patients was 20 hours. It is possible to eliminate the variable of time since ingestion by utilizing the rate of removal (half-life) of salicylate to extrapolate from the observed salicylate level at any time following ingestion to a theoretical zero-time salicylate concentration (So). There was found to be an excellent correlation of So with the clinical severity of salicylate intoxication. A nomogram that eliminates the necessity for calculating So is presented. Calculation of So or use of the nomogram provide better means than were previously available for predicting the severity of salicylate intoxication on the basis of measurements of blood salicylate. Particularly noteworthy was the fact that four patients who died despite intensive conventional treatments had So values which greatly exceeded those found in patients who recovered. This observation suggests that an So of greater than 160 is essentially incompatible with life and that measures for rapidly reducing salicylate levels, such as hemodialysis or exchange transfusion, should be instituted. The data are not sufficient to permit conclusions to be drawn concerning the influence of various factors on salicylate removal and toxicity. However, the use of So and salicylate half-life should facilitate such studies.