Studies in experimental salicylate poisoning were carried out to ascertain the therapeutic value of gastric lavage and to compare it with the effect of emesis. Fasted dogs, 6 to 10 kg in weight, were given sodium salicylate in tablet form, and lavage was performed or emesis induced at varying time intervals following ingestion. Salicylate recovery was determined in the lavage and emesis fluids. The data indicate that in the recovery of ingested salicylate: 1) Initial aspiration of the stomach before introduction of fluid is equally as important as succeeding lavage. 2) Lavage carried out within 15 minutes is no more effective than emesis induced within 30 minutes of ingestion; 3) lavage carried out 1 hour after ingestion is inefficient and far less effective than emesis produced within the same time interval. 4) Spontaneous emesis is not as effective as induced emesis. 5) Emesis induced immediately, or even when delayed as long as 1½ hours after ingestion, appears to be the more effective and thus the preferable form of therapy. 6) Neither lavage nor emesis under the most optimal conditions are consistent in their effectiveness, so that all patients after either form of therapy should be followed carefully for signs of increasing drug intoxication, and treated if a specific form of therapy is available.