Mortality and morbidity in fire victims is largely a function of injury due to heat and/or smoke. While degree and area of burn together constitute a reliable numerical measure of cutaneous injury due to heat, as yet no satisfactory measure of inhalation injury has been developed. In this study, with fluid resuscitation and pulmonary infection eliminated as variables, dose-response curves were constructed as a measure of inhalation injury by exposing burned and unburned animals to smoke of constant temperature and toxicity under conditions similar to the fire situation. In these animals, the natural history of inhalation injury: 1) proved to be a relatively simple function of smoke and burn dosage; 2) appeared to simulate and therefore aid interpretation of the inhalation injury syndromes seen in human fire victims; 3) indicated that within limits [COHgb] measured immediately after injury was directly proportional to, and might prove to be a clinically valuable measure of, absorbed dose of smoke. While fluid resuscitation and pulmonary contamination with bacterial pathogens may be eliminated experimentally, such is not the case with the vast majority of fire victims admitted to burn services with associated inhalation injury. Fluid resuscitation and inhalation of a Pseudomonas aeruginosa aerosol were therefore included serially in a study of animals with inhalation injury and burns large enough to require fluid resuscitation. In these animals it was demonstrated that: 1) pulmonary edema occurred in association with too little rather than too much fluid therapy; 2) after aerosol inoculation, fatal bacterial pneumonia was difficult to produce when inhalation injury was associated with no or only small burns, but common when associated with no or only small burns, but common when associated with a burn large enough to require fluid resuscitation.