The effects of delay in surgical treatment and the choice of operation on morbidity associated with empyema thoracis were evaluated in 122 consecutive patients. Patients (71 from a private practice and 51 from an inner-city trauma/indigent care facility) eligible for study were divided into treatment groups of chest tube only (CT = 39) and open drainage (OD = 19), or decortication (DC = 65). Delay in treatment was defined as greater than 3 days from recognition of empyema to CT and greater than 14 days to OD or DC when chest tubes were inadequate or were not used initially. Delay in OD significantly increased total illness (p = 0.023), days until removal of chest tubes (p = 0.037), and hospital stay (p = .048), but did not affect postoperative stay. Delay in DC increased total illness (p = 0.0001), but did not affect other variables. Delay in CT increased mortality from 3.4 percent to 16 percent. Delay did not increase mortality in OD and DC. DC was superior to OD in patients requiring major operation in total illness days (DC = 36.1 vs OD = 106.1) (p = 0.0005), days until removal of tubes (DC = 7.5 vs OD = 78.3) (p = 0.0001), and postoperative stay (DC = 11.6 vs OD = 17.3) (p = 0.018). Overall mortality was lowest in the DC group (6.1 percent). Delay in treatment increases morbidity and DC is more effective than OD in reducing morbidity and mortality when surgical intervention is necessary.