A policy of selective management of anterior abdominal stab wounds was evaluated in 333 consecutive patients. Laparotomy was performed based upon initial clinical assessment in 165 patients. There were six deaths (3.6%), major complications in 23 patients (14%), and 28 negative laparotomies (17%). Eighteen patients developed indications for laparotomy after 10.7 +/- 2.2 hours of observation. There were no deaths, major complications in two patients (11%), and four negative laparotomies (22%). One hundred fifty patients were observed and discharged after 1.8 +/- 0.1 days. Sensitivity, specificity, and positive and negative predictive values for initial clinical evaluation were 91%, 85%, 92%, and 83%, respectively. Length of hospitalization in the initially operated and in the delayed group was 9.5 +/- 0.6 and 10.6 +/- 2.6 days (NS), respectively. Most serious intra-abdominal injuries will declare themselves on initial clinical assessment. The remainder are less severe injuries and these patients can be safely observed without undue sequelae due to delay. The use of clinical assessment alone to prompt laparotomy is also cost effective based on analysis of hospital days for initial illness. No procedures are necessary to define penetration.