The practice of administering perioperative antimicrobials to uninfected patients undergoing elective surgical procedures has been subjected to careful prospective investigative scrutiny during the past two decades. A wound classification system was developed, helping to identify procedures for which prophylaxis did not provide benefit to the patient. Guidelines for prophylaxis in hysterectomy were proposed and risk factors were sought. Pharmacokinetics and spectrum of antibacterial activity, although variables of paramount importance in predicting success in the treatment of established infections, were discovered to have questionable impact on prophylactic efficacy. Duration of antimicrobial administration has gradually diminished from the total length of hospital stay to a single preoperative dose without increased risk of infection; efficacy may be related to route of administration. The wisdom of using newer agents for prophylaxis is addressed and questioned.