We undertook a prospective randomized observer-blinded study comparing the ability of preoperative showers with chlorhexidine gluconate (Hibiclens), povidone-iodine (Betadine), and a lotion soap (Safe 'N Sure) to diminish the staphylococcal skin flora of patients. By block randomization, patients scheduled for an elective cardiac operation or coronary artery angioplasty were assigned to shower with one of the study skin cleansers either once (evening only) or twice (both evening and morning) before the procedure. Semiquantitative samples for culture were obtained from the subclavian and inguinal sites on the evening before the procedure (baseline culture) and again the next morning before the operation. The chlorhexidine skin cleanser consistently reduced staphylococcal colony counts at both the subclavian and inguinal sites before the procedure. This reduction was significant for patients showering both evening and morning (p less than 0.05). The use of the povidone-iodine skin cleanser inconsistently affected skin flora. Patients using lotion soap either experienced no change or had an increase in colony counts. Chlorhexidine is more effective than povidone-iodine in diminishing skin colonization with staphylococci in patients before operation. Repeated applications of chlorhexidine are superior to a single shower with this agent.