A model for evaluation of the residual activity of agents used for hygienic hand disinfection, which relates closely to the "in-use" setting, is described. We looked for residual activity following hand disinfection by a standard method with soap and water and four skin disinfectants in common use. Two test organisms known to have caused cross-infection and having clinically-significant antibiotic resistance were used: a strain of vancomycin-resistant Enterococcus faecium and one of gentamicin-multiply-resistant Enterobacter cloacae. Both organisms survived well on artificially inoculated hands. Residual activity on previously disinfected hands against both test organisms was most marked for alcoholic chlorhexidine ('Hibisol') which showed a 4 log10 reduction within 1 min of inoculation. Chlorhexidine digluconate skin cleanser ('Hibiscrub') and povidone-iodine surgical scrub ('Betadine') also had marked residual activity, but less than alcoholic chlorhexidine. 60% Isopropanol showed no residual activity and, except for marginal activity against E. cloacae, soap and water showed little residual activity. The residual activity of alcoholic chlorhexidine may be useful to control cross-infection with those multiply-resistant strains which survive well on fingers previously washed with soap and water.