A technique for assessing the immediate and prolonged efficacy of surgical scrubs and alcoholic hand rubs is described. A mean baseline count is obtained from all volunteers and logarithmic reductions in resident skin flora immediately after one or more applications, and after wearing gloves for 3 h, are measured. Loose-fitting surgical gloves are used for sampling resident flora. Preparations were applied using a standard technique for 2 min, apart from one test with 70% isopropanol (IPA) in which the application time was 30 s. Two studies are described, one of which compared four chlorhexidine scrubs, and the second 70% IPA, 7.5% povidone-iodine scrub, 2% triclosan cleanser and unmedicated bar soap. In spite of their constituent similarity, the four chlorhexidine scrubs varied considerably in efficacy and user acceptability. A 2 min application of 70% IPA was the most effective treatment, and gave log10 reductions of 1.65 for immediate and 1.58 for prolonged effect. This was marginally more effective than a 30 s application, but the difference was not significant. 'Hibiscrub' was the most effective aqueous formulation and gave reductions of 1.01 for immediate effect and 1.16 for prolonged effect. The test described could be used by reference centres and manufacturers to assess the efficacy of new and existing surgical hand disinfection formulations.