Immersion method, cup method (modified closed patch method) and dropping method (localized immersion method in which the test solution is dropped on interdigital webs) were studied.The results obtained by the cup method were not correlated to those obtained by the immersion method. The dropping method was considered as being suitable for evaluation of skin-roughness caused by detergent solution, by which the relation between detergent concentration and skin-roughness was shown.Five surface active agents were evaluated by the dropping method. Sodium laurylsulfate (SLS) and cetyldimethylbenzylammonium chloride (CDMBC) caused skin-roughness in the low concentration. The action of sodium lauryltrioxyethylenesulfate (SLTOS), which was usualy used in liquid shampoo formulation, was milder than SLS. Poly-oxyethylenelaurylether (POL), which was non-ionic one, and distearyldimethylammonium chloride (DSDMC), which was cationic one possessed rather high molecular weight, did not cause skin-roughness.Deffating rates of these surface active agents were measured and compared with the results of skin-roughness evaluetion by the dropping method. There was a rough correlation between skinroughnese and deffating rate in lanolin.