Closed patch testing and the nitrocellulose replica method are performed as useful clinical methods for the evaluation of human skin irritation by cosmetics and topical medicaments. Comparison of the sensitivity between microscopic scoring by nitrocellulose-replica method and visual scoring by closed patch test in the detection of skin irritation, however, has not been well studied with statistical analysis. Here, we evaluated human skin irritation by carboxylic acids. Alcohols, esters and aldehydes, with different chain length (C8-C18), using both methods. The results of closed patch testing showed that, although the score of skin irritation for carboxylic acids (C8, C12), alcohols (C8) and aldehydes (C8), tested at a concentration of 0.5 m-2.0 m. significantly increased with increasing concentration of the test compounds, ester compounds scarcely caused any irritation on the surface of the skin occluded. In addition, an increase of carbon chain length in the test compounds made it impossible to detect skin irritation. In contrast, the nitrocellulose-replica method could evaluate skin reactions against very weak irritants that gave no macroscopic alterations on the skin surface in the closed patch test. However, the scoring system is somewhat subjective and should be improved to make the analysis more objective.