We have tested the characteristics of the method of Iwata and Nishikaze (Clin Chem 25: 1317, 1979). The linearity, sensitivity, and precision are satisfactory and the reactivity of benzethonium chloride with various proteins (albumin, immunoglobulins) is the same. The method has been compared with Meulemans's technique (Clin Chim Acta 5: 757, 1960), routinely used in our laboratories, by analysis of 82 samples of cerebrospinal fluid (CSF) and 119 samples of urine. Our results for cerebrospinal fluid agree well with those of Iwata and Nishikaze (r = 0.976; y = 0.992x - 0.013), but we find their method unsuitable for urinary protein determination, probably because of interfering compounds in urine.