Diffuse reflectance infrared spectra of differently mercerized cotton poplin fabrics are interpreted with the aid of fuzzy clustering and multivariate chemometrics methods (PCA and SIMCA). Fuzzy clustering analysis of spectra from blank, slack, and stretch mercerized samples shows that three clusters can be identified, and this conclusion is supported by the PCA display method and SIMCA classification. Spectra from slack mercerized samples, treated at three different temperatures, are also readily discriminated by the same three chemometrics methods.Mercerization is a well known and important treatment commonly applied to cotton fabrics to enhance their luster and dyeing properties [ 1 l, I 5, 25, 38 ] . It involves the application of concentrated caustic solutions at elevated temperatures. Significant changes result from this treatment, with different conditions producing different fiber and fabric properties. For this reason, it is useful to have rapid techniques that can discriminate differently mercerized samples. At present, principle tests that are used to determine the state of mercerization are based on staining techniques and microscopic examination [ 25,38 ] ; neither of these ap-proaches is accurate or quantitative. X-ray diffraction methods, which have been used to detect ccystallinity changes in cellulose [ 39 ], including detailed studies referring to mercerized cotton [ 33, 34 J, are comparatively time consuming and costly.Infrared spectroscopy is a recognized technique for studying cotton cellulose. Since the late 1940s, when Rowan et al. [ 31 published a paper on the first infrared spectrum of a film of regenerated cotton cellulose, vibrational spectroscopy has been used extensively in this 6eld [ 23,24 ] . In 1950, Forziati et al. [ 8 ] showed that during preparation of a Nujol mull, the grinding of cotton fibers produced some changes in the structure of cellulose, which were reflected in the IR spectra.