A statistical evaluation was made of reflectance measurements obtained on raw cotton. A large variation was observed in unblended cotton samples either greige or dyed while slight blending reduced the variation, significantly. Instrument variation also calculated from part of the data gathered showed only slight (negligible) variation. Homogenization of samples prior to quantitative measurements was indicated and use of reflectance measurements from the wavelength of maximum absorbance seemed most sensitive to sample preparation changes.
