Soil color is widely recognized as a primary property of soil identification. Soil physical, mineralogical, and chemical properties can be derived from assessing soil color. The current framework of soil classification, at the second hierarquical level of the Brazilian System of Classification, requires that the soil colour of the diagnostic B horizon be determined by comparison with the Munsell color charts, or equivalent. This process of classification is extremely subjective and time-consuming, limiting the accuracy, transferability, and usefulness of the procedure. This paper aims at a comparison of the conventional soil color determination with the colormeter method to evaluate possible implications in soil classification. Five soil scientists were invited to determine the hue of 80 soil samples by comparison with the Munsell color chart. The hue of the samples was obtained with a colormeter as well, also according to Munsell's notes. The statistical analyses determined correlation coefficients, precision index, clustering forms, and mean tendencies of hue determinations. Correlation coefficients varying from 0.68 to 0.94 showed that soil scientists produced precise data, fundamental for the validity of the study. However, the measurements realized by soil scientists were not coincident, and were quite different from the colormeter measures. The values of Precision Index indicated an agreement of only 8.75 % in moist samples and 17.5 % in dry ones. The divergences in hue determination would lead to mistakes in soil classification. The evaluation of average tendencies indicated that the pedologists overestimated color values in relation to colorimeter, resulting in alterations in soil classification.
hue; colorimetry; Munsell book of color; pedology
