Abstract Differences in fabric may explain why relationships between aggregate stability and soil color have never been reported for a relatively narrow-ranging population of soils. Therefore, three size fractions (1–2, 0.25–1, and 0.05–0.25 mm) of 25 topsoils (75 samples) from a Regosol–Umbrisol soilscape (Sierra Nevada, SE Spain) were characterized for fabric by scanning electron microscopy, aggregate stability by wet sieving, and CIELAB color (hab, L*, C*ab) by diffuse spectral reflectance. A stability-index ranking from 0.17 to 0.98 paralleled a ranking of hab (88.5 to 66.1), L* (48.3 to 30.2), and C*ab (7.6 to 18.9): that is, the progressive increase in stability was related to increased soil redness, darkness, and chromaticity. The reason was an increase in binding agents such as clay, organic C, and free forms coating the surfaces of the aggregates. The values of the correlation coefficients (0.55 to 0.82), factor-analysis variances (73 to 88%), and multiple-regression coefficients (0.40 to 0.65) indicated, however, that much of the variability in stability cannot be statistically explained by color changes. Size fractions of soils with a similar binding-agent content did not significantly (P