The reliability of reflectivity measurements for the remote sensing of crop productivity may be compromised by possible confounding effects of variation in soil color. An experiment was conducted to determine the influence of selected soil characteristics on reflectance in a broad range of soils typical of those found in southern Alberta. The reflectance of ground (< 2 mm) subsamples of 36 soils previously relocated to a common field site at Lethbridge Alberta was measured indoors with an artificial light source and a 60-degrees field of view. Reflectance was measured between 400 and 1100 nm at 5-nm intervals. Reflectance of ground soils was strongly correlated to that of measurements in the field (r2 = 0.99), though values from the former were consistently higher. Soil reflectance declined with increasing organic carbon content in a quadratic relationship. Furthermore, there was a positive linear correlation between reflectance and soil carbonate content. Variation in organic C and carbonate accounted for 60% of the variability in reflectance among soils, based on multiple regression analysis. These findings confirm that soil carbon exerts a significant effect on reflectance and that reliable estimates of crop productivity by remote sensing requires correction for variable soil reflectance.