Factors involved in estimating dry green biomass (DGB) by canopy spectral reflectance were evaluated at 6 dates during the 1976 growing season. In situ measurements were taken in the Mixed Prairie of western South Dakota [USA] on replicated pastures in different range condition classes using a modified hand-held biometer. Canopy reflectance and calibration panel reflectance were determined at 0.675 and 0.800 .mu.. Factors considered in estimating DGB via stepwise multiple regression were the canopy reflectances, calibration panel reflectances, time of day and coefficient of variability among vegetation samples in a pasture. Canopy reflectance readings were included as both a ratio of the 2 wavelengths and as 2 separate variables in 2 sets of analyses. Canopy reflectance readings alone were not acceptable estimates of DGB (R2 = 0.029 for ratio and 0.042 for the linear combination). The coefficient of variability samples within a pasture improved the association (R2 = 0.233 and 0.231) while further inclusion of both calibration readings resulted in a marked improvement in estimation of DGB (R2 = 0.633 and 0.899). These calibration readings corrected for sun angle and diffuse cloud cover so that time of day of measurement was not an important variable.