A simple hand-held reflectometer was used to estimate the shoot dry matter of pearl millet (Pennisetum glaucum), groundnut (Arachis hypogaea), cowpea (Vigna unguiculata), weeds and grass swards non-destructively. While the instrument was able to predict shoot dry matter well for single, standing millet and cowpea plants and proved useful for a growth analysis of millet, its reliability was unsatisfactory for groundnut. For millet, the slope of two separate regressions between the difference of reflectance ratios and shoot dry matter, taken 10 days apart, was almost identical. This suggests a possible simplification in future crop growth models. The usefulness of the instrument for estimating dry matter in natural species mixtures such as weeds and grasses depended on the homogeneity of the mixture and the uniformity of their physiological state.