Annuals or perennials. Leaf-blades mostly linear and flat; ligule short, scarious or membranous, glabrous or ciliate. Inflorescence composed of racemes, these digitate or borne upon an elongated central axis, rarely solitary, sometimes with secondary branchlets; rhachis flat or triquetrous, bearing the spikelets in appressed groups of 1-5 or more, their lower glume abaxial; pedicels terete or triquetrous, usually smooth or scaberulous. Spikelets lanceolate to oblong-elliptic, flattened on the front, convex on the back; lower glume small or suppressed; upper glume membranous, as long as the spikelet, or much shorter and exposing the upper lemma; lower floret barren, represented by a prominently nerved membranous lemma as long as the spikelet (rarely much reduced), usually hairy, typically with the hairs forming stripes between the 1st and 2nd lateral nerves and along the margin; upper floret bisexual, the lemma chartaceous to cartilaginous, finely longitudinally striate, with its hyaline margins enfolding and concealing most of the plea, ± dorsally compressed, usually rounded on the back, subacute to acuminate, rarely rostrate; caryopsis oblong, plano-convex in section, mostly acute to subacute.

A genus of about 200 species in tropical and warm temperate regions, particularly in the Old World; 14 species occur in Pakistan, 2 of them in cultivation only, 1 an alien weed.

Digitaria is not an easy genus to name. It is traditionally subdivided according to the nature of the hairs, but this requires a microscope, and the different kinds of hair form a continuous spectrum rather than discrete groups, with a sprinkling of glabrous variants to complicate the situation. Reliance is here placed early on in the key upon spikelet grouping. This works reasonably well, but is sometimes deceptive. Ternate spikelets often have the longest pedicel fused to the rhachis, so that the spikelets seem to be alternately solitary and paired; sometimes one of the spikelets is vestigial. Binate spikelets are often associated with the development of secondary branching, which can be misleading; in particular the pedicel is sometimes forked so that the spikelets appear ternate, but the pedicels of truly ternate spikelets all arise from the same node.