Annual or usually perennial herb; in drier situations with slender, more solid stems, erect, ± much-branched, to c. 30 cm; in wetter places ascending or commonly prostrate with stems c. 0.1-1 m long, rooting at the nodes, ± fistular, with numerous lateral branches; when floating very fistular, the stems attaining several metres in length and over 1 cm thick, with long clusters of whitish rootlets at the nodes. Stem and branches green or purplish, with a narrow line of whitish hairs down each side of the stem and tufts of white hairs in the branch and leaf axils, otherwise glabrous, striate, terete. Leaves extremely variable in shape and size, linear-lanceolate to oblong, oval, or obovate-spathulate, 1-9 (-15) x 0.2-2 (-3) cm, blunt to shortly acuminate at the apex, cuneate to attenuate at the base, glabrous or thinly pilose, especially on the lower surface of the midrib; petiole obsolete or up to c. 5 mm. Inflorescence sessile, axillary, solitary or in clusters of up to c. 5, subglobose (or somewhat elongate in fruit), c. 5 mm in diameter; bracts scarious, white, deltoid-ovate, mucronate with the excurrent pale midrib, glabrous, c. 0.75-1 mm; bracteoles similar, 1-1.5 mm, also persistent. Tepals oval-elliptic, equal, 1.5-2.5 mm, acuminate to rather blunt, white, glabrous, shortly but distinctly mucronate with the stout, excurrent midrib, the margins obscurely denticulate. Stamens 5 (2 filaments anantherous), at anthesis subequalling the ovary and style, the alternating pseudostaminodes resembling the filaments but usually somewhat shorter. Ovary strongly compressed, roundish, style extremely short. Fruit obcordate or cordate-orbicular, 2-2.5 mm long, strongly compressed with a narrow, pale, somewhat thickened margin. Seed discoid, c. 0.75-1 mm, brown, shining, faintly reticulate.

Distribution: A common species, very widepsread in the tropics and subtroprics of both Old and New Worlds in waste and cultivated ground, especially in damp or wet conditions.

No specimen of Alternanthera nodiflora R. Br. has been seen from Pakistan, and the record by Hassanain quoted by Stewart (but with no reference) must therefore remain suspect.