A study has been made of the Indo-Malaysian species of Cnestis. The mutual length ratio of sepals and petals, — brevi- and aequipetaly —, is the main differentiating character for the species; there are no transitions. The areas of distribution overlap in the Malay Peninsula (fig. 1); brevipetalous types are known from the Malay Peninsula, Sumatra, Borneo and Celebes, aequipetalous types from Burma, Siam, Indo-China and the Andaman Islands, the Malay Peninsula and the Philippines. Fruits are of two different shapes: beaked in aequipetalae of the Andamans, Burma, Siam, and Indo-China, pear-shaped in remaining aequipetalae and in brevipetalae. Leaves tend to be longer and jugae more numerous in brevipetalae than in aequipetalae. Other characters do not have so clear a separating value, such as texture and indumentum of leaflets, indumentum of inflorescence, texture and indumentum of petals, length of stamens, type and length of pistils, length ratio of stamens and pistils. However, even on the strength of these characters there is some reason to distinguish both groups mentioned above. As to the indumentum of petals there is a remarkable cline in a decreasing sense from the Philippines to continental Asia, the Andamans and the Malay Peninsula and back to the east through the brevipetalae of Malay Peninsula, Sumatra, Borneo and Celebes. Brevi- and aequipetalae have been considered to represent two species, viz Cnestis platantha Griff. and Cnestis palala (Lour.) Merrill. The latter one has been divided into two subspecies, viz subsp. palala with beaked fruits and subsp. diffusa (Blanco) Andreas with pear-shaped fruits. For their area of distribution see fig. 1. In many respects some plants of the Andamans, Burma, Siam, Indo-China (and the Malay Peninsula) are different from the remaining aequipetalae, but not in a uniform way as to the various characters. Although there are some arguments for a further taxonomic subdivision, we did not think it advisable to introduce such a division at present. Our classification differs from the division as given by Schellenberg (1938). This was caused by the material on one hand, being more heterogeneous than Schellenberg described it, and, on the other hand, by the fact that some of the diagnostic characters used by him, in our opinion were not fit for use as such. Therefore a revision of Schellenberg’s system of the genus Cnestis seems desirable.