This species is characterized by its small, deciduous, bidentate stipules that enclose well developed pilose colleters, which are almost as long as the stipule, and leave a well developed scar when they fall; its thin-textured, elliptic to oblanceolate, usually not particularly large leaves; iits rather short corymbiform inflorescences with the flowers subsessile and grouped into several dense clusters; its small corollas; and its small ellipsoid fruits. It is rather commonly collected. Occasional plants have axillary as well as terminal inflorescences (e.g., Antilahimena & Marcelin 5587, Bernard 574) but otherwise match Psychotria subnubila. Such axillary inflorescences are characteristic of the species of Bremekamp's Mapouria Group I, but those species have very different pyrenes.
Bremekamp recognized three varieties of this species. He included glabrous to "subglabrous" plants in var. subnubila, plants with pilose pubescence that dries ferrugineous on the inflorescence axes and the abaxial surface of the leaf midveins in var. rufo-pilosa, and plants with obovate leaves and velutinous pubescence that dries ferrugineous on young stems, petioles, the veins on the lower leaf surface, and the inflorescences in var. rufo-velutina. The pubescence visible in the image of the type specimen of var. rufo-pilosa corresponds to the pubescence called hirtellous here on other Madagascar Psychotria species, and is ferrugineous as Bremekamp described it; no modern material that corresponds to this form has been seen, although several specimens have sparse pilosulous pubescence that dries whitened distributed across all of the lower leaf surface (e.g., Antilahimena 6992, Razafindraibe & Antilahimena 38). The form of the pubescence of the type of var. rufo-pilosa does correspond to that of several other specimens that are densely ferrugineous-hirtellous on stems, petioles, the lower leaf surface, and the inflorescences (e.g., Kotozafy 53, 54, and 1012), and these are presumed to represent var. rufo-velutina. The plants that are pubescent on the undersides of the leaves do seem to represent the same species as the far more common glabrous plants, and are from scattered geographic localities and are here considered to represent represent local variation within Psychotria subnubila and are not separated.
However the plants that are densely hirtellous on the stems and leaves have different stipules, which are are triangular and entire and leave a thinner scar with shorter paler trichomes. These key out to Psychotria subnubila except for their stipules, and are also similar to Psychotria integristipulata. The type of var. rufo-velutina has not been seen to confirm if it is a different plant that is conspecific with Psychotria subnubila, or if it matches these specimens keyed as var. rufo-velutina. Psychotria integristipulata was classified in Mapouria by Bremekamp, but lacks domatia similarly to species of Bremekamp's Psychotria and Bremekamp did not actually see any mature fruits of Psychotria integristipulata to confirm its generic classification. These hirtellous plants are here provisionally included in Psychotria integristipulata but should be compared to the type of var. rufo-velutina.
As provisionally circumscribed here, Psychoria subnubila is found in humid middle elevation forests and includes a range of variation in leaf size and shape and degree of development and number of flowers in the inflorescences; eventually it may be found to include more than one species. It is similar to Psychotria decaryi, from generally lower and/or dry areas and further south, and these may deserve further study as to the details of their separation. Psychoria decolor is also similar, and its type may not be distinct. Bremekamp separated these based on number of flowers in their inflorescences, but this varies within Psychotria subnubila apparently more than he saw (e.g., material from Ambotavy area).