6. Mussaenda densiflora H. L. Li, J. Arnold Arbor. 24: 455. 1943.
密花玉叶金花 mi hua yu ye jin hua
Climbers, ca. 2 m tall; branches terete, densely yellowish brown villous or hirsute. Leaves opposite; petiole 5-7 mm, densely hirsute; blade drying membranous or thinly papery, broadly oblong-lanceolate or narrowly oblong-elliptic to usually ovate, 8-13 × 3-6 cm, adaxially moderately to sparsely strigose, hirsute, or hirsutulous, abaxially moderately to densely hirsute, pilose, or strigose with pubescence denser on principal veins, base obtuse to rounded and sometimes abruptly attenuate, apex acute or acuminate; secondary veins 8-10 pairs, tertiary venation visible and reticulate; stipules persistent, 8-10 mm, densely villous to hirsute, deeply 2-lobed, lobes lanceolate, narrowly triangular, or linear. Inflorescences capitate or subcapitate with 2-5 subglobose heads, 2-6 × 2-6 cm, densely villous, subsessile to pedunculate; peduncle to 3.5 cm; bracts lanceolate, narrowly triangular, narrowly elliptic, or linear, 10-15 mm. Flowers sessile, biology not noted. Calyx densely hirsute; hypanthium portion urceolate to ellipsoid, 2-3 mm; lobes linear-lanceolate to usually linear, 8-15 × 0.2-2 mm, sometimes 1 lobe of 1 or 2 flowers per inflorescence expanded into white calycophyll, blade ovate, 5-6 × 3-4 cm, moderately to densely strigose to pilose on both surfaces, base obtuse to cordulate, stipe 8-15 mm, apex acute to acuminate. Corolla yellow, outside moderately to densely hirsute; tube 22-25 mm, sometimes contracted around anthers; lobes narrowly lanceolate or orbicular-lanceolate, 6-8 mm, inside densely yellow papillose, long acuminate. Berry ovoid, 8-9 × 6-7 mm, glabrous, with calyx limb deciduous. Fl. May.
Sparse thickets; 300-800 m. Guangxi [Vietnam (Tsang 29049, P!)].
This species is said in the protologue to be similar to Mussaenda subsessilis Pierre, and it may be that with more specimens the distinctions between them (smaller calycophylls and longer narrower corolla lobes in M. densiflora) may turn out to be endpoints of continuous variation. H. H. Hsue and H. Wu (in FRPS 71(1): 300. 1999) described the stipules as caducous, but this has not been seen on specimens studied. The protologue described the corolla tube as 3-3.5 cm, but this does not match specimens seen, including the type; this may have been the length of the entire corolla that was labeled incorrectly in the final article.