An arching shrub to 8 or 10 ft high and as much wide; prickles sparse, short and straight, much thickened at the base and sometimes laterally compressed. Leaves 5 to 8 in. long; rachis usually glandular and prickly, slender. Leaflets seven or nine, elliptic or elliptic-ovate, acute to obtuse at the apex, usually not more than 11⁄4 in. long on the flowering laterals, but to twice that length on strong shoots (which may bear an inflorescence at the apex), medium green above, underside greyish, glabrous except for the hairy midrib and main veins, sometimes glandular on the blade. Flowers in June, up to twenty or even more in a lax cluster, deep purplish pink passing to white at the centre, about 2 in. wide; bracts large and leafy. Pedicels 1⁄2 to 2 in. long, more or less densely clad with glandular bristles, which may extend onto the rather narrowly ellipsoid receptacle. Sepals about as long as the petals, glandular at the margin, expanded at the apex and with a few lateral appendages. Petals downy on the back in the type but only very slightly so in authentic cultivated plants. Fruits flagon-shaped, 1 to 2 in. long, dark red, crowned by the sepals.
A species of limited distribution in China; discovered by Wilson in N.W. Hupeh in 1901 and introduced by him at the same time to Messrs Veitch’s Coombe Wood nursery, where it first flowered in 1909. It is one of the most handsome of the Chinese roses, recalling R. moyesii and R. davidii in its fruits and perhaps most nearly allied to the latter.
R. caudata Bak. – Very near to R. setipoda and from the same area. It was accepted as a distinct species by Boulenger, on the ground that the inflorescence is shorter than the subtending leaves (longer in R. setipoda). It was described from a plant in Miss Willmott’s garden, raised from the seeds collected by Wilson during his first journey for the Arnold Arboretum (1907-8).
R. hemsleyana Täckholm R. setipoda sens. Rolfe, (?) not Hemsl. & Wils. – This species, if such it be, was described from a plant at Kew, which had been raised from the same batch of seed as R. setipoda (W.1047). It differs primarily in its chromosome number (it is hexaploid, while R. setipoda sens. strict. is stated to be tetraploid), but also in having sepals with unusually well developed lateral appendages, as in the dog roses. A similar specimen was collected by Wilson in the wild in 1907 and is referred to R. setipoda in Plantae Wilsonianae (W.272). R. hemsleyana is figured in Bot. Mag., t. 8569, as R. setipoda of which it is probably no more than a form.