Small tree to 10 m, branches somewhat pendulous. Branchlets dark brown, villous when young. Buds ovoid, reddish brown, scales with slightly pubescent margins. Leaf blade 6–12 × 3.5–7 cm, elliptic to ovate, sparsely pubescent along midrib above and along the midrib plus lateral veins below; base rounded to cuneate, apex acuminate; margins with irregular small teeth; petiole 2–4 cm long, villous. Inflorescence a 4–8-flowered corymb, 4–6 cm across; pedicels 2–4 cm, pubescent. Flowers 2.5–3 cm diameter, May-June in China. Sepals triangular-lanceolate, densely hairy above, falling before fruit is ripe; petals white, obovate, 1.2–1.5 cm; stamens 25, unequal, shorter than petals; styles 4 or 5, slightly longer than stamens. Fruit red, ovoid or subglobose, 1–1.5 cm diameter, September (China). (Gu et al. 2003; Grimshaw & Bayton 2009).
Distribution Bhutan (perhaps) China SW Sichuan, SE Xizang, NW Yunnan.
Habitat Mixed forest in valleys, 2400–3800 m asl.
USDA Hardiness Zone 4-8
RHS Hardiness Rating H6
Conservation status Least concern (LC)
This is one of the more taxonomically troublesome plants in the orbit of M. baccata. As seen in cultivation, albeit rarely, it is a small tree showing good disease resistance, with abundant white flowers from pink buds, and bright red fruits which persist well (Fiala 1994; Grimshaw & Bayton 2009). At this point, any semblance of certainty ends.
Malus rockii was described by Rehder (1933) from a series of specimens collected in Yunnan by the indefatigable Austrian-American plant collector and ethnographer Joseph Rock (Rock 5346, 6842, and 11552 of 1922–23). Rehder noted its affinities with M. baccata, distinguishing it by the much hairer leaf underside and the way in which the calyx falls rather late. He also noted its likeness to hybrids between M. baccata and either M. domestica or M. prunifolia, although these have a persistent calyx. It was introduced to cultivation at the Arnold Arboretum through Rock’s seed collection 23380 of 1932, again from Yunnan. Material from this collection as grown at the Arnold is tetraploid (H. McAllister pers. comm. 2020) and hence presumably apomictic.
Multiple taxonomic treatments reflect the uncertainties around this plant. The Flora of China (Gu et al. 2003) treats it as a full species, but includes the Bhutanese M. baccata var. himalaica within the concept; but Rushforth (2018), with considerable field experience in these areas, considers that they differ in several characters. Taylor (1983) treated the cultivated plant from Rock 23380 as a cultivar, M. sikkimensis ‘Rockii’, an approach currently adopted in the Arnold Arboretum catalogue (Arnold Arboretum 2020). The RBG Edinburgh catalogue treats it simply as a synonym of M. baccata (Royal Botanic Garden Edinburgh 2020). A specimen from an Edinburgh tree (now gone) raised from Forrest 13001 of 1914 (Mekong Valley, Yunnan) has been determined in Paris as M. rockii (Muséum National d’Histoire Naturelle 2012). More recently, seed has been collected as M. rockii several times in Yunnan: such wild-origin material is being grown at Kew (M. Foster 93051) and Edinburgh (CLD 874 – Grimshaw & Bayton 2009). Any identification within this difficult and poorly understood group is subject to dispute and revision. For example Keith Rushforth (pers. comm. 2020) feels that KR 2824 (ibid. M. Foster 93051), while collected as M. rockii, probably does not represent this taxon; he considers that CLD 1463 is the closest recent introduction to M. rockii.