Hydrangea heteromalla D.Don

This wide ranging, variable, and relatively hardy species has always been in the second rank of garden hydrangeas, at least in popularity. Its great distinction is height: most often forming a mighty shrub, it is capable of producing a genuine small tree, especially in an open position with good formative pruning. The toothed, more or less sharply pointed leaves are a matt dark green above, woolly-white beneath, and sometimes have red petioles. The flowers are in flat corymbs ringed by well-developed sterile flowers, typically white or cream and sometimes developing pink or red tones, otherwise ageing pale brown. Good forms have large, plentiful inflorescences. It will grow in shade but like many hydrangeas flowers best in full sun. Some gardeners have found it to be a prolific self-seeder (Bean 1981; Haworth-Booth 1984). H. heteromalla lies at the centre of a baffling species complex. Variation patterns in wild plants are poorly understood, and there is ongoing uncertainty over how many (if any) other species should be segregated. Our aim here is to reflect this uncertainty honestly.

It has long been understood that the H. heteromalla complex has affinities with H. paniculata (eg. Diels 1900). Molecular evidence supports this close relationship (De Smet et al. 2015; Yang et al. 2024). Inflorescence form is the single most reliable way to distinguish them, with a single elongated axis in H. paniculata, both typical forms with long panicles and ‘flat panicle’ forms from northern Japan. In H. heteromalla s.l. there are usually about 3 shorter axes, and the resulting corymb is more or less flat (Wei & Bartholomew 2001; McClintock 1957).

Hydrangea heteromalla was first described from the western end of its range in Nepal by Don (1825); it was also illustrated but not described as H. vestita around the same time (Wallich 1826). Many other rather similar taxa were subsequently described from across a swathe of Asia between the western Himalaya, north-central China and northern Vietnam, either as full species or at various infraspecific levels, differing mainly in leaf shape and hairs. This raises the question of whether these are distinct entities or merely points on a multifactorial continuum. To this day it has not been satisfactorily answered. In the only remotely modern revision of Hydrangea, McClintock (1957) treated them all as belonging to a variable H. heteromalla, recognizing no infraspecific taxa at all. Her view is still just about tenable, although in some molecular phylogenies this would leave H. heteromalla paraphyletic (eg. Yang et al. (2024) have H. bretschneideri as sister to H. paniculata). On this view, Western horticulture is left with the problem of how to name those wild variants which in gardens seem significantly different from the typical species, most importantly H. bretschneideri and H. xanthoneura. Groups are an ideal answer to this conundrum; Bretschneideri Group is often used (e.g. Edwards & Marshall 2019) although Xanthoneura Group seems never to have been published.

However, some of these taxa have never been abandoned by botanists. The Flora of China treatment (Wei & Bartholomew 2001) nicely illustrates multiple botanical opinions. The published text reflects Wei Zhaofen’s multi-species view, accepting in China alone HH. bretschneideri, dumicola, hypoglauca, macrocarpa, sungpanensis and xanthoneura, in addition to H. heteromalla. Co-author Bruce Bartholomew, supported by editor Wu Zhengyi, noted that he would have preferred to sink all these into H. heteromalla. A trawl of research literature (pers. obs. 2024) plus a look at the numbers of images and herbarium specimens attributed to segregate species on the PlantPlus website (Chinese Academy of Sciences 2024) shows some of these species concepts (especially H. bretschneideri and H. xanthoneura) to be alive and well in China. In the West too, Plants of the World Online currently accepts HH. bretschneideri, hypoglauca, macrocarpa, and xanthoneura (Royal Botanic Gardens, Kew 2024). It is worth noting that while H. hypoglauca fits morphologically within the heteromalla complex, one plastid DNA study places it in Section Chinenses (Yang et al. 2024), hinting perhaps at a hybrid origin. We include below a key to all species recognized in the complex by Flora of China, based on Wei & Bartholomew (2001). Do not expect too much of it.

Reflecting the uncertainty we include separate, brief pages for H. bretschneideri and H. xanthoneura which are definitely in cultivation, even if there are issues around defining their morphological limits. H. bretschneideri extends much further north in central China than does H. heteromalla s.s., which is confined to Yunnan, Sichuan and southern Xizang (Wei & Bartholomew 2001). H. bretschneideri is reputedly hardier than typical H. heteromalla of gardens, has peeling bark, tends only to make a shrub to 3 m or so, and while the leaf undersurface may have some white hairs, especially at first and on the veins, it does not have the velvety, white indumentum of H. heteromalla s.s.. H. xanthoneura, which extends further east from Yunnan and Sichuan than typical H. heteromalla, into Guizhou, has a tree-forming reputation, has leaf undersurfaces without hairs except on the veins, and unlike H. bretschneideri has conspicuous lenticels even on first year branchlets (Wei & Bartholomew 2001; Edwards & Marshall 2019; Haworth-Booth 1984). In truth, while it is quite easy to point to ‘typical’ examples of each species, the dividing lines seem quite hazy. Several varieties in the complex, most described by Alfred Rehder (1912) and in Sargent (1913) but now largely ignored, shifted confusingly between species and have not always been sunk into the species in which they were first described (Royal Botanic Gardens, Kew 2024). Other arguably distinct segregate species have been in cultivation, but are rarely identified as such today, making it harder to know where they cling on, if at all. For example H. dumicola W.W.Sm. (needle-tipped white hairs along the length of veins beneath; bluish anthers) was described from two George Forrest collections (Smith 1917) and also introduced by Forrest (Foster 2023); it made a small tree in one of Maurice Mason’s Norfolk, UK, gardens (Foster 2023). A plant so-labelled at the Savill Garden, Surrey, UK, had thin, ovate-lanceolate leaves with very sparse, spreading, needle-like hairs on the veins beneath, but the anthers had already aged brown when examined (pers. obs. June 2024). Horticulture suffers when wide-sense species names are adopted without first providing appropriate infraspecific names (such as Groups) for plants which are distinct in the garden.

Typical H. heteromalla has apparently been grown in Europe for longer than most other hydrangeas. First introduction is usually given as being in 1821, from the Himalaya, without further details (Bean 1981; Haworth-Booth 1984). If this is correct (and primary sources prove hard to find) introduction to Britain via the Calcutta Garden seems most plausible. Bean (1981) felt that by his day most or all garden stocks derived from much later (little-documented) introductions. Ernest Wilson collected many plants from the complex (Sargent 1913) but apparently none attributable to H. heteromalla in the narrowest sense. Late 20th century collections made on botanical expeditions and labelled H. heteromalla are too numerous to list (see for example Royal Botanic Garden Edinburgh 2024). Those from the Himalaya (notably ENEP 191 – Nepal, 2001) are likely to be narrow sense H. heteromalla, while some from southwest China might equally represent other arguably distinct species. Crûg Farm have made numerous commercial collections, many in conjunction with Dan Hinkley: BSWJ 2142 (West Bengal, India), HWJK 2721 (NE Nepal, bordering Tibet), and HWJ 526 & 938 (Fan Si Pan, N Vietnam, 2003); BWJ 7657 (Sichuan-Yunnan border, 2000) also belongs somewhere in the complex.

The 1990 Cangshan, Lijiang, Dali (CLD) Expedition to NW Yunnan made several collections of this complex, simply as Hydrangea spp. Groups of original seedlings planted by Jim Russell at the Yorkshire Arboretum, UK, provide an unusual opportunity to study variation within wild seed batches decades after collection, as well as to pick out desirable individuals for vegetative propagation (J. Grimshaw, J. Sutton pers. obs. 2024). Most individuals from all batches conform well to H. heteromalla in the narrow sense, with leaves broadly ovate, broadly cuneate to subcordate at the base and quite abruptly, shortly acuminate, with short, broad, forward pointing teeth; leaf undersides are densely, rather stiffly white-woolly, and the major veins tend to curve strongly towards the apex. However, single individuals from each of CLD 799, 1442 and 1503 have a quite different ‘look’ in leaf, with narrower leaves, elliptic to obovate, more definitely tapering at both ends, with straighter veins, thinner, spreading teeth and a thinner texture; they have similar hairs across the lower leaf surface, but less densely, without the woolly feel of the others. Russell was a careful man with plenty of experience collecting Hydrangea and other seed in the field: we must assume that these odd plants came from the same parent as the typical ones, that there was no curatorial mix-up. Perhaps they represent hybrids with another member of the complex, or even some other hydrangea? Interestingly a group of three plants from B&L 12294 collected in the same area also have leaves of this type.

The entire complex seems hardy and growable in the British Isles (Foster 2023), Belgium (Botanical Collections 2024) and milder parts of France. In colder areas of central and northern Europe the hardier H. bretschneideri comes to the fore – it is even hardy in southern Finland (Arboretum Mustila 2024). None are very widely grown in North America, although Pacific Coast gardens suit them (Dirr 2021). Specimens from many sources, including KR 9681 & 9913 from India, as well as Chinese provenance plants which might include segregate taxa, grow at UBC Botanical Garden, Vancouver (University of British Columbia 2024). It seems likely that only hardier members of the complex are suitable for the Atlantic seaboard. At the US National Arboretum, Washington, DC, whose catalogue distinguishes segregate species, only H. bretschneideri and H. xanthoneura grow in the public garden (US National Arboretum 2024). While the Arnold Arboretum catalogue is inconsistent in its treatment, available details strongly imply that only these same two taxa are in the living collection (Arnold Arboretum 2024).

No well documented hybrids between the heteromalla complex and other hydrangeas are known, either in the wild or in gardens. Maurice Foster (2023) emphasizes that he has never seen signs of hybridity or introgression between the heteromalla and aspera complexes in natural populations, the former always growing at higher elevations, without overlap. Some have postulated that ‘Nepal Beauty’ (see below) might be such a hybrid, but there is no conclusive evidence. Similarly, gardeners have sometimes suggested that early flowering, flat panicled H. paniculata ‘Dharuma’ might be a hybrid with H. heteromalla (van Gelderen & van Gelderen 2004) but it proves to be a pure, hexaploid H. paniculata of the sort found in northern Japan (Beck & Ranney 2014).

We list all cultivars belonging to H. heteromalla in the widest sense together under ‘Hydrangea heteromalla Species Complex Cultivars’. Some may belong to segregate species (noted where this is suspected), some may be hybrids between them. Inherent uncertainty makes grouping them the only useful course. Few cultivars are grown in North America; few are close to mainstream even in Europe.