Hydrangea involucrata Siebold

With its roughly hairy leaves and striking ‘drumstick’ floral buds, Hydrangea involucrata is a most distinctive small to medium shrub. Plagued by a reputatation for tenderness and poor flowering, it has been grown too little; contemporary experience in milder parts of our area is far more positive. Its ability to flower on the current years shoots is very valuable. A modern proliferation of cultivars with mutant flower / inflorescence form, as well as colour variants, adds to its versatility in the garden.

Both leaf surfaces, young stems and inflorescences are covered in rough, bristly, more or less appressed hairs. It gives the impression of a smaller-leaved relative of the H. aspera complex, a relationship confirmed by molecular data (De Smet et al. 2015). However, the extraordinary, almost spherical inflorescences tightly enclosed in large bracts to give a drumstick or lollipop appearance resemble those of Section Cornidia, which in turn is closely related to the Asperae. This feature inspires the Japanese vernacular name, translating roughly as ‘ball hydrangea’. The only potentially confusing species is its close Taiwanese relative H. longifolia (q.v.), which shares this inflorescence form but has distinctive 2-branched hairs; it is very rare in gardens. McClintock (1957), who had only seen a photograph of the type specimen of H. longifolia, treated the two species as synonymous. As a result many authors continue to list H. involucrata as occuring in Taiwan.

The bracts fall away as the infloresence opens to a more or less domed corymb. In typical plants they have lacecap form, with bright lavender fertile flowers, and 4–6 white to pale lavender sterile flowers which age green. A later-blooming species, peak flowering is usually in August both in Britain (Foster 2015) and North America (Dirr 2021) but this varies a little with climate and to some extent with clone. On many clones it will flower on both old and (later) the current years shoots, extending the season. The species has long been cultivated in Japan, where most cultivars originated, some as wild finds. These include various double-flowered and mop-headed forms as well as colour variants. The botanical forms f. hortensis (Maxim.) Ohwi (most flowers sterile), and f. plenissima Tuyama (flowers double, with 8–10 petaloid sepals) have been used for groups of mutants (Ohba 2001), but individually named cultivars are more relevant to gardeners (see below).

H. involucrata was named scientifically by the doctor, naturalist and student of all things Japanese Philipp von Siebold (1829). He is generally credited with its introduction to Europe, although the precise date is harder to place. A mention of it flowering in the Botanic Garden in July (Siebold 1829) probably refers to his own Dejima Botanic Garden, Japan, rather than Ghent, now in Belgium, where his earliest European introductions were grown (Compton & Thijsse 2013). Edwards & Marshall (2019) give 1864 as the earliest known date of Western cultivation. Siebold & Zuccarini (1870) illustrate both a wild type plant and a double-flowered cultivar, although ‘Hortensis’ – the first double known in the West – was probably not introduced until 1906 (Bean 1981). Most other cultivar introductions from Japan are relatively recent. Commercial collections of wild-type plants include BSWJ 4790 of 1997 and 11578 of 2006, from the Chiba Peninsula (Crûg Farm Plants 2023); Dan Hinkley has collected in the same area (Larson 2018; Crûg Farm Plants 2023).

Ideal growing conditions include good light levels, with some shade at least in the middle of the day, and moisture throughout the growing season; full shade does not seem to inhibit flowering, though, even in the UK (Foster 2015). Some cultivars will take more sun than others without the leaves bleaching, especially in cooler maritime areas such as America’s Pacific Northwest (Larson 2018). Ishii (2014) makes a distinction between relatively sun-tolerant coastal plants and mountain-provenance plants which require more shade, drawing a parallel with the different needs of H. macrophylla and H. serrata. She gives ‘Yoraku’ and ‘Mihara Kokonoe’ as examples of coastal types, ‘Tama Kangashi’ and ‘Kokonoe’ (neither described here) of mountain types, although it is unclear how clear-cut this distinction is, and where many cultivars fit in this scheme. Light intensity affects flower colour in some double cultivars (Foster 2015).

Older literature is full of concerns around hardiness. Bean (1981), writing from an early to mid 20th century London perspective, considered it ‘not very hardy, […] often killed back more or less in winter’ and thought it better suited to western Britain. Recent experience is more positive, perhaps due to both climate change and de facto selection of hardier clones, with some succeeding even in more continental climates (Foster 2015). Even so, the species is able to flower on current year’s shoots when cut to the base, the plant remaining under 60cm tall (rather than 1–2 m) when grown this way (Haworth-Booth 1984). Pruning without loss of flowering is thus possible, but in no way essential. Leaf hairs tend to trap flower debris and other detritus, which is worth shaking off to avoid necrotic patches forming (Foster 2015).

In North America the species can be grown on both seaboards. There are well established plants in Washington Park, Seattle (Larson 2018). In the Northeast Dirr (2021) records plants at the Scott Arboretum, PA (sometimes cut to the ground by frost), Cornell Botanic Gardens, NY, and the Coastal Maine Botanical Gardens (the cultivar ‘Gyokudanka’); he claims a ‘total absence’ from Southeastern gardens, but has himself flowered it in Athens, GA.

H. involucrata is most unusual in having a chromosome number 2n=30, rather than 2n=34 in most of the H. aspera complex, and 2n=36 in most other hydrangeas (Mortreau et al. 2010). While this does not preclude hybridization, it makes fertile hybrids with potential for ongoing breeding (especially in the female line) much less likely. Apart from the possibility of novel-looking plants, the long flowering season of H. involucrata would be a desirable trait in hybrids (M. Foster pers. comm. 2023). Several experimental hybrids exist; most are poorly documented and none seem yet to have gained cultivar names or any significant distribution. Mark Fillan (Devon, UK) and Ed Bowen (Rhode Island, USA) have made H. involucrata hybrids with members of the H. aspera complex and with H. sikokiana. Some of these are illustrated by Bowen (2011). Hybrids with the more distantly related H. arborescens have also been made (Jones & Reed 2006); only crosses using H. arborescens as the seed parent produced seedlings, and then only a few, using H. a. ‘Dardom’ (‘Annabelle’ crosses yielded nothing). Only early, uninspiring reports of flowering were published, and nothing more has appeared in the literature or nursery trade. There was evidence of very low levels of pollen fertility in one these hybrid plants.

The following cultivars are those most likely to be seen in our area. More continue to arrive from Japan, and naming is an ongoing issue. Double flowered forms are particularly hard to distinguish and potentially confused in the West. The French nursery Sous Un Arbre Perché (2023) offers one of the widest selections.