Trees or sometimes thicket-forming shrubs, deciduous, occasionally evergreen in cultivation in very mild conditions. Bark smooth, or fissured, or scaly. Branches often many, slender, ascending, sometimes from low on the trunk. Twigs slender, often zig-zag; buds ovate or long-pointed, smooth, many-scaled. Leaves alternate, ovate to lanceolate, with 6–34 pairs of lateral veins, latereal veins often neatly parallel and sometimes impressed; leaf-margin unlobed but serrate, often doubly so; petiole short (4–30 mm). Autumn colour yellowish, sometimes red. Plants monoecious, wind-pollinated; male flowers in short catkins on old wood with many overlapping bracts, developing in the autumn and opening from the bud just before the leaves in spring; stamens 3–12. Female flowers paired within terminal or axillary racemose catkins, developing in the spring, 2–50 cm long and becoming pendulous in fruit through summer; bracts leaflike (each formed of a bract and 2 bractlets), overlapping, toothed on one side or with 2 or 3 lobes, green finally ripening brown or whitish. Nutlet longitudinally ribbed, enclosed in some species by the infolded bract. (Li & Skvortsov 1999; Grimshaw & Bayton 2009; Bean 1976).
Along with the genera Corylus, Ostryopsis and Ostrya, Carpinus belongs to the Coryloideae sub-family of the Birch Family, Betulaceae. Within this cluster the genus, as currently understood, is closest to the hop-hornbeams (Ostrya), which differ in their fruit-bracts that form a sealed bladder totally enclosing the nutlet, and in their male catkins which, like those of the hazels (Corylus), are exposed through winter and do not remain protected within a bud. In Carpinus itself, four east Asian species are easily distinguished and are probably more likely to be confused with Ostrya; these trees – C. cordata, C. fangiana, C. japonica and C. rankanensis, currently placed by most taxonomists within the sub-genus or Section Distegocarpus (Siebold & Zucc.) Nakai – share the sometimes scaly bark seen in Ostrya, plus long leaves with numerous neatly parallel side-veins, and their greenish-white fruit-bracts enfold the nutlets and look rather like bunches of hops, though they are not sealed to form bladders. Recent phylogenetic analyses (Li 2008; Dong et al. 2022) have in this case tended to confirm the casual observer’s expectation, and suggest that Section Distegocarpus is closest to the ancestral forms of the Coryloideae, and that both the protective bladder of Ostrya and the aerodynamic wing of the nutlets of the remainder of Carpinus – Section Carpinus – are modifications of the sheath-like bract still seen in hazels. It is possible that taxonomic fashion will come to reflect this position by preferring to elevate Distegocarpus to genus level, but, for now, only two members of the Section have published names within Distegocarpus (as a genus). All four of these ‘false hornbeams’ – to invent a vernacular equivalent – are in cultivation, and all four are memorably ornamental trees, but none is yet common in western gardens.
The rest of Carpinus – Section Carpinus, the ‘true hornbeams’ – form a much larger but relatively uniform group. Two species are accepted in this account from North America and two from Europe into western Asia – plus the latter pair’s stable hybrid – but the group’s heartland extends from the Himalaya eastwards and northwards as far as Japan. Within the latter regions as many as seventy species have been described (Li & Skvortsov 1999), the great majority from isolated and often tiny highland refugia within China – such as C. insularis N.H. Xia, K.S. Pang & Y.H. Tong, named from the mountains of Hong Kong in 2014 (Holstein & Weigend 2017). Even more recently, C. langaoensis Z. Qiang Lu & J. Quan Liu was described from Langao County, Shaanxi in 2017 (Holstein & Weigend 2017), C. tibetana Z. Qiang Lu & J. Quan Liu was described from south-east Tibet in 2018 (Lu et al. 2018) and C. gigabracteatus Z. Qiang Lu was described from south-east Yunnan in 2020 (Lu 2020). Some of these populations have already been assessed by the IUCN as Critically Endangered (Shaw et al. 2014): C. putoensis W.C. Cheng is known from a single wild tree, estimated to be at least 250 years old, on the subtropical Chinese Zhoushan Islands; it is a polyploid of apparent hybrid origin (Dong et al. 2022) and its scions are cultivated in about six Asian collections (Shaw et al. 2014). C. tientaiensis W.C. Cheng has about 20 known specimens in Zhejiang, and is known to be in just one (oriental) collection (Shaw et al. 2014). C. hebestroma Yamomoto is endemic to the Hualian Xian mountains of eastern Taiwan; it was collected for the first time by Cédric Basset and Marion Rivière in 2013 and by 2022 a few small seedlings were in cultivation in France (Basset 2015; Pépinière des Avettes 2022). C. henryana var. oblongifolia (Hu) Rushforth, from Baohua mountain in Jiangsu and comparably threatened (Wang & Wang 2019), is not yet known to be in cultivation (Shaw et al. 2014).
Although a few of these Asiatic ‘true hornbeams’ are quite distinctive – Carpinus microphylla Z.C. Chen ex Y.S. Wang & J.P. Huang, described in 1985 from the Tianyang Xian mountains of western Guangxi, is a shrub to three metres tall with leaves only about a centimetre long (Li & Skvortsov 1999), and is not yet cultivated in the west – the identification of most taxa hangs on tiny differences in the shape and pubescence of the leaves, bracts and nutlets; conversely, some accepted species such as C. londoniana H.J.P. Winkl. can be highly variable in foliage characteristics (Li & Skvortsov 1999). There is no real consensus among contemporary botanists about which names do represent good species, or are better regarded as subspecies or varieties or others, or indeed treated as synonyms altogether. Recent phylogenetic analyses have also suggested that many forms which have long been understood as separate species are too close in evolutionary terms for DNA analysis in its current state to be useful in separating them: this is particularly true of a group of taxa for which C. turczaninovii Hance is the earliest published name and which includes several forms which are – relatively – familiar in western gardens, under names such as C. fargesiana H.J.P. Winkl., C. henryana H.J.P. Winkl., C. pubescens Burkill, C. shensiensis Hu and C. stipulata H.J.P. Winkl. (Dong et al. 2022). It seems likely that taxonomic fashion will respond to this situation by lumping together various of today’s Carpinus species; for now, this account follows the latest taxonomic checklist for the genus (Holstein & Weigend 2017) which acknowledges 45 species (including Section Distegocarpus and the two taxa in Section Carpinus named since 2017). Of this total, slightly more than half have a history – sometimes tenuous – of cultivation in the west; the uncollected species are all Chinese endemics.
This detour from the garden path into the thickets of academic botany should at least have convinced readers that the hornbeams are one of the most difficult genera of trees to come to terms with. (W.J. Bean rather jingoistically observed that the tangled shrubby growth of Carpinus orientalis ‘much impeded some of the advances of our men’ during the Crimean War of 1853–6 (Bean 1976); the same sensation must surely have been experienced, metaphorically, by many a more recent student of this genus). But difficult is not necessarily dull: though they lack showy flowers or colourful fruit, and though their leaf-shapes and habits all tend to closely resemble each other, hornbeams are often pretty trees and can be easy to grow. Some, including the common European Carpinus betulus L., are masters at surviving in heavy clay, which can flood in winter and bake almost as hard as concrete in summer: roots than can cope with these conditions will equally well be able to manage the dry and compacted conditions under roads and pavements. (However, the large numbers of hornbeam species from eastern Asia tend to expect ample rainfall through the growing season and can be drought sensitive; some are also of limited hardiness, and the habit of many will preclude their use in restircted spaces until novel selections begin to emerge). No hornbeam is known to be calcifuge. All seem resistant to browsing by deer – though are vulnerable to rabbits and Grey Squirrels – and suffer from remarkably few pests and diseases in general; the rapid decline and death from 2019–21, from unknown causes, of one of Britain’s biggest hornbeams, at Brede in East Sussex, was a remarkable event (Tree Register 2022).
As relatively low if spreading forest trees, all hornbeams are also used to growing in shade and can be planted underneath other, bigger specimens, though they seldom object to full sun; they have the elegance characteristic of shade-tolerant plants, with level or drooping planes of dense and neatly-arranged foliage. C. betulus is not at all salt-tolerant, and specimens miles from the sea can look brown and tatty after stormy weather when the trees around them stay green, but thriving specimens of C. turczaninovii, collected in Gansu in northern China, provide two of the highlights of that most windswept of historic gardens, Highdown in West Sussex, UK. Many hornbeams have good autumn colour – yellows predominate – and in winter they share the grace of some maples, with their often smooth grey bark, fine twigs, and many, slender branches. (Another Acer, A. carpinifolium, is celebrated for carrying leaves which look just like the ‘average’ hornbeam’s, but maples can of course be distinguished by their opposite budding). The various Chinese Carpinus from low latitudes but great elevations, which are used in the wild to high levels of solar irradiation as the delicate new leaves expand, tend like many other plants from these habitats to produce abundant protective anthocyanins, which give a red or purple tint to the young growths and which, in a vigorous young tree, can last through the growing season. These are the brightest colours that hornbeams have to offer, and a peculiarity within the genus seems to be that the red can persist underneath the leaf after the upper surface has turned green, at least in juvenile specimens. This feature is seen in some populations of C. omeiensis Hu & W.P. Fang (KR 280), in at least one plant of C. kawakamii Hayata at the Sir Harold Hillier Gardens (International Dendrological Foundation 2022), and in a Tom Hudson collection from northern Vietnam whose foliage features agree with C. londoniana. In mild Cornish winters at Tregrehan, a seedling from this collection is fully evergreen and red-leaved year-round so far (T. Hudson pers. comm.). As is the case with the purple-leaved Sycamore, the colour is particularly attractive – at once brilliant and subtle – when the light shines through the bipigmented leaf.
The many rather parallel lateral veins of the leaves of C. betulus are a characteristic signature of the genus as a whole, as is the way these veins are slightly impressed to create a pleated effect, and the tendency of the fine teeth tipping each vein to be interspersed with several smaller teeth to create an intricate fringe. In other species, the intermediate teeth may be relatively tiny, less numerous, more distant or absent, and all the teeth may be minutely whisker-tipped. The number of vein pairs is also a useful feature for species identification, though as these diminish near the leaf-tip the total can be hard to count consistently – observers would be well advised to consult multiple leaves from multiple branches. The ovate leaves of C. betulus tend to the oblong, with slight ‘shoulders’; the leaves of some oriental hornbeams are much more slender and elegantly longer-tipped.
By contrast, the almost symmetrical three-lobed shape of the fruit-bract of C. betulus is an the exception within the genus, and is only really shared with the southern Chinese C. tientaiensis Cheng (Li & Skvortsov 1999), a species not known to be in cultivation in the west. The ‘false hornbeams’ in Section Distegocarpus have very different, hop-like bracts, while the commonest shape for the bract within Section Carpinus is D-like, with a slight auricle or fold at the base of the straight side to clasp the nutlet and with a few jagged teeth on the curved side, which near the base can sometimes be big enough to count as lobes. Recent phylogenetic analyses (Dong et al. 2022) have shown that a similarity in bract shape does not necessarily indicate a close evolutionary relationship, as taxonomists used to assume: the American species C. caroliniana Walter and C. tropicalis (Donn Sm.) Lundell are closest to the east Asian C. viminea Wall. and have very similar bracts, but C. betulus turns out to be quite close to the east Asian C. tschonoskii Maxim, whose bracts are unlobed.
The word ‘carpinophile’ – a lover of hornbeams – was coined, tongue in cheek, by Michael Dirr (Dirr 1983). Implicit in the gentle comedy of the orginal usage is the understanding that such an understated genus might win some affection among gardeners, but would hardly seem likely to inspire love, and in 1985 Keith Rushforth characterised Carpinus as ‘a neglected genus’ (Rushforth 1985). But in subsequent decades, against expectation, these have become ‘fashion trees’, much in the way that Abies and Picea were to the landowning aristocracy of a hundred years earlier – and in marked contrast to some other extensive, difficult and unshowy genera, such as Celtis or Salix. Latter-day curators of an Edwardian pinetum will struggle to source replacement plants for many of its original subjects, but recent editions of the RHS Plant Finder (Royal Horticultural Society 2020) list the great majoriy of the hornbeam species which have ever been grown in the west, most of them available from more than one specialist nursery. In turn, planters of modern arboreta tend to include a good suite of hornbeam species. In such cases, the relative position of the cart to the horse can be hard to decide: nurseries will respond to demand, but will also be keen to promote plants which they find relatively easy to grow. (The smaller-leaved asiatic hornbeam species are also favourites among bonsaists). The UK National Collection of Carpinus has been built up since the 1970s at the Beale Arboretum in the grounds of the West Lodge Hotel on the northern outskirts of London.
Outside such big collections, the cultivation of Carpinus has largely been a matter of selecting sports of C. betulus, plus – in the United States and Canada only – of the indigenous C. caroliniana. Neither of these species has proved particularly obliging, and the named variegated, cut-leaved, weeping and ‘purple’ clones tend to be poor things. Of the compact or erect sports, C. betulus ‘Fastigiata’ is by no means svelte, but is architecturally very satisfying and remains by far the most popular hornbeam to plant. In recent years, Cédric Basset and Manon Rivière have introduced some variegated Japanese garden forms of oriental Carpinus species to France. There have been few attempts to create new ornamental hornbeams through hybridisation; at the Centre for Landscape Plant Development in Kansas, Susan Wiegrefe has experimentally crossed C. betulus, C. caroliniana, C. orientalis Mill. and C. viminea Lindl., in the hope of creating shapely trees which will be hardy in the coldest parts of the United States (Dirr 2009).
There will probably be some impetus to introduce more Chinese hornbeams into cultivation for the sake of ex situ conservation, but these minor species are unlikely to add very much to the group’s suite of garden features. Within such a difficult genus, it is also inevitable that a few plants already in cultivation cannot yet be identified to species level. These include SICH 2415 at Kew and Coombes 901, collected outside Kunming in China in 2004, which is semi-evergreen in southern England but leafs out very late in spring (A. Coombes pers. comm.) and which has reached 4 m at Penrice Castle in south Wales (Tree Register 2022). A miniature hornbeam at Caerhays in Cornwall, obtained from a commercial source under the non-existent name ‘Carpinus nimpoli’ (Williams 2020; 2021), has a reddish flush and is drought-sensitive, suggesting a Chinese origin. The online species list for Tremenheere, also in Cornwall (Tremenheere Sculpture Garden 2022), includes C. chingiana C.C. Yang, a questionable taxon named in China in 1997 (Holstein & Weigend 2017), but this plant is likeliest to have been the unrelated Tilia chingiana (N. Armstrong pers. comm.). C. tsaiana Hu, a tall-growing species from subtropical southern China, is proposed by Randy Stewart (Stewart 2010) as an ornamental tree of promise for the deep south of the United States, but does not yet seem to be in cultivation.
It is particularly hard to devise a reliable key for a genus which is at once so uniform and so variable. Resources include the ‘Keys to Carpinus and Ostrya’, published by Keith Rushforth in 1987 (Rushforth 1987) and adapted from the key in the 1959 Flora Reipublicae Popularis Sinicae, and Jan De Langhe’s ‘Vegetative Key to Species Cultivated in Western Europe’, which is accessible online (De Langhe 2022). The key presented here draws on both these sources but reflects the very close evolutionary relationships revealed by recent phylogenetic studies, and attempts to indicate the apparent variability of wild populations. However, only the species known to have been introduced to the west are featured.
|1a||Fruit-bracts wrapped around the nutlet on both sides. Leaf-base often deeply cordate; veins parallel and in (many) more than 15 pairs (Section Distegocarpus)||2|
|1b||Fruit bract flat at least on one side, to form an aerodynamic wing. Leaf base rounded, rarely shallowly cordate; veins usually in less than 18 pairs (Section Carpinus)||5|
|2a||Leaf with 24–34 pairs of veins; fruiting catkin 45–50 cm long||Carpinus fangiana|
|2b||Leaf with 15–24 pairs of veins; fruiting catkin to c. 20 cm long||3|
|3a||Leaf base usually deeply cordate||4|
|3b||Leaf base shallowly cordate to rounded||Carpinus japonica|
|4a||Leaf with 15–20 pairs of veins||Carpinus cordata|
|4b||Leaf with 20–25 pairs of veins||Carpinus kawakamii|
|5a||Fruit-bract 3-lobed, almost symmetrically||Carpinus betulus|
|5b||Fruit-bract asymmetrical, with 1 or 2 small basal lobes or none||6|
|6a||Fruit-bract with a small basal lobe or lobule or auricle on each side||7|
|6b||Fruit-bract (narrowly) D-shaped, irregularly and sometimes coarsely toothed on the curved side but not lobed, though usually with an auricle folded over the nutlet on the straight side side||11|
|7a||Petiole to 3 cm, leaf apex a fine rather jaggedly toothed tail, fruiting catkin long (to 20 cm) and lax||8|
|7b||Petiole to 2 cm, often very short; leaf seldom so markedly fine-tipped, fruiting catkin more compact, to 12 cm long||9|
|8a||Leaves flush greenish||Carpinus laxiflora|
|8b||Leaves often flush red or purplish||Carpinus viminea|
|9a||Fruit-bract small, 10–20(–25) mm long; leaf often long and slender, with a short (4–7 mm) thick petiole||Carpinus londoniana|
|9b||Fruit-bract larger, 20–30 mm long; leaf ovate to oblong, with a slender petiole usually at least 8 mm long||10|
|10a||Leaf slightly oblong, with 10-14 vein-pairs||Carpinus × schushaensis|
|10b||Leaf ovate, with (6–)12–18 vein-pairs||Carpinus caroliniana, C. tropicalis|
|11a||Fruit-bract large (20–35 mm long); leaf rather large (to 12 cm long), sometimes slightly oblong||Carpinus tschonoskii|
|11b||Fruit-bract usually less than 25 mm long; leaf often small and/or long-pointed||12|
|12a||Fruit-bract and rachis of fruiting catkin with dense hairs||13|
|12b||Fruit-bract and rachis of fruiting catkin only sparsely hairy||14|
|13a||Leaf with a rather distant serration; dense yellow or rusty hairs under veins||Carpinus monbeigiana|
|13b||Leaf with a close double serration; only rather sparse, whitish hairs under veins||Carpinus kawakamii|
|14a||Teeth of leaf usually with conspicuously mucronate tips||15|
|14b||Teeth of leaf seldom with conspicuously mucronate tips||16|
|15a||Leaf-toothing more or less simple||Carpinus faginea|
|15b||Leaf-toothing usually double||Carpinus polyneura (and allies)|
|16a||Leaf often short (c. 5 cm long) and never very long-pointed||Carpinus orientalis|
|16b||Leaves sometimes slender and long-pointed leaves||17|
|17a||Fruit-bracts very short (7–10 mm)||Carpinus hebestroma|
|17b||Fruit-bracts usually more than 12 mm long||Carpinus turczaninovii (and allies)|