Deciduous tree 8–20(–25) m tall, to 1 m dbh. Branchlets yellow-green at first, becoming grey. Leaf blade obovate (rarely oblong-obovate) 10–19 × 6–10 cm; lower surface pale green with dense, wavy, silvery grey hairs (in young trees, leaves on trunk have only trichomes along midvein); upper surface dark green, glossy, glabrous; secondary veins 8–12 on each side of midvein; base cuneate; apex rounded with a small notch, or mucronate. Petiole 2–4.5 cm; stipular scar ⅙–¼ as long as petiole. Flower buds ovoid, ~3.5 cm, with pale yellow hairs. Flowers appearing before leaves, 15–33(–36) cm across, erect or pendulous, slightly fragrant. Tepals 10–14(–17), in 3 whorls, pale red to pale purplish red, obovate-spoon-shaped or narrowly obovate, 8–10 × 3–4.3 cm, fleshy, apex rounded to notched. Stamens 1–1.9 cm; filaments purple; connective exserted and forming a 0.5–1 mm mucro; anthers 7–9 mm, dehiscing laterally, base broad. Gynoecium green, terete, 1.8–2 cm, glabrous; stigmas purple. Fruiting peduncle thick and strong, 7–10 mm in diameter, with residual hairs on nodes. Fruit terete, 8–15(–17) × 2–3 cm, usually wrinkled; mature carpels blackish purple, hemispherical or subglobose, 1.2–1.4 × ~0.9 cm, densely tuberculate, apex shortly beaked. Seeds nearly reniform, irregularly orbicular, or obovate, 10–12 × 6–8 mm, bilaterally flat; testa reddish brown. Flowering April-May (China, UK), fruiting September (China). Hexaploid 2n=114. (Xia, Liu & Nooteboom 2008).
Distribution China Sichuan, Yunnan
Habitat Wet broad-leaved forests; 1400–3000 m.
USDA Hardiness Zone 7-9
RHS Hardiness Rating H4
Conservation status Vulnerable (VU)
This is one of the classic ‘tree magnolia’ species whose large pink flowers on naked twigs in spring seem to play chicken with frosts. Inviting comparisons with parachutes (Gardiner 2000) or less favourably with ‘pink Kleenex tissues […] toilet paper’ (Cook 1999), the flowers of a big specimen in a good year make a fine show, especially from a distance, where blemishes are less obvious. The flowers tend to hang downwards, making the paler pink inner surfaces of the tepals more visible than in many species. Not widely hardy in our area, it suits milder maritime climates such as the British Isles and the Pacific Northwest.
It is most similar to M. dawsoniana, but differs in the much hairier leaf undersurface, in having on average more tepals (10–14 as opposed to 9–12) and in tending to make a standard tree rather than a bushy multistemmed specimen. In both these species flowers tend to be held horizontally, later hanging, in contrast to the more upright flowers of M. campbellii and M. sprengeri., although there is considerable variation among flowers (Treseder 1978).
The species was described scientifically from collections made in Sichuan by Ernest Wilson for the Arnold Arboretum (W 914 & 932 of 1908, Rehder & Wilson in Sargent 1913), although it seems that it had previously been collected by Père David (Sichuan 1869). The specific epithet honours the frosty-natured but highly competent Charles Sprague Sargent, first Director of the Arnold. Wilson saw it only in fruit, and its flowers were described scientifically only long after the name had been published. However, it is clear that Wilson’s local colleagues were well aware of this tree, its impressive flowers and the use of its bark in medicine (Howard 1980). Seed was also collected, but along with Wilson’s other magnolias, which were proving tender in Massachusetts, the Arnold’s material was sent to French nurseryman Leon Chenault for grafting, distribution in Europe and eventual return to North America.
M. sargentiana grows very large. Wilson reported seeing a tree 24 m tall with a girth at 1.8 m of 3 m and a bole unbranched to 5 m. Its branches were very numerous and wide spreading, forming a massive flattened ovoid crown. In southern England large, rather spindly and twiggy trees of 15 m or more can be seen. Plants labelled var. robusta (see below) tend to branch from near the ground, making a more bushy tree. The glossy green leaves are in practice extremely variable in shape but the covering of downy hairs beneath is distinctive.
It is impossible to discuss M. sargentiana in Western cultivation without at the same time wrestling with the controversial var. robusta, not least because the overwhelming majority of garden plants are placed there. It was described from Wa Shan, Sichuan at the same time as the species, from W 932a, collected at a slightly higher altitude in the same area as W 932 (Rehder & Wilson in Sargent 1913). As described, it differed in having larger fruits and longer, narrower leaves than the type variety as originally described (the upper end of the range given above). Garden plants raised from the seed collection under this number proved very different to Wilson’s collections of var. sargentiana, and those whose main focus is on cultivation have usually emphasized its distinctness, sometimes to the point of believing that it might be a distinct species (Johnstone 1955; Treseder 1978), or perhaps a natural hybrid (Callaghan & Png 2010). Botanists involved mainly with wild plants have tended to downplay its significance.
Based on detailed comparsion of plants from the two Wilson stocks in Britain, Treseder (1978) laid out the distinctive features of var. robusta as follows (versus var. sargentiana in brackets): flowers very large, to 30 cm across, broad tepals not tending to recurve (versus to only 20 cm across, narrower drooping tepals); flower buds prolifically produced right down to ground level, usually curved (versus sparse and high in the crown, more symmetrically shaped); leaves oblong-obovate, in general longer and narrower than those of var. sargentiana; fruit large, stout, oblong, to 20 cm long (versus slender, to 13 cm); branchlets lighter in colour. Given all this, it is no surprise that the Wilson var. robusta and its progeny dominate garden stocks. As an extreme example of the paucity of var. sargentiana, a grafted specimen at Caerhays from the original distribution is only half the height of the largest var. robusta specimens there, and has flowered just once (2009) in Charles Williams and his father’s lifetimes (C. Williams pers. comm. 2022). Additionally it is sometimes claimed that var. robusta flowers more quickly from seed, although first flowering of the two varieties in Britain came at about the same time (var. sargentiana in 1932 at Nymans, var. robusta in 1931 at Caerhays – Savage 1980). Perceived differences in tree form and size vary between accounts.
In sharp contrast, botanists revising the genus as a whole have downplayed the significance of var. robusta. Dandy (1928) felt that it did not warrant varietal status, reiterating this view later in life (in litt. to Neil Treseder – Treseder 1978). Chen & Nooteboom (1993) did not uphold var. robusta, finding nothing other than the size differences in flower, fruit and leaf to distinguish them. Flora of China (Xia, Liu & Nooteboom 2008) merely lists it among synonyms, and the name is not current in research literature. Interestingly, Wilson himself came to believe that this was not a clear cut distinction worthy of varietal status (manuscript published posthumously in Howard 1980).
Our own feeling is that this is less of a controversy than it might seem. The garden stocks are different, and we have a validly published name by which to distinguish them. If there is insufficient evidence to make the case for var. robusta as a distinct, extant taxon in the wild, so be it. It will be interesting to see how our garden perception of the species changes if and when further wild introductions are made. As a postscript, while Callaghan & Png (2010) hypothesised a hybrid origin, they were unable to refind the tree or other potential parents of a hybrid on Wa Shan in 2009.
The finest older specimens in Europe are mostly found in southern England and Wales; most are attributed to var. robusta, for example an enormous but elegant specimen at Burncoose, Cornwall (23 m × 305 cm, 2014). However, the example mentioned above of var. sargentiana at Nymans, W Sussex, may be the one surviving today (15 m × 202 cm, 2016 – The Tree Register 2021). Elsewhere in Europe it suits maritime parts of France and the Low Countries. In North America this is a tree for the Pacific Northwest, with many good examples in Seattle and Vancouver.
This species – mainly the garden stock labelled var. robusta – has been an important parent of garden hybrids.
Flowers larger and a darker, redder pink than any other M. sargentiana clone. Propagated from a tree at the Strybing Arboretum, San Francisco, recorded as originating with Viscount Cranborne; distributed by Edwards Nursery, Palo Alto, CA.
Synonyms / alternative names
Magnolia sargentiana 'Chyverton Dark Form' nom. inval.
Deep purple buds; selected by Nigel Holman, Chyverton, Cornwall, before 1989.
Very large, white to pale pink flowers; Holman, before 1989.
Synonyms / alternative names
Magnolia sargentiana 'Vico Multipetal'
19–27 pale pink tepals per flower, the outer six reflexing. One of 50 seedlings of var. robusta raised by Ambrose Congreve at Mount Congreve, Waterford, Ireland; named by Sir Peter Smithers in 1983. A distinctive plant starred by Philippe de Spoelberch, Belgium (pers. comm. 2021).
As discussed in detail above, garden stock derived from W 923a, with flowers, fruit and leaf length at the upper end of the size range, has long been labelled var. robusta. These are the dominant forms in western cultivation.