A deciduous tree not uncommonly attaining a height of 50 to 70 ft in the wild and a girth of 7 to 9 ft (one exceptional specimen seen by Wilson was 80 ft with a trunk 10 ft in girth at 6 ft above the ground); branchlets glabrous, greenish yellow at first, becoming grey or greyish brown. Leaves obovate or less commonly oblong-obovate, 4 to 7 in. long, 21⁄4 to 4 in. wide, notched, rounded or acuminate at the apex, obliquely wedge-shaped at the base, dark glossy green and glabrous above, undersides paler and usually densely covered with long greyish hairs except on the more or less glabrous midrib; leaf-stalks slender, glabrous. Flower-buds ovoid, acute, covered with greyish hairs; flower-stalks downy. Flowers about 8 in. across, borne before the leaves in March-April. Tepals ten to thirteen, oblong-obovate or oblong-lanceolate, purplish pink on the outside, paler or white within; in a specimen collected in the wild on Mt Omei in 1939 the tepals were 23⁄4 to 43⁄16 in. long, 13⁄16 to 15⁄8 in. wide, and are of about the same dimensions in cultivated plants. Fruit-cones cylindrical, 4 to 5 in. long, often twisted owing to the uneven development of the carpels. Seeds usually one in each carpel, orange-scarlet.
A native of W. Szechwan (Wa-shan and Mt Omei) and of N. Yunnan; introduced by Wilson in 1908 during his expedition for the Arnold Arboretum and described from fruiting specimens collected at the same time. Wilson never saw this species in bloom, but a flowering specimen was collected by C. L. Sun in 1939 on Mt Omei, about thirty miles east of the type-locality (Pl. Omeienses Vol. 1, p. 1944 and Plate 8).
M. sargentiana first flowered at Nymans, Sussex, in April 1932 and received an F.C.C. when shown from there in 1935. When sending material from this tree to Kew in 1943, James Comber, the garden manager, wrote that it was ‘a seedling raised by J. Nix esq. from Wilson’s seed, and given to me when visiting Tilgate’. This tree, which still grows in the Walled Garden, must be one of the few in this country raised direct from the wild seed. Most of the older trees, both of the type and the var. robusta, came from Chenault of Orleans, who received seedlings from the Arnold Arboretum which he multiplied by grafting.
In cultivation M. sargentiana makes a tree which has already (at Caerhays) attained the average height of wild trees, but not yet their girth. Its flowers resembled those of M. dawsoniana, and are no less beautiful, but that species can always be distinguished by its shrubby habit and deeper green, veiny leaves. M. sargentiana in its typical state is not so widely planted as the var. robusta described below, partly because it does not seed freely and partly perhaps because it takes longer than the var. robusta to reach the flowering state. A tree at Caerhays, Cornwall,pl. 1921, measures 50 × 33⁄4 ft, with a bole of 11 ft; another, six years younger, is 38 × 3 ft (1966).
var. robusta Rehd. & Wils. – This variety was described from fruiting specimens collected by Wilson in the area from which came the type-specimens of M. sargentiana; seeds were also taken and distributed under field number W.923a. According to the original description it differs in its longer and narrower leaves, which in Wilson’s specimens are 51⁄2 to 81⁄2 in. long, 21⁄2 to 31⁄2 in. wide, with larger fruits (43⁄4 to 71⁄4 in. long). The grafted plants sent out by Chenault as var. robusta agree very well with this description in their elongated leaves and in bearing fruit-cones 7 to 8 in. long. They are also distinct in other respects. They make spreading bushy trees, usually dividing into three or more stems at 3 or 4 ft above the ground; in mature specimens the leaves are mostly broadly notched at the apex; and the flowers are strikingly different from those of typical M. sargentiana, being up to 12 in. across, with more numerous and much larger tepals (ten to sixteen, 8 in. long, 3 in. wide). If more herbarium material were available – it is very scanty and incomplete – it might prove that the var. robusta is linked by intermediates to the typical state of the species. But cultivated plants raised from the original wild seeds are certainly very distinct. Of the two, the var. robusta is much to be preferred, the flowers being of a clearer pink and better displayed.
M. sargentiana var. robusta first flowered in Britain at Caerhays, Cornwall, in April 1931. A flower sent by J. C. Williams to Kew on April 14 of that year is still preserved in the Herbarium and has with it a description drawn up by Dr Stapf while the flower was still fresh. Photographs of the Caerhays plant, taken on the same occasion, are reproduced in New Flora and Sylva, Vol. 5, figs, iv and v. In France, the var. robusta had flowered earlier, by 1923.
M. (sargentiana var. robusta × sprengeri var. diva) ‘Caerhays Belle’. – A hybrid of great promise, raised at Caerhays, Cornwall. See Journ. R.H.S., Vol. 91, p. 285 and fig. 151.