Tree to 40 m × c. 1 m dbh. Crown conical, round-topped in maturity. Bark of young trees smooth at first, orange- or reddish-brown, becoming exaggeratedly flaky after c. 4–6 years, exfoliating in large papery flakes even along branches in a manner similar to many Betula species; bark of old trees remains flaky in the upper parts of the trunk and on branches, but ceases to be flaky in the bole of old trees and becomes grey, developing into irregular narrow plates between fissures and ribs. First order branches long, spreading; second order branches spreading or weakly assurgent. Branchlets horizontal or assurgent, stout, first year growth shining red or maroon, rarely purple, increasingly matt and smooth with age, ridged and grooved especially when young, glabrous, or puberulent at first. Vegetative buds ovoid-globular, initially covered in a dense white resin; bud scales keeled. Leaves rigid, thick, densely set on the shoots, assurgent above the shoots (pectinate in shade), (1–)1.5–2.5(–2.8) cm × 1.5–2 mm, twisted at base, margins flat, apex obtuse or acute (rarely emarginate), pruinose- or glaucous-green with a central groove above, some (~7) broken lines of stomata on upper surface, with two greenish- or glaucous-white stomatal bands beneath. Pollen cones pendant, 2–3 cm long, yellowish with purple microsporophylls. Seed cones short-pedunculate, ovoid-cylindrical, apex obtuse, 5–8 × 2.5–4 cm, violet-purple when young, pruinoise purplish-brown at maturity; seed scales cuneate-flabellate or reniform-flabellate, 1.3 × 1.5 cm at midcone; bracts slightly exserted at maturity; rachis dark purple-brown. (Farjon 2017; Debreczy & Rácz 2011; Fu, Li & Mill 1999).
Distribution China S Gansu, S Qinghai, NW Sichuan, SE Xizang (Tibet)
Habitat Sub-alpine coniferous forests at 3500–4700 m asl, in regions characterised by long, cold winters and short, cool summers, with much precipitation falling as snow. It occurs in pure stands and with A. fargesii, A. recurvata, Larix potaninii, Picea asperata, P. crassifolia, P. linzhiensis, and several Juniperus species. High altitude forms of Betula species are the most common broadleaf associates.
USDA Hardiness Zone 5-7
RHS Hardiness Rating H7
Conservation status Vulnerable (VU)
Discovered by Wilson in 1904, and introduced by him in 1910, Abies squamata has always remained extremely rare in collections. Wilson’s introduction, W 4079 (not 1079 as widely cited) (Sargent 1917) was for many years the sole representative of the species in gardens; it has been supplemented by new material only in the very recent past. Other early collectors targeted it – there are Cheng and Harry Smith collections in several herbaria, some of which were gathered in the autumn months – but there is no evidence in literature that these were accompanied by seed consignments. In the UK and Ireland Bean (1976) could only report nine known trees; by the time the supplement was published (Clarke 1988) this list had reduced to four, and one of these was a previously overlooked tree at Durris near Aberdeen, meaning that six of Bean’s nine had died.
Both here and in North America A. squamata seems to have suffered high mortality rates, perhaps because it is somewhat at odds with our general understanding of the blue-coned firs of south west China. We tend to think of this informal group as prefering maritime climates, with abundant precipitation and without the extremes of temperature associated with continental regimes. A. squamata, however, is a tree of the subalpine zone. It is the highest-occuring fir, and reaching as high as 4700 m asl it is among the highest growing tree species on earth. It is distributed on the eastern margins of the Tibet-Qinghai Plateau and in the very high mountains of adjoining provinces: this is a very dry area, characterised by long, cold winters with abundant snowfall, and short, cool summers with relatively low rainfall. Debreczy & Rácz recall seeing A. squamata here growing ‘on river terraces with willows and Tibetan Sea-buckthorn’ and elsewhere in forests in the company of several Junipers and Spruces (Debreczy & Rácz 2011). As a consequence of its distribution, A. squamata is more tolerant of extremes of temperature, and of drier climates, than many of its close relatives; on the other hand it can be sensitive to moisture and humidity, and as a young plant is susceptible to significant damage by late frosts, exacerbated by very slow growth rendering it vulnerable for much longer (Rushforth 1987). Although exceptionally hardy, the first few years are perilous, as leaders and side shoots are regularly injured, often fatally, before young trees finally ascend above the danger zone.
It was scarce enough that within a few years of its introduction the Hillier Nurseries were only able to distribute grafted plants, using a Wilson original as a stock plant (Bean 1976), and it may be that no original trees survived in North America. Jacobson (1996) calls it ‘exceedingly rare’ here, and plants and scions were regularly reintroduced from Britain through the middle decades of the 20th century (USDA records; biodiversitylibrary.org).
Flaky Bark Fir can be regarded among the most ornamental of the Chinese firs, and perhaps of the whole genus, on account of its bark, which on growth aged four to six years or older becomes exaggeratedly flaky and remains so for many years, exfoliating in irregular, large, papery flakes in a manner reminiscent of some birches and the Paperbark Maple, Acer griseum. Today the Sir Harold Hillier Gardens contain probably the densest concentration of genuine trees in cultivation, a claim undermined only by the possibility that they represent a single genotype. The best examples are in Scotland, with superb, individual original trees at Dawyck in the Borders and at Durris in Aberdeenshire (Tree Register 2021). Both have been propagated and variously distributed in recent decades; it would be worth ensuring their progeny are labelled and catalogued such that they can be traced to their parents, as in the case of young trees at Ashmore House in Perthshire, traceable to the Durris tree (pers. obs. 2020).
As others have observed, ‘rare’ is a dangerous adjective. Employed lightheartedly it can cause something to become desired and revered far beyond its merits – in extreme cases it can awaken a hitherto dormant kleptomaniac. Scions from the original tree at Dawyck were grafted by George Kirkpatrick in the RBG Edinburgh nursery c. 1990, with two resulting plants being repatriated to Dawyck. These were in the nursery there when David Knott arrived as Curator in 1993. One vanished shortly afterwards in mysterious circumstances and has never been tracked down; the other was planted near the Chapel and is now a thriving young tree of c. 7 m (D. Knott, pers. comm. 2018; pers. obs. 2020).
Flaky Bark Fir is, undeniably, a rare plant. It has not been subject to the intense collecting that has driven some cycads, cacti, ferns, and orchids to extinction in their natural habitats, but the beauty, and scarcity, of the very few genuine examples in collections does lead to occasional outbreaks of ‘wishful thinking’ among dendrologists. It has been sought from the wild by various means since the Bamboo Curtain was raised in the 1980s: the Chinese Academy of Foresty sent out seed in the late 1980s, but this has been shown to represent A. recurvata; several UK nurseries, for example the Duchy of Cornwall and Dulford Nurseries in Devon, have raised firs with flaky bark and assumed these to be A. squamata, but these have been shown to be A. holophylla. Western collecting teams have sought A. squamata in Sichuan: SICH 1354 was introduced as A. squamata but is certainly not this species; it differs in several characters and appears to be a high-altitude form of A. fargesii. More recently the BCHM collecting team introduced multiple collections labelled A. squamata in 2011; these were widely distributed to institutions in western Europe and North America, but resulting plants are remarkably variable (pers. obs. c. 2015–2021). Most plants of BCHM 132 seem to better fit A. fargesii, while BCHM 127 includes some more like A. fargesii, and others that appear genuine, combining the approximate vegetative characters of A. squamata with a rather frost-prone and generally lack-lustre constitution which seemed to characterise Wilson’s A. squamata in youth on both sides of the Atlantic (Bailey 1923).
All these various red herrings prompt the obvious question: why is genuine A. squamata so elusive? In all likelihood there are several factors involved. In Plantae Wilsonianae Wilson suggests ‘This remarkable Silver Fir is the most alpine of the species occurring in the neighbourhood of Tschien-lu [modern Kangding] where it forms pure forests’ (Sargent 1917). It seems probable that subsequent collectors, following Wilson’s directions to this locality, have then overlooked the crucial question of altitude, for west of here A. squamata is not the only indigenous fir. Both A. recurvata and A. fargesii occur here too, with A. squamata overlapping with both toward their upper limits in some areas, and then forming pure stands only at very considerable altitude (Debreczy & Rácz 2011; Fu, Li & Mill 1999). To further complicate matters, A. fargesii shares the characters of maroon, red, or purple shoots, and variably flaking bark; indeed, some selections of this species are almost as good as A. squamata in this regard (pers. obs.). A. recurvata shares the character of obtuse or acute needle apices, but so do very high altitude forms of A. fargesii; one could be looking in the right place and still come unstuck.
This might explain a degree of confusion in ostensibly authoritative literature: the shoots are usually described as glabrous or only minutely hairy at first (e.g. Farjon 2017; Debreczy & Rácz 2011) which concords with original trees of W 4079 (pers. obs.). In his monograph, however, Liu describes old purple shoots as densely brown-hairy, but notes that young specimens in cultivation have glabrous red shoots (Liu 1971); this suggests that at least some of the specimens studied by Liu were wrongly identified, better fitting A. fargesii. In his original description Masters describes young branches as dark-pubescent, but does not qualify this by adding whether this intial pubescence is sparse or dense, retained or not, and the accompanying illustration shows no hairs at all (Masters 1906)! After examining several herbarium sheets Rushforth observed that the shoots are mostly glabrous, but that two collections (Cheng 1783 and W 3019) have hairy shoots (Rushforth 1986) which again calls their identity into question.
The incidence of common characters between this species and A. fargesii is worthy of re-examination, especially given the various recent introductions that strongly resemble the latter. Our idea of what A. squamata ought to look like is based on an extremely narrow sample set – Wilson’s original gathering – and even those trees thought to be distinct genotypes, for example the Dawyck and Durris trees, are remarkably alike. It may be that A. squamata is a more variable entity than horticulture would like it to be, and that Wilson, as he was known to, sought out the very best form that he could; a tree with more exaggeratedly flaking bark than most. We should not dismiss the possibility, either, that A. squamata might be nothing more than a very high-altitude form of A. fargesii – a variable species with a large and diverse range – that has adapted to very high altitudes in an area which also happens to be a very dry part of south west China, leading to rather striking differences with those relatively low forms of A. fargesii we are most familiar with.
Whether they merit distinction at species rank or as a cultivar, those very few clones of W 4079 that persist in cultivation remain highly sought after, and where they are grown well few firs are more beautiful.
A cultivar introduced from the Flora Wonder™ range of Buchholz & Buchholz c. 2009, this is likely a naming of material ultimately traceable to a Wilson original. It is safe to assume that the genetic base of A. squamata in collections is extremely narrow, and this is probably physically and genetically identical to many trees labelled simply Abies squamata. If ever the genuine article is reintroduced, being able to distinguish old trees – and grafts from them – will be useful.