See page 152, as A. delavayi var. forrestii. Recent investigation of all the available material leads to the conclusion that this fir should rank as a species, with a geographical area well demarcated from that of A. delavayi, extending from the Lichiang range of Yunnan, where it was first discovered, north and west to the area of the Tsangpo bend. It is variable, and contains one or perhaps two subspecies (Rushforth in litt.).
From New Trees
Abies forrestii Coltm.-Rog.
This species was described by Bean (B152, S24) and Krüssmann (K35) under the name Abies delavayi var. forrestii. Abies forrestii var. georgei (Orr) Farjon was also described by them (B153, S24, K35), under the name A. delavayi var. georgei (Orr) Melville. Two additional varieties have since been recognised, and a key to these is provided below. Var. smithii Viguié & Gaussen is not known in cultivation.
Cones large, 7–10(–14) cm long; young shoots glabrous and shiny; China (southwest Sichuan, northwest Yunnan, southeast Xizang)
Cones (usually) < 8 cm long; young shoots densely rusty brown-pubescent
Leaf resin canals medial; China (northwest Yunnan, southeast Xizang)
Leaf resin canals marginal
Bract scales conspicuously longer than seed scales, apex acuminate with lanceolate cusp; China (southwest Sichuan, northwest Yunnan, southeast Xizang)
Bract scales equal to, or only slightly longer than seed scales, apex rounded with short, central cusp; China (northwest Yunnan)
var. ferreana (Bordères & Gaussen) Farjon & Silba
A. chayuensis W.C. Cheng & L.K. Fu
A. rolii Bordères & Gaussen
A. yuana Bordères & Gaussen
The minor characters separating these so-called varieties should perhaps not be taken too seriously, despite the plethora of synonyms they have generated (see Farjon 2001). According to Rushforth (1987a), seed of var. ferreana was collected by Forrest and Kingdon Ward, but authenticated specimens are scarce. A recent collection was that made by the Kunming–Gothenburg Expedition (KGB 503) in 1993, from which a specimen is growing at Edinburgh. Material under this name has recently been offered by a nursery in Northern Ireland.
var. georgei (Orr)
A. georgei Orr
A. delavayi var. georgei (Orr) Melville
There has been some confusion over this fir. The leading characters given by Orr in describing A. georgei
were: branchlets densely clad with short, rusty hairs; bract-scales of cones with long, exserted cusps and of a distinctive shape, being nearly parallel-sided in the basal part and tapering into a triangular cusp, against spathulate in A. forrestii.
But prominently exserted bracts are a feature of some collections of A. forrestii
, and the branchlets are densely rusty-pubescent in A. forrestii
(see below). Keith Rushforth tells us he has seen only three herbarium sheets which agree with the type of A. georgei
(F.22547) and as yet no cultivated tree that agrees with it has been found (a tree at Wakehurst Place, Sussex, has cones with the bract-scales of A. georgei
, but the shoots are only weakly haired). The trees grown as A. georgei
are probably all from seed gathered by Forrest’s native collectors in 1931, shortly before his death, and those seen by Rushforth are either A. forrestii
or its var. smithii.
Incidentally, the Forrest number 30853, mentioned on page 153, belongs to a specimen with unripe cones; it was cited by Orr when he described his species, but according to Rushforth does not agree with the type specimen.
var. smithii Viguié & Gaussen
A. delavayi var. smithii (Viguié & Gaussen) Liu
A. georgei var. smithii (Viguié & Gaussen) Cheng & Fu
This variety was described in 1929 from a specimen collected by Dr Rock in the Lichiang range of north-western Yunnan, whence came also the type of A. forrestii
. It resembles the typical state of that species in cone characters, differing chiefly in its densely hairy branchlets, but also in its more resinous buds and shorter, bloomed leaves (Rushforth in litt
.). Two of Rock’s seed-introductions came as A. forrestii
, but the seedlings raised were variable in the pubescence of the shoots, some having them quite glabrous (E. L. Hillier, Conifer Conference Report
(1932), p. 228). Some trees grown as A. delavayi
belong to this variety, which was actually introduced by Forrest under number 10255 some years before the var. smithii
was described.In his monograph, Liu considers the difference between A. forrestii
and the var. smithii
to be of no significance and merges them as A. delavayi
, which is the name under which A. forrestii
is treated in his monograph, and was the correct one to use at the time that work was published, since A. forrestii
dates from 1929, and A. forrestii
was not reduced to the level of a variety (of A. delavayi
) until 1932. But according to the most recent edition of the International Code of Botanical Nomenclature
, the correct name for A. forrestii
considered as a variety of A. delavayi
is var. forrestii
, since in distinguishing A. forrestii
its authors automatically constituted the ‘autonym’ A. forrestii
, and such autonyms have priority over the name by which they were established (ICBN
(1983), Article 57.3).The following are either A. forrestii
or its var. smithii
: Wakehurst Place, Sussex, 68 × 4 ft (1974); Borde Hill, Sussex, in The Tolls, 66 × 33
ft (1981); Stourhead, Wilts., pl.
1924, 72 × 6 ft (1982); Hergest Croft, Heref., pl.
1924, 46 × 51
ft (1978); Werrington Park, Cornwall, in the Garden, 52 × 33
ft (1977); Lamellen, Cornwall, 50 × 61
ft (1977); Dyffryn Gardens, near Cardiff, 60 × 51
ft (1984); Edinburgh Botanic Garden, probably from F.6744 (not
W.4078), 62 × 51
ft (1981); Cortachy Castle, Angus, 84 × 91
ft and three others of a good size (1981); Tigh-na-Bruach, Invermoriston, Invern., 62 × 7 ft (1978); Headfort, Co. Meath, Eire, 60 × 73
ft and 70 × 71
ft, and another in the American Garden, 70 × 6 ft (1980); Birr Castle, Co. Offaly, Eire, 59 × 51