68 BALFOUR—NEW SPECIES OF RHODODENDRON. 
may fall off, as do the hair-bases on the intermediate areas, and 
the leaf becomes quite glabrous. 
A character of attunement having a development such as 
has been described may be expected to show arrests or ex- 
acerbations in relation to environmental conditions, and the 
long series of specimens which Forrest has collected exhibit 
the indumentum-character in various stages. In some of the 
specimens the young leaves with yellow indumentum do not 
appear, possibly because of the stage of development of the plant 
from which they were plucked; in others the indumentum 
becomes bright brown, not dark red, or it is only buff-coloured. 
In what I have described as the typical form the agglutinate 
patches of indumentum after it splits appear as so many “‘ scabs ” 
on the surface, and give it a by no means attractive look, but 
in other cases the splitting proceeds in undulate lines along the 
long axis of the leaf, giving the surface an areolate aspect with- 
out reaching the stage of isolation of the agglutinate patches. 
In some specimens the indumentum, so long as it is in the 
yellow stage, is continuous and unsplit, in others it is split at 
this stage. The difference between the appearance of the 
leaves in typical cases where the scab-like patches are well 
developed and cases in which there are only a few undulate 
lines of splitting is very conspicuous, and when these forms of 
splitting are associated with long and short leaves respectively, 
or with broad and narrow leaves, the question of specific identity 
of the forms naturally arises, particularly when, as in so many 
of the specimens that have been collected, only foliage or foliage 
with fruit is shown. I believe that they all belong to one 
species, and that the variation in leaf-size and indumentum- 
splitting is to be attributed to climatic conditions of the habitat. 
Forrest’s opinion upon this we should much like to have. 
He has called attention to the foliage-character of the plant 
as giving it value horticulturally. 
There is no doubt about the affinity of this remarkable 
Rhododendron. It belongs to the Taliense series, which in- 
cludes eglandular forms with elliptic or oval or oblong-oval thick 
fleshy leaves showing rounded or cordulate base and a dense 
tomentose bistrate under-leaf indumentum often agglutinate in 
its upper stratum, usually persistent—but here splitting and often 
shed in agglomerate scale-like patches—small calyx, campanu- 
late or funnel-campanulate 5-lobed (sometimes more) corolla, 
ro (sometimes more) stamens with hairy filaments, a glabrous 
ovary and glabrous style. Rh. taliense, Franch., was the first 
described species of the series, and it seems to be unknown 
outside the Tali Range in Mid West-Yunnan. In 1913 Forrest 
brought from the Chungtien plateau, in the east of N.W. 
