BALFOUR—NEW SPECIES OF RHODODENDRON. 77 
guineum. The construction in Rh. haematodes is this :—There 
are a number of hairs with long cord-like stems of many cells 
from which thin cylindric thick-walled ascending and divari- 
cate branches proceed at intervals to the top, where are many, 
and these, particularly the topmost ones, become rufous-brown. 
These long-stemmed branching hairs interweave and form the 
loose upper stratum of indumentum. Attached to the epidermis 
between these are many short-stalked hairs with many branches 
radiating from the top of the stalk—the branches uncoloured, 
of broad thin-walled vesicular cells—and they form a lower 
stratum of indumentum. They are invisible under the upper 
stratum of the branched stems of the longer hairs. We thus 
have two quite distinct strata of the indumentum. There is 
no such development of an upper stratum in Rh. sanguineum, 
and this is what Franchet means when he says of RA. sanguineum 
—‘‘ the leaves without tomentum beneath even when young.”’ 
The tomentum, the woolly surface, in Rh. haematodes is formed 
by the upper stratum. The whole indumentum of Rh, san- 
guineum may be taken as the equivalent of the under stratum 
in Rh. haematodes. In Rh. citriniflorum, as has been stated, 
there is a beginning of the formation of an upper stratum. In 
no specimens of Rh. haematodes which have come under my 
observation does the upper stratum of indumentum fall off in 
the older leaves. The end branches of its hairs sometimes 
become matted together in part, and the surface network of 
hairs is obscured: that is all the change I have seen. There 
is not in Rh. haematodes, as happens in species with bistrate 
indumentum such as members of the Hodgsoni series, a 
removal of the usually coloured upper stratum as the leaf 
oldens, exposing the close usually grey under stratum. If 
there were we should get a leaf-surface like that of Rh. 
sanguineum. Rh. sanguineum shows what for purposes of 
contrast—though I express no opinion upon order of evolu- 
tion—we may call a primary condition of one stratum of 
white compact hairs; Rh. haematodes shows a more developed 
condition of an upper stratum of looser tomentose hairs above 
a lower stratum, and both persist ; Rh. Hodgsoni, Hook. f. 
shows an evolution of strata like that of Rh. haematodes, but 
the older leaves lose the tomentose upper stratum and expose 
the compact white under stratum and thus revert to the 
primary condition. This is the difference In construction 
of indumentum to which Franchet directs attention in the 
last paragraph of his note. 
It should be borne in mind that Rh. sanguineum and 
Rh. haematodes are species which do not touch geographically. 
The latter is limited to the Tali Range and its vicinity about 
