118 BALFOUR—NEW SPECIES OF RHODODENDRON. 
Rh. Forrestii was discovered by Forrest in I905 growing 
in the Tsedjiong Pass on the Mekong-Salween divide in lat. 
28° 10’ N. at 10,000-11,000 ft. elevation but owing to the de- 
struction by the Lamas of the greater part of his collections 
of that year only one fragmentary specimen of the plant reached 
this country under No. 699—sufficient in the case of so striking 
a novelty to sanction description as a new species. Two 
prominent features in the plant at once arrest attention—the 
small elliptic or slightly obovate leaves with dark almost black- 
purple under surface and the deep crimson colour of the 
solitary terminal flowers. In course of the explorations which 
Forrest has made during the dozen years that have elapsed 
since the discovery of this plant he had not come across its 
like again until this year 1918 when he sends home to Mr. 
J. ©. Williams in a postal packet specimens of it under No. 
16,689 from a station at a higher altitude a little further 
south than that of the original find. This sending gives 
material for a better knowledge of the species—although un- 
fortunately the important fruit is still wanting—and enables 
me to add to the published description of the species hitherto 
incomplete. 
In the years between 1905 and rgr8 Forrest has found the 
plants cited here under Rh. repens. They differ from Rh. 
Forvestii in the arresting characters mentioned. They have 
larger coarser leaves prominently obovate and pale green not 
deep black-purple on the under side and the flowers are brighter 
crimson. Specimens No. 13,259 and 13,442 are in fruit which 
IN 13,259 is twice the length of that in 13,442, but there is no 
other apparent character of difference and the plants conform 
generally with the other green-leaved forms. From seeds of 
No. 13,259 plants have been raised by Mr. J. C. Williams at 
Caerhays Castle and these plants in specimens presented to 
us by him have leaves with green under surface like those 
of the parent. On one plant I found a leaf evidently an 
early formed one which has a bright-red under surface. This 
suggests that the seedling leaves in this plant possess this 
character which is not shown by the leaves produced by the 
adult. The development of such a type of leaf in the seedling 
stage 1S a recognised habit in many Rhododendrons and in 
back-breaks ” and water-shoots the same seedling type of 
leaf commonly appears. In Forrest’s Tsarong plant from 
Ka-gwr-pw No. 14,534 I find such a water-shoot. We have 
then evidence upon which to found satisfactorily the statement 
that the plant we are calling Rh. repens is one of those Rhododen- 
drons which begin life with leaves deeply red or purple-tinted 
below, and that the coloration is not maintained during its 
