THE GENUS CODONOPSIS, WALL. 377 
Kew Bulletin, 1907, p. 9; and I cannot do better than refer to them. Some 
species, too, have been noticed to contain a milky juice, but as only a few 
species are known in the living state it is uncertain whether it exists 
throughout the genus. 
The majority of species of Codonopsis in cultivation appear quite hardy and 
do not require any special attention, although otherwise they scarcely recom- 
mend themselves to the gardener owing to their unattractive appearance. 
One species C. eonrolrulacea, which is distinctly the prettiest in cultivation, 
requires some protection and is perhaps more suitable for pot-culture, see 
Bot. Mag. tab. 8178. The remaining species in cultivation are: C. lanceo- 
lata, C. ovata, C. rotundifolia, C. ussuriensis, C. Tangshen, and C. viridiflora. 
As to the distribution of this genus:—O wing to the small amount of 
material at present to hand we can form only a general idea ; but this, how- 
ever, is very interesting and shows that the genus is found almost exclusively 
at considerable elevations, sometimes reaching an altitude of 13,000 feet. 
The general area marked out at present is a semicircular band, having its 
most north-easterly station yet recorded in the north of the Sikhora Alin 
Mts., on the watershed of the river Amur, about 141° E. and 51° N., thence 
running along these mountains as far south as Vladivostock. One specimen 
has been recorded from the mountains near Pekin. It next appears on the 
mountains of the northern island of Japan, and passes along the mountains in 
the centre of the south island. There is then a slight break in the line, as its 
next station is in the mountains of western Hupeh, and thence seems to be 
generally spread over the Yung-ling Mts. as far as the province of Kansu to 
the north and down through Yunnan in the south. It can then be traced 
through tthe mountains of the Shan States in Northern Burma, where it 
reaches its southernmost point at a latitude of 21° N., to the Khasia Hills of 
Assam. It is next recorded from the Phari and Chumbi district, and also 
further north from the Tsangypo watershed and Lhassa. It has been collected 
all along the southern slopes of the Himalayas through Kashmir to the 
watershed of the Kurram River in Afghanistan, and its most western repre- 
sentative yet recorded is from the Alexandrovski and Alatau Ranges in 
Russian Turkestan at about 79° E. and 43? N. 
One cannot help thinking from the striking variation in the small amount 
of material at present available, and the large, though not as yet always 
connected, area of distribution, that this will probably prove a large and 
remarkable genus. 
My thanks are due to Mr. L. A. Boodle for having examined the morphology 
of the floral whorls, and to Mr. Е. W. Rolfe for the accompanying figures. 
