ot 
attracts attention by its glossy snow-white stems. This plant 
I found in a deep valley amongst the dunes on a shifting 
sand-desert almost devoid of vegetation; there were only a 
few specimens and only in this particular valley. This illus- 
trates the very scattered occurrence of the various species; 
one must examine a great number of localities before a 
thorough knowledge of the flora of the desert is acquired. 
Each locality presents only a few species of trees, shrubs or 
herbs. Other examples of this will be referred to in chap. 11. 
It will be seen from the above description that the flora of 
the shifting Sand-desert is extremely poor. Looking over the 
desert from the top of a barchan, the eye is attracted by the 
scattered rough tussocks of Aristida and the switch or brush- 
like dwarf trees and shrubs, standing widely apart, particu- 
larly Ammodendron and Calligonum the most frequent and 
most enduring. 
Only on closer examination does one find the herbaceous 
plants which hide in the valleys. These plants, already re- 
ferred to, are ill-adapted to withstand sand-covering and are 
often smothered. But they endure desiccation, heat and the 
tear and wear of the drifting sand. This tear and wear is perhaps 
the reason why the plants of the more stable sand-desert are 
very rarely seen in the true shifting desert. 
From the shifting sand-desert where the barchans hold 
their sway, we turn to a more stable type of sand-desert, 
the Hummock-desert (see p. 78), which RappE compared to 
a dead sea with a swell on, and which he regards as a tran- 
sitional form between the barchans and the flat deserts. 
The sand hills are round and generally rather low, RADDE 
states from a few feet up to 4—5 fathoms, i. e. about 10 me- 
tres at the most, which is, however, a considerable height. 
Where the hills are grouped closely together, the valleys be- 
tween them are basin-shaped. Smaller crescent-shaped bar- 
chans may be present so that the sand-drift is not every- 
where perfectly stabilised. The movement must however be 
slight in many parts, as indicated by RADDE finding on the hill- 
tops a lichen (Urceolaria indurata Wain.), a thin black and 
white coating over the surface of the sand. 
The vegetation consists partly of the same species as on 
