103 



whereas on the more stabilised sands these give place to 

 perennials and undershrubs. This is only natural, for if the 

 perennials have not, like Aristida, unusual powers of resisting 

 sand-drift, they will sooner or later be smothered by a shift- 

 ing barchan and very few, if any, will produce seeds during 

 the first year; on the other hand, the annual plants with a 

 shorter growth-period will have a greater chance of surviving 

 and ripening seeds. 



The difference in the numbers of annuals in Hummock- 

 deserts and Desert-plains appears to be due to the annual 

 halophytic Chenopodiaceae which occur in the depressions 

 of the Hummock-desert, but seem to be of less importance 

 in the Desert-plains. 



Finally, attention is directed to the trees of the desert 

 (the switch trees and shrubs) which as emphasised in the. 

 preceding pages, play the most promiment part and attain 

 the richest development in the shifting desert, while they 

 deteriorate where the sand is stable. Sand-drift seems to be 

 a condition essential for vigorous growth in their case and 

 also with Aristida pennata. 



Making a mental comparison between the sub-formations 

 of the sand-desert described, we see that they have many 

 features in common both floral and biological, but that 

 the differences between them are not altogether to be neglect- 

 ed. The most important common feature which unites them 

 and which causes them to be regarded as sub-formations and 

 not as formations is first the general occurrence of the desert- 

 trees, though under a somewhat different form; secondly that 

 the soil is sand, which is saline only in the depressions so 

 that the true halophytes play a comparatively minor part 

 except in these places. 



The different forms of sand desert are evidently histori- 

 cally related in that the one must have originated from the 

 other. What has been the course of development? It has 

 been already pointed out that RADDE (1899 p. 16) following 

 OBRUTSCHEW regards the Desert-plains" the covered Sand-steppe" 

 as the last stage in the metamorphosis of the sand; that the 

 sand-hills, while being covered by vegetation, are gradually 

 being levelled down through the agency of water, wind and 



