— 263 — 



mountains had been included, a number of Chenopodiaceae 

 (e. g. Nanopluihim, Alexandra, Ofaiston), Calligoimm spp. and 

 Ammodendron would be I'egarded as endemic. Bui as already 

 indicated Ihe natural conditions and the vegetation are rather 

 difTerent in these more northern areas. 



The distribution of the endemic species amongst the 

 various growth-forms will be seen from the following table 6: 



Table 6. 



The endemic species are thus distributed over the growth- 

 forms in about the same proportion as the aggregate floi-a 

 of the desert. The following variations may be of some 

 interest: 1) There are no endemic aquatics and marsh-plants 

 whatever; 2) The endemic species include comparatively 

 greater numbers of fanerophytes and hemicryptophytes than 

 the aggregate species; 3) The endemic species include com- 

 paratively fewer annuals than the aggregate. 



The first variation is not surprising since it is well 

 known that aquatics and marsh-plants have a wide distri- 

 bution (see e. g. Drude p. 317). 



The last point, that there are comparatively few endemic 

 therophytes, may be explained thus: the therophytes are 

 as a rule spring-plants, and many of these have a wide 

 distribution because over considerable tracts of regions with 

 winter-rains they find almost the same conditions during the 

 short vegetative period. In accordance with this at least 28 

 p. cent, of the endemic therophytes are late-flowering, in 

 comparison with 21 p. cent, of the total therophytes of the 

 whole flora (comp. table 4, p. 162). The second point of 

 variation suggests that the natural conditions of Transcaspia 

 have been favourable to the development of fanerophytes and 

 hemicryptophytes. This does not seem unnatural since the 

 fanerophytes are one of the most prominent and apparently 



