370 EEPOET — 1882. 



the desert a certain green appearance. In spring the hot days are followed 

 by cold nights, so that the difference between the average temperature of 

 day and night is enormous. In summer this difference is not so great, in 

 consequence of the intense heat of the sand. Dew has never been observed. 

 In the middle of September begin the long night frosts. In Januaiy the 

 thermometer falls 86 degrees under zero Fahrenheit ( — 38 Centigrade), 

 but there is little snow. In general, the climatic contrasts between the 

 Turkestan deserts and the French Sahara are much stronger than the 

 differences between their respective latitudes would lead one to expect. 

 Thus, for instance, there are about 12 degrees difference between the 

 latitude of Biskra and that of the Turkestan deserts, and still, at Biskra 

 frost is almost unknown, the mean annual temperature being 70 degrees. 



It is evident that, with such climatic discrepancies, the vegetation of 

 the Turkestan deserts, compared with that of the Sahara, must offer great 

 difference, but here also this difference is much stronger than it ought to 

 be, taking particulai'ly into account the high temperature of the Turkestan 

 summers ; and still, among the plants quoted in the above-mentioned 

 paper, there is not a single species common to the Turkestan deserts and 

 the Sahara. The flora of Kara-Kum seems to be so poor that, compared 

 with that of the Sahara, the Algerian desert appears sometimes like a 

 flourishing garden, particularly if we consider, not the large barren portions 

 of the central and western Sahai'a, but the region situated near the 

 northern boundary of the French Sahara, as among others, between Biskra 

 and the oasis of Sidi-Okba and Zadja, where I was struck by the variety 

 and beauty of very peculiar plants which, like brilliant jewels, glitter in 

 the midst of the sea of sand.^ 



Now, if from the Turkestar deserts we direct our steps to the east^ 

 we reach the long chain of moTintains which separate Siberia from Central 

 Asia, a chain composed of the mountain groups collectively called Altai, 

 Sayan, and Jablonovo'i, the tviO first of which I have visited, but without 

 crossing them, in order to descend into the desert of Gobi. This immense 

 desert— the largest in the worH after the Sahara — begins almost imme- 

 diately at the southern foot of the just-mentioned Siberian mountain 

 range, and extends to the soith to the chain of Kuenlun and its eastern 

 ramifications, having from mrth to south about 1,800 miles, and about 

 4,000 miles from east to vest — viz., from the mountainous chain of 

 Changan to the country of Tarkand. The geographical position of the 

 Gobi desert is, in the average,j)etween the .35th and 45th degrees of northern 

 latitude, and consequently amost under the latitude of Italy — a fact ren- 

 dering still more remarkably the climate of the desert, which exemplifies 

 in a most extraordinary wajj and even more than in the Turkestan deserts, 

 the influence of eastern longtudes, combined with the powerful radiation 

 of large, more or less flat, surfaces ; for, though under the latitude of Italy, 

 but about 40 degrees more \) the east, the Gobi offers the strongest con- 

 trasts between the seasons : iie summer reminding one of tropical heat, and 

 the winter of the cold of thdpolar regions, and that not only on account of 

 its intensity, but also its dij-ation — a fact of which Colonel Prejevalsky 

 gives us the following striing example: — In the mountainous tract of 

 Gansu, at a height of notmuch more than 3,000 feet, on May 16, the 

 thermometer indicated 25 ffigrees (—4° Centigrade), and on May 28, the 

 snow falling copiously foriied on the soil a coating 5^ inches (16 centi- 



' I have given an account cJthis curioixs desert flora in my last work upon Algeria, 

 pp. 299-300, and 303-306. 



