536 THE FRESH- WATER LOCHS OF SCOTLAND 
vast accumulations of dunes, 300 to 400 feet high. In addition, 
there is another system of sand-dunes at right angles to the first, 
running from east to west, built by winds blowing from north to 
south during the winter. The sand-dunes, therefore, form a network, 
in which are depressions with a clay bottom swept free from sand. 
By the autumn the lakes are half empty. ^ Previous to 1876 Lob Nor 
was placed in nearlv all the maps at a position which agreed with 
the accounts of ancient Chinese geographers. In that year the 
Russian, Prejevalsky,-^ discovered two closely connected lake-basins, 
Kara Buran and Kara Koshun, whose waters were fresh, fully one 
degree farther south and considerably to the east of the site of 
the old Lob-Nor. These lake-basins he nevertheless reg-arded as 
being identical with the Lob-Nor of the Chinese. This identification 
was disputed by Baron von Richthofen ^ on the ground that the Lob- 
Nor, the salt lake of the Chinese geographers and the terminal lake of 
a water-system, could not be filled with fresh water, and that this 
lake must be a modern formation. In 1895 Sven Hedin * ascertained 
that the River I'arim empties part of its waters into another lake, 
or rather a string of lakes (Avullu-koll, Kara-koll, Tayek-kcill, and 
Arka-koll), situated in the latitude given to the Chinese Lob-Nor, 
and thus so far justified the views of Richthofen and confirmed the 
Chinese accounts. At the same time he advanced reasons for believing 
that Prejevalsky's lake-basins, the southern Lob-Nor, are of quite 
recent origin — indeed, he fixed upon the year 1720 as the date of their 
formation. Besides this, he ai-gued that there exists a close inter- 
relation between the northern Lob-Nor lakes and the southern Lob-Nor 
lakes, so that as the water in the one group increases it decreases to 
the same proportion and volume in the other. He also considered 
that the five lakes of northern Lob-Nor are slowly moving westwards 
under the incessant impetus of wind and sand-storm (bivran). 
These conclusions were afterwards controverted by the Russian 
geographer Kozloff, who visited the Lob-Nor region in 1893-94, 
before Sven Hedin. In 1900 Hedin, following up the course of the 
Tarim, discovered at the foot of Kurruk-Tagh the basin of a 
desiccated salt lake, which he holds to be the true Lob-Nor of the 
Chinese geographers ; and at the same time he found that the Kara- 
Koshun or Lob-Nor of Prejevalsky had extended towards the north, 
but shrunk towards the south. Thus the old Lob-Nor no longer 
exists, but in place of it are a number of much smaller lakes of 
newer formation. It is interesting to note that Dr Stein, in his paper 
1 Sven Hedin, Central Asia and Tibet, vol. i. p. 419, London, 1903. 
2 Verh. Ges. Erclk. Berlin, Bd. v. p. 121, 1878. 
^ From Kulja across the Tian Sha?i to Lob-Nor, pp. 98-145, London, 1879. 
^ Through Asia, vol. i. pp. \b et seq., vol. ii. pp. 864 et seq., London, 1898. 
