XXIV PREFACE. 
In studying the great map compiled by Klaproth from caravan itiner- 
aries and other sources, I was puzzled by the great number of lakes dotted 
over the plains between the Aral Sea and the Siberian steppes—bodies of 
water without outlets and with no inlets of appreciable size. Their con- 
tinued existence in an arid region seemed inexplicable. And when later 
I crossed Mongolia, the poverty of the pasturage—barely supporting the 
present very sparse population of Mongols, even on Tamtchin-tala, the very 
cradle of Genghis Khan and his hordes—seemed evidence of a marked change 
in natural conditions since the times when these same plains poured forth 
the successive waves of humanity that first threatened China, and ultimately 
conquered that country and devastated the Eurasian continent. 
I passed the winter of 1864-65 in Irkutsk, where I learned from Russian 
officers, who had campaigned on the plains north of the Aral Sea, that the 
countless lakes still existed, but that they were continually and slowly 
shrinking in size. Some of them on the Siberian borderland, which had 
been lakes in the eighteenth century, had actually dried up and were now 
the sites of towns. 
It occurred to me that if the vision were reversed, one looking back 
through time would see the lakes gradually enlarging and coalescing till in 
some remote century they might appear asa large inland sea. There seemed 
to me to exist a relation between the buried cities of the Tarim basin, the 
diminished pasturage and population of Mongolia, the vanished Han-hai 
(dried sea) of the Gobi, the shrinking of the lakes of the Aralo-Caspian 
undrained depressed area, and the overwhelming movements of barbarian 
hordes toward China and Europe. 
The subject had the fascination of a mirage, in which dissolving glimpses 
of a vanishing world mirrored the parallel progress of nature and man 
toward desolation within and destruction without. A great inland sea 
shrinking to disappearance presupposed former conditions favorable to its 
creation and to the maintenance, on its border, of a vegetation and popula- 
tion incompatible with the present aridity. What might have been the 
origin of such a sea and how far back in the perspective of geological time, 
and what the cause of the apparently progressive change? 
Agassiz’s beautiful theory of a polar ice-cap in the glacial period had 
already been established, and it seemed possible that climatic influences 
that could produce such continental accumulations of ice might also have 
caused the Caspian and the Aral to coalesce and expand to form a large 
inland sea, and that with the passing away of the climate of the glacial 
period there would necessarily begin the shrinkage of the sea and the pro- 
gressive desiccation of Central Asia. And in this fateful progress should we 
see a resulting struggle for life lasting fifteen centuries, from the third cen- 
