National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

In times of flood, however, when the flow per second has been known to rise to
17,000 cubic feet, the Domain is allowed to divert as much of the surplus as it can
use or store, provided always that 90 per cent of the minimum amount remains in
the main stream. Accordingly the Domain receives more than 10 per cent of the
total flow—from 20 per cent to 25 per cent according to Mr. Nikrashevich. For
the sake of being on the safe side, let us call it 15 per cent of the total. This
amount of water supports to-day an agricultural population numbering 15,000
souls and a further body of workmen, artisans, and tradespeople, numbering
with their families not over 10,000.

In 1904 there was danger that the available water-supply would prove some-
what scanty for the thorough irrigation of the crops of even this number of people.
In other words 15 per cent of the water under the most favorable conditions of
irrigation can not support over 25,000 people; and all the water can not support
over 166,000, or in round numbers 200,000 people, twice the present population.
But even this liberal estimate leaves us with a deficit of 300,000 souls, when
compared with the 500,000 whom we have taken as the population of the oasis
in antiquity. Whatever the exact figures may be, it seems highly probable that
the present water of the Murg-ab can not possibly support so large a population
as that which actually peopled the oasis in antiquity. There appears to be no way
of explaining this discrepancy except by supposing that in earlier times the flow
of the Murg-ab was larger because of greater rainfall, or because of less rapid
evaporation due to lower temperature.

Note to Chapter XII.—This chapter was written in September, 1904, and is here printed
in its original form, the only changes being a recasting of the first paragraph, and a few
verbal alterations. During the three years and a half which have intervened between
the writing of the chapter and the time of going to press, March, 1908, the writer has had
the opportunity to make a somewhat extensive study of the climatic problems here touched
upon. The results of the later studies are set forth in a volume entitled "The Pulse of
Asia" (Houghton, Mifflin, and Company, 1907). Evidence gathered from the ruins of
Chinese Turkestan and from the fluctuations of the Caspian Sea, as well as from other
sources, indicates that in the regions from the Caspian on the west to China Proper on the
east all Central Asia has undergone a series of climatic pulsations during historic times.
There seems to be strong evidence that at the time of Christ or earlier the climate was much
moister and more propitious than it now is. Then during the first few centuries of the
Christian era there appears to have been an epoch of increasing aridity. It culminated
about 500 A. D., at which time the climate appears to have been drier than at present. Next
came an epoch of more propitious climate which reached its acme about 900 A. D. There is
a little evidence of a second epoch of aridity which was especially marked in the twelfth
century. Finally, in the later Middle Ages, a rise in the level of the Caspian Sea and the
condition of certain ruins render it probable that climatic conditions once again became
somewhat favorable, only to give place ere long to the present aridity.

It is not possible in this note to point out the various lines of evidence upon which
the conclusions are based. They are given in full in the work referred to above. It is
evident that if the conclusions are valid, the Merv Oasis must have passed through
several climatic vicissitudes since the time of the earliest kurgans. In that case it should
have experienced a corresponding series of epochs of expansion and prosperity during
times of more propitious climate, and of contraction, distress and retrogression during
times of unpropitious climate. No definite conclusion as to the matter can be reached as
yet, since our knowledge of the history of Merv is still fragmentary, and especially since
there has never been any full study of the distribution of ruins of different ages, whereby