858 
THROUGH ASIA 
the lake is muddy, cold, and perfectly fresh ; whereas the 
water which flows out is clear as crystal, is some degrees 
warmer, and has a soupgon of salt in it, all phenomena so 
simple as to stand in no need of explanation. 
The large lake of Issyk-kul in Semiryechensk presents 
a problem which has been a standing puzzle to geologists, 
hydrographers, and travellers. The river Chu, a stream of 
some magnitude, flows across the perfectly level plain to 
within a couple of miles or so of the west end of the lake ; 
but instead of entering the lake, as would naturally be 
expected, it flows on towards the north-west and pierces 
the great mountain-range of Ala-tau. Nor does it in any 
way add to the volume of the lake, except occasionally 
during periods of very unusual flood, and then it does 
sometimes send off a small side-branch. Various complex 
theories, partly geological, partly hydrographical, have been 
invented to account for this. I too have hit upon a theory, 
which enjoys the advantage of being at any rate simple. 
The relation of the Chu to the Issyk-kul is exactly 
paralleled by the relation of the Hadick-gol and the 
Koncheh-daria to the Bagrash-koll. 
The distance between the point where the delta of the 
Hadick-gol enters the lake and the point at which the 
Koncheh-daria runs out of it is only 15 or 16 miles. 
Between the two points the lake is shallow and overgrown 
with luxuriant reeds ; whereas its middle and eastern parts 
are deep and free from vegetation. Moreover the Hadick- 
o-ol sends off a long’ delta-arm tow’ards the Koncheh-daria. 
On the way to Kara-shahr we passed, at the distance of 
about an hour and a half, a dry channel, which branched 
off from the Hadick-gol and joined the Koncheh-daria. I 
inquired into its character, and was told, that every fifth or 
eighth year the Hadick-gol overflows its banks, and a 
portion of its flood-water makes its way along that dry 
channel direct to the Koncheh-daria without passing through 
Bagrash-kdll. The contour of the ground, I may remark, 
is almost perfectly level, and its elevation above the surface 
of the lake trifling. 
Assume now, that the overflow occurs fifteen times in 
