t the winter extremes go far below
this range: during the prevalence of north-east winds the thermometer
drops to -20 deg., or even lower, on the surrounding steppes, while on
the Ust-Urt plateau a temperature of -30 deg. is not uncommon. Again,
the July isotherm of 75 deg. crosses the middle section of the Caspian,
nearly coinciding with the January isotherm of 25 deg., while that of 80
deg. skirts the southern shore of the sea, nearly coinciding with the
January curve of 40 deg., so that the mean annual range over the
northern section of the sea is 60 deg. and over the southern section 40
deg. The former section, which is too shallow to store up any large
amount of heat during the summer, freezes for three or four months along
the shores, effectually stopping navigation on the lower Volga, but out
in the middle ice appears only when driven there by northerly winds.

The prevalent winds of the Caspian blow from the south-east, usually
between October and March, and from the north and north-west, commonly
between July and September. They sometimes continue for days together
with great violence, rendering navigation dangerous and driving the
sea-water up over the shores. They also, by heaping up the water at the
one end of the sea or the other, raise the level temporarily and locally
to the extent of 4 to 8 ft. The currents of the Caspian were
investigated by the Knipovich expedition; it detected two of special
prominence, a south-going current along the west shore and a north-going
current along the east shore. As a consequence of this the temperature
of the water is higher on the Asiatic than on the European side. The
lowest temperature obtained was 35.24 deg. on the bottom in shallow
water, the highest 70.7 deg. on the surface. But in March the
temperature, as also the salinity, was tolerably uniform throughout all
the layers of water. Another interesting fact ascertained by the same
expedition is that the amount of oxygen contained in the water decreases
rapidly with the depth: off Derbent in the middle section of the sea the
amount diminished from 5.6 cc. per litre at a depth of 100 metres (330
ft.) to 0.32 cc. per litre at a depth of 700 metres (say 2300 ft.). At
the same spot samples of water drawn from the bottom were found to
contain 0.3 cc. of sulphuretted hydrogen per litre. In the southern
section of the sea the decrease is not so rapid. In this latter section
Spindler ascertained in July 1897 that the tempera