480 DOVE ON THE PERIODICAL VARIATIONS 
sea-climate (figs. 1, 3), and this appears justified from the form 
of these curves being entirely different. In the illustration (fig. 
4) representing the pressure of the dry air, the European and 
Asiatic curves are, on the contrary, similar to each other, and 
pass so gradually into one another, that a separation would ap- 
pear arbitrary. The supposition which immediately presents | 
itself, viz. that from winter to summer the atmospheric pressure 
diminishes regularly as the temperature increases, is actually 
realized in Asia. But this decrease generally occurs in Europe 
only until April, the pressure then increases until autumn, and 
attains a second minimum in November, when it again quickly 
rises. That this deviation is owing to the admixture of vapour, 
is proved by the circumstance, that after elimination of its ela- 
sticity, a uniform increase and decrease of the atmospheric press- 
ure, corresponding to the changes of temperature, is likewise 
evident in the EKuropzan curves of fig. 4. Consequently, what 
the total pressure of the atmosphere loses in Asia by the thermal 
rarefaction of the air, is not wholly compensated by what it gains 
by the increase in the elasticity of the vapour produced by the 
rise of temperature (fig. 4). On the coasts of the Atlantic, on 
the contrary, an over-compensation takes place (fig. 3), the total 
pressure gains more by the increase of elasticity of the vapours 
than it loses by the rarefaction of the air. It is on this account 
that in the most decided sea-climate, in Iceland, the annual curve 
of the atmospheric pressure has an entirely opposite form to that 
in the interior of Asia. Europe forms the transition from the 
one extreme to the other. Should it be required to trace a line 
of demarcation between the continental and the sea-climates, it 
would be where the decreasing pressure during the summer 
passes into an increasing one. St. Petersburg, Moscow, Nico- 
laieff, are on one side of this line; Wilna, Cracow, Ofen, on the 
other. 
The application of the hypsometric formula presupposes the 
elimination of periodical variations of the atmospheric pressure 
at the stations of observation, which condition Ramond attempted 
to satisfy for the daily period by proposing 12 a.m. as the hour 
of observation. The want of this elimination becomes the more 
necessary the greater the deviation of the periodical oscillations, 
and still more requisite when this oscillation takes place in oppo- 
site directions at the stations to be compared. Now such is the 
case when we compare with each other places on the one and the 
