IN THE PRESSURE OF THE ATMOSPHERE. 481 
other side of the line of demarcation between the continental and 
the sea-climate. As these annual oscillations amount at Barnaul 
to 81 lines, it will be evident that considerable errors might be 
committed by not considering these relations in the levelling 
with the barometer. As journeys are generally undertaken during 
the summer months, it results that a much greater elevation 
above the Europzean stations will be ascribed to the Asiatic sta- 
tions than they in reality possess*. The numerous anomalies 
which have been observed on the frontier of Europe and Asia 
between the barometrical and trigonometrical levellings, may 
receive more or less explanation from these relations. We were 
certainly not previously acquainted with such considerable and 
regular curves as those at Barnaul and Nertchinsk, except from 
the district of the monsoons, and were consequently not con- 
vinced of the necessity of a correction for the annual period. 
But the assumption, that the high land of Asia forms a boun- 
dary towards the north for this phenomenon of the periodical 
variation of the atmospheric pressure, is refuted by observa- 
tions carried on for several years; we now know that it is exhi- 
bited with the same energy at the Altai as in the low land of 
the Ganges, and is far more considerable there than at the 
mouth of the Mississippi; and that the barometrical relations of 
Bombay are repeated at Nicolaieff; annual means only there- 
fore can be employed for barometrical levellings of the plains. 
What is represented in the drawings is expressed in the nu- 
merical values of the Tables at pp. 489 to493. I have based 
their calculation on the most trustworthy and recent observations 
which have been accessible to me. All of them, with the ex- 
ception of some localities in Table VI., are corrected for the 
thermal expansion. _ Table IX. leaves most to desire, because we 
do not yet possess the means for eliminating the daily oscillation 
resulting from the elasticity of the vapour. For the calculation 
of the averages, I have selected those hours which appeared to 
give the nearest approximation ; I must not however enter into 
the details of the calculations. The number of hygrometrical 
observations being limited, it appeared to me most appropriate, 
in order to determine the pressure of the dry air, to combine 
the results for the elasticity of the vapour, for such as were ex- 
tant, with the longest series of annual results for the barometer, 
even when hygrometrical observations did not exist for every 
* See Galle in Poggendorfl’s Annalen, vol. xlviii. pp. 58,379. 
