of Edinburgh, Session 1868 - 69 . 471 
suddenly shifted to W. or N.W., and the weather then to have be- 
come colder and clearer and the air drier, it is certain that a storm 
of greater or less magnitude has passed over the region, and since 
such storms are attended with great fluctuations of the barometer, 
it is plain that if the observation of pressure was made during 
these changes, it is worse than useless as a datum for the determi- 
nation of the height of the place. It should, therefore, be alto- 
gether rejected. 
These cases are given as examples of the method by which 
observations, as made by travellers, should be critically examined 
before they are made use of in calculating heights. It is pro- 
bably from inattention to these simple directions — travellers not 
recording the required data, which can all be recorded without 
instruments, and computers not giving weight to such observations 
when recorded — that a large number of the grosser discrepancies, 
given in works of Physical Geography, have arisen. Many of the 
larger errors are, of course, due to the use of imperfect instruments 
and a want of practice in the observer. 
An illustration of errors in the statement of heights may be given. 
The following places are situated in the neighbourhood of the Ural 
Mountains ; the heights are those given by the most recent author- 
ities, and a column is given showing the number of years for which 
the averages of mean annual pressure have been calculated : — 
Place. 
Lat. N. 
Long. E. 
Height 
in 
Feet. 
Years 
of 
Average. 
Mean Pres- 
sure at 32° 
and sea-level 
Bogoslovsk, . . 
O / 
59 45 
° / 
60 2 
600 
26 
Inches. 
29-862 
Nijni-Tagilsk, 
57 57 
59 53 
730 
21 
30-088 
Catherinenburg, . 
66 49 
60 35 
800 
18 
29-835 
Zlatoust, . . . 
55 10 
59 40 
1200 
28 
29-835 
From the above annual mean pressures it is evident that the height 
of Nijni-Tagilsk is over stated, the true height being, probably 
about 250 feet less than what has hitherto been assigned to it. 
Since it has, unfortunately, been the general practice not to 
publish the original observations, but only the heights deduced from 
them, it will be impossible, except in a comparatively small number 
