ANEROID BAROMETER IN GEOLOGICAL SURVEYING. 135 
At such times the observer should proceed with great caution, or 
what is far better, cease work entirely. 
An explanation of these irregularities may be found in the 
fact that cyclones and anticyclones are not the simple eddies that 
they are commonly supposed to be, but contain within them many 
smaller ones, each with the characteristics of the larger, but which 
pass over us so swiftly, or are so shallow, that they escape notice. 
It is these small eddies that give rise to the cumulus clouds and 
cirrus stripes of the anticyclone and to the cirrus stripes and clear 
spaces of the cyclone, while to the larger ones, or groups of them, 
we owe our thunder storms, and tornadoes. They are technically 
known as surges. 
Having shown that by the exercise of the same intelligent 
attention which must accompany any scientific work, the sources 
of error which inhere in the aneroid may be kept under control, 
and that results sufficiently accurate for the purpose indicated 
may be reached, we come next to consider the method of pro- 
cedure in the field. 
Having assumed a datum plane, far enough below the surface 
at the starting point so that no negative readings need be recorded, 
we run a series of base lines crossing the area at intervals of five 
or ten miles, using every precaution to guard against errors, and 
establishing stations at frequent intervals, so well described that 
they may be recognized when again met with. These base lines 
should all be doubled, that is, each should be run in the opposite 
direction, readings being taken at the stations established during 
the first traverse. A comparison of the two sets of readings will 
usually enable the observer to eliminate all errors, if the work 
has been well done, but if wide differences are still found, a third 
traverse, more carefully made, should correct them. These base 
lines will furnish the means of establishing the water slopes of 
the streams that cross the area, and these, together with the base 
lines themselves, will enable the observer to frequently check 
his work while making the traverses with which the details are 
filled in. 
No topographic survey can be absolutely correct, that is, the 
