130 LTTE fOOLI ALS OF NG OL OGNA 
weight of the levers, etc., bearing on the diaphragms will always 
be the same, and so this source of error will be eliminated. 
The index of a well-made aneroid should be exceedingly fine, 
flattened vertically, and should revolve as close as possible to 
the graduated arc, but even when both these conditions are 
fulfilled the reading may be varied as much as twenty feet by 
changing the line of sight, hence the instrument should always 
be held so that the line of sight will be perpendicular to the dial. 
Errors due to changes in temperature — All first-class instru- 
ments are supposed to be compensated for changes in tempera- 
ture, but the compensation is never perfect. Each aneroid 
should be tested under varying temperatures, and the ratio of 
error noted. It will generally be found so small that it may be 
disregarded except in extreme cases, but if the ratio is large the 
instrument should be rejected. 
Errors due to changes in atmospheric pressure.— The atmosphere 
is filled with eddies, formed by ascending and descending cur- 
rents, which are called cyclones and anticyclones respectively. 
These eddies generally originate over the elevated plateaus of 
the cordilleran system, spread out until their diameter is measured 
by hundreds of ‘miles, and move in an easterly direction across 
the continent, with an average velocity of twenty to thirty miles 
per hour. 
In the eddies formed by ascending currents, or cyclones, as the 
air rises, it is replaced by surface currents which set in from all 
directions, with a spiral motion, toward the center. A careful 
oy 
study of an area marked “low” on the daily map issued by 
the National Weather Bureau will make this clear. As the 
cyclone moves toward some easterly point a wind from that 
direction tells of its approach, while a change in direction from 
east, through north or south, to some westerly point tells of its 
passage. 
As soon as the particles of air come within the influence of 
the eddy, they are drawn upward, at first slowly, then more 
rapidly, until near the center they acquire the almost vertical 
motion of the vortex. This upward tendency of the air lessens 
