10 
apart (about one-tenth of a meter) and calculate the pressure per 
unit area on such a plane at a depth of about 1 decimeter. The pres- 
sure of the atmosphere being subject to comparatively slight and 
compensating variations can be totally disregarded throughout 
hydrodynamic works. (See p. 45.) We have given by definition 
values of pressure = weight per unit area = 
area, height, density, acceleration of gravity. 
area 
Since the area values cancel, we have 
pressure=h g @. 
But it has been determined that g h=D, where D equals 1 dynamic 
decimeter. 
Substituting: ri 
p=qD 
Now it remains to find a suitable system of units of pressure based 
upon the value equal to a water column 1 dynamic decimeter high 
and possessing a mean density q. 
The most common example of natural pressure with which we are 
familiar is that of the atmosphere. It has been a practice, long 
established, to balance the perpendicular column of the atmospheric 
envelope against an equal cross-sectional area of mercury. This is a 
well-known experiment of any physics laboratory in which mercury 
has come to be adopted because of its great density; other liquids 
being forced to too great a height by the balance. We employ exactly 
the same equation, of course, as evolved in the case of a motionless 
ocean; in fact, we might imagine finding the pressure at various 
depths in the sea, theoretically, by means of a balanced column of 
mercury. 
Tt has been found that at 0° C. and 45° latitude at sea level, the 
normal height to which mercury is forced by the ever pressing air 
envelope, is 0.76 meters, sometimes termed an ‘atmosphere.’ 
Since the acceleration of gravity at 45° latitude is known, viz, 9.8 
meters, and the density of mercury at 0° C. is 13.59, let us calculate 
the pressure p in meter-ton-second units—i. e., the system upon 
which previous dynamic figures have been based. Substituting in 
p=qgh, we have p=13.59 X 9.8 X 0.76 =101.218. 
VY. Bjerknes has used this quantity of 101.218 as a guide in deciding 
upon the value ascribable to p. He has selected as a unit suitable 
for hydrodynamic computations, the nearest integral number of 
10 to 101.218, viz,.100, and has called this a bar. A bar is approxi- 
mately the pressure exerted by a column of water 10 meters in height; 
therefore the pressure of 1 meter of water is very nearly equal to the 
