16 
uvoward. 
The vabor pressure at deck level is usually less than 
at the sea surface. In this case the 
average vertical distribution of vapor oressure is seen to be a con- 
tinuous decrease from the sea surface to deck level and beyond, the 
deercase being very steep -t the surface. Actually the change of 
vapor pressure or temperature from the surface to an elevation a 
centimeter or two above is aout as much as the change fron the latter 
to deck level. The gradient at deck level is already so slight that 
the elevation of deck level has little effect on temnerature or hu- 
midity measurementsS.. 
An instantaneous vertical distributian has no such regular shape, 
and at any fixed level temperature and humidity usually fluctuate 
rapidly. This is often noticed in using a psychrometer, and is simply 
a manifestaticn of the irregular turbulent motion. 
The intensity of mixing varies with elevation above the surface,” 
and at a given elevation depends on the wind force, on the stability 
of the air, and on the roughness characteristics of the surface. The 
last two factors are fortunately usvally unimportant for elevations 
sufficiently low over the open ocean. Hence the intensity of mixing 
can for practical purposes be expressed in terms of the wind force. 
Wind force is in most ways preferable in this work to wind speed. 
The degree of turbulent mixing in the air is intimately associated with 
the shearing stress between the air and water, and the latter is di- 
rectly evidenced in the pepe anes of the sea surface. Hence the 
wind force estimated from the appsarance of the sea according to the 
instructions in Circular M of the United States WYeather Bureau is ex- 
cellently suited as an indication of turbulent mixing. It may be 
noted that such a scale based on the appearance of the surface is an 
