UPPER AIR—GOLD AND HARWOOD. 267 
The remarkable feature is the relatively high temperature and low 
value of H. in March and September. This peculiarity and the fact 
that T. is least near the equator suggest that the general nature of the 
process may be as follows. The cool air above the equator moves 
polewards, and in the natural course descends again to feed the trade 
winds, Owing to the irregularities of the earth’s surface, the change 
of seasons, and the very considerable difference between the northern 
and southern hemispheres, the process will be neither regular nor 
symmetrical. Consequently, the equatorial cold air will encroach on 
the advective region of temperate latitudes, and such encroachments 
will produce anticyclonic regions. The advective atmosphere would 
be reached there at a higher level, and initially at a lower tempera- 
ture than in the average state, but the temperature would be gradually 
raised by absorption of thermal radiation to the normal value for that 
latitude. 
The fact that H, has minimum values in March and September, 
when equatorial temperatures are highest, appears at first to be con- 
trary to this view; but the first effect of increased temperature will 
be to increase the strength of the trade winds, and as at the same time 
there is a transference of air across the equator to the southern hemi- 
sphere, a transference which can be made only through the upper 
return current, there will be a deficiency of descending air, and the 
equatorial cold air will encroach less than usual on the northern ad- 
vective region. The reverse process would be expected to occur in 
September, but the autumnal transference of air to the northern 
hemisphere will be initially much more intense toward the great con- 
tinental regions than to the Atlantic and European area, and it may 
well be that the equatorial current again encroaches less than usual 
on that region. It may be expected that the value of H, in Asia 
and America will not show the September minimum. 
The explanation of the discontinuity in the temperature gradient 
appears to be this. The fall of temperature is governed mainly by 
convection, and a necessary condition for convection to persist is that 
the radiation shall exceed the absorption in the upper layers of the 
convective system. A limit is therefore set to the height to which 
convection can extend, and at this limit the discontinuity in the fall 
of temperature occurs. It has been shown that the observed height 
is about the same as the limiting height of the convective system 
found from theoretical considerations based on the experimental 
knowledge of the radiating power of the atmosphere. 
The results of the observations of wind velocity may be briefly sum- 
marized as follows: In general, the velocity increases with height, the 
greater part of the increase up to 2,000 meters taking place in the 
layers immediately above the surface; 75 per cent of the total increase 
takes place in the first 160 meters. Above 500 meters numerous cases 
