154 
quency distribution of thunderstorms 
here shows a night-time maximum ai 
this season. The more probable con- 
clusion is that the impulses generating 
these storms originate not in the 
lower layers where the stability is so 
marked, but in the upper layers. One 
factor which readily suggests itself 
as an important one is radiational 
cooling of the moist air aloft. At 
the top of these moist currents is nor- 
mally found another stable zone, fre- 
quently an inversion. Cooling of 
moist air aloft becomes intensified 
when clouds form at the boundary 
surface, for clouds act as black bodies. 
The process thus envisioned demands 
that convective stirring occurs in the 
layer cooled from above, just as it 
must occur in a layer heated from be- 
low. However, the processes of re- 
moving heat aloft and supplying it 
from below probably lead to convec- 
tion of a somewhat different nature. 
Heating from below, when associated 
with steep lapse rates, may result in 
rapid upward motions of small ele- 
ments, while cooling from above is 
probably slower acting and is trans- 
ferred slowly downward. 
For a thunderstorm to become en- 
ergetic and produce sizable amounts 
of precipitation it must have available 
an ample supply of moisture. Since 
in summer the maximum concentra- 
tion of moisture is generally in the 
surface layers, it follows that what- 
ever the origin of the convection, over- 
turning must eventually take place 
throughout these lowest layers if the 
storm is to become of appreciable 
intensity. Thus if convection aloft is 
caused, let us say, by radiational cool- 
ing of a cloud layer, we must draw 
upon some supply of energy to carry 
this convection into the moisture-rich 
surface layers. This energy some- 
times appears in layers which are 
conditionally unstable and moist. 
AIR MASS ANALYSIS 
The problem of forecasting summer 
showers therefore depends not only 
upon the stratification existing aloft 
in the early morning when soundings 
are made, nor entirely upon the 
changes brought about by diurnal 
heating (see Arts. II, VIII), but also 
upon the advection of moist and dry 
tongues aloft. The best method of 
estimating these advective changes 
lies in the isentropic analysis. In this 
manner we are able to detect in ad- 
vance the likelihood of soundings of 
any of the above types being trans- 
formed into other types through 
advection. 
When reliable isentropic charts are 
at hand, together with the corres- 
ponding cross sections, it is possible 
to obtain information of forecasting 
value which is not easily obtained 
from other charts. The problem of 
shower and thunderstorm forecasting, 
therefore, revolves chiefly about the 
determination of the lapse rate and 
the source and availability of mois- 
ture. Tongues of dry and moist air, 
as shown by the isentropic charts, 
may be identified from day to day by 
means of these charts. In summer it 
is found almost invariably that thun- 
derstorm activity and showers are 
associated with the moist tongues, 
while the dry tongues are free of 
convective precipitation. Furthermore, 
by making use of cross sections one 
is able to form an idea of the repre- 
sentativeness of the chosen isentropic 
chart. Sources of moisture indicated 
within the frictional layer may be 
found to be shallow and not represen- 
tative of conditions in higher isen- 
tropic surfaces. The presence or in- 
vasion of dry air aloft would then 
counteract the possibility of showers 
and thunderstorms. Thus, though an 
energy diagram may offer indications 
of possibilities for shower activity, 
one should use the isentropic chart to 
