112 
precipitation during the colder sea- 
sons. 
Superior or subsiding air, (S): The 
assumption that the dryness of this 
air mass is due to subsidence is not 
necessarily always correct in summer, 
because rapid surface heating of rela- 
titvely dry air at that season may 
produce a deep column of air whose 
moisture content is far from the sa- 
turation values. When such an air 
mass is cooled during the night by 
surface radiation some slow sinking 
may occur in the layers near the sur- 
face and those layers which are not 
cooled by radiation will show low rela- 
tive humidities by the time of the 
airplane observation the next day. 
Since S air is recognized only on the 
basis of relative humidities less than 
AIR MASS ANALYSIS 
40%, the source of this air can there- 
fore be traced to subsiding polar or 
tropical air or to rapid daytime heat- 
ing of either of those air masses. 
The range in specific humidity and 
equivalent potential temperature is 
considerable at all elevations for the 
dry type of air called S, so it can be 
assumed from this evidence also that 
the source of S air may be either 
polar or tropical or a stratification of 
air from both sources with a tendency 
for horizontal mixing along the isen- 
tropic during the night and vertical 
convection during the day resulting 
in a dissipation of the concentration 
of moisture. The orographic effect 
also plays an important role in the 
development of S air east of the 
Rocky Mountains. 
AtR MASS CLASSIFICATION 
Our study indicates that it is possi- 
ble to make some reasonable standard- 
ization of air mass classification for 
synoptic purposes but that any classi- 
fication falls far short of definitely 
identifying the thermodynamic prop- 
erties of the different air masses. In 
other words the mean values of each of 
the various properties (see figs. 12 to 
15 of original article) show definite 
groupings for the different air masses, 
and the individual values show definite 
limits for the properties of fresh trop- 
ical and fresh polar outbreaks. How- 
ever, the modifying influences affect- 
ing air masses, especially in the 
summer months, result in a very wide 
range of equivalent-potential tempera- 
tures as indicated on the frequency 
distribution charts (figs. 5 and 6 of 
original article). It can be seen from 
a study of these charts that the classi- 
fication is relative for any one day, 
and no definite limits of equivalent- 
potential temperature have been in 
use. This seems regrettable for sta- 
tistical purposes but in _ practical 
synoptic work it can hardly be avoided. 
It is usually the practice to label the 
air masses differently on either side of 
a front, and since polar air is some- 
times modified very rapidly, it happens 
that modified polar air behind a cold 
front on one weather map may have a 
higher equivalent-potential tempera- 
ture than a modified tropical current 
behind a warm front on a map a week 
later. Since all tropical air is polar 
air modified by surface effects, any 
classification of air masses must be 
only a compromise as to number of 
types. The author is of the opinion 
that no purpose is served by increasing 
the number of types. 
THE AiR MASS CYCLE 
It is possible to identify from the 
mean seasonal properties for the dif- 
ferent air masses, two distinct cycles 
of transition from polar to tropical 
air, one a moist cycle with rapid addi- 
tion of moisture, the other a dry cycle 
with a marked subsidence and slow 
isentropic mixing in the early stages. 
