580 
that a supposed slowing down of the cyclone without 
change in its structure would lead to increasing surface 
pressure tendencies. From this it can be concluded 
that the speed of a given cyclone is stable as long as 
its total three-dimensional structure remains the same. 
The pressure tendencies are then also stable although 
they are made up as small differences of large opposite 
contributions of horizontal divergence and convergence. 
A real increase in the divergence effects of the upper 
wave would come from a lowering of the level of non- 
divergence and an inherent increase of that part of 
the atmosphere in which the air current is supercritical. 
According to (8), this may take place through one of 
the following changes of parameters in the upper wave: 
(1) an increase of the speed v, of the upper westerlies, 
(2) a decrease of angular wave length 27/n, and (3) 
travel towards higher latitudes. In all these cases the 
compensating mass-divergence effects from low levels 
automatically become greater as the speed of the cy- 
MECHANICS OF PRESSURE SYSTEMS 
has corroborated these findings. A summary of the 
careful and extensive work in that field was published 
in 1948 by Miller [20]. From that report we also know 
that the vertical motion of extensive air masses usually 
is less than 3 cm sec even at the level of nondiver® 
gence, where the maximum upward and downward 
values of momentum occur. Higher values of vertical 
motion up to 10 cm sec should occur over narrow 
zones near fronts, while the occurrences of updrafts 
and downdrafts of several meters per second are re- 
stricted to small parts of individual convective clouds. 
Typical patterns of temperature distribution in the 
extratropical cyclone can be seen from the sample 
cyclones described later in this article. The most fre- 
quent development of the temperature pattern in the 
lower tropospheric part of the vortex can be repre- 
sented schematically by the maps and profiles in Fig. 
2. The incipient cyclone (Fig. 2a) coincides with the 
apex of a warm tongue formed on a “front’’ across 
Fig. 2—Successive stages of development of a frontal wave to an occluded vortex. 
clone increases. The occurrence of excessive values of 
barometric tendencies is thus automatically avoided. 
The slowing down which is normally observed in 
deep and extensive cyclones is associated with the 
great depth of atmosphere moving in a closed cyclonic 
flow pattern. If that pattern has a subcritical eccen- 
tricity, which is the more frequent case, it will maintain 
mass convergence in the eastern half and mass diver- 
gence in the western half. The influence of an upper 
wave pattern can then only barely overcompensate 
the divergence effects below, and the resulting pressure 
tendencies will be small. A final reversal of tendencies, 
and a retrograding of the cyclone, will result if the low- 
level mass-divergence effects overcompensate those 
from the upper wave pattern. 
The distribution of horizontal divergence also deter- 
mines the -vertical motion, which in the “smoothed”’ 
cyclone model always goes upward in the front half 
and downward in the rear (Fig. 1). This model feature 
agrees with the observed distribution of cloudiness and 
precipitation in the cyclones. Modern aerological analy- 
sis of the field of vertical motion carried on by the 
Department of Meteorology at New York University 
which the horizontal temperature gradient reaches a 
maximum. The frontal surface rises towards the cold 
side at an angle of inclination averaging around one in 
a hundred. It conserves its identity from day to day 
and moves along at a speed determinable from the 
winds through the kinematic boundary condition. The 
wave amplitude increases as the cyclone matures (Fig. 
2b) and the central pressure decreases. Next follows 
the ‘‘occlusion”’ process (Fig. 2, c and d) during which 
the warm tongue is lifted from the ground, first near 
the center, later also farther out. The occluded front 
formed at the junction of the two cold wedges tends 
tO wrap around the cyclone center as part of a spiral, 
and the same shape is found for the warm tongue in 
all levels of closed cyclonic circulation. In the profile 
in Fig. 2c, which is placed at a short distance south of 
the cyclone center, the occluded front is a warm-front 
type, that is, the cold wedge behind the occlusion is 
less cold than the one in front. This is also true of 
Fig. 2d, but it can usually be assumed that farther 
south the occlusion is a cold-front type. Where the 
transition from one occlusion model to another takes 
place the occluded front on the map must show a little 
