EXTRATROPICAL CYCLONES 
the old stationary Hudson Bay mimimum. But the 
young cyclone over the Great Lakes has all the poten- 
tialities of development inherent in the solenoid field 
concentrated along the frontal wave all through the 
troposphere. As a hydrostatic consequence of the wave- 
shaped pattern of isotherms, the upper current above 
the closed center is likewise wave-shaped with an anti- 
cyclonic bend over the warm-front area of upgliding. 
The frontal cyclone deepened at the rapid rate of 14 
mb per twelve hours during November 10, and ended 
up as a storm center of 975 mb over northern Labrador 
on November 12. We shall briefly consider the applica- 
tion of the various theories of upper-air divergence 
which may claim to explain the deepening. Considering 
first the formation of the upper pressure crest which 
precedes the surface cyclone, we shall see the nature 
of the interplay between lower and upper layers. The 
fact that the upper pressure crest moves at a speed of 
only 9 m sec in a current of 70-80 m sec shows that 
the upper wave cannot be a free one. The time and place 
of the first appearance of the pressure crest on the 300- 
mb map (November 9,0300Z) make it likely that the up- 
per crest was formed by the upward motion connected 
with the beginning frontal upglidmg im the nascent 
frontal wave. Once the upper wave is established, upper 
divergence will be located in the area between the pre- 
existing upper pressure trough and the new upper pres- 
sure crest ahead of it. With the trough in north- 
northeast-south-southwest orientation and the pres- 
sure crest in northwest-southeast orientation, the half 
wave length between them shortens northward and 
makes the upper divergence particularly strong in that 
region. This is also the region where the rapid deep- 
ening of the frontal cyclone takes place. So far, the 
dynamics involved in the deepening process conform 
with the principles of Fig. 1. With that basic pattern 
accepted, we must also admit the existence of the fol- 
lowing modifying processes. 
The upper pressure crest is continually being fed from 
below by the rismg motion in the front half of the 
cyclone and therefore is forced to maintain the same 
slow speed of propagation as the surface vortex. When 
the curvature of the anticyclonic bend of upper isobars 
becomes sufficiently strong, the fast upper current is 
unable to follow the isobars in gradient wind fashion, 
and horizontal divergence must result on both sides 
of the upper crest line (see p. 590). This component of 
upper divergence modifies the model in Fig. 1 in the 
sense of extending the pressure fall farther ahead of 
the surface center. 
The upper-air divergence of the Ryd-Scherhag 
theories (p. 586) is also superimposed on the divergence 
and convergence inherent in the wave pattern. It can 
best be judged by comparing the geostrophic wind 
at two successive inflection pomts. The average geo- 
strophie wind between the 30,000-ft and 29,000-ft con- 
tours at the inflection point southwest of the Great 
Lakes was 60 m sec, and between the same contours 
at the inflection point over Labrador it was 42 m 
sec. Therefore it can be concluded that the Great 
Lakes cyclone was situated under a “delta” of the up- 
597 
per current, and that an upper divergence pattern in 
the style of Fig. 8 would be superimposed. This upper 
pattern would tend to produce pressure falls in the 
northern half and pressure rises in the southern half of 
the delta. It may be counted in favor of this reasoning 
that the Great Lakes cyclone proceeded northeastward 
to northern Labrador and not along the 300-mb con- 
tours ahead of the storm which would have meant east- 
northeastward propagation. 
The sample cyclone described above had its early 
development near the ground, where the frontal up- 
gliding of the warm air and the downgliding of the cold 
were a direct consequence of the preceding fronto- 
genesis. Such wave cyclogenesis is typical for all the 
frontogenetical areas of the middle and high latitudes. 
Another type of cyclogenesis occurs at times imde- 
pendently of any pre-existing low-level front. The cyclo- 
genetic mechanism in that case seems to lie in the 
dynamic instability of the upper tropospheric wester- 
lies, which leads to the meandering of the current and 
the subsequent formation of an upper cold low. Cyclo- 
genesis of this kind is described by means of two 
synoptic examples in the following article by H. Palmén. 
In one example, that of November 4, 1946, the upper 
low did not reach a frontogenetical area, and did not 
extend down to the surface. In the other case, that of 
November 17-18, 1948, a frontal wave action can be 
discerned but, in contrast to the case of November 7-10, 
1948, the wave motion started first in the upper tropo- 
sphere and later extended to low levels. 
CONCLUSION 
Summing up our state of knowledge of extratropical 
cyclones, we may classify their formation as being due 
either to unstable frontal wave action or to unstable 
growth of an upper wave trough. A subsequent com- 
bination of both processes is quite frequent, and all the 
strongest cyclones on record seem to have that double 
origin. This article has dealt mainly with the unstable 
frontal wave action, which is in itself a large subject. 
The descriptive study of cyclogenesis and cyclone 
growth is carried out daily by all the weather forecasters 
of middle latitudes, and an adequate report on their 
experiences would have exceeded by far the scope of 
this article. The theoretical study of the model of frontal 
cyclogenesis has been the work of a small number of 
scholars of dynamical meteorology. Viewed in retro- 
spect, the contribution of H. von Helmholtz to this 
field of research appears to be of outstanding im- 
portance and to represent the foundation upon which 
contemporary and future theories should build. The 
Helmholtzian concept of dynamic instability of the 
inertial motion in isentropic surfaces points out which 
properties of the atmospheric fields of mass and motion 
on a rotating earth are conducive to the growth of 
perturbations from infinitesimal to finite amplitudes; 
and it seems obvious that such a fundamental theoreti- 
cal principle must have its applications to the early 
phases of the life cycle of cyclones. However, there is 
still a wide gap to be bridged between the existing 
theory of dynamic instability and the applications called 
