THE INSTABILITY LINE 
it is assumed that the wind at all levels is instan- 
taneously adjusted to the gradient value. Instability 
lines normally occur near the axis of a warm tongue as 
represented in the horizontal distribution of mean vir- 
tual temperature through a layer, say, 20,000 ft in 
depth. The curvature of the component of gradient 
wind flow perpendicular to the axis of the warm tongue 
must, because of this temperature distribution, become 
more anticyclonic (or less cyclonic) with height. The 
component of gradient flow across the axis must there- 
fore tend to increase with elevation, and to carry the 
upper colder air over the lower-level position of the 
axis of warm air, thus decreasing the vertical stability. 
Also, in a manner pointed out by Rossby [10], a decrease 
of mean virtual temperature northward along the axis, 
as is generally observed, must cause an increase of 
geostrophic wind from the west and thus normally act 
in the same direction as the effect of curvature to pro- 
duce greater eastward advection of colder air aloft 
as compared to lower levels. These considerations do 
not, of course, take account of important nongradient 
components of flow that must exist. 
Conditions near the center of a warm tongue are also 
favorable for upward vertical motion if the energy of 
the solenoidal field can be realized, and these same 
conditions provide a means for eviction of air aloft 
because of nongeostrophic components of flow as 
pointed out by Durst and Sutcliffe [2]. Release of latent 
heat of condensation tends to maintain the solenoidal 
field. Some low-level convergence into the warm axis 
can be accounted for by frictional effect in the pressure 
trough. Further low-level convergence is a possibility 
in the layer above surface frictional effects in the pres- 
ence of dynamic instability as pointed out by Solberg 
[12]. However, it is not known from observation that the 
horizontal component of dynamic instability (as first 
discussed in this country by Fulks [4]) is of general 
importance in the case of instability lines, the require- 
ment being that there is appreciable anticyclonic vor- 
ticity in the horizontal wind field. Probably this factor 
is important in the immediate vicinity of the squalls 
but, m general, mstability in the vertical appears to 
be predominant. 
Some Remarks on Tornadoes 
Tornadoes occur normally along instability lines and 
must therefore involve some of the same dynamic 
factors, though not all instability lines are accompanied 
by tornadoes. Studies or discussions of the mechanics 
of tornadoes by Bigelow, Humphreys, Jakl, Showalter, 
and others in this country, Wegener in Germany, some 
Russian investigators, and others have given some hint 
of the factors involved, but much more work remains 
to be done. Basically, there are two main factors to 
be explained: 
1. Source of energy. This has usually been assumed to 
be vertical instability, such as is known to exist in the 
vicinity of instability lines. This explanation is, how- 
ever, not sufficient unless it can be shown that the 
difference between instability lines which produce tor- 
nadoes and those which do not is a matter of degree 
651 
of vertical instability, or that other factors are absent 
in one case and not the other. Showalter [11] suggested 
that precipitation carried aloft and falling ahead of 
the squalls, cooling the air at higher levels, might be 
a means of producing a high degree of instability more 
or less spontaneously. In any case some nearly spon- 
taneous local release of energy seems necessary. An- 
other possible means of creating a vertically unstable 
lapse rate is for the convectively unstable air mass 
ahead of the squall line to be lifted bodily. This would 
itself require some source of energy and seems less likely 
than the cooling by precipitation aloft for which con. 
vective instability is also important m producing a 
steep vertical lapse rate. An important possible source 
of energy other than vertical instability, which cannot 
be ruled out without further study, is that of the al- 
ready existing wind circulation in the vicinity of tor- 
nado formation. Existing kinetic energy fed into the 
whirl would appear to be at least a contributing factor 
once the tornado is formed. 
2. Source of rotation. It seems necessary that the 
rotation of the tornado results from low-level hori- 
zontal convergence within an already existing field of 
cyclonic (or occasionally anticyclonic) vorticity. The 
convergence appears to take place first in the upper 
portion of the warm moist air, roughly 2000-5000 ft 
above the ground (above surface frictional effects), 
and the whirl appears to extend rapidly upward and 
to bore downward to the ground. Tornadoes normally 
occur along or near a line of cyclonic shear, but this 
shear line is usually between the warm air moving 
rapidly northward and the wedge of rain-cooled air 
moving eastward. Since the required vorticity, to be 
effective, must be entirely within the warm air because 
any injection of cooler air at lower levels would tend 
to damp out convection, it is unlikely that the line 
of cyclonic shear is a direct source of the vorticity. A 
more likely source of vorticity is in the interaction 
between the warm and cool air, especially im warm 
tongues that must break off from the northward-moy- 
ing warm air mass and flow into individual convective 
cells. This turning from northward to westward would 
produce cyclonic curvature in the warm air as well as 
some shear, cyclonic on the southern side of the tongue 
and anticyclonic on the northern side, the latter being a 
possible factor in producing the rarely observed anti- 
cyclonic rotation. In this connection it is also necessary 
to consider the irregular nature of outflowing tongues 
of cold air as described by Williams [15] and their 
possible effect on warm air flow. Another source of 
rotation that has long been considered is that of the 
roll cloud of the thunderstorm, the roll being presumed 
to extend down to the ground, resulting in a cyclonic 
whirl on one end, or an anticyclonic whirl if the other 
end should reach the ground; this theory has not been 
fully refuted or confirmed, but appears unlikely in the 
light of descriptive reports. 
Some Recent Studies 
Tepper [13] recently proposed that a “pressure-jump 
line” is an important part of the mechanism of squall 
