1238 
are still uncertain. The two atmospheric phenomena to 
which these studies show the most resemblance are 
tropical cyclones and tornadoes. Generally speaking, 
aside from qualitative details, the experiments lead 
only to conclusions that result also from elementary 
theory and go back to the seventeenth and eighteenth 
century beginnings of the discussion. An example of 
this is the conclusion that necessary conditions for 
vortices of these types are concentrated regions of 
ascent or descent (thermally produced or due to the 
initial field of mechanical motion) combined with initial 
absolute circulations in fluid circuits enclosing the re- 
gions. One fairly general result obtained by experiment, 
that as far as I know was not anticipated m theory, 
concerns a general tendency of concentrated vortices 
to arrange themselves so as to end at a boundary, for 
example at the earth’s surface (Fujiwhara [21], Hale 
and Luckey [30]). Thus one might generate a horizontal 
vortex and have it break and arrange itself so as to 
meet the surface vertically. (See Wegener [70] for tor- 
nado developments of this kind, or Fujiwhara for 
possible synoptic applications.) This result is, of course, 
probably related to Helmholtz’ proposition that a vor- 
tex filament cannot have a free end in the interior of a 
fluid in continuous motion. 
However, one gets the impression that, at least with 
respect to the tornado problem, there is something 
worth studying in much more detail in some of these 
experiments. The phenomena in Letzmann’s work, or in 
that of Dines and Weyher in air with water vapor as 
the tracing element, show almost complete qualitative 
similarity to actual tornadoes, even to the hollow sheath 
apparent along the tornado tube in many photographs, 
or to the swellings traveling along the tube as noted in 
a few descriptions of the natural phenomenon [70]. 
It is a curious fact in this connection that hardly any- 
where in these investigations, or for that matter mm 
most of those to be mentioned later, has any real 
attempt been made to make numerical measurements 
of even the simplest and most obvious field quantities. 
The principal exception among the investigations men- 
tioned above is that of Lunelund [43], who generated 
vortices in a liquid with a free surface by rotating a 
simple shaft below the surface. He then measured the 
geometry of the free surface on photographs, mainly for 
comparison with the form to be expected for a Rankine 
combined vortex (‘vr vortex in the outer part and 
solid rotation in the inner part). Another exception is 
an investigation carried out by Sttimke [62] at Gottingen 
in 1940 on the anticyclonic motions induced by a slow 
expansion motion from the center of a rotating pan. 
The measured quantity, the slope of the free surface, 
showed good agreement with Sttimke’s theory of the 
phenomenon. 
Experiments Involving Discontinuities—Cold Fronts 
and Cyclone Systems 
The study of discontinuous fluid motions begun by 
Helmholtz in the 1850’s and illustrated in the demon- 
strations of Oberbeck [47] in 1877 has had a special 
importance in meteorology since it culminated in the 
LABORATORY INVESTIGATIONS 
work of Margules and the polar-front studies by the 
Norwegians. Here also the climate of opinion on the 
meteorological side was influenced, though probably 
not essentially suggested, by a number of experimental 
investigations. Those of special interest to us fall into 
two groups: an early one which stemmed from studies 
of squall lines (Béen) and later, explicitly, of the cold 
front; and the other one, much later (after the idea of 
cyclone families along the polar front had been con- 
structed by V. and J. Bjerknes), which was concerned 
with systems of vortices developing along similar dis- 
continuity surfaces. 
In the first group we have the cold-front propagation 
studies with liquids by Schmidt [56, 57, 92] around 1910 
and by Ghatage [27] in 1936, the interna’ wave studies 
in 1923 by Defant [14] which included some internal 
hydraulic jumps, the more recent work on jets by 
Spilhaus [60], and the demonstration possibilities im- 
dicated by Haynes [382] and Dinkelacker [76]. These 
studies have had some historical significance in the de- 
velopment of ideas on cold-front and squall-line phe- 
nomena. For example, the nose of the front (Béenkopf) 
(Fig. 4) as discussed by, say, Koschmieder [88], seems 
F 
Fic. 4.—Cold-front forms produced by Schmidt in a glass- 
walled trough 181 cm long, 31 em high, and 4 cm wide by allow- 
ing a salt solution to flow into clear water. The density ditf- 
ferences increase toward the bottom of the figure. The pictures 
were obtained by exposing sensitized paper behind the tank to 
a magnesium flash. (After Schmidt [56].) 
to have been suggested primarily by the experimental 
work. Not very much has actually been done, except by 
indirect reasoning, to verify observationally the occur- 
rence of this characteristic. (See, however, Weickmann 
[94].) Similarly the velocity formulas for the front eval- 
uated from the work of Schmidt, Ghatage, and the 
more recent work of Yih [72] have not been carefully 
