1242 
latitudinal distribution of the zonal motion, under some 
heating conditions, shows considerable resemblance to 
atmospheric zonal wind profiles at moderately high 
levels. This and other resemblances mentioned in the 
LIMITS OF 
WINTER LATITUDE 7 TSSERMED Portis 
CROSS SECTION Ory \ oe 
AT 12 KM 
(AFTER WILLETT) 10° \ jf 
i 
a i 
MEAN FLOW 
X- 
7 28 MAR 1947 INK 
4 APR 1947 PELLETS 
SUMMER 60° 
CROSS SECTION \ INET 5 15 
AT 12 KM re KEs 
(AFTER WILLETT) ] 70 
SS KE, = 160 ERGS 
: 102) 04) 06) 08) M10 
=10 =08 =06 =04 -02 ve 
EASTERLIES Ce 
WESTERLIES 
Fre. 9.—Mean estimated curve of u/C;, in the hemispherical 
shell of Figs. 7 and 8 for 16.27 rpm and 30 v indicated heating 
compared with the mean zonal winds for winter and summer at 
12 km along a cross section (from Willett; see [22]) running 
from Havana to Aklavik, N. W. T. (After Fultz [22].) 
original paper [22] suggest that a study of the mechanics 
of these systems in as great quantitative detail as pos- 
sible will have real significance in suggesting working 
principles which would be applicable to the earth’s 
atmosphere. The opportunity will then exist to establish 
areas in the conditions of the experiment that give the 
closest resemblances to planetary cases. More or less 
data for comparison exist at least for atmospheric 
layers on the earth, the sun, Mars, and Jupiter. With 
the controllability conferred by the experiment, one 
can then compare these areas with conditions which do 
not give similarity and have some hope of separating 
out the way in which the essential factors are operating. 
For example, it would be of the greatest interest, if it 
should turn out to be possible to generate such motions, 
to compare conditions required for producing motions 
divided sharply into zones, like those of Jupiter, with 
the conditions mentioned above that give motions re- 
sembling to some extent those of the earth’s atmosphere. 
In many respects the possibility of making measure- 
ments in an experiment is restricted by purely instru- 
mental or observational difficulties. Any practicable 
experimental size is so small, for instance, that measure- 
ments of any velocity field to a density comparable with 
tive inertia force to a characteristic Coriolis force. In addition, 
as pointed out to me by R. R. Long, it is a factor in a ratio 
of characteristic fluid velocity to the velocity of Rossby long 
waves relative to the basic current. The remaining factor in 
this ratio depends on geometrical quantities only. It would 
thus appear that this parameter measures the approach of a 
given system of currents to several properties peculiar to 
large-scale meteorological motions. For example, as u2/r(Qu), 
it indicates the degree of approximation to quasi-geostrophic 
motions. 
LABORATORY INVESTIGATIONS 
meteorological synoptic measurements is possible only 
by indirect means. In this case the utilization of the 
double refraction property of certain viscous fluids in 
deformation flow is one of the most promising possi- 
bilities [22]. Other devices which we have found to be of 
such great utility as to allow measurements to be made 
which would otherwise be almost impossible to secure 
include such an instrument as the ‘“rotoscope’’ shown 
in Fig. 10. This instrument is modeled on one used by 
ma ia, Bee da 
Fia. 10.—General view of rotoscope setup for observing the 
rotating hemispherical shell apparatus. The hemisphere is at 
the bottom and is viewed in a first-surface mirror (partly con- 
cealed by the rotoscope) so as to obtain a convenient direction 
for the line of sight. Part of the image of the hemisphere can 
be seen in the Dove reversing prism which is rotated in its bar- 
rel by the variable speed motor at right. 
D. Thoma [19, 23] for observing the relative flow in 
centrifugal pumps. Provided observation is carried out 
on a line of sight coinciding with the axis of rotation of 
an object, the image of the object can be reduced to 
apparent rest by rotating the Dove reversing prism 
in the rotoscope at the proper rate. Effectively, there- 
fore, observations can be carried out in a relative co- 
ordinate system rotating at any desired rate. The 
advantages of such a system are difficult to appreciate 
without one’s having had experience with attempts to 
make reasonably precise measurements on a rotating 
object, but these advantages have been demonstrated 
again and again in some of our more recent investiga- 
tions. These experiments have concerned a mechanically 
driven vortex in a two-fluid layer occupying the hemi- 
spherical shell in a slightly modified version of the 
original apparatus. A large amount of data on wave 
velocities and various types of instability points in this 
system has been collected very rapidly by this means. 
Such data could hardly have been obtained so well in 
any other way.® 
Suggestions for Future Research 
The impression one receives in reviewing this long 
series of experimental work is that to make it really 
5. The problem in this investigation was suggested by 
Rossby’s pole-flight result for anticyclones [52]. At the present 
time it seems almost certain that an exactly analogous phe- 
nomenon has been demonstrated in the experiments. The 
results will be presented at length elsewhere [23, 42]. 
