EXPERIMENTAL CLOUD FORMATION 
culations produced in unstable layers. In this modi- 
fication the top of the chamber was a long glass plate, 
which could be moved across at any desired rate, 
being driven by a small electric motor. The top plate 
produces a shear, analogous to a wind varying with 
height. 
It was found that unless the excess of temperature 
at the base over that at the top was above the limiting 
value found in the investigation with the top at rest, 
no structure was observable in the smoke, no matter 
what the rate of shear. In all the experiments with 
shear, care was therefore taken to ensure that the 
vertical gradient of temperature exceeded the limit 
appropriate to the depth of chamber. 
In the first experiments made in these conditions, 
the shear was given a high value. Figure 8 represents 
Fig. 8.—The effect of shear, by motion from left of right of 
the top plate, in a chamber of depth 8 mm, with temperature 
difference of 58C. 
the flow pattern in a layer 8 mm in depth, with a 
temperature difference of 58C, and with the top plate 
moving at arate of 10 em sec. The chamber is filled by 
a set of double rolls, with descent at the common 
boundary of the two rolls in a pair, and ascent at the 
Fig. 9.—Transverse rolls formed by small shearing motion, 
with temperature difference of 90C and rate of motion of top 
plate of 9 cm sec. 
other boundary. The central boundary of a pair, at 
which there is descent, is not so clearly defined as the 
boundary which separates adjacent pairs. This is clearly 
1259 
shown in Fig. 8. Rolls which are directed along the 
direction of shear are called “longitudinal” rolls. 
When the rate of motion of the top plate is decreased 
to a considerably lower value, the longitudinal rolls 
are replaced by “transverse” rolls, or rolls perpendicular 
to the direction of shear. Careful observations were 
made to investigate whether these rolls had alternating 
sense of rotation, and it was established that adjacent 
transverse rolls do undoubtedly rotate in opposite di- 
rections. Figure 9 shows a system of transverse rolls. 
With a still smaller rate of shear, and with the 
vertical gradient of temperature held constant, it was 
found that no rolls, either longitudinal or transverse, 
were formed, but that the original polygonal structure 
with no shear was distorted, as shown in Fig. 10. The 
Fre. 10.—Distortion of the cell pattern by small shear. 
observations made by Chandra [4] and by Dassanayake 
[5] indicate that the maximum rate of shear which 
will produce transverse rolls increases as the vertical 
gradient of temperature increases. Dassanayake made 
a special effort to maintain a constant vertical tem- 
perature difference of 40C with a chamber 10 mm 
deep, and to approach as nearly as possible to the 
limiting shear above which the rolls are longitudinal, 
and below which the rolls are transverse. He succeeded 
in obtaining photographs showing both transverse and 
longitudinal rolls, occurring simultaneously, and super- 
posed over a part of the chamber, when the top plate 
was moved at a rate of 0.5 em sec~. The photographs 
were, however, not suitable for reproduction, being 
lacking in marked contrast. 
With a difference of 40C in temperatures at top 
and bottom of the 10-mm chamber, motion of the 
top plate will give transverse or longitudinal rolls, 
according as the rate of motion of the plate is less than, 
or greater than, 0.5 em sec. The mechanical arrange- 
ment used by Chandra for moving the top plate was 
insufficiently reliable to permit of accurate control of 
the shear, and his observations were limited to high 
values of the temperature difference and high rates 
of shear. We can therefore analyse the phenomena, in 
chambers of depths of 7 mm or more, as follows, for a 
given difference of temperature between top and bot- 
tom of the plate: 
1. When there is no shearing motion, the chamber 
