EXPERIMENTAL CLOUD FORMATION 
By SIR DAVID BRUNT 
Imperial College of Science and Technology 
Historical Note 
The motions which occur in unstable layers of fluid 
in which there is no general motion of the fluid have 
been described independently by a number of writers. 
The “cellular” convection which occurs in unstable 
fluids has been associated with the name of Professor 
Henri Bénard, on account of the thorough study which 
he made of the phenomena which he described in 
detail [1]. The present writer gave the name of ‘““Bénard 
cell” to the typical convection cell which occurs in 
unstable layers initially at rest [3]. Bénard’s observa- 
tions were anticipated about twenty years earlier by 
James Thomson [17] (the brother of Lord Kelvin), who 
observed a “‘tessellated structure” in cooling soapy 
water in a tub casually seen in the yard of an inn. 
Thomson later reproduced the phenomenon in the lab- 
oratory. Weber [18] also observed a structure, similar 
to that described by Thomson, in a layer of alcohol 
,and water on the slide of a microscope, but Weber 
gave a wrong interpretation of his observations. A 
correct interpretation was given later by Lehmann in 
Molekularphysik, Vol. 1, p. 279. Many other subse- 
quent writings by various authors followed the acci- 
dental discovery of convection in a cellular form, but 
only the earliest discoverers need be mentioned at this 
stage. Further brief accounts of independent discoveries 
of the same phenomena in a variety of liquids will be 
found in Nature (April and May, 1914). 
In Fig. 1 are shown the lines of flow in a central 
vertical section of the typical Bénard convection cell. 
When the fluid has a large shearing motion, the axis 
of convection is replaced by a vertical plane of con- 
vection, and the fluid is divided into a series of parallel 
rolls, rotating im opposite directions, and Fig. 1 now 
h 
2b 
Fie. 1.—Cireulation in the Bénard convection cell. 
gives the general features of the circulation, transverse 
to the axis of shear, for a pair of adjacent rolls. 
The first detailed investigation of the effects of shear 
was made by Terada [16], who carried out a series of 
investigations with a variety of liquids in which the rate 
of flow increased with height, but without change of 
direction. Terada’s paper contains a number of striking 
photographs of these longitudinal (downwind) rolls, 
and the author draws attention to the opposite sense 
of rotation of adjacent rolls, stating that the rolls have 
a circular cross section. 
Phillips and Walker [13] showed, by experiments in 
air, that rolls transverse to the direction of shear can 
be formed when the shear is small, and Graham [8] 
a pupil of Walker, later confirmed this result. 
b) 
The Nature of the Simple Convection Cell 
When instability in a fluid initially at rest breaks 
down, the fluid is seen to become divided into a number 
of polygonal cells. In a liquid, the motion normally 
consists of ascent in the centre of the cell, outward 
motion at the top, downward motion at the outer 
margin, and inward motion at the bottom of the cell. 
The simplest illustration of such motion is obtained 
by pouring cheap gold paint into a small vessel to a 
depth of, say, a quarter of an inch. In cheap gold paint 
the liquid medium is usually benzene or some other 
highly volatile liquid, while the “gold” is in the form 
of thin flakes. The evaporation of the liquid leads to 
such rapid cooling at the top of the layer that marked 
instability is immediately produced. As the thin flakes 
tend to set themselves parallel to the motion, they are 
Fie. 2.—Convection cells formed in gold paint. 
not visible at the centre or periphery of the cell, where 
they are seen edgewise from above. The cell will thus 
appear to have clear liquid at the centre and at the 
1255 
