460 SECTIONAL TRANSACTIONS.—Af. 
Eberhard effect. The microphotometer will introduce further errors, but 
the largest errors of all probably arise in the final process of reducing the 
microphotometer tracings, the difficulties of drawing in the continuous 
background and eliminating blends being often insuperable. With spectro- 
graphs of ordinary dispersion, an error of 20 per cent. is to be expected in 
many cases. This seriously limits the usefulness of such observations in 
deriving values of atmospheric pressures, compositions and opacities, and 
of stellar rotations. 
EVENING. 
Prof. F. Linxe.—Cloud evolution (with cinematograph demonstration, 
8.30). 
Tuesday, September 12. 
Mr. A. GraHAM.—The instability of air layers (10.0). 
Walker and Phillips have shown that the vortex patterns produced, when 
a thermally unstable layer of air is subjected to a double shear, show close 
similarity to certain cloud formations in the sky ; in the sky, however, there 
is ordinarily a single shear. ‘The paper describes some experiments with 
an unstable layer subjected to single shear. 
The upper surface of the experimental channel was a long strip of plate 
glass ; it was drawn steadily over a short, hot iron plate, thus producing 
a single shear in the air layer between them. As in the double shear experi- 
ments straight vortices could be formed, aligned either transversely to or 
along the shear; in addition, there were formed a square pattern having one 
set of diagonals along the shear, and some transitional patterns ; hexagons 
were obtained without shear and also with a certain value of the shear. 
There is a steady change from hexagons into longitudinal rolls through the 
other patterns. All these patterns have their counterpart in the sky. 
An investigation was made into the patterns formed in the absence of shear. 
If the lower surface of an air layer is rapidly warmed, cells having ascending 
air at the centres are produced ; if the upper surface is rapidly cooled, the 
cells have descent in the centres. In the sky cells are formed having ascent 
in the centres and also cells with descent, thus it appears that they should 
be formed under the above conditions. If in the laboratory the temperature 
difference between the top and bottom surfaces of an air layer is large the 
cells have descent in the centres, whereas with a liquid layer the cells have 
ascent ; this phenomenon is apparently due to the fact that the cool upper 
surface of an air layer has a lesser viscosity than the warm lower surface 
and is therefore less stable, while for a liquid it is the reverse. 
The experiments considerably strengthen the theory that many cloud 
patterns are due to thermal instability and not to Helmholtz waves. 
Mr. E. Tititotson.—High focus earthquakes in the International 
Seismological Summary (10.20). 
From June 1914 to March 1928 there were twenty-four so-called shallow 
focus earthquakes, whilst there were about three times as many deep focus 
earthquakes and approximately 11,000 normal tremors published in the 
International Seismological Summary. One of the ‘ shallow ’ focus earth- 
quakes had a focus 0°04 of the earth’s radius above normal, and so it has 
been suggested for this and other reasons that the normal depth of -an 
earthquake is about 160 miles below the surface. More recently, however, 
