1889 - 90 .] Mr C. Michie Smith on Surface Tension. 
119 
seen distinctly only in one particular direction, and care must be 
taken to place the axis of the camera in that direction. If the 
camera could he placed vertically over the dish it would greatly 
simplify the measurement of the plates, since the scale- value would 
be the same in all directions. This, however, is seldom practicable, 
and usually the axis of the camera has to be inclined to the vertical 
at a considerable angle, so that the scale-value varies not only for 
different parts of the plate, but also for different directions. The 
corrections to be applied cannot easily be calculated, but they can 
be got by actual measurement. For this purpose a sheet of paper 
was taken of the size of the dish, and on it a number of circles were 
described with a radius of 1 cm. This was then laid on the surface 
of the mercury and photographed. The scale value for any part of 
the mercury surface in any direction can be obtained by measuring 
the diameter of the corresponding circle in the same direction. 
The results so far obtained can only be looked on as preliminary, 
but they are sufficient to test the value of the method. At present 
the value of the wave-lengths is probably not trustworthy to within 
less than 2 per cent., but there seems no reason to doubt that this un- 
certainty can be reduced to at least one-half by the introduction of 
certain modifications in the apparatus. An error of 2 per cent, in the 
wave-length corresponds to an error of approximately 6 per cent, in 
the value of T, and the extreme values obtained differ from each other 
by somewhat less than this in the case of mercury. Quincke’s 
results for mercury* which are usually accepted, vary from 511 to 
572 dynes per centimetre, with a mean value of 540; so that his 
extreme values differ from each other by about 1 2 per cent. 
In his paper read at last meeting of the Society, Professor Tait 
showed that if we take account of the surface-flexural rigidity (E), 
the equation for the wave-length takes a form which can be 
written : — 
T _9P± 2 ■ 4ff 2 E 
2t rt 2 4 tt 2 X 2 ' 
4?r 2 
With small values of A. the factor ^ — becomes very large, and 
unless E is excessively small it ought to be possible to detect its 
influence by comparing the results got for two forks of very different 
Pogg. Annal cv. p. 1. 
