1889-90.] Mr C. Michie Smith on Surface Tension. 117 
photographed. The distance between the ripple-crests is easily 
measured in such a photograph, but there is some risk of not 
measuring exactly at right angles to the lines of crests. 
The method of exciting the ripples was found to be of great 
importance. In the first experiments the driving-fork of a Helm- 
holtz vowel-sound apparatus was taken, and the dipper attached to 
one of the prongs was replaced by a light arm of aluminium, bent 
downwards at right angles near the end. The point of this was 
smoothed and tapered, and was allowed to dip slightly into the 
liquid. When the fork vibrated, the point set up a series of 
circular ripples which were fairly steady and uniform. A number 
of photographs were taken with this arrangement, but on measuring 
them it was found that they gave for mercury a wave-length of 
only 0T04 cm. for 256 vibrations per second, corresponding with 
T = 159. The value for T given by Quincke is 540, and the 
difference was evidently far too great to be accounted for by errors 
of observation. During these experiments it was noticed that very 
perfect ripples were produced when a vessel containing mercury 
was laid on the sounding-box of a tuning-fork, and a large series of 
measurements were made in this way. Using the same fork of 256 
vibrations per second the photographs gave A = 0T08 cm., which 
differed but little from the previous results. Experiments were then 
made to determine whether the shape, size, or material of the dish 
had any influence on the length of the ripples, but the slight varia- 
tions observed in the value of A were only such as might easily be 
explained by differences in the temperature or in the cleanness of 
the surface. The tuning-fork was then changed for one giving 512 
vibrations per second, but this, too, gave approximately the same 
value for T. Finally, a fork giving 768 vibrations per second was 
used, and it was found that with a round glass vessel a result was 
obtained which differed widely from previous results. On sub- 
stituting a rectangular ebonite dish for the glass one it was found 
to be impossible to get ripples distinct enough to be photographed 
by simply laying the dish on the sounding-board, and consequently 
experiments were tried with the edge of the dish pressing against 
one prong of the tuning-fork low down. This proved most effective, 
and the first photograph gave A = 0075 corresponding to T = 528 
— a result differing only about 2 per cent, from Quincke’s mean value. 
