THE ELECTRICAL RESISTANCE OF THIN LIQUID FILMS. 
469 
method. Results were sometimes obtained agreeing to about 0'5 per cent., but 
generally the difference was greater, and sometimes it was as large as 6 per cent. The 
fact that the most discordant results were obtained with films which had exhibited 
a considerable number of black specks at the top shows the advantage of working 
in the middle of the film. We may here remark that the effect of these black specks 
upon the electrometer was carefully noted. They generally travelled round the film 
at varying speeds, and it was found that a speck much larger than was usual was 
required to produce any effect on the electrometer by its passage over the nearest 
needle. 
Observations with three Electrometers. —It may be desirable to state why three 
needles were used for the films rather than two. The object in view was twofold. In 
the first place, when a film begins to thin, it often happens that its upper part, the 
region between needles 1 and 2, shows recognizable colours, and is in a fit condition 
for observation and measurement, before the colours of the region between needles 
2 and 3 become well defined, and thus time is gained by attacking the upper part of 
the film while the lower part still remains thick. Again, it happened not unfrequently 
that during the observation of the region between needles 1 and 3, a thickening of the 
film would occur from below, the colours rising steadily from the lower cup and being 
separated by a sharp fine of (apparent) discontinuity from the portion—often perfectly 
uniform in tint—towards which they were advancing. Such a “ thickening from 
below” was always carefully watched, and as soon as it approached needle 3 the 
electrical connexions were readjusted and the examination transferred to the region 
between needles 1 and 2. 
In the second place, it appeared very important to be able to prove by direct 
experiment that the electrometer method of measuring the conductivity of a film was 
to be relied on, and that the results obtained were not dependent upon the conditions 
of the instrument employed. The galvanometer method of testing the electrometer 
measures was not, as has already been stated, always applicable. It was therefore 
determined to measure the resistances of the film between needles 1-2, 2-3, and 1-3 
respectively by three independent instruments. We were fortunate in securing, 
through the kindness of Professors W. G. Adams and F. Guthrie, the loan of two 
quadrant electrometers similar to the one already in use. The electrometers were 
arranged as shown in fig. 7. 
The first operation was to compare the electrometers with each other, which was 
effected in the following manner :—A simple circuit was formed, traversed by a current 
of constant strength, the resistance in it being about 15,000 ohms. Two points in the 
circuit, separated by resistances gradually increasing from 1,000 to 12,000 ohms, were 
connected in succession with the electrometers, and the deflections, right and left, noted. 
The numbers so obtained enabled comparison curves to be drawn, by which deflections 
of electrometers (2) and (3) could be referred to scale readings of electrometer (1). 
The comparison of the electrometers was repeated each day while the experiments 
3 P 2 
