THE ELECTRICAL RESISTANCE OF THIN LIQUID FILMS. 
455 
further tested by more than 500 observations made in September and October, 1880. 
It is hardly necessary to state that these revisions were carried out without any 
reference to the electrical experiments with which the results obtained by them were 
to be combined. As, however, the conclusions arrived at in this paper with respect to 
the specific resistance of soap films depend only on the final series of experiments made 
during September and October, 1880, it is satisfactory to note that the table of colours 
in no way depends upon these experiments, but that they are used below only to test 
the results arrived at a year before. 
The following details will serve to show the accuracy attained in the two sets of 
experiments, in the calculation of which the table as finally adjusted was used : — 
Experiments made in September and October, 1879.—The films were examined 
under very varying hygroscopic conditions, and it was thought advisable to see 
whether the value of the quotient was affected by the varying dilution of the films. 
The value of the quotient would have been about 2 per cent, greater for pure water 
than for a standard solution. Twenty films were selected, which gave tolerably 
constant specific resistances, and the means of the ratios given by observations on each 
colour were calculated for each film. The mean of these numbers for the ten films 
which had the lowest specific resistance, and were therefore (as will be shown hereafter) 
the more dilute, was 1'004, for the other ten films it was 0’994. A precisely similar 
result was attained when the quotients were compared film by film instead of colour 
by colour. As the effects of evaporation or absorption seemed, therefore, to appear on 
the means of a large number of observations, and as we had not at that time all the 
data for calculating the theoretical quotient in each case, the mean of the quotients for 
each colour was divided by the mean of all the quotients. These numbers are inserted 
in column YII. of Table IT. 
Experiments made in September and October, 1880.—These experiments being, for 
the reasons already stated, the more valuable as tests of the table of colours, we give 
the following rather fuller details :—510 observations were made, and (omitting 
fractions) in the case of 52, 84, and 95 per cent, of these the difference between 
the quotient and unity was less than ’01, ’02, and ‘03 respectively. Hence in 84 per 
cent, of the observations the mean value adopted differed from both of those from 
which it was derived by less than 1 per cent. The mean value of the mean quotients 
for all the colours was 0'9985. The ratio of the two cosines was the same in all cases, 
viz.: 1'157. Both these numbers were calculated with the refractive index of the 
standard solution. 
In comparing the new table with that of Newton, it is best to take from the latter 
only such colours in the second list as have on each side of them colours which also 
occur in our own list, i.e., to confine the comparison to those cases when no question as 
to the boundaries of the colours can arise. 
In Table II.—Column I. contains the names of the colours. 
Column II. the symbols. 
