on a Permanent Electric Current. 179 
then removed by dissolving in nitric acid, after which the 
glass was again dried and weighed. In addition to this the so- 
lution of silver was filtered and treated with hydrochloric acid. 
The precipitate was filtered off, and the silver reduced by burn- 
ing with the filter paper. The amount of silver on the glass 
was thus estimated in two ways. According to the weight lost 
by the plate the amount of silver appeared to be 4:3 mgr., while 
the amount obtained by the chemical process was only about 
25 mgr. There are good reasons for thinking the former 
value too great and some reasons for thinking the latter too 
small. Giving the latter double weight in taking the mean we 
set tot2X2'5 
fe) 
=8'1 mer. for the amount of silver in the film. 
The area covered by this on the glass was about 20 sq. cm. 
Taking the specific gravity of silver to be 10°5, we get for the 
thickness of the film 
0031 
= ————_- = 20000148 cm. 
20 & 10°5 
t 
_ This value is more than 3°6 times as large as that obtained 
by the resistance method. In order to make perfect accord be- 
tween the results obtained with the two kinds of silver plates, 
the thickness would need to be rather more than four times as 
great as that obtained by the resistance method, but consider- 
ing all the difficulties of the case, it seems to me that the large 
discrepancy still existing is within the limits of experimental 
error. In presenting the results of all the experiments in tab- 
ular form further on I shall give the results obtained with these 
silver films as calculated on the basis of the larger value, i. e., 
0000148 cm., found for the thickness. 
Mention is made above of a fourth plate of silver. This 
was also of a very thin film, but the silver was fastened to the 
glass with shellac instead of being deposited from a solution. 
The silver was in the same state as that of the thickest plate, 
and the results of measurements with it accord sufficiently well 
with those obtained with that plate. As the resistance method 
was employed in estimating the thickness, it does not seem 
” 
P 
worth while to publish the results obtained. 
Tron. 
_ Measurements have been made with three separate plates of 
iron. The first two plates were made early in the research and 
the quantitative results, like all others obtained at that time, are 
hardly reliable enough to be worth publishing. 
The dimensions of the third strip were as follows: 
