THE EARTH’S MAGNETIC FIELD IN INTERNATIONAL UNITS. 
449 
when the cells were simply kept in a water-bath standing in the room, and no 
precautions were taken to keep the temperature of the bath constant. Since, 
however, the temperature of the room seldom differed from between 17° and 20°, 
no rapid changes in temperature took place. 
The difference between the agreement of the two kinds of cell being so marked, 
it was decided to use the cadmium cells to the exclusion of the Clark cells. The 
cadmium cell has the further advantage that its temperature coefficient is only 
about one-thirtieth of that of the Clark cell, and that there is much less lag when 
the temperature is changing. Again, it was found that although a cadmium cell 
was accidentally short-circuited, its E.M.F. regained the normal value in the course 
of a few hours. The temperature coefficient of the cells was measured between 16° 
and 26°, and the value obtained agreed with that given by Jagee, and Wachsmuth.’^' 
It has been shown by these observers that, at temperatures below 15°, the cadmium 
cell is unreliable, hence in aU the observations the temperature of the cells was kept 
above this temperature. In the actual observations, the cells were placed in a water 
or oil bath, which was kept stirred, and the temperature was noted at each reading. 
The Silver Depositions. — In a paper! by the author and the late Mr. J. W. Rodger, 
it was first jDomted out, that, if a solution of silver nitrate is electrolysed for some 
time, a silver anode being used, the weight of silver deposited by the passage of 
unit quantity of electricity gradually increases, this increase amounting in time to 
about 1 part in 1,000. It was, however, found that freshly prepared solutions of 
silver nitrate gave the same weight of deposit, whether the salt used was simply 
“ pure re-crystallized,” as obtained from Messrs. Hopkins and Williams, or whether 
this salt had been frequently re-crystallized or even fused. These observations have 
been confirmed by subsequent observers, and KahleJ has published the results of a 
most extended and careful set of observations carried out at the Reichsanstalt, on 
the behaviour of the silver voltameter. He is not, however, able to explain the 
change which takes place when the solution has been used. 
More recently, Richards, Collins, and Heimrod§, have attacked the problem 
and shown that the increase in the weight of the deposit is due to the formation, 
during electrolysis, of some substance in the neighbourhood of the anode, and 
that it is this substance which, diffusing through the liquid, reaches the kathode 
and causes the deposit of the additional silver. These observers seem to adopt 
the explanation tentatively given by Rodger and Watson, that the effect is due 
to the formation of silver ions having a larger electro-chemical equivalent than 
the ordinary silver ions, that is to the formation of complex silver ions. 
In order to prevent the diffusion of these anomalous ions to the neighbourhood of 
* Loc. cit. 
t ‘Phil. Trans.,’ A, vol. 186, 631, 1895, H. 
X ‘ Wied. Ann.,’ vol. 67, 1, 1899. 
§ ‘ Proc. Amer. Acad, of Arts and Science,’ vol. 35, 123, 1899 ; also ‘ Z., Phys. Chem.,’ vol. 32, 321, 1900. 
VOL. CXCVIII.—A. 3 M 
