CLARK CELL AS A STANDARD OF ELECTROMOTIVE FORCE. 
591 
this semi-solid mass longer to reach the equilibrium condition than is required for the 
liquid in No 2. 
To obtain the observations on December 22, the heating apparatus had been 
turned off and the windows of the room left open on the two previous nights. 
On December 21 there was a thick coating of ice on the bath which continued for 
one or two days; longer observations were made on December 22. Dr. Schuster’s 
cell, No. 97, had been taken to London a few days previously. 
A number of observations on the difference between the Cell No. 90 and the 
standard were made during the absolute determination of the E.M.F. Several of 
these confirm the fact that No. 90 gained its temperature equilibrium more slowly 
than No. 1. Thus, between the morning and evening of August 14, the tempera¬ 
ture of the bath rose by 1°. The difference between No. 90 and No. 1 increased from 
2 to 5 units. 
Other observations again showed that it takes some time for No. 1 to change its 
value. Thus, on August 15, the bath was at 18°‘l, and the air in the test tube 
round No. 1 was about 0 U ‘5 higher; the difference in E.M.F. between Nos. 90 and 1 
was 5. No. 1 was then taken out of the test tube and placed in the bath with 
No. 90. In about three hours the difference had fallen to 3, at which it remained 
for some time. 
Similar effects were noted again during the early part of this year. On January 19 
the room had been cooled down to about 9° C. About a week later the difference 
between No. 90 and No. 1 was 11, the temperature being 13°. The next day the 
difference was 8, and it was noticed that there was a thick compact layer of crystals 
at the bottom of the clear liquid. The cell was slightly shaken so as to mix up the 
sulphate of zinc which was clearly saturated below, but not above. The next day, 
Tuesday, January 26, the cells were together. On January 28, the temperature 
having risen to 15°, the difference had increased to 4 ; this increased to 5 the next 
day, the temperature of both cells being 16°, and some time more elapsed before the 
cells again came together. 
It will be seen that the other cells are fairly near the standard. 
Mr. Elder’s H cells, 3a-5, are slightly too low, and it would appear as if their 
temperature coefficients were different. 
The observation for No. 3, on November 14, shows distinct change, but we cannot 
attach great weight to the one isolated observation. It must be remembered, also, 
that it is difficult to read the temperature with great exactness. 
It would appear as if Mr. Callendar’s cells, Nos. 6-10, had risen in value 
between February and August. This is quite possible, for the insulation was 
originally somewhat defective—the wires from both poles of each cell were fastened 
down to a piece of deal. 
It is probable, however, that their temperature, specially on November 14, is not 
very accurately known. The temperature of the room on that day, as indicated by a 
