364 Messrs. Threlfall and Pollock on the Clark Cell 



current-density of 0*00012 per square centimetre of mercury 

 surface. The more nearly corresponding curves, taking current- 

 density as the criterion of correspondence, are to be found by 

 using one of the Clark-cell curves for 0*001 ampere. In this 

 case the latter is infinitely superior, and is even superior when 

 giving 0*01 ampere with ten times the current-density. The 

 trouble and difficulty involved in setting up a Daniell cell 

 would quite independently render it useless for our purpose, 

 so the comparison was on all accounts pushed no further. 



It will be noticed that the greatest fall of terminal P.D. 

 was observed on August 9th and amounted to 0*006685 volt, 

 being 0*001468 greater than the mean. Now, on July 26th 

 occurred the smallest observed value of the cell at rest, this 

 was 1*435415 volt. Suppose that in practice these causes of 

 error (here observed on different days) conspired, we should 

 have taken a real P.D. of 1-428730 volt for 1*430663 volt. 

 The error thus introduced would have been *002 volt nearly, 

 or an error of about 0*14 per cent. 



Similarly taking the greatest observed terminal P.D. at 

 rest and the smallest fall, the error the other way if these 

 causes conspire amounts to calling a real P.D. of 1*43186 

 only 1*430663, or about 0*08 per cent. 



It remains to discuss the effect to be expected from changes 

 of temperature. The temperature-coefficient of the Clark 

 cell is of course known, but we have as yet given no evidence 

 as to any possible temperature-change of A other than what is 

 to be expected from the known change in the cell itself. To 

 test this matter, on November 1st the large cell was placed in 

 a very large vessel filled with water at a temperature of 

 about 38°C, and left there for a considerable time. The bath 

 was well wrapped up, and during 37 minutes the temperature 

 fell to 34°* 2 C. when tested by a standard thermometer. The 

 fall of temperature being so slow allows us to hope that the 

 temperature of the cell during the experiments did not differ 

 to any considerable extent from that of the water. More- 

 over the E.M.F. of the cell had again become constant and 

 had fallen through 0*00729 volt. By comparison with the 

 40 cells, on October 20th, it was found that the E.M.F. of the 

 master cell was 1*43356 volt at 17° C. By a comparison of 

 the large cell with the master cell at the time of the experi- 

 ment and at a temperature of 21°*5 C, its E.M.F. was found 

 to be 1*42900 volt; consequently at 34°*2 C. the E.M.F. 

 was 1*42171 volt. The results are given in Table III. 



