270 



Dr. C. R. A. Wright on the Determination of 



Smaller plates. 



Larger plates. 



0, D, 



E,. 



0, 



L\. 



B a . 



•001738 -0006952 



•869 



•00385 



•000770 



•770 



•000960 



•0003840 



•960 



•001854 



•0003708 



•927 



•000509 



•0002036 



1-018 



•000999 



•0001998 



•999 



•0002112 



•00008448 



1-056 



•0005195 



•0001039 



1-039 



•0001075 



•00004300 



1075 



•0002136 | -00004262 



1-068 



•0000543 



•00002172 



1085 



•0001085 j -00002170 



1085 



•0000273 



•00001092 



1-092 



•0000546 -00001092 



1092 



•00001057 -00000423 



1097 



•0000262 ! -00000524 



1-097 



•00000533 -00000213 



1-100 



•0000108 -00000216 



1-100 



•00000215 -00000086 



1-103 



•00000535 



•00000107 



1103 



00000107 -00000043 



1-103 



■00000215 



•00000043 



1-103 



•00000052 -00000021 



1-103 



•00000105 



•00000021 



1-103 



104. From these figures the following values for R are de- 

 duced — the first four determinations only of each series being 

 employed, on account of the smallness of the differences be- 

 tween Ei and E 2 in the other cases: — 



of Clark's valuation (Proc. Roy. Soc. xx. p. 444), but also on the exactness 

 with which the B. A. unit of resistance is determined. If this latter be too 

 small — as appears probable from the experiments of Joule, and of the 

 writer and Mr. Rennie (Phil. Mag. March 1881, p. 169), from the results 

 of Rowland, and from the recent experiments of Lord Rayleigh and Prof. 

 Schuster — the true value of an average Clark's cell is below 1 '457 to 

 the same proportionate extent : thus, if the B.A. imit be really 0'99 

 earth-quadrant the E M .F. of an average Clark's cell is onlv 0-99 x 1-457 



second 

 Xl0 8 =l -442 x 10 3 C.G.S. units. In view, however, of the fact that the 

 question of the amount and even of the direction of the error (if any ) in 

 the B.A. resistance unit is not yet absolutely settled, it is assumed in this 

 paper that there is no error at all. 



For analogous reasons the value of J is assumed, as previously, to be 

 42xl0 6 ergs, the evidence in support of its having a higher value still 

 being not inconsiderable ; although the probability is that, if the B.A. 



resistance unit be only 0-99 — J"-3^_ > J is close to 41-o x 10°. The 



second 

 value of x (the electrochemical constant denned in § 7) deduced in § 9 as 

 the most probable, viz. -000105, is also adhered to, notwithstanding that 

 Mascart's recent experiments (Comptes Rendus, xciii. p. 50) tend to indi- 

 cate that this value is too large, -0001044 representing his final result : this 

 value is 0-8 per cent, lower than -00010527, the mean value deduced from 

 Kohlrausch"s experiments. If J be taken = 4L5x 10 ,; , and x he assumed 

 = •0001048 (mean of Mascart and Kohlrausch's results), the value of ^.1, 

 the factor for reducing gram degrees to C.G.S. E.M.F. units, becomes 

 4349, or 1-4 per cent, less than 4410, the value hitherto assumed, and still 

 retained in this paper. 



Prof. S. P. Thompson applies the term "Faraday coefficient'" to the 

 numerical value x (Journ. Soc. Arts, xxx. p. 34 | ; should this term be 

 generally accepted, the letter F might gracefully be used instead of x l ° 

 indicate the factor, just as J is used to indicate the Joule coefficient. 



