Chemical Affinity in terms of Electromotive Force. 171 



to the B.A. unit the same ratio as that found by Kohlrausch 

 (§ 36), viz. 1 to 1-0493, the value of the B.A. unit becomes 

 finally 



0*9550 x 1*0493 = 1-0021 earth-quadrant per second, 



exhibiting an error in excess of 0*0021 =0*21 per cent. 



Taking these results into consideration along with those of 

 Kohlrausch and Lorenz, discussed in Part II., the following- 

 values are finally deducible for the Siemens and B.A. resistance 

 units respectively in earth-quadrants per second : — 



Siemens unit. B.A. unit. 



B.A. Committee 0-9530 (a) 1-0000 



Kohlrausch 0*9717 1*0196 



Lorenz 0*9337 0*9797 (b) 



Rowland 0*9446 (a) 0*99112 



H. F. Weber, 1st method . . 0*9550 1*0021 (b) 



2nd „ . . . 0*9545 1*0016 (b) 



3rd „ . . . 0*9554 1-0025 (b) 



the values marked (a) for the Siemens unit being derived 

 from the observed value of the B.A. unit by dividing by 

 1*0493, the ratio found by Kohlrausch, and those marked (b) 

 for the B.A. unit being obtained by multiplying the observed 

 values of the Siemens unit by 1*0493. Mr. Fletcher's preli- 

 minary results as above stated give a value less than 1*0000. 

 Taking all these valuations into consideration, it would seem 

 that, in the existing state of our knowledge, it is just about as 

 probable that the B.A. unit is in error in deficiency as that it 

 is erroneous in excess. Accordingly, in the following experi- 

 ments it is assumed that the B.A. unit represents exactly one 

 earth-quadrant per second, and in consequence that the E.M.F. 

 of a Clark's cell is exactly, as determined by him, 1*457 x 10 8 

 C.G.S. units, or 1*457 theoretical volt. 



56. Admitting this to be so, however, it does not seem to 

 be possible that Joule's water-friction values of J are exact 

 (§ 34) : not only do the determinations of Hirn and Violle 

 concur in indicating values for J of from nearly 2 to 3 per 

 cent, higher than these water-friction values (§ 37), but, 

 further, four different sets of observations also indicate that if 

 the B.A. resistance-unit be correct (i. e. if it = 1 earth-qua- 

 drant per second), or if the Siemens unit be 0*9530 earth- 

 quadrant per second, the value of J must be from 1 to 2 per 

 cent, higher than these values, and, on the whole, upwards of 

 42 megalergs. These observations are, first, those of von 

 Quintus Icilius (Pogg. Ann. ci. p. 65) as corrected by H. F. 

 Weber {loc. cit. supra), which were based on an erroneous 



