462 Mr. J. Brown on the Role of 



general approximate agreement of the two values points to 

 this as probably the correct method, notably in cases 1, 2, 3, 4, 

 5, 14, 15, 17, 19, 21, 24, and less distinctly in 6, 11, 16, 27, 28. 



In considering the discrepancies between the theoretical 

 and observed values, and the cases where factors besides those 

 mentioned above come in, the following may be noted. The heats 

 of solution of IG1 and BrCl are not available. If positive, these 

 would decrease the theoretical values. As the heat of com- 

 bination for platinum chloride is not available, the theoretical 

 values where platinum forms the negative plate are somewhat 

 too high. 



25. The observed forces of those cells with negative plates of 

 iron and cadmium, 7 and 8, are much higher than the theo- 

 retical. Both these metals coat themselves at first in the 

 strong solution with a yellowish-white deposit, which may act 

 as a protective layer in the same way as the film of oxide on 

 " passive " iron is believed to do. Professor Letts was kind 

 enough to examine this deposit from the iron, and found it 

 to be iron iodate. Zinc also in very strong iodine-chloride 

 solution deposits a little light-coloured powder, and both iron 

 and zinc are less positive in strong than in moderately dilute 

 solution, the differences in both cases being about *4 Daniell. 



In 10, the heat of oxidation of iodine to HI0 3 probably 

 comes in ; but it seems difficult to estimate even approxi- 

 mately the theoretical value for this cell. " Lithanode " 

 (said to be chiefly peroxide of lead) as negative plate instead 

 of platinum also increased the E.M.F. of the IC1 cell by *2 D. ; 

 but manganese dioxide rather decreased it, as did also, to a 

 less extent, other oxidizing agents, such as platinum that had 

 been a positive pole in dilute sulphuric acid, and platinum 

 immersed in nitric acid. 



In 11, if the heat of amalgamation of potassium 2(K, Hg 12 ) 

 (Berthelot) be subtracted, 2*44 results, otherwise 3'8. The 

 true value apparently may lie between these. A similar 

 consideration applies to 27 and 28. 



In the case of magnesium, 9, 10, 25, 26, the observed 

 values are considerably below the theoretical. Wright and 

 Thompson have observed a similar discrepancy in the case of 

 magnesium immersed in its chloride *. 



* Phil. Mag. [5] xix. p. 210. Prof. Herroun suggests (Phil. Mag. [5] 

 xxvii. p. 211) that the metal acts on the water and coats itself with 

 hydrogen. 



If instead of the thermochemical value for Mg, CI we take Mg, 0, 

 H 2 0- H 2 , 0+H 2 , Cl 2 , the theoretical E.M.F. is reduced by about 1 Daniell 

 and approaches the observed values. A similar explanation may perhaps 

 apply in the case of potassium, 11, 27, 28, instead of that proposed above. 



