Notices respecting New Books. 1&5 



difference in the two curves. The original curve by this 

 adjustment of its coordinates is projected into a curve which 

 touches, or nearly touches, the line of fig. 3. After that, the 

 former falls below, and diverges widely from, the latter, until 

 the ordinates of the former are less than two-thirds of the 

 corresponding ordinates of fig. 3. To illustrate this difference, 

 it may be pointed out that the values of W (the number of 

 molecules of water formed for each pair of ions), as deter- 

 mined by the original tables ranged from 9*7 to 4*1. The 

 tables of this paper, although the conditions have been greatly 

 varied, reduce these limits and confine W between 9*4 and 6*1. 



I attribute this discrepancy to the nature of one of the 

 electrodes used in the first investigation. A large circular 

 hole had been turned out of the lower plate to permit the 

 passage of the ultra-violet light, as I explained ; and the cavity- 

 was covered by a coarse wire network of copper, which was 

 certainly far from being a plane surface. The silver electrodes 

 employed in the present research were carefully turned plane 

 surfaces. It is true that the lower of the zinc electrodes had 

 a series of narrow slits let into it for experiments with ultra- 

 violet light, but the results given above prove that these slits, 

 as was to be expected, produced no perceptible disturbance in 

 the observations. 



The discrepancy between the curves cannot be due to the 

 difference in the mode of generating the oases in the two 

 cases. For though caustic potash was used in the first 

 investigation as the electrolyte, yet, as explained above, 

 Table VIII. shows that it is indifferent whether barium 

 hydrate or caustic potash is used for the purpose of these 

 experiments. 



I must express my thanks to Professor Townsend for his 

 valuable suggestions and criticisms. I also wish to thank 

 Mr. H. B. Baker for kindly supplying me with very pure 

 barium hydrate and for some valuable advice. 



XVIII. Notices respecting JSeic Books. 



The Collected Mathematical Papers of James Joseph Sylvester. 

 Vol. I. Edited by H. P. Baker. Cambridge Universitv Press. 

 1904. 



HPHJS is the beginning of the long expected and patiently waited 

 J- for publication of the collected works of Sylvester. Il i- a 

 finely printed volume of 650 pages, and covers the first sixteen 

 years of his mathematical activity (1 8:37 to 1853). The papers are 



arranged chronologically, and are in this respect very instructive as 



