Manchester Memoirs, Vol. xlviii. (1904), No. IX 5 



each series, and I must also observe that the prints are 

 not as accurate representations of the mirrors or their 

 photographs as I should like, but they may be taken 

 generally as approximately accurate as regards the inten- 

 sities of the mirrors. 



It occurred to me that as the presence of platinum in 

 any form with the zinc always interfered with the delicacy 

 of the test in the Marsh-Berzelius apparatus that it should 

 be eliminated, so an apparatus was arranged with a rod of 

 pure zinc as the kathode, and with this apparatus the 

 electrolytic method was found to be as delicate as the 

 Marsh-Berzelius process, and gave results in which the 

 mirrors were rather more evenly deposited, and which has 

 proved by continued experience to be reliable for the 

 detection and approximate estimation of arsenic, whether 

 it exists as arsenic trioxide (AS4O6) or as arsenic pentoxide 

 (AS2O5), and so Lord Kelvin's suggestion has really given 

 the best process for the approximate estimation of arsenic. 



The Series B* Fig. i, Plate VIII., shows the mirrors 

 obtained by the Marsh-Berzelius method, and Series E and 

 F show the mirrors obtained by the quantities indicated at 

 the bottom when using my own zinc kathode electrolytic 

 method, and cooling the part of the tube on which it was 

 desired to deposit the mirrors ; series E was obtained 

 when using a current of 3 amperes, and Series F when 

 using a current of 5 amperes. The former gives the 

 mirrors less spread out than the latter, and I find it more 

 easy to accurately compare mirrors obtained by a 3, than 

 by a 5 ampere current. In some of the photographs of 

 the mirrors obtained by the 3 ampere current there may 

 appear not so much difference between the mirrors from 

 two different quantities of arsenic ; when, however, they 



* All the prints of the tubes are about one half, or more accurately ffths 

 of the diameters of the originals. 



