LSHBA.RBR] PRODUCTION OF HYDROGEN PEROXIDE 173 



gave similar results. The great dependence of the action upon tlie 

 amount of dissolved oxygen present is already shown by the following 

 test: — Some foil was cleaned and placed in boiling distilled wator, then 

 after boiling for some minutes, to expel all air, the heat was removed 

 and the flask tightly corked. The water was allowed to cool and shaken 

 frequently, but when opened and tested, no HjO^ could be detected. The 

 same foil, however, when placed in aerated water developed the peroxide 

 in a very few minutes. The greatest yield was obtained when the water 

 was cooled by packing the flask in cracked ice, and at the same time 

 passing an electric current through the water from a mass of aluminum 

 scrap to a small aluminum wire, which served as cathode. A continuous 

 stream of air was, at the same time, passed through the water. 



It is stated that zinc and iron also form the peroxide with water. 

 In order to test this, zinc in the form of a powder was treated in the 

 same manner as the aluminum, but no peroxide could be detected. 



The iron treated was in the form of fresh turnings, and although a 

 great deal of rust was formed, no peroxide could be detected. 



A second trial with zinc, in a coarse granulated form, gave a 

 marked peroxide reaction about equal, if not stronger, than in the case 

 of the aluminum. 



If the water was allowed to stand in contact witli the zinc, no air 

 being bubbled through, the peroxide was in every case completely de- 

 composed after , 7 or 8 hours, but in the case of the aluminum, this 

 decomposition took place only after several days. A more rapid action 

 of decomposition with finely divided substances prolmbly explains the 

 failure to gjet the test in the case of the trial with zinc powder. 



Other metals as copper, platinum and magnesium, were tried, but 

 the only one which gave a yield strong enough to be detected by the 

 starch KI reaction, was magnesium. Several other tests for the detec- 

 tion of the peroxide were tried, but the only one which was at all suc- 

 cessful was that known as Bach's. This is as follows : — Add 2 or 3 

 drops of 5^ oxalic acid to equal parts of the solution to be tested and 

 a .003^ solution of potassium bichromate, which contains 6 drops of pure 

 anilin per iitre. A pale pink colour denotes the peroxide 



Quantitative tests for the approximate strength of the yield of 

 H0O2 from the different metals, were carried out as follows : — a solution 

 of ordinary of/ ].(n-0Xide was diluted to 1 part in 2,000 and this was 

 again diluted to 1 in 30,000, 60,000, etc., up to 1 in 000,000 and ^■o]ution> 

 of each obtained. The starch KI test was applied to each of these, the 

 result being ]al)c]ed and kept for reference. The test was then applied 

 to the sample io be estimated and compared with those obtained from 

 the solutions of known strength. Considerable difficulty was experienced, 



