171 



took place, the residue should either l)e alkaline or coutaln potas- 

 sium inanganate. or permanganate, and that he could obtain no evidence 

 that either was the case. He brought forward experimental evidence to 

 prove that the evolved gases did not contain more than a mere trace of 

 chlorine and artirmed his belief that the odor and the property of Iduing 

 starch and potassium iodide paper was due to ozone. In 1894 McLeod 

 stated that when the gases Avere led through alkaline silver niti'ate solu- 

 tion, and this later acidihed. a precipitate was obtained which corresponded 

 in quantifj- with the alkalinity of the residue in the generator. He could 

 obtain no evidence of ozone. Some further work was done by these men 

 but they did not apparently succeed in settling the point at issue. 



Sodeau. in llXil. proved that the action of manganese dioxide, barium 

 sulphate, sand, and other supposedly inert l)odies increased the evolution 

 of oxygen not mechanically, Imt rheniically. 



EXPERIMENTAL. ' 



The apparatus used in the experimental part of the present investiga- 

 tion was very simple. Hard glass test-tubes tive inches in length, with 

 side necks, were used for heating the mixtures, these being placed in a 

 bath of Wood's fusible metal, heatetl in a thick cast-iron cup large enough 

 to accommodate tive tubes. A thermometer was also placed in the metal. 

 Short delivery tubes, with ends drawn to a narrow aperture, letl to a ves- 

 sel for collecting the evolved gases in test-tubes over water. 



The manganese dioxide used was Merck's "Artificial Pure," and pre- 

 vious to using was heated for several hours in an open dish over a free 

 flame, in order to remove moisture: it was then placed in a glass stoppered 

 bottle for keeping. Elmer and Amend's potassium chlorate was dried 

 for six hours at 105°-110° for this purpose. It was not labeled "C. P." 

 but tested free from chlorides lioth before and after drying. 



The first mixtures were made in tlie following molecular ratios of 

 manganese dioxide to potassium chlorate: 10:1. 2:1. 1:1. 1:2, 1:10. These 

 were ground together, placed in the tubes, and slowly heated. At 150°- 

 1G5° a gas was evolved from all, showing the presence of oxygen by means 

 of a glowing spark, and giving a strong odor of chlorine or chlorine oxide. 

 This odor is certainly not that of ozone and may be either chlorine or 

 chlorine oxide, or both. In this paper it will be provisionally called 

 chlorine. It was noticed that considerable moisture collected upon the 



