214 EEPORT— 1886. 



aud to preserve tliem for long periods without change, in order that 

 the results of series of experiments made at intervals of several days or of 

 weeks shall not be subject to unknown errors. In dealing with oxygen^ 

 the presence of nitrogen must especially be guarded against, for it com-- 

 biues more freely with oxygen (when the latter gas is present in excess) 

 under the influence of the electric discharge than is commonly known to 

 be the case. 



The mercury gasholder invented by Bunsen, which has been 

 described in ' Watts' Dictionary ' and elsewhere, is hardly suitable for 

 collecting gases in the rather large quantities that will be required. A 

 similar but much larger gasholder, in which the mercury was replaced 

 by sulphuric acid, has been tried. But, apart from the risk of air gaining 

 admittance through the sulphuric acid, in which it is to some extent 

 soluble, we find that even thoroughly washed oxygen, prepared from 

 chlorate of potassium, carries with it a sufficient quantity of suspended 

 matter to result in the presence of slight traces of chlorine tetroxide in 

 the gas after the gasholder has been refilled several times. After the 

 failure of this method of storing oxygen, an attempt was made to prepare 

 it by electrolysis of dilute sulphuric acid almost saturated with chromic 

 anhydride. When the superficial area of the negative electrode em- 

 ployed greatly exceeded that of the positive electrode, pure oxygen was 

 obtained in this way. When the evolution of oxygen was conveniently^ 

 rapid, however, some bubbles of hydrogen escaped the oxidising action 

 of the chromic acid and made their appearance. Finally, after the failure 

 of an attempt to store pure oxygen by compressing it in iron bottles, the 

 apparatus next described was constructed for producing the gas in smaller 

 quantities as required. 



In the diagram, A is a cylinder having a capacity of one litre. It can 

 be filled with mercury from a reservoir, not shown, through an india- 

 rubber tube O, the entrance of bubbles of air carried by the mercury being 

 prevented by the air-trap B. E is a flask connected to A by the tube 

 J H. In E is placed the material from which oxygen is to be produced. 

 G contains phosphorus pentoxide * to remove moisture as far as possible 

 from the gas before it is delivered through P into the receiver in which 

 it is to be collected. Beyond G is one of Mr. Cetti's patent vacuum taps. 

 As this will not prevent the passage of air in the direction a to b, however,, 

 it is trapped at C. This trap can be filled with mercury to any desired 

 level from a reservoii-, as shown at R and S. The only joints not made 

 before the blowpipe are those shown at J, H, and F. These are all 

 protected with mercury in the now familiar manner, the india-rubber 

 connections being well lubricated and firmly bound with iron wire. 



The materials from which oxygen is to be prepared having been 

 placed in E, and everything being in order, A is filled with mercury, t^ is 

 closed, t- and fi are opened, and the appai-atus is exhausted through P. 

 Oxygen is then generated in E until the whole apparatus, including A, 

 is filled. This process of exhausting and refilling is repeated at intervals 

 of a few hours two or three times ; and after the third operation a 

 specimen may be collected and examined. Such specimens have been 

 found to be very fairly satisfactory ; two samples of oxygen which had been 

 confined in A for several weeks contained respectively 99'97 and 99"9& 



' We find this substance to be admirably suited for removing suspended solid 

 matter from gases. 



