OF ARTS AND SCIENCES. 165 



bustion beyond the limit we have named made no difference in our 

 results, as was repeatedly shown. Again, in one instance, after de- 

 taching the condensation tubes and weighing them, they were again 

 put in place and the combustion continued three hours longer, during 

 which time the tubes gained no appreciable weight. In another in- 

 stance, when a suspicion arose that possibly some hydrogen might be 

 occluded on the walls of the globe, the condensation tubes having been 

 dismounted and weighed as before, the globe was also dismounted and 

 heated over the free flame of a Bunsen lamp to as high a temperature 

 as the glass would safely bear, over 300° C, and then, the apparatus 

 having been remounted, the combustion was continued for one hour. 

 Here again the condensation tubes gained only a small fraction of 

 a milligram in weight, an effect which might easily be accidental, 

 and which was wholly without influence on the result. 



Apparatus for preparing Hydrogen. 



In the preliminary determinations, the hydrogen was drawn from 

 a large copper generator charged with zinc and dilute sulphuric acid. 

 The zinc and sulphuric acid were wholly free from arsenic, and of the 

 best quality, but not absolutely pure; and the writer depended upon 

 an elaborate system of washers and driers for purifying and drying 

 the gas. He found the greatest difficulty in removing the last traces 

 of sulphurous oxide, which hydrogen prepared in this way always car- 

 ries. The presence of this trace cannot be detected by litmus paper, 

 but is immediately indicated by the production of hydric sulphide when 

 the gas is passed over heated platinum sponge ; and by interposing a 

 tube containing platinum sponge maintained at a low red heat, fol- 

 lowed by a set of potash bulbs, this impurity can be entirely removed. 

 It can also be removed by washing with a strong solution of potassic 

 hydrate alone, if the gas remains long enough in contact with the solu- 

 tion. It was found, however, that a series of potash bulbs was insuffi- 

 cient for this purpose ; but two of the long washers represented in the 

 background of Fig. 5 and in Figs. 7 and 8 (Plate), where the gas in 

 small bubbles travels up a tube 5| feet long, are sufficient to remove 

 the sulphurous oxide from even a quite rapid current of hydrogen gas. 

 A series of preliminary determinations was made with hydrogen gas 

 thus prepared and purified, and it was obvious from an inspection of 

 the results, as well as from the difficulties which were experienced in 

 keeping all the joints of this complicated apparatus tight, that the 

 irregularities arose from the diffusion of the air into the hydrogen at 

 some one or other of these joints. It was therefore next sought to 



