300 EXPERIMENTAL INVESTIGATIONS. 



the narrow platinum plates. The tubes will of course be full 

 of water, and the main channel full of atmospheric air ; this 

 will gradually be displaced by the hydrogen ascending from 

 the zinc, which hydrogen, in consequence of the curve at A, 

 will retain its position. When it is judged that the greater 

 portion of air has been expelled, the stopper at A, covered 

 with a little grease, is to be inserted ; the hydrogen now will 

 rapidly descend in all the tubes until the zinc is laid bare, and 

 then remain stationary. 



We have now a gas battery the terminal wires of which 

 will give the usual voltaic effects, the atmospheric air supply- 

 ing an inexhaustible source of oxygen, and the hydrogen being 

 renewed as required by the liquid rising to touch the zinc ; by 

 supplying a fresh piece of zinc when necessary, it thus becomes 

 a self-charging battery, which will give a continuous current ; 

 no new plates are ever needed, the electrolyte is never satu- 

 rated, and requires no renewal except the trifling loss from 

 evaporation, which indeed is lessened, if the battery be in 

 action, by the newly-composed water. There is an aperture 

 in the pedestal with a moveable slide, through which the 

 vessel B' can be removed, when necessary, to replace the zinc, 

 and the remaining part of the apparatus is never disturbed. 

 This battery would form an elegant substitute for the water 

 battery ; it would much exceed in intensity a similar number 

 of series of that apparatus ; it would be applicable to experi- 

 ments of slow crystallisation, and possibly to the telegraph. 

 Its construction is difficult, and makes its prime cost expen- 

 sive, but after that it is the most durable, the most easily 

 charged, and the most free from local action of any known 

 form. I have had one of ten cells constructed, shown at fig. 5 

 which succeeds perfectly, giving sparks, decomposing water, 

 &c., and is ever ready for use. Any number of such sets 

 might be united by adapter-tubes, or indeed it would be much 

 more economical, and reduce to a minimum the damage from 

 breakage, to have the main channels A A' made of varnished 

 wood or porcelain, with apertures into which separate glass 

 tubes might be cemented. 



