RLECTRICIIY. 153 



of electrical currents, their direction and strength. 

 These will only need the arrangement already described 

 under the head Galvanometer. For other experiments, 

 such as that of the electric light, there will be needed 

 some one of the many fixtures for holding the carbons 

 to be ignited. If this can be put into a lantern the 

 carbons may be projected at once upon the screen by 

 removing the objective and drawing the carbons back 

 until the image appears plainly upon the screen. This 

 image will be made much sharper by putting a dia- 

 phragm with about an inch aperture over the conden- 

 sers, which in this case serves for an objective. 



For the projection of spectra precisely the same 

 conditions need to be observed as for the lime lisht: — 

 Some regulator in the lantern, a slit in the focus of the 

 condensers, an objective to project the sHt and the 

 prism in the focus in front of the objective. The spec- 

 trum of metals is easy with this arrangement. Make 

 a small cavity 'n the end of the lower carbon stick, 

 a?id place a small bit of the metal whose spectrum is 

 wanted in it ; then bring down the upper carbon upon 

 it so as to complete the circuit and then raise it a little, 

 the metal will be at once fused and volatilized, emit- 

 ting its characteristic light, which will appear upon the 

 screen as bright bands. Silver, copper, zinc, iron, and 

 mercury give good spectra among the more common 

 elements. 



For the successful working of this method of spec- 

 trum analysis, not less than forty cells will be needed, 

 and fifty are decidedly better than forty. 



DECOMPOSITION OF WATER, 



This is effected by sending a current of electricity 

 from three or four cells through water that has been 



