Spectrum of Aqueous Vapor. 



223 



The source of the electrical energy I employed was twenty 

 thousand storage cells of the Plante type. The direct current 

 from these cells through a liquid resistance was used to pro- 

 duce the white spectrum ; and a glass condenser consisting of 

 300 (three hundred) plates of glass — each plate having a coated 

 surface of 16x20 inches — with a total capacity of about 1*8 

 microfarads — was charged by the cells to produce the four- 

 line spectrum. 



I had great difficulty at first in obtaining tubes which would 

 stand such powerful discharges. I began my work with end- 

 on tubes which were closed by plates of quartz luted on by 

 silicate of soda. The electrodes were hollow cylinders of 

 aluminum connected to thick pieces of platinum wires. These 



wires passed through the walls of the glass tubes, and were 

 immersed in large vessels of distilled mercury. After con- 

 siderable experience tubes were constructed which would resist 

 the disruptive and heating effects of the discharges. This 

 form of tube, however, was abandoned for another form, 

 shown in fig. 1. The chief peculiarity of this form of tube is 

 an X-ray bulb in place of the end covered with the quartz 

 plate. The reason of the adoption of this form of tube is this : 

 it was necessary to heat the entire tube to a high temperature 

 for a long period during the process of exhaustion, to drive 

 out the air and aqueous vapor before it. was filled with hydro- 

 gen, and this heating was impossible with a luted-on end. 



