Charged Bodies at Moderate Temperatures. 99 



§ 3. Testing of the Apparatus. — The insulation was tested 

 first of all before the gas was lit. First, the electrometer 

 alone was tested, then the electrometer and air-leyden 

 together. When this latter test was made, the outside box, 

 the cylinder, the case of the electrometer, and the wire gauze 

 were put in metallic connexion with one another. The 

 insulated terminal of the air-leyden was connected to that 

 of the electrometer and charged to the desired potential. 

 The rate of leak was almost exactly the same as that when 

 the electrometer alone was charged. 



The capacity of the air-leyden and its connexions was 

 approximately the same as that of the multicellular. It was 

 found that # 34 of an electrostatic unit of electricity raised 

 the potential of the electrometer by 1 volt when it was con- 

 nected to the air-leyden. Thus a drop of potential of 1 volt 

 per minute is equivalent to 

 •34 

 l^xloro am P eres > 



or # 6 189 microamperes. 



In the next instance the flame was applied to the outside of 

 the box. It was found that the rate of leak after half an 

 hour was not appreciably different from that at the end of 

 half an hour when no flame was applied; the original poten- 

 tials being the same in each case. In these pi'eliminary tests 

 the outside box of the air-leyden was of tinplate, and the 

 insulated plates, which were of course connected to the 

 insulated terminal of the electrometer, of zinc. In this way 

 it was shown that neither the zinc nor the tinplate was heated 

 to such a temperature as to cause an increased leak of elec- 

 tricity from that cause alone. The same test also showed 

 that the insulated part of the apparatus was sufficiently 

 guarded from the fumes of the flame to prevent increased 

 leak on their account. 



The next thing to do was to see whether or not the appa- 

 ratus was sufficiently sensitive. For this purpose the insulated 

 zinc plates were covered with a uranium salt, no heat being 

 applied. It was found that the leak of electricity was in this 

 case markedly increased. A very much greater leak was 

 also observed when a small piece of phosphorus was placed on 

 the insulated plates. 



The insulated zinc plates were next formed each in the 

 shape of a very shallow vessel, and both were filled with 

 water. The flame was applied, and the rate of leak observed 

 for about half an hour ; steam began to come from the open 

 end of the box after about ten minutes: it was found that 



H2 



