1890.] between Electrodes at different Temperatures. 



121 



employed to cover up the negative leg of the carbon ; in any case this 

 shielding destroys the effect. 



(11.) If, instead of shielding the negative leg of the carbon, a 

 mica screen is interposed between the negative leg and the side of 

 the middle plate which faces it, then the current produced in a gal- 

 vanometer connected between the positive terminal of the lamp and 

 the middle plate is much reduced. Hardly any effect under the same 

 circumstances is produced when the mica screen is interposed on that 

 side of the metal plate which faces the positive leg of the carbon. 



(12.) The position of the metal plate has a great influence on the 

 magnitude of the current traversing a galvanometer connected 

 between the metal plate and the positive terminal of the lamp. The 

 current is greatest when the insulated metal plate is as near as pos- 

 sible to the base of the negative leg of the carbon, and grea test of all 

 when it is formed into a cylinder which embraces without touching 

 the base of the negative leg. 



The current becomes very small when the insulated metal plate is 

 removed to 4 or 5 inches from the negative leg, and becomes prac- 

 tically zero when the metal plate is at the end of a tube forming part 

 of the bulb, which tube has a bend at right angles in it. Copious 

 experiments have been made with metal plates in all kinds of 

 positions. 



(13.) The galvanometer current is greatly influenced by the surface 

 of the metal plate, being greatly reduced when the surface of the 

 plate is made small, or when the plate is set edgeways to the negative 

 leg, so as to present a very small apparent surface when seen from 

 the negative leg. In a lamp having the usual commercial vacuum, 

 the effect is extremely small when the insulated metal plate is placed 

 at a distance of 18 inches from the negative leg, but even then it is 

 just sensible to a very sensitive galvanometer. 



(14.) If a charged condenser has one plate connected to the insu- 

 lated metal plate, and the other plate connected to any point of the 

 circuit of the incandescent filament, this condenser is instantly dis- 

 charged if the positively charged side of the condenser is connected 

 to the insulated plate, and the negative side to the hot filament. If, 

 however, the negative leg of the carbon horseshoe is shielded by a. 

 glass tube, this discharging power is much reduced, or altogether 

 removed. 



(15.) If the middle plate consists of a separate carbon loop, which 

 can itself be made incandescent by a separate insulated battery, then, 

 when this middle carbon is rendered incandescent and employed as 

 the metal plate in the above experiment, the condenser is dis- 

 charged when the negatively charged side of it is connected to the 

 hot middle carbon, the positively charged side of it being in con- 

 nexion with the principal carbon horseshoe. 



