EXPERIMENTAL ARRANGEMENT TO DETERMINE THE ELECTROMOTIVE POWER. 91 



as well as himself, has made experiments to determine the quantities of electricity set in 

 motion by known increments of heat. From these experiments, they conclude that 

 through the whole range of the thermometric scale, those quantities are directly propor- 

 tional to each other. 



335. But as thermo-electric currents are now employed in a variety of delicate physi- 

 cal investigations, and as there appears to be much misconception as to their character, I 

 propose in this memoir to show, 



1st. That equal increments of heat do not set in motion equal quantities of electricity. 



2dly. That the tension undergoes a slight increase with increase of temperature; a 

 phenomenon due to the increased resistance to conduction of metals, when their tem- 

 perature rises. 



3dly. That the quantity of electricity evolved at any given temperature is independ- 

 ent of the amount of heated surface ; a mere point being just as efficacious as an indefi- 

 nitely extended surface. 



4thly. That the quantities of electricity evolved in a pile of pairs are directly propor- 

 tional to the number of the elements. 



336. First, then, as to the comparative march of electric development, with the rise 

 of temperature, in the case of pairs of different metals. 



337. The experimental arrangement which I have employed is represented m Jig. 31. 

 A A is a glass vessel about three inches in diameter, with a wide neck, through which 

 can be inserted a mercurial thermometer, b, and one extremity of a pair of electro-motoric 

 wires. The wires I have employed have generally been a foot long, and one sixteenth 

 of an inch in diameter. The extremity, s, of the wires thus introduced into the vessel 

 ought to be soldered with hard solder ; their free extremities dip into the glass cups, d d, 

 filled with mercury, and immersed in a trough, e, containing water and pounded ice. By 

 means of the copper wires, ff, one sixth of an inch thick, communication is established 

 with the mercury cups of the galvanometer. The coil of this galvanometer is of cop- 

 per wire one eighth of an inch thick, and making twelve turns only round the needles, 

 which are astatic. The deviations were determined by the torsion of a glass thread, in 

 the way described in Chapter VIII. 



338. It is surprising to those who have never before seen the experiment, with what 

 promptitude and accuracy a copper and iron wire, soldered thus together, will indicate 

 temperatures. 



339. In the arrangement now described, when an experiment has to be made, the 

 vessel A A is to be filled two thirds full of water, the bulb of the thermometer being 

 so adjusted as to be in the middle of the vessel, and the soldered extremity, s, of the two 

 wires being placed in contact* with it, and a small cover with suitable apertures ad- 

 justed on the top of the vessel, so that the steam, as it is generated, may rush up along- 

 side of the tube of the thermometer, and bring the mercurial column in it to a uniform 

 temperature. The communicating wires, ff, are then placed in the cups, and the trough, 

 e, filled with water and pounded ice, and carefully surrounded with a flannel cloth. The 



* If the extremity of the thermo-electric pair be allowed to rest on the bottom of the glass vessel, no accurate results can 

 bfi obtained ; the pair does not then indicate the temperature of the water. 



