342 Messrs. Wright and Eennie on the Determination of 



doubtful if it would wholly remove the additional tendency to 

 superheating from this cause. 



48. In order to form some idea of the possible extent to 

 which the wire might thus become superheated by currents of 

 the magnitude used by Joule*, a platinum wire about 50 

 centims. long and 1*15 B.A. unit resistance was twisted into 

 a spiral, the ends of which were soldered to thick copper rods 

 surmounted with mercury-cups and passing through an india- 

 rubber cork to keep them together : this spiral was immersed 

 in a beaker containing about a litre of distilled water, the tem- 

 perature of which was read off to o, l degree by a thermo- 

 meter placed with its bulb almost touching the centre of the 

 spiral ; by means of a broad horizontal ring of metal attached 

 to a stout wire the water could be briskly agitated, to a much 

 greater extent indeed than would be safe with a calorimeter in 

 which splashing must be avoided. Currents of different 

 strengths were then passed through the coil for periods of time 

 varying from 20 minutes to an hour, a voltameter being also 

 included in the circuit, and the total gas evolved collected as 

 previously described. During this time the water was kept 

 continually stirred, and its temperature observed at equal inter- 

 vals of time, whilst the difference of potential existing between 

 the ends of the wire was observed as frequently as possible by 

 connecting the mercury-cups with a quadrant-electrometer 

 standardized by a Clark cell before and after the experiment, 

 precisely as above described. During the first two or three 

 minutes the electrometer-readings always diminished through 

 diminution of current by polarization of the voltameter-plates; 

 afterwards they remained nearly constant, gradually rising 

 through the increasing mean temperature of the wire, or 

 slightly sinking if the current-strength diminished through 

 continued use of the battery. During the first few minutes 

 the readings were accordingly taken every 10 or 15 seconds 

 without reversal, so as to obtain a large number of observa- 

 tions giving the mean potential difference during this period ; 

 subsequently the readings were taken by reversal at the rate 

 of one pair per minute, the readings being divided into a 



* In Joule's experiments the galvanometer made one turn only of 

 0-62723 foot radius, the angular deflection varying from 26° to 32°: hence 



0-fi979^ v ^0-48 

 the currents must have lain between u "*' '* x ou *° . I , tan 26° and 



0-62723x30-48 T , OOQ , on .. . . 2n . 



■ 5 . 1 . tan 32 y , where 30-48 is the number of centims. m a 



foot and I the horizontal terrestrial magnetic force. Taking 1 = 0-18, the 

 currents must therefore have averaged about 0*3 O.G.S. unit. The resist- 

 ance of the platinum-silver wire used was nearly equal to 1 B.A. unit = 10 9 

 C.G.S. units. 



