344 Messrs. Wright and Bennie on the Determination of 



Since the average temperature of the water surrounding the 

 wire was nearly the same throughout, not differing by more 

 than a few tenths of a degree from the mean 8°*75, the values 

 in the last column are fairly comparable with one another ; 

 whence it is evident that a notable rise in resistance of the 

 wire is produced by each successive increase in current- 

 strength. The resistance of the wire for a current of very 

 small value would clearly lie under 1-1495 x 10 9 ; so that the 

 increase in resistance caused by currents of from 0*24 to O30 



nri « ., . , , t 1*1695-1*1495 , 1*172-1*1495 

 O.U-.b. unit is at teast ., _. . QK to 1 Q , , or 



1-7 to 2-0 per cent. ^^^ L ' L ™ b 



50. In order to see what increase in mean temperature of 

 the wire over that of the surrounding water this would re- 

 present, the fourth experiment was repeated, using warm water 

 (at about 27°) to surround the coil : in this way the following 

 numbers were obtained : — 



C.G.S. Average Potential- 

 current, temperature, difference. Resistance. 



Water warmed . 0*1419 27*25 1-716 x 10 8 1*209 x 10 9 

 „ cold . , 0-1459 8-20 1-697 x 10 8 1-163 x 10 9 



Difference . 19-05 0-046 x 10 9 ~ 



Since an increase in resistance of 0*046 corresponds to 19*05 

 degrees, an increase of 0*0200 to 0*0225 would correspond to 8*3 

 to 9*3 degrees ; i. e. currents of 0*24 to 0*30 C.Gr.S. unit caused 

 a superheating of the wire above the temperature of the sur- 

 rounding water to an extent averaging 8°*3 to 9°*3*. If it 

 be admitted that in Joule's experiments a similar superheating 

 to the extent of 8° to 9° may have taken place, the effect on 

 the calculated resistance of the platinum-silver used would be 

 to cause the amount to be underestimated by about 8 x 0*031 

 = 0*25 per cent, to 9 x 0*031 = 0-28 per cent., 0*031 being the 

 percentage increase in resistance per degree of the alloy used 

 by Joule (this alloy was purposely not used by the authors, 

 in order to make the alteration in resistance, if any, more per- 

 ceptible) ; that is, the value of J ultimately deduced would be 

 about 0*25 to 0*28 per cent, too low. 



51. In order to see whether an increased amount of super- 

 heating would be produced by only stirring the water sur- 

 rounding the wire at intervals of one or two minutes instead 

 of continuously, some of the above observations were repeated 



* This increase in resistance with temperature amounts to about 0-21 

 per cent, per 1° C. Matthiessen found the average increase of pure metals 

 to "be about 0*37 per cent, per 1° C, that of an alloy of platinum and iri- 

 dium containing 33 per cent, of the latter being only O06 per cent, per 1°; 

 probably, therefore, this platinum wire was somewhat impure, very likely 

 containing iridium. 



