346 Messrs. Wright and Rennie on the Determination of 



Apparently, therefore, with the stronger currents the varnish 

 produced an increase in resistance about equal to that due to 

 an increase in temperature of from 4° to 5°; so that if the 

 varnish in Joule's experiments produced a similar effect, the 

 total superheating must have been near 12° to 14°, represent- 

 ing 037 to 043 per cent, of error in deficiency in the estima- 

 tion of R, and consequently of J. 



53. Yet another source of error, and in the same direction, 

 in these experiments of Joule's lies in the fact that, in compa- 

 ring the resistances of the B.A. unit and the experimental 

 wire employed, the method adopted consisted in determining 

 the angular deflections «, /3, and 7 produced in a tangent-gal- 

 vanometer on passing the current from a given electromotor 

 (1) through the galvanometer only, (2) through the galvano- 

 meter and B.A.-unit coil, and (3) through the galvanometer 

 and experimental wire — the resistance of the experimental 

 wire being called x, and that of the unit coil =1, 



_ (tan a —tan 7) tan /3 

 (tan a — tan /3) tan 7' 



Now, were the currents used so feeble that no appreciable 

 amount of heat was developed by them in either wire, or were 

 the two wires compared similarly sized and situated so as to 

 be heated to equal extents, the value of x thus deduced would 

 be subject to no other errors than the instrumental and obser- 

 vational ones ; but as the B.A. -unit-coil wires are imbedded 

 in solid paraffin whilst the experimental wire was immersed 

 naked in water (save for the film of varnish), it is evident that, 

 if any heat at all had been generated by the currents employed, 

 the B.A.-unit coil must have been mora heated than the expe- 

 rimental wire, as the heat developed could not pass away 

 readily through the badly conducting solid paraffin. The cur- 

 rents actually employed were sufficient to cause deflections of 

 36° to 37° with a galvanometer of nine turns and 17 inches 

 diameter, and of 34° to 50° with a galvanometer of eighteen 

 turns and 17 inches diameter; so that the currents were of 

 magnitude equal to from 



17x2-54 T , 0A , 17x2-54 T , Q7 



— — . 1 tan 34 to — — x 1 tan 37 



47rxl8 47rx9 



(where I is the horizontal magnetic terrestrial force =0*18, 



and 2*54 is the number of centims. in an inch), or averaged 



about 0*04 C.Gr.S. unit. In a wire of 1 B.A. unit resistance 



a current of this magnitude would produce per minute 



0*04 x 0-04 x 10 9 x 60 OQ ' , . , 

 jn — Ycfi —l'o gramme-degrees, which, assuming 



