406 Mr. It. Sabine on some Electrical Experiments 



(g) of comparatively low resistance, by means of which the 

 current moving in that side could, within a very small per- 

 centage of error, be observed. The battery-circuit was pro- 

 vided with a sliding resistance (r), by means of which, which- 

 ever section of the selenium plate was in circuit, the current 

 in it, as indicated by galvanometer (g), could be kept at a 

 constant value. In this way the following measurements were 

 made: — In the first series the current-strength in the selenium 

 was kept uniformly at 2*9 microwebers, and in the second at 

 0*42 microweber. The positions of 

 the selenium are indicated as "direct" 

 and " inverted" — " direct " when the 

 current moved in the selenium as 

 shown by the arrow (fig. 3), "in- 

 verted" when it moved the other way. 



Fig. 3. 



Resistance Current =2-9 micro w. 



Current =0*42 microw. 



measured 

 between 



Direct. 



Inverted. 



Direct. 



Inverted. 





meg. 



meg. 



meg. 



meg. 



a and b 



•3176 



•3181 



•3202 



•3244 



b „ c 



•3864 



•3858 



•3936 



•3893 



c „ d 



•2734 



•2734 



•2747 



•2782 



a „ c 



•6095 



•6097 



•6116 



•6116 



b „ d 



•4842 



•4846 



•4906 



•4900 



Call the resistance of the selenium 

 f direct 1 



when the current is 



between b and <? i j f 



\ inverted / ; tlie resistance of tne j unc - 

 b and c \ 

 6'andc'j 



resistance from a up to the junction b < f 



from the junction 



tions 



{ 



when the current is {i^eSedj ' the 



direct ) ? 

 inverted J ' 

 to the end 



and, lastly, the resistance 

 , fd direct \ 



\d f inverted J ' 



It is evident that by measuring the resistance between each 

 of the wires (a, b, c, and d), in turn, with two different cur- 

 rent-strengths, we have data which enable us to calculate the 

 mean resistances of the junctions as well as that of the sele- 

 nium. If, with the two different current-strengths, the calcu- 

 lated values of the junctions agree, we may assume that the 

 change resides in the selenium ; on the other hand, if the sele- 

 nium agrees in both calculations, the change must reside in 

 the junctions. 



The above tests furnish the following results : — 



