OF ARTS AND SCIENCES. . 235 



nite. With the whistling arc, there was always great unsteadiness, 

 which was very much diminished at the moment when the arc hecarae 

 quiet. At the same time the positive carbon suddenly brightened, 

 indicating a great rise of temperature. 



A quite interesting peculiarity was noticed in the curves for the 

 whistling arcs. As will also be seen from the equations, the lines all 

 converge approximately towards a point situated behind the axis of 

 ordinates, and corresponding algebraically to a negative separation 

 of about 9 X 32 inches, the corresponding algebraic value of the 

 resistance being about -^^ ohms. This property of the curves fur- 

 nishes a means of finding approximately the equivalent resistance 

 of any whistling arc, as follows. Divide the average value of the 

 inverse electromotive force, 15.13 volts, by the strength of the current, 

 which gives the intercept, a, corresponding to the current in question. 

 Draw a line from the point of convergence as given above, through the 

 upper point of the intercept, and it will be the curve for this current. 

 The ordinate corresponding to any abscissa gives the equivalent resist- 

 ance for that length of arc. The curve for the smallest current used, 

 3.27 amperes, alone shows much deviation from the rule. The curves 

 for the silent arcs do not show any such definite point of convergence. 



It is interesting to examine Edlund's early values fur the inverse 

 electromotive force in the light of the preceding results. His first 

 values are expressed in units which are purely arbitrary. They show 

 approximately, however, the same figure for a O through a considerable 

 range of variation in the strength of the current. His later results 

 give 9.7 as the value of the inverse electromotive force with a current 

 of 30 Buusen elements, the electromotive force of a Bunsen cell being 

 taken as a unit. With 50 elements its value rose to 15. It seems 

 probable that the former result was reached with a whistling arc, and 

 the latter with a silent one. It is difficult to explain the existence 

 of certain intermediate values, but the method by which they were 

 obtained is liable to great errors. 



If we consider Peukert's paper, we find that the results from which 

 he derives his value of 35 do not include any with a less current 

 than 10 amperes, and but one in which the length of the arc is less 

 than 2 mm., so that the arcs were probably all silent. The observa- 

 tions from which Frohlich constructs the curve giving the value of 39 

 are too irregular to show any difference that might exist between the 

 values of the constant for the silent and for the whistling arcs. Von 

 Lang states that the arcs measured by him were carefully kept from 

 hissing by manual adjustment. 



