228 PROCEEDINGS OF THE AMERICAN ACADEMY 



These facts might lead one to infer the probability of there being 

 a sudden rise iu the inverse electromotive force of the arc at the 

 point when whistling ceases. So far as we are aware, this matter 

 had not been studied previous to the observations which we have to 

 present. 



The quantity aC was originally supposed by Edlund to be of the 

 nature of an opposing electromotive force. It is true that this view 

 has not been accepted without question, and that the real cause of the 

 phenomena under consideration is still somewhat uncertain ; but we 

 shall, nevertheless, in this paper, for convenience of descri2)tion, refer 

 to this quantity a (7 as the inverse electromotive force of the arc. 



The object of the experiments described in the present paper was 

 to review the results of earlier observers on the quiet arc, to extend 

 the observations to the whistling arc, and to observe the effect on the 

 inverse electromotive force of the arc caused by variations in the 

 position and temperatuie of the electrodes, by the presence of volatil- 

 ized substances in the arc, and by variations in the density of the air 

 in which the arc is formed. 



A lamp was constructed for the purpose whose upper carbon was 

 movable in a vertical direction by a micrometer screw, having a pitch 

 of 55 of an inch. Wires ran from brass collars surrounding the car- 

 bons, and as near as possible to their extremities, to a voltmeter, which 

 gave the difference of potential between the carbons. The current of 

 electricity was usually furnished by a Brush dynamo machine. It 

 was measured in some cases by means of a carefully and frequently 

 calibrated Ayrton and Perry permanent magnet ammeter, and in 

 others by a Paterson and Cooper electromagnet ammeter, also cali- 

 brated by comparison with standard instruments. A similarly cali- 

 brated Ayrton and Perry permanent magnet voltmeter was employed 

 for the potential measurements. Reversals of readings were obtained 

 by a mercury commutator so arranged as to reverse both the current 

 and potential instruments simultaneously. The quotient of the differ- 

 ence of potential between the carbons divided by the current gave 

 the equivalent resistance of the arc. Care was taken that the arc 

 should be formed so as to pass between those points of the carbons 

 which were closest together, which was secured by filing them flat. 

 The separation between the points as measured by the micrometer, 

 therefore, gave the length of the arc. It was of course necessary to 

 make the measurements with a sufficient degree of quickness, so that 

 no sensible burning away should occur before they were finished. 

 No difficulty was found in doing this, as it took a number of seconds 



