240 PROCEEDINGS OF THE AMERICAN ACADEMY 



Whatever may prove to be the real cause of the inverse electro- 

 motive force of the arc, it is clear that it is intimately connected with 

 variations in the temperature of the carbons. Thus both the actual 

 temperature and the difference of temperature between the carbons 

 is greater with the silent than with the whistling arc, and at the same 

 time the inverse electromotive force of the former is the hisher. 

 Again, the temperature of the carbons is apparently lowered when 

 metallic salts are introduced into the arc, and the inverse electromotive 

 force falls at the same time. Edlund * found that by heating the nega- 

 tive carbon by a blast-lamp, the inverse electromotive force was raised ; 

 but this was not the case when the flame of the lamp was directed 

 against the positive carbon, and in fact there then seemed to be a 

 fall. It seems very probable that this last result came from the fact 

 that the positive carbon was so much hotter than the flame that the 

 effect of the latter was on the whole to cool the arc. We endeavored 

 to increase the temperature of the positive carbon in a diff'erent 

 manner, by surrounding it with a deep cup-shaped shield of fire-clay, 

 which projected over the end of the carbon so as to envelop the arc, 

 thus preventing radiation. At the same time a broad, flat, horizontal 

 plate of brass was attached to the lower and negative carbon. This 

 obstructed the access of air to the arc, and thus tended to keep up 

 its temperature. 



A current of 7 amperes was employed, with arcs varying from 

 ^ to ^f in. in length. As the carbons could not be filed in these 

 experiments, the length of the arc was obtained by screwing the 

 points into contact while the current was passing, and then separating 

 them to the required distance. A moderately high wire resistance 

 was placed in the circuit with the arc, so that there was no material 

 change in the strength of the current and condition of the dynamo 

 machine caused by this adjustment of the carbons, and the needed 

 measurements could be made rapidly. There was considerable periodic 

 fluctuation of the reading of the voltmeter, as had been noticed previ- 

 ously with the inverted arc, and two curves were drawn correspond- 

 ino- to the extreme readings. The current was sensibly constant. 

 The highest set of values for the equivalent resistances corresponding 

 to different lengths of arc are satisfied for the silent arc by the equa- 

 tion (25) r = 6.70 -j- .16 /, and for the whistling arc by the equation 

 (26) r = 3.02 -|- 1.10 l. The corresponding values of the inverse 

 electromotive force are 46.9 volts and 21.1 volts. The lowest set 



* Pogg. Ann., vol. cxxxiv. p. 250. 



