Cady and Vinal — Electric Arc. 97 



by means of rubber bands until it passed an opening in the 

 chute at a velocity of about 565cm. / sec. = 0*000177 sec. / mm. 

 By means of an achromatic lens a minute image of the arc was 

 formed on the plate as it passed this opening. 



A number of plates were dropped under varying conditions 

 of arc, the current being on the order of 0'3 amp. and the 

 length of arc l mm . 



In nearly all cases where the bolometer system showed a 

 deflection the record on the photographic plate was not a con- 

 tinuous line, but broken into a system of more or less periodic 

 dots or dashes. These occurred in trains of from five to 

 several hundred waves, the period being practically constant 

 throughout each train, but varying abruptly from one train to 

 the next. No evidence of damping could be seen. 



The lowest frequency observed was about 1,300 per second, 

 the highest about 43,000. In one case a change from a 

 frequency of 1,300 to one of 10,000 took place on the same 

 record. The velocity of the plates and clearness of the image 

 hardly permitted the identification of a frequency higher than 

 50,000, so it cannot be affirmed that higher frequencies were 

 not present. 



These experiments show that in the oscillating carbon - 

 copper arc in free air, the current and the light emitted from 

 the arc are varying constantly with a period which itself is 

 subject to sudden and great variations (cf . § 3). 



§ 10. Following is a brief account of experiments made 

 with a view to learning under what conditions the regularity, 

 frequency and intensity of these oscillations can be increased. 



In C0 2 and in N. Deflections of bolometer small and 

 unsatisfactory. 



In H. No arc, but only a glow, could be obtained except 

 when the current was over 0*3 amp. 



A little benzine vapor was then introduced by allowing the 

 hydrogen to bubble through a bottle of benzine. * This 

 enabled us to decrease the arc current to 0*2 amp., at which 

 current good oscillations occurred. However the use of ben- 

 zine was discontinued owing to the copious deposits of soot 

 and to the tendency of the arc to go out. 



In illuminating gas. This gave far better results than any 

 of the gases mentioned above. A silver cathode was chiefly 

 used. On account of the deposit of carbon dust from the 

 dissociated gas, the latter was diluted by the addition of a cer- 

 tain amount of nitrogen or air. It was in illuminating gas 

 that large deflections were observed for the first time when the 

 bolometer circuit was entirely disconnected from the arc 

 * Cf. Barreca, Electrician, Jan. 17, 1908. 



