102 PROCEEDINGS OF THE AMERICAN ACADEMY. 



At 2.5 amp. the left component appeared a little weaker and the 

 lines were less sharp than with less current. At low vapor density 

 there appeared a single line, slightly hazy on both sides. With small 

 amount of vapor and in double order condition, the lines appeared 

 hazy, and no re^'ersal was to be seen. This line, 4G80, has, however, 

 been observed at this current in single order condition as a single 

 line with two side components which broadened at certain times, all 

 three lines broadening and receding from each other. 



At 3.5 amp. the two components were far apart, and equal in inten- 

 sity. With small amount of vapor the reversal was less well marked. 



The results of this visual study of the arc may be summarized as 

 follows : — 



(1) All three lines are at times single. 



(2) All have been observed "reversed" but, 



(3) 4810 and 4680 are generally double or quadruple, while 4722 is 

 generally triple, and, 



(4) All are still more complex at times and show asymmetry, but 

 this asymmetry is no more often marked by stronger red satellites 

 than violet. 



Visual study of the inductance spark lines. A visual study 

 of the spark with inductance showed that the conditions could in this 

 case be more easily controlled and were more steady. 



4810, in a nearly single order condition, the right order of the 

 three being stronger than the left, with coil (b) as inductance in the 

 condenser discharge circuit, showed two lines, the right component 

 distinctly weaker than the left when the arrangement of the orders 

 would, if the two components were intrinsically of the same intensity, 

 make this right component the stronger. 



The central part of the image was under observation. This right 

 component, however, appeared stronger when the end of the spark 

 image was thrown upon the slit. With no inductance the fluting 

 was faintly visible (it had not been observed previously with the 

 Petitdidier instrument ^^) for the gap was small so that the spark 

 burned quietly. 



4722 showed the fluting more clearly than 4810, and with inductance, 

 the condition being nearly single order, there appeared two bright 

 central lines which were nearly equal at times. Whenever unequal, 

 however, the left line was the stronger. At the end of the image near 

 the terminals the line broadens out and resembles the disruptive spark. 



19 See page 97. 



