100 PROCEEDINGS OF THE AMERICAN ACADEMY. 



these being, of course, modified by the condition of the echelon, 

 whether of absolutely single or double order, or part way between the 

 two. However, interpreting the pattern is a simple matter in either 

 case. The echelon was so placed that in the field of the microscope 

 the lower orders lay at the left, the higher at the right. 



Zinc 4810. Upon starting the arc after a fresh piece of zinc was 

 put in, the whole field resembled that of a polychromatic source, 

 except that the normal diffuse echelon image was marked by several 

 fine lines similar to the pattern shown in Plate 1, 12, whether the posi- 

 tion of the echelon for monochromatic light be that of single or double 

 order. This structure always accompanied the arc when noisy, and 

 was present in 4722 and 4680, as well as 4810. It is clearly visible 

 in arc lines with the Petitdidier instrument and is similar to the 

 "fluted appearance" of spark lines. At low current and the single 

 order condition, eight components were visible in 4810, the two outer- 

 most poorly marked; the two innermost the sharpest of all. As the 

 vapor became less dense, the structure became less extensive. There 

 appeared two lines strongly marked, lying between two other wider, 

 less intense and less sharp satellites. Then the two outer satellites 

 faded, the stronger, inner pair at times receded from each other and 

 then again approached. Finally the reversed region (if, indeed, we 

 are justified in speaking of the phenomena as a " reversal") disappeared 

 and the two lines merged into one, which eventually became a single, 

 narrow line. 



The above phenomena were noticed in the Pfund arc at 2.5 amp., 

 with the upper pole (in error) positive. The condition was the 

 double order one, the lower order being slightly stronger. The same 

 phenomena appeared, however, at single order condition, and with 

 lower pole positive. 



At 3.4 amp. at another temperature such that the condition was 

 nearly single order, and with the lower terminal positive, the same 

 general phenomena appeared, but when the two line structure was 

 present and the vapor density was decreasing, the right component 

 became weaker than the left; whereas when judged by the fact that the 

 adjacent order was stronger, it should have been the stronger of the 

 two. However, the right component sometimes appeared stronger than 

 the left and was generally broader. At this point the current was re- 

 duced to I'.l amp. and the right component, although visible near 

 the lower, or positive terminal, disappeared at the center of the arc, 

 the line there being single. 



At 5.5 amp. and the lower pole negative, the phenomena of 2.5 



