1884. Spectroscopic Studies on Gaseous Explosions. 475 



violet two lines, X 3272 and X 3245 - 5. All three lines were very 

 strong, and the two ultra-violet lines were in some cases reversed. 

 These lines were also frequently developed when no copper linjng 

 was in the tube, probably from the brass of the small side tubes. 



Copper also gave a line in the indigo, A, 4281 about, decidedly less 

 refrangible than the copper line, X 4275, coincident apparently with 

 the strong edge of one of the bands developed when a copper salt is 

 held in a Bunsen burner. 



A lining of copper which had been electro-plated with nickel 

 developed only one nickel line, A, 5476, in the visible part of the spec- 

 trum, but gave by photography the following lines in the ultra- 

 violet : — 



3807-5 3524 3445 '5 



3783 3514 7 3432 



3775 3510 3422 



? 3641 3492 3413-2 



3618 -3 3461 '5 3391 -5 



3612-5 3457-7 3378*4 



3597-3 3453 3369*6 



3571-5 3451 3367 '4 



3565 



When nickel oxalate was put into the tube, lines with wave-lengths 

 3670-5, 3470-3 and 3389-6 in addition to the preceding were 

 developed. It is doubtful whether the line X 3451 be a nickel line. 

 That at X=3453 is ascribed to cobalt by Cornu, but it seems to be a 

 nickel line as well. 



When copper wire electro-plated with cobalt was put into the tube 

 cobalt lines appeared with the approximate wave-lengths : — 



4119 3594 3492 ? 



4089 3568 3474 



3995 3528 | 3462 



3909 3525 \ 3453 



3894 3523) 3431 



3872 3502 3411 



3845 3495 3404 



3601 



The lines X 3528 to 3522 form a continuous band in the photo- 

 graph, so that these three lines may not represent the whole group at 

 that spot. It is doubtful whether X 3492 be a cobalt line as well as a 

 Ni line. 



1$o other metal gave anything like the number of lines that were 

 given by iron, nickel, and cobalt. 



