26 Scientific Proceedings, Royal Dublin Society. 



Kayser states, in his paper, that the lines which Rowland gives 

 near the cyanogen band 3883 could not be seen in his photographs 

 owing to the band being very strong in them. Rowland 1 gives 

 15 lines in this part of the spectrum (3778 to 3876). All these 

 occur in the author's spark-spectrum, with but one exception 

 (3828-32). 



The possibility of some of the lines at least being due to im- 

 purities must be considered. Many of them are, as a matter of fact, 

 coincident with those of other elements, particularly of iron. The 

 author has carefully compared his spark-spectra of ruthenium 

 with those of the rest of the platinum metals, :md of iron, cobalt, 

 nickel, chromium, manganese, copper, calcium, and silicon. As 

 was anticipated from the very large number of lines in most of 

 these spectra, several of them were found to have similar wave- 

 lengths to many of the ruthenium lines ; but on carefully com- 

 paring the character of these lines with those with which they 

 coincided in the ruthenium spectrum, it was quite easy to see that 

 they were not the same. The author has only been able to identify, 

 in his ruthenium spark-spectrum, 19 lines due to iron, and 2 to 

 calcium. 



Exner and Haschek 2 have measured 2250 lines between the 

 limits above given ; some 1330 of these occur in the author's 

 photographs. The larger number of lines occurring in the photo- 

 graphs of these authors is, no doubt, chiefly owing to the fact that 

 different methods of producing the sparks were employed in the 

 two cases. 



The author desires, in conclusion, to acknowledge his indebted- 

 ness to Miss M. Hall for the very valuable assistance she has given 

 him in the work of making the micrometer measurements and of 

 calculating the wave-lengths from them. 



1 Astrophysical Journal, 1896. 



2 Sitzungsberichte der Kais. Akad. der Wiss. in "Wien., 1896. 



