1879.] 



Note on " Spectroscopic Papers." 



167 



potassium lines which are often seen as one, not merely from want of 

 dispersive power (which with prisms is usually sufficient at the violet 

 end), but because the lines are expanded until they meet. We at 

 first saw them reversed as one broad line which divided in two as the 

 potassium evaporated. 



Mr. Lockyer complains that we omitted to refer to his work on the 

 wave-length of the potassium lines. In our first communication on 

 the subject of reversals we named at the outset those who had been 

 previously engaged in the same field, and after referring to their work 

 in general terms we did not think it desirable subsequently to notice 

 the specific results obtained by each individual. As neither the exist- 

 ence of the two potassium lines nor the exact determination of their 

 wave-lengths, but only the identification of the dark lines we ob- 

 served with the bright lines was in question, we considered it un- 

 necessary to mention specially Mr. Lockyer's work on those lines. 



We owe it to the Society to explain the value which is to be 

 attached to the wave-lengths mentioned by us. We have never set 

 ourselves to determine exact wave-lengths ; but we sought to deter- 

 mine the conditions of reversal, and we used wave-lengths for the 

 convenience of indicating the lines of which we wrote. In very 

 few cases have we given any determinations of wave-lengths. In 

 general we determined the identity of a dark line with the cor- 

 responding bright line by observing that both had the same place 

 on the cross- wires, or pointer, or both gave the same reading of 

 the scale of the spectroscope. We then gave the wave-length of the 

 bright line as determined by some good authority, usually either 

 Thalen or Boisbaudran. When we could not identify the dark lines 

 in that way, we took the readings of known lines with our spectro- 

 scope, using, for lack of sunshine, most frequently Boisbaudran's 

 method and wave-lengths, and having drawn a curve, determined the 

 wave-lengths of the dark lines therefrom. It was in this way that we 

 obtained the wave-lengths of the two dark lines we ascribed to potas- 

 sium. It is, of course, just within the bounds of possibility that the 

 line for which we obtained the wave-length 4045 (it is misprinted 

 4044 in the table, but not in the text) may have been an iron line 

 reversed, but, as the principal well-known lines of potassium were at 

 the same time seen bright in the spectrum while those of iron were 

 not seen, it is far more probable, almost certain, that the line was 

 really due to potassium, and the last figure of the wave-length 

 wrongly determined. By giving exactly what we observed, we have 

 enabled our readers to judge for themselves what is most probable in 

 this case. To attempt to determine wave-lengths to any greater 

 degree of nicety would, so far as our purpose was concerned, have 

 been labour thrown away. Whenever the wave-length of a line 

 observed by us to be reversed is given differently from that given by 



