166 Mr. H. Kamage. Abnormal Changes in [Nov. 19 



ing (2 in figure) from 4603*46 to 4602*71, having its middle (measured) 

 at 4603-06. The middle of the broad bright line is, according to the 

 above, at wave-length 4601 *95, and the middle of the dark line (reversed) 

 is coincident with the bright line of the flame spectrum. 



Diagrams of the blue lithium line as it appears in photographs. The fine vertical 

 lines are merely shading lines ; the broad wings of the actual lines in ihe 

 spectra are continuous. 



(1) Bright line of flame, arc and spark spectra. 



(2) Reversed line from dense outer flame of arc. 



(3) Reversed line from inner cone of strong arcs and from near negative 



pole of weak arcs. 



(4) Bi'oad line with two absorption bands, from arc spectrum. 



(5 and 6) From spectrum of thin outer flame of arc with different exposures. 



One pole of a weak bar magnet was placed near the arc so as to 

 direct the expelled vapour in a direction perpendicular to the axis of 

 the collimator. After the large excess of the lithium salt had been 

 volatilised the image of the central portion of the arc was thrown on 

 the slit, and the spectrum photographed, when the ends of the carbon 

 poles, not widely separated, were red-hot. The lithium vapour in the 

 central core was doubtless at a very high temperature, and there was 

 only a thin layer of cooler vapour between it and the slit. Under 

 these conditions the bright line is again broadly expanded, and so also 

 is the dark line, for it now extends from 4603-55 to 4601*25 (mean 

 = 4602-40) on one photograph (3 in figure), and from 4603-68 to 

 460T22 (mean = 4602-45) on another. The bright wings are nearly 

 of the same intensity and extent. The middles of these two dark 

 lines are nearly the same as the wave-length 4602*37, given by Kayser 

 and Eunge, and by Ecler and Yalenta for the reversed lines. 



