592 PROCEEDINGS OP THE AMERICAN ACADEMY. 



is slight, aud that there are no sharp bands in it. Remarkable is the 

 fact that fluorite, which shows no marked absorption out as far as A 1200, 

 should be strongly phosphorescent. This is the case with some very 

 pure white plates which are used by Dr. Theo. Lyman in his work on the 

 photography of very short wave-lengths. They show no measurable 

 absorption of a selective sort, and permit the passage of light evenly as 

 far as A 1200, while they are more strongly phosphorescent under the 

 influence of the spark between metallic terminals than many colored 

 fluorites. 



Photographs of my own which extend as far as A 2100 show no 

 marked absorption aud no selective absorption whatever. These facts 

 simply point to the flatness of the absorption curve for this substance, 

 aud they are all the more remarkable when we consider the great 

 sharpness and intensity of the fluorescence lines and bands from these 

 same fluorites. 



Plate 2 shows three photographs enlarged from small plates like those 

 in Plate 1, and they show very clearly that many of the lines in the 

 fluorescence spectra are equally as sharp as the lines in the comparison 

 spark spectra. Some of these lines occur in groups as doublets and 

 triplets, having all of the characteristics as regards appearance and sharp- 

 ness which characterize the spectra of incandescent metals and gases. 



Figure 1 shows a portion of the fluorescence spectrum of a crystal 

 of fluorite excited by the spark between magnesium terminals. It is 

 especially characterized by three strong lines, one of them very strong 

 indeed. 



Figure 2 shows the fluorescence spectrum of the same crystal under 

 the exciting influence of the spark between iron terminals. Beside the 

 strong groups in the yellow, orange, and red, there may be seen iu this 

 photograph a large number of lines which are partially obscured by 

 the broad band. That these are not lines which have entered the 

 spectrograph from the light of the spark is proven by the fact that 

 the very strongest spark lines are missing from the fluorescence spec- 

 trum. Iu fact, every line which can be seeu in this spectrum is a true 

 line of fluorescence light. 



Figure 3 shows the fluorescence spectrum of the same crystal under the 

 exciting influence of the spark between cadmium terminals. A number 

 of lines have in this case been reflected in the crystal through poor ad- 

 justment, and these appear in the fluorescence spectrum. The strong 

 triplet A 5737 to A 5711 is however made up wholly of fluorescence 

 lines, and the other lines belonging to this spectrum can readily be sepa- 



