28 PROFESSORS G. D. LIVEING AND J. DEWAR 



edges of A, B, and a respectively ; and we have not been able to make out such an 

 exact correspondence between the lines of the water spectrum and those of A, B, and 

 a as M. DESLANDRES' words seem to imply. Nevertheless, the similarity of the 

 grouping is very remarkable, as may be seen from the accompanying map, on which 

 are given the lines of A, on a scale slightly reduced from Professor PIAZZI SMYTH'S 

 solar spectrum, side by side with the lines of the water spectrum. The corres- 

 pondence of B and of a to certain lines of the higher groups is less striking. We 

 have no doubt that the peculiar arrangement of the lines, commencing at the more 

 refrangible end with some closely set lines and continued in a series of doublets, is in 

 all these cases the result of a general law. 



The tables of wave-lengths which follow were obtained by measuring the distances 

 of the lines from those of iron photographed on the same plate through a part of the 

 slit. The whole of the lines previously measured have been re-measured, and the 

 numbers assigned as their wave-lengths corrected by reference to more recent measure- 

 ments of the iron lines. The scale is, however, that of ANGSTROM. On account of 

 the faintness of the rays at the two extremities of the spectrum a wide slit had to be 

 employed, as well as a lengthened exposure in photographing them, consequently 

 the photographed lines are broad and somewhat ill-defined, and groups of closely set 

 lines could not be resolved. The wave-lengths in the first and last tables are 

 therefore liable to greater errors than those in the intervening tables. Moreover, 

 in these parts, the spectrum is so weak that the strongest line is less intense than 

 the weakest lines of the region about X 3] 00, and it has been found necessary to 

 adopt a different scale of intensities for different sections. 



It is, perhaps, worth remarking that there is a broad diffuse line, the strongest 

 of the lines in that region, coincident with K of the solar spectrum. There is no 

 line coincident with H. 



