Prof. Bunsen on Gcesium. 247 



sented their reduced observations in a drawing. This drawing 

 does not agree at all well with the numbers in column 3, and it 

 is therefore evident that a different mode of interpolation must 

 have been used by Messrs. Johnson and Allen from that which 

 I have employed. In their paper they do not describe the method 

 which they adopted for reducing their observations ; the follow- 

 ing example shows, however, that their method cannot have 

 been a suitable one. According to the above statement, the 

 value of one division on the two scales from which the numbers 

 in columns 1 and 2 are derived is nearly equal, and the divi- 

 sions 100 and 50 correspond exactly to each other as the sodium 

 line stood at these points on both scales. Now Messrs. Johnson 

 and Allen read off on their instrument division 101 for the line 

 VIII. ; hence, when this is reduced to the other scale, we must 

 obtain a value nearly equal to 51 . In Messrs. Johnson and Allen's 

 drawing, however, the line VIII. has the limits 52*5 and 53*5. 



A glance at the foregoing Table shows that Messrs. Johnson and 

 Allen have observed a large number of lines which are not repre- 

 sented in our drawing. The greater number of these we have also 

 measured ; but they were not included in the drawing, because, as 

 we have already frequently remarked, we did not endeavour to ca- 

 talogue the lines completely, but to represent as truly as possible 

 the characteristic appearances of the spectra of the several sub- 

 stances. Owing to the difficulty of accurately distributing many 

 shades of the same colour in chromolithographic printing, the 

 weaker lines have necessarily been omitted in our drawings. 



I have still to observe, in reference to Messrs. Johnson and 

 Allen's remarks, that it is line XVI., and not line XV. as they 

 state, which is coincident with Fraunhofer's line E. 



In order to avoid errors similar to those which I have exposed 

 in the foregoing, and for the purpose of facilitating the nume- 

 rical comparison of the data of various spectrum-observations, 

 we give in Plate V. graphical representations of the observations 

 which are taken from the guiding lines given in the chromo- 

 lithograph drawings of the spectra published in our former 

 memoirs and in which the prism was placed at the angle of 

 minimum deviation. The ordinates of the edges of the small 

 blackened surfaces, referred to the divisions of the scale as ab- 

 scissas, represent the intensity of the several lines with their 

 characteristic gradations of shade. These drawings were made 

 when the slit was so broad and the flame of such a temperature, 

 that the fine bright line upon the broad band Ca a began to be 

 distinctly visible. This breadth of the slit was equal to the fortieth 

 part of the distance between the sodium line and the lithium line a. 

 Por the sake of perspicuity, the continuous spectra which some 

 bodies exhibit are specially represented on the upper edge of the 



