20 MR. W. DODGSON ON A 



ard instrument, which would be retained at a central ob- 

 servatory ; all other instruments would, before being used, 

 be compared with it. It would thus be possible to compare 

 together the indications of various instruments working in 

 different places, provided that these, before being issued, 

 had their coefficients determined at the central observatory. 



On a Graphical Method of Drawing Spectra. By Mr. 

 William Dodgson, Whitworth Scholar. Communi- 

 cated by Professor Roscoe, F.R.S. &c. 



Read March aist, 1876. 



Construction of Curve for ascertaining the Wave-lengths. 



A NUMBER of points are first plotted on the curve-paper, 

 the abscissae of these points being the micrometer-scale- 

 readings of certain lines as observed directly in the spec- 

 troscope, or being absolute or relative measurements taken 

 from an existing photograph ; whilst the ordinates of these 

 points are the corresponding wave-lengths, taken from the 

 determinations of some reliable observer. 



If there be a large number of points taken, the curve 

 ought not to be drawn actually through them all, or it 

 would probably be irregular; it therefore becomes necessary 

 to draw a mean curve, i. e. a continuous even curve which 

 shall pass near to all the given points, some being on one 

 side and some on the other. 



The following method was employed to determine the 

 wave-lengths of the well-defined edges of the bands in two 

 absorption - spectra . 



