ON OUR KNOWLEDGE OF SPECTRUM ANALYSIS. 317 



Report of the Committee, consisting of Professor DE^^^AK, Dr. 

 Williamson, Dr. Marshall Watts, Captain Abney, Mr. Stoney, 

 Professor W. N. Hartley, Professor McLeod, Professor Carey 

 Foster, Professor A. K. Huntington, Professor Emerson Key- 

 NOLDS, Professor Eeinold, Professor Liveing, Lord Eayleigh, 

 Dr. Arthur Schuster, and Mr. W. Chandler Egberts 

 (Secretary), appointed for the purpose of reporting upon the 

 present state of our Knoivledge of Spectrum Analysis. 



[Plate V.] 



General Methods of Obseeving and Mapping Spectka. 

 By W. Marshall Watts, B.Sc. 



The best method of measuring and mapping a spectrum must, of course, 

 depend on the object with which the spectrum is observed. If the 

 spectroscope is employed only as an auxiliary to the ordinary methods of 

 chemical analysis, and the object is simply to detei'mine the presence or 

 absence of a metal of the alkalies or alkaline earths — say lithium or 

 calcium — -very rough measurement only is needed ; indeed, in most cases, 

 the colour of the line or the general appearance of the spectrum is 

 sufficient. But if, on the other hand, the object is, for example, the 

 determination of the presence or absence of oxygen in the sun's 

 atmosphere, or the description of some new spectrum observed for the 

 first time, the case is altogether different ; the greatest dispersive power 

 that the circumstances of the case will allow must be employed, and 

 the position of each line must be measured with the utmost accuracy 

 attainable by the best use of the best apparatus at command. 



The spectrum may, of course, be produced either by difiPraction from 

 a diffraction-grating or by refraction through a prism. The splendid 

 diffraction-gratings furnished by Rutherford give results unapproached 

 by any other means when the source of light is sufficiently powerful ; but 

 the intensity of a diffraction spectrum is always so much less than that 

 of a dispersion spectrum that for most purposes of spectroscopy the 

 prism must be employed. 



For the ordinary purposes of chemical analysis, nothing can be better 

 than a strongly-built spectroscope, provided with one prism of 60° of 

 dense glass, and a photograph-millimetre scale, seen by reflection at the 

 first surface of the prism. 



It is not possible to construct instruments with exa,ctly similar scales, 

 and each instrument should therefore have its readings reduced to wave- 

 lengths by the method of graphical interpolation, to be presently 

 described ; but it is convenient to have these reflected scales as nearly as 

 possible similar to the one given in Bunsen's first paper.' On this scale 

 the Fraunhofer lines have the following positions : — 



A 17-5 B 28-9 C 35-0 D .50 E 709 i 75 F 90 

 G 127-3 H, 161-2 H2 1657 



and the Lithium, Strontium, and Thallium lines are as follows : — 



Li 31-5 Sr g 105-5 and Tl 67-8. 



' Phil. Mag. (Fourth Series) voL xxvi. p. 247. 



