Knowledge. 



With which is incorporated Hardwiclte's Science Gossip, and the Illustrated Scientific News. 



A Monthly Record of Science. 



Conducted by Wilfred Mark Webb, F.L.S., and E. S. Grew, M.A. 



MARCH, 1915. 



CHAPTERS IN SPECTRUM ANALYSIS. 



By W. MARSHALL WATTS, D.Sc. 



L — Law and Order in Spectra. 

 A. Line Spectra. 



(Continued from page 37.) 



Series in line spectra, similar to that of hydrogen 

 represented in Balmer's law, are found in many 

 spectra, such regularities being the rule and not 

 the exception. The most extensive of such series 

 is that observed in sodium, and known as the 

 Principal Series. The well-known yellow lines 

 seen in the spectrum of a sodium flame constitute the 



first term of this series — the double hne {sQno.iof- 

 The second term of the series is the double line 



and no 



/3303-07\ ^, .K- . . ■ /2853-041 

 \3302-47/ '• ^^^ ^^'^ *^™ ^' 12852-84/ 

 doubt all the other terms are double also, though the 

 pairs may be too close to be separated. It is well 

 known that the sodium lines are easily reversed, 

 being then seen as dark lines, e.g., the Fraunhofer 

 lines Dj and D.^ in the solar spectrum ; and 

 Professor Wood, by observing the absorption 

 spectrum of sodium vaporised in an iron tube 

 heated to a dull red-heat, has extended this series 

 of lines from the seven terms previously known in 

 the laboratory to no fewer than forty-eight terms. 

 This extensive series is represented in the diagram 

 (see Figure 60). 



It is a question of much interest to examine 

 whether the wonderful " law and order " exhibited 

 by the hydrogen spectrum, and shown by the 

 exact agreement of the observed wave-lengths 

 with those calculated from Balmer's formula, can 

 be traced in this still more extended series. We 

 see that the case is now somewhat less simple than 

 with hydrogen. In the first place, we have now to 



deal with a series of double lines, or rather with two 

 series of single Unes ending at the same convergence- 

 frequency, since the Hues of the pairs become closer 

 and closer together as we pass from red towards 

 blue. In the next place, we soon find that the 

 law of the series is not so simple as that of the 

 hydrogen series, and therefore requires a more 

 complicated formula for its expression. The most 

 satisfactory formula appears to be that employed 

 by Mogendorff in 1906, and by Hicks in 1910, 

 namely, 



O.F, = C,F.- '°*'5 



("■ + '' + ^) 



Balmer's formula for hydrogen, it will be 

 remembered, is a simpler form of this, namelv, 



^^ ^^ 109675 

 O.F. = C.F.--^ 



in which m is put equal to 1,2, 3, and so on, suc- 

 cessively, and n and c are constants. Since we have 

 to deal with pairs of lines, we must have two 

 formulae, one for the less refrangible and the 

 other for the more refrangible lines of the pairs. 

 The best values of the constants are : — 



For the less refrangible hne : 



^ = 0-U7408. c= -031328. 



For the more refrangible line : 



^ = 0-148204. c= -031380. 



The convergence - frequency, C.F., has the same 

 value for both components, namely, 41 448-67. The 



65 



