February, 1903.] 



KNOWLEDGE 



31 



It will be seen that not only is there a coiucideuce in 

 position of the bright with dark lines, but also a remarkable 

 agreement in the intensities of the two sets of lines— 

 the brighter the bright line the more intense the dark 

 one. 



It is quite evident that the chemistry of the heavenly 

 bodies can only be investigated by combining the work 

 of the laboratory with that of the observatory. In the 



I to —1 



Liall part of the Spectrum of the Sun compared with tliat of [ 



laboratory the spectrum of every known substance must 

 first be carefully mapped out, and the lines tabulated, for 

 ready comparison with the spectra encountered in the 

 observatory. In either case, the most convenient mode of 

 procedure is to record the spectra photographically, as 

 in the two illustrations already given. AVhen a spectrum 

 has been once photographed, it can be re-examined at any 

 time without further experimental preparations, and in 

 some cases, as in that of Nova Persei, a permanent record 

 is secured when the opportunity only occurs once. In 

 dealing with faint objects, the photographic method has 

 the additional advantage of integrating successive im- 

 pressions, so as to effectively increase their apparent 

 brightness, whereas the eye can only take account of 

 momentary effects of the rays of light, and does not see the 

 lines any brighter by continued observation. 



It is scarcely necessary to explain in detail here the 

 methods employed in the practical observation and 

 mapping of spectra, but there is one point of fundamental 

 importance upon which a few remarks may not be super- 

 fluous. Although experience may 

 render the observer so familiar with 4 

 the spectra of different substances 

 that he can often identify lines by 

 mere inspection, careful measurement 

 can usually alone give satisfactory 

 results. It is evidently necessary, 

 therefore, to adopt some system of 

 stating the positions of the lines 

 by which chemical substances are 

 identified, which shall be applicable 

 to spectra observed with any in- 

 strument whatsoever, in ordcu- that, 

 all records may be strictly compar- 

 able. 



In practice, the jiositioiis are stated 

 lengths, the number attached to each line tlien repre- 

 senting the length of the waves of light which produce 

 the Hue. These waves are excessively small, but they can 

 nevertheless be measured with great accuracy by the use 

 of dift'raction gratings. They are so small that the most 



I 2 



convenient unit of wave-length — often spoken of as an 

 " Angstrom Unit " — is the ten-millionth of a millimetre, 

 or " tenth metre," and in this system, the visible spectrum 

 ranges from about 3930 in the violet to 7600 in the red 

 end of the spectrum. Thus, when it is stated that the 

 wave-length of the D3 line of helium is 5875'98 it is 

 understood that it is that number of times a ten-millionth 

 of a millimetre. It is not necessaiy, however, to directly 

 determine a wave-length every time 

 3 that an observation is made, and 



5 indeed this cannot be done when 



5 a prismatic spectroscope is em- 



ployed, as is most frequently the 

 case. Certain standard wave-lengths 

 have been carefully measured, and 

 it is usual to determine any required 

 wave-length 'from these either by 

 inspection or by interpolation. The 

 most generally useful of these 

 reference lines are those given in 

 Rowland's " Tables of Solar Wave- 

 lengths,"* which state with great 

 accuracy the wave-lengths of many 

 thousands of the dark lines which 

 appear in the spectrum of the sun 

 or daylight. In addition to the 

 Tables, Prof. Rowland published a 

 photographic map having a scale 

 by which the wave-lengths can be 

 read off directly with considerable accuracy. Rowland's 

 map is said to be a " normal " one, for the reason 

 that equal distances on the map correspond to equal 

 differences of wave-length, which is not the case 

 with prismatic spectra. A small portion of this map, 

 corresponding in part with the photograph in Fig. '2, is 

 reproduced in Fig. 3, and by comparing the two it will be 

 seen that, although the photographs are not on the same 

 scale, the wave-length of any required line, brisjht or dark, 

 in Fig. 2 may be read off by the scale in Fig. 3. For 

 example, the three very strong lines of iron, forming a 

 " triplet," near the left of Fig. 2 are clearlv identical with 

 the lines 4045-98, 4063-76, and 4071-91 in the solar map. 

 Having a complete set of such maps, it is evidently an 

 easy method of obtaining wave-lengths to photograph the 

 spectrum of sunlight alongside that of the spectrum under 

 examination. Sometimes, however, it is more convenient 

 to use the spectrum of iron as a comparison, in which 

 case reference might be made to the admirable wave- 

 length map of this spectrum published by Messrs. Kaysor 



41 



A. portion of the Solar Spectrum, with Wavo-longtli Scale. (Rowlaml.) 



>u a scale 



of Wiive- 



aud Ruuge.t 



When 110 comparison spectrum has been photographed 

 in juxtaposition with that under investigation, and this is 



• PubUshed by the University ot Chie&go Press, 1896. 

 + All/,, d. Akad. d. M'i.is. zu Berlin, 1888. 



