POTASSlfM. KlT.IDIl'M, CAKSIT.M, AND LITHIUM 557 



spending with one of the lines between G and H. Of these rays the red 

 is the brightest, and therefore the general colour of luminous hydrogen 

 (with an electric discharge through a Geissler tube) is reddish. 



The correlation of the Frauenhofer lines with the spectra of metals 

 depends on the phenomenon of the so-called reversed spectrum. This 

 phenomenon consists in that instead of the bright spectrum correspond- 

 ing with a metal, under certain circumstances a similar dark spectrum 

 in the form of Frauenhofer lines may be obtained, as will be directly 

 explained. In order to clearly understand the phenomenon of reversed 

 spectra, it must be known that in the passage of light through certain 

 transparent substances these substances retain rays of a certain re- 

 frangibility. The colour of solutions is a proof of this. Light which 

 has passed through a yellow solution of a uranium salt contains no 

 violet rays, and after having passed through a red solution of a per- 

 manganate, does not contain many rays in the yellow, blue, and green 

 portions of the spectrum. Solutions of copper salts absorb nearly all 

 red rays. Sometimes colourless solutions also absorb rays of certain 

 definite refractive indexes, and give absorption spectra. Thus solu- 

 tions of salts of didymium absorb rays of a certain refrangibility, and 

 therefore an impression of black lines is received, 28 as shown in fig. 



spectra (see later). Consequently, these lines are the most characteristic. Only the 

 longest and most brilliant are given in our table, which is composed on the basis of a 

 collection of the data at our disposal for bright spectra of the incandescent and rarefied 

 vapours of the elements. As the spectra change with great variations of temperature 

 and vapour density (the faint lines become brilliant whilst the bright lines sometimes 

 disappear), which is particularly clear from Ciamician's researches on the halogens, there- 

 fore, until the method of observation and the theory of the subject are enlarged, par- 

 ticular theoretical importance should not be given to the wave lengths showing the 

 maximum brilliancy, and which only possess any significance in a practical respect for 

 the common methods of spectroscopic observations. 



!8 The ocular impressions of light (it is essentially the same with all other impressions 

 received by the senses) are all relative ; in those portions of a spectrum, received through 

 an absorptive medium, where there appears to be an absence of light, it may be only 

 rendered fainter, and for absorption spectra this is directly proved to be the case both 

 by experiment (by employing solutions of different strengths or strata of different thick- 

 nesses), and by direct measurement by the aid of the spectroscope for instance, by 

 Vierordt's apparatus, which is described in works on physics. The relative distinctness 

 of the dark lines in an absorption spectrum, and of the bright coloured lines in luminous 

 spectra of vapours and gases, which are self-evident in making observations, offer great 

 difficulty with regard to precise measurement, just as is the case, for instance, with the 

 relative brilliancy of the stars. 



The method of observing absorption spectra consists in taking a continuous spectrum 

 (one which does not give either dark lines or particularly bright luminous bands in the 

 spectrum) of white light for instance, the light of a candle, lamp, or other source. The 

 collimator (that is, the tube with the slit) is directed towards this light, and then all 

 the colours of the spectrum are visible in the ocular tube. A transparent absorptive 

 medium for instance, a solution or tube containing a gas is then placed between the 

 source of light and the apparatus (or anywhere inside the apparatus itself in the path of 

 the rays). In this case either the entire spectrum is uniformly fainter, or absorption 



