94 



SCIENCE. 



[N. S. Vol. XIX. No. 472. 



Preclit, series of groups of lines are had 

 which satisfy mathematical formula. 



"It was proposed by de Gramont, at the 

 International Congress in Paris, in 1900, 

 and agreed, that no new substance should 

 be described as an element until its spark 

 specti'iun had been measured and shown to 

 be different from that of every other known 

 form of matter. ' ' As Hartley * remarks, 

 'This appears to me to have been one of 

 the most important transactions of the con- 

 gress. ' Radium f was the first to be tested 

 by this rule. Exner and Haschek obtained 

 1,193 spark and 257 arc lines for Demar- 

 cay's europium. It must not be forgotten, 

 however, that by overlapping, lines in mix- 

 tures may be masked or appear, which are 

 absent, in those bodies of the highest state 

 of purity. It must not be forgotten that 

 pressure influences the spectrum, usually 

 producing a broadening of the lines, as 

 shown by Schuster, and that it may occur 

 symmetrically or only towards the least re- 

 frangible red. Lest we forget, the spectro- 

 scope failed a long time to show radium and 

 we knew it was there. It must not be for- 

 gotten, as Kriissg has shown, that the "'in- 

 fluence of temperature can not be neglected 

 and ignored, but must be considered by 

 every chemist who wishes to make correct 

 spectroscopic observations.' It is well 

 known to spectroscopists that band spectra 

 are obtained at temperatures intermediate 

 between those required for the prodiiction 

 of continuous spectra and line spectra. || 



■* Address before the Chemical Section of the 

 British Association, Southport, 1903. 



t Runge and Preeht, Am. Physik., TV., 12, 407, 

 190.3. 



X British Association, Report, 1880, 275. Vide 

 also Lockyer and Frankland, Proc. Roy. Soc, 27, 

 288, 1869. 



I ' The Influence of Temperature upon the 

 Spectrum; Analytical Observations and Measure- 

 ments,' LieUg's Annalen, 238, 57; Ghem. News, 

 56, 51. 



II ' Spectrum Analysis,' Landauer, English trans- 

 lation by Tingle, p. 70. • 



The explanations of these facts do not con- 

 cern us at present. 



It has been shown by the researches of 

 Newton, Dale, Gladstone, Jamin, Schrauff, 

 Landolt and others that the refractive 

 power increases in all liquids, except in 

 water, between 0° and 4° with the increase 

 of density — that is, with decrease of tem- 

 perature. Rydberg showed that various 

 solid bodies, such as quartz and aragonite, 

 follow the same law. There are some ex- 

 ceptions, however. Among these is glass, 

 as proved by Arago and Neumann prior to 

 Rydberg. "On a rise of temperature all 

 phenomena of absorption or emission are 

 displaced toward the violet with the glass 

 , prisms, but toward the red with quartz 

 prisms. These displacements are the great- 

 er the more refrangible the region of the 

 spectrum in which they occur." As the 

 result of a large number of observations, 

 Kriiss learned that by a variation of 25°, 

 marked changes would be observed in the 

 spectroscopic lines. From a table given, it 

 could be seen that errors may spring from 

 neglect of the temperature (of the instru- 

 ment?) in stating wave-lengths, since a 

 rise of 5° is sufficient to transfer the D^ to 

 the position D„. Roscoe obtained an en- 

 tirely new spectrum with the metal sodium, 

 whereby it appears that this metal exists in 

 a gaseous state in four different degrees of 

 aggregation, as a simple molecule, and as 

 three or four or eight molecules together. 

 Griinwald in a series of papers on his 

 theory of spectrum analysis* endeavors "to 

 discover relations between the spectra and 

 thus to arrive at simpler, if not funda- 

 mental 'elements.' " He came to the con- 

 clusion that 'all the so-called elements are 

 compounds of the primary elements a and 

 * ' Uber das Wasserspectrura, das Hydrogen — 

 und Oxygenspeetrum,' Phil. Mag., 24, 304, 1887. 

 ' Math. Spectralanalyse des Magnesiums vmd der 

 Kohle,' Monatshefte fur Chemie, 8, 650. 'Math. 

 Spectralanalyse des Kadmiums,' Monatshefte fiir 

 Chemie, 9, 956. 



