SCIENCE. 



221 



This method of inquiry has been tried also with potassium, 

 calcium, and some other metals, and with metallic salts. 



With potassium and calcium we get the same inversion 

 of phenomena, the yellow-green lines of potassium being 

 seen without the red ; while in the case of calcium the blue 

 line alone was seen. 



The fact that in these experiments we get, as before men- 

 tioned, vapors which at one and the same time exhibit dif- 

 ferent colors and different spectra at different levels in the 

 tube, at once suggests the phenomena of fractional distilla- 

 tion. 



It is also suggested, as a result of the application of this 

 new method, that in the case of a considerable number of 

 chemical substances not only the line spectrum is com- 

 pound in its origin, as I suggested many years ago, but that 

 a large number of the lines is due to molecular groupings 



To take an instance, the rlame spectrum of sodium gives 

 us, as its brighest, a yellow line, which is also of marked 

 importance in the solar spectrum. The-flame spectra of 

 lithium and potassium give us, as their brightest, lines in 

 the red which have not any representatives among the 

 Frannhofer lines, although other lines seen with higher 

 temperatures are present. 



Whence arises this marked difference of behavior? 

 From the similarity of the rlame spectrum to that of the sun 

 in one case, and from the dissimilarity in the other, we may 

 imagine that in the former case — that of sodium — we are 

 dealing with a body easily brokeu up, while lithium and 

 potassium are more resistant ; in other words, in the case 

 of sodium, and dealing only with lines recognized gener- 

 ally as sodium lines, the rlame has done the work of disso- 

 ciation as completely as the sun itself. Now it is easy to 



Fig. 3. — Position of Spectioscope for discovering Vapors cLse to die .Metal. 



of considerable complexity, which can be kept out of^re- 

 action by careful low temperature distillation. 



So much then for one method. Now for the other. 



In this I have attempted to gain new evidence in the re- 

 quired direction by adopting a method of work with a spark 

 and a Bunsen flame, which Col. Donnelly suggested I 

 should use with a spark and an electric arc. This consists 

 in volatilizing those substances which give us tiame spectra 

 in a Bunsen flame and passing a strong spark through the 

 llame, first during the process of volatilization, and then 

 after the temperature of the flame has produced all the sim- 

 plification it is capable of producing. 



The results have been very striking ; the puzzles which a 

 comparison of flame spectra and the Frannhofer lines has 

 presented us find, I think, a solution ; while the genesis of 

 spectrais made much more clear.' 



1 1 allude more especially to the production of triplets, their change 

 into quartets, and in all probability into Hutings, and to the vanishing .of 

 flutings into lines, by increasing the rate of dissociation. 



test this point by the method now under consideration, for 

 if this be so then (i) the chief lines and flutings of sodium 

 should be seen in the flame itself, and (2) the spark should 

 pass through the vapor after complete volatilization has 

 been effected without any visible effect. 



Observation and experiment have largely confirmed these 

 predictions. Using two prisms of 6o° and a high-power 

 eyepiece to enfeeble the continuous spectrum of the densest 

 vapor produced at a high temperature, the green lines, the 

 timings recorded by Roscoe and Schuster, and another 

 coarser system of Hutings, so far as I know not yet de- 

 scribed, are beautirully seen. I say largely, and not com- 

 pletely, because the double red line and the lines in the 

 blue have not yet been seen in the llame, either with one, 

 two, or four prisms of 6o°, though the lines are seen during 

 volatilization \i a spark be passed through the llame. Sub- 

 sequent inquiry may perhaps.show that this is due to the 

 , sharp boundary of the heated region, and to the fact that 

 j lines in question represent the vibrations of molecular 



