548 



NATURE 



[October 3, 1895 



en bloc, \ar>" enormously in the sun. the appearances 

 being — 



4922 (4921-3) = 30 times. 



4713(4712-5) = twice. 



" These are not the only facts which can be adduced to 

 sujigest that the gasfrom cl&veite is as complex as that from 

 broggeritc, but while, on the one hand, the simple nature 

 of the gases obtained by Profs. Ramsay and Cl^ve, and 

 by myself, must be given up, reasoning on spectroscopic 

 lines, the observations I have already made on several 

 minerals indicate that the gases composing the mixtures 

 are by no means the only ones we may hope to obtain." 



It will be seen that the laboratory separation of D3 

 from the lines 504S, 5016, and 4922 was complete, and 

 we now know that they belong to different series. 



These lines have now been differentiated by Runge and 



Paschen by a different but equally satisfactory method. 

 Nor is this all. The difference between the results 



obtained by Thalen and myself seemed susceptible of 

 i explanation by admitting a fractional distillation, accord- 

 I ing to which D3 and 447 came oft" first, and 4922, 5016, 

 t and 667 later on (Fig. 2). 

 j Here also 1 got the same result as in the diffusion 



experiment referred to by Drs. Runge and Paschen. 



They found similarly — 



Less bright. More bright. 



D3 5016 



6678 



.\11 these various lines of evidence tend therefore to 

 complexity, and there can be little doubt from the con- 

 vertrcnce of all these lines of work, the results of which 



FtG. I. — Diagram showing changes in intensities of lines brought about by var>-ing the tension of the spark, 

 (i) Without air-break. (2) Witn air-bre.ik. 



Later on, in the same month, I returned to this subject, 

 and showed that the lines at D3 and 447 behaved in one 

 way, and that at 6f)7 behaved in another. 



I wrote as follows ' : — 



■• I J In a simple gas like hydrogen, when the tension of 

 the electric current given by an induction coil is increased, 

 by inserting first a jar, and then an air-break into the 

 circuit, the effect is to increase the brilliancy and the 

 breadth of all the lines, the brilliancy and breadth being 

 greatest when the longest air-break is used. 



" 12) Contrariwise, when we are dealing w ith a known 

 compound gas ; at the lowest tension we may get the 

 complete spectrum of the compound without any trace 

 of its constituents, and we may then, by increasing the 

 tension, gradually bring in the lines of the constituents, 

 until, when complete dissociation is finally reached, the 

 spectrum of the compound itself disappears. 



agree among themselves, that we are in presence of at 

 least two distinct gases, the complete spectra of which are 

 those given by Drs. Runge and Paschen. 



The second point is that there is no connection what- 

 ever between either of these gases and argon. Argon is 

 of the earth, earthy, but helium and gas .\ are distinctly 

 celestial, even more celestial than 1 thought when 1 

 claimed for them last May' the dignity of "a new order 

 of gases of the highest importance to celestial chem- 

 istry'." It was supposed at first that the spectra con- 

 tained any number of common lines, next that there were 

 two coincidences in the red between the new gases and 

 argon ; one I found broke down with moderate dispersion, 

 the other has yielded to the still greater dispersion 

 employed by Drs. Runge and Paschen ; and, more than 

 this, 1 have not found a single coincidence between argon 

 and any line in the spectrum of any celestial body what- 



♦9fJ0l 



D, 



3S76. 



667. 



Fit;, a. — Diagram showing the order in which the h'nc^ .nppcar in spectrum when hrAggcrite is heated. 



" Working on these lines, the spectrum of the spark at 

 atmospheric pressure, passing through the gas, or gases, 

 distilled from broggerite, has been studied with reference 

 to the special lines C (hydrogen), 1)3, 667, and 447. 



" The first result is that all the lines do not vary equally, 

 as they should do if we were dealing with a simple gas. 



"The second result is that at the lowest tension 667 is 

 relatively more brilliant than the other lines ; on mcrcas- 

 in^; the tension, C and D, considerably increase their 

 brilliancy, 667 relatively and absolutely becoming more 

 feeble, while 447, seen easily as a narrow line at low 

 tension, is almost broadened out into invisibility as the 

 tension is increased in some of the lubes, or is greatly 

 brightened as well as broadened in others (F"ig. i). 



' /"rr-r. kny. Soc., vol. Kiii. p. 193. 



NO. 1353, VOL. 52] 



ever. This happens, as c\erybody knows, also in the 

 case of oxygen, nitrogen, chlorine, and the like. - 



The third point is as follows. -So far 1 have worked 

 upon some eighty minerals, and 1 have found the 

 yellow line in sixteen ; among the lines which 1 have 

 already reported 10 the Royal .Society are included all 

 the stronger ones in the various series determined by the 

 (ierinan physicists, but I can now add that in the region 

 over which my work has extended, there is scarcely a 

 single line in their series which I have not cither seen or 

 photographed in the spectrum of some celestial l)ody or 

 another. The following tables will show the results I 

 have already obtained with all the six series of lines 

 indicated by Drs. Runge and Paschen. 



* I*ivc. Roy. Soc., vol. Iviii. p. 117. 



