April 5, 1906] 



LETTERS TO THE EDITOR. 



NA TURE 



553' 



[Tlie Editor does not hold himself responsible for opinions 

 expressed by his correspondents. Neither can he undertake 

 to return, or to correspond with the writers of, rejected 

 manuscripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Spectral Series in Relation to Ions. 



As 1 have already shown (Physik. Zeitsclir., vi., 892, 

 1905), the observation of the Doppler effect on the Kanal- 

 strahlen permits the detection of spectra emitted by the 

 positive ions (Kanal-strahlen). I have made such re- 

 searches on Kanal-strahlen in hydrogen, potassium vapour, 

 and mercury vapour in cooperation with Messrs. S. 

 Kinoshita, K. Siegl, and W. Hermann. We shall give 

 details of our methods and measurements in separate 

 papers; but here I wish to state the principal results of the 

 res. arches, together with some general conclusions. 



I he series of lines (Ha, H/3 . .) of hydrogen is a first 

 subordinate series. Its lines are revealed by accurate 

 analysis as pairs or doublets. The difference of wave- 

 lengths of the two components is, as Michelson has found 

 (Phil. Mag., xxxiv., 280, 1892), 0-14 Angstrom unit for 

 Ha and 0-08 Angstrom unit for H/3 ; the measurements 

 made by Ebert (Wied. Ann., xliii., 790, 1S91) give 0-132 

 Angstrom unit for Ha. In every other first subordinate 

 series of doublets, and also in the case of hydrogen, the 

 difference in oscillation frequencies of the components of 

 the doublet is constant throughout the series ; this differ- 

 ence is for Ha 033, for Hj3 0-34 per 1 cm. path in vacuo. 

 My previous and recent observations lead to the conclusion 

 that the first subordinate series of doublets of hydrogen 

 has as carriers monovalent positive atom-ions, i.e. atoms of 

 hydrogen which have lost a single negative electron. 



A second subordinate series of lines of the hydrogen 

 has been observed in certain stars. From their spectral 

 position, Rydberg (Astrophys. ]ourn., vi., 233, 1897) has 

 calculated the principal series of hydrogen ; he gives to 

 the first line of it the wave-length 468788 Angstrom; this 

 line has been observed in stars with bright lines. It is 

 found also (somewhat displaced) in all spectrograms I have 

 taken of the kathode rays or Kanal-strahlen, partly alone 

 and partly in cooperation with Mr. Kinoshita. This line of 

 the principal series — it may be termed H/3 — shows also the 

 Doppler effect in the Kanal-strahlen, the quantity of the 

 effect being the same as for the first subordinate series of 

 doublets ; the principal series of the hydrogen, which is also 

 composed probably of doublets, has therefore the same 

 carrier as the first subordinate series, namely, the mono- 

 valent positive ion of hydrogen. 



In cooperation with Mr. Siegl I have further examined 

 another doublet of a principal series, namely, the second 

 doublet of the principal series of potassium (A 4047-36- 

 404429). Both components show the Doppler effect in the 

 Kanal-strahlen, and the amount is the same as that calcu- 

 lated for an atom of potassium which has lost a single 

 negative electron. Therefore, in the case of the alkali- 

 metals also, the principal series of doublets has monovalent 

 positive ions as carriers. 



In the spectrum of mercury hitherto only series of 

 triplets have been found — a first and a second subordinate 

 series (Kayser and Runge). Using a small concave grating 

 kindly lent to me by Prof. Runge, I succeeded, in cooper- 

 ation with Mr. Hermann, in examining the Doppler effect 

 on the lines of mercury. It was found that all components 

 of triplets, a.nd further all triplets of a series, have the 

 same positive ion as carrier, and, moreover, the lines of 

 the first and of the second series show the same Doppler 

 effect in type and quantity. Both series of triplets of 

 mercury have therefore the same carrier, the bivalent 

 positive ion of mercury; for the Doppler effect really found 

 agrees in quantity with the effect calculated theoretically 

 for an atom of mercury which has lost two negative 

 electrons. 



The lines of mercury not belonging to the series of 

 triplets show likewise a Doppler effect, but there is a 



NO I9OI, VOL. 73] 



difference in character and amount between them and the 

 lines of the two series of triplets. There are lines which 

 show a larger effect than the series of triplets ; the carrier 

 of these lines is therefore a positive ion of mercury of 

 higher valency, i.e. an atom of mercury which has lost 

 more than two negative electrons. To these lines of 

 higher valency belongs the line A 4078-1. 



Finally, there are lines in the spectrum of mercury which, 

 show a smaller Doppler effect than the lines of the series- 

 of triplets; their displacement is roughly 1-5 times smaller 

 than that of a line of a triplet of equal wave-length.. 

 Therefore they have as carrier not a bivalent, but a mono- 

 valent positive ion of mercury. To these lines belong the- 

 wave-lengths AA 233672, 433947-302164, 3984-08-284785. 

 It seems that X 253672 is the first component of a principal 

 series of doublets, that the pair AA 4339-47-3021-64 belongs, 

 to a first, and the pair AA 3984 08-284785 to a second sub- 

 ordinate series of doublets. This being so, mercury con- 

 firms the view that the principal and the first and second 

 subordinate series of doublets have monovalent positive 

 ions as carriers. 



Generalising the foregoing results, we have come to the- 

 following conclusions : — The carriers of the spectra of lines 

 of the chemical elements are positive atom-ions. All lines of 

 a series have the same carrier, and, moreover, the same 

 carrier may emit several series at the same time. The 

 carrier of the principal series and of the subordinate series 

 of doublets is a monovalent positive atom-ion ; the carrier 

 of the subordinate series of triplets is a bivalent positive 

 atom-ion ; ions of a higher valency emit likewise line 

 spectra, but the structure of these is not yet recognised.. 

 The spectrum of an element, for example, that of mercury, 

 may represent a mixture of several spectra, namely, of the 

 spectra of its monovalent, bivalent ions, and of ions of 

 higher valency. 



The foregoing results and conclusions are in striking-, 

 agreement with the results which were arrived at by 

 Runge and Paschen (Ber. d. Berliner Akad., 1902, 3S0, 

 720) in their researches on the Zeeman effect of series of 

 lines. They found that the principal series of doublets of 

 all elements examined (Na, Cu, Ag, Mg, Ca, Sr. Ba) show 

 in type and amount, when measured in oscillation fre- 

 quencies, the same Zeeman effect ; this also holds good for 

 the first and the second series of doublets. The first and 

 second subordinate series of triplets show in type and 

 amount another Zeeman effect than the series of doublets ; 

 but the series of triplets of different elements are again 

 in the same way broken up by a magnetic field. The 

 agreement of my results with those of Runge and Paschen 

 comes out in the following detail : — The lines of mercury 

 (AA 2536-72, 433947, and 2847-85) referred by me to series of 

 doublets must show in a magnetic field the known splitting- 

 up of the lines of the principal and of the subordinate 

 series of doublets. Runge and Paschen enumerate these- 

 lines under those which do not show the behaviour of 

 triplets, and, in fact, their statements on the magnetic 

 behaviour of those lines are concordant with that postulate- 



It follows from the Zeeman effect that the centres of 

 emission of series of lines are periodically accelerated 

 negative electrons. From the complexity of their mag- 

 netic splitting up we may draw the conclusion that these 

 centres of emission — the negative electrons — are coupled 

 in electrodynamic systems ; the electrodynamic structure 

 of these systems of negative electrons is for the emission of 

 series of doublets rather than for the emission of series of 

 triplets. 



We do not know of spectra of neutral atoms. It follows 

 from the foregoing results that the known spectra of lines 

 can only be emitted if the chemical neutral atoms have lost 

 negative electrons, and thus have become positive atom- 

 ions. Therefore the electrodynamic symmetry of the system 

 of negative electrons in the positive atom-ion is different- 

 from the symmetry in the neutral atom. Certain systems 

 of negative electrons have in the positive monovalent ion 

 an electrodynamic symmetry which enables them to emit- 

 radiation of electromagnetic energy ; this symmetry is 

 characterised by the emission of doublets. Losing two or 

 more negative electrons a neutral atom also gains an 

 electrodynamic symmetry capable of radiation ; but the- 



