1889.] of the various Species of Heavenly Bodies. 227 



safely that we are not dealing with the line spectrum of oxygen. 

 We have previously seen that it is not the negative-pole spectrum of 

 oxygen. 



In the same table (' L'Aurore Boreale,' p. 92), the aurora lines are 

 compared by Lemstrom with some of the lines or bands observed by 

 himself in the spectrum of rarefied air. The air lines which he gives 

 all agree in position with some of the nitrogen flutings mapped by 

 Hasselberg (' Memoires de 1'Academie Imperiale de St. Petersbourg,' 

 Series 7, vol. 32, No. 15). One of them is at wave-length 558, and this 

 he believes to be coincident with the aurora line 557. The intensity of 

 the line is not given, but Hasselberg gives it as a comparatively feeble 

 fluting at 557 (see fig. 16). Considering the absence of the brightest 

 nitrogen flutings from the spectrum of the aurora, the supposed 

 coincidences between some of Lemstrom's rarefied air lines and lines 

 in the aurora spectrum, which are far from perfection, may be dis- 

 regarded. 



The same objections apply to the lines in the rarefied air spectrum 

 which have been recorded by Sundell ; those which fall anywhere 

 near lines in the aurora are comparatively faint flutings or lines in 

 the spectrum of nitrogen ; at all events, flutings of the same or 

 greater intensities are absent, and there is no evidence to show that 

 the coincident ones retain their brightness as the others fade. 



Lemstrom then leaves the origin of the aurora spectrum as uncer- 

 tain as ever. There is no evidence to show that it is a spectrum of 

 air, or, indeed, of any other gas. If it be a spectrum of air, it is one 

 which has never been obtained experimentally, and one which can 

 only be put forward by making unphilosophical assumptions and care- 

 fully avoiding experiments. 



III. Gyllenskiold? s Observations and Conclusions. 



Still later observations of the aurora .which have been published 

 are those made at Cape Thordsen by M. Carlheim-G-yllenskiold.* Two 

 lists of lines are given, one from observations made with a Hofmann 

 spectroscope, and the other from observations made with a Wrede 

 spectroscope. The lines in the first list extend from blue to red, and 

 those in the second list from green to violet. The individual observa- 

 tions of different auroree with the lines observed in each are given. 

 36 auroras are recorded in which only 1 line was visible, 15 in which 

 there were only 2 lines, 6 with 3 lines, 15 with 4 lines, 5 with 6 lines, 

 4 with 7 lines, 1 with 8, 1 with 9, and 1 with 10 lines, so that 

 altogether, no less than 84 observations are recorded. 



The total number of lines seen were 32. Gyllenskiold's main con- 

 clusions are : 



* ' Observations faites au Cap Thordsen, Spitzberg, par 1'Expedition Suedoise.' 

 Vol. 2, 1. "Aurora Borealis," par Carlheim-G-yllenskiold. 



