42 SECTIONAL ADDRESSES. 



describing this work it was suggested as a possibility that this was the 

 origin of the red auroral line, and somewhat similar ideas have been 

 revived by McLennan in his recent Bakerian lecture. But there are 

 difficulties to be met. Photographically two yellow nitrogen bands come 

 in with intensity equal or superior to the red one, and these have no 

 counterpart in auroral spectra. Moreover, the wave-length data for the 

 red auroral line are far from being accurate enough for an identification 

 depending on a single coincidence only. One of the most urgent problems 

 in auroral work is an adequate wave-length determination of this red line 

 from a large-scale spectrogram. 



CORONIUM. 



A problem which has generally been classed with those we have been 

 discussing is that of the lines in the sun's corona, attributed to a hj^pothetical 

 coronium. In the light of our present knowledge it is not probable, 

 perhaps we may say not possible, that an unknown element coronium 

 exists. Attempts have not been wanting to identify these lines with 

 known elements. The latest is by Freeman, working in the Ryerson 

 Laboratory of Chicago, who seeks to attribute the lines to argon. He 

 thinks, for instance, that the strong visual green line, from which the 

 conception of coronium arose, may result from two different transitions 

 in the argon atom, being in reaJity double. One of his proposed transitions 

 would give the fifth line of a possible series, and the other the ninth 

 member of an actual series. But none of the earlier members of either of 

 these series are seen in the corona, and this seems fatal to the identification 

 proposed. We could not assign an observed line at X3771 as Hj (H iota) 

 if H , H^ and H^ and the other earlier members of the series were miss- 

 ing, yet this would be an analogous case. 



I think we must consider the origin of the strong lines of the corona as 

 an unsolved problem. The possibility of their being in reality heads of 

 molecular bands must be kept in view. 



Excitation of the Various Spectra. 



We have discussed these cosmical spectra so far chiefly from the 

 standpoint of the spectroscopist. It will now be of interest to consider 

 the probable mode of excitation of some of them. 



Let us consider first the polar aurora ; • this, as is well known, is closely 

 bound up with exceptional conditions of magnetic disturbance, and these 

 in turn are conditioned by solar influence. As regards the nature of this 

 influence, the theory of Birkeland, elaborated by Stormer, still holds the 

 field. The sun was regarded by them as emitting localised streams of 

 electrically-charged particles from limited areas of its surface. 



The unrivalled advantages of this theory are that it allows the solar 

 action to be emitted in a highly specialised direction, thus accounting for 

 the sudden commencements of magnetic storms all over the globe, and 

 their tendency to recur after the twenty-seven days of a solar rotation 

 have passed : and further, that by the earth's magnetic field the action 

 can be got round to the night side of the earth. But this theory in its 

 original simplicity has required a good deal of patching, and it is difficult 



