136 On a possible Ultra-Solar Spectroscopic Phenomenon. [Feb. 1, 



whilst it is as definitely systematic, has twice the range of the mean lunar- 

 diurnal variation ; and it is also subject to remarkable modifications which 

 accompany changes of phase of the moon. 



II. " Note on a possible Ultra-Solar Spectroscopic Phenomenon." 

 By C. Piazzi Smyth, F.R.S., Astronomer Royal for Scotland. 

 Received December 13, 1871. 



One great object with the solar-eclipse expeditionists at work to-day in 

 the far East is to trace spectroscopically the existence of any faint solar 

 luminous appendage to a further distance from the sun than the brighter 

 parts of the corona hitherto so identified. 



But much further they cannot go, on account of the large amount of 

 general atmospheric illumination during every lunar-solar eclipse. The 

 matter may, however, be taken up again during a terrestrial solar eclipse, 

 i. 2. an ordinary sunset below the horizon, if the sun be sufficiently far 

 below to terminate all aerial twilight. Under such circumstances, too, it is 

 that the zodiacal light, historically called the sun's atmosphere, is occasion- 

 ally seen stretching away to distances of 60, 90, and even more degrees 

 from the sun. 



Now I have recently been trying to test in this manner an alleged 

 foreign observation of the zodiacal light's spectrum, but in vain. Some 

 sort of faint bluing of the dark night sky there certainly was over the 

 north-western horizon when the sun was more than 18° vertically below, 

 and its spectrum I noted at the time, but only as being that of the very 

 last of the twilight. 



Perhaps it was only that ; but perhaps, also, it may have been the direct 

 light of some solar appendage approximating to the outer coronal region ; 

 for, on making up the accompanying plate* of nine observed spectra, and 

 comparing them with a tabular solar spectrum, a wide distance was mani- 

 fested between the place of maximum light in the solar continuous spec- 

 trum, as given by Fraunhofer, or near wave-length 5620, and in my so- 

 called residual twilight spectrum, which was wave-length 5300. 



I tested, therefore, lately Fraunhofer's value for the solar spectrum in 

 my own night-apparatus, by darkening its objective to all but extinction 

 with four thicknesses of linen cloth in the middle of a dull, grey day, and 

 found for the last visible portion of continuous spectrum wave-length 5700 

 nearly ; that is, Fraunhofer's value was confirmed, and my residual-twilight 

 spectrum left anomalous as regards the ordinary solar spectrum, but re- 

 markably agreeable with the corona spectrum, whose chief line has wave- 

 length 5322, and which spectrum has been shown by Mr. Lockyer to de- 

 crease its number of lines, and tend to form a faint continuous spectrum, 

 as it thins out with increase of distance from the sun's limb. 



* See plate 54, vol. xiii., " Astronomical Observations made at the Royal Observa- 

 tory, Edinburgh." 



