Uarkor.l ' ' '- 1 | Nov. 15, 



peating the operation three times and varying the width of the slil from 

 lime t<> time in each region ; but not, ;i Bingle bright line could be detected , 

 L then requested you to conic and take a glance through my spectroscope, 

 as had been previously agreed ; saying thai although 1 could sec dark lines 

 and a continuous spectrum, I was unable to detect a single bright line, and 

 knew not what to make of it. You were then lookingat theeclipse through 

 your ingenious little telespectroscope of two inches aperture. You came 

 to my instrument, looked at the spectrum, moved the observing telescope 

 over its whole length and remarked that the results in my spectroscope 

 agreed entirely with those in yours, and that in both the spectrum appeared 

 continuous, showing no bright lines whatever. 



My mind being thus relieved, I took my place again at the spectroscope, 

 and this time, placing the slit tangential to the moon's limb, I moved the 

 observing telescope from end to end of the spectrum, opening and closing 

 the slit at intervals ; but the spectrum appeared as continuous as before. 

 A.gain the image was adjusted so that the slit was once more radial ; and 

 tins time on a still different portion of the corona. On examining again 

 the spectrum, no bright lines appeared, except once for an instant, when 

 the slit passed over the small chromospheric prominence already noticed. 

 Warned by .Mrs. Draper's clear and distinct counting that the precious 165 

 seconds had two-thirds gone, I decided to devote the time still remaining 

 to a more careful observation of the dark Fraunhofer lines. Now, for the 

 first time, as I adjusted the width of the slit and its position on the corona 

 with more care, I observed that these lines did not pass clear across the 

 field, but were of a length corresponding to the width of the coronal 

 image on the slit. At the base of the spectrum, which corresponded to 

 the base of the corona, they appeared bright and sharp ; certainly quite as 

 much so as in the light of the moon similarly condensed ; though the con- 

 tinuous spectrum which formed their background was relatively brighter 

 than in moonlight. There was no difficulty in identifying them as Fraunhofer 

 lines from their general appearance and position : but some of them could 

 be identified beyond question. Such were h and F, which were especially 

 distinct, I), E and G, which were considerably less so. They faded gradu 

 ally out from the base of the spectrum upward, appearing to end where 

 the continuous spectrum of the corona was limited above. While thus 

 employed, a Hash of sunlight told us that totality had ended and that the 

 solar eclipse of 1878 was over. 



In discussing the results of the spectroscopic observations which have 

 now been detailed, I am, in the first place, quite at a loss to account for the 

 fact that no bright lines were seen by me, notwithstanding the persistent 

 efforts made to get them. The failure to observe them can lie accounted 

 for, as i; would seem, only on the ground that with the dispersive power 

 employed, the bright lines were too faint to be seen on the much brighter 

 background of the continuous spectrum. 



'I'he lessons to be drawn from these spectroscopic observations appear to 

 be few and simple. The absence of bright lines, or at least, of any which 



