106 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



tiflc importance as showing that the coioim could have been seen much eailier by employing a 

 more delicate method. 



Just before second contact 1 put my eye to the spectroscope. The right-hand side of the 

 spectrum, corresponding to the eastern limb of the sun, exhibited a long brilliant line in the 

 place of 1474, extending 10' or 12' from the edge of the sun ; on the other half, however, the line 

 was much fiiinter and not more than 3' or 4' in length. The magnitude of the difference was so 

 great a surprise to me that, although the appearance of the spectrum itself would have betrayed 

 any fault in adjustment, I immediately tested it by giving the equatorial a slight motion east and 

 west. The effect was what I anticipated— the shortening of alternately one and the other at the 

 base, but no other change. 



I had intended to examine all parts of the corona by rotating the spectroscope gradually 

 through 180° on the axis of collimation, but the phenomenon observed was so important that I at 

 once resohed to leave the spectroscope unchanged and to note the changes due to the moon's 

 motion alone. 



This exaggerated difference on the opposite sides of the sun lasted but a short time. I 

 observed the rapid equalization, but only made the single eutry in my note-book of equality at 200 

 seconds after beginning of totality. 



Just at the beginning of totality I had seen the lines C, D3, F, and near G extremely brilliant, 

 but very short. This was, of course, just at the middle of my spectrum band, along a line extend- 

 ing from red to ^'iolet, and they were confined to that side of the line corresponding to the eastern 

 limb of the sun. It is possible that some or all of these lines extended away from the middle of 

 the band as very faint bright lines, but I could not be sure of the fact after the eclipse from a 

 recollection of the hasty examination which I gave it. 



At 200 seconds from the beginning, experieuce had showu that all the significant changes 

 could be looked for only in the last few seconds of totality, and, as more than two minutes 

 remained, I set myself to a deliberate examination of the spectrum of the corona exclusive of 1474. 

 That which first struck me was the intensity of the colors: none were too feeble to give a 

 vivid impression of their hues, except perhaps the upper violet, which I ftiil to remember. Not 

 less striking was the marked difl'erence between this spectrum and one equally brilliant of sky- 

 light. Two lines only were seen, the bright 1474 and the dark D line, rather faint. The latter, 

 however, was so distinct that it could hardly liave been regarded as a single line even if unfa- 

 miliar; but the other familiar groups, C, E, h, F, and less certainly G as being much fainter, were 

 wanting. A streakiuess at first suspected in the green was not rendered more probable by subse- 

 quent inspection. This observation was made with the utmost care and deliberation, as I was 

 aware that it differed from the spectroscopic results reported at some former eclipses. I also made 

 a single estimate of the extent to which the coronal line could be traced from the limb of the 

 moon, and found it about 15 minutes of arc. 



As the end of totality approached I again turned my attention to the 1 474 line. At 280 sec- 

 onds I recorded the line on the left side of the spectrum brighter and on the right side shorter^ 

 that is, the line had now become longer and brighter on the western side. This difference increased 

 until returning sunlight put au end to the observations. At a time estimated as 20 seconds after 

 Mr. Fletcher ceased counting, that is, at 320 seconds from the beginning of totality, I noted C, P3, 

 and F as appearing but very short. I cannot recall seeing the hydrogen line near G. 



The conclusions to be derived from these observations are interesting. It is obvious that the 

 enormous change in the exteut to which the 1474 line could be traced east and west of the sun, with 

 very slight change of the moon's place, at once precludes the explanation hitherto accepted of a, 



