124 MEMOIRS OF THE NATIONAL ACADEMY OF SCIENCES. 



Tims it seemed to me not an improbable explanation of the corona to attribute it to diffraction, 

 thougU the reasoning upon which my conclusion was based was hardly of sufficiently definite char- 

 acter to justify its publication. When the expedition to observe the eclipse of last May was pro- 

 liosed, I thought, though there was no expectation in my mind of being one of the observers, that 

 the theory might be rendered sufficiently probable to recommend a crucial test of its validity, if 

 such could be devised. It was not difficult to find such a decisive test. According to the theory 

 the corona should be particularly rich in edge light of the sun, that is, in chromosphere light. Since 

 this is not subject to such a variation as light from the photosphere near the limb, the conclusion 

 as to rapid change in the distribution of such light during the first and last seconds of totality 

 is an essential part of the theoiy. On the other hand, if the bright lines of the coronal spectrum 

 indicate an atmosphere about the sun, there can be no change in it due to the motion of the moon. 

 This defines a pronounced and recognizable difference between the phenomena as demanded by 

 the two theories. How this difference was made observable and the results of observation' are 

 given in the first part of this report. 



It only remains to apply the theory to the explanation of the known characteristics of the 

 corona to find how far it escapes the defects of the old one and what it leaves unexplained. 



Prom it follows that the light of the corona consists in part of ordinary sunlight, a strong ad- 

 mixture of chromosphere light, together with such light as is reflected or radiated by particles of 

 solid or liquid matter which are carried above the photosphere by convection curreuts or by the 

 greater atmospheric disturbances which give rise to the prominences. As the strongest lines in 

 the ordinary solar spectrum correspond roughly with the brightest lines of the chromosphere, in 

 the composite spectrum of a corona thus formed the principle Fraunhofer lines must be wanting 

 and even some of them reversed. The reversal, however, cannot depend in any way upon the 

 vapor densities of the gases characterized by the lines, but almost exclusively on the brightness 

 of that particular constituent of the chromospheric light. Thus the calcium lines H and K may be 

 much more prominent tjian the hydrogen lines, as is probably the case. Again, the fact that the 

 1474 K line is recognized so much higher in the corona than the hydrogen lines leads to no such 

 contradiction as to the relative densities of the corresponding vajiors as we have seen to be the 

 case in the ordinary theory, since it only implies that the material indicated by 1474 K shines with 

 a greater brightness. Whether this implied intensity of the green line in the ordinary siiectrum 

 of the chromosphere is in accord with observations or not I think it impossible to assert. Cer- 

 tainly, in view of the facts that it is only found in the closest proximity to the photosphere and 

 that it lies near the most brilliant portion of the spectrum where diffused light would do most to 

 blot it out, w^e must consider its visibility under any circumstances as a proof of extraordinary 

 brightness. 



The diffraction theory also demands that the relative proportion of chromospheric light should 

 diminish in leaving the sun, hence the spectrum of the outer corona should more closely resem- 

 ble that of ordinary feeble daylight. Stone's observations of 1S74 seem to afford strong evidence 

 that this is the case, for he detected Fraunhofer lines at a distance from the sun though they were 

 absent at the base of the corona. 



Any variations in the intensity of the light near the limb of the sun must produce con'espoud- 

 iug variations in the brightness of neighboring portions of the corona ; a brighter area would give 

 rise to a streamer,, and an area of deficient light to a rift. We know from i)hotogTaphs of the solar 

 surfiice that there are such variations of enormous magnitude, especially in those latitudes where 

 sun spots are frequent. It is just at those regions that the rifts and streamers are most abundant 

 and well marked. It is evident from the nature of the argument that a diffraction corona would 



