SOLAR EGLirSE, MAY C, 1883. 107 



gaseous atmospliere extemliii*;' as far as implied by the spectroscope. On the other hauil, we can 

 find no explanation in difliision by our own atmosphere, for diffusion was absolutely insensible. 

 Still, as an einineut autliority has suggested that this might be an adequate explanation of the 

 phenomenon, it is well for inc to give categorically the reasons for rejecting it. 



First. The skj- was extremely clear, aud the moon during the eclipse sensibly black. Of 

 course, if there had been any considerable diffusion, the moon would have looked less dark near 

 the limb. 



Second. The bright lines of the chromosphere and corona stopped abruptly at the moon's limb, 

 although the diffusion here should have been more sensible than towards the outer corona, because 

 at the place where the change in illumination was most I'apid. It will be recognized that my 

 examination of this point was a careful one, for by it I assured myself of the adjustment of the 

 apparatus in the beginning of the observations. 



Third. All the photographs taken by the English and French observers showed a sensibly 

 black moon. 



Fourth. In a photograph of the coronal lines H and K, taken by the English observers, these 

 lines ended abruptly at the moon's edge, although the instrumental diffusion must have been much 

 greater than in mj- own apparatus because thej- made use of a heliostat. 



Before attempting any explanation of the observed pheuomeno7i, however, it is well to review 

 briefly all that study has taught us concerning the corona, for it is evident that no theory is tenable 

 which is contradicted by any established fact. la this review I shall depend implicitly upon Ran- 

 yard's "Observations made during total solar eclipses, " which forms Vol. XLI of the Memoirs ot 

 the Royal Astronomical Society, and my references will be to the pages of that work. This ad- 

 mirable work, for which students of solar physics owe Mr. Ranyabd a debt of gratitude, includes all 



useful observations to the eclipse of 1875 inclusive. For those of later date I shall refer to the original 

 publications. 



GENERAL HETIEW OF THE ItESVLlS OF OBSEETATIOSS ON THE COliONA. 



I. — Spectroscopic Observations. 



As we shall find much contradictory evidence in these observations, and many cases of negative 

 evidence even on occasions when definite advances in our knowledge have been made, it is necessary 

 in a general discussion to establish some criterion by which this evidence may be weighted. Un- 

 questionably the best method of determining the relative values of the observations is from the 

 efficiency of the apparatus used in each case. To do this we must first discuss the theory of the 

 spectroscope as applied to such work. This step seems to me advisable, not alone on account of 

 the reason given, but also because the apparatus used is often so badly devised that the theory is 

 evidently not generally known. The frequent description of non-essential constants of the spec- 

 troscopes used when the. essential ones are omi^tted is another proof of inadequate ai)prehension of 

 the optical principles involved. 



Of the three forms of spectroscope used in corona observations we will discuss, in order, [a] the 

 prismatic telescope; {b) the integrating spectroscope; (c) the analyzing spectroscope. 



(a) Prismatic telescope. — This consists of a system of prisms or a grating before the objective ot 

 a telescope or the naked eye. The office of the dispersive system is to form a virtual spectrum ot 

 the corona at an infinite distance. This image is to be observed by means of the telescope. It is 

 clear that a high magnifying power is not requisite for observing the coronal ring, since it measures 

 a half degree in its inside diameter; on the other hand, it is equally evident that the image should 

 appear as brilliant as possible. This last condition limits the magnifying power. According to 

 the well-known law governing telescopic vision of a surface, the power for maximum brightness 



