226 On the Temperature of Incandescence. 



a single physical experiment, to show that hydrogen can be 

 caused to glow with a red light or yield its characteristic spec- 

 trum by heat alone, or that it is actually highly heated by the 

 electric influence which causes it thus to glow when rarified in 

 a Geissler tube. On the contrary, the temperature of the sun 

 might be comparatively low, yet if still above the temperature 

 of the dissociation of water (approximately 4,800 degrees C.) 7 be 

 consistent with all the eruptive phenomena of the chromosphere 

 and protuberances, the chromatic phenomena of which incon- 

 sistent with that temperature may it seems to me, as yet be 

 attributed to purely electrical influences. If the maintenance 

 of the opening in the crust which we call a spot, were consistent 

 with a comparatively quiet condition of the surface upon which 

 the crust floats, the penumbra might perhaps be an image of the 

 the interior side of the crust seen through the sun's mass, or it 

 might be the less luminous reflection of the side of the opening 

 from the sun's actual surface. Were the mass of the sun metal- 

 lic, the image would appear almost as bright as the reflected 

 object, like the image of a ring of litharge, reflected from the 

 surface it surrounds, or of a button of silver, melted in a scori- 

 fier or cupel ; but if the sun be a mass of a transparent silicate, 

 the image would be far less brilliant than the side of the cavity 

 in the crust. And in this connection it may be remarked, that 

 in certain details many of the drawings of sun-spots which have 

 been published appear to afford, to a person prejudiced in its 

 favor, striking illustrations of this hypothesis, and the famous 

 representation of a sun-spot reproduced in almost all astronomi- 

 cal work?, and which was drawn by Prof. Langley without refer- 

 ence to such an hypothesis, would seem to support it conclu- 

 sively, for it shows a projection of the crust over an annular 

 chasm and its apparent image in a mirror-like surface beneath. 



If, whatever the size of the sun-spot, the penumbra be found 

 to have approximately always the same width, that it is an image 

 of the walls of the chasm would be strongly indicated, but it by 

 no means follows that it must be such an image to accord with 

 this hypothesis. 



On the contrary, while reflection would still occur to a greater 

 or less extent, the penumbra might actually be a thin film, of 



