THE FUEL OF THE SUN. 35 



the spectroscopic dilution of the continuous spectrum of the 

 denser photosphere, while the solid projectiles that must 

 proceed from them in every direction can only be seen dur- 

 ing a solar eclipse. 



The observations and drawings of Zollner and Kespighi 

 were, for the most part, made while my book was in the 

 press, and, like those of Secchi above quoted, were un- 

 known to me when I wrote; I was then only able to quote, 

 in support of my theoretical requirements, the evidences of 

 actually observed tangential ejection afforded by Sir John 

 HerschePs account of the great solar storm of September 1, 

 1859. 



Besides this direct tangential projection there are other 

 elements of motion contributing to the same result, such as 

 the whirl of the prominences on themselves, their motion 

 of translation on the sun's disk, and the rotation of the sun 

 itself. 



I must now bring this sketch to a close by stating that, 

 in order to submit the fundamental question of an universal 

 atmosphere to an experimentumcrucis analogous to that by 

 which Pascal tested the atmospheric theory of Torricelli, I 

 have calculated the theoretical density of the atmosphere of 

 the moon and of each of the planets, and compared the 

 results as severely as I could with the observed facts. As 

 Jupiter is 27,100 times heavier than the moon, and between 

 these wide extremes there are six planets presenting great 

 variations of mass, the probabilities of accidental coinci- 

 dence are overwhelmingly against me, and a close concur- 

 rence of observed telescopic refraction and other phenom- 

 ena with the theoretical atmospheric density must afford 

 the strongest possible confirmation of the soundness of the 

 basis of my whole argument. Such a concurrence exists, 

 and some new and very curious light is unexpectedly thrown 

 upon the meteorology of Mars and the constitution of the 

 larger planets. The latter, if I am right, must be minia- 

 ture suns, permanently red or white-hot, must be some- 

 thing like a photosphere, surrounded by a sphere of vapor 

 (the outside of which we see), must have mimic spot vor- 

 tices and prominences, and in the case of Saturn must 

 eject volle} r s of meteoric matter, some of which should 



