ORIGIN OF URANIUM AND THORIUM 123 



ail those which Miss Clerlie regarded as well established, but chromium and 

 uickel are perhaps present (ibid., p. 198). The identification of the solar ele- 

 ments rests chiefly on Rowland's great table of the solar spectrum, 1895 to 

 1897. In 1891 Rowland published a preliminary table (Johns Hopkins Uni- 

 versity circular) containing most of the elements enumerated in this paper. 

 In the six succeeding years he was able to add only two, namely, platinum 

 and ruthenium. Of these, platinum is represented by only one line 

 (2,923.180), which it shares with cerium, and this seems to me insufficient evi- 

 dence of platinum in the sun ; but ruthenium has five lines, of which one is 

 doubtful. Columbium (or niobium) appeared in the first list, but in the great 

 table is represented by a single doubtful line with a wave length of 4,2.32.111, 

 and I have left it out. Rowland photographed the spectrum of every element 

 known in 189.5 for comparison with the solar spectrum, excepting gallium, of 

 which he had no specimen, and he failed to find even a doubtful trace of 

 uranium or thorium. Mr P. G. Nutting has also been good enough to compare 

 for me Exner and Haschek's table of the spectrum of uranium with a 30-foot 

 reproduction of the solar spectrum, and finds no evidence of its existence. 

 Rowland also found erbium and neodymium in the sun, but they do not appear 

 in the table because their position in the periodic system is uncertain. Row- 

 land did not include nitrogen in his lists, but seems to have agreed with other 

 spectroscopists that the solar carbon lines are due to cyanogen. These lines 

 consequently prove nitrogen just as well as carbon (Kayser and Runge, 

 Wiedemann's Annalen, vol. 38, 1889, p. 80). Since Rowland's labors closed, 

 gallium has been found in the solar spectrum by Messrs Hartley and Ramage, 

 the presence of oxygen was established by Messrs Runge and Paschen, and 

 Sir William Ramsay has identified helium in clevite. Very lately the rare 

 element, europium, has been identified by Mr J. Lunt (Proceedings of the 

 Royal Society, series A, vol. 79, 1907, p. 118). Coronium has also been iden- 

 tified in the corona of the sun (Astrophysical Journal, vol. 10, 1899, p. 306), 

 and is believed to be a vei'y light inert gas. It is unknown on earth or in the 

 stars. The list of elements in meteors is that given by Sir Archibald Geikie, 

 with the addition of helium (Young's Astronomy). Some other simple sub- 

 stances (lead, silver, gallium) have been announced as determinable spec- 

 troseopically in meteorites, but spectroscopic traces in an accessible substance 

 do not seem comparable with stellar constituents observed under conditions 

 of very much greater difficulty. 



Rowland's determinations of wave lengths have been shown by Mr Kayser 

 to be less accurate than he supposed them (Philosophical Magazine, vol. f*. 

 1904, p. 568), but his identifications of the elements in the sun were made by 

 direct comparison between photographs of the solar spectrum and of the 

 spectra of the elements, and will presumably prove correct. 



In tlie nebulae only helium, hydrogen, and nebulium have been identi- 

 fied. Nebulium which in the earlier days of spectroscopy was confounded 

 with nitrogen, does not appear in the table, because it has not been found 

 on earth and its place in the periodic system is unknown. The nebulae 

 very likely contain other elements besides the three just mentioned, but 

 the nebulas are excessively tenuous bodies and must also be extremely 



