OP ARTS AND SCIENCES. 17 



Moreover, a high dispersion shows that very few lines of metals 

 are simple and short, hut, on the contrary, winged and nebulous, and 

 complicated by a great variety of reversal phenomena. A " line " is 

 sometimes half an inch wide on the photographic plate, or it may be 

 split into ten by reversals. 



At first, we believed that these reversals were due to defects in the 

 rulino - of the grating, but we are convinced that they are true phe- 

 nomena from the following experiments. 1st. The wings continue 

 when various portions of the grating are covered. 2d. They are the 

 same in three successive orders of spectra. 3d. They are very differ- 

 ent in different metals, and in some are not seen at all. 4th. We 

 arranged a flat grating, with collimator and projecting lens, each of 

 five feet focus, and found that with this apparatus the same phenom- 

 ena appeared. 



On pages 87 and 88 of " The Sun," Professor Young gives a list of 

 elements in the sun according to the best authorities, which is followed 

 by a list of doubtful elements. Some of these we have examined 

 with tho following results : — 



Cadmium. — The coincidence of the two lines given by Lockyer at 

 wave-lengths 4677 and 4799 is perfect. These are the only cadmium 

 lines near, and sun lines in the vicinity are not numerous. 



Lead. — The evidence for lead, due to Lockyer, is based upon three 

 lines at 4019.7, 4058.2, and 40G1.8. We have photographed these 

 Lines with the sun many times. They are broad and nebulous, and 

 often several times reversed. Lines in solar spectrum numerous and 

 faint. 4019.7 and 4058.2 certainly do not coincide. 4061.8 is very 

 difficult to pronounce upon ; it may coincide. 



Cerium, Molybdenum, Uranium, and Vanadium. — These four 

 metals may be classed together. Lockyer finds four coincidences 

 each for molybdenum and vanadium, three for uranium, and two for 

 cerium. The arc spectrum of each is characterized by great com- 

 plexity and vast numbers of lines. So numerous are the lines in fact, 

 that often on the photographs the total space occupied by them is 

 greater than the space not so occupied. A plate ten inches long may 

 contain a thousand or so. Evidently coincidences between these and 

 solar lines cannot fail to occur as matters of chance, and therefore 

 prove nothing. One can easily count a hundred or so such coinci- 

 dences without the slightest conviction that the connection is other 

 than fortuitous. Of course all this is nothing against the probability 

 of these metals being in the sun ; but at the same time those peculi- 

 arities of grouping, strength of lines, and other characteristics which 

 vol. xxin. (n. s. xv.) 2 



