OF ARTS AND SCIENCES. 7 



The scientific world seems largely to have accepted the wave-lengths 

 of Angstrom and Thalen as final. One eminent authority speaks of 

 them as the " ne plus ultra " of spectroscopic accuracy ; and any at- 

 tempt to revise or correct them may be looked upon as presumptuous. 

 However, we believe the time has arrived when the whole of Thalen's 

 work on metallic spectra must be re-examined. It is safe to say 

 that he has tabulated not more than one line in many metals where 

 several exist, and his positions are occasionally wrong by as much as 

 two wave-lengths. 



As yet no approach to the accuracy with which the solar spectrum 

 has been delineated has been attempted in metallic spectra, — a re- 

 markable fact, when we consider that the chief interest that attaches 

 to the study of the solar spectrum is in its connection with spectra 

 of terrestrial elements. 



The test of the existence of oxygen in the sun is the coincidence of 

 the bright lines of the spectrum of oxygen with bright lines or with 

 dark lines of the solar spectrum. If the bright lines of any metallic 

 vapor formed in the electric arc or the electric spark coincide with the 

 dark lines of the solar spectrum which is photographed directly above 

 the spectrum of the metal on the same sensitive plate, the evidence is 

 usually considered conclusive in regard to the existence of the metal 

 in the sun. In the case of iron, where hundreds of lines of the metal 

 coincide with dark lines in the solar spectrum, not only in exact posi- 

 tion but in general grouping and character, the evidence cannot be 

 doubted by any one who has carefully examined it. When a ma- 

 jority of the lines of any metal coincide with dark lines in the solar 

 spectrum under high dispersion, not only in position but in group- 

 ing, while a few of the metal lines have no representatives in the 

 solar spectrum, there is a probability that the corresponding lines 

 wanting in the sun have been obliterated by superposed lines or bands 

 of other metals. In our paper " On the Existence of Carbon in 

 the Sun," we have called attention to a case of such obliteration. 

 It is probable, also, that the non-appearance of certain lines in the 

 sun may be due to certain conditions of temperature. We have 

 discussed this point more fully in the paper on Carbon, above re- 

 ferred to. 



The same remarks apply to the coincidence of the lines of any ele- 

 ment with the supposed bright spaces in the sun. The value of the 

 test of coincidence increases with the number of coincidences. If an 

 element has only two or three lines, and these two or three agree in 

 position with dark lines in the solar spectrum, the evidence of the 



