METHODS IN MODERN PHYSICAL ASTRONOMY. 481 



in its relations with bodies. We then bring the two sciences together 

 by performing on light an operation parallel to the one that has been 

 made on matter. Chemical analysis by light was performed in posse 

 on the day when its rays were regarded in the light of chemical spe- 

 cies. 



This great idea of the specific character of luminous rays is due to 

 Newton. It was introduced into science when that great genius gave 

 his explanation of the action of the prism on white light. The founda- 

 tions of spectrum analysis were laid at that time, and the study of it 

 might have been begun then at once ; but the human mind does not 

 proceed with so penetrating and absolute a logic as that. The succes- 

 sive and often fortuitous acquisition of revealing facts had to be left 

 to time. It must, however, be said that, when the facts were presented, 

 their real significance would have been overlooked, notwithstanding 

 the genius of the experimenters, had not the grand idea of Newton 

 illuminated them with its brilliant light. The conception of the indi- 

 viduality of the rays made its way so slowly into our minds that it 

 bore its fruits, as it were, unknown to us. But history, whose vision 

 goes back to the beginnings, will have to assign their respective parts 

 to the causes which have been influential toward the end. The allot- 

 ment will in no degree diminish the admiration which is due to the 

 creators of the marvelous instrument. They have given a body of life 

 to what was slumbering in posse / they have thus shown themselves 

 worthy continuers of the work of Newton. 



You know that this spectrum analysis made its appearance very 

 suddenly in science. You may recollect the emotion that affected us 

 all when it was announced that the solar atmosphere had been chem- 

 ically analyzed, and a list of the metals it contained was published. 

 You are, however, too well acquainted with the history of science to 

 suppose that a method as complete as the one that was thus announced 

 had no antecedents. The antecedents in fact existed, and they were 

 even numerous. With the labors which contributed to the constitu- 

 tion of the definitive method were associated the names of Sir John 

 Herschel, Talbot, Miller, Wheatstone, Swan, Masson, Foucault, etc. ; 

 but Kirchhoff and Bunsen were able to make a synthesis of all these 

 efforts, and they gave the method its general and practical form. 

 When spectrum analysis presented itself to the scientific world, it held 

 in one hand cresium and rubidium and in the other hand a list of the 

 metals recognized in a star ninety-three million miles away. Why, 

 then, should we be surprised at the enthusiastic reception that was 

 given it ? 



At first it was believed that the incandescence of gases was one of 

 the conditions of elective absorption by them. A French physicist, 

 judging that the phenomenon related rather to the gaseous condition 

 than to the temperature, was led to believe that the earth's atmos- 

 phere, as well as the atmosphere which was supposed to exist around 



VOL. XXII. 31 



