Notices respecting New Books. 



55 



proved both experimentally and theoretically that a gas has the 

 power of absorbing the very same rays of light which, when ignited, 

 it gives off. The positive spectrum is from the rays of light produced 

 by the gas ; the negative spectrum is due to absorption of the same 

 rays when white light from some extraneous source is passed through 

 the gas. Kirchhoff hence arrives at the conclusion that the sun con- 

 sists of an ignited solid or liquid central body, throwing off intense 

 white light, and of an atmosphere containing many of our terres- 

 trial elements in a state of vapour, yet capable of absorbing parts of 

 the white light. The inductive proof of this interesting conclusion 

 is one of the most beautiful problems which the philosophy of Bacon 

 has presented ; and we shall gladly dwell upon it. 



Of the 676 lines registered in Kirchhoff's Map and Table, 141 are 

 marked as corresponding with lines in the spectra of certain metals ; 

 in 8 cases there is coincidence with the lines of two metals ; and of 

 the coincidences, 60 occur with lines in the very complicated spec- 

 trum of iron. Taking iron as an instance, it has been proved that 

 iron in the sun may produce the negative spectrum corresponding to 

 that of iron, which we detect with surprise in the solar spectrum. 

 Yet it may be that, of the 676 lines observed, 60 have by mere 

 chance occurred in the precise positions in which iron would have 

 produced them. Required the probability that the correspondence 

 is due to chance, not to the presence of iron in the sun. 



The 676 lines in the Professor's Map occur in a length of 1250 

 millims. ; so that the average interval between each two lines may 

 be taken at 2 millims. The method of observing coincidences was 

 so far exact that, when a bright line of the iron spectrum came 

 within \ millim. of a dark line of the solar spectrum, there seemed 

 to be coincidence. We may express this in the adjoining figure by 

 supposing the dark and, in reality, infinitely narrow lines A, B, C, 

 &c. to be arranged at equal intervals of 2 millims., each being spread 



TNT 





out \ millim. on each side. Thus each dark line has an apparent 

 width of 1 millim., and the lines altogether occupy half the length 

 of the spectrum. 



Now suppose an infinitely thin bright line of the iron spectrum to 

 fall by pure chance upon such a spectrum. It is obviously just an 

 equal chance that it will fall upon one of the dark spaces aa', bb' , cc', 

 &c, as upon one of the bright spaces a'b, b'c, &c. If it fall like N 

 upon a bright space, there will be no coincidence ; if it fall, however, 

 like M, ever so little within the boundary of the dark space, there 

 will be at least apparent coincidence. Each of the 60 iron lines in 



