26 



Mr. G. J. Stoney on the Physical [June 20, 



each result by the author, since the proofs must in many cases be entirely 

 omitted. It Avill be convenient to do this by numbers. 



The probability 4, then, is to be understood to imply that the matter in 

 hand appears to the author to be fully made out. He would, for example, 

 assign this probability to the wave-theory of light, and to the main features 

 of the theory of the molecular constitution of gases which have been worked 

 out by Clausius and others within the last twenty years. The number 1 

 will be used where an hypothesis agrees so well with such of the pheno- 

 mena as are known, that it is concluded that it must be either the true ac- 

 count of them, or bear some intimate relation to the true theory ; 2 will 

 indicate that we have good ground to conclude our hypothesis to be the 

 true theory, although at the same time the evidence is too scanty or con- 

 flicting to free us from hesitation ; 3 will indicate a proof so strong that we 

 should be very much surprised if anything were eventually to disturb it ; 

 4, as has been already stated, will mark a conclusion fully made out ; and 

 to complete the series, 5 may be used for that demonstrative proof of which 

 few subjects of inquiry are susceptible. 



Observations with the spectroscope have made known to us that the sun's 

 outer atmosphere, that is, the part of the atmosphere which extends out- 

 side the photosphere, is a mixture of many gases, amongst which hydrogen, 

 sodium, magnesium, calcium, chromium, manganese, iron, nickel, cobalt, 

 copper, zinc, and barium — all of them permanent gases in consequence of 

 the temperature — have been detected. Now it is shown to be a necessary 

 consequence of the molecular constitution of gases that in such an atmo- 

 sphere, decreasing in temperature from within outwards, the various consti- 

 tuent gases are not everywhere equally mixed, but that in the upper regions 

 those which have the lightest molecules rise the furthest, so that the gases 

 overlap one another in the order of the masses of their molecules (proba- 

 bility 5). It also follows from a consideration of the vapour-densities and 

 atomic weights of the chemical elements, with probabilities which range 

 from 4 to 1, that those which are present in the sun's atmosphere have 

 molecules with masses increasing in the order in which their names have 

 been printed above, the molecules of hydrogen being the lightest. This, 

 then, is the order in which the boundaries of these gases would be met with 

 in descending from the surface of the sun's atmosphere downwards. 



This result is abundantly confirmed, and in its main features raised to 

 probability 4, by observations with the spectroscope. Each constituent of 

 the solar atmosphere is opake to those rays which it emits when incandes- 

 cent, and which constitute its spectrum. In this way all the light of these 

 particular wave-lengths which has been emitted, either by the photosphere, 

 or by the lower and more intensely heated strata of a gas in the solar at- 

 mosphere, is stopped in its passage outwards, and the gas substitutes for it 

 the much more subdued light which emanates from its own upper and 

 therefore coolest stratum. Now if the view enunciated in the last paragraph 

 be true, these outer layers of the respective gases, from which the rays as 



