284: THE NEBULAE HYPOTHESIS. 



most of necessity. That an a 2^riori ixiference which prob' 

 ably seemed to many readers wildly speculative, should be 

 thus conclusively justified by observations, made without 

 reference to any theory, is a striking fact ; and it gives yet 

 further support to the hypothesis from which this a jyriori 

 conclusion was drawn. It may be well to add that Kirch 

 hoff, to whom we owe this discovery respecting the consti- 

 tution of the solar atmosj^here, himself remarks in his me- 

 moir of 1861, that the facts disclosed are in harmony with 

 the Xebular Hypothesis. 



And here let us not omit to note also, the significant 

 bearing which Kirchhoff's results have on the doctrine con- 

 tended for in a foregoing section. Leaving out the barium, 

 copper, and zinc, of which the quantities are inferred to be 

 small, the metals existing as vapours in the Sun's atmo- 

 sphere, and by consequence as molten in his incandescent 

 body, have an average specific gravity of 4* 25. But the 

 average specific gravity of the Sun is about 1. How is 

 this discrepancy to be explained ? To say that the Sun 

 consists almost wholly of the three lighter metals named, 

 Avould be quite unwarranted by the evidence : the results 

 of spectrum-analysis would just as much warrant the asser- 

 tion that the Sun consists almost wholly of the three heav- 

 ier. Three "metals (two of them heavy) having been al- 

 ready left out of the estimate because their quantities ap- 

 pear to be small, the only legitimate assumption on which 

 to base an estimate of specific gravity, is that the rest are 

 present in something Uke equal amounts. Is it then that 

 the lighter metals exist in larger proportions in the molten 

 mass, though not in the atmosphere ? This is very un- 

 likely : the known habitudes of matter rather imply that 

 the reverse is the case. Is it then that under the condi- 

 tions of temperature and gravitation existing in the Sun, 

 the state of liquid aggregation is wholly unlike that exist- 

 ing here ? This is a very strong assumption : it is one for 



