342 HUMBOLDT'S COSMOS : 



years for the transmission of light to us. The most diligent and 

 successful observer in this part of astronomy, M. Peters, has 

 carried the determination of parallax in some cases even to 

 the tenth of a second; thus indicating distances 'which we 

 may well hesitate to translate into numbers either for space 

 or time.* Such research is likely to be carried yet farther, 

 but the main results are probably now ascertained. We may 

 name as one of these results, the proof of the great difference 

 in the magnitude of stars, derived from the want of any pro- 

 portion between their degree of brightness and their parallax. 

 This difference might have been supposed probable, but it is 

 thus rendered almost certain. 



The actual magnitude, however, of any of the Stars is a 

 problem hitherto insuperable ; and probably destined ever to 

 remain so, seeing that the best telescopes do not give to them 

 any real disk or angular diameter. The only approach to a 

 solution is by comparative photometrical observations on the 

 light of our own Sun, and of certain conspicuous stars ; 

 a method open to various sources of error in its progress, and 

 rendered doubtful in its results by our ignorance of the rela- 

 tive intensity of light emitted from these different bodies. 

 The intrinsic brightness of a Centauri has been estimated at 

 2-J- times, that of Sirius even at 63 times the light of the 

 Sun ; yet we are not entitled to draw thence any direct con- 

 clusion as to the comparative magnitude of these great globes. 

 The only inference attainable is that stated above, of great 

 diversity of size among them, corresponding in this respect 

 to the conditions of our own planetary system. We do not 



* M. Peters's observations are recorded in Struve's ' Etudes d' Astronomic 

 Stellaire' a work remarkable among all others of our time for its profound 

 views in this department of the science. Had we room for it, we might give, 

 what we do not find in Baron Humboldt's volume, an account of the refined 

 method of investigation by which M. Struve obtains, first, the relative mean 

 distance of the stars classed under different magnitudes ; and then, by reference 

 to the distances actually determined, the absolute mean distance of each of these 

 classes of stars. 



