72 COSMOS. 



periments of the accurate observer, Bouguer, that a difference 

 of -gipth in the intensity of the light is necessary to render 

 objects visible. To use his own expression, we only negatively 

 see mountain-tops from which but little light is radiated, and 

 which stand out from the vault of heaven in the form of dark 

 masses ; their visibility is solely owing to the difference in the 

 thickness of the atmospheric strata extending respectively to 

 the object and to the horizon. Strongly illumined objects, such 

 as snow-clad mountains, white chalk cliffs, and conical rocks 

 of pumice-stone, are seen positively. 



The distance at which high mountain summits may be 

 recognized from the sea is not devoid of interest in relation to 

 practical navigation, where exact astronomical determinations 

 are wanting to indicate the ship's place. I have treated 

 this subject more at length in another work, 19 where I con- 

 sidered the distance at which the Peak of Teneriffe might be 

 seen. 



The question whether stars can be seen by daylight with the 

 naked eye through the shafts of mines, and on very high 

 mountains, has been with me a subject of inquiry since my 

 early youth. I was aware that Aristotle had maintained 2 * 



19 Humboldt, Relation hist, du Voyage aux Regions equinox. 

 torn. i. pp. 92-97; and Bouguer, Traite d' Optique, pp. 360 

 and 365. (Compare also Captain Beechey in the Manual 

 of Scientific Enquiry for the use of the Royal Navy, 1849, 

 p. 71.) 



20 The passage in Aristotle referred to by Buffon occurs in 

 a work where we should have least expected to find it De 

 Generat. Animal., v. i. p. 780, Bekker. Literally trans- 

 lated, it runs as follows: "Keenness of sight is as much 

 the power of seeing far, as of accurately distinguishing 

 the differences presented by the objects viewed. These two 

 properties are not met with in the same individuals. For 

 he who holds his hand over his eyes, or looks through a 

 tube, is not on that account more or less able to distinguish 



