88 COSMOS. 



optical problem, excited the strongest interest in the mind of 

 Bessel, whose too early death was so unfortunate for the 

 cause of science. In his long correspondence with myself, he 

 frequently reverted to this subject, admitting that he could not 

 arrive at any satisfactory solution. I feel confident it will not 

 be unwelcome to nay readers, if I subjoin, in the form of a note, 

 some of the opinions of Arago, 39 as expressed in one of the 



the reflection of solar rays on the surface of separation of the 

 strata of different density, of which we suppose the atmo- 

 sphere to be composed. In fact, if we suppose the sun to be 

 situated on the horizon, the surfaces of separation in the 

 direction of the zenith will be horizontal, and consequently 

 the reflection would likewise be horizontal, and we should 

 not be able to see any light at the zenith. On the supposi- 

 tion that such strata exist, no ray would reach us by means 

 of direct reflection. Repeated reflections would be necessary 

 to produce any effect. In order, therefore, to explain the 

 phenomenon of diffused light, we must suppose the atmo- 

 sphere to be composed of molecules (of a spherical form, 

 for instance), each of which presents an image of the 

 sun somewhat in the same manner as an ordinary glass 

 ball. Pure air is blue, because, according to Newton, 

 the molecules of the air have the thickness necessary to 

 reflect blue rays. It is therefore natural that the small 

 images of the sun, reflected by the spherical molecules of the 

 atmosphere, should present a bluish tinge; this colour is not, 

 however, pure blue, but white, in which the blue predomi- 

 nates. When the sky is not perfectly pure and the atmo- 

 sphere is blended with perceptible vapours, the diffused light 

 is mixed with a large proportion of white. As the moon is 

 yellow, the blue of the air assumes somewhat of a greenish 

 tinge by night, or, in other words, becomes blended with 

 yellow." MSS 0/1847. 



39 Ifun des Effets des Lunettes sur la Visibilite des etoilp*. 

 (Lettre de M. Arago a M. de Humboldt en Dec. 1847.) 



" I/ceil n'est doue que d'une sensibilite circonscrite, bornee. 

 Quand la lumiere qui frappe la retine, n'a pas assez d'inten- 

 site, 1'ceil ne sent rien. C'est par un manque d'intensite que 

 beaucoup ftetoiles, meme dans les nuits les plus profondes 

 echappent a nos observations. Les lunettes ont pour effet, 



