Frazer.] iOO [jyiay i, 1574. 



tlius it happens that in a clear moon-light night the sky is much more 

 strikingly blue than the same sky would be at mid-day. 



When the Moon shines in the day-time we must suppose that the rays 

 she sends to us are affected in precisely the same way as at night. If 

 she appear white (as is the case) it must be owing to an addition to this 

 light of the constituents v/hich it has lost, viz., blue. We know that 

 these waves are coming to the eye from every part of the sky, and there- 

 fore from that part occupied by the disc of the Moon, and hence the in- 

 ference is natural that this contribution from the store of the Sun's light 

 just makes up what was necessary to produce white light, and that as 

 this accession can go on after the setting of the Sun, and until the 

 twilight circle has passed over the Moon, the whiteness of the latter will 

 commence to fade as the thickness of shell of direct rays diminishes, and 

 the maximum of deviation from the color (under given conditions of the 

 atmosphere) will be reached just after the Sun has reached a point in the 

 heavens whence the last direct ray tangent to the earth's surface falls in 

 l^ie upper limits of the atmosphere on a line joining the Moon with the 

 eye of the observer. 



But there is a practical mode of testing this hypothesis, which is de- 

 pendent upon the polarization of the sky light in directions perpendicu- 

 lar to the Sun's rays. 



When the Moon is in her first quarter she lies in just this direction 

 from the observer ; and since the blue light from the Sun, which, added 

 to her OAvn, causes her to appear white, is polarized, the Moou when 

 viewed through a Njcols' prism by day ought to appear orange. 



This observation has been many times repeated by me, and the results 

 are precisely those anticipated. 



Owing to the fact that there is always some unpolarized light received 

 in this direction the change of color is not quite so marked as is that 

 from day to night, still the change is very striking and unmistakable. 



There is another cause for the blue color of the sky which is the effect 

 of contrast in the eye. If all the light which was reflected was white 

 light and very generally diffused over the firmament, the effect of the 

 bright yellow orb of the Sun or Moon would be to tinge this light with 

 blue so far as the subjective phenomenon was concerned. But that this 

 does not explain the whole of the phenomenon is evident from the fact 

 that the blue light obtained by Tyndall from his decomposition tubes 

 was also polarized in a direction perpendicular to the path of the beam. 



