

refractive and dispersive Powers of the Atmosphere, fyc. 203 



Of the varying spectra of the Moon, and the composition 

 of moonlight, I shall speak hereafter. 



The next consideration is, the effects which the above-de- 

 scribed variety in the colour of the light of the celestial bodies 

 will have on our astronomical observations : and it appears 

 to me that this effect will be very considerable. Sir William 

 Herschel has already noticed this circumstance, and has ob- 

 served, that the prismatic power of the atmosphere should 

 not be overlooked, as in observations on very low stars it 

 must make a great difference in the correction which might 

 become necessary. He has stated the measure of two dia- 

 meters of e Sagi'ttarii, and from his observations thereon de- 

 duced the refraction of the extreme rays as being about ^ ¥ + 

 the mean refraction*. 



To me it appears evident that the different stars will, in 

 consequence of the different composition of their light, require 

 very different corrections ; and that tables of refraction should 

 not only be made out for each observatory, in order to apply 

 to the mean refractive power of the air in different places as 

 before described ; but also, that some rules should be ap- 

 pended for applying a correction for each particular class of 

 stars, according to their predominant colour. 



As a natural deduction from the foregoing facts, we must 

 infer that the real declinations of Sirius and of Aldebaran, for 

 example, cannot have been determined by the application of the 

 same correction to the apparent places of both of them ; since 

 Sirius is composed of a large quantity of the most refrangible 

 colours, while Aldebaran is particularly deficient in them. 



We cannot, I think, easily pass by the above facts, without 

 inquiring into their particular causes, which will be found to 

 involve us in some exceedingly curious speculations. 



The difference in the colour of the light of the fixed stars 

 is probably owing to a real difference in the proportions in 

 which the several rays are combined. Thus Sirius sends 

 forth more of the violet ; Aldebaran more of the red ; and 

 perhaps Capella more of the yellow rays. Atair and the 

 dull white stars perhaps have more of the yellow and blue 

 together, and less of the red, so as to make green light; 

 and so onf. The stars differ too in the intensity or brilliancy 

 of their light ; the largest apparently not being always the 

 brightest. And it is curious to observe that some stars seem 



♦ Phil. Trans, vol. lxxv. On Double Stars, by Dr. Herschel. 



f I have worded it thus on the very doubtful supposition of three pri- 

 mitive rays, merely in conformity to received opinion : phaenomena are 

 decidedly against that notion ; for I cannot separate their green light in 

 the spectrum. 



C c 2 to 



