of the Atmosphere, and on the Peculiarities of Stars. 33 1 



neously produced all over the star at once: it seems sometimes 

 to ascend from the lower limb upwards, like the drawing up 

 of a curtain, or like the progressive ascent of light thrown 

 on to a disk by the gradual inclination of a reflecting plane 

 mirror, occupying (though the time is perhaps a deception of 

 our organs) on different occasions + ^" of time in passing 

 from the lower to the upper edge of the spectrum: but it does 

 not recede again in the same direction; it suddenly seems to 

 vanish, either leaving the disk of the star of its usual colour 

 and brilliancy, or giving place to a blueish colour. In Sirius 

 I have noticed that indigo, violet, and greenish white change 

 with red. I have been enabled to determine the avenige 

 number of the fits of redness in Antares as being three in S" of 

 time; and generally lasting from |" to 1", though sometimes 

 longer. In Sirius the red is paler, there is more variety, and 

 the fluctuations are much more rapid. Another thing ob- 

 served in the expanded spectrum is, that the intensity of the 

 alternating red light, like that of the general light of the star, 

 varies inversely as the size of the spectrum ; so that instead 

 of a deep and intense red, such as we see with the naked eye, 

 or when we adjust the telescope to the proper focus; we see a 

 more diluted red or pink light, alternating with other di- 

 luted colours according to the prevailing colour of the parti- 

 cular star that we are viewing. 



I have not devoted a separate column for the colours and 

 relative prevalence of the fluctuations of the stars, correspond- 

 ing to the fourth and fifth columns of my last table, Philoso- 

 phical Magazine, p. 208, as it seemed unnecessary. It may 

 suffice to observe generally, that it is only certain stars which 

 show this phaenomenon in any considerable degree, and most 

 of those have been before pointed out. This coloured fluctu- 

 ation, when observed in the prism, presents a very curious and 

 beautiful phaenomenon ; and what is very remarkable is, that 

 whatever colours anv particular star may afford in the disper- 

 sive lens, the same colours will be produced by viewing the 

 star in an ordinary telescope, and simply vibrating it at the 

 time of observation : and this led me to discover a new method 

 of measuring the proportions of the colours, hereafter to be 

 described. A great degree of fluctuation interferes considerably 

 with bur endeavours to estimate the refrangibility of different 

 stars, founded on the dispersion of their colours ; because it 

 makes the spectrum, as viewed in the glass, vary so perpe- 

 tually, that we cannot easily measure the deviation of the ex- 

 treme rays with the micrometer ; and therefore in stars which 

 are liable to an inordinate degree of this fluctuation we must 

 always make greater allowance foi probable error. The third 

 T I 2 column. 



