Colours of Stars. 469 



eliminate that most annoying element of uncertainty, " per- 

 sonal chromatic equation/' and establish the alleged periodi- 

 city, here will be a truly important field for spectrum analysis. 

 But though, in the present state of our knowledge, the 

 Professor is fully justified in calling this " one of the most 

 remarkable cases in the heavens/' we cannot say, considering 

 the recent date of the inquiry, that instances yet more decided 

 may not yet remain to be disclosed. 



If such phenomena attain the rank of established facts, 

 there would seem no other way of accounting for them but by 

 supposing that a change takes place in either — 1, The com- 

 position or temperature of the luminous material, or— 2, The 

 condition of an encompassing atmosphere, or — 3, The acci- 

 dental interposition of unknown media in intervening space. 

 The latter hypothesis, however, which has been suggested by 

 Sir J. Herschel in the case of Sirius, is of course excluded in 

 periodic variations ; and the proof of any one such instance 

 would go far to displace it altogether. The two previous 

 suppositions would come strictly within the range of spectrum 

 analysis. It is much to be regretted that with the exception 

 of Arcturus, whose change of colour is very questionable, all 

 the suspected objects are possessed of an inferior degree of 

 light ; and a material difficulty is thus introduced in the use of 

 the spectroscope ; but if this could be overcome, and if a clear 

 instance of change could be brought under examination, it 

 would be most interesting to ascertain whether any correspond- 

 ing effect was produced upon the dark bands ; and either an 

 affirmative or a negative result would advance our knowledge 

 of the composition of light. In the absence of any con- 

 spicuous instance, what we have now to do is to endeavour to 

 lay so firm a hold upon less obvious phenomena as to place 

 them beyond the reach of suspicion ; and to exercise our 

 ingenuity in devising the best means of so grappling with 

 them. The beautifully executed diagram of tinted discs in 

 the Sidereal Chromatics, will give very valuable help in 

 training the eye to the discrimination of colour, though the 

 plan certainly involves, as has been suggested, some use of 

 " qualifying adjectives/' since the colours of stars are fre- 

 quently incapable of being referred to tints of such a simple 

 character. Other modes of experimenting yet remain, and 

 ought to be tried if possible. 



It follows from the comparisons of Professor P. Smyth at 

 Teneriffe, that in our latitudes no determination of colour can 

 be strictly relied upon beyond 10° of S. Decl. ; but this 

 leaves abundant room for the inquiry; and with a view of 

 saving trouble to those who may be disposed to enter upon it, 

 the following suspicious instances have been selected from our 



