DOUBLE STARS. 283 



white light appears green when a strong red light is brought 

 near it ; and that a white light becomes blue when the stronger 

 surrounding light is yellowish. Arago, however, with his 

 usual caution, has reminded us of the fact that even though 

 the green or blue tint of the companion star is sometimes the 

 result of contrast, still on the whole it is impossible to deny 

 the actual existence of green or blue stars. 20 There are 



mer. This is the effect of contrast ; but as red is scarcely 

 ever pure, it may as correctly be said that red is the com- 

 plementary of blue : the colours nearest to the solar spectrum 

 reciprocally change." (Arago, MS. of 1847.) 



20 Arago, in the Connaisance des Temps pour Van 1828, 

 pp. 299-300 ; and in the Annuaire pour 1834, pp. 246-250 ; 

 pour 1842, pp. 347-350 : " Les exceptions que je cite, 

 prouvent que j'avais bien raison en 1825 de n'introduire 

 la notion physique du contrasts dans la question des etoiles 

 doubles qu'avec la plus grande reserve. Le bleu est la 

 couleur reelle de certaines etoiles. II resulte des observations 

 recueillies jusqu'ici que le firmament est non seulement par- 

 seme de soleils rouges etjaunes, comme le savaient les anciens, 

 mais encore de soleils bleus et verts. C'est au terns et a des 

 observations futures a nous apprendre si les etoiles vertes et 

 bleues ne sont pas des soleils deja en voie de decroissance ; si 

 les differentes nuances de ces astres n'indiquent pas que la 

 combustion s'y opere a differens degres ; si la teinte, avec exces 

 de rayons les plus refrangibles, que presente souvent la petite 

 etoile, ne tiendrait pas a la force absorbante d'une atmosphere 

 que developperait 1'action de 1' etoile, ordinairenient beaucoup 

 plus brillante, qu'elle accompagne." "The exceptions I have 

 named proved that in 1825 I was quite right in the cautious re- 

 servations with which I introduced the physical notion of con- 

 trast in connexion with double stars. Blue is the real colour 

 of certain stars. The result of the observations hitherto made 

 proves that the firmament is studded not only with red and 

 yellow suns, (as was known long ago to the ancients,) but also 

 with blue and green suns. Time and future observations must 

 determine whether red and blue stars are not suns, the bright- 

 ness of which is already on the wane; whether the varied 

 appearances of these orbs do not indicate the degree of com- 



