Colours of Stars. 441 



same eye at different times from varying conditions of the 

 retina. To say nothing of the probability that a wearied eye 

 would be less sensitive to slight differences of tint than a fresh 

 one, it is well known that the long- continued impression of 

 light of any decided colour is succeeded, upon its removal, by 

 the appearance of the complementary hue — a fact which may be 

 illustrated in a pleasing manner by closing one eye, and look- 

 ing with the other at a white object through a piece of strongly 

 tinted glass ; this having been continued for a sufficient time 

 till the sight is accustomed to it, let the glass be suddenly 

 taken away, when the complementary colour will fill the whole 

 field of vision, to an extent that will be fully manifested by 

 opening the closed eye, which of course will see only white 

 light. In exactly the same way, the eye which has been long 

 gazing upon a bright yellow star, on turning the telescope to 

 a white one, will see it tinged with the complementary purple, 

 or, if the star was of a red hue, with the corresponding green. 

 The cause of this phsenomenon may probably be, that diminution 

 of sensibility under a long-continued and unvaried stimulus 

 which is common to all our perceptions. The retina becomes 

 gradually less responsive to the action of any colour, just as it 

 is to the action of strong white light, from prolonged exposure 

 to its unmixed influence ; and therefore when light composed 

 of various tints is subsequently let in upon it, it fails in the 

 adequate perception of that hue to which it has become as 

 it were deadened, and catches chiefly the impression of the 

 other colours in the compound, until the retina has had time 

 to recover its normal condition . On this account, the observa- 

 tion of colour should never be attempted after micrometrical 

 measurement, in which artificial illumination is employed ; nor 

 indeed at any time when the retina has just been previously 

 stimulated by lamp or candle-light. So Struve I., who paid 

 great attention to colours (while his assistant Knorre could dis- 

 tinguish none !), has cautioned us that tints observed by day- 

 light are not to be depended upon ; the impression of the blue 

 background predisposing the eye to ascribe a tinge of comple- 

 mentary orange to the star, exactly as we have seen the fight 

 of a cloudy sky, penetrating through a hole in a window of 

 greenish glass, appear distinctly of a lilac colour. 



Another inquuy, and rather a troublesome one, springs 

 out of this relation of the retina to colour. Since a coloured 

 light of predominant intensity will obviously tinge all lesser 

 lights in its neighbourhood with the complementary hue,* a 



* It is probable that intensity of hue may occasionally overbalance mere 

 quantity of uncoloured light. At least the Blight green tint of Sirius when brought 

 into the same field with Arcturus in Chacomac's ingenious experiment on their 

 relative brightness, may be reasonably ascribed to that cause. 



