The retina may be so excited by a strong light that it cannot perceive fainter lights (11): if it perceive these they appear coloured: hence candle-light by day appears reddish, thus resembling, in its relation to fuller light, the spectrum of a dazzling object; nay, if at night we look long and intently on the flame of a light, it appears to increase in redness. 87. There are faint lights which, notwithstanding their moderate lustre, give an impression of a white, or, at the most, of a light yellow appearance on the retina; such as the moon in its full splendour. Rotten wood has even a kind of bluish light. All this will hereafter be the subject of further remarks. 88. If at night we place a light near a white or greyish wall so that the surface be illumined from this central point to some extent, we find, on observing the spreading light at some distance, that the boundary of the illumined surface appears to be surrounded with a yellow circle, which on the outside tends to red-yellow. We thus observe that when light direct or reflected does not act in its full force, it gives an impression of yellow, of reddish, and lastly even of red. Here we find the transition to halos which we are accustomed to see in some mode or other round luminous points. VIII. SUBJECTIVE HALOS. 89. Halos may be divided into subjective and objective. The latter will be considered under the physical colours; the first only belong here. These are distinguished from the objective halos by the circumstance of their vanishing when the point of light which produces them on the retina is covered. 90. We have before noticed the impression of a luminous object on the retina, and seen that it appears larger: but the effect is not at an end here, it is not confined to the impression of the image; an expansive action also takes place, spreading from the centre. 91. That a nimbus of this kind is produced round the luminous image in the eye may be best seen in a dark room, if we look towards a moderately large opening in the window-shutter. In this case the bright image is surrounded by a circular misty light. I saw such a halo bounded by a yellow and yellow-red circle on opening my eyes at dawn, on an occasion when I passed several nights in a bed-carriage. 92. Halos appear most vivid when the eye is susceptible from having been in a state of repose. A dark background also heightens their appearance. Both causes account for our seeing them so strong if a light is presented to the eyes on waking at night. These conditions were combined when Descartes after sleeping, as he sat in a ship, remarked such a vividly-coloured halo round the light. 93. A light must shine moderately, not dazzle, in order to produce the impression of a halo in the eye; at all events the halos of dazzling lights cannot be observed. We see a splendour of this kind round the image of the sun reflected from the surface of water. 94.