If the object is large, its centre remains uncoloured. Its inner surface is then to be considered as unlimited (195): it is displaced, but not otherwise altered: but if the object is so narrow, that under the above conditions the yellow border can reach the blue edge, the space between the outlines will be entirely covered with colour. If we make this experiment with a white stripe on a black ground,[1] the two extremes will presently meet, and thus produce green. We shall then see the following series of colours:-- Yellow-red. Yellow. Green. Blue. Blue-red. 215. If we place a black band, or stripe, on white paper,[2] the violet border will spread till it meets the yellow-red edge. In this case the intermediate black is effaced (as the intermediate white was in the last experiment), and in its stead a splendid pure red will appear.[3] The series of colours will now be as follows:-- Blue. Blue-red. Red. Yellow-red. Yellow. 216. The yellow and blue, in the first case (214), can by degrees meet so fully, that the two colours blend entirely in green, and the order will then be, Yellow-red. Green. Blue-red. In the second case (215), under similar circumstances, we see only Blue. Red. Yellow. This appearance is best exhibited by refracting the bars of a window when they are relieved on a grey sky.[4] 217. In all this we are never to forget that this appearance is not to be considered as a complete or final state, but always as a progressive, increasing, and, in many senses, controllable appearance. Thus we find that, by the negation of the above five conditions, it gradually decreases, and at last disappears altogether. [1] Plate 2, fig. 5, _left_. [2] Plate 2, fig. 5, _right_. [3] This pure red, the union of orange and violet, is considered by the author the maximum of the coloured appearance: he has appropriated the term _purpur_ to it. See paragraph 703, and _note_.--T. [4] The bands or stripes in fig. 4, plate 1, when viewed through a prism, exhibit the colours represented in plate 2, fig. 5. XV. EXPLANATION OF THE FOREGOING PHENOMENA. 218. Before we proceed further, it is incumbent on us to explain the first tolerably simple phenomenon, and to show its connexion with the principles first laid down, in order that the observer of nature may be enabled clearly to comprehend the more complicated appearances that follow. 219. In the first place, it is necessary to remember that we have to do with circumscribed objects. In the act of seeing, generally, it is the circumscribed visible which chiefly invites our observation; and in the present instance, in speaking of the appearance of colour, as occasioned by refraction, the circumscribed visible, the detached object solely occupies our attention. 220.