Let a short, lighted candle be placed at twilight on a sheet of white paper. Between it and the declining daylight let a pencil be placed upright, so that its shadow thrown by the candle may be lighted, but not overcome, by the weak daylight: the shadow will appear of the most beautiful blue. 66. That this shadow is blue is immediately evident; but we can only persuade ourselves by some attention that the white paper acts as a reddish yellow, by means of which the complemental blue is excited in the eye.--Note D. 67. In all coloured shadows, therefore, we must presuppose a colour excited or suggested by the hue of the surface on which the shadow is thrown. This may be easily found to be the case by attentive consideration, but we may convince ourselves at once by the following experiment. 68. Place two candles at night opposite each other on a white surface; hold a thin rod between them upright, so that two shadows be cast by it; take a coloured glass and hold it before one of the lights, so that the white paper appear coloured; at the same moment the shadow cast by the coloured light and slightly illumined by the colourless one will exhibit the complemental hue. 69. An important consideration suggests itself here, to which we shall frequently have occasion to return. Colour itself is a degree of darkness _σκιερόν_; hence Kircher is perfectly right in calling it _lumen opacatum_. As it is allied to shadow, so it combines readily with it; it appears to us readily in and by means of shadow the moment a suggesting cause presents itself. We could not refrain from adverting at once to a fact which we propose to trace and develop hereafter.--Note E. 70. Select the moment in twilight when the light of the sky is still powerful enough to cast a shadow which cannot be entirely effaced by the light of a candle. The candle may be so placed that a double shadow shall be visible, one from the candle towards the daylight, and another from the daylight towards the candle. If the former is blue the latter will appear orange-yellow: this orange-yellow is in fact, however, only the yellow-red light of the candle diffused over the whole paper, and which _becomes visible in shadow_. 71. This is best exemplified by the former experiment with two candles and coloured glasses. The surprising readiness with which shadow assumes a colour will again invite our attention in the further consideration of reflections and elsewhere. 72. Thus the phenomena of coloured shadows may be traced to their cause without difficulty. Henceforth let any one who sees an instance of the kind observe only with what hue the light surface on which they are thrown is tinged. Nay, the colour of the shadow may be considered as a chromatoscope of the illumined surface, for the spectator may always assume the colour of the light to be the opposite of that of the shadow, and by an attentive examination may ascertain this to be the fact in every instance. 73.