Hence we may safely deduce the following rules to determine before-hand the colours of all spectra. 1.  The direct spectrum without any lateral light is an evanescent representation of its object in the unfatigued eye. 2.  With some lateral light it becomes of a colour combined of the direct spectrum of the central object, and of the circumjacent objects, in proportion to their respective quantity and brilliancy. 3.  The reverse spectrum without lateral light is a representation in the fatigued eye of the form of its objects, with such a colour as would be produced by all the primary colours, except that of the object. 4.  With lateral light the colour is compounded of the reverse spectrum of the central object, and the direct spectrum of the circumjacent objects, in proportion to their respective quantity and brilliancy.

2. Variation and vivacity of the spectra occasioned by extraneous light.

The reverse spectrum, as has been before explained, is similar to a colour, formed by a combination of all the primary colours, except that with which the eye has been fatigued in making the experiment:  so the reverse spectrum of red is such a green as would be produced by a combination of all the other prismatic colours.  Now it must be observed, that this reverse spectrum of red is therefore the direct spectrum of a combination of all the other prismatic colours, except the red; whence, on removing the eye from a piece of red silk to a sheet of white paper, the green spectrum, which is perceived, may either be called the reverse spectrum of the red silk, or the direct spectrum of all the rays from the white paper, except the red; for in truth it is both.  Hence we see the reason why it is not easy to gain a direct spectrum of any coloured object in the day-time, where there is much lateral light, except of very bright objects, as of the setting sun, or by looking through an opake tube; because the lateral external light falling also on the central part of the retina, contributes to induce the reverse spectrum, which is at the same time the direct spectrum of that lateral light, deducting only the colour of the central object which we have been viewing.  And for the same reason, it is difficult to gain the reverse spectrum, where there is no lateral light to contribute to its formation.  Thus, in looking through an opake tube on a yellow wall, and closing my eye, without admitting any lateral light, the spectra were all at first yellow; but at length changed into blue.  And on looking in the same manner on red paper, I did at length get a green spectrum; but they were all at first red ones:  and the same after looking at a candle in the night.