There are various modes of producing coloured images in objective experiments. In the first place, we can fix coloured glass before the opening, by which means a coloured image is at once produced; secondly, we can fill the water-prism with coloured fluids; thirdly, we can cause the colours, already produced in their full vivacity by the prism, to pass through proportionate small openings in a tin plate, and thus prepare small circumscribed colours for a second operation. This last mode is the most difficult; for owing to the continual progress of the sun, the image cannot be arrested in any direction at will. The second method has also its inconveniences, since not all coloured liquids can be prepared perfectly bright and clear. On these accounts the first is to be preferred, and deserves the more to be adopted because natural philosophers have hitherto chosen to consider the colours produced from the sun-light through the prism, those produced through liquids and glasses, and those which are already fixed on paper or cloth, as exhibiting effects equally to be depended on, and equally available in demonstration.
As it is thus merely necessary that the image should be coloured, so the large water-prism before alluded to affords us the best means of effecting this. A pasteboard screen may be contrived to slide before the large surfaces of the prism, through which, in the first instance, the light passes uncoloured. In this screen openings of various forms may be cut, in order to produce different images, and consequently different accessory images. This being done, we need only fix coloured glasses before the openings, in order to observe what effect refraction produces on coloured images in an objective sense.
A series of glasses may be prepared in a mode similar to that before described ([284]); these should be accurately contrived to slide in the grooves of the large water-prism. Let the sun then shine through them, and the coloured images refracted upwards will appear bordered and edged, and will vary accordingly: for these borders and edges will be exhibited quite distinctly on some images, and on others will be mixed with the specific colour of the glass, which they will either enhance or neutralize. Every observer will be enabled to convince himself here again that we have only to do with the same simple phenomenon so circumstantially described subjectively and objectively.