THIS process was invented by Professor R. W. Wood, of Johns Hopkins University, and is a beautiful application of the phenomena of diffraction by gratings. It is, however, merely a laboratory process, as the results are only capable of being seen by one person at a time or of being projected on a very limited scale, nor, so far as the author is aware, are the necessary diffraction gratings commercially available. On the other hand, it is a comparatively easy process and capable of indefinite reproduction.

The theory involved is simple, and no explanation will be required for those familiar with the phenomena of gratings. A grating is a sheet of metal or glass ruled with extremely fine lines, ranging from 2000 to 20,000 per inch. When white light falls on a glass grating it is split up into a series of spectra, as shown in Fig. 24.

Fig. 24.

The light passing straight through the grating is seen as a white central band C; on each side are seen the spectra with the violet end nearest the central image. The spectra are called the first, second, third, etc., to the nth order spectra, according to their position as regards the central beam; and these spectra become longer and fainter the higher their order. Theoretically there is no limit to the number of spectra that can thus be formed, but the first order spectrum is the only one used in this process, because it is the brightest. The distance from the central beam depends entirely upon the number of lines per unit or inch, and the finer the ruling, that is, the greater the number of lines per inch, the further removed are the spectra from the central image. It is obvious, then, that we can so choose the rulings that any three colors may be on the same plane, as roughly shown in Fig. 25, in which the light and the gratings are supposed to be behind the page, and we are looking on the spectra formed on the paper. With three gratings, that which gives the image A has the coarsest ruling, that which gives B being finer, and C being finer still. If these rulings are correctly chosen, the red of A, the green of B, and the blue of C will all appear on the same vertical line 00.

Fig. 25.

It has already been pointed out that from the three lights red, green and blue we can produce all colors, so obviously in this case we are in a position to reproduce all colors by the use of three gratings, either by projecting the spectra on one plane or by superposing the gratings; and this is the fundamental basis of the diffraction process. The three gratings are printed on one bichromated gelatine film.

The rulings chosen by Professor Wood had 2000, 2400 and 2750 lines per inch for the red, green and blue respectively, but one is not tied to these particular numbers, finer gratings being used in the same ratio.

Three constituent negatives are required, taken in the usual way through the red, green and blue selective filters. From these negatives three positives in silver must be made as usual; from these positives the final result is obtained. The support for the final picture must be plate glass, as ordinary glass is not flat enough. The sensitive mixture is:

Gelatine 40 g Potassium bichromate, saturated solution 50 ccm.

Distilled water 1000 ccm.

Soak the gelatine in the water for fifteen minutes, then dissolve by heat and add the bichromate solution, and filter while hot. The glass must be thoroughly cleaned and polished. Enough of the solution is poured on to cover the surface, the excess drained off for about ten seconds, and then the plate placed on a leveled slab to set and dry in the dark. The coating should be thin and the plates will dry in about two hours.

There are two methods of printing, by contact or by projection. Of the two possibly the latter is the easier, although, unless an arc lamp is available, the exposure is unduly prolonged; with an arc it is approximately from one to two minutes. The main difficulty is the registration of the images, though this may be gotten over by making ink dots on the glass side of the positives and bringing these into register. It is as well to support the two positives in a clip, such as is used for holding lantern slides for binding, and, with the aid of a focusing magnifier, bring the two images into registration; then, with a pen, put a dot of ink on the glass of the positive nearest the eye corresponding with some clearly marked object at one side of the picture, and also make another dot at the top or bottom of the positive. Then reverse the positives, putting that which was nearest the eye farther away; again make the images coincide and put on ink dots to register with the first ones; repeat this for the third positive.

A lens must be used to project the image of the positive on the sensitive plate. This lens may be supported in a V-shaped block of wood or clamped in an ordinary retort clamp, as this latter enables one to adjust it laterally and horizontally. Assuming that the positives are to be reproduced the same size, then, as is well known, there must be just double the equivalent focus between the lens and the positive and the same distance between the lens and the plate. No elaborate means are required for holding the sensitive plate, but it must be held rigidly in the same position for the three exposures. The positives are placed near the arc; it should be possible to place all three in approximately identical positions, and this can be effected if they are placed in a printing frame and jammed down into one corner. A sheet of yellow glass, or a fixed-out plate deeply stained with tartrazin, must also be provided.