ALTHOUGH the results of the screen processes are often of exceeding beauty, and are in some ways far in advance of any other method of automatically recording the colours of nature, they have not hitherto been applied directly to reproduction on paper. For this we must still rely on the three-colour negatives.

In Chapter XXXIX (Photography In Natural Colours), we saw that by selecting three special colour points in the spectrum, a mixture of these will suffice to convey all the colour sensations produced by the intermediate colours. The exact points to be selected for the fundamental colours vary according to different authorities: Clerk Maxwell chose 630, 530, and 457; Konig 675, 508, and 475; Grunberg 665, 506, and 482. Sir W. Abney, working with more accurate instruments than were at Clerk Maxwell's disposal, does not differ very widely from the latter, but selects a red near the red lithium line, a green near the E line, and a violet near the blue lithium line. The nerve fibrils stimulated by the particular coloured rays have also considerable sensitiveness towards the rays of adjoining parts of the spectrum. Thus the nerves most sensitive to the green are also to a considerable extent sensitive through the orange to the red on one side, and to the blue on the other, the blue sensitiveness also applying to the green. This sensitiveness is expressed by Clerk Maxwell and others by curves, which indicate at highest point the greatest degree of stimulation and taper off according as the sensitivity of the particular set of nerves decreases.

If we wish to obtain three negatives, which are to be used for the expression of natural colour, they must correspond with these three curves. Such a combination will then be more or less interpreters of the three sets of colour nerves in the human eye. The greatest densities of each when used for transparencies and superposed over each other must give a white light.

Chromatic Plates

Except for some photo-mechanical work, it is no longer the custom to sensitise three different sets of plates for the three selected colours red, green, and blue-violet, but to employ for all alike a plate sensitive to the whole spectrum. Such plates are usually stained with cyanine, pinachrome, or still better in a solution of pinachrome and pinacyanol, in the proportions of 3 to 2. Prof. Namias, while recommending ethyl-violet and erythrosine in the proportions of 3 to 2, remarks that pinachrome gives a band which runs from ultra violet to nearly all the red with two inconsiderable minima, one between G and F (blue-green) and the other between D and E (yellow-green). Plates treated with ethyl-violet and erythrosine must be exposed through the red fitter for at least double the time of plates treated with pinachrome. A pinachrome solution (1 gram to 500 c.c. pure alcohol 950 and 500 c.c. distilled water) is diluted for use with fifty times its bulk of pure water, and the plates immersed in the dark for fifteen minutes. After this they are rinsed, plunged for a moment in an alcohol bath, and dried in the calcium-chloride box. With the arc light, pinachrome is almost perfect as a sensitiser. The amateur will be well advised, in view of the considerations involved, if he contents himself with the ready-prepared panchromatic plates and the filters recommended by the makers concerned.

The Colour Curves According To Clerk Maxwell.

Fig. 71. The Colour Curves According To Clerk Maxwell.

Colour Filters

The exact tints of the three filters, orange, green, and blue, will differ slightly according to the light employed, and the dyes with which the plates are sensitised. Earlier workers were satisfied with glass filters tinted orange red, chromium green, and cobalt blue. Glass filters reliable enough for many purposes may still be obtained, and they have the advantage over the aniline dyes of being less liable to lose colour under the action of light. A good selection of filters may be obtained from firms such as Wratten & Wainwright, or Sanger Shepherd & Co. The advanced worker will do well to consult the Atlas of Absorption Spectra, lately completed by Dr. Mees and S. H. Wratten. Merely for the sake of example we will give filter dyes suitable for plates sensitised in pinachrome.

Blue

Methylene (1/2 per cent. solution) . . . . 50 min. Water......... 50 ,,

Green

Methylene (1/2 per cent. solution) ... 30 min. Auramine G. „ „ ,,... 180 „

Orange

Erythrosine (1/2 per cent. solution) . . . . 35 min. Metanile Yellow (sat. solution) at 6o° Fahr. . . 40 „ and these solutions may be adapted for liquid filters, dry filters, or for staining the plates themselves.

But the question of filters is scarcely one on which it is wise to dogmatise. Three different classes of filters are in common use for colour work, and all seem, in the hands of skilful operators, to produce excellent work:

(1) Filters that divide the spectrum into three bands, from some point in the red to D, from D to F, and from F to H, forming three zones without overlap or gaps.

(2) Filters forming three zones, which considerably overlap each other.

(3) Filters whose zones contain gaps or lacunae.

The first system is the older one and might almost be called the orthodox method, but the second has very strong support and is rapidly gaining favour.

Exposures For Landscape

The ordinary landscape dark slides are unsuitable for three-colour work, owing to the impossibility of securing exact registration of the three negatives. No two dark slides can be depended upon to hold the plates in precisely the same relative position. During the withdrawal and replacement of a slide the camera is certain to shift, and, however slight this movement, it will destroy any hope of getting the negatives to correspond. A "repeating back" must be employed comprising three screens of the proper colours, which fit on to a dark slide containing one long, narrow plate, on which the three exposures are successively taken, a spring catch denoting when each third of a plate is in position. There are difficulties in gauging the relative exposures for the three screens, which may vary from 4:2:1 to 3:2:1 1/2 according to the type of plate and the intensity of light. One great difficulty about these three exposures in the open air is that, owing to atmospheric changes, the light may vary in quality. If one plate is taken in clouded weather and the next in sunshine, the results will be inharmonious. The exposures must also be such that plates of corresponding density are obtained; and they must be accurate within 2 per cent.