The dyes that should be used for this process are natural carmine or lanafuchsin BB or SL for the red impression, indulin blue for the blue, and acid yellow, mikado yellow, or quinoline yellow for the yellow. The dyes are made up in three per cent solutions in distilled water; but if the carmine be used, then 4 g should be rubbed up into a paste with a little water, and ammonia added drop by drop until a perfectly clear deep red solution is obtained. About 5 ccm of the ammonia will be required, and then water should be added to make the solution up to 100 ccm and it will be ready for use.

As was pointed out, the printing colors are the comple-mentaries of the filter colors, and this applies to all sub-tractive processes, so no further mention of this will be made. The first time of staining, the plates will require about twenty minutes to attain full intensity; but in subsequent staining only about ten minutes. The print-plates may be used for an almost indefinite number of pulls; exactly how many has never been determined, this naturally being dependent on the care with which they are used.

It is actually immaterial in what order they are pulled; but it will be found easier to make the order blue, red, yellow or red, blue, yellow. The reason for this is that it is much easier to obtain registration of the outlines of objects in one of these series than if the yellow be put down first, for this looks so faint against the white paper that it is not easy to secure accurate registration of the outlines. After repeated use, the print-plates become deeply stained even in the whites, and while this is no disadvantage in practice as regards the dyes transferring, it is a great disadvantage from the point of view of easy registration, as it makes it difficult to discern the outlines of the objects. When this occurs, the plate should be immersed in a bath of:

Potassium permanganate 2 g.

Sulphuric acid, 10 per cent solution 10 ccm.

Distilled water 1000 ccm.

As soon as the dye has bleached out, the plate should be immersed in a five per cent solution of sodium bisulphite until all brown stain disappears, and then well washed.

The paper to which the dye image is transferred is plain gelatine-coated paper, and this gives one a wide choice of surfaces. One can easily obtain a suitable paper by fixing out bromide or developing paper, washing and treating with five per cent solution of formaldehyde for ten minutes, and drying without washing. After the print-plate has been soaked in the dye solution for a sufficient time, it should be rinsed with water until the drainings are practically colorless. A sheet of paper should be soaked in water for about five minutes, left for half a minute after the surface moisture has been blotted off, then the print-plate squeegeed down and the two left in contact. The paper should be uppermost, a thin piece of wax paper placed over it and the squeegee used well, stroking from the middle of the print out to the sides; roller squeegees are not so suitable for this as the old fashioned flat ones. The paper and plate should be left in contact for about fifteen minutes, and then a small corner of the paper turned back and the image examined to see if it is deep enough. Naturally, a little experience is required to tell this, but one soon becomes expert enough. If the pull is not thought sufficiently intense, the corner must be squeegeed down again and the whole left a little longer, and again examined.

The next plate is treated as described for the first, but as at first one is a little awkward in obtaining register, a piece of thin celluloid is placed over the first pull. The thinnest celluloid that can be obtained should be used, and creases or folds in this should be avoided; the celluloid should be cut about an inch longer than the print, as this gives one a convenient handle to catch hold of, and it should be placed over the first print so that a little space is left at the top of the print with no celluloid. The idea is that this space enables one to temporarily clip or hold the second plate in contact with the print, after registration has been obtained, and then withdraw the celluloid without shifting the relative positions of the second print-plate and the paper. The easiest way to work is to form a temporary desk by supporting a piece of thick glass on two piles of books or blocks of wood, and place under this a mirror or white card that will reflect the light through the paper and the second plate. Then, with a magnifying glass, it will be found fairly easy to obtain coincidence of the outlines. Then, if the plate and the paper are firmly held at the top, the celluloid may be slipped out and the two well squeegeed together.

The third print is obtained in exactly the same way as just described. Under no circumstances must the one pull be allowed to dry before the next impression is superimposed. Therefore, it may be as well to place the print-plate with the adherent paper between sheets of stout blotting paper well dampened with water.

Working in this way, particularly if the gelatined paper has been well hardened with formaldehyde, there is but little danger of want of exact register from expansion or shrinkage of the paper. Only in the case of very soft paper stock is there any danger. In such cases, the paper should be soaked in water for at least half an hour before making the first pull. A still more radical remedy is to pin the paper with the gelatine side down on a board, and paint the back with a celluloid varnish, such as:

Scrap celluloid 20 g.

Amyl acetate 50 ccm.

Acetone 450 ccm.

Methyl alcohol to 1000 ccm.

Two or three coats of this thinly applied will prevent any expansion or contraction. The paper, after this treatment, must be left at least twenty-four hours to thoroughly dry.

If, when the three impressions are pulled in superposition, it is found that one or other color predominates, and of course the depth of the staining is dependent on the time of contact between the paper and the plates, one of the plates may be again soaked in the dye solution and again superimposed in register and left for a short time. Thus the excess of red may be killed by a second staining up with blue, and excess of blue by a second staining up with yellow or red. If the whole picture is too weak, all three plates may be reapplied. If it is too dark, it may be moistened in water and squeegeed in contact with a gelatine coated plate for a short time, and some of the excess dye will migrate into this plate. A negative must not be used for this, as the silver image prevents the even transfer of the dye, but naturally a plate with the silver removed can be used, and by careful and frequent examination of the corner of the print as already advised, one can soon tell when the reduction has proceeded far enough.

The prints thus obtained are sufficiently permanent to withstand months of exposure to any daylight that may be met with in an ordinary room; but they may be made more permanent by five minutes immersion in a three per cent solution of cupric sulphate. This turns the reds into a more violet hue, and it is hardly necessary.