Special sliding backs can be obtained commercially, fitted with the three filters and made to take three plate-holders, or with some plate-holders, particularly of the English book-form pattern, the filter may be placed in actual contact with the sensitive surface; then naturally its thickness must be allowed for in focusing.

One important point in the choice of filter fittings, particularly metal ones, is that there should be no abnormal pressure on the glasses, as this may cause strain and consequent degradation of definition. It should be possible to turn the filter round, or shift it, with the lightest pressure of the fingers. Neither is it advisable to use cells screwing into the lens hood, as this is almost certain to shake the camera and there is much loss of time in changing.

For photomechanical work, in which long-focus lenses are nearly always used with half-tone screens, the glass must be optically worked, as carefully, in fact, as the lenses themselves. Such glasses are known commercially as "optical flats," and are very costly if of any size. They must all be absolutely the same thickness and be so arranged that they are always perpendicular to the axis of the lens.

The ordinary filters may be used for making the separation negatives for photomechanical work, as if there is not absolute coincidence of size, this can be corrected by the operator when making the screen negatives from the transparencies, though he will not be pleased at having to do this.

It may possibly be as well to interpolate here a note as to the making of the constituent negatives for photomechanical purposes. The use of the panchromatic gelatine plate for this work is largely on the increase, and in some cases the slow panchromatic plate is used for making the color separation and the screen negatives in one; but the usual practice is to make the separation negatives first, from these a set of transparencies, and then the screen negatives. It may be noted that the transparencies for such work should be as little like a lantern slide as possible. They should be fully exposed, quite "soft" in character, and with practically no bare glass except in the very deepest shadows. Full exposure should be given to the plates, and it is better to use slow negative rather than transparency plates for this work, as giving a longer range of gradation and less tendency to brilliancy. The exposure should be full and development not pushed too far, so that the highest densities are quite transparent.

In many commercial process establishments collodion emulsion, and even the wet plate process, still hold their own for the making of the separation negatives, and also the combined separation-screen negatives. Usually the emulsion is obtained commercially with its special sensitizers, and the makers issue instructions for the making of the filters, which are usually of the liquid cell type, for use with the same. On the other hand the method of sensitizing already advised may be adopted, or the dye may be added to the emulsion, and in this case 80 ccm of sensitol violet stock solution should be added to 1000 ccm of the plain emulsion, and the plates washed in running water or under a rose tap for fifteen minutes. The washing increases the sensitiveness of the plates about five times.

By some writers it has been proposed to use different plates for the different color separations, that is to say, an ordinary, non-color-sensitive plate for the minus yellow negative; an orthochromatic plate for the minus red negative; a panchromatic or red-sensitive one for the minus blue negative. This plan may at first sight appear to have certain advantages, but this method is not one that should be adopted. It is a well established fact that the degree of contrast, or gamma, differs with different kinds of plates, and in fact with different batches of the same kind of plate, to say nothing of the development velocity of the plates, and one of the most important essentials in making separation negatives is to have them of the same degree of contrast as far as possible. That is to say, in the three negatives the range of densities of a black and white scale should be the same; and with three totally different kinds of plates this is almost an impossibility. One kind of plate should be used for all three separation negatives, and they should be, as already pointed out, as far as convenient, developed together. The adoption of this plan will save no end of after manipulation and dodging in getting concordant results; and it may be taken as an axiom that hand work, except for the removal of purely mechanical defects, such as pinholes, etc., cannot be successfully executed with color negatives.