(c) The Sanger Shepherd Process.—This is employed mostly for lantern transparencies. As in the Ives process, three negatives and three transparent positives are made. But instead of coloured glasses being used to give effect to the positives the positives themselves are dyed, and placed one on the top of another in close contact, so that the light from the lantern passes through them in succession. We have therefore now quitted the realms of harmony for that of discord, in which prisoners are made; and Mr. Shepherd has had to so arrange matters that in every case the capture of prisoners does not interfere with the final result, but conduces to it.
In the first place, three negatives are secured through violet, green, and red screens. Positives are printed by the carbon process on thin celluloid films. The carbon film contains gelatine and bichromate of potassium. The light acts on the bichromate in such a way as to render the gelatine insoluble. The result is that, though in the positives there is at first no colour, patches of gelatine are left which will absorb dyes of various colours. The dyeing process requires a large amount of care and patience.
Now, it would be a mistake to suppose that each positive is dyed in the colour of the screen through which its negative was taken. A moment’s consideration will show us why.
Let us assume that we are photographing a red object, a flower-pot for instance. The red negative represents the pot by a dark deposit. The positive printed off will consequently show clear glass at that spot, the unaffected gelatine being soluble. So that to dye the plate would be to make all red except the very part which we require red; and on holding it up to the light the flower-pot would appear as a white transparent patch.
How then is the problem to be solved?
Mr. Shepherd’s process is based upon an ordered system of prisoner-taking. Thus, as red in this particular case is wanted it will be attained by the other two positives (which are placed in contact with the red positive, so that all three coincide exactly), robbing white light of all but its red rays.
Now if the other positives were dyed green and violet, what would happen? They would not produce red, but by robbing white light between them of red, green, and violet, would produce blackness, and we should be as far as ever from our object.
The positives are therefore dyed, not in the same colours as the screens used when the negatives were made, but in their complementary colours, i.e. as explained above, those colours which added to the colour of the screen would make white.
The red screen negative is therefore dyed (violet + green) = blue. The green negative (red + violet) = pink. The violet negative (red + green) = yellow.