With regard to screen plates made by other processes, there are not at present any on the market, though several patents have been taken out, such as Joly's (i9,388 of 1895) for dyed threads laid on a transparent support, Beeton and Gambs' patent (20,834 of i906) for a woven tissue, and Szczepanik's (17,065 of 1908) on the same lines.
The Szczepanik-Hollborn screen plates are based on the affinity of basic dyes for acid tissues and acid dyes for basic tissues ; the former have a decided preference for collodion and acid dyes for gelatine. Three solutions of gelatine or other colloid are dyed, dried, powdered, mixed in proper proportions, and dusted on to a moist collodion plate. The dyes pass from the colloid into the collodion, and the former is washed off, leaving a mosaic of coloured areas. In the second form only two colours are applied in this way and the third applied by a dyeing bath.
There are certain factors in the manufacture of screen plates which are extremely important for the successful reproduction of colour, and the first is that the screen itself should be free from colour when examined by white light ; that is, it should be of a neutral shade. This has been defined by Mees as the " first black condition," and it must be attained by adjustment of the areas of the colour units, and not by varying the depth of staining which controls their absorp tions. The second point is that the photo-chem ical effect of the spectrum through the screen elements and compensating filter must corre spond with the luminosity curve, though this point may be of less importance, as the retina is able to perceive small variations from the correct curve. The third point is the total visual absorp tion, as this affects the duration of exposure, for, naturally, the greater the absorption of light the longer will be the exposure. The above mentioned authors state that in order to fulfil the first black condition the green area must transmit about two-thirds of the light transmitted by the plate, and under the best conditions for the other filters it may occupy half the area of the plate, and therefore half the plate will only transmit one-third of the incident green light or two-ninths of the incident white light, so that the whole plate will transmit only one-sixth of the incident white light as a maximum.
The question of invisibility of the filter ele ments depends upon what has been termed the " period," which is twice the distance of separa tion for lines equal in width to the spaces ; and this must be less than r 5. of the distance from the eye ; and if this be taken as 20 cm. the screen period will be 0.2 mm. ; therefore the separate filter elements will be invisible if they are not larger than AT in. (.066 mm.), or approximately in., this naturally applying to results examined in the hand. When, however, it comes to the projection of the pictures, one has to take into consideration the magnifica tion, and assuming this to be 4o diameters or a xo-ft. screen, and the nearest observer to be 12f ft. away, the screen elements must not be more than in., which requires the actual screen elements to be in. When dealing with irregular grain screens in which the units are distributed by a dusting-on method, then there may be clumping of the grains ; as many as twelve grains may be dumped together, and therefore the unit may become much larger.
The resolving power—that is, the power of the screen plate to resolve a coloured object into its form and colour—may be divided into three heads. First, the objects are resolved both in form and colour, when the images of the objects are as large as or larger than the screen period ; secondly, the images may be of the same size as the screen elements, then they will be resolved as to form but indeterminate as to colour ; thirdly, if the images are smaller than the screen units, they may or may not be resolved, accord ing to the resolving power of the emulsion.
The spectral absorptions of the filter elements can only be those which are generally recognised in three-colour work—that is, having a slight overlap in the yellow and blue, considering the taking of negatives only ; but as, in the majority of cases, the screen is also used for viewing, the spectral absorption should be as pure as pos sible, and should agree as nearly as possible with the three fundamental colours—red, green, and blue. A compromise has therefore to be made, and probably the red should transmit from the extreme red to A 5,900, the green from A 5,900 to A 4,900, and the blue from A 5,000 to A. 4,000.