When I look over Stevens' and Norrish's paper, "Border Effects Associated with Photographic Reversal Processes," I am convinced that they have proven that halation ("the scattering of light in the
film") is the primary cause of Sabatier border effects, though they also state that bromides and oxidation byproducts may enhance the border effect. In my article "Controlling the Sabatier Effect" I point
out that "print solarization" might be a good general term for the multiple effects derived by re-exposing an emulsion during development, which include fog, sabatier reversal, and border effects caused by diffusion
At first, when reading the Stevens and Norrish article (from the January 1937 issue of The Photographic Journal), I doubted halation took place in paper at all, as it is so often spoken of as a
film phenomenon. Referring to Mees, Stevens and Norrish state that halation is caused by: "(1) Reflection of the light at the surface of the crystals, predominant with coarse-grained emulsions." and "(2)
Diffraction at the edges of the crystals, predominant with fine-grained emulsions." If these are indeed the causes of halation, then I see no reason why halation should not also take place in a paper emulsion.
But we must bear in mind that all of Stevens' and Norrish's tests were done with glass plate negatives, not prints.
On page 21 of the The Photographic Journal of 1937, Stevens and Norrish state, "The production of these photographs [showing the Sabatier effect] is only made possible by the action of two separate but closely related facts. Firstly, the developed part of the plate is desensitized, and secondly this desensitization is extended for a short distance outside the image of the object." Their paper is concerned with the second, or so-called border effect. They do not address the cause of the first effect, desensitization of certain areas of the emulsion. This desensitization was not effectively addressed until William Jolly, et al., published their first paper "An Explanation of the Sabatier Effect" in 1985.
On page 23, Stevens & Norrish report, "We have concluded...that while the potassium bromide produced in the first development can restrain the second, the effect is not permanent and can be washed out of the plates. No desensitization or destruction of the latent image by potassium bromide has been found, and...it is possible readily to produce both the Sabattier effect and the line under circumstances when potassium bromide is not present." "The line" referred to is of course the Sabatier border effect. Later in their paper they point out that Sabatier reversal takes place even with developers that do not cause a release of bromide. They proceed, by ingenius experiments, to prove that the cause of the Sabatier border effects must be diffusion halation.
On page 26, Stevens and Norrish state, "...potassium bromide causes no such permanent desensitization as is found in the Sabattier effect, and no destruction of the latent image. Any influence that soluble bromides may have on the effect is almost certainly caused by a local retardation of the development, which influence is removed by washing the plate before the second development." Stevens and Norrish are not trying to prove that no bromide reduction is taking place; rather that bromide is not a primary cause of Sabatier reversal, which is why they washed their test plates between the first and second developments. The fact is, with most developers, bromide is produced as a byproduct of development and it does have an inhibitory effect on development where it is present, and Stevens and Norrish clearly acknowledge this.
On page 25, when they discuss the results of their slit experiment, they note, "...the density of the line at the edge of the heavily exposed slits is depressed below the density of the point of reversal. This additional desensitization is almost certainly due to the restraining action of oxidation byproducts diffusing from the main image." On page 30 they repeat this assertion: "This additional desensitization is almost certainly due to the diffusion of reaction products from the main developed image." On page 32, in their summation, they state: "It must be admitted, in view of the known restraining action of bromides and oxidation products, that the lines may on occasions be reinforced by this effect...."
In my experience with print solarization, the presence of bromide definitely contributes to the formation of the reversed image. I encourage this most of the time by not washing my prints between the first and second development. On many occasions I see areas reversing in the tray during the second development--this is not desensitization of the emulsion but clearly is a destruction of image values created during the first development, which I believe could only be caused by bromides released during the latter period of development.