I've been thinking about the repulsion problem. Repulsion of oil, the beading of medium, an effect like painting water on wax, has been a noted problem for fine oil painters for centuries.
The traditional solution to this problem is to rub a raw onion or raw potato onto the picture between layers, and the few traditional painters that I could find still do this. I also found an article report from an egg tempera painter with a similar problem, which he solved by applying a layer of ox-gall (a fluid most commonly associated with watercolours). He had also found that the powerful solvent acetone worked but stopped doing this after he was told that the integrity of the underlayer might be damaged by this. Dali recommended a raw potato, then stated that it's best to take the precaution of rubbing the potato through a fine silk rag, then washing the picture with water, drying it then apply retouching varnish.
I wanted to know what caused this problem and what the onions and potatoes did, and I've come up with a theory. I read from a conservator that all oil paintings crack, it's just a matter of time. This made sense when I thought about it.
I imagined that each layer of paint was not the smooth cellophane of plasticised oil that a glazing artist might imagine, but full of gaps. The gaps might be caused by uneven oil application or solvent evaporation. These gaps, like holes in a bag of rubber bands, the bands like the long oil molecules, grow to become cracks over the years. The fatter, more polymerised, oils would be more long and bandlike, more flexible but have smaller holes.
These holes also give a tooth to overpainted layers, making the application of future oil possible without repulsion. Where the paint is applied evenly and with no solvent, the gaps do not appear or are too small to have this tooth, and so the repulsion of oil occurs.
If this theory is correct, the adhesion problem can be solved by either damaging the existing paint surface to roughen it, or applying a new layer that adds tooth. I'm not sure what the onion does but the potato deposits a layer of starch which must act as this tooth. What it does not do then, is dissolve or digest or absorb the oily surface below (which is what many painters of old seemed to think). Washing the starch away will only stop it working, which might explain why Dali appended the rather odd postscript about adding retouching varnish (which would add the tooth and disrupt the paint surface alone, without the need for the potato at all).
Acetone would damage the molecules to add these microscopic cracks. Perhaps the acids in the onion do the same but I suspect it only deposits starch too, especially as fine sweet Spanish onions were preferred. Depositing a thin tooth layer that is not disruptive to pigments or oils is the solution then, and if starch does not discolour, and to my knowledge it does not, then the good old raw potato should be fine.
If this theory is true then it would be ironic that the smoothest, most stable and crack resistant areas of oil paint are the ones that are impossible to paint over unless the painter first damages that perfect surface.
