Now "I'm" curious; why would you want to do this?

The fewer components there are in a given solution, the easier it would be to keep it in balance. For example, in an extreme case, if the solution could be subdivided all the way down to a single ionic compound dissolved in water per reservoir, mere osmotic pressure alone could keep the concentration at a target level.

I'm an experimenter by nature, so I'm just curious as to whether anyone already has any data along these lines.

It sounds like a viable theory to me. I don't see why it wouldn't work, aside from the trouble of making the system more complicated as well.

It would probably be easiest to implement in drain to waste, else you'd need some automated valve to control the return solution to the correct reservoir. Plus you'd always get a certain amount of lechate from the other solution from the medium.

Complicated? Absolutely.

Would it work? Probably.

Would it work better? I'm skeptical.