That's right, I said dependent. Does anyone have any experience dealing with such beasts. I haven't been able to find a single mention of them in any textbook on PDEs. The thing I'm really curious to know is whether the method of separation of variables works as usual, e.g. if the dep vars are f and g, can I legitamately write [itex]f = h(x)\varphi(y)[/itex] [itex]g = f(x)\psi(y)[/itex] you may assume of course that the PDE has been linearized.