“…There we avoid such techniques, by appealing to the natural change of coordinates associated with a cylindrical surface. Actually, the choice for the domain in [9] enabled us to pass to cylindrical coordinates, by a one-to-one mapping, and then to extend to cylindrical coordinates the technique used in [4] for Cartesian coordinates. Indeed, changing from a Cartesian coordinate system (x 1 , x 2 , x 3 ) to a cylindrical one, (r, θ, x 3 ) ≡ (y 1 , y 2 , y 3 ), we return to a situation similar to the one of a flat boundary, where the cylindrical coordinates play now a formal role similar to that of the Cartesian coordinates in the flat boundary case.…”