Returning to one aspect of the original post, there is one way that the down-curved barrel MIGHT have improved accuracy in the absence of rifling. I recall reading an extract of an 18th century ballistics treatise, possibly by Robins, in which the author discussed the spin induced on a bare ball from contact with the side of the bore, and how it appeared in his trials that the orientation of the spin from the last contact with the bore was random in direction, and thus also the curve induced by that spin, enlarging the dispersion of the shots. He then related that some of the better gunmakers would counter this by giving the muzzle of a "gun" intended primarily for ball a downward bend just sufficient to ensure that the last contact of the ball with the bore would be at the top, giving the ball a consistently-oriented backspin. This produced two salutary effects: the spin, and thus the induced curve, was now consistently oriented, and that curve was in the upward direction, producing a flatter trajectory than would otherwise have been the case. It seems possible at face value that the curved bore of these pistols might produce a similar effect.
Regards,
Joel