The patent is expired. Nobody needs permission now.
I think, to address the original question, the answer is, broadly speaking, yes. If you don't keep the axes coincident, edge distortion will move toward the center of the virtual image.
I should explain virtual image. The way a telescope of any kind works, you focus the eyepiece on the plane that the main objective is focussed on. If you, for example, put a sheet of paper in the focal plane of a main objective lens, standing to the side (so you aren't in the way of the image) you will see a sharp image projected onto the paper by the main objective lens. If you remove the sheet of paper, there is still a virtual image on the same focal plane, but you just can't see it when air is all there is on that focal plane because air is transparent rather than reflective, as the paper is. Now imagine the paper is back in place and the eyepiece focussed on it from behind. Then, when the paper is removed, the eyepiece will be focussed on the virtual image and will see a clear, sharp image put there by the main objective lens.
So, now, looking through the eyepiece, suppose you started to tilt the main objective. The focal plane would move along an arc equal to the focal length in radius, and you would be looking at the image at an angle through the main objective. This will distort it some and tend to make brightness uneven if this is a simple convex lens. If it is a flat field lens, like a photographic enlarger lens, designed to project a flat, undistorted image of a flat object or a very distant field of view onto a flat surface, then tilting the main objective will also cause you to have to refocus the eyepiece. None of that is desirable in a telescoping sight. So, I imagine, coaxiality is maintained.