Are you sure the minor axis always aligns with the cylinder axle and goes through the 3D center? I've heard that before and thought it was weird, but when I looked it up, it sounded like it was an approximation that's only completely true when the axis/axle intersects the "center of vision". Since people generally put their focal point/center of vision directly on what they're looking at, in real life it will probably always seem true. Same for if you're specifically calling this out in a demonstration, you'll probably put the circle/ellipse near the center of the camera's view. But if you have a cylinder closer to the edge and tilted, I believe they can get misaligned, even without lensing effects. I originally learned some of this from Marshall's 1994 perspective series. Near the end of one part, in what I think was a Q&A, he mentions the minor axis doesn't always go through the center either, but it's "so close". Another source was the "Vanishing Point: Perspective for Comics" book, which says making the minor axis parallel to the axle is usually a good enough approximation, and much easier. So it's not usually important, but maybe it is if you're practicing drawing a bunch of arbitrary cylinders at various angles. It confused me for a while.