19MP is a significant, visually obvious improvement over 13MP when individual pixel quality is similar in both images.
I imagine so. When I'm out in the field next with an interesting static subject to shoot, I'll try and remember to take a single 5D shot with the 24TS, camera horizontal, and 3x20D shots from the same position, camera vertical, using shift to its maximum, and compare images.
The aspect ratios will not be exactly the same. The 20D stitched images should have marginally less height. However, depending on the required cropping, the resolution differences should be fairly similar to those when taking a single shot with each camera with identical lens and f stop, then cropping the 5D image to the same FoV as the 20D image.
However, I accept that such differences in resolution are probably not significant on smallish prints. But my printer is capable of producing a 24"x36" print from a single 35mm frame. The less interpolation the better.