Hi, Wanderer,
Note that, with the lens focused at infinity, the second (rear) nodal point will be precisely the focal length in front of the focal plane. Thus, for EF lenses, the distance from the 2nd nodal point to the flange reference surface will be:
f-44.00 mm
where f is the actual focal length of the lens. (For focus at infinity, that will generally be the "marked" or "advertised" focal length.)
(44.00 mm is the "back flange", or "register", distance for the EOS camera series, the design distance from the mount flange reference surface to the focal plane.)
Yes, for a 35-mm lens (for example), that will be in thin air behind the physical lens. This is what "retrocus design" makes possible. (Otherwise, we couldn't have short focal length lenses on an SLR.)
One can get a pretty good approximation of the location of the exit pupil visually. If you look into the aft end of the lens, at what you think is the "iris", what you are really seeing is the the virtual image of the iris as seen from behind the lens, which is in fact defined as the exit pupil. If you can visually estimate how "deep" that is into the lens, you have your answer.
Note that to be certain you are looking at the (virtual image of) the iris and not some other circular outline, it is helpful to have the lens stopped down.
You can do that by having the lens on the body, setting a farily small aperture, pressing the DoF preview button to make the lens stop down to that aperture, then (while still holding the DoF preview button), dismounting the lens. The iris will stay where at the chosen aperture.
Yes, this is a little tricky to do if the 2nd nodal point is to the rear of the entire physical lens.