Upload & Sell: On
I've conducted a few more tests using a different lens (my old 28-105) on my 50D, 40D and 5D. I also compared results of my 24-105 on the 50D, 40D and 5D using a controlled lighting situation.
In a nutshell: the CA/fringing is entirely dependent on the situation, lighting, flash/no flash, strong vs. weak back lighting, etc. It really has little or nothing to do with the camera itself but, rather, the interaction of lens, high density sensor and lighting conditions.
Once I realized which combination produce which result, I was actually able to generate CA/fringing when ever I wanted to and ... conversely ... I was able to reduce or eliminate CA/fringing whenever I wanted to.
FYI: my 28-105 is tack sharp on the 50D ... amazingly so ... and is equally sharp on the 5D, but the 50D renders slightly more detail. At the widest angles, the 28-105 produces LESS CA/fringing (width-wise) than the 24-105 on all 3 cameras. The least CA/fringing was, perhaps predictably, was with the 5D.
When I have the time I'm going to do some more tests using my 1Ds II ... though I imagine the results will be similar to the results from the 5D.
Thank you all for your comments.