In his article on noise and resolution, Mark Segal uses a Canon 24-105 zoom lens for a resolution test between the Canon 1Ds and 1Ds3, and finds very little apparent difference between the two cameras.
I wonder if that is a result of the limitations of the lens?
It seems to me that if you want to examine the resolution of different cameras, it would make more sense to use a prime lens that is well-known for its sharpness, such as the Canon 135L, Canon 200 2.8L II, or the new Zeiss ZF primes.