It seems that in order to accurately measure the impedance of the element itself, you should be using a half-wave (multiple) section of cable cut for the mid point frequency, instead of some random length of cable.
Any length other than a half-wave multiple should be considered an impedance transformer the way I see it. By using a half-wave multiple, the unknown impedance of the element is brought back to the analyzer.
If the DB224 harness is made up of true odd-multiple quarter wave transformers, then the impedance of the element itself works out to 112.5 ohms on paper.
If I were to assume the elements are 50 ohms, then the impedance at the feed point connector is 112.5 ohms.
I have seen at least 2 different drawings of the harness, which both work in theory. The differences are mainly the types of cable used and number of junctions involved.
It might be worth mentioning that selection of cables with a velocity factor of .66 is important if you want the cables to reach the elements and not end up with slack to deal with on the mast.