They made the pri inductances the same, instead of the turns.
This because you do not get the same henries per turn when you get further away from the core.
The linkage is not as good out there away from the center of the E core.
The reason the DCR is less for the most turns is because the split primary has one section with a way larger mean turn radius.
On a UTC twin coil type, they can make the turns and the inductances the same, this probably improves common mode noise rejection, I do not know.
This is the only transformer in my dissection files with the turns mis matched
on purpose.
And it was the second hardest to dissect, due to the nasty black epoxy that they used.
It was hard as a rock, and if you pulled to hard, the wire breaks.
The worst was the Langevin from Brazil, for the tube pre.
Nasty pine pitch type wax, and millions of turns of ultra fine wire.