GCNP Boundary: Another NPS comparison: Segment I, east off the Shivwits

I have been taking a closer look at the somewhat ambiguous place on Segment I (the Shivwits rim), where it leaves the Shivwits and crosses the Esplanade to reach the rim of Parashant-Andrus Canyon. Here is the GCNP website map, with the area circled in red:

I have been comparing it to the official 1975 map, which is admittedly a bit slap-dash:

But look right where that lower, horizontal, orange line points--  match to my first green cross on the NPS map. On the official map, the line continues NNE to the rim, then bends back NW to point to the word "Boundary". My line of green crosses tries to follow the NNE part of that line. Contrariwise, the GCNP map does what the BLM map does, drops off due east, crosses over Mollies Nipple (not shown on the NPS map; out of modesty, I guess) to reach Parashant rim near the river. Here is BLM:

But the official map doesnt show any sort of due east indentation. The official line rather seems to go around that last point then, with a little bulge, head for the side canyon one higher up, as my green crosses try to show. This is supported by a close look at the township lines on the BLM and official map. On the official map, that bulgy line is way on the west side of T31N R10W, not cutting it in half as does the BLM map line. My green-cross line is supported by the grazing allotment maps (see my 9/13/10 entry) as well, which show a NE-trending line in T31N R10W for the 8/9-228 "afton snyder" boundary.

There is also a misleading aspect to the BLM & NPS maps, which shows up best on BLM: Their north-south boundary on the east side of the last point before the line goes due east is not located on the rim. In that area, there is a treed slope before the much-more precipitous Kaibab drop-off is reached. Moving the Park boundary about ½ mile east would bring it to the sharper rim and point it that much more toward the second side canyon up.