Thank you for your interesting and thought-provoking anwers.
FT - I remember from 'theory of flight' (many years ago!) that mainplane dihederal aided in stability, and for low wing aircraft this was why they had dihederal. However, we were taught that high wing aircraft have excessive stability due to the pendulous nature of the underslung fuselage, and therefore most have either no dihederal (Cessna 152/172 etc) or varying degress of anhederal (BAE146, C5, Starlifter, Harrier, etc) in order to reduce the relative stability (and make it easier to manouver/roll I guess). Is this not the case? Does the same not hold true for any lifting surface?
John Farley - I guess from above comment, the axis would be in roll (of the rear fuselage?!).
The theme that seems most consistent from all answers is to avoid the wash of the mainplane, and you state particularly at full flap. Interestingly, the DC-10/MD-11 seems to have the highest nose attitude (although it's AOA is probably no different from Boeings etc, unless MD knew something that the rest didn't) of any airliner on approach, so perhaps it generates more turbulant flow behind the wing/flaps and hence the significant dihederal of the tailplane. So what about the L-1011?