>Again, my understanding is that the main and pinion shaft load
>is axial and not longitudinal.   They don't care if they are spun
>clockwise or counterclockwise.
>

Don't want to match wits with Todd regarding transmissions, I know I'd lose.
But his statement caught my attention.
When torque is applied to the gears *attached* to the main and pinion shaft
a load pushing outward in a radial pattern would result.  The gears would
try to push the shafts apart instead of transferring power from one shaft to
another.  This would indicate the load is not completely axial.
I don't think my observation changes anything, the same radial loads would
be applied regardless of rotation.

FWIW, PSD.