Interesting occurrence... i did a few more hypothetical calculations and it seems that a 165mm x 10mm wall thickness carbon TUBE, responds with the same in deflection to an identical load, as a 200mm x 10mm wall thickness RHS in e-glass....

Using autoCAD, if you draw a 9mm thickness circle around the 200mm e-glass RHS - to represent the buildup of minimum 9mm minimum thickness circular bearing, the total diameter of the bearing surface is 300mm. So if you wanted to use a 300mm ID PVC pipe as a mold for the deck bearings, you would end up with exactly 9mm wall thickness at the corners of the RHS for the bearing buildup in solid glass - which should be ok as it would also be bonded to the RHS spar in these locations anyway..

As to laminate stress, we get less than half the stress in the glass 200mm RHS section compared to the carbon 165mm round section - which gave us the same deflection. This proves deflection is driving the design, not stress.

Carbon to carbon - if we compare the same 165mm x 10mm carbon tube, to a 165x165x10mm Carbon RHS, we get 59% less stress and deflection -from the improved section properties. The RHS section would be heavier as it has more section area, and the front and rear fairings would mean a thicker mast overall, but certainly not 59% more...

So i think square is the way to go... build up the round bearings, and choose your material based on weight and cost...