Edit: Rewrote some of this to make it clearer and to emphasize practical applications, but the update got lost in the recent server crash.
Nerd Alert! To finish my little math exercise, I looked at the effect of hollowing on the tranverse moment of inertia (TMI) and specific TMI. Unfortunately, this part of the story is more complicated, and for that reason, I'll limit the discussion here to TMI about the CM rather than the tip.
Doing the math: We start once again with a solid cylindrical blank of radius R, axial length L, and uniform density b and hollow out its center to form a tube with inside radius Ri. In the first installment, we found that the tube's mass, AMI, and specific AMI could be expressed rather simply in terms of the corresponding blank properties and the "radius ratio" X = Ri / R, like so...
M = M0 (1 - X2)
I3 = I30 (1 - X4)
J3 = I3 / M = J30 (1 + X2)
where the subscript "0" refers to the blank, and the subscript "3" refers to the symmetry axis. It's worth noting that 1 - X4 is always greater than 1 - X2 in hollow tubes, where by definition 0 < X < 1.
To apply the same approach to the tube's TMI about the CM, we need another key proportion -- the "aspect ratio" A = L / R. Then
I1cm = I1cm0 [3 (1 - X4) + A2 (1 - X2)] / (3 + A2)
where the subscript "1cm" refers to any axis passing through the tube's CM perpendicular to the symmetry axis.
For pancake-like low-aspect tubes with A << sqrt(3) = 1.732, as in the big yellow rotor below (X = 0.82, A = 0.19), the way hollowing affects TMI is largely independent of aspect ratio. In these cases,
I1cm ~ I1cm0 (1 - X4),
which varies with X just as AMI does after hollowing.
But as A increases to "taller" values more typical of throwing tops (A ~ 2), the TMI about the CM shifts progressively toward
I1cm ~ I1cm0 (1 - X2),
which varies with X as mass does after hollowing.
For specific TMI about the CM,
J1cm = I1cm / M = J1cm0 [3 (1 + X2) + A2] / (3 + A2),
which increases with radius ratio X for any aspect ratio A. For low-aspect tubes with A << 1.732, this reduces to
J1cm ~ J1cm0 (1 + X2),
which varies with X just as specific AMI does after hollowing. Again, the way hollowing affects TMI is largely independent of aspect ratio. But taller aspect ratios progressively slow the rise of J1cm with X.
Some physical implications: For starters, hollowing has roughly the same effect on TMI about the CM as it does on AMI in pancake-like low-aspect tubes. Ditto for specific TMI and specific AMI. In such cases, the TMI/AMI ratio about the CM will change very little with hollowing. But at throwing top-like aspect ratios on the order of A ~ 2, hollowing tends to reduce TMI/AMI. This important ratio has a strong but complicated impact on top behavior, including responsiveness. Perhaps Jorge will have something to say about TMI/AMI in throwing tops.
In contrast, the way hollowing affects mass, AMI, and specific AMI depends only on radius ratio. Aspect ratio has nothing to do with it in blanks with cylinder-like symmetry, regardless of the lateral profile.
CM location: Finally, what does hollowing do to CM location? Well, boring out a cylindrical blank can't shift the CM from its original location. In non-cylindrical blanks, the CM will shift toward the blank's widest diameter. And if we add a tip and a stem and some spokes to turn the tube into a finger top, the top's overall CM will lie somewhere between the collective CM of all the additions and the CM of the tube itself.