I might be wrong in all this but I am assuming it's the "wandering" shank that has caused the effect shown and not the bit as such?
Lozz.
Most shanks I have come across are Hexagonal, the angles don't match, don't matter how much it "wanders" it won't make a pantagonal hole as regular as that.
Yes the shanks when new are hex, and yes the shank can, it can make semi circular grooves in the hole if it is wandering.
For semi circular grooves the maximum number of grooves attainable would be dictated by the bit circumferance divide by the shank diameter, I have in my posession an old hex shanked standard single chisel bit rock drill circa 1972 fro South Crofty, the tungsten insert measures 35 mm cross, it's circumferance is 110 mm, the hex shank measures 22mm across the flats, 110 over 22 equals 5, this means that in theory that the hole can accomodate 5 semicircular grooves made by shank abrasion centered around the holes initial circumferance made by the bit.
I think the question now is to establish wether the grooves were formed by the above method or by the the bit itself kicking around in the hole, Geoffs explanation in softer granite would hold true, in tombstone granite it might be more difficult, however not impossible if the bit was biased eg: a chunk of the tungsten tip had broken away from one side of center.
I'm running out of ideas now, good thread though by Roy.
Lozz.