Scotty,
this flies in the face of all of J North's quotes (posted at 8:25pm) and I think the reason is what your paper megaphone experiment demonstrates is the effect of a horn, and a conical one at that. This might be valid if the sound source were within the throat of the whizzer, but you'll notice there is a phase plug there on the Zu FRD.
In the case of the whizzer, the source is the whizzer cone itself, which is smaller than the main cone and thus has intrinsically wider dispersion.
Russell