Al, focusing on the losses at these velocities and thinking that re-accelerating the column of air in the intake tract is a bad thing misses the point entirely.  That's like saying a supercharger will drop horsepower because it takes power from the engine to turn it.

        Here is a link to a pretty good explanation of how to select the correct intake runner length:

        http://www.hotrod.com/projectbuild/113_9907_efi/

        They say that the runner length should equal the third harmonic wavelength.

        Here is a snip from the article:

Chrysler testing resulted in a formula to calculate where the ram effect will come into play. To wit: N x L = 84,000, where N represents the desired engine rpm to tune for and L is the length in inches from the opening of the ram tube to the valve head. Shope explains: "Let's say you're running at Bonneville with an engine that develops peak horsepower at 8,400 rpm and want to tune for maximum ram effect at that level. Then, L should equal 10 inches, as in 8,400x10 inches=84,000." To achieve ram tuning at 5,500 rpm, simply divide the constant, 84,000 by 5,500 rpm. The result of 84,000/5,500=15.27, the ideal distance for the intake tract as measured from the opening of the ram tube to the valve head.

        I would guess that you would use the rotor RPM multiplied by 6 in this formula. As I understand it, the shaft turns three times the rotor speed, so you would insert two times the shaft speed of a rotary engine. Using the example above, a rotary would need a 7.85" ram tube to get max performance at 5,500 RPM. This assumes that I assumed correctly. :^)

        I would also guess that you would want the runners to make an abrupt change in diameter at the ends to promote the reflection of the pressure wave. This would imply a "log" manifold or a plenum of some sort to feed air from the  throttle body. Not a smooth transition like a wye or a very gentle flare.