[aerohead]

freebeard asked about this, and I don't remember what thread it was in, so:
From Hucho's 2nd-Edition:
* Palowski, Figure 1.38, page-31,
For his four velocities tested, all the drag minimum curves intersect at a leading edge radius of approx. 1-2/3rds inch ( 42.3mm ), and remain saturated. No actual vehicle parameters are provided other than 'square'-bodied.
* On page-32, Volkswagen's research for what became the Vanagon, indicated radius optimization with the radius = 4% of the body width.
* On page-181, Figure 4.99, for RV/ caravan trailers, the drag bottoms out @ radius = 40% of the square-root of frontal projected area ( ostensibly, the tow vehicle wake is shielding 60% of the trailer's face, in this specific test scenario ).
* For commercial vehicles, on page-326, Figure 8.5, the optimum radius is @ 5% of body width.
* NASA's best-case-scenario indicated for the trailer to be exactly the same frontal area as the rear of the tow vehicle, with zero gap.