August 10/06:
The only other hot-weather bi-directional runs I did in warm weather before changing transmisisons was the alternator/no alternator test.
Unfortunately, it's a little complicated by the fact that I tested at 70 km/h, which was the one speed on the speed vs. mpg graph where the data seemed to get a little out of line (literally):
(should be an image above this line)
I'm inclined to think that 71.0 jump at 70 km/h was testing error, since in the alternator vs. no alternator test, the control runs (5 bi-directional) at 70 km/h in 71F temps (vs 58F for the speed runs) averaged 70.8 (i.e. same results, at higher temps, which you wouldn't expect to see).
(should be an image above this line -)
All of this is a long-winded way of saying that the FE improvement from the transmisison is likely closer to 5% (5.6%, based on the single 70 km/h point comparison) than 8.8% I claimed before.