Probably the lower pressure has a bit more drag and ends up slipping just enough to cause a slight rotation difference. But I guess I will just have to try lowering the air pressure in the xB and see if my odometer changes on my trip around Ocean drive as well as how much my mileage changes as a result - might be nice to have a soft ride too. Generate some real data . . . simple test would be to mark the tire with a line of gooie stuff and roll it for a couple of revolutions and measure the distance traveled very accurately then change the air pressure and try again. I calibrate the scooter and bike speedos by measuring the distance of 1 revolution in millimeters and program that into the bike computer speedo for the wheel size. Gives a pretty good result considering a millimeter resolution.
I would have agreed with you if you had said:
"more air pressure added to an underinflated tire DOES slightly change the diameter of a tire."
i can see your point there, but to continue splitting hairs we have to determine what "underinflated" means. my tires are max rated for 44 psi, yet the (auto) manufacturer's suggested pressure is 32. below which of these numbers is the tire "underinflated" as would affect diameter?
if there is an upper limit on the amount of "stretch" in a tire's design, i bet it's at the tire manufacturer's maximum, not the automaker's. in which case, since most auto manufacturers ratings are significantly lower than the tire's max rating, most cars may see a FE benefit from both slightly larger diameter and lower rolling resistance by going with higher pressures.
i suspect there's a not insignificant diameter difference between the recommended 32 and the 48 i'm running in my tires. note my earlier comment: i measured a 2 mm radius difference between 38 and 48 psi - though admittedly the only way to know for sure how much of that is diameter change as opposed to "de-bulging" of the sidewall would be to run a controlled distance test as jangeo suggests.
also, if there is a difference between 32 and 44/48, i suspect it's not so much from "slip" (which i would expect to see at much lower than 32) as it is from an actual (if small) diameter/circumference difference.
My approach has been to go up one tire size (from 165/70/13 to 175/70/13) and then to increase tire pressure to (from 44 to 55 psi) to deliberately try to cause the tires to rise up in the center to reduce the tread width actaully in contact with the road to the same width that I had with the smaller tire size. My results have been that, while I am getting the same MPG that I got with my smaller sized tires, I am getting MUCH improved cornering and braking performance. I suspect the lateral loading of the sidewalls while cornering or braking is causing the tread to flatten out and putting the entire width of the tread in contact with the road during these events. All I know is that the car now stops is a very short distance and that it is practically impossible to lock up my tires while braking on dry pavement. And I have added 10 to 15 mph to my cornering speeds, thus preserving that oh so important kinetic energy while negotiating curves (its also a lot more fun!).