mpwilliams wrote:

> EPA mileage estimates are determined through dynamometer testing in a
> laboratory, and among the factors reducing actual gas mileage that are not
> considered in these tests are rolling resistance of unpowered wheels/tires,
> aerodynamic drag, non-productive operation (engine running with no forward
> progress, as occurs at traffic controls or in stop-and-go traffic), ambient
> temperature (affecting intake air density), and altitude (also affecting
> intake air density). City and Highway mileage ratings determined in the
> laboratory are adjusted downward by 10 and 22 percent, respectively, in an
> attempt to compensate for these and other factors, but Consumer Reports have
> consistently determined, using their own mileage testing protocol, that the
> EPA mileage estimates posted by automakers for new vehicles are optimistic,
> sometimes by as much as 30-40 percent.

And yet, my '93 Accord consistently got 10-20% above EPA estimates on the
highway. Go figure.