Thanks all for the discussion.  (this is an '89 2.1 WBX by the way)

I don't understand all of the responses, for example I am not sure what
torque has to do with my question.  Sure there are some torque differences
in 3rd vs 4th gear, I can't understand how that might affect fuel
consumption.

But what David wrote makes good sense.   In 3rd gear there would be more
losses in the engine itself (piston / conn rod weight reversal, valve
springs, compression events, oil/water pump).  So that suggests that the
MPG would be lower while driving in 3rd gear.

But then as a counter argument (if I read it correctly), the engine might
run more efficiently at higher RPMs, at 3400 in 3rd vs 2400 in 4th.

I think we need to find some graphs of fuel consumption vs engine speed.
Ideally they would be for the WBX, but perhaps all engines have similar
characteristics.  About 5 different plots, from 1000 RPM to 5000 RPM with
each plot being a different load on the engine.  I think this is what I am
searching for / asking about.  It is almost like reverse engineering the
ECU.

Roland

On Sun, Mar 25, 2012 at 3:25 AM, David Beierl <dbeierl@attglobal.net> wrote:

> At 09:56 PM 3/24/2012, Roland wrote:
>
>> consumed in 3rd gear vs 4th gear?  Now the immediate answer is none!  It
>> takes the same energy to drive at 45 mph in 3rd or 4th.  The wind
>> resistance is the same, the drive-line loss is the same, other friction
>> like tires are the same.
>>
>
> But it takes more energy to run the engine faster.  The inertial losses
> from reversing the pistons' direction of travel, the frictional losses from
> scrubbing the piston rings up and down in their bores, the energy lost by
> compressing the fuel/air charge and operating the valve train, the
> additional output of the water and oil pumps - these parasitic losses all
> increase more-or-less linearly with increasing rpm.
>
> It would be interesting for someone with an OBDII engine to do some
> calculations of no-load operation at various rpm, using the rpm and engine
> load numbers to determine the relative amount of fuel required to merely
> spin the engine at those rpm.  I wish I'd done it while I still had my
> Honda.
>
> Yours,
> David
>