thats not the best example. you are looking at different final drives, different tires, and different motors as well.

However to remedy your point: I put a d16z6 (originally belonging in a 2400lb 94 si) and put it into a 2100lb 94 cx. My mpg was notably better than that of the si and believe me i romped on that car all the time. hit vtec every time i drove the car and went 80 every time i was on a highway.
gas logs are still available in my garage.

the bottom line to this topic is that every case is different and there are too many counter examples to make a hard fast rule in the applications, but we do all know how things work in theory.

Wait, when did we go from GM 3.8 cars to Civics?

Those are polar opposites regarding weight. You can probably throw an extra 50% weight on a GM 3.8 car and barely see a difference in FE. OTOH, you can trim 100 pounds from a Civic and measure a difference...

^ yes. I've seen the same with my civic compared to my Odyssey 3.5 v6. The Ody doesn't really change with loading, but the civic's mileage changes with minor weight changes. Full tank is about 80 lb heavier than empty, and averages 2-3 mpg lower. Family on board brings a much larger penalty.

i agree.