[cloudjunkie]

I just want to point out that the aerodynamic drag on a moving body increases quadratically (i.e., to the second power like V^2) with velocity and not exponentially (i.e., e^V). This distinction is very important.

I also want to comment on the issue of the effect of headlights on MPG. At a speed of 55 MPH, the engine of a normal sedan needs to provide on the order of 10 HP to achieve a constant speed. Now, 10 HP is equivalent to about 7.6 kW (that's kilowatts) of power to overcome aerodynamic, road, and mechanical resistance. Now, my headlight bulbs say they use 100 W of power each, so having both headlights on requires about 200 W of power. So the proportion of energy the headlights use compared to the energy required to keep the car moving is

200/7600 = 2.6%

Thus, the engine will use roughly 3% more power to supply the energy to turn the alternator and power the headlights (energy must be conserved, the source of energy for the alternator is supplied by the belt that the engine is turning). If a car gets 30 MPG at 55 MPH, it wouldn't be unreasonable to expect a maximum decrease on the order of 1 MPG by having the headlights on. I spend roughly about 20% of my driving distance with me headlights on; thus, I would expect roughly a 0.2 MPG decrease in total gas mileage. That's pretty small, and I would have to question whether driving without headlights is worth the trade-off in safety.

Of course, your MPG after a head-on collision is technically infinite, unless your car is on fire in which case you get 0 MPG, but at that point you might not be worried about your MPG.

Lastly, the accord of basjoos gets about 60 MPG, so his observation of experiencing a 2 MPG decrease in gas mileage with headlights on is consistent with the rough calculation performed here.