Petr and Carol:

Good point about hybrids having the ability to recover energy through braking. This is because there is a chargeable battery on board. A simple fuel cell vehicle does not have this capability unless it, too is 'hybridized' to allow charging of a battery or a capacitor. I have not seen this as a part of most proposed fuel cell vehicles, e.g. the 'skateboard' design promoted by GM.

Heat engines such as external combustion (steam) or internal combustion are theoretically only able ever to get to 50% efficiency in recovering the energy locked into their fuel's chemical bonds. However, the best IC engines to date hardly get to even HALF this figure. Gas engines have been getting more and more efficient by among other things:
1) Reducing weight of rotating masses within engine
2) Increasing compression ratio
3) Making fuel burn more complete
4) Hybrid vehicles turn engine on only when needed and attempt to use engine at its most efficient RPM setting, although the cars allow the engine to run directly with their electric engine for maximum on-demand acceleration.

The next 'big' step is to reduce the size of the engine, and its complexity, by leaving out the cam shaft and camgear linkages, using electrically driven solenoids controlled by the onboard computer. The technology already exists, but putting together an engine that won't be able to run without its computer has yet to be tested on the market (despite new airplanes that can't be controlled without their electrical drive units, there being no physical links between control yoke and control surfaces in several of them, including the Concorde since the 70's). In mechanical engineering terms, this stuff is 'way cool'.

As far as fuel cells go, I believe there is no theoretical 'ceiling' on their efficiency (beyond the obvious one of 100%). What needs to be considered is overall capital cost, the longevity of the components, and their efficiency in the various stages that takes us from fuel to motion.
Technically speaking, for an overall comparison of efficiency between IC and FC, you have to compare apples to apples, start with a fuel, look at miles per gallon, then take into account construction, maintenance, and recycling costs. When you start with a FC fuel, you have to extract the hydrogen from it (Reforming), Convert the hydrogen to DC electricity, 'invert' the electricity to AC, and run it through a motor. There are several different fuel cell technologies, and they are all 'moving targets' due to evolving R & D.

There is a separate issue as to storing raw hydrogen, which is a resourceful little molecular critter and tends to leak right through metal.

Right now the main source of hydrogen in quantity is from natural gas. In future it can come from electrolysis from electricity from any source. I don't know how many Kilowatt Hours it takes to produce a gallon of hydrogen, but I think it is a pretty solid figure, i.e., we're not going to find a way to suddenly improve on it, in other words, we're already at peak efficiency when it comes to electrolysis.

There is a publicly traded fuel cell company that is constantly advertizing how they are making progress in fuel cell vehicles. I don't remember their name, but I remember talking to a fuel cell 'guru' in Anchorage who was the brains behind our post office being powered by a trendsetting fuel cell installation. He said their system allowed a test vehicle to operate for a day and then they had to be pulled out and replaced. A fuel cell is a lot like a battery in that chemical energy is converted directly to electricity and fluids are interacting with solids. When the temperature is not at the best, when you're starting up, when you're at top acceleration, some of the molecular activity gets 'messy' and you get deposits built up on the solid parts, and all at once efficiency goes down and you're in a world of hurt as a vicious cycle ensues, the less efficienty you get the more chemical deposits you get and so on. Sometimes the effects are reversible, and sometimes not. And fuel cell parts like battery parts contain heavy metals and stuff that is tough to recycle.

Are we ready to hug an internal combustion engine yet? Pause and reflect how much fun we've had with them, how much horse manure we haven't had to shovel in the last hundred years.

You can already simply and efficiently and cleanly burn natural gas in an internal combustion engine without going the fuel cell route. You're back in the heat engine efficiency limit, but fuel cells haven't proven themselves as any better in the practical world.

What I'm getting at by the above windage is that fuel cells are real neat, but not yet ready for prime time, and in themselves they do not address an energy supply problem, they kind of 'stir the pot.'

Analogy Coming:
You know how there are all these diet plans out there, but the simple truth of losing weight is that if you expend more calories than you digest, you will lose, and if you take in more than you burn off, you will gain weight.

The ideal ways to utilize energy are to:
Build light, don't move more mass than you have to.
Don't build more horsepower in than you really need.
Don't drive when you can ride or walk.
When you leave a room, turn out the lights.

But you want a vehicle to last. You may save on a miles per gallon basis with a light vehicle, but the average person I know says: "What good does it do for me if I save money on the commute, but one night I hit a moose and it comes through my tin foil car and puts me in the hospital?

I've personally come to the conclusion that a lot of American cars look solid and durable, but in fact are not. The Hummer comes to mind. I just drove a Toyota Camry across country at all kinds of speeds, it was quiet, the music was great, and the MPGs were about the same as my motorcycle was in the 80's.

I reiterate that market forces will do the job of getting us from now to the future not perfectly, but a lot better than government, unless government simply concerns itself with encouraging R & D (which it should).

I think a gas tax would be a good start, and it should be indexed to our expenses in the mideast.