I read on a "Random Space Facts" page that the Galileo space probe used 67 gallons of propellant to get to Jupiter, getting about 36,000,000 miles per gallon.
This is indeed a 'random' fact. It is so out of context that I can see all kinds of misuse brewing. There is a slight disclaimer, but even this misses the point: "...although Galileo's usage of fuel was not at all continuous, but rather occurred in discrete bursts."
Don't be fooled.
I will take this in two directions: first, the pedantic engineer's approach - Galileo was launched from the Shuttle Atlantis and then boosted with a trans-injection stage, and so all the fuel for those two craft would have to be included for a more meaningful number. But even this more thorough treatment misses my second and central point, that of the pedantic physicist, who says: "wait a second, we're talking about an orbit here!"
To see what I mean, let's consider how much fuel the Space Station has used so far during its travels. We should be able to measure some distance, and then divide by the amount of fuel used, right? But hold on, the distance keeps changing. What is the 'end point' for orbital travel? The Station racks up miles without using any fuel.
So, for objects in orbit (and Galileo was always in orbit around something - first the Earth, then the Sun, and finally Jupiter), talking about fuel efficiency is tricky, and division of the fuel used by the distance covered is not a useful measure.
Since each orbit is something that can (more or less) be maintained without using fuel, what counts in space is how much fuel, or energy, you need to change from one orbit to another. When you hear discussions of about 'circularizing orbits,' 'aerobraking,' 'retrofire,' 'delta V,' or even more technical terms like 'flyby' or 'Hohmann transfer' and the esoteric 'Lyapounov tubes,' these are all referring to changes in orbits, or methods for getting from one orbit to another. A great deal of spacecraft mission design (in fact, probably the largest factor in design) deals with how to get into the desired orbit with the least effort.
Changing orbits takes a lot of energy. It's not just "float here, OK, now float over there." That's what the whole Shuttle return-to-flight safety debate is centering on right now. If Shuttle were to go into an orbit to be able to service the Hubble, it is impossible to then change the orbit to get to the Space Station. The Shuttle simply does not have enough fuel to do it. This is what makes movies that show the Shuttle going to asteroids ridiculous.
Getting all 23,981 kg of Hubble to its 611 km altitude orbit was very much at the edge of Shuttle's capabilities in 1990. It is no coincidence that Hubble orbits at an inclination equal to the latitude of KSC - this is the 'least energy' orbit inclination for launches from KSC. In fact, for Shuttle to get to the higher inclination used by the Station (51.6 degrees), it has to give up about half of its payload-to-orbit performance. This was one of the primary drivers for the development of the Shuttle's new super-lightweight pressurized tanks.
And why is the Station at this high inclination? Because the first piece of it Zarya, was launched from Russia' Baikonur (Tyuratam) cosmodrome. It took all of the Proton rocket's power to move it from Baikonur's 46 degree latitude into the standard Russian orbital inclination of 51.6 degrees. Remember: once there, it's almost impossible to change the inclination.
Senator Buford P. Frink (D-LA): "Now just hold on here for a minute, Mr. NASA-administrator-man... are you telling this here co-mmittee that you can put a vee-hicle into orbit around, um, ah, Jupiter, using only 67 gallons of gas-o-leen, but all your thousands of pencil-necked nookular engineers can't come up with a more efficient car or truck?"
Mr. NASA-administrator man: "Well, Senator Frink, NASA would be happy to produce more vehicles that got 36 million miles to the gallon, but you have to remember that Galileo's fuel cost us $21 million dollars per gallon."