ggplot2's mpg dataset--what's the "fl." variable?

Question

In ggplot2's built-in mpg dataset there is variable called "fl.", which is a factor with levels: "c", "d", "e", "p", & "r".

Does anyone know what those letters are supposed to stand for? Needless to say, googling those letters has yet to give me any relevant leads...

library(ggplot2)
data(mpg)
str(mpg)
?mpg

[Note: There was a similar question on SO re: the mtcars dataset, which gave me the impression that this would be an appropriate forum for this sort of question.]

Answer 1

The fuel:

• e: ethanol E85, note(subset(mpg, fl=="e") pulls up only "new" american cars, and that fuel economy is much lower than the corresponding presumably gasoline models, which lines up with the lower energy content of ethanol)
• d: diesel
• r: regular
• p: premium
• c: CNG (note as far as I know the civic is basically the only passenger car that runs on CNG in the US).

Note, I have no reason to know this other than an educated guess based on the rest of the data, but here is some graphical evidence:

ggplot(mpg, aes(x=fl, y=hwy)) + geom_boxplot() + facet_wrap(~cyl, nrow=1)

Notice how e is consistently low d is consistently high at least where there is more than 1 data point (diesel has higher energy content) and p is consistently higher than r (premium allows cars to run at higher compression ratios and efficiency, though actually premium has lower energy content than regular) for each cylinder category (facets are # of cylinders).

---

UPDATE: as per @naught101, this now appears to be documented.