In 1990, Bob Pease of National Semiconductor published a “design idea” in the June 14, 1990 issue of Electronic Design for the use of a FET for reverse-polarity protection, with the advantage that it will typically have much lower forward voltage drop than a diode.  Pease mentions this on page 164 of his book Troubleshooting Analog Circuits., and the circuit is shown in figure 13.5. Electronic Design published another very similar design idea by Jim Walker in the March 31, 2005 issue.
My friend Rich and I were somewhat puzzled by the source and drain of the FET being reversed from the expected orientation, so I wanted to look up the original article to see if there was any detailed explanation given. The Electronic Design web site has an online archive, but it only goes back to October 1990. The local public libraries don’t seem to have it. However, I’m not that far from Stanford University, and their engineering library does have it from 1984 to present.
I drove over to the campus, parked in the garage, and fed $2.25 in quarters into the parking meter for an hour and a half. As soon as I’d done that, someone pointed out to me that parking is free on weekends.
When I got to the library, it was easy to find the bound volumes. It’s somewhat irritating that they’ve been bound, as they didn’t have enough inside margin before binding, and have nearly none after, which makes it difficult to get a good photocopy. However, I had a bigger problem. The photocopy machines in the library don’t accept coins directly; you have to purchase a card. The only card vending machine in the engineering library was out of order, and there is no backup method in place.
I’ll have to try again some other Saturday.
I wonder whether they’d get upset if I took in my notebook along with my Plustek OpticBook 3600 book scanner. This is a scanner that can image very nearly up to the corner of the scanner, so you don’t have to open the book more than 90 degrees to get a good (flat) scan of a page.
I use that circuit all of the time, I found it in an Maxim IC app note.
The drain and source are oriented (its a P channel fet) such that the body diode provides the initial path for current. The gate is hooked to ground. Once the current is flowing through the body diode, then the gate-source is forward biased and the source-drain turns on giving the low voltage drop.
If power is applied reversed, no current can go through the body diode, so that no vgs can exist.
http://pdfserv.maxim-ic.com/en/an/AN2012.pdf
My comment dissapeared, but here is a link to my source for that circuit.
http://pdfserv.maxim-ic.com/en/an/AN2012.pdf
Sorry, Les, Akismet thought that your comments were spam. Not sure why. Thanks for the info.
I just thought of something. One product I designed using this circuit, uses a 9 volt battery. If the battery is inserted backwards, the unit is protected. The problem occured when the people building the product would insert a battery backwards, then place the unit in a programming fixture injected power into the device after the reverse voltage protection circuit. The power from the programmer turned on the mosfet, which then allowed reverse voltage from the battery to enter the circuit.
It didn’t matter how many times I told them there was no need to insert a battery to program the unit, but now I digress.