Depends on who is doing the testing
From your question, it seems that you're working with a fairly loose set of requirements and are simply asking for some friends to do some basic checking. However, this may not be the case so here are some general guidelines for a quick test plan based on the experience of the testers.
Experienced testers familiar with the product
Just point them at it, and tell them what's changed over a few emails. You should supply the user stories or requirements that you used so that they can infer the new behaviour.
Inexperienced friends who are just checking it out
You need to be a bit more specific. Perhaps a bunch of screenshots showing expected behaviour for a given use case on one browser and then get them to replicate that behaviour on a variety of other browsers a differing version levels.
The browsers should be resized in all manner of crazy configurations to ensure that the pages scale properly.
You must ensure that they enter all manner of rubbish data. Supply them with examples of nasty stuff (see http://xkcd.com/327/) to put into text fields.
Make sure that they attempt to use the site using the keyboard only and that it still works.
That should cover most of the biggies.
Consider automated testing
If you're not confident in your friends doing the testing, consider writing a few regression test scripts that perform some clicking around and parse the returned HTML. Your developers should be able to knock up basic HTTP-client type scripts in no time that just test simple behaviour.
You can get progressively more complex in your approach later as your application gains complexity.