Basically, when a camera takes a photo, it cannot capture the full range of light in the shot.
A picture is obviously made up from the very light sunny bits, down to the dark shadows, and all that is inbetween.
The average camera can only grab a limited range (very expensive ones have wider ranges) so you take a number of shots, each with a different exposure setting.
You then end up with shots capturing the dark areas, and the lightest ones all in greater detail than if you took just one shot to cover them all.
Then you can merge them all, and take the "best bits" from each, to create one single image that shows the whole range better.

I hope that makes sense!

This is a "normal" shot taken at the same time as the different exposure ones I used to create mine.

http://img520.imageshack.us/img520/2372/origck5.jpg