A very basic description...the information content of recorded music is not constant. For instance, periods of silence contain very little information, while loud, thunderous orchestral crescendos contain a lot of information. If you encode your MP3s at something like 128 or 192 kbps, the data rate is constant, and the file uses up more space than it needs to in encoding simple portions of the music. But with VBR encoding, the software looks ahead to see how complex the upcoming section of the song is and encodes it at the appropriate data rate. So you can cut down file sizes, or alternatively keep a similar size but have better quality, because the bits are used where they're needed.
Edit...Argghh, why type when you can just link...
