Basically, if you have a page accessible by multiple URLs, from search engine’s prospective you’re bloating its database by serving duplicate content. So to avoid this, you have to set
<link rel="canonical" href="..."> with
href pointing to the original URL of this page. Search engines will check the url in href attr and won’t put a duplicate content penalty on your pages.
Useful thing to remember and use!
And I just found an interesting post on this topic by Anne van Kesteren with interesting discussion going in comments.