Depending on the title of an article, I can sometimes guess the opinion of the author and so I'm more interested in the HN commentary. If the article is something I know very little about, I will read that first.
You may be interested in some previous submissions of this question and the evolution of the replies over time:<p>1533 days ago : <a href="http://news.ycombinator.com/item?id=145261" rel="nofollow">http://news.ycombinator.com/item?id=145261</a> : 5 comments<p>1240 days ago : <a href="http://news.ycombinator.com/item?id=429802" rel="nofollow">http://news.ycombinator.com/item?id=429802</a> : <i>Many</i> comments<p>974 days ago : <a href="http://news.ycombinator.com/item?id=861970" rel="nofollow">http://news.ycombinator.com/item?id=861970</a> : A poll with 17 votes<p>8 days ago : <a href="http://news.ycombinator.com/item?id=4031889" rel="nofollow">http://news.ycombinator.com/item?id=4031889</a> : 5 comments<p>I'm sure there are more, my search used specific terms to reduce the number of items returned.
It varies depending on whether or not I know the source (Hacker News displays the domain from which the link comes) and the subject of the article. Sometimes I read the article first, if I know the source is reliable or if I know I am interested in the subject, and sometimes I read the comments first to figure out if other HN participants, more knowledgeable than I am about the subject, think the article's treatment of the subject is worthwhile or not.
Comments, to see if the article is worth reading.<p>It's not uncommon to find an article that is little more than some glib anecdotal observation and all the interesting content is in the comments.