For cheap thrills, I entered http://www.perlmonks.org/?node=Best%20Nodes into The Wayback Machine. Unfortunately, only five dates came up, as shown below. I added today's "Best Nodes" for comparison:

• Mar 05 2007 Best nodes: month 83-135 votes, year: 121-194 votes
• Oct 10 2007 Best nodes: month 73-105 votes, year: 113-230 votes
• Dec 08 2007 Best nodes: month 71-132 votes, year: 115-238 votes
• Feb 14 2008 Best nodes: month 84-126 votes, year: 111-245 votes
• Apr 24 2008 Best nodes: month 77-111 votes, year: 111-156 votes
• Dec 08 2010 Best nodes: month 42-94 votes, year: 72-144 votes

Perhaps the most reliable indicator here is the lowest "best node" because the highest one often tends to be an "outlier" (that is, a freakishly popular node) while the lowest ones in the top 10 and top 20 lists tend to have similar values.

Based on these numbers (and my own random observations) the voting at Perl Monks seems to have declined considerably, especially in the past year or so. Maybe we lost quite a few users after getting hacked in July 2009?