Peter Suber, responding to a J Cheminfo paper, wondered what proportion of chemistry journals in the DOAJ charge author-side fees. Since I was in that mode, as it were:

Hybrid journals are those that offer OA-for-a-fee, so of course all of those charge fees. "Open" above refers to Gold OA journals, roughly half of which charge author-side fees in this chemistry subset. This is broadly consistent with the overall DOAJ listing (as of December 2007) and also with several other studies that Peter mentions.

I still can't solve the tables bug; if you want the numbers, view source -- I've commented out a simple table that displays fine unless Moveable bloody Type gets hold of it. If you want to see how I generated the numbers, grab this spreadsheet. I first cut-and-pasted from the DOAJ subject listings into a text editor, then used the replace function to introduce tabs before "hybrid" or "open" and between "publication fee" and the entry for each journal. Then I used the replace function to delete all lines between "hybrid/open" and "publication fee", to simplify the Excel formula... you'll see what I mean if you look at the spreadsheet.