I have two examples of critical apparatus documents ready to work on now, so I started work on the XSLT and CSS to render them.

Working from Ray's 2004 Henry VIII edition as a model, I set up the XQuery to harvest the correct apparatus information from the database, and set up the XML output so the apparatus is internal to the document (and is marked as such in the <variantEncoding> tag). Right now, I'm suppressing the orthographic variants just by hiding them in the CSS, so they can be displayed if necessary. I have a couple of remaining issues (<del> tags in the lemma are giving the whole lemma, or sometimes the whole line, a grey background for some reason), but the basics are there.