Merge/Condense vocal parts arbitrarily, on a per-'master staff' basis

[Jeff Epstein]

I have been playing with LilyPond on and off for a while now. I'm a programmer
and musician, so I feel like I have a head start. But it's--impressive--heavy
duty stuff. I'm still really doing only the basics with LilyPond.

Anyway, I arrange acapella music. I want to get away from arranging by pencil
and paper, once and for all, and switch over completely to LilyPond. I am unable
to figure out one major need, which I hope someone out there might be able to
help me with.

---

I can compose the following in LilyPond, no problem. Five vocal parts for an
acapella piece, each on their own staff. Here's an example:

http://jeffyepstein.com/posts/merge_parts_question/before.jpg

What I need to do, is to merge parts together, arbitrarily, differently within
each "master staff" (each set of staves, on each new line on the
printout...sorry, don't know the right terminology).

Here's what I'm looking to do:

http://jeffyepstein.com/posts/merge_parts_question/before.jpg

1-4: So from measures 1 through 4, it should be the same (one staff per part).
5-9: But on measures 5 through 10, it should be sop1+sop2 on one staff,
alto1+alto2 on the second, and tenor by himself.
10-14: Finally, measures 10-14 should be sop2 on the top staff, and
sop1+alto1+alto2 on the second.

And so on. This will condense the printout by a LOT, and it feels necessary if
I'm going to dive further into LilyPond as an acapella arranger.

Perhaps I'm missing something easier? I would greatly appreciate your advice on
how to either do what I'm trying here, or perhaps I just need to think about
this problem differently?

Thank you!!!