Unequal, changing time signatures in parts

[Erik]

Hello,

I have a manuscript from my organ teacher I'm attempting to transcribe. So 
far, so good, but the beginning of one composition has me stumped, despite 
looking for hints in the archive, snipits, and documentation.

The piece begins with a somewhat unmeasured section; the bass is a slightly 
modified version of Adoro Devote, and the accompaniment is a tone cluster 
composed of tied whole notes. The cluster starts with a high d, and then after 
each phrase of the plainchant the next lower semitone is added (cis, c, b 
bes...fis) so that at the end of the chant, the full cluster is sounding. It 
is very beautiful.

The problem I'm having is reconciling the whole note notation in the cluster 
with the varying meter of the chant within Lilypond; the bars in the chant are 
9/8, 8/8, 10/8, 8/8, 10/8, 6/8, 8/8. 8/8. So, I can't get the tone cluster 
additions to line up properly with the chant phrases.

In the manuscript, my teacher used vertical arrows in the score to emphasize 
the aligment of the tones where a new cluster tone enters. I'm happy to place 
barlines instead, as drawn arrows across three staves sounds pretty exotic.

Once this section concludes, the rest of the score changes time frequently, 
but all parts are in the same time so I'm home free.

I've looked at staff measure length and the arguments of \time, but can't seem 
to find a way to affect one staff without messing up the other.

Any solutions or hints welcome.

Thanks,

Erik