The comments from David Evans and Nicholas Simons make a lot of sense,
but how do you decide whether to seal or not?  If you think you have
good leather, and don't seal it, what tests should be done on the
assembled stack?  Should you then test the air flow through each pouch
(past MMD articles by Craig Smith and Phil Dayson cover this), in which
case what level of flow should you aim for?  Or do you rig up the stack
and test for repetition, in which case what rate should it be capable
of, and how do you go about creating the repeat signal and measuring
the achievable rate?

If it then fails your tests you have to strip it down again (partly
at least) to seal the pouches.  It almost seems that sealing first,
and being prepared to enlarge bleeds (which are usually "get-at-able")
is the easier option.  Note also Craig Smith's article in the MMD
Gallery/Technical which compares some sealing methods for longevity:

  http://www.mmdigest.com/Gallery/Tech/leakage1.html 

Like many of these discussions, the threads seem to peter out before
covering all the angles, and testing and measuring methods seem to be
not covered in detail, so it would be useful, this time, if the experts
would add their comments.

John Farmer, UK
http://www.muscanic.com/