I did get a couple of replies to my question:
Kim Worley at Baylor College of Med. genome center says "We've
worked with combining data in assemblies. Different data and
different assemblers or versions of assemblers work better in
some combinations and worse in others." They also have
an unpublished tool to scaffold and gap fill with this data.
Scott Emrich has experience in this with insect/arthropod genomes
n collaboration with the Celera assembler group at U Maryland,
and offered to help us apply this with the Daphnia genome.
There is also this recent summary paper on uses/problems for
genome assembly with second/next generation data.
Assembly of large genomes using second-generation sequencing
Michael C Schatz, Arthur L Delcher and Steven L. Salzberg
Genome Res., May 2010 doi:10.1101/gr.101360.109