I am grateful to Robert Bremer for a very comprehensive explanation of
the thinking on the nature of manuscripts and the nature of serials. If
forthcoming documentation indeed excludes manuscript coding, we won't
add it. (Hey, we did without it for lo these many years.)

Diaries, ledgers and books of minutes we would probably do as books,
but the pieces in hand were microfilms of government reports titled
"Blue Book," written out each year by colonial officials (one was
even a printed form filled in by hand) and submitted to superiors--not
"published" in the sense you describe.  I think we have all
traditionally cataloged such items as serials if they have numbering or
dating.  I doubt they would be searched as manuscripts (and I don't
think theses will either), but I didn't want to ignore either their
serial nature nor their manuscript form.  We will add a note about
their nature and purpose, and their limited distribution, and let that
suffice.

Though I was intrigued with Joel Hahn's suggestion to add a book 006
with code t, too.  And Joe Orth, I hope OCLC's explanation has also
answered your question about the relationship between rules and coding!

Best,  Frieda Rosenberg (UNC-CH)
       <friedat@EMAIL.UNC.EDU>

.  Likewise, cataloging rules would exclude certain combinations
> of type of record (Leader/06) and bibliographic level (Leader/07), such as
> manuscripts and serials.  We have added relationships to PRISM validation
> for these invalid combinations to assist users and avoid input of incorrect
> records.
> The CONSER Editing Guide update number 3 for spring 1996 does include a
> revision page for Type of record (Leader/06) which indicates that codes d,
> f, p, and t are not applicable to serials cataloging.