Though I'm not a cataloger by trade, let me take a stab at John
Radencich's query about superfluous dates.  In general, I'm in favor
of transcription, on the theory that the extra date information might
mean something to someone.  (For instance, having an issue's exact
date is always handy if you're trying to claim it.)
     In the case of the Journal of Geophysical Research, this is a title
published in several different parts, and which is also prone to
splitting and title changes and other nefarious actions.  Each part
comes out on a different day of the month.  So while there are no other
August issues of JGR: Planets, there are other August issues of JGR, and
having the day could distinguish among them (say, in an incomplete
citation).
     Other titles that add these extra dates may not have such good
explanations, but you never know.  Would a Druidic monthly publish on
the 22nd day of the month, so as to hit the solstices?

(My wife probably will be surprised to hear I'm giving out advice on
dating now...)

Bob Persing
Serials Supervisor
O'Neill Library, Boston College
(617) 552-3213   PERSING@BCVMS