In the last 2 months, our loader programs for records from Yankee and 
Lindsay have handled 3781 bibliographic records.  These records 
contained 474 490:0 fields, of which the loader was able 
automatically to resolve a tad over 81%.  Here's what happened to the 
remainder after I stared at them for a while (except in the one case 
noted, the 490 field became a traced series in every case--sometimes 
440, sometimes 490/8XX):

16 Added 643 field to existing authority record
12 Added 430 field to existing authority record
12 Series became traced, no change needed to authority record 
(usually caused by extreme abbreviation in 260 field of provisional 
vendor record)
6 New authority record needed
14 Deleted the 490 field ('series' in provisional vendor record 
duplicates information present elsewhere)
2 Added volume to existing records
3 Series matched 430 field, and so was flipped (loader programs are 
at present not allowed to do this automatically)
13 Only problem was with series numbering
2 Series became traced, but no new authority record (wasn't confident 
in vendor data)
6 Series matched multiple authority records
1 No action taken (series did NOT become traced; I could not begin to 
guess from the vendor record what to make of this one)

In almost every case, changes to LC/PCC name authority records were 
made only to the local version of the record, because they were made 
without consultation of the item.  (There was one "Oxford" series 
published by OUP whose authority record had no 643 field at all, and 
for that I made an exception.)  Similarly, new authority records were 
not contributed.

It's not visible from the above statistics, but most of the 490:0 
fields appear in the 'provisional' vendor records; the full records 
that replace them only occasionally have 490:0 fields.  Meaning that 
the these records probably don't reflect the change earlier this year 
to LC's series tracing policy.  So we'll consider this two months' 
worth of data to be a 'base', and we'll look at things again in a few months.

By the way: If we fold the 13 "only problem was with series 
numbering" into the "successfully handled" pile (because the series 
is now traced and there's nothing that can be done with the numbering 
at this stage, anyway), and also add in the 3 "successful match on 
430 field" (which I think I will soon allow the programs to handle on 
their own), the success rate becomes a somewhat more respectable 
84.8%.  Because of the additional 643 and 430 fields, if the same 
records were processed again the success rate would be 88%.  (But 
sadly, given the nature of series, it's more likely that the same 
ones will never come up again, so perhaps we shouldn't be so optimistic.)

Other loader programs handle Harrasowitz, Touzot, Aux amateurs and 
Casalini. Together they processed 527 records, containing 29 490:0 
fields, 58.7% of which were resolved by program. For the 490:0 fields 
unresolved by program (in all cases the series became traced):
7 New authority record needed
1 Split 490 into 2 830 fields: one has a matching authority record, 
one does not
2 Series traced, but no authority record created
2 Added 430 to authority record

Although the numbers here are much smaller and it may be rash to 
generalize, there is the suggestion that these unresolved 490 fields 
reflect a larger proportion of series for which there is no authority 
record (over half!) than do records from the English-language vendors.

Gary L. Strawn, Authorities Librarian, etc.
Northwestern University, 1970 Campus Drive, Evanston IL 60208-2300
e-mail: [log in to unmask]   voice: 847/491-2788   fax: 847/491-8306
Forsan et haec olim meminisse iuvabit.