I’ve been very busy with other projects recently, so there’s been a long silence on this blog. Having a few free hours this weekend, however, I’m putting up a note on some of the work I did this summer.
This note might interest people who are curious about which CIA records are available online, or who are looking for Garrison files in the ARC. It’s more “inside baseball” coverage, so caveat lector.
NIA vs RIF records
Long ago I put up a couple of notes about a large set of records on the Mary Ferrell Foundation website which I called “old RIF” records. The notes are available here and here, and if you want to follow this latest note, you’ll probably have to take a look at them.
As I explain in my page on JFKARC content, documents in the Collection can be divided into “indexed docs” and “unindexed docs.” Unindexed docs are not listed in the JFK database. However, the metadata for everything else is there.
(See the page cited for a description of what kind of documents are not indexed.)
Indexed documents are supposed to have a printed “Record Information Form” (RIF sheet) attached on top of them, listing all the metadata for that document. This is the data that went into the JFK database. One of the metadata items for each indexed document is a unique record number (RIF number) which can be used to look up and reference ARC documents, or order them from NARA if you have a burning research question.
However, there is also a very large set of CIA documents available on the MFF which does not have RIF sheets attached. Instead, these docs have a different sheet attached, called a NARA Identification Aid (NIA). The information here is largely the same type of stuff that one finds on the RIF sheets, but in a somewhat different format. See the two notes I cited above for examples of this.
CIA provided NARA with NIA forms for their records until they finally decided to bite the bullet and redo everything in the standard RIF sheet format in February 1995. As a result, CIA records released from 1993 to 1994 all had NIA sheets, not RIF sheets.
It is possible, however, to go from the NIA sheets to the RIF sheets, and hence to determine which record you are looking at. One can do this by checking what I call the “accession number” which is printed on both the NIA sheets and the RIF sheets. I describe how to do this in the two notes cited above.
When I did this, I noticed that many of the records with NIA sheets were not otherwise available on the MFF website. MFF estimated at one point that it had about only about a third of documents in the JFK collection online, but this was clearly not counting the NIA records, of which there were about 10,000.
Adding these into the count, I found that many more CIA records were available online than MFF indicated. For example, the JFK database tells us that there are 177 records with RIF numbers beginning with the prefix 104-10001, but the MFF RIF search gives links to only 32, all of which have RIF sheets attached to them.
If you check the NIA records on MFF, however, you will find that there are 140 more records with NIA “accession numbers” and descriptions, that match up with the 104-10001 RIF numbers and their associated metadata in the JFK database.
Actually, there are even more NIA records than that, but some of these are duplicates of the records with RIF sheets. The duplicates are mostly records which were re-released later, with fewer redactions, and sometimes have other corrections as well (duplicate pages, missing pages, weird errors that put irrelevant docs in, etc). These changes are sometimes, but not always, noted on the later versions with RIF sheets.
There are also dozens of NIA records that do NOT match up with RIF records, and these are the subject of this note.
A corrected count of CIA records not online
Using the amazing JFK database explorer, one can compile a complete list of CIA records that are not yet online. (All hail Rex Bradford, creator of the JFK DBE and master of the MFF website!)
Because the NIA records are not factored in, however, this count is off. I have a corrected count which I will put up if anyone is interested (send me a note or leave a comment).
Eventually, of course, NARA has promised to put all these records online, but that project is going mighty slow. They are currently working on lengthy and boring documents from the Warren Commission. It would be nice if we could get these much more interesting CIA docs online. And it may actually be possible, through more mining of the non-pareil MFF collection.
The message of the almost missing disk: 104-10189
As noted above, RIF numbers for the CIA can be identified by lengthy prefixes such as 104-10189. The 104 is the agency number, meaning it was provided to NARA by the CIA. 10189 means it is from disk 189, one of the hundreds of floppy disks NARA supplied CIA to save its document metadata on.
Most of these disks hold up to 500 RIF sheet files. In the case of disk 189, there are 449 sheets, with RIF numbers 104-10189-10000 to 104-10189-10448. Each sheet belongs to a document, with 1 to many pages, and each sheet contains a standard set of metadata. The metadata is all available through the JFK database, now online via 6 excel files.
The documents, however, seem to be almost all missing from the MFF collection (and NARA’s online records as well). 416 records with this prefix are nowhere to be found (meaning 32 are available at MFF). Are there other disks like this? Yes, there are. Disk 104-10116 had 450 RIF sheets on it. Only 42 are available on MFF. There are many other cases where disks list hundreds of records that are not online.
But are they really not there? To answer this question takes real digging. As it turns out, the 104-10189’s missing records are not randomly missing. This is the disk that held the RIF sheets for a specific box of documents: the Domestic Contact Division’s files on Jim Garrison’s investigation into the Kennedy assassination.
A big chunk of the CIA records in the ARC are devoted to Garrison’s highly controversial probe. The DCD files amounted to 6 volumes. Then there was the Counterintelligence Staff’s investigation of the investigation (six more volumes), and apparently there is a third set of records on the “Garrison Case”, compiled by the CIA Office of Security (yet another six volumes). Many of these files are extremely redundant of course.
The DCD’s Garrison files make up the majority of the entries on disk 104-10189. Why weren’t these released like the other records? Well, they were. Look at https://www.maryferrell.org/showDoc.html?docId=55185 and you will find the entire first volume of the DCD files.
Skipping through docids 55185 to 55189 will get you the first five volumes. Volume six is at 55192, but this has a new disk number: 104-10170.
These records were all released in large units, each volume being packed into one file with hundreds of pages, and MFF scanned the files as single pdfs. The NIA sheet identifies these pdfs correctly, but the NIA number does not map onto any RIF sheets in the collection. Since they were released to NARA as units, not as single documents, they cannot be identified as ARC docs using their accession numbers.
The ARRB, the federal board that was created to assemble and oversee the release of the JFK documents, was, perhaps inadvertently, the cause of this strange result. The ARRB ruled that if CIA files had documents with redactions in them, CIA had to release them as individual documents, not in file units. This made for quite a mess.
Files make sense as a set. When the file is atomized and released as individual atoms or molecules, making sense of it is inevitably difficult, and putting it back together is quite a task. On the other hand, we now have the JFK database, which gives us a key to the unit, but we don’t have the RIF sheets that help us sort through the unit!
The brute force solution
If you want access to the individual documents in these files by RIF number, there are two solutions. Number one, wait for NARA to finish wading through the dead sea of JFK documents and post everything online. (How long did Moses and the Hebrews spend in the wilderness?)
Number two, take the brute force route and go through the file, comparing each page with the RIF sheets for disk 104-10189 (and 104-10170 for volume 6). I am now in knee deep in this slough of despair, but when it is done I will have another 418 RIF docs. But can it be done? Yes, it is not as difficult as you might expect. Tune it to the next installment on the blog to see how.