Filling the gaps in the ACRS: Three sets of FBI records

The Assassination Collection Reference System (ACRS) is NARA’s online database of finding aids for the JFKARC. This post discusses three sets of FBI records in the ARC which are not listed in the ACRS. Using the Mary Farrell Foundation’s online collection, ARC documents posted online at NARA, and ARRB record notices published in the Federal Register, I have put together an excel file of the missing data which is available here.

The gaps in the ACRS

The background for this post is given in two earlier notes: “Missing” RIFs at NARA and Gaps in the ACRS: A note from NARA

Summarizing these notes, the ACRS should have metadata for all records added to the ARC after 1992. The metadata was provided to NARA in two forms: 1), a cover sheet for each record in the Collection, called a RIF (Reader Information Form) giving the metadata in tabular format; 2), computer disks on which the agency providing the record input the RIF metadata using NARA’s software.

Unfortunately, as NARA staff explained to me, some of these disks were damaged when NARA received them, so that they could not be merged into the ACRS database. This means that none of the records from agencies such as the Secret Service and NSA are listed on the ACRS, a very unfortunate situation for researchers.

The FBI disks

This also happened for three disks of FBI record metadata: Disks 124-10203, 124-10204, and 124-10223.

What do these numbers mean? 124 is the agency prefix for the FBI. 10XXX is the “disk number,” the number of the disk that the FBI gave NARA. For example 10203 would be the 203rd disk of RIF data which the FBI gave to NARA.

Each disk held metadata for up to 500 records, which were numbered serially on the disk so that using the prefix, disk number, and record number (NARA calls this the “control number”) gives you a 13 digit RIF number, which is (theoretically) unique for each record in the Collection: in this case, 124-10203-10000, 124-10203-10001, up to 124-10203-10499, the last record on disk 10203.

Although these three FBI disks were scrongled, the documents are still there at NARA, each with a RIF sheet on top, so that NARA could reenter the data in the ACRS itself if it wanted to. It has chosen not to do so, apparently because of the cost. If researchers want this data, they will have to go to NARA’s College Park facility in Maryland, where the entire Assassination Records Collection is stored.

Recompiling the metadata: 3 sources

Fortunately for the curious but cash-strapped researcher, much of the RIF metadata for the three disks can be recovered from three sources:

  1. the Mary Ferrell Foundation’s online collection of ARC records;
  2. The ARC records posted at NARA from 2017-2018;
  3. The record notices which the Assassination Records Review Board (ARRB) published in the Federal Register (FR) from 1995 to 1998.

In fact, the latest version of the MFF collection now incorporates all of source two, the 2017-2018 records posted at NARA. For reasons explained below, however, I will still separate these two sources, using MFF to designate the FBI records online at MFF prior to the NARA releases.

The MFF records (prior to adding the NARA records) have the most complete metadata for these three disks. They have complete RIF forms for many documents, and many “duplicate copies” from all three disks. I will explain what these duplicates are in a another post.

Most of the NARA records are missing their RIF forms, and are shorter, compared to the versions in the MFF Collection, for reasons I will also explain in another post. There are, however, 11 records with RIF forms here that were not originally present in the MFF, so I can add this data as well.

The ARRB record notices give only a few bits of data, including the ARC record number, the date a document’s release was voted on, and the number of text bits which the ARRB agreed to postpone releasing. Nonetheless, this data fills in a few missing record numbers and helps us understand what the NARA releases were all about.

Putting together the information from these three sources, one can see that there were RIFs for 1500 records on the three disks, with 500 RIFs on each disk:

  1. 124-10203-10000 to 124-10203-10499
  2. 124-10204-10000 to 124-10204-10499
  3. 124-10223-10000 to 124-10223-10499

One record is not mentioned anywhere: 124-10223-10202 is not in the MFF collection or the NARA releases, and is not listed in FR notices. Probably released without redactions by the FBI in the early days of the collection. The FR notices also have two RIF numbers which seem to be from disk 10223, but which are probably typos.1

NARA released 842 pdfs of records in the three series during 2017-2018. There was extensive duplication in the releases, so that there were actually only pdfs from 466 unique records posted online at NARA. Most of these pdfs did NOT have RIFs, but 11 of them did have RIF sheets not available from the MFF.

As a result, there are 77 records for which I have only the FR data (78 counting the missing record 10202). The rest of the records I have complete sets of metadata. This gives us a fairly complete picture of what was in the three series and how they were handled by the ARRB.

What’s on the disks

Repeating again the link from above, the excel file for all of this data is here.

In addition to the RIF metadata, the excel file also gives data from the ARRB’s Federal Register notices, including: the number of the FR notice where I found the record, the number of text postponements in the document as listed in the FR notice, whether or not the record has Grand Jury information, and the release date for the text postponements.

I also list the filenames for documents released by NARA in 2017-2018, and the docids for records online at the MFF. This should be sufficient for anyone who wants to find the documents online.

Overall, these are typical RIF forms from the FBI. The FBI had its own style for filling out NARA’s RIF forms, and the RIFs on these disks are typical of this style. For example, the FBI doesn’t often use the title field, leaving it blank for most records. Instead, it uses the filenumber field to give the FBI file number of the document, the record series field to show which FBI office produced the document, and the subject field to give a fairly clear idea of the content of the record (once you can read the esoteric FBI abbreviations).

Putting this all together, the disks are a general mishmash of FBI documents. This is frequent in the ARC. Just because the FBI input these records on the same disk, using consecutive numbers, doesn’t mean that the documents are related to each other. There are runs of related material, but then they switch to other files or topics. Not totally scrambled, but don’t look for much order here. In terms of overall subjects though, these disks are largely files on organized crime. Here are the three most frequent file numbers for documents on the disks:

92-6054 La Cosa Nostra 251 records
92-2717 Angelo Bruno 142 records
92-3182 Gus Alexander 136 records

This is over a third of the records in this set. There are probably another 400-500 records in this category as well, so about two thirds of the records are organized crime. The others are Cuban related (MDC, Orlando Bosch, Louis Henry Jones, etc.) and almost a hundred records on Susan Frank (a CPUSA activist later expelled from the Party for her pro-Chinese Communist activities).

Numerous redactions from the Frank records were released by NARA in 2017-2018, and these newly unredacted documents have been useful source material for recent research, such as Darren Trombay’s 2019 article “From Old Left to New Left: The FBI and the Sino-Soviet split” in the International Journal of Intelligence and CounterIntelligence.

Given the much larger number of organized crime records, the failure to include the records from the three disks in the ACRS has surely had a negative effect on research into organized crime. Eventually NARA will fix this (I hope). For those who want access to the RIF data right away, download the excel file!

  1. 124-10223-20106 and 124-10223-40124. The latter may be a typo for 124-10223-10124, no idea what the former was supposed to be.