[03-23 update: After the first version of this note went online, NARA released a third batch of pdfs on 03-20. I have revised the note to discuss this new batch]
This post is a survey of the March 18 and 20 releases from the JFK Assassination Records Collection. NARA posted pdfs for 2182 records on March 18 in two batches, then a third batch of 161 records on 3/20. Except for a single locked drawer (labeled NSA), a faded raincheck for May 2027 (initialed MLK), and a few crumbs here and there, the ARC cupboard is bare.
The 3/18 releases
On March 18, pdfs of 2182 records in the ARC were posted at NARA in two batches. The link to these is here. Below is a table counting how many records, from which departments and agencies (D/A), had redactions released in those pdfs. Counts are listed by D/A prefix, D/A name, record count, and percent of total:
D/A prefix | D/A name | count | pct |
---|---|---|---|
104 | Central Intelligence Agency | 1483 | 68% |
119 | Department of State | 1 | <1% |
124 | Federal Bureau of Investigation | 139 | 6% |
157 | Senate Select Committee on Intelligence Activities | 68 | 3% |
176 | John F. Kennedy Library | 54 | 2% |
177 | Lyndon B. Johnson Library | 18 | 1% |
178 | Gerald R. Ford Library | 19 | 1% |
180 | House Select Committee on Assassinations | 232 | 10% |
194 | US Army Staff, ACSI G2 | 152 | 7% |
197 | Pentagon Telecommunications Center | 1 | <1 |
198 | Office of the Secretary of the Army | 11 | .5% |
202 | Joint Chiefs of Staff | 4 | <1% |
Is this good or bad? How close did NARA come to releasing all the redactions? Well, compare the following table, which gives MY count of redacted records. This table is based on the 2 lists that I posted last week (here and here):
D/A prefix | D/A name | count | pct |
---|---|---|---|
104 | Central Intelligence Agency | 1482 | 68% |
119 | Department of State | 1 | >1% |
124 | Federal Bureau of Investigation | 230 | 10% |
144 | National Security Agency | 207 | 9% |
157 | Senate Select Committee on Intelligence Activities | 68 | 3% |
176 | John F. Kennedy Library | 54 | 2% |
177 | Lyndon B. Johnson Library | 23 | 1% |
178 | Gerald R. Ford Library | 21 | 1% |
180 | House Select Committee on Assassinations | 229 | 10% |
194 | US Army Staff, ACSI G2 | 151 | 07% |
197 | Pentagon Telecommunications Center | 1 | >1% |
198 | Office of the Secretary of the Army | 11 | >1% |
202 | Joint Chiefs of Staff | 4 | >1% |
As the table shows, I claimed there were a total of 2482 records that still had redactions, and provided a complete listing of each one. The 2482 records I listed exceed the 2182 pdfs NARA posted. So which records were not posted?
Here is a table summing up the differences between what I listed and what NARA DIDN’T post in the first two batches:
D/A prefix | count |
---|---|
104 | 5 |
124 | 105 |
144 | 207 |
157 | 1 |
176 | 1 |
177 | 5 |
178 | 2 |
Summing up, there were 326 records on my list that did not have a pdf posted at NARA on 3/18.
But wait! was my list 100% accurate? Before I answer, let me ask that a somewhat different way: did any records come out that were NOT on my lists? Alas, some did. Here is a count of these:
D/A prefix | count |
---|---|
104 | 5 |
124 | 8 |
157 | 1 |
176 | 1 |
180 | 3 |
194 | 1 |
This shows that 19 records which I did not think were redacted had pdfs posted at NARA. I have now checked all these records, and the 2022-2023 versions of the records did indeed have redactions. That means I missed redactions in some records. How many records did I miss? Fewer than 1 percent of the total. Good enough.
The 3/20 releases
On March 20, NARA posted what it said was another 161 pdfs. Since this was the third set of records posted at NARA this time around, I will call these batch 3. When I downloaded batch 3, however, I wound up with 164 pdfs. It is quite hard to figure out what is going on here, for now I will just post a count of the prefixes and record counts I got in my download:
D/A prefix | count |
---|---|
104 | 1 |
124 | 124 |
135 | 14 |
144 | 2 |
157 | 1 |
178 | 1 |
180 | 2 |
198 | 1 |
206 | 18 |
The 135 prefix is for Pike Committee records. The 206 prefix is for PFIAB records. The Pike Committee was the House version of the Church Committee. PFIAB was the Presidential Foreign Intelligence Advisory Board. There are not many ARC docs available on line from either of these sources, probably because there simply aren’t that many in the Collection.
Some of the Pike Committee records in batch 3 appear online for the first time. Some of the PFIAB records were available at MFF, but were NOT posted online by NARA from 2017-2023.
The PFIAB records online at MFF had redactions. Now those redactions are being released. It seems NARA missed these redacted records in doing its earlier releases. They did not occur in any previous record lists, and were not mentioned in the D/A transparency plans that told us what was redacted in the ARC.
This is a good sign, actually. It means NARA is going rigorously through the collection to make SURE everything is open.
What about the other records?
The one CIA record in this batch WAS previously released, perhaps NARA (or CIA) just realized that there was still redacted text in it, so they released it on 3/20, instead of 3/18. Better late than never.
There are a surprising number of FBI records in batch 3. Some of these were previously available online at MFF. Those previous versions were redacted, but like the PFIAB records, they never appeared on any 2017-2023 record lists and were NOT mentioned in the FBI’s transparency plan. So again, this is another set of overlooked records with redactions that are being released in full.
The other records in batch 3 are similar cases. These are records with overlooked redactions that are being released in full. There are no redactions in the pdfs online.
Summary of the 2025 releases
Virtually every record that we knew about is released in full. I will discuss this “in full/not in full” question in later posts, especially for the CIA records, which constitute the bulk of the releases.
The largest set of records that did NOT appear this time were NSA records (record prefix 144-). No NSA records were posted at NARA on 3/18. Two NSA records were posted on 3/20. I haven’t looked at them yet.
Earlier I counted 207 NSA records still redacted in the DOD transparency plan, that’s still the largest difference between my count of redacted records and what actually got released.
The NSA records are actually RELEASED, don’t be confused about this. It’s just that they still have text redacted. I think it is possible that these NSA records, which contain sensitive communications intel, will keep their redactions for a while.
Otherwise my lists of redacted records were very close to what NARA released this time. Just looking at the totals of records on the lists isn’t the whole picture of course.
If you look at the totals, you might think I got the CIA records almost on the nose. I said 1482 were redacted, and 1483 were released. The difference was actually bigger. Five of the CIA records I said were redacted were not released, and SIX CIA records were released that were not on my list.
Those 5 CIA records on my list were redacted, no question. Why didn’t they get posted this time? Don’t know. Will they be released in full later? Don’t know. I’ll discuss all this is a future post.
Figuring out what’s going on with the FBI records is very difficult. The main problem is with what I will call “multi-record pdfs.” Way back in 2017, NARA started releasing FBI records in pdfs that combined multiple records. It was not simple to figure out where one record stopped and another record started in these cases, and it was sometimes unclear which records one was even looking at.
This problem has come back to bite researchers trying to figure this out. I believe that the FBI itself may have been confused in a couple of cases. It will take a while to figure out what happened with these records, but I will venture to say that several FBI records with MLK documents in them were not released. There are also some FBI records released this time which have tax information that is redacted by law. This cannot be released unless the law is changed.
Remaining are five or six DOD or CIA records. As I have mentioned in other articles and posts, there was one record which recorded “nuclear yield” per bomber, an abbreviated way of describing how many nuclear bombs U.S. bombers were carrying, and how powerful the bombs were. This record was not re-released. I hasten to add that the record is released, all 23 pages, with only three numbers, the yield numbers, redacted. I don’t think these three numbers will be released any time soon.
A couple of other records which redacted discussions of nuclear strategy were not re-released this time either. When will THOSE redactions get released? Not in your lifetime? Hey, don’t say that. May you live to be a hundred.
And of course, the SSNs ALL came out. Are you glad? Not me, I think it’s a disgrace, a blot on the ideal of government transparency. How will the Trump administration handle this crazy snafu? Don’t know, just glad it wasn’t my SSN that got published. At the NATIONAL ARCHIVES. Sheesh.
Overall, I think I did a damn good job of counting which records had redactions. I seem to have missed a few. Hardly any, really.
Prior to the the 2025 releases, many news stories reported a figure of 3600+ records. This figure was wrong, as I noted in my earlier posts on this subject (see here for the posts).
Rex Bradford at MFF, however, thinks that my count is pretty close to the true number (see here for Rex’s note). So that makes two of us, at least.
Of course more of these “overlooked redacted records” could show up, but I doubt there will be many of them. And if they do show up, NARA’s policy is clear: post an unredacted pdf toot sweet!
Thankyou for taking up this subject in such detail.
I read the whole article above, and by the end, I get some of the fine distinctions that you have been observing.
Good job!
Thanks!
Hi Robert,
I for one didn’t get the three bonus pdfs when downloading batch 3, so I wonder what that’s all about. I did, however, notice that several of the “multirif” pdfs in batch 2 and 3 are duplicates (I used WinMerge for verification) and therefore quite redundant.
Keep up the good work!
Br Mark
Hey Mark,
Interesting about the duplication, sounds like it’s worth a post!
Well, I counted 32 duplicate files, most of which were fairly large, so it does noticeably affect the total page count (for those that are into that sort of thing).