MKDELTA: What the CIA Records Actually Say

In July 1960, the CIA's deputy director for plans signed an instruction controlling the issue of “MKDELTA materials”. An operating division had to make a written request, the deputy director had to approve each issue, and unused material had to be returned. The technical office was to keep an issuance record. The instruction is attached to the CIA Inspector General's 1963 report, PDF pp. 43–44.

Three years later, the Inspector General described MKDELTA as the control system for operational use of material developed through MKULTRA research. The same report summarised a 1960 internal review that counted nine operations using disabling drugs over eight years. Yet a February 1975 CIA Inspector General memorandum reported that the operations directorate said the MKDELTA instruction had never been implemented in fact, and that the directorate held no operational records for it. 1963 IG, cover p. 1 and PDF pp. 18–19; 1975 IG memo, CIA release C01434878, PDF p. 2.

That denial does not settle the question. Seven months later, the same Inspector General's office sent the director, under the heading “MKULTRA/MKDELTA”, a list of “summaries of actual uses”, which remains withheld. Later in 1975, Senate staff read out a 1962 field dispatch that described “the other completed MKDELTA operation”. The rule is documented. Use of related materials is reported, and at least one 1962 operation was called MKDELTA in field messages read into the Senate record. Whether the rule's written approval and issue record were followed in any identified case is not established by the records examined here.

The CIA's 1963 Inspector General drew a bureaucratic line. MKULTRA covered research and development of chemical, biological and radiological materials intended to affect human behaviour. MKDELTA denoted the deputy director for plans' system for controlling their operational employment. The second name was not a guarantee that every experimental idea became a working field tool. CIA IG, 1963, cover p. 1 and report PDF p. 3.

The operational policy came before MKULTRA's 1953 research authority. The 1963 report refers to an earlier DD/P Notice No. 220-1 dated 20 October 1952, and the 1960 instruction says it rescinds a notice with that date. The CIA Inspector General's own January 1975 drafts also give 20 October, and so did a Senate questioner citing the 1960 instruction to Sidney Gottlieb in October 1975. The date 29 October 1952 first appears in the February 1975 versions of the Inspector General's review. The change is unexplained and is probably a slip. The original 1952 notice has not been located, but every earlier record seen gives 20 October. 1963 IG, PDF pp. 3, 43. Gottlieb deposition, 15 October 1975, printed p. 50. IG draft, 29 January 1975, para. 5. IG draft, 17 January 1975. February 1975 versions, PDF pp. 4, 10.

The names sit in a longer, overlapping history. The Church Committee traced earlier CIA interrogation work through BLUEBIRD (approved in 1950) and its 1951 renaming as ARTICHOKE. It described security and medical offices as important to that channel. MKULTRA was a research and funding route authorised in 1953; MKNAOMI concerned stockpiled incapacitating and lethal materials and delivery equipment; MKSEARCH later continued parts of MKULTRA's research. These labels identify different responsibilities and periods, not one uninterrupted chain of field approvals. Drug interrogations also ran outside MKDELTA. A November 1954 medical-branch report describes a completed “modified ARTICHOKE type interrogation” in which the subject was injected with a drug. The Church Committee found that the Office of Security “consistently sought” to bring the technical division's psychochemical operations under ARTICHOKE. Church Committee report in the August 1977 hearing, PDF pp. 70–71 and 93 (printed pp. 387–388, 410); Senate material on MKSEARCH, PDF p. 171. CIA medical branch memorandum, 15 November 1954, release C00149554.

A 1977 Department of Defense summary instead called MKDELTA the CIA's first “project” for operational use and MKULTRA its successor, adding that MKDELTA may never have been implemented. The memo describes these CIA projects as they appear “from the available documents”, and its “never … implemented operationally” phrasing closely follows the CIA's 11 February 1975 paper. It is a cautious secondary description, not independent confirmation or a case record. DoD memorandum, 20 September 1977, in the September 1977 Senate hearing, printed pp. 157–159.

The 22 July 1960 instruction defined MKDELTA materials as certain chemical and biological materials capable of producing behavioural or physiological changes; it expressly mentioned “K and L tablets”. It excluded the materials it called “harassment materials”. The instruction limited use to official duties that clearly justified it. It required personal written DD/P approval for each issue, not a standing licence to use a stockpile. A division or staff chief or deputy chief had to sign the request, technical officers assessed feasibility, and the technical office was to record the date, recipient and approved request. Unused material had to be returned. CIA IG appendix, PDF pp. 43–44.

That is a meaningful control design on paper. It also creates a test: for a particular operation, can we find the signed request, written DD/P decision, issuance entry and field result? This investigation found no complete matched chain. The instruction itself says earlier procedures may have differed, so imposing its 1960 rules on every 1950s action would be an anachronism. Its exclusion of harassment materials also means the CIA's larger harassment count cannot simply be labelled MKDELTA-approved use.

The 1963 Inspector General quoted a review written by Sidney Gottlieb in 1960. Gottlieb reported nine operations using disabling drugs in eight years; roughly two-thirds involved interrogations. He reported no use of lethal substances. He also reported about 100 operations using harassment materials, explicitly outside the MKDELTA regulation. These are an internal retrospective count by category, not a complete list of named subjects, countries or approvals. “No lethal use reported” describes the report's knowledge; it is not proof that no such use occurred. The 1960 report itself has not been released, and by September 1975 the Inspector General's office noted that it had no copy. The nine-operation figure therefore survives only as the 1963 summary. CIA IG, PDF pp. 18–19. IG memorandum “MKULTRA/MKDELTA”, 15 September 1975, p. 2.

The Church Committee later said MKDELTA was a special procedure to govern the use of MKULTRA materials abroad, and that “such materials were used on a number of occasions.” It said overseas use began in 1953, possibly earlier, but that destroyed files prevented a full reconstruction. Citing a 1957 Inspector General survey, it reported six drugs developed for operational use and used in six operations on 33 subjects, and said that by 1963 “the number of operations and subjects had increased substantially.” Those figures come from different reports, categories and years. They are not one running total, and none of them is a ledger of approvals. They do stop Gottlieb's nine-operation figure from being read as a lifetime ceiling. Church Committee report, reprinted in the 1977 hearing, printed pp. 391–392/PDF pp. 74–75.

In his 17 October 1975 Senate deposition, Gottlieb remembered some drug interrogations and described an approval chain running up to the deputy director for plans. He did not recall every specific approval or know what the deputy director did after a request reached him. Senate staff read from an older memorandum about two field trials using psychochemicals alongside interrogation; Gottlieb's reported assessment in that memo was favourable, but the deposition is not the original trial file. His memory gaps and the absence of a matching signed MKDELTA issue record prevent this testimony from proving that the formal procedure was followed in those cases. Gottlieb deposition, PDF pp. 26, 49, 53.

On 18 October, Senate staff showed Gottlieb a 1962 interrogation file that repeatedly used the MKDELTA name. A cable dated 18 May carried MKDELTA as a cable indicator. A separate 25 June 1962 cable described one interrogation: two sleepless nights “due A-2”, then P-1 (LSD) given around a meal, followed by a “paranoid panic reaction”. A dispatch about the series, as read into the record, said: “The other completed MKDELTA operation can be considered a success in that it induced a paranoid reaction in the presence of an unwitting psychiatrist.” On the indicator, Gottlieb said that people often put the wrong designator in, one that “covered some aspect of the operation”, and that “somebody made a mistake, I think.” He said he would assume that “the term MKDELTA interrogation meant P-1/A-2 interrogation”, and he recalled no others. The original 1962 messages have not been released. This transcript is the clearest available sign that field officers used the MKDELTA name for an actual operation. It does not show the deputy director's written approval or the technical office's issue record for that case. Gottlieb deposition, 18 October 1975, release C00095171, printed pp. 76–80, 83.

The 1963 Inspector General reported that approval of MKDELTA use “is now accomplished within the chain of command of the DD/P”, and that there was “an extremely low rate of operational use of the controlled materials.” Both are general statements. Neither identifies a case. CIA IG, printed pp. 19, 21.

The CIA's own technical assessment undercuts the popular idea of a reliable “truth drug”. The 1960 review said no effective truth serum, knockout pill or recruitment pill was then known, and described the field as high risk with low yield. The 1963 IG said drugs had become a support to broader interrogation methods rather than a dependable mechanism of control. A drug could disturb or disable someone without making their answers true. CIA IG, PDF p. 19.

A November 1957 memorandum requested planning for special interrogation of a defector. Its approval text was conditional: actual use required further specific approval by the Chief of Counterintelligence Staff, the Director of Security and an agency medical officer. The original scan is hard to read at the approval line; the proviso is also read into the later Senate deposition. The visible record documents a proposed technique and an approval barrier, not administration to that person. CIA memorandum, 13 November 1957, PDF pp. 1–2. Gottlieb deposition, PDF p. 63.

A January 1959 memorandum concerned a proposed P-1 (LSD) interrogation. It says the technical division had approved its part, but that security and medical concurrence still needed attention. The writer recommended approval subject to controls and wanted confirmation of exactly which technique would be used. Again, a proposal and recommendation are not evidence that the drug was given or that a MKDELTA issue was logged. CIA memorandum, 20 January 1959, p. 1.

These files make the control question concrete. They show officials discussing specific cases and conditions. They do not close the gap between permission, issuance and what happened to a subject. A complete record for one case would do more to resolve the MKDELTA question than another general description of the programme.

The channels were not interchangeable in practice. In his 1975 deposition, Gottlieb said his technical office did not normally clear P-1 operational work with the ARTICHOKE committee, and ordinarily did not involve the security office unless a particular case required it. He described a route up through the deputy director for plans but could not say what happened to every request after it reached that level. The 1957 proposal, by contrast, expressly reserved further counterintelligence, security and medical approvals. This is evidence of differing routes and retrospective recollection, not proof that either route completed the MKDELTA issuance process in a named case. 17 October 1975 deposition, PDF pp. 50–53; 1957 request, PDF p. 2. 1957 scan.

On 5 February 1975, CIA Inspector General Donald Chamberlain sent the director a review of behavioural-drug work. His cover memo said: “We have been informed by ADDO that although the DDO has an instruction on behavioral drugs -- MKDELTA -- this was established against the event that some such materials became available and has never been implemented in fact. There are no operational records in the DDO in this connection.” ADDO was the associate deputy director for operations, and DDO the operations directorate. CIA IG cover memorandum, release C01434878, PDF p. 2.

The released file shows how that statement entered the congressional record. It holds two versions of the attached paper. The 5 February version, marked “old”, describes the 1952 policy without saying anything about implementation. The revision dated 11 February 1975 adds one sentence: the policy “was in anticipation of the development of behavioral drugs, but was never implemented operationally.” The Church Committee cites the 11 February version as the CIA's memorandum to the committee. The new sentence is not attributed to anyone. Because it was added after the cover memo recorded the operations directorate's statement, it appears to repeat that statement rather than report a separate finding. The two versions also differ elsewhere. On CIA-sponsored work at Edgewood, the 5 February text says testing “involved twenty volunteers”. The 11 February text says Edgewood “did not progress to testing materials on human volunteer subjects under the work sponsored by CIA.” The revision was more than a copy-edit, and the file does not say which version is correct. February 1975 versions, PDF pp. 4, 7, 10, 12. Church Committee report, printed p. 387, note 4.

Later CIA papers neither repeated the February statement nor reconciled it with other records. On 15 September 1975 the Inspector General's office sent the director a memorandum titled “MKULTRA/MKDELTA”. It quoted Gottlieb's 1960 finding that the clandestine services “had up to that time shown little inclination to use the end products operationally”. It then said: “Specific records are scarce, but below are listed summaries of actual uses”. The list is withheld, so it is not known whether those uses were made under MKDELTA. Two days later the Office of Security told the Inspector General that its own files referred to “only a few specific instances where drugs were utilized overseas”, and that these were not a full list. That memo does not mention MKDELTA and covers drug use overseas generally. IG memorandum, 15 September 1975, p. 2. Office of Security memorandum, 17 September 1975.

The most defensible reading is a narrow one. In 1975 the operations directorate held no operational records for MKDELTA. Its statement may also mean that the formal approval-and-issue procedure was never used as written. The records examined here cannot test that narrower meaning. Read broadly, as a claim that the MKDELTA name or procedure was never used in the field, the statement conflicts with the 1962 messages read into the Senate record, which call one operation MKDELTA. Those messages survive only as read aloud, and Gottlieb doubted the cable indicator. The 1957 and 1960 counts show that related drugs were used in operations, but neither shows the MKDELTA procedure at work. Nothing found here proves that any particular use received the written approval the 1960 instruction required.

The CIA Reading Room was offline for maintenance at this article's evidence cutoff. The copy of the 1975 file linked here matches the digest of archived captures of the CIA-hosted file made between 2023 and 2025.

Most MKULTRA project files were destroyed in 1973. Recordkeeping was already thin: the 1963 IG noted missing planning and approval records for some test work. In 1977, officials found seven boxes of financial records that had survived outside the destroyed project files. That survival shows why “all records were destroyed” is wrong; the destruction still prevents a complete accounting of overseas operational use. 1977 Senate hearing, PDF pp. 8, 13; Church Committee appendix, printed pp. 391–392 and 404–405.

The missing approval chain is the decisive limit here. It bears on responsibility: a written gate could constrain a dangerous practice only if requests, decisions and material issues were actually recorded and followed. The documented inability to match those steps also hampers later scrutiny of who was exposed and on whose authority. That is a conclusion about the evidence trail, not a claim that a particular unnamed person was dosed without the required sign-off.

High confidence: MKDELTA was a written CIA operational-control procedure, distinct in purpose from MKULTRA's research authority, first set by a notice dated 20 October 1952. The 1960 instruction specified narrow materials, personal approval, return and recordkeeping. The 1975 “never implemented” statement came from the operations directorate. Its broad wording was added in a revised paper that went to the Senate. Moderate confidence: related materials were used in a limited number of operations, mainly interrogations. The 1957 and 1960 internal counts, the congressional investigation and participant testimony all report use, though their accounts overlap and are incomplete. At least one 1962 operation was called MKDELTA in field messages, as read into the Senate record. Contested: the 1975 statement, read broadly, conflicts with those 1962 messages. Its narrower meaning, that the formal procedure was not used, cannot be tested on the records found. Unresolved: how many operations occurred in total, and whether any identified operation followed the MKDELTA approval and issue procedure.

The conclusion would change if any of these records were released: the 1962 field messages; the Inspector General's withheld September 1975 list of “actual uses”; the technical office's issue register; Gottlieb's full 1960 report; or a matched deputy-director approval tied to a field report. Until then, the honest finding is a documented rule, reported use, and an unclosed gap between the two.

Method note

This investigation checked primary CIA and Senate records against their page images, along with specialist interpretations and current search results, through 25 October 2026. The 1952 notice, the 1960 Gottlieb report and a complete operation ledger were not found in the routes searched. The CIA Reading Room, the Senate Intelligence Committee site and the DoD file were offline or blocked at their official URLs during this check. CIA releases were read from Internet Archive mirrors of the Reading Room, and the 1977 Senate hearing from an archived capture of the Senate file. Both the key 1975 CIA file and the hearing match the digests of archived captures of the official copies. The DoD memorandum was read in its reproduction in the September 1977 Senate hearing, hosted by a third party. Search results and later summaries did not substitute for case evidence. This article was corrected on 25 October 2026. The changes fix a misquoted Church Committee phrase, revise the treatment of the 1952 date, explain where the 1975 “never implemented” wording came from, and add the 1962 and September 1975 records that were previously missing.