Dear Mr. Blakey:
Attached you will find pages 1-8, 14, 17-18, 24-25, 57-59, 62a-63, 94, 118, 130-133 from the report by the Inspector General on plotting against Castro. These pages have been sanitized and declassified in compliance with a request by Mr. Goldsmith, and may be used by you in your hearings. It should be noted that it is Agency policy to not release reports of the Inspector General outside the Agency, in order to protect the unique fact-finding advisory function that they fulfill within the Agency. Were the internal confidentiality of the reports to be compromised the effectiveness of the function could be impaired. Because so much of the information from the 1967 report was placed officially in the public domain by the interim report of the Church Committee, Alleged Assassination Plots Involving-Foreign Leaders, it can be released in this form on this special occasion.
Principal Coordinator, HSCA encl.
25 April 1967
This reconstruction of Agency involvement in plans to assassinate Fidel Castro is at best an imperfect history. Because of the extreme sensitivity of the operations being discussed or attempted, as a matter of principle no official records were kept of planning, of approvals, or of implementation. The few written records that do exist are either largely tangential to the main events or were lout on paper from memory years afterward. William Harvey has retained skeletal notes of his activities during the years in question, and they ere our best source of dates. of the Office of Medical Services, has a record of whom he met and when and cryptic references to the subjects discussed of TSD, has a record of two or three dates that are pertinent. and were involved in only the technical aspects of operational planning, and their participations were short-lived. Although fragmentary, their records are a help in establishing critical time frames. Operational files are useful in some instances, because they give dates of meetings, the substances of which may be inferred from collateral information. For the most part, though, we have had to rely on information to us orally by people whose memories are fogged by time. Their recollections of dates are particularly hazy, and some of them -1- are no longer able to keep the details of one plan seprate from those of another. We interviewed everyone whom we could identify as likely to be knowledgeable, with the exceptions of Mr. Dulles and General Cabell. A complete list is attached at Tab A. We did not go on fishing expeditions among the mere possibles. To have done so have risked making witting a number of employees who were previously unwitting and; .in our estimate; would have added little to the details available from those directly involved. There are inconsistencies among the various accounts, but most of them can be resolved by collating the information furnished by all of the identifiable participants in a particular plan and by then checking it against specific dates that can be fixed with fair certainty. We believe that this reconstruction of what happened and of the thinking associated with it is reasonably sound. If there are significant inaccuracies in the report, they are most likely to occur in faulty ordering of the sequence of events. People still remember much of what happened; but they can no longer recall precisely when. It became clear very early in our investigation that the vigor with which schemes were pursued within the Agency to eliminate Castro personally varied with the intensity of the U.S. Government's efforts to overthrow the Castro regime. We can identify five separate phases in Agency assassination planning, although the transitions from one to another are not allways sharply defined. Each phase is a reflection of the then prevailing Government attitude toward the Cuban regime. a. Prior to August 1960: All of the identifiable schemes prior to about August 1960, with one possible exception, were aimed only at discrediting Castro personally by influencing his behavior or by altering his appearance. b. August 1960 to April 1961: The plots that were hatched in late 1960 and early 1961 were aggressively pursued, and were viewed by at least some of the participants as being merely one aspect of the over-all active effort to overthrow the regime that culminated in the Bay of Pigs. c. April 1961 to late 1961: A major scheme that was begun in August 1960 was called off after the Bay of Pigs and remained dormant for several months, as did most other Agency operational activity related to Cuba. d. Late 1961 to late 1962: That particular scheme was reactivated in early 1962 and was again pushed vigorously in the era of Project MONGOOSE and in the climate of intense administration pressure on CIA to do something about Castro and Cuba. e. Late 1962 until well into 1963: After the Cuban missile crisis of October 1962 and the collapse of Project MONGOOSE, the -3- aggressive scheme that was begun in August 1960 and revived in April 1962 was finally terminated in early 1963. Two other plots were originated in 1963, but both were impracticable and nothing ever came of them. We cannot overemphasize the extent to which responsible agency Kennedy officers felt themselves subject to the ^ administration's severe pressures to do something about Castro and his regime. The fruitless and, in retrospect; often unrealistic plotting should be viewed in that light. Many of those we interviewed stressed two points that are so obvious that recording them here may be superfluous. We believe, though, that they are pertinent to the story. Elimination of the dominant figure in a government, even when loyalties are held to him personally rather than to the government as a body, will not necessarily cause the downfall of the government. This point was stressed with respect to Castro and Cuba in an internal CIA draft paper of October 1961, which was initiated in response to General Maxwell Taylor's desire for a contingency plan. The paper took the position that the demise of Fidel Castro, from whatever cause, would offer little opportunity for the liberation of Cuba from Communist and Soviet Bloc control. The second point, which is more specifically relevant to our investigation, is that bringing about the downfall of a government necessarily requires the removal of its leaders from positions of power, and there is always the risk that the participants will resort to assassination. Such removals from power as the house arrest of a--Mossadeq or the flight of a Batita should not cause one to overlook the killings of a Diem or of a Trujillo forces encouraged but not controlled by the U.S. Government. There is a third point, which was not directly made by any of those we interviewed, but which emerges clearly from the interviews and from review of files. The point is that of frequent resort to synecdoche--the mention of a part when the whole is to be understood, or vice versa. Thus, we encounter repeated references to phrases such as "disposing of Castro", which may be read in the narrow literal sense of assassinating him, when it is intended that it be read in the broader, figurative sense of dislodging the Castro regime. Reversing the coin, we find people speaking vaguely of "doing something about Castro" when it is clear that what they have specifically in mind is killing him. In a situation wherein those speaking may not have actually meant what they seemed to say or may not have said what they actually meant, they should not be surprised if their oral than was intended. The suggestion was made to us that operations aimed at the assassination of Castro may have been generated in an atmosphere stress in intelligence publications on the possibility of Castro's demise and on the reordering of the political structure that would follow. We reviewed intelligence publications from 1900 through 1966. The NIE's on "The Situation and Prospects in Cuba" * * * * * * * * have brief paragraphs on likely successor Castro were to depart the scene. In each case the treatment is no more nor less than one would expect to find in comprehensive round-ups such as these. We conclude that there is no reason to believe that the Operators were unduly influenced by the content of Intelligence publications. Drew Pearson's column of 7 March 1967 refers to a reported CIA plan in 1963 to assassinate Cuba's Fidel Castro. Pearson also has information, as yet unpublished, to the effect that there was a meeting at the State Department at which assassination of Castro was discussed and that a team actually landed in Cuba with pills to used in an assassination attempt. There is basis in fact for each of those three reports. a. A CIA officer passed an assassination weapon to an Agency Cuban asset at a meeting in Paris on 22 November 1963. The weapon was a ballpoint pen rigged as a hypodermic syringe. The CIA officer suggested that the Cuban asset load the syringe with Black Leaf 40. The evidence indicates that the meeting was under way at the very moment President Kennedy was shot. b. There was a meeting of the Special Group (Augmented) in Secretary Rusk's conference room on 10 August 1902 at which Secretary McNamara broached the subject of liquidation of Cuban leaders. The discussion resulted in a Project MONGOOSE action memorandum prepared by Edward Lansdale. At another Special Group meeting on 31 July 1964 there was discussion of a recently-disseminated Clandestine Services information on a Cuban exile plot to assassinate Castro. CIA had refused the exile's request for funds and had no involvement in the plot. c. CIA twice (first in early 1962 and again in early 1932) supplied lethal pills to U.S. gambling syndicate members working in behalf of CIA on a plot to assassinate Fidel Castro. The 1961 plot aborted and the pills were recovered. Those furnished in April 1962 were passed hy the gambling syndicate representetive to a Cuban exile header in Florida, who in turn had them sent Cuba about May 1962. In June 1962 the exile leader reported that a team of three men had been dispatched to Cuba to recruit for the operation. If the opportunity presented itself, the team would make an attempt on Castro's life--perhaps using the pills.
This report describes these and other episodes in detail; puts them into perspective; and reveals, that while the events described by Drew Pearson did occur and are subject to being patched together as though one complete story, the implication of a direct, causative relationship among them is unfounded. Gambling Syndicate The first seriously-pursued CIA plan to assassinate Castro had its inception in August 1960. It involved the use of members of the criminal underworld with contacts inside Cuba. The operation had two phases: the first ran from August 1960 until late April or early May 1961, when it was called off following the Bay of Pigs; the second ran from April 1962 until February 1963 and was merely a revival of the first phase which had been inactive since about May 1961. Gambling Syndicate - Phase 1 August 1960 Richard Bissell, Deputy Director for Plans, asked Sheffield Edwards, Director of Security, if Edwards could establish contact with the U.S. gambling syndicate that was active in Cuba. The objective clearly was the assassination of Castro although claims that there was a studied avoidance of the term conversation with Bissell. Bissell recalls that the idea originated with J.C. King, then Chief of WH Division, although King now recalls having had only limited Knowledge of such a plan and at a much later date--about mid-1962. same way. A memorandum for the record prepared by Sheffield on 14 May 1962 states: "No monies were ever paid to Roselli and Giancana. Maheu was paid part of his expense money during the periods that he was in Miami. (Giancana is "Gold.") * * * * * * * was introduced (in true name) to Roselli as an employee of Maheu, the explanation being that * * * * * * would handle the case for Maheu, because Maheu was too busy to work on it full time himself. No one else in the Office of Security was made witting of the operation at this time. Edwards himself did not meet Roselli until the summer of 1962. At this point, about the second half of September, Shef Edwards told Bissell that he had a friend, a private investigator, who had a contact who in turn had other con%acts through whom syndicate elements in Cuba could be reached. These syndicate elements in Cuba would be willing to take on such an operation. As of the latter part of September 1960, Edwards, * * * * * * *, and Bissell were the ones in the Agency who knew of a plan against Castro involving U.S. gangster elements. Edwards states that Richard Helms was not informed of the plan, because Cuba was being handled by Bissell at that time. With Bissell present, Edwards briefed the Director (Allen Dulles) and the DDCI (General Cabell) on the existence of a plan involving members of the syndicate. The discussion was circumspect; Edwards deliberately avoided the use of any "bad words." The descriptive term used was "an intelligence operation." Edwards is quite sure that the DCI and the DDCI clearly understood the nature of the operation he was discussing. He recalls describing the channel as being from A to B to C." As he then envisioned it, "A"'was Maheu, "B" was Roselli, and "C" was the principal in Cuba. Edwards recalls that Mr. Dulles merely nodded, presumably in understanding and approval. Certainly, there was no opposition. Edwards states that, while there was no formal approval as such, he felt that he clearly had tacit approval to use his own judgment. Bissell committed $150,000 for the support of the operation. (Comment: In the light of this description of the it is appropriate to conjecture as to dust what the Director did approve. It is safe to conclude, given the men participating and the general subject of the meeting, that there was little likelihood of misunderstanding--even though the details were deliberately blurred and the specific intended result was never stated in unmistakable language. It is also reasonable to conclude that the pointed avoidance of "bad words" emphasized to the participants the extreme sensitivity of the operation.) During the week of 25 September 1960, * * * * * * and Maheu went to Miami where Roselli introduced only Maheu to "Sam Gold" at a meeting tying it to a recollection that Castro frequently drank tea, coffee, or bouillon, for which a liquid poison would be particularly well suited. January - February 1961 Despite the decision that a poison in liquid fore would be most desirable, what was actually prepared and delivered was a solid in the form of small pills about the size of saccharine tablets * * * * remembers meeting with Edwards and * * * * * in Edwards' office to discuss the requirement. The specifications were that the poison be stable, soluble, safe to handle, undetectable, not immediately acting, and with a firmly predictable end result. Botulin comes nearest to meeting all of those requirements, and it may be put up in either liquid or solid form. * * * * * * * * states that the pill form chosen because of ease and safety of handling. (Comment: The gangsters may have had some influence on the choice of a means Of assassination. * * * * * says that in his very early discussions with the gangsters (or, more precisely, Maheu's discussions with them) consideration was given to possible ways of accomplishing the mission. Apparently the Agency had first thought in terms of a typical, gangland-style killing in which Castro would be gunned down. Giancana was flatly opposed to the use of firearms. He said that no one could be recruited to do the job, because the chance of survival and escape would be negligible. Giancana stated a preference for a lethal pill that could be put into Castro's food or drink. Trafficante the courier") was in touch with a disaffected Cuban official with access to Castro and presumably of a sort that would enable him to surreptitiously poison Castro. The gangsters named their man inside as * * * * * * * who was then * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * The gangsters said that * * * * * * had once been in a position to receive kickbacks from the gambling interests, had since lost that source of income; and needed the money.) When Edwards received the pills he dropped one into a glass of water to test it for solubility and found that it did not even disintegrate, let alone dissolve. * * * * * * took them hack and made up a new hatch that met the requirement for solubility. Edwards at that point wanted assurance that the pills were truly lethal. He called on * * * * * to make an independent test of them. Edwards gave * * * * money to buy guinea pigs as test animals. * * * * * * has a record of a conversation with * * * * * * * on 6 February 1961. It may have related to the tests, but we cannot be sure. What appears to have happened is that * * * * * tested the pills on the guinea pigs and found them ineffective. Gambling Syndicate - Phase 2 William Harvey, Chief of * * * * * was briefed In February 1961 (by authority of Richard Bissell) on phase one of the gambling syndicate operation. That briefing was in connection with a sensitive operation that Bissell had assigned to Harvey. Harvey describes it thus: Early in the Kennedy administration, Bissell called him in to discuss what Harvey refers to as an Executive Action Capability; i.e., a general stand-by capability to easy out assassinations when required. Harvey's notes quote Bissell as saying, "The White House has twice urged me to create such a capability." Bissell recalls discussing the question of developing a general capability with Harvey. He mentioned the Edwards/gambling syndicate operation against Castro in that context; but he now thinks that the operation was over by then and that reference to it was in terms of a past operation as a case in point. It was on this basis that Harvey arranged to be briefed by Edwards. Harvey's fixing of the date as February was only after review of events both preceding the briefing and following it. He says now that it might have been as early as late January or as late as March 1961. After some discussion of the problems involved in developing an Executive Action Capability, Bissell palced Harvey in charge of the effort. Harvey says that Bissell had already discussed certain aspects of the problem with * * * * * * * * and with * * * * * *. Since * * * * * * was already cut in, Harvey used him in developing the Executive Action Capability, although never with respect to Castro. We did not question * * * * * * * * on his knowledge of the program for creating an Executive Action Capability, but Harvey's mention of him in this connection may explain a notation by * * * * * * that Harvey instructed * * * * * * * to discuss techniques with * * * * * without associating the discussion with the Castro operation. Harvey states that after the decision was made to go ahead with the creating of an Executive Action Capability, and while he still discussing its development with Bissell, he briefed Mr. Helms fully on the general concept but without mention of the then ongoing plan to assassinate Castro. The Executive Action program came to be known as ZRRIFLE.. Its principal asset was an agent, QJWIN, who had been recruited earlier by * * * * * * * * * for use in a special operation in the Congo (the assassination of Patrice Lumumba) to be run by * * * * * * * * * * * * * * * * * * * made a survey of the scene, decided he wanted no part in an assassination attempt, and asked to be released--which Bissell granted.) The project name, ZRRIFLE,. first appears in the files May 1961, although the first recorded approval is dated 19 February 1962. The new DD/P (Helms) on that date authorized Harvey, by - 38 - memorandum, to handle the project on a special basis. Accounting for expenditures was to be by general category and on Harvey's certification. The initial approval was for $14,700, consisting of $7,200 for QJIN's annual salary and $7,500 for operational expenses. Project ZRRIFLE was covered as an operation (ostensibly to develop a capability for entering safes and for kidnapping couriers). It continued on a course separate from the Edwards/gambling syndicate operating against Castro until 15 November 1961. Harvey has a note that on that date he discussed with Bissell the application of the ZRRIFLE program to Cuba. Harvey says that Bissell instructed him to take over Edwards' contact with the criminal syndicate and thereafter to run the operation against Castro. Harvey adds that as a completely unrelated development, shortly after this discussion with Bissell he was told by Helms that he was to be placed in charge of the Agency's Cuba task force. Late 1961 - Early 1962 Harvey recalls that he was very busy with a number of things in the period that followed the discussion with bissell that led to his taking over Edwards' Castro operation. He was turning over his responsibilities in. He was working with NSA on the Martin/Mitchell defection case. He was reading in on Cuba operations and briefed the DDCI, General Carter, who said he understood the situation and in due time might brief the Director, Mr. McCone. It is not known whether General Carter did or did not brief Mr. McCone. There is no indication that General Carter was further briefed on the full details of the assassination plot against Castro. May 1962 The Attorney General obviously was told of CIA's operational involvement with gangster elements, because he requested a briefing on the details. On May 1962 Sheffield Edwards and Lawrence Houston met with Attorney General Robert Kennedy and, as Edwards puts "briefed him all the say. " Houston says that after the briefing Kennedy "thought about the problem quite seriously." The Attorney General said that he could see the problem and that he could not proceed against those involved in the wiretapping case. He spoke quite firmly, saying in effect, "I trust that if you ever try to do business with organized crime again--with gangsters--you will let the Attorney General know before you do it." Houston quotes Edwards as replying that this was a reasonable request. Edwards says that among the points covered was that of Roselli's motivation. The Attorney General had thought that Roselli was doing the job (the attempt at assassination of Castro) for money. Edwards corrected that impression; he was not. Houston recalls that during the meeting with the Attorney General the latter asked for a memorandum record of the meeting. Edwards believes that the request was made later and by telephone. A memorandum was prepared and was signed by Edwards. Both Edwards and Houston recall having had a hand in writing it. A transmittal buckship from Houston to Kennedy notes that the request was made on 11 May, which suggests that Edwards is correct in his belief that the request was made by telephone after the 7 May briefing of the Attorney General. The memorandum is dated 14 May 1962- It was typed in two copies only with the original being sent to Attorney General Kennedy and the other copy being retained by the Director of Security. It typed by Edwards' secretary, * * * * * * * * * *. It does not state the purpose of the operation on which Kennedy was briefed, but it does make it clear that the operation was against Castro and its true purpose may be inferred from the memorandum. Edwards states that the briefing of the Attorney General and the forwarding of a memorandum of record was canned out without briefing the Director (John McCone), the DDCI (General Carter), or the DD/P (Richard Helms). He felt that, since they had not been privy to the operation when it was under way, they should be ]protected from involvement in it after the fact. As noted previously, Houston had briefed the DDCI on the fact that there was a matter involving the out of the meeting * * * * and AMLASH were informed that President Kennedy had been assassinated. * * * * * was visibly moved over the news. He asked, "Why do such things happen to good people?" The contact report does not state the time nor the duration of the AMLASH meeting, but it is likely that at the very moment President Kennedy was shot a CIA officer was meeting with a Cuban agent in Paris and giving him an assassination device for use against Castro. * * * * * * * states that he received an OPIM cable from FitzGerald that night or early the next morning telling him that everything off. We do not find such a cable in the AMLASH file. There is a record in the file that * * * * * * * was due to arrive back in Washington at 1810 hours, 23 November. The AMLASH project was probably about as widely known within the Clandestine Services as any other project of a similar nature. However, we can identify only four people who how of the justdescribed episode involving a hypodermic syringe and Black Leaf 40. * * * * * * * knew all of the story, * * * * * * * knew most of it, * * * * * * * * * much of it. Fitzgerald did not mention this aspect of the AMLASH operation when he first briefed us on it. When we went back him later with specific questions, he said he remembered something about Black Leaf 40, but nothing ,whatever about a device for administering it. * * * * * * said he had the impression that FitzGerald (Text in book turn sideways) him on the Pearson story. d. Pearson, Anderson, and Greenspun (in Los Vegas) are newspapermen with a newsworthy story. Pearson has already published much of it. e. Maheu does have good reason for not wanting aired further. Unfavorable publicity might cause him to lose his lucrative client, Howard Hughes. There might be some value to be gained from endorsing his suggestion that he approach * * * * * * * and perhaps Roselli and urge discretion. What do other components of Government know about this operation? Former Attorney General Robert Kennedy vas fully briefed by Houston and Edwards on 7 May 1962. A memorandum confirming the oral briefing was forwarded to Kennedy on 14 May 1962. The memorandum does not use the Word "assassinate," but there is little room for misinterpretation of what was meant. Presumably the original of that memorandum is still in the files of the Justice Department. It should be noted that the briefing Of Kennedy was restricted to Phase One of the operation, which had ended about a year earlier. Phase Two was already under way at the time of the briefing, but Kennedy was not told of it. As far as we know, the FBI has not been told the sensitive - 130 - operational details, but it would be naive to assume that they have not by now put two and two together and come out with the right answer. They know of CIA's involvement with Roselli and Giancana as a result of the Lag Vegas wiretapping incident. From the Chicago newspaper stories of August 1963, and from Giancana's own statement; it appears that they know this related to Cuba. When Roselli's story reached them (Roselli to * * * * * * * to pearson to Warren to the FBI), all of the pieces should have fallen into place. They should by now have concluded that CIA plotted the assassination of Castro and used U.S. gangster elements in the operation. There is some support for this thesis in The conversation I had with Sam Papich on 3 May 1967 when I told him of the expected meeting. between Roselli and Harvey. Sam commented that Roselli and Giancana have CIA "over a barrel" because of "that operation" He said that he doubted that the FBI would be able to do anything about either Roselli or Giancana because of "their previous activities with your people." Can we plausibly deny that-we plotted with gangster elements to assassinate Castro? No, we cannot. We are reasonably confident that there is nothing in writing outside of the Government that would confirm Pearson's story of the gambling syndicate operation; but there are plenty of non-gangster witnesses who could lend confirmation. a. Maheu can confirm that Shef Edwards told Roselli that Edwards had told the Attorney General of Roselli's activities on behalf of the Government. b. * * * * * * * * * * * * can confirm the pill and three-man team elements of the story. c. * * * * * * can confirm the pill element of Phase One. d. If an independent investigation were to be ordered, the investigators could learn everything that we have learned[. Such an investigation probably would uncover details unknown to us, because it would have access to the non-CIA participants. Can CIA state or imply that it was merely an instrument of policy? Not in this case. While it is true that Phase Two was carried out in an atmosphere of intense Kennedy administration pressure to do something about Castro such is not true of the earlier phase. Phase One was initiated in August 1960 under the Eisenhower administration. Phase two is associated in Harvey's mind with the Executive Action Capability, which reportedly was developed in response to White House urgings. Again; Phase One had been started and abandoned months before the Executive Action Capability appeared on the scene. When Robert Kennedy was briefed on phase One in May 1962; he strongly admonished Houston and Edwards to check with the Attorney General in advance of any future intended use of U.S. criminal elements. This was not done with respect to Phase Two, which was already well under way at the time Kennedy was briefed. The Pearson story, which is now causing us so much distress, includes one detail that is found only in Phase Two: the three-man team. What measures might be taken to lessen the damage? We see little to be gained from personal approaches now to Maheu, * * * * * *, or Roselli. Maheu has much to lose and might be able to prevail upon * * * * * * and Roselli not to spread the story further. It is questionable whether any such urging would be effective with Roselli, because Roselli stands only to gain from having the story of his CIA connection known and accepted. We cannot now suppress the story, because it is already out and may boil up afresh from the Garrison case. If we were to approach any of the participants and urge discretion upon him, and if this became known, it would merely lend credence to a tale that now sounds somewhat improbable.