The Track of Overthrow From Within
In fact, though not all of the Kennedys’ opponents knew it, the accommodation track was not the only one being explored by the Kennedy brothers. On March 14, Robert had sent his brother a memo urging a combined program to stop Cuban subversion abroad and to appeal within Cuba to elements of the Cuban military:
John McCone spoke at the meeting today about revolt amongst the Cuban military. He described the possibilities in rather optimistic terms. . . .Do we have evidence of any break amongst the top Cuban leaders and if so, is the CIA or USIA attempting to cultivate that feeling? I would not like it said a year from now that we could have had this internal breakup in Cuba but we just did not set the stage for it.21
The Bundy memo of April 21 envisaged a total of three possible options: a) forcing "a non-Communist solution in Cuba by all necessary means," b) insisting on "major but limited ends," c) moving "in the direction of a gradual development of some form of accommodation with Castro."22
There are abundant indications in the newly released CIA documents that the CIA, along with other agencies, became pan of a new U.S. strategy aimed at promoting revolt from within Cuba, particularly among the Cuban military. This inter-agency plan was called AMTRUNK inside the CIA, and "Operation Leonardo" by its original authors, George Volsky of USIA, the Cuban exile Nestor Moreno, and Tad Szulc of the New York Times. Szulc, who had excellent connections inside the Kennedy White House, presented the plan to the State Department Cuban Coordinator, Robert Hurwitch in early 1963, when the State Department and the White House pressured the CIA "to consider a proposal for an on-island operation to split the Castro regime."23
The CIA’s own documents make it abundantly clear how distasteful this White House-backed plan was to them. Old disagreements from the Bay of Pigs operation were revived: the White House preference was to use participants in the original Castro revolution, notably men close to Manolo Ray and Huber Matos; and such men were anathema to the more right-wing Cubans who had defected earlier and been championed by the CIA. By 1963 Ray and Nestor Moreno, both close to Szulc, had formed the anti-CIA and anti-Castro group JURE, which not only rejected CIA influence but was suspected by CIA of trying to penetrate its JMWAVE operations. The links of Moreno and Volsky to JURE became key arguments in the CIA’s case for disliking AMTRUNK.24
By April 5, 1963, JMWAVE Station Chief Theodore Shackley was ready to recommend that the whole AMTRUNK operation "be terminated at the earliest possible moment:"
The AMTRUNKers admit to being anti-KUBARK [CIA] and to be working "with" KUBARK now only because there was no alternative if they were to accomplish their mission. . . .[Redacted, a key AMTRUNK member] believes he is receiving special attention because of his [Washington] connections, and he will not hesitate to go behind KUBARK’s back to AMTRUNK-1 [Volsky]. . . or higher authority, if the operation or KUBARK handling of AMTRUNK does not progress to his liking.25
This recommendation to terminate was supported at Headquarters, whose return cable to JMWAVE on April 10 "concurred that the AMTRUNK operation should be terminated for a number of reasons, including the fact that CIA could not at that time be certain that hostile elements [these, in CIA’s view, included Volsky and Szulc] were unaware of the plan."26
Nevertheless, after the decision recorded in the April 21 Bundy memo, the CIA continued to support the AMTRUNK operation until March 1964.27 In the Johnson era, however, the purpose of AMTRUNK appears to have changed completely. Instead of infiltrating agents to woo Cuban military leaders, AMTRUNK operations in early 1964 had become the depositing inside Cuba of Belgian FAL rifles for the assassination of Castro.28 Along with this change in AMTRUNK’s purpose, the CIA JMWAVE Station terminated the involvement of Nestor Moreno, the plan’s original author "in the sensitive aspects of AMTRUNK in November 1963."29 AMTRUNK in other words was by this time subordinated to the Cubela/AMLASH operation, which had become similarly diverted from politics to assassination (see Chapter VI).
The CIA’s continued support of AMTRUNK appears to have been unwilling; and Headquarters soon implemented Shackley’s alternative recommendation of giving AMTRUNKers cash to mount their own independent operations.30 In June the Standing Group approved a sabotage program of raids by exiles, "to nourish a spirit of resistance and disaffection which could lead to significant defections and other byproducts of unrest."31 It was hoped that the pressures on the economy would contribute to "internal discontent that would take appropriate political and military forms."32 This "track two" concept of "autonomous operations," as distinguished to the "track one" of CIA’s support of Artime, was proposed by Walt Rostow of the State Department (a political ally of Lyndon Johnson). A principal beneficiary proved to be JURE, the group which CIA suspected of being behind AMTRUNK.33 Because "track two" supplied resources to JURE for military operations, it had the effect of de-emphasizing the political objectives of the original Plan Leonardo.
Both the plans for an internal military-based coup and the supporting infiltration and sabotage missions were hereafter given the CIA code name AMTRUNK. The renewed CIA sabotage operations became operational in August 1963. As part of this program, a new exile group, with U.S. Army training and advisers, launched raids on August 18 and October 21 as "Comandos Mambises," from the CIA ship "Rex," a former subchaser.34
Rolando Cubela, himself an Army Major, was by CIA accounts approached in 1963 because of his contacts in the Cuban military. His case officers were also part of an operation (which can only be AMTRUNK)
to penetrate the Cuban military to encourage either defections or an attempt to produce information from dissidents, or perhaps even to forming a group which would be capable of replacing the then government in Cuba.35
As mentioned above, in 1964 AMTRUNK teams were used by the CIA to supply assassination rifles with long-distance scopes to Cubela (AMLASH).36
The CIA’s redirection of AMTRUNK exemplified their long-term disagreement with the Kennedy White House over policy objectives. Arthur Schlesinger has argued that, since 1961:
The CIA wished to organize Castro’s overthrow from outside Cuba, as against the White House, the Attorney General’s office and State who wished to support an anti-Castro movement inside Cuba. The CIA’s idea was to fight a war; the others hoped to promote a revolution. Any successful anti-Castro movement inside Cuba would have to draw on disenchanted Castroites and aim to rescue the revolution from the Communists. This approach, stigmatized as Fidelismo sin Fidel, was opposed by businessmen, both Cuban and American, who dreamed of the restoration of nationalized properties. But the CIA alternative was probably dictated less by business interests than by the agency’s preference for operations it could completely control—especially strong in this case because of the Cuban reputation for total inability to keep anything secret.37
To this preference for control can be added another one. The CIA, despite its fiasco at the Bay of Pigs, was still hoping to reassert itself as the preferred agency for paramilitary operations, which had accounted for the biggest item in its annual budget. In this respect AMTRUNK, an inter-agency operation, was not one to its liking: for by all accounts the key co-ordinating role was given, not to the CIA, but to the Department of the Army under Cyrus Vance and his aides Joseph Califano and Alexander Haig.38
Given the norma] CIA penchant for secrecy, it is the more remarkable that the CIA, at the Brazil meeting in September, took the suspected blabbermouth Cubela into its AMTRUNK planning. According to the CIA’s IG Report of 1967,
Cubela discussed a group of Cuban military officers known to him, and possible ways of approaching them. The problem was, he explained, that although many of them were anti-Communist, they were either loyal to Fidel or so afraid of him that they were reluctant to discuss any conspiracies for fear they might be provocations. Cubela said that he thought highly of [redacted, apparently Major Ramon Guin Díaz] (AMTRUNK-[short redaction]) who was hiding [redacted, identified by the Cubans as the infiltrated CIA agent "Miguel Día
z"]. ["Díaz"] had been sent to Cuba to recruit [Guin] in place, and had done so. Cubela said he planned to use [Guin] but was concerned about [Guin’s] "nervous condition" and the fact that he drank heavily. Cubela was told to assist [Guin] in [Guin’s] intelligence assignments, but not to help [Guin] leave Cuba—as Cubela proposed 39
According to a later memo from Helms to Rusk, Ramon Guin "was recruited by a CIA agent in August 1963 inside Cuba as a Principal Agent to recruit high-level military leaders."40 By all accounts the October 29 meeting of FitzGerald with Cubela continued to focus on what Richard Helms, the senior CIA official cognizant of the AMLASH meetings, later called in testimony "the political action pan of it. . . have a group to replace Castro."41
Excluding the CIA: The Secret Attwood Initiative
Robert Kennedy’s penchant for pro-active operations, even if rationalized as a "stick" to encourage Castro to behave reasonably, was clearly unhelpful to unblocking the accommodation track. Sabotage missions in particular had been denounced in September, not only by Castro, but also the Soviet Union 42
Nevertheless the accommodations track, even if interrupted from time to time, seems never to have died under Kennedy. On June 3 the Special Group agreed that it would be a "useful endeavor" to explore "various possibilities of establishing channels of communication to Castro."43
Shortly afterwards a public suggestion by Castro that Cuba might consider normalization of relations was rebuffed by John Kennedy at a press conference. The President attacked Cuba as a Soviet satellite. It is possible however that another cause for concern was the fear of some experts, apparently unfounded at this time, that Castro might be tilting towards Beijing in the increasingly evident Sino-Soviet split.44
Despite this public rebuff, in September the President approved secret contacts at the UN in New York between a Special Advisor to the U.S. Delegation, William Attwood, and the Cuban Ambassador to the U.N., Carlos Lechuga. On September 5 Lisa Howard told Attwood she was convinced that Castro wanted to restore communications with the United States, and she offered to arrange a social gathering in her apartment so that Attwood could meet informally with Lechuga. (It is not clear if Howard was simply reacting to her Castro interview, or whether the Cubans had proposed talks on September 5. as suggested by the Schweiker-Hart Report.)45 (Note that September 5 was two days before the CIA resumed contact with Cubela in Brazil; Attwood comments laconically that "the CIA must have had an inkling of what was happening from phone taps and surveillance of Lechuga.")46
A week later Attwood went to Washington and saw Harriman, a man with whom he had traveled to India in 1959. Harriman was interested in the proposed approach to Lechuga; and he requested a memo which Attwood submitted to him on September 19. Attwood’s memo transmitted information from Guinea’s U.N. Ambassador that Castro was unhappy about his dependence on the Soviet Union "and would go to some length to obtain normalization of relations" with the U.S. It proposed a discreet inquiry to achieve three objectives: "a. The evacuation of all Soviet bloc military personnel, b. An end to subversive activities by Cuba in Latin America, c. Adoption by Cuba of a policy of non-alignment."47 The President gave his approval via Ambassador Adlai Stevenson at the U.N., but it was understood that Attwood would report directly to McGeorge Bundy in the White House. The CIA and the State Department were to be excluded. (Stevenson’s response to Attwood’s memo was that "Unfortunately the CIA is still in charge of Cuba.")48
In addition to knowing Harriman, Attwood had interviewed Castro in 1959 as an editor of Look magazine.49 On becoming Kennedy’s Ambassador to Guinea, he was exposed to the neutralist initiatives of Guinea’s President Sekou Toure and Ghana’s President Kwame Nkrumah, both of whom were on good terms with Castro. Attwood monitored Cuba as an Advisor to the U.S. Delegation at the 1962 Session of the UN General Assembly.50 It was the Ghanaian Ambassador to the UN who in March of 1963 had obtained a Cuban visa for Lisa Howard; and it was the Guinean Ambassador to Cuba who in September told Attwood that Castro, dissatisfied with his Soviet relationship, was looking for a way to escape 51
The first meeting between Attwood and Lechuga took place on September 23, 1963, at a cocktail party hosted for this very purpose by Lisa Howard.52 (Note that this meeting occurs just four days before Oswald, in Mexico City, is supposed to have made contact with Silvia Durán, whom the CIA had reported in early 1963 to be Carlos Lechuga’s mistress.)53 The meeting was productive, and produced a series of informal contacts broken only by Kennedy’s death on November 22.
Attwood saw Robert Kennedy the day after his rendezvous with Lechuga. Robert told Attwood that a Havana visit would be too risky. It was bound to leak. . . .But the general idea was worth pursuing. He told Attwood to stay in touch with Bundy and his staff man on Cuban affairs, Gordon Chase. The Attorney General consulted his brother, who declared himself willing to normalize relations if Castro ended the Soviet bloc military presence on his island, broke ties with the Cuban Communists, and stopped the subversion of Latin America.54
Robert Kennedy proposed that direct U.S. contacts with a special Castro emissary, as proposed by Attwood. should take place at a neutral site in Mexico, with Lisa Howard serving as a go-between.55 We do not yet know if Thomas Mann, the U.S. Ambassador in Mexico, or Win Scott, the CIA Station Chief, were in any way consulted about, or alerted to, the projected meeting.
UN Ambassador Adlai Stevenson contributed to Attwood’s initiative with a speech suggesting "that if Castro wanted peace with his neighbors, he could have it if he stopped trying to subvert other nations, stopped taking orders from Moscow and started carrying out the original democratic pledges of his revolution.56
The Conflict Between the AMLASH and Attwood Initiatives
On October 24, at Attwood’s urging, the President saw the French journalist Jean Daniel, who was about to interview Castro in Havana (Note that this is just five days before the meeting with AMLASH in which FitzGerald presented himself, falsely, as a representative of Robert Kennedy.)
The President is not known to have mentioned the problem of the Cuban Communists to Daniel, but complained that Castro had "agreed to be a Soviet agent in Latin America." "The continuation of the blockade,’ Kennedy said, ‘depends on the continuation of subversive activities.’ Then: ‘Come and see me on your return from Cuba. Castro’s reactions interest me."‘57 Daniel went on to wait three frustrating weeks in Havana before seeing Castro.
On October 11, and again six days later, Cubela in Europe had asked to meet a high-level U.S. government official, "preferably Robert F. Kennedy," for "assurances that the U.S. Government would support him if his enterprise were successful."58 On October 29, five days after the President’s meeting with Daniel, Desmond FitzGerald met with Cubela in Paris, using the AMLASH case officer Nestor Sanchez as an interpreter.59 According to the CIA’s I.G. Report, the contact plan for the meeting, a copy of which was in the AMLASH file, had this to say on its cover: "Fitzgerald will present self as personal representative of Robert F. Kennedy who traveled Paris for specific purpose meeting Cubela and giving him assurances of full U.S. support if there is change of the present government in Cuba." FitzGerald claimed he discussed the planned meeting with the DD/P (Helms) who decided it was not necessary to seek approval from Robert Kennedy for FitzGerald to speak in his name.60 Helms, for whom the I.G. Report was prepared, later confirmed that he had not consulted the Attorney General.61
Sanchez’ report of the meeting does not mention assassination. It says that FitzGerald told Cubela U.S. support "will be forthcoming only after a real coup has been effected and the group involved is in a position to request U.S. . . . recognition and support."62 Nevertheless both FitzGerald and Cubela agree that assassination was discussed. FitzGerald recalled that Cubela wanted "a high-powered rifle with telescopic sights."63 Cubela, conversely, told his interviewer Tony Summers that "it was the CIA who brought up the idea of assassination in the first place—and he who resisted."64
Even if assassination was not the purpose, th
is meeting between a high-level CIA official and a known assassin was extraordinary, perhaps unprecedented. Normally the CIA uses covers and (when assassins are involved) intermediaries or cut-outs. In the well-studied case of the Giancana-Roselli-CIA plots against Castro, the CIA even used one cut-out (Maheu) to contact another (Giancana). Cubela’s inability to keep a secret had become known to the CIA a year earlier; and two CIA officials (Shacklev and Langosch) later testified that they had warned FitzGerald against this meeting.65 Their fears were well-grounded. Earlier that same month the FBI had learned of the renewed CIA-Cubela contact (in a report that was not transmitted to the CIA).66
There is perhaps one other case where the CIA in 1963 prepared to abandon its normal guidelines of plausible deniability, and it too raises questions of the CIA’s loyalty to the Kennedys. In 1962 Robert Kennedy’s representative James Donovan, a New York attorney, along with John Nolan of Kennedy’s staff, had negotiated the release of the Bay of Pigs prisoners. In April 1963 Donovan and Nolan returned to Cuba, to conclude their negotiations with Castro personally. Their mission concerned a few prisoners, including some CIA men, who remained to be released. But the occasion led predictably to the possibility of normalizing the relations between the two countries. Arthur Schlesinger links the success of the Donovan-Nolan mission to the important interview given by Castro to Lisa Howard in late April.67
Desmond FitzGerald of the SAS staff does not appear to have looked favorably towards this step on the accommodation track. In early 1963 the staff arranged for the CIA’s Technical Services Division to purchase a wet suit, and contaminate it with tuberculosis bacilli and the spores for a disabling skin disease. The plan was for Donovan (who was not informed of the plot) to give the suit to Castro, his companion in scuba diving 68
Oswald, Mexico, and Deep Politics Page 11