The Technical Services Division, meanwhile, was working on poisons which might be used for the murder, after Giancana had protested that a gangland-style killing would never work. In a separate but related effort in August 1960, the CIA's Office of Medical Services was given a box of Castro's favorite cigars and told to treat them with a lethal poison. They were ready in October, and delivered to someone in the Agency—it is not known to whom—the following February. The cigars may have been intended for Castro during his trip to the United Nations in September 1960. According to David Wise and Thomas B. Ross in The Espionage Establishment, a CIA officer told Michael J. Murphy of the New York Police Department that the Agency had planned to assassinate Castro with a box of exploding cigars, but then had changed its mind. Perhaps the box of cigars referred to by Murphy was the same one the office of Medical Serviced had treated with botulinus toxin by October 7. Perhaps not. In any event, the Technical Services Division prepared botulinus toxin pills in February 1961, tested them successfully on monkeys, and delivered them to Colonel Edwards, who passed them on to Rosselli in Miami. Late that month or early in March, Rosselli told the CIA that the pills had been given to a man in Castro's entourage, but that he had returned them after he lost his job, and with it his access to Castro. A second attempt in April failed when the agent got "cold feet," and after the collapse of the Bay of Pigs invasion, the Maheu-Rosselli operation went into a dormant phase.
Early that fall, however, the Mafia plot to kill Castro was reactivated after Bissell, in a meeting with both Kennedy brothers held in the Cabinet Room, was "chewed out" for "sitting on his ass and not doing anything about getting rid of Castro and the Castro regime." This time Bissell bypassed Colonel Edwards and gave the job to a veteran of clandestine operations, William Harvey. Earlier in 1961, Bissell had asked Harvey to organize a unit within the DDP which might recruit agents to carry out assassinations on call—described with the euphemism "executive action," the very phrase, interestingly, which Allen Dulles later used in his memoirs to describe the "'Murder Inc' branch of the KGB." Harvey organized the group, and on November 16, 1961, he and Bissell discussed the possibility that ZR/RIFLE, the "executive action" group, might be used for killing Castro. Bissell also told Harvey about the Mafia plot, and later Harvey briefed Helms.
In early April 1962, acting on Helms's explicit orders, Harvey asked Colonel Edwards to put him into contact with John Rosselli, and a few days later the two men were introduced in Miami by the man Edwards had assigned as Rosselli's case officer, James O'Connell. Harvey got off on the wrong foot with Rosselli by telling him to break contact on the Castro operation with Robert Maheu and Sam Giancana. Harvey had apparently decided the two men were superfluous and untrustworthy as the result of an episode eighteen months earlier, in October 1960, when the CIA-Mafia plot was first getting under way. At that time, Maheu, as a favor to Giancana, had hired a private detective to tap the Las Vegas phone of one of Giancana's girlfriends in order to discover if she was being unfaithful to him. The tap was discovered by a maid, the detective was arrested by local police, and Maheu was told to square it or else.
Later, in April 1961, with the permission of Colonel Edwards, Maheu told the FBI that the tap was connected to an operation he had undertaken for the CIA, and Edwards confirmed his story. The problem refused to go away, however, and the following year, in 1962, the Las Vegas wiretap episode helped the FBI to learn the rough outlines of the plot to kill Castro. This all struck Harvey as a perfect example of an operation going out of control, and he decided that the first step was to get rid of the clowns, Maheu and Giancana. Rosselli did as Harvey asked, and the two men met again in New York on April 8, 1962. Before the end of he month, Harvey delivered four poison pills to Rosselli in Miami. In May, Rosselli reported that the pills were inside Cuba, and later, in June, that a three-man team had been sent in to kill Castro.
But that was as far as things went. By September 1962, when Rosselli told Harvey another three-man team was to be sent to Cuba, Harvey had concluded that the operation was going nowhere. He had run the operation with extreme security; not even the men who worked for him on Task Force W (the CIA's end of Operation Mongoose) knew what he was up to, or where he was going when he disappeared for a few days every month or two. Bissell had given him the Rosselli operation, Helms told him to give it a shot, Harvey decided on his own it was a will-o'-the-wisp. In February 1963, Harvey told Rosselli the operation was over.
Harvey's replacement by Desmond FitzGerald and the scuttling of the Rosselli operation did not end but only redirected the CIA's attempts to kill Castro. One of FitzGerald's early inspirations was fanciful and impractical, appealing to his temperamental fondness for the clever and the ingenious. It called for the Technical Services Division to rig an exploding seashell, which would be placed on the sea floor in an area where Castro liked to go skin diving. Like many CIA people, in love with the subtle and the artful, FitzGerald was fascinated by gadgets and resented skeptics who dourly suggested they would cost too much or would fail to work or weren't even needed at all. He was downright petulant at times. When Sam Halpern once protested that a fancy new communications device just wasn't going to work, FitzGerald said, "If you don't like it, you don't have to come to meetings anymore."
Halpern protested that the seashell plan was inherently impossible to control. How could they be sure that Castro would be the one to find it? Besides, the best assassinations do not appear to be assassinations at all, while Castro blowing up on the ocean floor would point a finger directly at the United States. Similar protests had been made about the plan to give Castro a box of poisoned cigars. He might hand them all out to a delegation of visiting schoolteachers. If the idea was to kill Castro, they had to find something which would get him and no one else. FitzGerald's ideas weren't turning out any better than the earlier ones, such as the proposal to provide Castro with a poisoned wet suit to be delivered by James B. Donovan, an American lawyer negotiating the release of the Bay of Pigs prisoners. The Technical Services Division had duly purchased a suit and contaminated the breathing apparatus with tubercle bacilli and the suit itself with fungus spores which would cause a chronic skin disease called Madura foot. Critics of this plan claimed that its authors had neglected the most elementary points: for example, the fact that it was in effect a gift from the United States (the idea was to keep it secret), or Donovan's feeling about being the gift-giver in this plot. If he didn't know, after all, he might try on the suit himself. As it happened, Donovan gave Castro a wet suit entirely on his own, and the CIA's wet suit was destroyed.
But FitzGerald did not abandon the problem. Eventually he came up with a serious effort to use a major in the Cuban army, in contact with the CIA since 1961, named Rolando Cubela. Cubela was on intimate terms with Castro, and often saw and talked to him in his office or at official functions. He and some of his friends bitterly resented the Russian presence in Cuba and felt that Castro had betrayed the revolution. From the CIA's point of view, he was an ideal conspirator, a man with a public reputation as a leader in the fight against Batista, close to Castro, spokesman for a circle of dissidents, and ambitious. On top of that, Cubela had already proved himself as an assassin. In October 1956, he shot and killed the chief of Batista's military intelligence, Blanco Rico, a deed which haunted him thereafter and even resulted in a nervous breakdown. Rico had been picked as a target not because he was ruthless or cruel, but because he was a fair, temperate man; he reflected credit on Batista as a leader. Cubela was convinced that Rico knew why he was being killed, and believed that Rico had smiled at him at the very moment Cubela pulled the trigger.
The CIA was well aware of Cubela's political and mental history, but decided to use him anyway, since he was perfectly situated to engineer the one thing which might actually get rid of Castro—a palace coup. From the beginning, Cubela insisted that a coup had to include Castro's "execution." The word "assassination" disturbed him; he preferred to say he would "eliminate" Castro. At various times he asked the CIA to provide him with exotic assassination devices and more mundane sniper rifles, and the CIA undertook to give him what he wanted. With Helms's approval, FitzGerald personally met with Cubela in Paris on October 29, 1963, despite protests from subordinates who said that no high CIA official should expose himself in such a manner. Cubela had requested a meeting with Robert Kennedy, but FitzGerald satisfied him with the claim that he was Kennedy's personal representative. Not quite a month later, on November 22, 1963, Cubela's case officer gave him a specially prepared "pen" which might indetectably inject a deadly poison into Castro; the CIA recommended Blackleaf 40, a widely available toxin which Cubela was to procure on his own. The Church Committee's assassination report says that Cubela dismissed the poison pen as a toy and insisted the CIA could surely come up with something "more sophisticated."
At the end of the meeting, the CIA case officer learned that Kennedy had just been shot in Dallas. During the ensuing tension and uncertainty, the Cubela plot was allowed to lapse for a matter of months. Christmas, 1963, came and went; nothing happened. Early in 1964, the CIA-Cubela plotting was revived, and two caches of arms—one in March, the second in June—were landed in Cuba for Cubela's use. That fall Cubela requested a sniper rifle, and the CIA told him the United States no longer wanted to have any role in the "first part" of his plan—that is, in Castro's assassination. Why did the CIA change its mind at this late date? The record provides no persuasive reason, but it may have been because Lyndon Johnson was quietly sounded out—so quietly that he may not have known that he was being asked—and he wanted no part of assassination. Clearly, Johnson had not known about the earlier Mafia plots, and Helms did not tell him about the CIA's relationship with Cubela during Johnson's own tenure as President. The important point here is that the CIA's direct involvement in Cubela's assassination plans came to an end at a time when they seemed not to have the President's sanction.
The subject of assassinations was a painful one for CIA people. On no other subject did they fight so hard to keep the secrets, and in particular the secret of presidential authority. On this point the testimony of high-level CIA officials before the Church Committee was elusive in the extreme. Helms in particular remembered next to nothing, and dismissed the rest. He never believed the Mafia plot was going anywhere. He let Harvey proceed only to see if Rosselli really had assets in Cuba. Cubela's plan to "eliminate" Castro was indulged to see if he and his associates could really put together anything in the nature of an honest plot. The committee had obtained the CIA inspector general's report of 1967, but the memories of those involved halted pretty much where the documents came to an end. The Church Committee's report was detailed and lawyerly, proceeding point by point in a logical and yet a confusing manner; discussions of closely related events are sometimes scores of pages apart.
But even when one has reassembled the story in its proper order, the picture one gets is fragmentary, occasionally vivid and complete on minor points, more often bald and out of focus. The primary reason for this is the tendency of CIA officials to suffer memory lapses on all those points, which were very numerous, that had not survived in the files. In addition, of course, Eisenhower, both Kennedys, both Dulleses, General Cabell, and other high officials had died. Livingston Merchant and Admiral Arleigh Burke were too ill to testify. Some of the lower-level officials—William Harvey, Justin O'Donnell, Sidney Gottlieb, and others—testified at length but did not really know who gave the orders or when, and would not have presumed to ask.
The idea of assassination itself did not seem to trouble CIA officials who testified. The wisdom of the undertaking was something else again. It was stupid, foolish, ridiculous, unworkable; worse than a crime, a blunder—the regular spiel. Everyone had his own adjective, none of them flattering. The best they could muster by way of justification was "the climate of the time," the Kennedys' hysteria on the subject of Castro, the eager willingness of the Cubans who were recruited seriatim to do the job. But all the same, they shook their heads in dismay. More than anything else, it seemed to be the sheer difficulty of assassination—that is, of a genuinely secret assassination—that left them wondering.
But on the question of presidential authority, there is no such equanimity. One exception said that no one in the CIA doubted for a minute that Eisenhower and Kennedy "jolly well knew," but others, more closely involved, did more than simply squirm in their chairs. Several different men, in fact, showed dramatic signs of psychological stress in discussing this point.
It is inconceivable that Richard Helms would ever betray himself in so unmistakable a manner. But in his testimony before the Church Committee, Helms more than once revealed an uncharacteristic degree of irritation with the committee's insistent return to the question of authority. He was being as clear as he could: the Kennedys wanted Castro out of there, the CIA did not go off on its own in these matters, the Agency was only trying to do its job. What more could he say? Senator, how can you be so goddamned dumb? This isn't the kind of thing you put in writing.
And despite the Church Committee's diligent search, they never did find anything in writing. The committee did learn, however, of three separate occasions when one or both Kennedys discussed the assassination of Castro in a manner indicating that it lay heavily on their minds. The first occasion occurred in March or early April of 1961, just before the Bay of Pigs invasion, at the height of the first Rosselli effort to poison Castro, when President Kennedy asked his friend Senator George Smathers what he thought the Latin American reaction would be to the assassination of Fidel Castro. Smathers said he told Kennedy the murder would be blamed on the United States and on Kennedy personally, and that he, Smathers, was therefore against it. According to Smathers, Kennedy immediately responded that he was against it too. But in mid-March 1961, before their conversation, the CIA had already given botulinus toxin pills to Rosselli in Miami, and a second batch were to be handed over on April 21.