Lessons Unlearned

The CIA's Internal Probe of the Bay of Pigs Affair

Michael Warner

The Bay of Pigs invasion met its ignominious end on the afternoon of 19 April 1961. Three days after the force of Cuban migrs had hit the beach, the CIA officers who planned the assault gathered around a radio in their Washington war room while the Cuban Brigade's commander transmitted his last signal. He had been pleading all day for supplies and air cover, but nothing could be done for him and his men. Now he could see Fidel Castro's tanks approaching. "I have nothing left to fight with," he shouted. "Am taking to the woods. I can't wait for you." Then the radio went dead, leaving the drained and horrified CIA men holding back nausea. 1

Within days the postmortems began. President Kennedy assigned Gen. Maxwell Taylor to head the main inquiry into the government's handling of the operation. 2 Director of Central Intelligence (DCI) Allen Dulles asked the CIA's Inspector General (IG), Lyman B. Kirkpatrick, Jr., to conduct an internal audit. A humiliated President Kennedy did not wait for either report before cleaning house at CIA. He accepted resignations from both Dulles and Deputy Director for Plans Richard Bissell (although both stayed at their posts until their successors were selected a few months later).

Lyman Kirkpatrick subsequently acknowledged that his Survey of the Cuban Operation had angered the handful of senior Agency officers permitted to read it, particularly in the Directorate for Plans (the Agency's clandestine service and covert action arm, referred to here as the DDP). 3 The IG's Survey elicited a formal rejoinder from the DDP, written by one of Bissell's aides who was closely associated with all phases of the project. These two lengthy briefs, written when the memories and documentation were fresh, were intended to be seen by only a handful of officials within the CIA. They shed light on the ways in which the CIA learned from both success and failure at a milestone in the Cold War.

Lyman B. Kirkpatrick, Jr.

Did Kirkpatrick build a fair case against the Bay of Pigs operation? If he did, what can be inferred about the rejection of his Survey by Dulles, Bissell, and other Agency principals? Historian Piero Gleijeses has noted that the White House and the CIA were like ships passing in the night during the planning for the Bay of Pigs invasion; they assumed they spoke the same language with regard to Cuba, but they actually were imprisoned by mutually exclusive misconceptions about the invasion's likely outcome. The Kennedy administration believed the assault brigade would be able to escape destruction by melting into the countryside to wage guerrilla warfare. According to Gleijeses, CIA officials, from Dulles on down to the branch chief who ran the operation, professed this same belief but tacitly assumed President Kennedy would commit US troops rather than let the Brigade be overrun. 4 A close reading of the IG's Survey and the DDP's response supports Gleijeses's thesis and hints that an analogous misunderstanding within CIA itself hampered planning for the invasion and contributed to the communications breakdown with the White House.

Shooting the Messenger?

The Eisenhower administration and the CIA had decided in late 1959 that Fidel Castro was a tool of Communism and an ally of the Soviet Union. Bissell contended in February 1961 that popular discontent with Castro's regime could be galvanized into active resistance only by an external shock. The spring of 1961 was seen as the last opportunity to administer such a shock (without actually committing US troops) before Castro's military received more shipments of Eastern Bloc weapons. A CIA-trained force of Cuban exiles would seize an isolated area along Cuba's southern coast, allowing migr political leaders to return to the island and offer the populace a democratic alternative to Castro. Assuming the migr force gained control of the air and consolidated its beachhead, the Brigade's aircraft (obsolescent but potent B-26 bombers allegedly purchased on the black market) would then negate the Cuban Army's numerical superiority and demonstrate Castro's impotence to the Cuban people. Over the next few weeks, Cuba's populace and military would finally mount an active resistance to him, setting in motion his eventual downfall. If worst came to worst, however, the Brigade could be evacuated by sea, and elements might be able to "go guerrilla" in the not-too-distant Escambray Mountains. 5 These assumptions proved disastrously mistaken.

Allen Dulles had ordered Kirkpatrick to investigate the failed invasion three days after the Cuban Brigade surrendered. Kirkpatrick subsequently called the events surrounding the Bay of Pigs affair one of the most painful episodes of his long service with CIA. 6 He had been named IG by Dulles in 1953 after being crippled by polio. Although Kirkpatrick was rumored to covet the job of Deputy Director for Plans and to resent his bad luck, there was no doubt about his competence and concern for improving the Agency's functioning. His judgments commanded responsible consideration.

The IG's team of three investigators quickly set to work, reviewing the voluminous documentation and interviewing approximately 125 CIA and military officers associated with the project, codenamed JMATE (originally JMARC). Kirkpatrick himself played an unusually active role in compiling and evaluating records and interviews for the study. After six months of research and drafting, the IG Staff completed its thick report and had it ready for submission to DCI Dulles. 7

At this point, Kirkpatrick made a serious tactical error. He set aside Copy #1 of the Survey for DCI-designate John A. McCone, rather than for Dulles, and gave McCone his copy before he had given copies to Dulles or Bissell. 8Both McCone and Dulles were angered by this breach of protocol. Kirkpatrick's faux pas naturally stimulated gossip about his motives. The IG Survey was critical of the DDP and would not have been enthusiastically received in any event, but the IG's premature presentation of the Survey to McCone had piled insult on injury. Soon after taking office, McCone allowed Bissell to prepare a formal rebuttal to the IG. 9

Bissell's assistant, C. Tracy Barnes, drafted the DDP's response, completing it in January 1962. Barnes was well qualified to present the DDP's case, although hardly an objective observer. One of the Directorate's two Assistant Deputy Directors (Richard Helms was the other), Barnes had set aside his usual duties for a year to concentrate on the Cuban operation. Although he rarely imposed operational direction himself, he often reviewed and approved decisions in Bissell's name. 10 Barnes thus had gained a comprehensive view of (and significant responsibility for) the project, obtaining wide knowledge of its details as well as working with many of the policymakers involved.

Tracy Barnes

Investigation and Controversy

The most notable feature of the IG's Survey of the Bay of Pigs operation is that it says little about the Bay of Pigs invasion per se. Kirkpatrick later insisted that Dulles had ordered him to "stay out of national policy decisions"--that is, to restrict his probe to the performance of the CIA and not to pass judgment on decisions taken by higher authority. 11 Whatever Dulles's orders had been, the Survey stated on its first page that its purpose was "to evaluate selected aspects of the Agency's performance" in the attempt to overthrow Castro, and that those aspects did not include the operation's "purely military phase." The Taylor report had already evaluated the US Government's conduct of the entire operation. Kirkpatrick's Survey did not presume to judge the actions of other departments, let alone those of higher authority, and thus concentrated on the phases of the operation that CIA controlled. Nor did the Survey examine the totality of CIA activities within Cuba or directed against it from abroad; among other things, Kirkpatrick did not examine in depth the functioning of the Havana station or the Santiago base, the development of foreign intelligence assets and liaison contacts, Division D's technical collection programs, or counter-intelligence work against the Cuban services.

The inspectors concluded that the operation's unorthodox command structure ensured that vital information would not be properly disseminated and that decisionmakers would entangle themselves in minutiae. Operational details fell to Branch 4 (Cuba) of the DDP's Western Hemisphere Division (WH), but Jacob Esterline, chief of Branch 4, reported to DDP Bissell and Tracy Barnes rather than to the chief of WH, J.C. King (although King was regularly informed and often consulted). To confuse matters still further, Branch 4 had no direct control over the Brigade's aircraft, which were managed by a separate DDP division that also took some orders directly from Deputy Director of Central Intelligence (DDCI) Charles P. Cabell, a US Air Force general who liked to keep his hand in the planning of airdrops and other missions. These odd command relationships were accompanied by similarly ad hoc arrangements in other phases of the operation. 12

Kirkpatrick's inspectors also criticized Branch 4's mishandling of intelligence on the political resilience and growing military capabilities of the Castro regime. Although the branch already had its own Foreign Intelligence Section, it nonetheless established a separate "G-2" unit subordinate to its Paramilitary Section, which planned the actual invasion. This decision was "a grave error," in the IG's opinion, because it allowed the project's most important analysts to become so engrossed in the invasion planning that their objectivity and judgment suffered. Even worse, there was no one to audit the "G-2's" analyses: Branch 4's Foreign Intelligence Section could not see all the available sources and was not privy to the invasion planning. These circumstances:

undoubtedly had a strong influence on the process by which [Branch 4] arrived at the conclusion that the landing of the strike force could and would trigger an uprising among the Cuban populace. This conclusion, in turn, became an essential element in the decision to proceed with the operation, as it took the place of the original concept, no longer maintainable, that the invasion was to be undertaken in support of existing and effective guerrilla forces. 13

The IG Survey also criticized CIA Headquarters' micromanagement of the Agency effort to bolster the indigenous "guerrilla forces" operating in Cuba in the months before the Bay of Pigs invasion. The CIA's air supply effort accomplished little; the Agency's maritime supply operation looked no better. CIA efforts to train and infiltrate rebel leaders wasted months and produced no appreciable results. The air operation in particular suffered under the personal attentions of DDCI Cabell. In one especially embarrassing foul-up, agents in Cuba requested a drop of not more than 1,500 pounds of weapons and sabotage equipment; thanks to Cabell, they received 1,500 pounds of other unrequested materiel, plus 800 pounds of rice, 800 pounds of beans, and 160 pounds of lard! 14

Senior Agency officials often gave short shrift to the operation in the press of daily business, and more-junior officers working full-time on it had too little authority and no view of the full picture. The project staff was shorthanded from the beginning despite its rapid expansion (the work to be done expanded even more rapidly), and its managers did not insist that DCI Dulles honor his promise to put the CIA's best talent on the effort. Finally, the Agency's plans were left behind by its assumptions and never caught up. The CIA kept building its Cuba project ever bigger as the likelihood of popular resistance to Castro faded in the distance. In the autumn of 1960, Agency officers envisioned a strike force to assist the failing rebellion; by the following spring, it had become clear that there was no more rebellion. The only solution was to create a rebellion by shocking the Cuban people. 15 In the end, the shock was too ephemeral to damage the Castro regime, let alone threaten its survival. But no one with significant authority seemed to understand this dilemma, and no one at the lower levels who grasped it could do much about it.

The IG Survey suggested that the Agency's principals--Bissell in particular--had been derelict in their duty to advise the White House of the growing possibility of disaster. "When the project became known to every newspaper reader, the Agency should have informed higher authority that it was no longer operating within its charter." The DDP [Bissell], "a civilian without military experience, and the DDCI, an Air Force general, did not follow the advice of the project's paramilitary chief, a specialist in amphibious operations," to insist that Kennedy revoke his cancellation of the D-Day airstrike. "And the President made this vital, last-minute decision [to cancel] without direct contact with the military chiefs of the invasion operation." Faced with a choice between "retreat without honor and a gamble between ignominious defeat and dubious victory," states the IG Survey, "the Agency chose to gamble" and accommodated its plans to whatever restrictions were imposed by the White House. 16

The IG Survey ended with a brief set of conclusions and recommendations. Kirkpatrick's team believed the CIA had failed to notice that the project had progressed beyond the Agency's capabilities and responsibility:

The Agency became so wrapped up in the military operation that it failed to appraise the chances of success realistically. Furthermore, it failed to keep the national policymakers adequately and realistically informed of the conditions considered essential for success.

In addition, the Agency had misused some of its Cuban partners, failed to build resistance to Castro "under rather favorable conditions," and neglected crucial information on Castro's strength. 17

Kirkpatrick's team had produced a detailed but flawed appraisal of the Agency's performance in the Bay of Pigs operation. The Survey's rambling argument obscured some of its more important insights. For example, the Survey did not explicitly conclude that the CIA's allegedly bungled effort to foster an anti-Castro insurgency helped ensure that popular resistance to the regime would collapse by early 1961--and that an invasion would be the only option left for Agency planners. The IG Survey also missed other opportunities to strengthen the logic behind its conclusions. Important judgments were scattered almost randomly across a haphazard overall structure, which, combined with the internal disorganization of certain sections, surely left readers wondering how some of the evidence collected by the IG's staff supported the Survey's key judgments. These weaknesses in the Survey gave its opponents easy targets.

Tracy Barnes responded to the survey by attacking its assumption that the invasion was doomed from the start. More clearly written (although no better organized) than the Survey, Barnes's lengthy analysis insisted that JMATE was not given a real chance to succeed. Instead of proving that the plan was irredeemably flawed, Barnes argued, the Survey had busied itself with highlighting trivial mistakes and raising false issues in an effort to show that the Agency alone was responsible for the disaster.

Arguing that defeat on the beach was by no means foreordained, Barnes suggested that any serious inquiry would have looked at what actually happened instead of judging that Castro would have won anyway. Once that questionable hypothesis was set aside, said Barnes, it then became clear that all the problems encountered before the invasion had not mattered much because, despite all these obstacles, the Cuban Brigade had actually been trained and landed. The pre-invasion setbacks had only slowed the Brigade's preparations; they did not diminish its fighting ability. Alleged mistakes by CIA "were not in the actual event responsible for the military failure." The Brigade could not hold its beachhead because its ammunition was lost at sea to Castro's T-33 jets--aircraft that the Agency had planned to destroy but was not allowed to attack at the critical moment. 18 CIA's error was not in mishandling the Brigade but in misperceiving Castro's ability to rally his forces and crush the landing. Barnes argued that Kirkpatrick had missed this point:

It is impossible to say how grave was [CIA's] error of appraisal since the plan that was appraised was modified by elimination of the D-Day airstrike. Had the Cuban Air Force been eliminated, all these estimates might have been accurate instead of underestimated.

Left to right: Allen Dulles, Richard Bissell, President Kennedy, John McCone. April 1962

Turning to the specifics of the IG Survey, Barnes complained that the Survey was little more than a list of niggling and ultimately inconsequential errors committed by the DDP. The organization and staffing of the Bay of Pigs operation had followed standard practices, according to Barnes; arrangements that the IG Survey had criticized had both logic and custom to recommend them, and it was not clear that alternatives would have worked any better. Barnes conceded that the operation's security precautions, logistic procedures, and training efforts fell short of perfection, but he argued nonetheless that they had been done about as well as they could have been.

Barnes's analysis seemed to make a telling case against the IG Survey, exposing every weakness and factual error in the IG's effort. Nevertheless, he had begged as many questions as he answered. His analysis offered almost no concessions to the IG's findings, defending virtually everything done by the DDP--even the infamous "rice and beans" supply drop mentioned earlier. 19 It sometimes seemed as if Barnes was describing a model operation. Ultimately, however, the sheer magnitude of the disaster thwarted Barnes's efforts to shift blame away from the Agency and forced him into the refuge of inconsistency. Barnes seemed to want it both ways. He defended the DDP against charges of unorthodox practices by citing the unique nature of the Cuban operation, in which standard procedures did not always suffice. At the same time, Barnes disputed Kirkpatrick's insinuations of complacency at the top by asserting that the Bay of Pigs operation was an ordinary project in many respects and that the Agency's principals did not need to do much beyond the ordinary call of duty.

The fundamental dispute between Kirkpatrick and Barnes, however, was over the operational plan itself. Was it a good one gone awry (Barnes's view), or a wild gamble that never should have been tried (Kirkpatrick's)? In taking this contrary view, the IG Survey implicitly supported the Taylor commission's speculative judgment that the Cuban Brigade was too small to have maintained its foothold, even with proper air support. CIA planners knew that the 1,500-man Brigade could face as many as 13,500 well-armed troops of Castro's regular army within 12 hours of its landing and would also face several thousand militia troops, albeit of questionable loyalty and fighting prowess. 20 Both Taylor and Kirkpatrick concluded that the Brigade could not have held its 40-mile-wide beachhead--even with air superiority--much longer than it actually did.

The IG Survey's argument and conclusions hinged on the assumption that the Brigade was simply too weak to hold its wide beachhead--a point both obvious and infuriating to Barnes and the DDP. Kirkpatrick had indeed analyzed the Agency's performance apart from the larger context of policy decisions made in Washington on the eve of the invasion. If the invasion had been doomed from the outset, Kirkpatrick implied, then its planners in the Agency should not delude themselves with the excuse that President Kennedy's last-minute cancellation of key airstrikes had wrecked the operation. Kirkpatrick dismissed this alibi, arguing that such logic begged the question of why the project had so little margin for error that it could be spoiled by one hasty decision. The CIA's mishandling of the operation from the beginning had produced "pressures and distortions" and inattention to the developing dangers--leading to grave errors of judgment and finally to disaster. 21

In the end, Kirkpatrick and Barnes were talking past each other. Barnes was correct in saying that CIA could not be judged in isolation from the motivations, anxieties, and misapprehensions affecting policymakers in the White House and other agencies. On the other hand, Kirkpatrick was correct in arguing that CIA should be judged on its mediocre performance in those areas that it ran. Both assertions were true, but they did not fully grasp what had happened at the Bay of Pigs.

A Missing Assumption

Piero Gleijeses's recent analysis suggests a way beyond this impasse. The basic error in the US Government's planning, according to Gleijeses, was the lack of any real effort to outline and assess the consequences that would follow from a failure by the Brigade to hold its lodgment. CIA bears primary responsibility for this omission. The Agency's principals accepted two general assumptions: that Castro was too weak to crush the invaders, and that President Kennedy would land the Marines and finish Castro once and for all if it seemed the Brigade was doomed. Beyond these two certainties, Bissell later explained to Gleijeses, specific planning was pointless because the actual situation on the island would be too fluid as Cuban politicians and Army officers mounted their challenges to Castro:

In most covert operations I know of' [,] particularly those that have a large paramilitary component, the planning for later stages is very incomplete. The outcome of the first stages of the operation is usually so difficult to predict (especially in operations like PBSUCCESS [in Guatemala] and the Bay of Pigs, in which there is very heavy reliance on psychological warfare) that it wouldn't have seemed sensible to have planned the later stages. One can plan the first phases, but not what happens next. 22

This is what indeed had happened in Guatemala in 1954; Headquarters had all but lost hope that the CIA-trained invading force could overthrow the leftist government of Jacobo Arbenz, when suddenly the Guatemalan Army turned on Arbenz, who stepped down and fled. 23 Experience had taught Agency officials to expect a certain amount of chaotic uncertainty after the initial stages of any paramilitary covert action, and not to try to hold events to rigid plans and timetables. There were no such rigidities built into JMATE. "Arms were held in readiness for 30,000 Cubans who were expected to make their way unarmed through the Castro army and wade the swamps to rally to the liberators," noted the IG Survey with a hint of sarcasm. "Except for this, we are unaware of any planning by the Agency or by the US Government for this success." 24

CIA had re-learned one lesson from PBSUCCESS--coups are chaotic--but the Guatemalan operation held another lesson of equal or greater importance. PBSUCCESS succeeded not because the CIA-trained rebels won on the battlefield or frightened Arbenz into fleeing, but rather because the migr invasion of Guatemala, combined with the Guatemalan Army's concern over Arbenz's leftward drift and fear of US military intervention, gave Army leaders a pretext to force Arbenz from power. 25 CIA-orchestrated airstrikes and ground maneuvers had played an indirect role in changing the Army's mood, to be sure, but Agency personnel in Guatemala City itself had initiated the crucial face-to-face meetings that ultimately prodded the Army's indecisive leadership to act, and had met repeatedly with vacillating Guatemalan colonels, insisting that they save themselves and their nation by toppling Arbenz before it was too late. This "K-Program" to influence the Army had proceeded with the support of US Ambassador John Peurifoy. 26 What, then was the second lesson from PBSUCCESS? Very simple: divide and conquer. Get your adversaries fighting among themselves.

JMATE had no "K-Program"--no significant CIA or diplomatic effort to persuade Cuban Army leaders to depose Castro. It is difficult to tell exactly how Havana station was dealing with the Cuban military in 1960 because the station cables have been destroyed. 27 Nevertheless, surviving records from Headquarters, the Havana station, and the Brigade training sites suggest that CIA's principals did not expect the Ambassador, the chief of station, or any American in Havana to influence the Cuban Army. The possibility of turning the Army against Castro looked too remote to consider. An unsigned DDP analysis from February 1960 compared the earlier situation in Guatemala with the contemporary scene in Cuba:

Arbenz, a professional Army officer, had left the armed forces of Guatemala virtually unchanged--and could not rely on them in the crisis; Castro has largely liquidated [deposed Cuban dictator Fulgencio] Batista's armed forces, filled key military posts with his trusted followers, and introduced a system of intense ideological indoctrination. 28

Fidel Castro had drawn his own lesson from the Guatemala operation, and he was determined to leave no opening for the sort of "chaos" that PBSUCCESS had exploited.

CIA's Havana station had little opportunity to persuade Castro's new army in any event. The IG Survey noted that the station reported creditably on political, economic, and Communist Party matters, but found that "its agents in Cuba lacked access to high-level military sources" when Headquarters asked for more military reporting in late 1960. 29 Castro's secret police kept a close watch on station and Embassy personnel, and in October 1960 they caught three Technical Service Division technicians redhanded as they were installing listening devices at the New China News Agency. 30 The slim opening for mounting a "K-Program" in Havana slammed shut in January 1961, when the outgoing Eisenhower administration severed relations and closed the American Embassy. Thus JMATE proceeded without one particular capability that had proved vital to PBSUCCESS.

The possibility of personally persuading Cuban Army officers had been discounted in the earliest days of the operational planning, but CIA had another arrow in its quiver. Bissell probably believed that Castro would be dead at the hands of a CIA-sponsored assassin before the Brigade ever hit the beach. This expectation perhaps kept Bissell and Barnes overoptimistic about JMATE, but project officers themselves were not privy to assassination plotting and thus should have been looking for some way of working within Cuba to influence the loyalty and effectiveness of Castro's military.

They did not have any such plan--a fact made uncomfortably clear in hindsight. Lacking direct contact with Castro's army, project officers by March 1961 had convinced themselves that the mere survival of the Brigade on Cuban soil would suffice to turn much of the military against Fidel. Grasping at straws--and tacitly assuming that they were trying to replicate the dynamic that had operated in Guatemala seven years earlier--the DDP analysis now portrayed Castro's thorough reorientation of Cuba's armed forces as a source of weakness for Castro and strength for the CIA:

It is our estimate that [Castro's] forces, if confronted by a trained opposition element with modern weapons and a unified command, will largely disintegrate. It is significant that most of the leaders of the anti-Castro insurgent groups are Army officers who once fought with Castro against Batista. The Army has been systematically purged, and most of it is now serving in labor battalions or on routine garrison duty. There is great resentment in the Army at this downgrading, the subordination to the Militia, and the imprisonment of such popular leaders as Huber Matos. 31

This estimate was wishful thinking disguised as analysis. The Agency had "no intelligence evidence" that there was anyone in Cuba who "could have furnished internal leadership for an uprising in support of the invasion," noted the IG Survey. 32 JMATE thus coasted along on the tacit assumption that something good would happen within the Cuban Army, once the battle was joined and the migr Brigade demonstrated its staying power. (S)

At least one DDP leader had the experience to have recognized this error and the authority to have acted upon it. Ironically, that man was A/DDP Tracy Barnes, who had commanded the CIA's LINCOLN task force at the climax of PBSUCCESS, and who been Bissell's aide for JMATE. Yet the long apologia for JMATE that Barnes wrote in response to the IG's Survey seemed deaf to the real lesson of PBSUCCESS and the way in which it was unlearned during the planning of the Cuban operation.


The disconnect between what CIA wanted Cuba's Army to do and how the Army would be persuaded to do it was a major flaw in the invasion planning. This defect, in turn, distorted the Agency's advice to President Kennedy. It made Allen Dulles and Richard Bissell overconfident, and thus contributed to the disastrous misunderstanding explored in Piero Gleijeses's recent analysis. CIA officials did not spot this omission before the Bay of Pigs, and the controversy over the IG Survey obscured the lesson and ensured that few Agency principals would understand what had gone wrong.

Forgetting history kept Barnes and Kirkpatrick talking past one another in their respective reviews. Barnes had turned his apologia into an attack on the IG Survey and the Inspector General's motives. The DDP would have served itself and CIA better by drafting a careful analysis of the operation, particularly the way in which the assumptions contained in the JMATE plan evolved on their own without conscious revision and constant comparison with current intelligence and policy directives. Kirkpatrick, for his part, had approved a rambling report and then bungled its presentation to CIA's principals, thus incurring lasting resentments and helping to ensure his report would not be heeded. Neither the IG nor the DDP prepared clear insights that could instruct Agency leaders and planners. More attention to the need to understand the Bay of Pigs invasion might have prevented a generation of CIA officers from believing that one more airstrike would have saved the Brigade.

What difference did history make? Richard Bissell, Tracy Barnes, and the DDP had forgotten one of the crucial lessons of PBSUCCESS. As a result, CIA convinced itself that 1,500 brave and well-trained men--with no help from American diplomats and intelligence officers in Havana--could hold 40 miles of beach against Castro's toughened military long enough to spark a coup or a general uprising. Dulles and Bissell then sold this plan to the White House, apparently believing that the details did not matter much anyhow because Castro would either be assassinated or President Kennedy would send in the Marines to rescue the Brigade. Fidel Castro and his Soviet allies, however, had studied the 1954 events in Guatemala and resolved to avoid Arbenz's mistakes. The result was the surrender on Blue Beach on 19 April 1961, when the lessons of history meant plenty for the men trapped and taken prisoner.


1 The scene at CIA Headquarters is described by David Atlee Phillips, an eyewitness, in The Night Watch: 25 Years of Peculiar Service (New York: Atheneum, 1977), p. 109. The text of the final radio message appears in the sanitized version of the Taylor committee report to President Kennedy; see Luis Aguilar, ed., Operation Zapata: The "Ultrasensitive" Report and Testimony of the Board of Inquiry on the Bay of Pigs (Frederick, MD: University Publications of America, 1981), p. 28.

2 General Taylor's Board of Inquiry comprised himself, Attorney General Robert F. Kennedy, Chief of Naval Operations Arleigh Burke, and DCI Allen Dulles.

3 Lyman B. Kirkpatrick, Jr. The Real CIA (New York: Macmillan, 1968), p. 200. Peter Wyden, Bay of Pigs: The Untold Story (New York: Simon & Schuster, 1979), p. 324.

4 Piero Gleijeses, "Ships in the Night: The CIA, the White House and the Bay of Pigs," Journal of Latin American Studies, 27 (February 1995), pp. 37-42.

5 Unsigned memorandum [probably from Col. Jack Hawkins, Western Hemisphere Division] to Jacob Esterline, Chief, WH/4, "Policy Decisions Required for Conduct of Strike Operations Against Government of Cuba," 4 January 1961, cited in Gleijeses, "Ships in the Night," p. 17. For Bissell's explanation of the need for a "shock," see pp. 10-11.

6 Kirkpatrick, The Real CIA, p. 184.

7 CIA Inspector General, "Inspector General's Survey of the Cuban Operation," October 1961, CIA History Staff files, HS/CSG-2640.

8 Lyman B. Kirkpatrick, IG, to John A. McCone, DCI, "Inspector General Survey of the Cuban Operation (dated October 1961)," 16 February 1962, Executive Registry Job 80B01676R, box 20, folder 1.

9 Richard Bissell, with Jonathan E. Lewis and Frances T. Pudlo, Reflections of a Cold Warrior: From Yalta to the Bay of Pigs (New Haven, CT: Yale University Press, 1996), p. 183.

10 Deputy Director (Plans), "An Analysis of the Cuban Operation," 18 January 1962, section VI, p. 13. Barnes's official title was Assistant Deputy Director (Plans) for Psychological and Paramilitary Action.

11 Kirkpatrick, The Real CIA, p. 200.

12 IG Survey, pp. 36-38.

13 Ibid., p. 78.

14 Ibid., pp. 98-99.

15 Ibid., pp. 51-66.

16 Ibid., pp. 61-63.

17 Ibid., pp. 143-148.

18 DDP Analysis, sec. I, pp. 8-10, 14-15.

19 Ibid., sec. IX, pp. 4-5.

20 The estimate of the size of Castro's forces was provided by Allen Dulles to the Taylor committee; see Aguilar, ed., Operation Zapata, p. 351. The Taylor committee's judgment that the assault brigade was doomed--a judgment from which Dulles and Admiral Arleigh Burke dissented--can be found on p. 29.

21 IG Survey, pp. 34-35.

22 Gleijeses, "Ships in the Night," p. 29.

23 Nicholas Cullather, Operation PBSUCCESS: The United States and Guatemala, 1952-1954 (Washington: Central Intelligence Agency, 1994), p. 75.

24 IG Survey, p. 60.

25 Piero Gleijeses, Shattered Hope: The Guatemalan Revolution and the United States, 1944-1954 (Princeton, NJ: Princeton University Press, 1991), pp. 338-342.

26 Cullather, PBSUCCESS, pp. 64-65, 75-77. See also Western Hemisphere Division, "K-Program," Annex E in "Debriefing Reports, Project PBSUCCESS," no date [1954], Latin America Division Job 79-01025A, box 167, folder 6.

27 DO Latin America Division's inventory of retired records indicates the cables to and from Havana station from 1958 to 1961 were in Job 65-00196R, which was destroyed in 1979. Havana station dispatches survived in LA Division Job 78-02 163R, box 3; a sampling of those from October 1960 showed no contacts with the Cuban Army.

28 Unsigned memorandum, Directorate of Plans, "Covert Action against Cuba," no date [probably 24 or 25 February 1960], Latin America Division Job 85-00106R, box 1, folder 5.

29 IG Survey, p. 75.

30 David R. McLean, "Western Hemisphere Division, 1946-1965," Clandestine Services History Program [CSHP] study number 324, December 1973, CIA History Staff, pp. 233-235. See also Nathan Nielsen, "Our Men in Havana," Studies in Intelligence, Vol. 32 (spring 1988), p. 1.

31 Western Hemisphere Division, "Anti-Castro Resistance in Cuba: Actual and Potential," 16 March 1961, Latin America Division Job 82-00679R, box 3, folder 4.

32 IG Survey, p. 60.

Michael Warner is Deputy Chief of CIA History Staff.