A coup. A backstop.

Chapter 01 / 1961–1963

A Brazilian crisis, a hemispheric calculation

João Goulart reached the presidency through a constitutional succession contested from the outset. When Jânio Quadros resigned in 1961, military opponents tried to prevent Vice President Goulart from taking office. A compromise reduced presidential power under a parliamentary arrangement; a January 1963 plebiscite restored presidential government. That history made military intervention a lived possibility. It also meant that the armed forces were divided institutions whose officers could claim competing obligations to discipline, constitutional succession and national security.

Goulart, widely known as Jango, sought structural reforms in an unequal society. Land reform, labor mobilization, economic nationalism and the political role of enlisted servicemen collided with the interests and fears of landowners, business groups, conservative politicians and much of the officer corps. Inflation and economic instability aggravated the conflict. His supporters were not one Communist organization; his opponents were not one American-controlled instrument. Civilian institutions, unions, students, churches, governors and military commands all helped create the crisis in which an outside power intervened.

Washington’s concern combined Cold War ideology with economic and strategic interests. Brazil’s size made its direction unusually consequential. U.S. officials worried about Communist influence, foreign-policy independence and conflict over American-owned businesses. They could seek accommodation, use economic leverage, cultivate opponents or prepare to support a forcible change. These were connected possibilities, but their adoption changed over time. A December 1962 options paper retained overthrow for consideration while recommending that the immediate effort change Goulart’s policies. Kennedy accepted that recommendation and sent a special representative. Discussion of removal was already real; it was not yet the March 1964 naval deployment.

A CIA assessment of plotting in March 1963 shows that Brazilian officers were discussing action and that Washington was following them. An embassy cover memorandum from November described a contingency plan emphasizing armed intervention. These records place Brother Sam within a longer interventionist history. They do not establish that every Brazilian conspiracy originated at CIA headquarters, or that every contemplated measure was executed. The distinction becomes essential when the last days are reconstructed: preparatory influence, intelligence access, a decision to move ships and the conduct of a coup are different links in the chain.

Sources: S16 · S17 · S18 · S19 · S25 · S28

Chapter 02 / March 1964

The ambassador asks for a backstop

During March, political demonstrations and military indiscipline narrowed the space for compromise. Goulart’s reform rally on 13 March was followed by a large opposition march in São Paulo on 19 March. The sailors’ rebellion and the government’s handling of it intensified officers’ fear that hierarchy itself was being dismantled. Opposition governors offered territorial bases and political cover. Within the armed forces, some officers had long wanted a coup; others moved against the president as the crisis unfolded. Their motives overlapped without becoming identical.

Ambassador Lincoln Gordon interpreted the events in increasingly catastrophic terms. In the cable FRUS dates 28 March, received in Washington on 27 March, he argued that Goulart was moving toward dictatorial power in collaboration with Communists. That is strong evidence of what Gordon believed and wanted senior officials to act upon. It does not independently authenticate an imminent Communist seizure. His own account still acknowledged the possibility that Goulart could finish his term through constitutional elections, though he considered that increasingly unlikely.

Gordon recommended strengthening forces prepared to resist Goulart. His requests included preparations for covert supply, petroleum and a possible naval presence. Washington’s first reaction was selective. Officials questioned why a small submarine arms delivery would help an army that already possessed weapons, and why a carrier force was immediately appropriate. They took petroleum supply more seriously. The March 28 meeting memorandum records an administration willing to aid Goulart’s opponents while disputing the usefulness, timing and visibility of particular instruments. It is an interventionist discussion with unresolved operational choices.

Gordon also described measures already under way, beyond a future military contingency. Paragraph 16 of his appraisal lists covert support for opposition street rallies, discreet messages of U.S. concern and encouragement of anti-Communist sentiment in Congress, the armed forces, labor, student, church and business circles. That contemporary description establishes active political intervention. It does not supply a complete CIA budget, recipient list or accounting of effects, and the crowds’ existence alone cannot measure their dependence on American help.

On 29 March Gordon explained that unidentified arms could aid paramilitary formations or friendly military units, and that their availability might stiffen resolve. He also distinguished military attaché Vernon Walters’s intelligence contacts from operational contacts whose identity remains redacted. This matters for agency attribution. Walters, a U.S. Army officer with extensive Brazilian contacts, was not simply another name for the CIA. CIA personnel participated in policy and intelligence work, and Gordon described covert activities; the later Brother Sam task force nevertheless ran through the military chain. A coherent account follows these relationships without collapsing them into an omnipotent single actor.

Sources: S01 · S02 · S03 · S15 · S20 · S25

Chapter 03 / 29–31 March 1964

A signal expected; a march begins

Walters’s report of 30 March conveyed a conversation with General José Pinheiro de Ulhoa Cintra. Conspirators intended to act that week, with a signal to follow. The attaché expected advance knowledge and promised to report. The document establishes unusually close access to people organizing removal of the president. It also preserves uncertainty: a decision about the day was still pending, and the report distinguished a confident source’s prediction from an event already accomplished.

The move that began the overthrow came from General Olímpio Mourão Filho in Minas Gerais. On 31 March his troops started toward Rio de Janeiro, supported politically by Governor Magalhães Pinto. Brazilian historical accounts describe multiple centers of plotting and limited coordination. Mourão’s initiative forced other commanders to decide under pressure. A history that makes the entire episode the smooth execution of one master schedule has difficulty explaining the hesitation, personal calculations and hurried shifts in allegiance that followed.

Secretary of State Dean Rusk had sent Gordon instructions the previous night. The United States wanted opponents of Goulart to secure a claim to legitimacy through congressional action, governors’ declarations or other evidence of constitutional justification. Rusk worried that Goulart might retreat enough to blunt military action. The concern was not simply whether an illegal overthrow should be avoided: officials sought a politically usable basis for helping the forces they preferred. Their private language reveals how constitutional claims could become instruments of intervention.

General Amaury Kruel’s Second Army in São Paulo became a crucial part of the balance. Washington’s reports treated its movement as a central unresolved question, while civilian governors and politicians maneuvered over what government might follow. The existence of U.S. support preparations could matter to participants’ expectations. Yet determining exactly who received what assurance, and whether it changed an individual decision, requires a communication trail. Walters’s access and Gordon’s advocacy are established without assuming a guarantee reached every wavering officer. The coup’s immediate success depended on Brazilian commands abandoning or declining to defend Goulart, and on loyalists failing to organize an effective response.

Sources: S04 · S06 · S15 · S21 · S25 · S28

Chapter 04 / 31 March–1 April 1964

Brother Sam moves from contingency to orders

On 31 March Washington began converting preparations into movement. A State Department telegram listed three forms of assistance: petroleum tankers from Aruba, a naval task force, and an assembled ammunition shipment for possible airlift. The Joint Chiefs assigned the code name Brother Sam. The operation belonged to a U.S. diplomatic and military support effort for anti-Goulart forces. It was not a CIA amphibious invasion, even though CIA officials were involved in the wider policy and covert-action setting.

The fine print changes the story. Corrections to the initial cable placed the first tanker off Santos on 13 April. The naval force was to include a carrier and two guided-missile destroyers, followed by four destroyers and tankers. The ammunition package was about 110 tons, with other equipment including CS agent. The corrected aircraft figure was six cargo planes, alongside tanker and fighter support. These were preparations and planned capabilities. Unloading fuel and dispatching the ammunition airlift still required a further political decision and a request from a group Washington could regard as legitimate.

In his telephone report to Johnson that afternoon, Under Secretary George Ball explicitly distinguished ships starting south, tankers being loaded, and ammunition waiting for a decision. Johnson urged his officials to prepare and do what was needed. The call is direct evidence of presidential willingness to back Goulart’s opponents. It is stronger than an innocent-observer account permits. It is narrower than a claim that Johnson ordered American troops into combat that day. The original audio allows the sequence and tone to be heard alongside the written transcript.

A 1 April White House meeting record then moves the naval component beyond a paper plan: Defense Secretary Robert McNamara said the task force had sailed that morning, with arrival near Santos expected around 11 April. Ammunition was being assembled in New Jersey. Officials also considered commercial fuel possibilities. The distinction between diversion, assembly, sailing and delivery is not semantic. It tells us which instruments could have influenced the immediate contest, which were available for a longer struggle, and why a force still far from Brazil could not have won a battle there on 1 April.

Sources: S05 · S06 · S09 · S23 · S24 · S29

Chapter 05 / 1–2 April 1964

The Brazilian government falls before the fleet arrives

The balance shifted rapidly on 1 April. Army commands moved or declared against Goulart, officers who might have resisted chose accommodation, and the president’s supporters failed to turn their numbers into coordinated force. In the morning teleconference, Gordon distinguished confirmed developments from rumors and reported that momentum at that moment needed no special American encouragement. In the afternoon he relayed Castelo Branco’s statement that U.S. logistical support was unnecessary. Those statements do not erase prior encouragement or planning. They identify what was happening when the planned supplies and warships had yet to arrive.

Goulart left Rio for Brasília and then the south. Leonel Brizola urged resistance in Rio Grande do Sul; Goulart chose not to pursue a military confrontation. Brazilian accounts emphasize the failure to assemble loyalist commands and an effective general strike. The choice to avoid a wider conflict occurred amid a collapsing position, not in an empty room. Assessing the separate influence of a feared American intervention requires evidence of what Goulart knew and how he weighed it. The existence of Brother Sam alone cannot supply his private calculation.

Political removal acquired a civilian constitutional façade. Early on 2 April, Senate President Auro de Moura Andrade declared the presidency vacant, and Chamber President Ranieri Mazzilli took office. Goulart was still in Brazil. U.S. reporting acknowledged that fact, and Gordon privately called the juridical situation questionable. Ball likewise described legality as doubtful at the National Security Council. Washington’s preference for the outcome did not prevent its officials from recognizing weaknesses in the legal explanation offered for it.

Mazzilli’s installation did not put real power securely in civilian hands. Military leaders were already shaping the new order. The sequence is therefore more specific than either “the Americans did everything” or “a constitutional vacancy was filled.” Brazilian troops and political actors displaced an incumbent who had not yet left the country; the United States prepared material support and worked to give the successor arrangement an acceptable international position. The rapid collapse of organized resistance made the most conspicuous military component of that support unnecessary before it could reach its intended destination.

Sources: S08 · S10 · S11 · S12 · S13 · S22 · S25 · S27

Chapter 06 / 2 April 1964–1985

Recall, recognition and a regime that remained

On 2 April Gordon asked that the naval task force be recalled without appearing in Brazilian waters or attracting publicity. He wanted fuel tankers to keep coming until supply and remaining resistance were clearer. Washington did not instantly execute every part of his request. At the National Security Council that day, officials decided the force should continue for the moment; McNamara said it remained far from Brazil. The following morning he and Rusk agreed it was time to turn it around. The FRUS editorial note identifies Joint Chiefs recall instructions at 11:30 a.m. on 3 April.

That chronology preserves a real action and a real limit. Brother Sam was more than a speculative planning file: military movement was ordered and a force sailed. The force was recalled before intervention in the coup’s fighting. The cited record does not establish delivery of the prepared ammunition airlift. Separate routine military-aid cargoes were discussed on 2 April; their existence should not be recast as proof that Brother Sam’s emergency arms package had already been used. Petroleum movement and the cancellation of the naval force likewise should not be merged into a single instantaneous cancellation of every aid relationship.

Washington treated the successor arrangement as a government with which it could continue relations, while urging a stronger constitutional presentation. Ball’s later oral history recalled an early message effectively recognizing it and Johnson’s irritation at not being consulted. FRUS editors could not locate the telegram he described. The contemporary reports and later recollection should be read together without pretending they have identical evidentiary weight. Support for the outcome is independently established by policy decisions and private discussion, regardless of that missing message.

The victors did not simply restore the politics that had preceded the crisis. A military command exercised real power, an institutional act enabled purges, and Castelo Branco became president in April. Early repression targeted organized labor, students and the left, with arbitrary arrests and torture. Later institutional acts further restricted political and judicial guarantees; military rule lasted until 1985. Brazilian officials made those subsequent decisions. Recognizing their responsibility does not make the outside support irrelevant: the regime Washington welcomed became a durable dictatorship, exposing the distance between the democratic language of the intervention’s advocates and the political order that followed.

Sources: S11 · S12 · S13 · S25 · S26 · S27

Chapter 07 / 1970s–2026

What the released record changed

Brother Sam became a problem for an older public account in which Washington had merely watched Brazilians settle their own affairs. Declassified cables, Joint Chiefs records and the presidential recording disclosed advocacy, close access to conspirators, conditional supply arrangements and a naval deployment. Later releases placed those final days within policy discussions extending back into the Kennedy administration. The record supports a substantial American role in preparing to help an overthrow. That conclusion does not depend on accepting every later participant’s defense or every critic’s most expansive accusation.

The files also make a mechanically controlled American coup difficult to sustain in its strongest form. Brazilian officers disagreed, Mourão’s movement precipitated decisions, and the allegiance of major commands changed as events unfolded. Gordon was advocating measures that Washington initially questioned. The naval force’s timetable lagged behind the decisive collapse. To claim that American combat forces accomplished the overthrow requires evidence of action that these records do not show. To claim outside support was causally irrelevant because the fleet was unused overlooks diplomacy, covert political activity, intelligence relationships and the possible influence of an available backstop.

The hardest question is the counterfactual: would the coup have failed, been delayed or cost more lives without U.S. support? The documents show intent and capability more clearly than they measure that alternative history. A strong enabling account points to assurances, supply preparations and diplomatic acceptance as ways to reduce insurgents’ perceived risk. A strong domestic-causation account points to Brazilian command decisions and loyalist disorganization as sufficient for the outcome actually observed. Both must explain the same dated sequence. Neither gets to resolve the unobserved alternative merely by repeating its preferred label.

Evidence that would materially sharpen the assessment includes authenticated CIA operational records with authorization and delivery trails; communications showing precisely what key Brazilian commanders were told before deciding; Brazilian command diaries; and contemporaneous evidence of what Goulart knew about American preparations. The guide’s claims distinguish what is established from what those records could still change. There was a Brazilian civil-military coup and an American effort to help its preferred side prevail if assistance became necessary. The unused fleet limits the form of intervention that occurred. It does not absolve the policy that sent it toward Brazil.

Sources: S01 · S03 · S05 · S06 · S13 · S19 · S21 · S25 · S28 · S30 · S31