Brazilian Republic

At its end the republic covered almost exactly the territory of modern Brazil, some 8.5 million square kilometres, and governed about 33 million people. At its beginning it governed 14.3 million, most of them illiterate, most of them rural, and only a small minority able to vote. In between, the country absorbed close to three million immigrants, laid most of its railway network, destroyed two millenarian communities in the interior with artillery, lost and then squandered the world rubber trade, built a capital city's boulevards over the houses of its poor, and produced a modernist movement and a generation of rebel officers who between them invented the idea of a Brazilian interior worth governing.

It matters because it is the laboratory in which modern Brazil's institutions were assembled: the federative pact, the secular state, civil marriage, the professional army as political arbiter, the coffee economy, and the patronage system that Brazilians call coronelismo. It matters too because the republic's central problem — a state that could not reach most of its citizens, and a citizenry that could not reach its state — outlasted it. The republic fell in 1930, but the arrangement of local bosses, state governors and a federal government that bargained with both survived well into the 1950s, and its vocabulary survives today.

Two phrases used for the period are retrospective. República Velha, the Old Republic, took hold after 1930, when the regime that replaced it needed a past to be new against; contemporaries called the country simply "the Republic". Café com leite, coffee with milk, is a later shorthand for the dominance of São Paulo and Minas Gerais, and no document establishes it as a pact. Both labels describe something real. Neither was what the participants called their own politics, and the story reads differently once that is clear.

At a glance

How did a monarchy fall in a single day?

The coup of 15 November 1889 took about eight hours and killed almost nobody. Marshal Deodoro da Fonseca, a 62-year-old army officer with a reputation for bluntness and a liver complaint that most of Rio believed would kill him within the year, rode out to the Campo de Santana, the parade ground behind the railway station, at the head of a column of troops. Units of the garrison joined him through the morning. The imperial cabinet met, discovered that the army would not fire on the army, and resigned. By evening a group of civilians and officers had proclaimed the republic from the balcony of the city council chamber. Pedro II, who was at Petrópolis, returned to Rio, was informed that he had been deposed, declined to let anyone fight for him, and sailed for Portugal with his family on 17 November.

The speed was possible because the monarchy had already lost the three groups that mattered. The army had been accumulating grievances for two decades: officers resented being used as a police force against abolitionists and election rioters, resented the promotion of men with court connections over men with combat records, had been left underfunded since the war with Paraguay had made it a national institution, and had been inflamed by the "military question" of the 1880s, in which two officers were court-martialled and then amnestied by public agitation. The Church had been estranged since the 1870s, when the state's arrest of two bishops in the affair of the Freemasons left a mark that the court never erased. And the planters of São Paulo, who had been the monarchy's most reliable supporters, had converted to immigrant wage labour in the 1880s and no longer needed an emperor to protect slavery; abolition in May 1888 removed the last interest that bound them to the throne. None of this made a republic inevitable. It made the monarchy cheap to overthrow.

What replaced it was not a movement but a coalition of the aggrieved. The civilian republicans were a small, urban, mostly professional group — lawyers, journalists, doctors, positivists, a few Jacobin radicals — with no party organisation capable of winning an election and no following in the countryside. Deodoro himself was a monarchist who had agreed to dismiss the cabinet, not to abolish the dynasty, and spent the rest of his life insisting that he had acted to save the army's honour. The republic was the army's work: the one institution the empire had taught to think of itself as national, turned against the dynasty it had served.

What did the provisional government do before there was a constitution?

For fifteen months the country was governed by decree. The provisional government abolished the monarchy's political police, separated church and state in January 1890, instituted civil marriage and secular cemeteries, and — in a measure that says a great deal about how the republicans understood citizenship — declared every foreigner resident in Brazil on 15 December 1890 to be Brazilian unless he formally opted out. The "great naturalisation" was aimed at the immigrant labour force and at the Portuguese merchant community, and it drew protests from the Portuguese and Italian governments before it was quietly narrowed.

The economy was the harder test. Rui Barbosa, the finance minister, tried to create credit where the monarchy had created none, licensing banks to issue notes against government bonds. The result was a speculative boom — the encilhamento, the "saddling" — in which companies were floated on paper, shares changed hands at multiples of their value, and prices rose through 1890 and 1891. When it broke, the banks failed, the currency fell, and the republic acquired the reputation for financial incompetence that its opponents never let it lose. Rui Barbosa's own later career as a liberal constitutionalist was spent partly in arguing that he had been misunderstood.

The Constituent Congress, elected in September 1890, produced a constitution promulgated on 24 February 1891 that was the most decentralised in Latin America. It created the United States of Brazil out of the old provinces, gave each state the right to tax exports, to borrow abroad, to maintain its own military police and to elect its own governor, and it reserved for the union only the import tariff, the postal service and the right to issue currency. Congress was bicameral: a Chamber of Deputies, at least four per state and one per 70,000 inhabitants, and a Senate of three senators per state and three for the Federal District, elected for nine years and renewed by thirds. The president served four years and could not succeed himself.

The bill of rights was, on paper, generous for its time. Article 72 declared all persons equal before the law, abolished titles of nobility, guaranteed public religious worship, recognised only civil marriage, made public instruction secular and forbade any church to receive state funds. The suffrage was another matter. Article 70 made voters of "citizens over 21" who registered, and excluded the destitute, enlisted soldiers, members of religious orders under a vow of obedience — and the illiterate. In a country where perhaps four out of five adults could not read, that clause did most of the work of exclusion. Women were not named in the article and were excluded in practice; the first Brazilian woman to register a vote did so in Rio Grande do Norte in 1927, and women only received the federal franchise in 1932, after the republic had fallen.

One further clause mattered more than anyone intended. Article 14 made the army and navy "permanent national institutions" and declared the armed forces "essentially obedient, within the limits of the law, to their hierarchical superiors". The phrase came from officers who wanted protection against governments that might disband them, and from civilians who wanted a check on the president. It supplied, for the next forty years, a constitutional argument for the army to decide when an order was lawful — which is to say, when to obey and when to govern.

Deodoro was elected president by Congress on 25 February 1891, and lasted nine months. He had no talent for sharing power with a legislature, and the Congress elected alongside him had no intention of being decorative. When it began investigating his government's finances and the military budget, he dissolved it on 3 November 1891, declared a state of siege and announced that he would govern alone. The navy revolted; the garrison split; within three weeks he resigned, on 23 November, and went home.

Why did the 1890s turn into a civil war?

Vice-President Floriano Peixoto took office on Deodoro's resignation in circumstances that were, by the letter of the constitution, illegal: article 42 required a new election if the presidency fell vacant before half the term had elapsed. Floriano did not call one, and the legal objection never went away. He governed as a soldier's president — "the Iron Marshal" — with the support of the Jacobins, the radical nationalist republicans who saw plots everywhere, and against the men who had made the republic and now felt excluded from it.

The reckoning came in 1893. In September the navy rose in Guanabara Bay, led by admirals who had served the empire and who now demanded Floriano's removal; in the south, the Federalist revolt against the positivist governor of Rio Grande do Sul, Júlio de Castilhos — a state that had seceded as the Riograndense Republic in 1836 and had never quite accepted Rio's terms — spread into Santa Catarina and Paraná and became a war between the state's two rival families with a private army on each side. The naval revolt was broken by an improvised government fleet, which included armed merchant ships bought in the United States and escorted into Brazilian waters by United States cruisers, and by the rebels' own incompetence. In March 1894 the surviving ships fled to Argentine men-of-war in the Plata and were interned. The land war in the south dragged into 1895 and was fought with a savagery that José Maria Bello, the historian who lived through it, called the cruellest of Brazil's civil wars; prisoners were executed on both sides, and the federal government intervened on the side of the state machine.

Floriano handed over in November 1894 to Prudente de Morais, the first civilian president, who ended the war by amnesty and tried to reduce the army's political weight. He established a General Staff on the German model in 1896, began to professionalise the officer corps, and was immediately handed a war he did not want. In the backlands of Bahia, a preacher called Antônio Conselheiro had gathered a community that the republic's newspapers had decided was a monarchist army in waiting. Prudente's government sent 9,542 men to destroy it.

The 1890s set the pattern that the rest of the republic followed. Civilian politicians governed through the states; the army considered itself the guarantor of the constitution and intervened when it disapproved; the urban middle class produced journalists and conspirators; and the rural poor, when they organised at all, were treated as a security problem rather than a political one. Every later crisis — 1904, 1910, 1922, 1924, 1930 — was a variation on this arrangement.

What happened at Canudos?

Canudos was a settlement of perhaps 20,000 to 30,000 people on the Vaza-Barris river in the Bahian sertão, in a bend where the river runs dry for most of the year. It was founded in 1893 by followers of Antônio Vicente Mendes Maciel, known as Conselheiro, a wandering preacher in a long shirt and a staff who had spent 20 years repairing churches and cemeteries across the north-east and denouncing the republic as the work of the anti-Christ. His followers called the place Belo Monte. They built some 5,000 houses of mud and wattle, a church, a school, and a communal economy in which land was held in common and labour was shared. They came because the droughts of the 1880s and 1890s had destroyed their crops, because the republic had imposed a civil registry and new taxes that reached into the sertão for the first time, and because the local priests had turned against Conselheiro and his people had turned against the priests.

The war against them was made by local political feud and metropolitan panic. Around Canudos a quarrel between two landowning families had divided the region, and Conselheiro's refusal to deal with either side made him an enemy of the winners. In Rio de Janeiro the press reported that the community was armed by monarchist exiles and intended to march on Salvador. The first expedition, in November 1896, was a hundred men under a lieutenant; it was ambushed and fled. The second, in January 1897, was 500 men with two field guns; it was broken in a day, and its commander killed. The third, in February, was 1,300 men under Colonel Antônio Moreira César, a florianista with a reputation for shooting prisoners; it attacked entrenchments without reconnaissance on 2 March, was shot to pieces, and lost its colonel. The sertanejos, who had captured rifles and guns from the first two expeditions and had numbers on their side, were now fighting a war they had not wanted and winning it.

The fourth expedition was a full division. Some 9,542 officers and men were eventually committed, with artillery, telegraph and a supply train, under General Artur Oscar; they reached the town in June, were pinned on a hill in front of it for a month, and were nearly destroyed before reinforcements arrived. The siege that followed lasted through September. On 1 October an assault cost the army 587 dead and wounded and the defenders perhaps 5,500 killed or missing; on 4 October dynamite and kerosene finished the last resistance, and the town was burned. The army's wounded in the July–October fighting came to 4,193. Nobody counted the dead of Belo Monte, and the published estimates have ranged from about 5,000 to 25,000 for a century.

Canudos entered Brazilian memory through a book written by an engineer who had gone there as a war correspondent. Euclides da Cunha's Os Sertões, published in 1902, is a geological, ethnographic and military account of the campaign which ends by concluding that the republic had committed a crime against a people it did not understand — and which simultaneously describes those people in the language of racial science. The book's argument, that the sertão was the true Brazil and that the coast had destroyed it out of ignorance, became one of the founding propositions of Brazilian national thought. The war itself taught the army two lessons it never forgot: that the interior was vast, hostile and ungoverned, and that the army was the only institution capable of entering it.

How did federalism work in practice?

The 1891 constitution gave the states the export tax, and the export tax made the states rich. São Paulo, whose coffee left through Santos, could fund a railway network, an immigration programme, a school system and a military police force that by 1906 was training with French instructors. Minas Gerais, whose coffee and cattle went by rail to Rio, did much the same at a slower pace. The poorer states of the north-east had little to tax and became clients of the federal treasury, which is where the politics of the republic was really decided.

The linking institution was the National Guard, created in 1831 as a militia of propertied citizens and used ever since as a rural police. Its officers were commissioned locally, which is why prosperous landowners were addressed as coronel whether or not they had ever commanded anything. The Guard was made the army's second line by decree in 1918, its units dissolved, and what remained was wound up in 1922; the title outlived the institution by decades, as titles do. Victor Nunes Leal, whose Coronelismo, Enxada e Voto (1948) is the classic analysis, insisted that coronelismo was not simply local thuggery but a bargain between three levels of a weak state: the local boss delivered votes and order, the governor delivered roads, schools, police appointments and credit, and the president delivered protection and federal money. It was, in his phrase, an exchange of advantage between a public power that was growing stronger and the decaying social influence of the landowners.

The bargain was necessary because the electorate was tiny and entirely manageable. With illiterates excluded, the voters of a municipality might number a few hundred, and those few hundred could be rounded up, paid, pressured or simply counted twice. Congress verified its own members' elections, which gave the majority a legal instrument for expelling the opposition — the degola, the beheading — and the state governors, who controlled the police and the electoral boards, decided who would sit. In 1910 there were about 627,000 registered voters in a population of some 22 million, under 3 per cent; through the 1920s the figure never rose above 3.4 per cent. Elections were held, results were published, and the outcome was usually known in advance.

The system was formalised under Campos Sales, president from 1898 to 1902, whose política dos governadores offered the state oligarchies federal non-interference in exchange for congressional discipline. Campos Sales had inherited a currency crisis and a suspended foreign debt; he negotiated a funding loan with the London house of Rothschild that froze debt service and gave the government room to deflate, and he used the breathing space to build the alliance that governed Brazil for the next thirty years. The arrangement was not a constitution and had no author. It was a set of understandings that held because everyone at the top of it benefited, and because the people outside it had no way to complain that counted.

Who actually governed — and was it coffee with milk?

The table shows both the pattern and its limits. São Paulo and Minas Gerais held eight of the eleven presidencies between 1894 and 1930, and the two states' machines, the Paulista Republican Party and the Minas Republican Party, were the arbiters of most successions. But Nilo Peçanha came from Rio de Janeiro, Epitácio Pessoa from Paraíba, and Hermes da Fonseca was a gaúcho marshal rather than a Minas planter. Rio Grande do Sul, with its cattle, its rice and its disciplined republican machine under Júlio de Castilhos and then Borges de Medeiros, was a third power that could not be ignored; Bahia and Pernambuco mattered in the north-east.

Coffee paid for it. Brazil produced about three-quarters of the world's coffee at the end of the 1890s, and coffee supplied 63 per cent of the country's export earnings in 1891. The crop's centre of gravity moved west with the railways: São Paulo grew 25 per cent of Brazil's coffee in 1880, 40 per cent in 1888 and 60 per cent by 1902, and the labour for it came from Europe — some 201,000 immigrants entered the state between 1884 and 1890, and more than 733,000 between 1891 and 1900, most of them Italians, subsidised in passage and settled on the colonato system, in which a family worked a few thousand trees in return for a house, a plot of food crops and a share of the harvest.

Overproduction was built into the system, because a coffee tree takes four years to bear and every planter planted in the same years as his neighbours. When prices fell, the state intervened. The Taubaté agreement of 26 February 1906, signed by the governors of São Paulo, Minas Gerais and Rio de Janeiro, set minimum prices for the surplus, to be bought with loans raised abroad, and committed the federal government to create a currency board — the Caixa de Conversão — to stabilise the exchange rate at a fixed parity. It was the first of a series of coffee defence schemes that continued, with interruptions, until 1930, and it worked in the narrow sense that it kept the planters solvent; it also locked the republic into a single export, a fixed exchange rate and a permanent dependence on foreign loans. Brazilian governments borrowed in London in 1898 and again in 1914 to hold the currency, and in 1929 the whole structure met the Wall Street crash.

The other Brazil existed alongside this one and was not consulted about it. The north-east, whose sugar could not compete with Cuba and Java, exported people instead: labourers moved south to the coffee zone and to the Amazon, and the region's political weight declined with its revenue. Rio de Janeiro, the capital, was a city of some 800,000 in 1900 with the sanitary conditions of a much poorer place. The south had cattle and colonists; the Amazon had rubber, briefly. The republic's politics was conducted as if the country were the triangle between Rio, São Paulo and Belo Horizonte, and for most purposes it was.

What did the republic offer the poor?

Rio de Janeiro was the republic's shop window, and in 1903 the mayor, Francisco Pereira Passos, began to rebuild it. Avenida Central was driven through the old colonial centre, hundreds of buildings were demolished, tramways and sewers were laid, and the poor were pushed out of the centre into the hills and the suburbs. The same years brought a public health campaign under Oswaldo Cruz, who had been made director-general of public health in 1903 and who attacked yellow fever by sending teams door to door to clear standing water and kill mosquitoes. Yellow fever receded. Smallpox did not, because vaccination required consent, and consent was not forthcoming.

In the winter of 1904 an epidemic killed some 3,500 people in the capital. Cruz proposed a law making vaccination compulsory, with sanitary inspectors empowered to enter houses, issue certificates and fine the recalcitrant; the law passed on 31 October 1904, the regulations followed nine days later, and on 10 November the city rose. For six days crowds fought in the streets, overturned trams, cut telegraph wires and built barricades; the Positivist Church's leaders, monarchist officers, opposition journalists and medical men who doubted the vaccine all found themselves on the same side. A military conspiracy attempted a coup and failed. The government declared a state of siege on 16 November; when it was over 30 people had been killed, 110 wounded and 945 arrested, and nearly half of those arrested were shipped to Acre, where many were put to forced labour. Rodrigues Alves kept Cruz in his post but dropped the compulsory vaccination. Smallpox deaths in Rio fell to nine in 1906; a new epidemic in 1908 killed more than 6,500.

The revolt is usually remembered as a riot of the ignorant against science. It was more specific than that. The law gave inspectors the right to enter homes, to inspect women's bodies, and to impose a certificate on anyone who wanted to work, marry or travel, and it arrived in the middle of a remodelling programme that had already expelled the poor from the city centre. The argument in the streets was about liberty and dignity as much as about vaccines, and the government's retreat — which cost thousands of lives four years later — is evidence that the protesters' objection was political, not medical.

Six years later the navy produced the republic's most disciplined popular protest. In November 1910 the crews of the new dreadnoughts Minas Gerais and São Paulo, led by a sailor called João Cândido, seized their ships and trained the guns on the city. Their demand was the abolition of flogging, a punishment abolished in the army but still used at sea, and they got it, along with an amnesty that was later dishonoured: the leaders were discharged, some were imprisoned, and Cândido died in poverty. The "Revolt of the Lash" was the most successful armed action by working men in the republic's history, and it was settled in four days by negotiation.

Labour organisation was otherwise treated as a police matter. São Paulo's textile and clothing factories, where most workers were women, saw strikes in 1903, 1906 and 1912; in July 1917 a walkout in a São Paulo mill turned into a general strike in São Paulo, Rio, Santos, Recife and Porto Alegre, with demands for an eight-hour day, higher wages and the right to organise. The government's response was repression, and the employers' was blacklisting, but the strike did produce the first hesitant labour legislation: the Eloy Chaves law of 1923 created pension funds for railway workers, the beginning of Brazilian social insurance. For the rural majority — the sharecroppers, tenants, cowboys and squatters who made up most of the population — there was nothing at all.

What did the rubber boom do to the Amazon?

For about 30 years the Amazon was one of the most profitable places on earth. Rubber from wild Hevea trees, tapped by hand and smoked over palm-nut fires, was the only large-scale source of the material that the bicycle, the tyre and the electrical industry needed, and Brazil supplied most of the world's rubber. Manaus, a thousand miles upriver, acquired electric trams, a floating dock, a customs house built on a British prefabricated frame and an opera house that opened in 1896 with a performance of La Gioconda. Belém, at the river's mouth, doubled in population between 1890 and 1900, from about 50,000 to 96,560, and more than doubled again by 1920; Manaus went from 50,300 in 1900 to 75,704 in 1920.

The boom's labour system was debt peonage. A tapper was advanced tools, food and credit by a patron, and by the time he had paid for them at the patron's prices he owed more than the season's rubber was worth; the debt passed to his children. The trade's worst edge lay upstream in the Putumayo region, where a Peruvian company supplied the London market with rubber collected by indigenous workers under a regime of terror that a British consular investigation exposed in 1910–11. The Amazon's rubber was cheap only because the trees were wild and the workers had no alternative, and both conditions were about to change.

The change was botanical. In 1876 the Englishman Henry Wickham carried Hevea seeds out of the Amazon to Kew, and the seedlings were grown in Ceylon and Malaya, where rubber could be planted in rows on estates with indentured labour from India and China. Estate rubber was capital-intensive but predictable; Amazonian rubber was neither. Real wages in the Amazon in 1912 were still more than eight times those in Asia, and the cost of moving rubber down the rapids of the Madeira — where, according to consular reports, between 5 and 25 per cent of the rubber shipped was lost — was ruinous. Brazilian production peaked around 1912 and then lost its world primacy to Asia within a decade. The boom ended not with a crash but with a long slide into a subsistence economy.

Acre, the rubber region on the border with Bolivia, was annexed in 1903 after a war fought by Brazilian tappers and a local leader, Plácido de Castro, against Bolivian authority. The Treaty of Petrópolis of 17 November 1903 gave Brazil the territory in exchange for about £2 million and a commitment to build a railway around the Madeira rapids; the Madeira-Mamoré line, built between 1907 and 1912, cost the lives of hundreds and by some accounts thousands of workers, and was obsolete before it was finished, because the rubber it had been built to carry was no longer competitive. Henry Ford tried again in 1928, planting a rubber plantation at Fordlândia on the Tapajós; it failed too, in part because the trees were attacked by South American leaf blight, which the Asian plantations did not have. The Amazon's rubber economy never recovered, and the region's marginalisation in the twentieth century dates from that collapse.

Why did the Contestado rise, and why did it take an army to end it?

The Contestado was a war in the pine forests of the Paraná–Santa Catarina border, and it began with land titles. The two states had disputed the region since the colonial period; the Supreme Court had ruled for Santa Catarina in 1904 and Paraná had refused to accept the ruling, so both states had been issuing deeds to the same land for a generation. Into that legal chaos came the São Paulo–Rio Grande railway, built by Percival Farquhar's Brazil Railway Company, which received a 15-kilometre strip of land on either side of the line, and the Southern Brazil Lumber and Colonization Company, which bought the timber on it. Families who had farmed and grazed the region for a century — posseiros with no written title, or with titles from the wrong state — found their land sold, leased or fenced, and the timber cut.

The first fighting came in October 1912, at Irani, where a hundred men of the Paraná state militia attacked a gathering of pilgrims and were beaten off, with the loss of their commander. The pilgrims were followers of a healer called Miguel Lucena Boaventura, known as José Maria, who preached that the republic was the reign of the anti-Christ and that the monarchy was God's true government; he was killed at Irani, and his followers refused to accept it. Through 1913 and 1914 the movement grew around a series of fortified settlements — Taquarassu, then Santa Maria — whose inhabitants called themselves the Peers of France, after the guard of Charlemagne in the popular romances that José Maria had read aloud. At its height perhaps 20,000 people lived in the redoubts and controlled some 28,000 square kilometres.

The state militias could not cope, and in 1914 the federal government gave the command to General Fernando Setembrino de Carvalho, who had just finished a campaign against the followers of Padre Cícero in Ceará. He came with 7,000 men — nearly half the Brazilian army — and fought a war of encirclement and starvation. Through the second half of 1914 his columns closed the roads, burned the crops and shelled the redoubts; women and children came out to surrender in their thousands; Santa Maria fell in April 1915 after a two-month siege, with 600 defenders killed and 5,000 houses burned. The last bands were hunted down through 1915 and 1916, and the military campaign was over. The cost was put at about US$250,000, a figure the army did not enjoy explaining.

The dead have never been counted. Estimates run from a few thousand to 20,000, and the difference is not a matter of bookkeeping: nobody in authority had an interest in counting peasants. What the army took from the war was a conviction that the interior was full of fanatics, that the republic had abandoned them, and that the officer corps was the only institution capable of acting. The Contestado, like Canudos, became part of the army's institutional memory, and the officers who had served there — many of them trained in Germany and writing in the new journal A Defesa Nacional — carried their critique of the republic into the revolts of the 1920s.

What did 1922 mean?

In February 1922 a group of São Paulo artists and writers held a festival in the city's municipal theatre. The Semana de Arte Moderna ran for three nights, 13, 15 and 17 February, with an exhibition of painting, sculpture and architectural models in the foyer and lectures and concerts on stage. It was paid for largely by members of the coffee elite, and it was designed to scandalise them. Anita Malfatti's expressionist paintings had already been denounced in 1917 by Monteiro Lobato, the country's most influential literary journalist, as the work of a madwoman or a fraud; the Semana put her on the wall again, alongside Di Cavalcanti, Victor Brecheret and the music of Heitor Villa-Lobos, and the audience responded with catcalls, neighing and whistling. Mário and Oswald de Andrade read poems about São Paulo's streets and immigrants, and Graça Aranha told the audience that Brazilian art had to stop copying Europe.

On 5 July a group of young officers rose in the Copacabana garrison. It was a hopeless gesture: the government had warning, most of the conspirers were arrested, and the handful who marched out onto the Avenida Atlântica were shot down in the street, with two survivors. The episode became the founding myth of tenentismo, the lieutenants' movement, and the officers who survived it spent the next eight years trying to work out what it had been for. In July 1924 the tenentes took São Paulo for three weeks and were shelled out of it; in October the same year they rose in Rio Grande do Sul, and a column of some 1,500 men under Luís Carlos Prestes began a march that lasted until February 1927.

The Prestes Column was the republic's strangest military fact. It marched perhaps 24,000 kilometres — the figure usually given is 25,000, and the English edition of the standard modern study calls it fifteen thousand miles — through the interior of 13 states, fought dozens of engagements, was never destroyed, and never took a city it could hold. Its purpose was to keep the army's discontent alive and to spread it, and in that it partly succeeded: the column's officers came home convinced that the republic could not be reformed from the coast, and several of them, Prestes included, went on to join the Communist Party and to attempt an insurrection in 1935. Its larger effect was imaginative. Jacob Blanc's history of the column argues that the march turned the interior into a political subject — a place that the coast had to reckon with, and a mirror in which the republic's failure to govern its own territory became visible.

The two 1922s belong together, not because the modernists and the lieutenants liked each other — most of them did not — but because both were revolts of the young against a public sphere that had no room for them. Both attacked the same target: the rhetorical culture of the oligarchic republic, its parnassian poetry, its legalistic oratory, its confidence that Brazil was a European nation on the coast with a colonial hinterland attached. And both were absorbed, in the end, into the state that replaced the republic. The modernists got jobs in the Vargas cultural apparatus; the tenentes got posts in the army and the bureaucracy that dismantled the federalism they had marched for.

How did the Old Republic end?

The end came through the coffee price. The Wall Street crash of October 1929 destroyed the foreign loans that financed the coffee defence, and the price of coffee collapsed. Washington Luís, the Paulista president, had already broken the unwritten rule of the succession: instead of supporting a Minas candidate in 1930 and preserving the rotation, he nominated another Paulista, Júlio Prestes, the governor of São Paulo. Minas Gerais, Paraíba and Rio Grande do Sul formed the Liberal Alliance, with Getúlio Vargas, the governor of Rio Grande do Sul, for president and João Pessoa, the governor of Paraíba, for vice-president. Their programme was modest — a secret ballot, an independent electoral court, amnesty for the 1922 and 1924 rebels, and an end to the degola — but it was enough to make the election a contest.

The election of 1 March 1930 gave Júlio Prestes 1,091,709 votes out of 1,890,524 and a majority in Congress. The opposition denounced fraud: in the municipality of Osasco, 3,095 registered voters had somehow produced 6,018 votes for Prestes, and the congressional verification committee had thrown out 14 deputies elected in Minas and the whole Paraíba delegation in favour of the deputies from Princesa, a town that had declared itself independent of João Pessoa's government. On 26 July João Pessoa was murdered in a Recife café by a man whose motive was probably personal. The alliance had its martyr and its excuse.

The revolt began at 17:30 on 3 October 1930, in Porto Alegre, and spread through the south and the north-east. There was little fighting: the army's senior commanders were unwilling to fight for Washington Luís, and the state police of Minas and Rio Grande did most of the moving. On 24 October a military junta deposed the president in Rio and installed itself; on 3 November Vargas arrived by train and took office as provisional president. The decree that gave him his powers, nº 19.398 of 11 November 1930, dissolved the national Congress and every state legislature in the country, suspended the constitutional guarantees, excluded the acts of the provisional government from judicial review, and gave Vargas the legislative as well as the executive power of the union. The First Republic ended not with a constitution but with a decree.

It ended, moreover, without a social revolution. The coronéis who had delivered the votes for the old system delivered them for the new one; the state machines survived; the army that had made and unmade governments since 1889 remained the ultimate arbiter. What changed was the level at which politics was conducted. Vargas's regime would centralise the federation, intervene in the states, incorporate the urban working class into a corporatist labour system and, in 1937, abolish the constitution altogether. But it did so with the men and the methods of the Old Republic, and it took another 15 years and a world war for the country to acquire a democratic constitution again.

What did the Old Republic leave behind?

It left a state. The 1891 constitution's federalism, its secularism, its civil marriage and its bill of rights survived 1930 in substance; the 1934 and 1946 constitutions built on them. It left the army with a doctrine of guardianship that took the 1964 coup to exhaust and the 1988 constitution to dismantle. It left the coffee economy's railways, ports, immigration networks and financial habits, and it left a labour market in which the rural majority had no rights at all — the problem that Vargas's legislation addressed by incorporating some workers and ignoring the rest. It left, finally, a political culture in which the state is the prize and the municipality is the currency, which is what coronelismo means once the colonels are gone.

The labels came afterwards. República Velha is a name given by the regime that succeeded it, and like most such names it flatters the successor: it makes 1889–1930 a single block of stagnation, when the period was in fact one of the fastest rates of urban and industrial growth in Brazilian history. Café com leite flattens a complicated bargaining system into a dairy product. República da Espada (1889–94) and República Oligárquica (1894–1930) are textbook divisions that get the army's continuing role wrong in the second period and the civilians' role wrong in the first. The most useful description is the one the republicans themselves used: a republic that had abolished the monarchy and the state religion but not the property structure, the illiteracy or the local power that had underpinned both.

Historians have argued about the period since the 1930s, and the argument has moved. The generation of the 1930s — Gilberto Freyre, Sérgio Buarque de Holanda, Caio Prado Júnior — read its politics as a mask for economic interests. The generation of the 1960s and 1970s, working in the shadow of the military regime, studied the state, the army and the economy, and found in the Old Republic the origins of dependent capitalism. Since the 1980s the work has been social and regional: the sertão, the ex-slave population, the immigrant colonies, the women of the textile factories, the messianic movements, the municipal bosses. The republic looks less like a failed democracy now than like a state that was never designed to include most of its people, and which knew it.

Common myths about the Brazilian Republic

Timeline of the Brazilian Republic

What the map shows next

At 1891 on the map the map shows the republic at the moment its constitution took effect: the twenty states drawn out of the old provinces, the Federal District at Rio, and a territory that still excludes Acre, which would be bought from Bolivia in 1903. It is worth looking at the size of the states — Minas Gerais, Bahia, São Paulo — because their size is the federal pact.

At 1930 on the map the same territory is at the end of the coffee republic, in the year of the revolution. The next drawn year for this territory is 1946 on the map, when the republic reappears after the Vargas dictatorship and the war; the years in between are drawn under the Vargas-era state, which abolished the federal pact, and the military republic of 1964–85 has its own entry. The republic of 1985 onwards inherits the same borders and, in 1988, a constitution that finally extended the vote to everyone the Old Republic had left out.