Maciel was born on 13 March 1830 in Quixeramobim, in the interior of Ceará, into a family that ran a modest backlands business and was locked in a bloody feud with the powerful Araújo clan. His grandfather taught him, and he did well at Latin, French and mathematics. He married Brasilina Laurentina de Lima in 1857 and worked as a salesman, a schoolteacher and a lay counselor. The marriage broke apart in 1861, and he walked out of ordinary life. For the next three decades he moved from village to village across Ceará, Pernambuco, Sergipe and Bahia, digging cemeteries, raising dams and collecting money for churches. He had vowed to build twenty-one and completed the work in twelve towns. To poor sertanejos he was a benefactor who arrived on foot. To the landowners whose workers followed him, he was something else.

In 1893 he stopped moving. The site he chose, near Monte Santo on the Vaza-Barris, was hard to reach and easy to defend, and within two years several thousand people had joined him. Belo Monte held property in common, did without currency and civil marriage, and banned taverns and prostitution. Its people came from across the region's ethnic and economic range — small farmers, freedpeople and their children in the years just after abolition, drought refugees, families with nowhere else to go. By the time the army arrived the settlement had grown to some 5,200 houses of mud and straw, making it, by several accounts, the second-largest urban center in Bahia. Conselheiro preached that the long-vanished Portuguese king Sebastian would return and that monarchs ruled by God's right. In a Republic four years old and unsure of itself, that was enough to make him a threat.

The first force sent against Canudos was about a hundred men under Lieutenant Pires Ferreira. The Canudenses met them at Uauá on 4 November 1896 and drove them off. Ferreira came back three weeks later with 104 men and attacked the settlement itself on 21 November; that failed too, and the soldiers retreated after severe losses. The federal government took over. On 3 March 1897 Colonel Antônio Moreira César attacked with 1,300 troops — three infantry battalions, cavalry and artillery. Two days of fighting killed him and broke his column, and the retreat that followed, completed by 6 March, was a rout. The news landed in the coastal cities as a scandal. Something in the interior was beating the Brazilian Army, and the press had already decided what: a monarchist conspiracy plotting to restore the throne. The rumors ran well ahead of any evidence.

The fourth expedition set out on 16 June 1897 under General Arthur Oscar de Andrade Guimarães, and grew until more than 8,000 soldiers were in the field. This time the army did not try to rush the settlement; it surrounded it and shelled it. Conselheiro died on 22 September. The fighting ended in the first days of October — on 5 October, by the reckoning of the historian Walnice Nogueira Galvão — after dynamite bombs were thrown into the last defended houses. Men who accepted an offer of surrender, on the promise that their lives would be spared, were surrounded and hacked to death in front of hundreds of witnesses. What was left to give up, in Robert M. Levine's account, was roughly three hundred women reduced by hunger to walking skeletons, their children, and a few old men. The soldiers then smashed, leveled and burned all 5,200 buildings.

How many died is genuinely unknown. Euclides da Cunha, the engineer and journalist who reached Canudos with the fourth expedition, put the total near 30,000 — 25,000 residents and 5,000 attackers — and that is still the figure most often quoted. Levine's estimate is around 15,000; Joel Singer's is 5,000. The gap is not carelessness. Nobody counted the people of Belo Monte while they were alive. Da Cunha turned what he saw into *Os Sertões* in 1902 — *Rebellion in the Backlands* in English — and it remains the foundational Brazilian account. It is also a book saturated with the racial science of its moment: da Cunha believed the mixed-race people of the interior were doomed by their ancestry to backwardness, and he wrote them that way. As Peter Robb has pointed out, nearly every correspondent who covered the war was embedded with the army; a flawed witness is very nearly the only witness there is. Galvão's summary of what the Republic actually won is drier: the regime was solidified and the specter of monarchical restoration was exorcised at last.