At the edge of a field in Zande country, far enough from the huts that no neighbour's witchcraft can overhear, a man sits with a fowl pinned under his foot. He twirls a small brush into a container filled with a liquid drug, presses it through a filter of folded leaves, and works the liquid into the bird's beak. The drug is benge, a poison drawn from the root of a forest creeper.
Beside him sits another man, on whose behalf the entire ritual has been initiated and it's his turn to take over the ritual. He needs an answer, so he begins to speak, not to the bird, but to the poison inside it. "If my wife's illness comes from that man's witchcraft", he says, "let the fowl die." The bird shudders, beats its wings, and lies still. An answer has come. Or not.
Nothing is quite settled yet. This was just the first part of the ritual as a second fowl is dosed and the question is turned upside down, "If the first answer was true, let this one live". Only when the two birds disagree in the right way, one dying and one living, has the oracle truly spoken. A single death proves little. The second bird is the ritual's answer to doubt, reaffirming the first. In the Zande view the poison was not merely poisonous. It knew things. When a chief was present, its verdict could carry the weight of law.
Between 1926 and 1930, the anthropologist E. E. Evans-Pritchard watched this performance again and again, over three expeditions that added up to twenty months. He photographed the operators crouched in the dust, and in 1937 he published 'Witchcraft, Oracles and Magic among the Azande.'
The Crime
A woman grows thin in a darkened house, wasting away with an illness that can not be diagnosed even by a spiritual healer. Her family searches for answers. What has caused the illness? Why has it refused to leave? And could someone have brought it upon her?
In many pre-colonial African societies, illnesses and other misfortunes could be attributed to witchcraft. When this happened, most likely all herbs or efforts made by a spiritual healer, have been sought and neither worked. Then, it became a question of who might be responsible for their suffering. Suspicion could fall on anyone: a neighbour? A relative? Someone the family had recently disagreed or quarrelled with? However, more often than not, there was never physical evidence to examine these suspicions and no witness could testify as to what happened.
This was one of the situations in which poison oracles and other forms of ordeal were consulted. Peter Leeson and Christopher Coyne, who studied the sasswood ordeal in Liberia, found that the poison was generally reserved for serious accusations and cases in which ordinary evidence was missing or inconclusive. The ordeal provided a means of reaching a verdict when the alleged offence could not be established through conventional investigation.
Of course, not every offence required an oracle. In many African societies, ordinary disputes were brought before elders, chiefs and councils, who heard the parties, considered the evidence and determined the appropriate punishment or compensation. Among the Yoruba, for instance, the Ogboni, also known as Osugbo in Ijebu and Egba towns, was a council of senior men and women with political, judicial and spiritual responsibilities. Its authority extended to serious matters of public order and had so much influence, it could remove a king from power.
But witchcraft presented a different problem. How do you investigate an offence that is believed to have happened through unseen forces? How do you identify the person responsible for an illness when there is no physical evidence to connect that person to it? The ordinary methods of settling disputes could not always answer these questions. This is where the oracle became a divine intercessor. It was consulted to reveal what human witnesses and physical evidence could not.
The Oracle
The Zande kept a ladder of oracles for different occasions. At the lower end was the rubbing board, two small pieces of wood that could be carried anywhere and consulted at almost any time. It was convenient but its answers were never particularly reliable. Above this came the termite oracle, which required two branches from different trees to be placed among termites, and the answer was determined by which branch the insects ate. Like the rubbing board, its answers were also not reliable and the Zande knew it. Yet, they found reasons to explain the failures of their oracles without abandoning them. Hence, misgivings were often heard: a taboo had been broken, the poison was weak, the question was badly put. Evans-Pritchard had a name for such excuses, 'secondary elaborations'.
Down the River
The Zande were only one people among many. Further south, along the Zambezi, David Livingstone encountered other communities that used poison ordeals to determine guilt.
However, the way these tests were conducted, and what happened to the accused afterwards, could differ considerably. The gentlest of these tests kept human beings out of danger altogether.
Among the Barotse, for instance, the accused did not have to drink the poison themselves. Livingstone recorded that the ordeal medicine was administered to a cock or a dog, and the animal's reaction was used to determine the innocence or guilt of the person accused. Whether the animal vomited or purged provided the answer. The test therefore placed the immediate physical danger on the animal, rather than the person whose guilt was being investigated. The accused, however, still faced the consequences, a fine or exile or a beating, but nobody had to own the decision. No chief had condemned anyone and no family had chosen revenge. The bird had died, and the oracle had spoken.
Farther along, among the Banyai, the accused had no such protection. Livingstone had reached the village of a chief named Monina when a witch-doctor arrived. That morning, all of Monina's wives went out into the fields fasting. A husband who suspected his wife of bewitching him could call for the ordeal, known as muavi. The women would each drink an infusion prepared from a plant called goho, raising a hand to heaven as they swore their innocence.
The outcome determined their fate. Those who vomited were declared innocent and allowed to return home, where they slaughtered a cock as an offering of thanks to their guardian spirits. Those who purged were pronounced guilty and burned to death.
Livingstone was startled by what he witnessed, or rather, what he was told was about to take place, for he had not watched it himself but had merely been told how it would go. He had come to believe that women in the region enjoyed considerable influence, and he could not understand why they would submit to an ordeal that could end in their deaths. The explanation he received was that the women had such confidence in the medicine that they willingly sought the test. They believed themselves innocent and trusted the muavi to expose only the guilty.
But there is another perspective we can choose to see it from. A woman might have believed completely in the power of the medicine and volunteered to take the test. She might also have felt that refusing it would make her appear guilty. Livingstone's account tells us that the women were said to go willingly, even eagerly. It does not tell us how much freedom they had to refuse.
Red Water, White Bean
Thousands of kilometres west, in present-day Sierra Leone and Liberia, another poison ordeal was practised. Here, the drink was known as red water and was prepared from the bark of the sasswood tree. The mid-nineteenth-century writer John Leighton Wilson described how the accused stepped forward before a gathering of onlookers, confessed the evil of their past lives, called on God three times and drank the mixture. The result determined their fate. Those who vomited were declared innocent, while those who died or showed certain other reactions were considered guilty. The people who gathered the bark treated the tree with great reverence, speaking to it as something that could not lie.
But there was a problem. The person preparing the poison also had considerable control over the outcome. In her 1853 account, Sarah Hale complained that those who administered the ordeal could manipulate the mixture, making it strong enough to kill an accused person or weak enough to spare them. Leeson and Coyne also refer to accounts of specialists adding ingredients that would induce vomiting. In other words, the person responsible for preparing the test could influence the very verdict it was supposed to reveal.
The use of poison ordeals was not limited to the sasswood tree. In other parts of Central and East Africa, similar practices were known by different names, including nkasa around the Congo and mwavi around Lake Malawi. The plants, preparations and rules varied from one community to another, but the underlying purpose was familiar: to use an ordeal to establish guilt where ordinary evidence was considered insufficient.
On the coast of what is now southeastern Nigeria, among the Efik of Old Calabar, the poison came from a bean rather than tree bark. Known as esere, or the Calabar bean, it was obtained from the seeds of a climbing plant. The seeds were crushed and soaked in water to produce a milky liquid, which the accused were made to drink. As with the other ordeals, the body's reaction was taken as evidence of innocence or guilt.
The bean was also used in a particularly dangerous form of dispute settlement. Older accounts describe opponents dividing a single bean between them, each swallowing half. The dose was known to be capable of killing both. It was a deadly way of settling a quarrel, and one that left little room for either man to withdraw once the test had begun.
When the Ordeal Became a Weapon
So far, we have seen how poison ordeals were used to settle accusations that ordinary evidence could not resolve. But there was another side to these practices. The outcome did not always depend on the poison alone. It also depended on the person who prepared it, the authority who ordered the test and those who enforced its verdict. When these powers fell into the wrong hands, an ordeal meant to establish guilt could become a means of eliminating rivals, punishing enemies and taking possession of another person's property.
In the 1880s, the trader Fred Moir travelled to the village of the Ngoni chief Chikusi, about fifty miles south of Lake Nyasa, hoping to establish trade relations and secure peace. Moir was involved in the management of the Livingstonia Central Africa Company, the predecessor of the African Lakes Corporation. What he encountered was a large-scale administration of the mwavi ordeal.
Chikusi had summoned his wives and thousands of villagers to a plain within sight of Moir's tents. A medicine man stood before them in a headdress of feathers that extended about a foot above his head, with a cape of animal tails across his shoulders. In his hand was a thin stick tipped with a small ball. Six of the chief's wives were brought forward. They had been accused of stealing the chief's goods and committing adultery.
Bark was pounded in a wooden mortar, while the medicine man danced between the women and the mixture, calling on the mwavi to kill them if they had committed the offences and spare them if they were innocent. The poison was then poured into drinking gourds and handed to the women two at a time.
One of the women, who had been laughing hysterically from fear, called out to another person in the crowd to give her baskets of goods to the chief if she died. The response was chilling. In retort, the man told her to die if she wished, adding that he would not take her goods because they belonged to the medicine man. Two of these six women died and their bodies were left on the plain, while young men armed with spears and clubs rushed off to kill their relatives and seize their property. That night, hyenas fought over the corpses. The frightened woman survived, and a goat was slaughtered to celebrate her innocence.
The ordeal did not end there. After an interval, Moir watched twenty-nine of the chief's wives and twenty-three young men being taken to the drinking place. About seventy more wives followed. When Moir finally left, protesting against what he had witnessed, some two hundred people were gathered on the plain waiting to drink, with more still expected.
Moir suspected that the ordeal was being manipulated for personal gain. He observed that anyone who accumulated cattle or other property could attract the envy or suspicion of others, including the chief himself. An accusation of witchcraft could then be raised against the person and their family. If the accused died, their relatives could be killed and their property seized. And because the ordeal was believed to reveal the truth, few dared to question its verdict. Moir himself admitted that he could not be certain the mixture had been manipulated on the occasion he witnessed, but he believed such abuses were common.
In the central highlands of Madagascar, the poison ordeal had become an instrument of royal authority. Among the Merina, it was known as tangena, after the poisonous nut from which the preparation was made. The accused swallowed three pieces of chicken skin and then took the poison. If they survived and vomited up all three pieces, they were declared innocent. Failure to recover all three, or death, meant guilt.
Even here, however, the ordeal did not expose everyone to the same risk. Nobles and free people were generally tested only after animal substitutes, including a fowl and a dog, had succumbed to the poison. Enslaved people on the other hand were made to take the poison themselves.
Under Queen Ranavalona I, who ruled the Merina state from 1828 to 1861, the tangena became one of the principal instruments through which the ruling elite administered justice and dealt with suspected enemies. It was used mainly in accusations of witchcraft and sorcery, but Christians were also subjected to it during the persecutions of her reign. Enslaved people were among its greatest victims. When illness, misfortune or death occurred in an enslaver's household, suspicion could fall on the enslaved people living there, and many were repeatedly subjected to the ordeal.
The historian Gwyn Campbell, drawing on earlier estimates rather than any surviving comprehensive record, puts the average number of deaths from tangena at about 3,000 annually between 1828 and 1861. He also cites an estimated fatality rate of between 33 and 50 per cent among those who underwent the ordeal. These figures should be treated as historical estimates, but they give some indication of the scale of the deaths.
In 1863, Ranavalona's son and successor, Radama II, officially abolished the tangena ordeal. Its use did not disappear immediately, however. Campbell records that it continued clandestinely in some parts of the Merina kingdom and openly in other parts of the island.
The Logic Behind the Poison Ordeals
Why would anyone drink poison? Leeson and Coyne's answer is that belief changes what the cup means. If people believe the poison can identify the guilty, an accused person has to consider the consequences of refusing the test. Someone who is guilty may confess, pay compensation or flee rather than risk drinking. Someone who believes themselves innocent may be willing to undergo the ordeal, sometimes even requesting it to clear their name. Leeson and Coyne call this an "institutional placebo effect". The poison cannot identify guilt by itself, but belief in its power can influence how people behave. The specialists administering the ordeal can observe who volunteers, who refuses and who confesses, and use that behaviour to make judgements about the accused.
Chemistry may have also helped. One proposed explanation for the Calabar bean holds that its poison is absorbed through the mouth, so someone who gulps it down, sure of herself, is likelier to vomit it up, while someone who holds it in her mouth afraid takes in more. An explanation is not a defence, though. Leeson and Coyne themselves call the method crude, and Moir's explanation shows what the same machinery did when the man at the centre wanted a particular outcome. Leeson and Coyne explain that the specialist could also manipulate the mixture to influence the outcome. A person believed to be innocent might be given a weaker preparation or one containing ingredients that induced vomiting, while a stronger dose could be administered to someone the specialist believed guilty. The test therefore might be presented as a way of discovering the truth, but those who controlled it could also determine who suffered its consequences.
After the Verdict
Not every verdict ended in death. Among the Azande, a person identified by the oracle could be approached by a messenger carrying the wing of the fowl that had died during the consultation. The accused was expected to take water into his mouth and blow it over the wing, saying that he knew nothing of the harm and wished the sick person well. The gesture was intended to "cool" the witchcraft and persuade the accused to desist from harming the sufferer. Evans-Pritchard explains that, while the sick person remained alive, this procedure offered a way to prevent retaliation and resolve the accusation without further violence. If the person died, however, the situation changed: the dead person's relatives could feel compelled to seek vengeance.
The same process could therefore have very different consequences. Among the Azande, the ritual could contain suspicion and avert a feud. In other communities, an ordeal could end in the death of the accused.
In Liberia, trial by ordeal had been outlawed by the formal legal system long before the twentieth century was over. In 1916, the Supreme Court ruled against lethal ordeals, holding that forcing an accused person to undergo such a test violated the constitutional protection against self-incrimination. The Court reaffirmed its position in 1940 and again in 2005. Yet the practice continued, partly because the Ministry of Internal Affairs had continued to issue licences to traditional ordeal specialists under the Hinterland Regulations. On 28 June 2008, the Ministry banned the issuing of those licences.
The following year, the deaths of people accused of witchcraft in southeastern Liberia brought renewed attention to the practice. In June 2009, seven suspected witches in River Gee County were subjected to an ordeal. At least two died after drinking a poisonous concoction; the bodies of the others were later found butchered. In October, an assistant minister of Internal Affairs announced that the government had halted trial by ordeal. The ban, however, did not bring an immediate end to the practice. Reports of further deaths emerged in later years, including cases in Grand Kru and Grand Gedeh counties in 2012.
The poison itself found another use. In Edinburgh, the Scottish physician Robert Christison tested the Calabar bean on himself. In February 1855, he ate part of a seed to investigate its effects. When he began to experience severe dizziness and weakness, he tried to rid himself of the poison by swallowing the water he had just used for shaving, which induced vomiting. He survived, and published his findings in On the Properties of the Ordeal-Bean of Old Calabar, Western Africa. His paper was an important early scientific investigation of the bean, although it was not the first written account of it.
In 1876, the German ophthalmologist Ludwig Laqueur introduced physostigmine, the active compound in the Calabar bean, as a treatment for glaucoma. The substance constricts the pupil and helps reduce pressure within the eye. Decades later, in 1935, the African American chemist Percy Julian and his colleague Josef Pikl achieved the chemical synthesis of physostigmine at DePauw University. Their work assembled the molecule in eleven steps. A plant whose poisonous seeds had been used in ordeals in West Africa had become the source of a compound with an important place in the history of ophthalmic medicine.
Livingstone, writing about the Banyai, could not help thinking of home. He told his men how Scotland had once tested witches with water. The suspect was bound hand and foot and thrown into a pond. If she floated she was guilty and burned, and if she sank and drowned she was pronounced innocent. His companions were as astonished by the wisdom of his ancestors as he had been by theirs. In the end, every society has had to decide what to do with a grave charge that nobody saw.
