Curare: The Arrow Poison, Its Plants and Its Science

Table of Contents

1. Poison Darts of the Amazon and the Guianas
2. Early European Accounts
3. Chondrodendron and Strychnos: The Curare Plants
4. Extract First, Plant Later, Molecule Last
5. Brodie, Waterton and the Breathing Donkey
6. Quaternary Ammonium and Curare-Like Action
7. Harold King and Tubocurarine (1935)
8. Acetylcholine at the Motor Nerve Ending
9. How Curare Blocks the Receptor
10. Richard Gill’s 1938 Expedition
11. Key Research Papers
12. Connections
13. Featured Videos

1. Poison Darts of the Amazon and the Guianas

“Curare” is not one substance but a family of arrow and dart poisons made by many different peoples across the northern half of South America, from the forests of the upper Amazon to the highlands and river valleys of the Guianas. What these poisons have in common is their effect rather than a single recipe: a wound from a coated point leaves the animal unable to move. A monkey or bird hit high in the canopy loosens its grip and falls; a running animal slows and collapses. For hunters working with light darts that could not kill large prey by force alone, a paralysing coating turned a small wound into a catch.

The word itself reached European languages in several spellings — curare, urari, ourari, wourali, woorara — which reflect the different ways travellers heard and wrote down the local names. One common account traces it to wurari in the Carib languages, including that of the Macusi people of Guyana. Another, given by the anaesthetist T. Raghavendra in his history of neuromuscular blocking drugs, connects “ourari” to Indian words meaning “bird” and “to kill”. The etymology varies from source to source, and the honest summary is that the name comes from the region’s own languages and that no single derivation is settled.

Ethnography of the preparation

In many travellers’ accounts, the making of curare was specialised knowledge, held by particular people within a community. Travellers’ descriptions agree in outline: bark and stems of certain forest vines were scraped, the scrapings were boiled for a long time, and the liquid was concentrated by evaporation into a dark, tar-like paste that could be stored and later smeared on points. Different groups added other plant material, and the finished poisons varied in strength and in the vessels they were kept in — bamboo tubes, small clay pots or gourds — which later gave European pharmacologists a rough way of sorting the samples that reached them. Claude Bernard, drawing on the travellers’ accounts, wrote in 1865 that the preparation was complex and that the poison “very speedily kills an animal if introduced under the skin”. That is as far as this page goes: the details of plant choice and processing are recorded here as history only.

Back to Table of Contents

2. Early European Accounts

The first European reports of poisoned arrows in the Americas came quickly after contact. Peter Martyr d’Anghera, an Italian-born chronicler at the Spanish court who collected the reports of returning sailors and officials, described poisoned arrows in the letters gathered as De Orbe Novo, including letters dated to 1516. He never saw the weapons himself; he wrote down what others told him, and his accounts mix observation with rumour. Historians of the drug, Raghavendra among them, cite Martyr’s letters as an early written European reference to the poisons that would later be called curare.

Eighty years later, Sir Walter Raleigh’s The Discovery of Guiana (1596), the book he wrote after his voyage up the Orinoco, also mentioned the poisoned arrows of the region. Raleigh’s book was written partly to promote further expeditions in quest of gold, and it is not a careful scientific source, but it helped fix the image of the deadly Guiana arrow in the English imagination.

From travellers’ tales to samples

The eighteenth and early nineteenth centuries brought more careful observers. The historian M. R. Lee lists among the explorers who brought curare to European attention the French scientist Charles-Marie de La Condamine, who travelled down the Amazon in the eighteenth century, and the Prussian naturalist Alexander von Humboldt, whose South American journey at the turn of the nineteenth century took him into the Orinoco region. Bernard himself, explaining in 1865 what was known when his own experiments began, credited “the interesting accounts of Alex. von Humboldt and of Roulin and Boussingault” for the little that Europe then understood. Those accounts described the poison, its preparation and its speed, but they did not explain how it killed.

Samples did reach Europe. They were dried, kept in their original containers and passed between collectors, museums and laboratories. This is why curare could be studied in Paris and London long before its botany was clear: experimenters had the poison in hand, but often not the plant, and rarely any reliable record of which plants had gone into a particular sample.

Back to Table of Contents

3. Chondrodendron and Strychnos: The Curare Plants

Two plant families supply the paralysing ingredients of most curares. The first is the Menispermaceae, the moonseed family, a group of mostly climbing plants. Its most famous curare plant is Chondrodendron tomentosum, a large woody vine (liana) of the tropical forest. Lee’s history describes d-tubocurarine, the principal active compound of the curares used in medicine, as coming from the liana Chondrodendron.

The second is the Loganiaceae, in which the genus Strychnos provides curare ingredients in other regions, particularly in the Guianas and the Orinoco basin. The genus is a striking example of how plant chemistry can cut both ways. Other Strychnos species are the source of strychnine — one of the alkaloids isolated by the French pharmacists Pelletier and Caventou — a poison that causes violent convulsions, the very opposite of curare’s limp paralysis. The curare alkaloids of Strychnos and the convulsant strychnine come from related plants, yet act on the nervous system in opposite directions.

Why the botany was hard

Several things made it difficult for outsiders to pin the plants down. A single curare could contain extracts of more than one plant. The makers did not always share their knowledge with visitors. The vines grow high into the canopy, so flowers and fruits — the parts botanists rely on for identification — were hard to collect, and a sample of bark or stem alone could not easily be matched to a named species. And curare from different regions really did come from different plants, so an identification that held for one people’s poison did not necessarily hold for another’s. Raghavendra summarises the modern view: the curares are derived from plants of the Menispermaceae (Chondrodendron) and the Loganiaceae (Strychnos), with the moonseed vine the source of the curare that entered medicine.

Back to Table of Contents

4. Extract First, Plant Later, Molecule Last

M. R. Lee, writing on curare in the Journal of the Royal College of Physicians of Edinburgh in 2005, compared its history with that of another South American medicine, quinine: “Like quinine, at first came the extract but no plant, and later the plant but no chemical compound.” Lee describes the whole journey from first European contact to a defined drug as taking about three centuries, and lists a chain of explorers and scientists along the way: La Condamine, Humboldt, Brodie, Waterton, Bernard, Dale, Walker and King.

The parallel with quinine is close. Powdered cinchona bark was used against fevers in Europe for nearly two centuries before Pelletier and Caventou isolated quinine from it in 1820; the bark came first, the botany was argued over, and the pure molecule arrived last. Curare followed the same order but took even longer, because the plant source was buried inside a mixed, processed paste rather than sold as recognisable bark.

Three stages in brief

• The extract. From the sixteenth century onward Europeans knew curare as a finished poison in a tube, pot or gourd, and from the early nineteenth century they could experiment with it on animals.
• The plant. Botanical identification came gradually through the nineteenth and early twentieth centuries, and was still being argued over in the late 1930s (section 10).
• The molecule. A pure, crystalline active compound, tubocurarine, was isolated only in 1935 (section 7), and the chemical messenger it blocks was identified the following year (section 8).

Claude Bernard’s work sits in the middle of this story. He knew neither the plant nor the molecule; he had only the extract. That was enough, because his question was not “what is curare?” but “what does curare stop working in a living body?”

Back to Table of Contents

5. Brodie, Waterton and the Breathing Donkey

The first clue to how curare kills came from English experiments in the early nineteenth century. The surgeon Benjamin Brodie found that small animals given curare stopped breathing, but that they survived if their lungs were inflated with bellows until the poison wore off. This was a profound observation. It meant that curare did not destroy life outright; it stopped the muscles of breathing, and if breathing was supplied from outside, the animal could recover.

The naturalist and traveller Charles Waterton, who had collected curare (which he called “wourali”) in Guiana, took part in a famous demonstration in England: a donkey given the poison was kept alive by artificial ventilation through an opening in the windpipe, and it recovered. Raghavendra’s history recounts both the bellows experiments and the donkey; the story became one of the best-known episodes in the early history of the poison.

Why breathing for the animal worked

In modern terms, these experiments showed three things at once. First, curare’s effect was on movement, including the movement of the diaphragm and the chest muscles that draw air into the lungs. Second, the heart was not the first organ to fail, since an animal whose heart had stopped could not have been saved by bellows. Third, the effect was temporary: given enough time, the body cleared the poison and the muscles worked again. These are exactly the properties that, more than a century later, made curare usable in surgery — when the patient’s breathing is supported by the anaesthetist. That later history is told on the legacy page.

What Brodie and Waterton could not say was where in the chain from brain to nerve to muscle the poison acted. That was the question Claude Bernard answered in the 1840s and 1850s, using a frog, a nerve and a muscle. His finding, in brief: the heart kept beating and the muscles still contracted when stimulated directly, but stimulating the motor nerves no longer made the muscles move. The site of the block lay between nerve and muscle. The full account, in Bernard’s own words and with the later debate over his interpretation, is on the poison experiments page.

Back to Table of Contents

6. Quaternary Ammonium and Curare-Like Action

Bernard had located the block, but he had no idea what in curare caused it. The next step came from chemistry, in Edinburgh in the late 1860s. The physician-pharmacologist Thomas Fraser and the chemist Alexander Crum Brown took naturally occurring alkaloids, including strychnine, and converted them chemically into their quaternary ammonium forms — molecules in which a nitrogen atom carries four carbon groups and a permanent positive charge. They found that these modified compounds paralysed animals in the manner of curare, even when the parent alkaloid acted quite differently. Raghavendra records Fraser and Crum Brown’s finding that quaternary ammonium compounds have curare-like activity.

The result is often cited as one of the earliest demonstrations of a structure–activity relationship: the idea that a drug’s effect depends on a definable feature of its chemical shape, so that changing the shape in a known way changes the effect in a predictable way. Strychnine, a convulsant from a Strychnos plant, became a curare-like paralysing agent once its nitrogen was converted to the quaternary form — a chemical bridge between the two opposite poisons of the same plant genus.

Why the charge matters

A permanently charged nitrogen gives a molecule a particular character in the body. Charged molecules cross fatty cell membranes poorly, so quaternary compounds tend to stay outside cells and act on structures at the cell surface — such as the receptors that receive a nerve’s chemical message. The same feature explains why quaternary curare compounds enter the brain poorly, a point later examined in a 1947 study titled “The lack of cerebral effects of d-tubocurarine” (see the legacy page). And the positively charged nitrogen turned out to be the very feature that lets curare compounds fit the receptor normally occupied by acetylcholine, which also carries a quaternary nitrogen — though that link could only be made once acetylcholine’s role at the muscle was known (section 8).

Back to Table of Contents

7. Harold King and Tubocurarine (1935)

For all the experiments done with curare in the nineteenth century, no one had isolated its active principle in pure form from a sample of known kind. That changed in 1935, when the British chemist Harold King, working at the National Institute for Medical Research in London, published “Curare alkaloids. Part I. Tubocurarine” in the Journal of the Chemical Society. Raghavendra notes that King’s starting material was a curare sample held in a museum — a reminder that the collections built up from travellers’ samples over the previous century had become a scientific resource.

The name tubocurarine reflects “tube curare”, the name given to curare shipped in bamboo tubes, the type of sample from which King worked. The “d-” in d-tubocurarine refers to the particular mirror-image form of the molecule found in the plant. With a pure compound in hand, pharmacologists could for the first time give exact, measured amounts of a single substance and compare its effects with those of crude curare, and chemists could begin to work out its structure. King’s paper is the point at which the “molecule last” stage of Lee’s three-step history arrived.

What was still missing

King’s isolation did not by itself settle the plant question. A museum sample of tube curare came with no reliable record of exactly which vines had gone into it. The link between tubocurarine and a named plant species, Chondrodendron tomentosum, was part of the later story of Richard Gill’s expedition and the standardised extracts that followed (section 10).

Back to Table of Contents

8. Acetylcholine at the Motor Nerve Ending

To understand what curare blocks, physiology first had to answer a question Bernard could not: how does a nerve make a muscle contract? In the 1920s Otto Loewi had shown that nerves to the heart act by releasing a chemical substance, and Henry Dale, who had long studied the actions of acetylcholine, led a laboratory in London that went on to ask whether the same chemical carried the signal from ordinary motor nerves to skeletal muscle. The story of Loewi and Dale is told on the Loewi and Dale page.

In 1936 Dale, Wilhelm Feldberg and Marthe Vogt published “Release of acetylcholine at voluntary motor nerve endings” in the Journal of Physiology. The title states the finding: when a motor nerve to a voluntary muscle is stimulated, acetylcholine is released at its endings. This was the chemical messenger in the gap between nerve and muscle — exactly the place where Bernard, ninety years earlier, had located curare’s block. The same year, Dale and Loewi shared the Nobel Prize in Physiology or Medicine for their work on the chemical transmission of nerve impulses.

Joining Bernard to Dale

With this result, the pieces of the curare puzzle fitted together. Bernard had shown that curare stops the nerve’s message from reaching the muscle while leaving both the nerve and the muscle themselves able to work; his successors, led by Alfred Vulpian, had argued that the block lay at the specialised ending where nerve meets muscle, the motor end-plate; and Dale’s group now showed that the message crossing that junction is acetylcholine. Lee’s summary of the modern understanding is that curare acts by blockade of acetylcholine at the neuromuscular junction. The 19th-century debate over Bernard’s own interpretation is covered on the poison experiments page.

Back to Table of Contents

9. How Curare Blocks the Receptor

The neuromuscular junction is the point where a motor nerve fibre ends on a muscle fibre (the site is drawn and animated on the site’s neuromuscular junction page). When a nerve impulse arrives, the nerve ending releases acetylcholine into a very narrow gap. On the far side, the muscle surface carries nicotinic acetylcholine receptors — protein channels that open when acetylcholine binds to them. The open channels let charged particles flow into the muscle fibre, the fibre is electrically excited, and it contracts. An enzyme, acetylcholinesterase, then breaks the acetylcholine down within a few thousandths of a second, so the signal is short and precise.

Tubocurarine works by occupying the same binding sites on the receptor without opening the channel. It is a competitive antagonist: it competes with acetylcholine for the receptor, and while it is bound, acetylcholine cannot act there. When enough receptors are occupied, the nerve still fires and still releases acetylcholine, but too few channels open to excite the muscle, so the muscle does not contract. W. C. Bowman, reviewing the history of neuromuscular block in the British Journal of Pharmacology in 2006, describes the nondepolarising blockers of which tubocurarine was the first as reversible antagonists of acetylcholine at its receptor, and notes that their site of action was first determined by Claude Bernard in the mid-nineteenth century.

What this explains about the old observations

• Bernard’s frog. The nerve still conducted and the muscle still responded to direct stimulation, because neither was damaged; only the chemical handover between them was blocked.
• Brodie’s bellows and Waterton’s donkey. The muscles of breathing are skeletal muscles worked through the same kind of junction, so they fail with the rest; the heart, driven by its own pacemaker, keeps beating. Supplying breath from outside bridges the time until the poison is cleared.
• Reversibility. Because the drug binds and unbinds rather than destroying the receptor, its effect fades as it is removed from the body, and raising the amount of acetylcholine at the junction can tip the competition back. Bowman’s review also describes newer drugs of the class, such as rocuronium, and a cyclodextrin molecule used to reverse its block.

The same nicotinic receptor at the muscle surface is the target of the antibodies found in most people with myasthenia gravis, a disease of fluctuating muscle weakness; the history of how that disease came to be understood runs alongside curare’s (see Myasthenia Gravis: History and Discovery).

Back to Table of Contents

10. Richard Gill’s 1938 Expedition

By the late 1930s the chemistry and physiology of curare were in place, but medicine still lacked a reliable supply. Crude curare samples varied from batch to batch, and without a known plant source and a consistent product, clinicians had no way of knowing what strength they were dealing with. The American explorer Richard Gill set out to close that gap. According to N. Elsherbini and S. B. Backman, who in 2024 published a study of previously unknown correspondence between Gill and the Montreal anaesthetist Harold Griffith, Gill returned from his 1938 Amazon expedition with more than 11 kilograms of curare.

From that material, Chondrodendron tomentosum was identified as a plant source of curare, and a standardised curare extract was produced by a drug company — the product that made consistent clinical use possible. It was a standardised extract of this kind that Griffith and his resident Enid Johnson used in Montreal in January 1942, the first use of curare to relax the muscles during general anaesthesia, published that year in Anesthesiology. That turning point, and what followed from it, are told on the legacy page.

The dispute over the source

The identification was not the end of the argument. Elsherbini and Backman report that Gill himself disputed the claim that curare derives solely from Chondrodendron tomentosum. The disagreement echoes the whole botanical history of curare described in section 3: different peoples used different plants, a single poison could combine several, and a laboratory identification based on one batch could not speak for every curare in South America.

The dispute did not slow the medical story, because what clinicians needed was a consistent product of known strength, not a complete botany of every curare. Within a few years of Gill’s expedition, purified tubocurarine and then fully synthetic drugs — among them gallamine, made in the laboratory of Daniel Bovet — carried the action of the forest vine into operating rooms around the world, completing the path from arrow poison to medicine that Lee describes.

Back to Table of Contents

Key Research Papers

1. Lee MR. Curare: the South American arrow poison. J R Coll Physicians Edinb. 2005;35(1):83-92. PubMed PMID: 15825249
2. Raghavendra T. Neuromuscular blocking drugs: discovery and development. J R Soc Med. 2002;95(7):363-7. PubMed PMID: 12091515
3. Bowman WC. Neuromuscular block. Br J Pharmacol. 2006;147 Suppl 1(Suppl 1):S277-86. PubMed PMID: 16402115
4. King H. Curare alkaloids. Part I. Tubocurarine. J Chem Soc. 1935:1381. DOI: 10.1039/jr9350001381
5. Dale HH, Feldberg W, Vogt M. Release of acetylcholine at voluntary motor nerve endings. J Physiol. 1936;86(4):353-80. PubMed PMID: 16994763
6. Elsherbini N, Backman SB. The connection between Dr. Harold Griffith and Richard Gill: new insights into the history of curare use in anesthesia from previously unknown correspondence. Can J Anaesth. 2024;71(12):1664-1671. PubMed PMID: 38653929
7. Black J. Claude bernard on the action of curare. BMJ. 1999;319(7210):622. PubMed PMID: 10473481
8. Cousin MT. Vulpian and not Claude Bernard first proposed the hypothesis of the motor end-plate as the site of action of curare. Anesthesiology. 2002;97(2):527-8. PubMed PMID: 12151956
9. Barbara JG. [Claude Bernard and his successors on curare: epistemological questions at stake]. J Soc Biol. 2009;203(3):227-34. PubMed PMID: 19833069
10. Griffith HR, Johnson GE. The use of curare in general anesthesia. Anesthesiology. 1942;3(4):418-420. DOI: 10.1097/00000542-194207000-00006
11. Smith SM, Brown HO. The lack of cerebral effects of d-tubocurarine. Anesthesiology. 1947;8(1):1-14. PubMed PMID: 20281549

PubMed Topic Searches

1. PubMed: curare history
2. PubMed: Chondrodendron tomentosum
3. PubMed: Strychnos curare alkaloids
4. PubMed: tubocurarine and the nicotinic acetylcholine receptor
5. PubMed: history of neuromuscular blocking agents

Further Reading

• Bernard C. An Introduction to the Study of Experimental Medicine, translated by H. C. Greene (1927; Dover reprint 1957). Full text at archive.org
• Bernard C. Leçons sur les effets des substances toxiques et médicamenteuses. Paris: Baillière; 1857.
• Bernard C. “Du curare.” Revue des Deux Mondes, 1 September 1864.

Back to Table of Contents

Connections

• Claude Bernard: Curare, Carbon Monoxide and the Milieu Intérieur
• Claude Bernard: Life and Career
• Curare and Carbon Monoxide: Claude Bernard’s Poison Experiments
• Claude Bernard’s Legacy: Milieu Intérieur, Experimental Medicine and Curare in Anaesthesia
• Daniel Bovet: Antihistamines, Curare, and Blocking the Body’s Own Signals
• Otto Loewi & Henry Dale: Acetylcholine and the Chemistry of Nerve Signals
• Cinchona and the Plant Alkaloids of Pelletier and Caventou
• Paracelsus: The Dose Makes the Poison
• The Neuromuscular Junction: How a Nerve Commands a Muscle
• Myasthenia Gravis
• Myasthenia Gravis: History and Discovery
• Pharmacology

Back to Table of Contents

Featured Videos