Greek fire allowed Byzantine warships to send burning liquid toward enemy vessels through tubes at their bows. It could keep burning when the liquid landed on the sea, spreading flame between ships at close quarters. On a wooden galley crowded with rowers and fighting men, the flames could injure a crew and disrupt its ability to manoeuvre or defend itself.

The weapon appeared in the seventh century, as the eastern Roman Empire fought to defend its remaining territories against the expanding Arab caliphate. Its own writers used names such as liquid fire, sea fire and Roman fire. They considered themselves Romans, although outsiders increasingly described the Greek-speaking empire as Greek. The name ‘Greek fire’ reflects that outside usage.

No authenticated recipe has reached us. Accounts of campaigns, imperial instructions and military equipment nevertheless tell us how the Byzantines used the fire and why they guarded it. Those accounts also require caution when later stories claim that the liquid ignited on contact with water, burned underwater or could never be extinguished.

A weapon born from the struggle for Constantinople

The eastern Roman Empire still drew on Roman naval traditions when its position in the Mediterranean deteriorated during the seventh century. Muslim armies took Syria, Palestine and Egypt, formerly important Roman provinces, and the caliphate began deploying powerful fleets in the eastern Mediterranean. Constantinople stood behind formidable walls beside the Bosporus, the narrow waterway joining the Black Sea to the Sea of Marmara. The walls could not by themselves protect the harbours or the supply ships entering the capital. The empire needed a way to fight them in confined waters.

Using machinery to direct fire at an enemy was not new. At Delium in 424 BCE, the Boeotians attacked an Athenian wooden fortification with a hollow beam attached to bellows and a cauldron of burning fuel. Thucydides describes the device in his history. Air driven through the apparatus fed a mixture of charcoal, sulphur and pitch. The Boeotian machine projected flame through a tube centuries before the Byzantines developed their shipboard system for directing burning liquid. It did not propel liquid fuel from a warship.

Much of the story of Greek fire’s beginnings comes from Theophanes the Confessor. Writing as a monk in the early ninth century, he described campaigns that had taken place more than a hundred years earlier and credited Kallinikos, an architect from Heliopolis in Syria, with preparing a naval incendiary. Heliopolis is Baalbek in present-day Lebanon. In Theophanes’s account, Kallinikos fled to the Roman Empire, and the fire helped destroy Arab ships in a Byzantine victory. The Prosopography of the Byzantine Empire identifies the testimony and the later tradition surrounding it. This testimony does not establish that Kallinikos invented every component of the fuel and its delivery equipment.

Byzantine dromons projecting Greek fire at Arab ships outside Constantinople
Byzantine dromons blast an Arab fleet with flame before Constantinople’s sea walls; manuscript illumination. Source: Wikimedia Commons

Greek fire’s supposed debut is commonly placed during the Arab siege of Constantinople dated 674 to 678. The dates and nature of this siege remain disputed. Theophanes compiled his account long after the events, and his entries combine details that are difficult to reconcile chronologically. Historian Marek Jankowiak, in a study of the first Arab attempt on the capital, instead argues for a major Arab campaign in 667 to 669. The evidence associates naval incendiaries with the Byzantine struggle against Arab fleets in the later seventh century. It does not establish a first demonstration in 672 or allow each engagement to be reconstructed in detail.

Theophanes himself complicates the inventor story by mentioning ships fitted with fire-carrying devices before he introduces Kallinikos. He may have combined separate incidents, mentioned an older incendiary device or arranged the account around a memorable name. We cannot determine which explanation is right. Kallinikos remains central to the tradition, although we have no reliable description of his workshop, assistants or exact contribution.

What was Greek fire made of?

A petroleum product is the leading candidate for the principal fuel. Light petroleum, often called naphtha, can travel through pipes, float on water and burn while it remains on the surface. Resin or pitch might have helped the liquid cling to a ship’s timber or sails. Neither proposed additive can be confirmed from a Byzantine recipe. No securely authenticated Byzantine formula survives for the substance discharged from naval siphons, and online lists of five or six ingredients are modern reconstructions, not recovered seventh-century formulas.

The empire had access to regions where petroleum could be obtained. Constantine VII, writing in the tenth century, described petroleum sources near the northeastern coast of the Black Sea. A 2019 archaeological study by Valeriy Naumenko examines extraction and transport evidence around the Kerch Strait, including tall-necked jugs associated with the movement of petroleum. The finds help establish a plausible supply context for petroleum. It does not reveal the contents of a sealed Greek-fire container or establish that oil from one particular source reached a named imperial naval workshop.

Some proposed recipes add quicklime on the grounds that it releases heat when water is added. Quicklime does behave that way, but its chemistry does not demonstrate that Byzantine siphons used it or that Greek fire ignited spontaneously when it struck the sea. Sulphur, saltpetre, calcium phosphide and several other substances also appear in modern theories. Knowledge of these substances does not establish that Byzantine incendiary makers combined them. Describing the weapon as a gunpowder-like explosive creates a further problem by projecting later technology back into the seventh century.

The claim that Greek fire ‘burned on water’ is easier to explain than the claim that water itself caught fire. Petroleum can remain afloat and burn above the surface. To sailors watching flame spread across the water between ships, that difference would hardly have made the attack less frightening. Medieval descriptions of flames supposedly intensified by water may preserve an experience of the weapon, exaggerate it or confuse different incendiaries. They cannot settle its chemical composition.

Maciej Kokoszko and Mirosław Leszka, in a peer-reviewed discussion of Byzantine liquid fire and its chemical uncertainties, distinguish the liquid from the machinery used to deliver it. A usable shipboard weapon depended on the fuel and on reliable equipment, trained crews and regular supplies. Secrecy may have protected this working system as much as any individual ingredient. With no surviving complete naval siphon and no securely identified sample of its fuel, modern formulas remain hypotheses.

Advertisements

Constantine VII and the secret nobody was meant to share

Constantine VII addressed this secrecy directly in the mid-tenth century. In the manual he composed for his son Romanos, known by its Latin title De Administrando Imperio, he explained how an emperor should respond if foreign rulers asked for liquid fire or its projectors. He offered no chemical instructions. He gave Romanos a sacred origin story that justified refusing the request.

Constantine VII crowned in a court ceremony symbolizing imperial guardianship of Greek fire
Coronation of Constantine VII Porphyrogennetos before court and clergy Source: Wikimedia Commons

In the version Romanos was instructed to tell, an angel had given the secret to Constantine the Great, the first Christian Roman emperor, on the condition that it be manufactured in the imperial city alone and never shared with foreigners. The surviving Greek text in chapter 13 threatens anyone who breaks that command with religious condemnation and recounts divine punishment of a man who revealed it. The account shows how the court justified secrecy; it provides no historical evidence that an angel supplied a military formula in the fourth century.

Constantine VII’s story cannot be reconciled literally with the attribution to Kallinikos in the seventh century. The emperor’s purpose was different from that of a historian investigating invention. He was advising his heir how to turn down requests from other rulers. By representing liquid fire as a sacred inheritance, he made disclosure a breach of duty. The chapter uses comparable arguments when discussing other objects that foreigners might demand, among them imperial ceremonial possessions.

Even a government determined to conceal the ingredients would have needed people to obtain fuel, make and repair vessels and tubing, and teach crews to use them. Taking possession of a projector would not necessarily give an enemy that knowledge. A study of the Bulgarians’ encounter with naval fire discusses how Khan Krum captured thirty-six bronze siphons and stocks of fire after taking Mesembria in 812. Historians disagree over whether the Bulgarians subsequently learned to make or deploy the weapon. Possessing a captured siphon was different from knowing how to manufacture the fuel, maintain the equipment and train crews to use it.

How Byzantine ships fired it

At sea the fire was principally associated with the dromon, a galley driven by both oars and sails. Larger tenth-century dromons carried rowers, sailors and troops whose actions had to be closely coordinated in battle. The ships developed out of older Roman Mediterranean naval practice; they were not invented solely to carry the incendiary. Their most distinctive equipment was a siphon, generally positioned toward the bow, through which a crew could direct burning liquid at an opposing hull.

Bow siphons project liquid fire along the waterline to cripple ships before boarding
Dromons firing bow mounted siphons into an approaching fleet Source: Wikimedia Commons

Exactly how the projector worked remains unknown. Texts mention siphons and later images show flame issuing from tubes, but archaeologists have never recovered a complete shipboard Greek-fire apparatus. Engineers have proposed systems combining a fuel reservoir, a pump to generate pressure, pipes and a nozzle where the liquid was ignited. Some designs require the petroleum to be heated before release. Such reconstructions test what the descriptions permit. No Byzantine construction instructions survive. Claims about the precise arrangement of valves, a corrosion-resistant bronze nozzle or one particular pressure vessel cannot be verified.

The major study of Byzantine naval equipment by John Pryor and Elizabeth Jeffreys, The Age of the Dromon, examines naval terminology and tactics alongside the military manuals, images and possible mechanical designs. A published scholarly review notes the authors’ caution about treating prescriptive military texts as straightforward descriptions of battles. Manuals sometimes reused earlier material, and instructions for ideal conduct did not always reflect what a fleet could achieve at sea. An emperor’s prescription for naval fighting is different evidence from a combatant’s account of a battle.

A stream of burning liquid offered its greatest advantage at close range, where opposing galleys were already vulnerable to missiles, grappling and boarding. It could set fittings or rigging alight, drive defenders away from part of a ship or interrupt the rowers just as the vessel needed to manoeuvre. The flames threatened timber and tarred materials, while the tightly packed crew had little space to escape. Byzantine vessels still needed archers, spearmen and boarding troops. The siphon worked alongside the weapons already used in galley combat.

Distance, wind and sea conditions limited its usefulness. A moving galley was not a stable platform from which to aim a short-range jet, and a badly directed flame might threaten the ship releasing it. An enemy could try to avoid the bow, move beyond the weapon’s reach or protect exposed surfaces. Success still depended on seamanship, fleet size and command decisions. Its presence alone could not secure victory.

Advertisements

Leo VI, naval manuals and the hand siphon

Leo VI, emperor from 886 to 912, also wrote about naval incendiaries. His Tactica, composed around the turn of the tenth century, includes naval instructions commonly discussed under the title Naumachika. Fire weapons appear there among many other forms of equipment and manoeuvre. The passages show that Byzantine commanders continued to consider them useful long after the period associated with Kallinikos. The instructions record recommended tactics rather than eyewitness accounts of their use.

Leo’s treatise and later texts mention the cheirosiphon, a name meaning ‘hand siphon’. Associated with siege warfare, fortifications and fighting on land, it seems to have been more portable than the large naval projector. Some manuscript drawings depict men directing flames from elevated structures. The word ‘hand’ alone cannot establish that one soldier could carry a scaled-down naval siphon complete with its pump and reservoir. Scholars have suggested different operating principles, and medieval military illustrations are not reliable engineering plans.

Portable cheirosiphon spraying Greek fire from a tower during a siege
Soldiers operating a hand siphon from a siege tower against defenders Source: Wikimedia Commons

Fire could be used against the timber of a siege tower, a gate, a protective screen or a shield, and defenders could also burn machines advanced toward their walls. The Byzantines employed containers of combustible substances as well as devices that poured or hurled fire. Sources did not always use consistent names for these materials. A pot described by a later writer or a modern museum as containing ‘Greek fire’ need not have held the same liquid that the imperial navy pumped through a bow siphon.

Material evidence helps distinguish the wider use of incendiaries from the particular Byzantine naval weapon. Peter Pentz’s study of a medieval workshop at Hama in Syria discusses a workshop linked to the manufacture of ceramic fire containers. The Hama workshop illuminates medieval incendiary production but provides no recipe for the Byzantine naval siphon fuel.

What happened when Greek fire was used in battle?

Despite the doubts about its first appearance, naval fire remained in use after the seventh century. Accounts of the Arab siege of Constantinople in 717 to 718 describe Byzantine attacks against ships supporting the besiegers. The incendiary helped the defenders contest the waters around the capital. Other factors also shaped the siege’s failure: the city had strong land defences, besieging forces faced supply difficulties and severe weather, and fighting elsewhere affected their position. The incendiary contributed to the defence but cannot explain the siege’s failure by itself.

In 941, a Rus fleet attacked Constantinople under Romanos I Lekapenos. The Rus were connected to trade and political networks along the rivers between northern Europe and the Black Sea, with centres including Kyiv. Liutprand of Cremona, a western diplomat, later described Byzantine fire driving attackers from their boats and into the water. His story conveys the danger the weapon posed to men aboard smaller vessels, although the dramatic detail must be assessed as literary testimony. Liutprand also reports the use of Greek fire in western Mediterranean diplomacy during roughly the same period.

Byzantine fleets continued to carry incendiaries during the tenth and eleventh centuries; sources mention fire used again against the Rus in 1043. A well-known image in the twelfth-century Madrid Skylitzes shows a Byzantine warship sending flame toward another vessel during the revolt of Thomas the Slav, fought from 821 to 823. The illumination records a twelfth-century Byzantine representation of an earlier naval battle. It cannot establish the exact fittings of a ninth-century ship or show a moment witnessed by the artist.

Anna Komnene gives a particularly detailed account of the equipment in action. Writing in the twelfth century about the reign of her father, Alexios I, she describes a battle against Pisans near Rhodes late in the eleventh century. Metal animal heads decorated the ships’ bows, and tubes carried fire through the creatures’ mouths. In Book XI of the Alexiad, Anna explains that the stream could travel sideways or downward instead of merely rising with the flame. Her account includes a first shot that failed, a later hit and the arrival of a storm. The missed shot and the storm show the problems of operating the weapon under battle conditions.

Anna’s description places the siphons amid actual manoeuvring and missed shots, rather than presenting them as infallible machines. It still tells us nothing definite about the fuel’s ingredients. Greek fire did not give Byzantium an enduring monopoly on naval victories. Other powers developed fleets capable of disputing sea routes and defeating Byzantine ships, even when the empire retained incendiary equipment.

Advertisements

How did the Byzantines lose Greek fire?

It is difficult to identify when the original naval siphon system disappeared. As Byzantine sea power changed, references to fire weapons became harder to interpret, partly because writers used similar names for different substances. The capture and sack of Constantinople by the Fourth Crusade in 1204 devastated the empire’s institutions and resources, including those that supported its navy. Yet no surviving document says that the last makers of Greek fire died in the sack or that its final recipe was destroyed then. The sack alone cannot date the end of the weapon.

Fires and smoke as Latin crusaders seize Constantinople in 1204
Crusaders enter burning Constantinople during the 1204 sack Source: Wikimedia Commons

Popular stories sometimes explain the loss by claiming that a hereditary guild guarded the recipe until its members were killed. No securely identified source establishes that guild or the supposed sequence of events. Restricted groups may have transmitted specialist knowledge, but we cannot demonstrate a family monopoly lasting four centuries. Nor can we establish that the formula was never written down. What survives contains no complete, authenticated recipe for the shipboard Byzantine incendiary.

The name itself makes the later record difficult to follow. Writers in medieval Latin and Arabic could call several incendiary substances ‘Greek fire’, including liquids thrown in pots instead of pumped from ships. A late reference to the name is therefore insufficient proof that an imperial naval siphon or its original fuel was still being used. Silence in the sources, on the other hand, cannot prove that all similar incendiary methods had ceased. The question concerns the fate of a specific naval system, not the end of using burning oil in warfare.

Modern researchers have tested how a petroleum fuel might be projected and ignited. John Haldon and Maurice Byrne’s 1977 technical study brought engineering analysis to the question, and subsequent reconstructions have examined pumps and the behaviour of heated fuel. The experiments establish that some proposed mechanisms can work. Even a working device that shoots a powerful jet cannot prove that Byzantine craftsmen used the same mixture, pressure, temperature or precautions.

The Byzantines developed and maintained an effective shipboard fire weapon, used it over several centuries and treated its operation as an imperial secret. Naval accounts and military manuals make those points clear. They leave other questions open: what precisely went into the fuel, how its earliest projectors worked, how much Kallinikos contributed and when the original system ceased to be made. The written evidence explains why Byzantine commanders valued the weapon and why their enemies feared it, while leaving its manufacture and mechanism beyond secure reconstruction.