The Mystery of Greek Fire: The Byzantine Medieval Flamethrower

You almost stumble on the last step. The dekarchos hisses, ‘You there – watch your feet. One false move and we’re all toast.’ One of the siphōnarioi had brought a torch to light your way across the fortifications, but the officer grabbed it, cursing and hurriedly smothered it. Before the thick darkness fell, you manage to snatch a brief look into the cellar – dozens, maybe hundreds of sealed amphorae like the one you are carrying, stacked in ranks and packed with rags to keep them in place. And something else too, toward the corner at the back – the strange, snake-like coil with a big round belly, the terrifying monstrous mouth of shining brass, and the large handle like that of a well-pump. Though you were plunged into darkness, the smell remained – indeed, it swallowed you in a sickly-sweet stench of pitch, olive oil and sulphur.

Hurriedly setting the final amphora into place with its others, grateful to be rid of its sloshing weight, you scramble back up the uneven steps onto the quayside. You can see them, off in the darkness of the strait, their lanterns bobbing as they blockade well outside arrow-shot – dozens of Arab ships, dispatched by the Caliph. Tomorrow, the four-year siege ends. You’ll be on a many-oared Byzantine dromōn, maybe even the one being fettled right here on the dockside, at your station in the bows, working the infernal machine that spews liquid death upon the enemy. You’ll have to recall every ounce of your training to stay level-headed amidst the carnage – you shudder at the thought of what this new weapon could do to another person. Come the morrow, you and the dekarchos will either be at the bottom of the Bosporus and the Arabs will be dining in Constantine’s palace – or you’ll be legend, the saviours of New Rome, the secret wielders of Greek fire, though none will know your names.

The first Muslim siege of Constantinople lasted four years. The Umayyad Caliphate’s armies and navies established permanent bases on mainland Greece and the outlying islands, constricting the Eastern Roman Imperial capital tighter and tighter from 674 CE until the spring of 678. The future of Byzantium looked bleak. Yet the decisive factor in the final showdown between the Arab and Byzantine navies was a terrifying new weapon: the newly-perfected Greek fire devastated the Arabian fleet, burning ships and sailors alike, setting the waters of the bay itself alight. Greek fire has subsequently entered the annals of legend, but its becoming a subject of cult fascination for later Europeans, and spawning multiple depictions in print and on the silver screen. But was Greek fire all that it has become in the telling? Many even ask, was Greek fire real? Here, we’ll do our best to sift the myth from reality in the hidden history of Greek fire.

The Genesis of Marine Fire

Combing through historical sources of 7^th century Byzantium, most historians agree that the Byzantine navy acquired a new type of chemical weapon that was first deployed at, or shortly before, the naval battles that ended the first Arab siege of Constantinople. The Byzantine historian Theophanes ascribes this weapon to the invention of a Byzantine-Jewish chemist named Kallinikos. Fleshing out the story a century and a half later, Emperor Constantine VII Porphyrogenitus writes in his manual of statecraft that Kallinikos fled the Arab invasion of his home at Heliopolis in the Byzantine Levant (modern Baalbek, Lebanon), bringing with him to the capital the result of his life’s endeavours: a substance called pyr thalassion (‘marine fire’). This substance could be ejected through a nozzle (siphōn) as a projectile weapon, and, in the words of Theophanes, was used to great effect to ‘set fire to the boats of the Arabs, and burned these with their men aboard’.

After the invention of Greek fire and its use against the Arab navies in 678 CE, it became a staple of Byzantine naval warfare for a critical generation. It played another critical role at the second Arab siege of Constantinople in 718 CE, the final concerted attempt for an Islamic Caliphate for 800 years. The Muslim navies again attempted to blockade the City, but this time were much more cautious, wary of the new pyr thalassion – but defecting Egyptian Christians who had been pressed into service in the Islamic navy gave Emperor Leo III vital information about the disposition of the ships. Lacking crew and revealed by the Egyptians, the Muslim fleet was burned to cinders with Greek fire. It appears throughout the 9^th and 10^th centuries, recorded as being used against Islamic raiding fleets in the Eastern Mediterranean, and is frequently wielded by the Constantinopolitan central government against various internal rebel fleets.

Greek Fire Within And Without

The Rus’ occasional raids into Byzantine territory through the late 10^th and 11^th centuries, often undertaken to improve the diplomatic position of the Kievan Grand Princes, were often met with Greek fire. Liutprand of Cremona records that in 941 CE, a huge fleet of 1,000 Rus’ ships were driven off by only 16 Byzantine dromōns crewed by the fearsome siphōnarioi – we can likely adjust the numbers somewhat to account for exaggeration, but the contemporary Russian Primary Chronicle speaks of a ‘winged fire, and by means of a certain tubing the Greek generals flung the flames onto the Russian warships’, with the few remaining Rus’ survivors returning home to tell their countrymen that ‘the Greeks have a fire which flies through the air like lightning’. Thirty years later, the invasion of Grand Prince Sviatoslav into Bulgarian territory was cut off by a Byzantine fleet bearing Greek fire which sailed up the Danube, forcing the Kievan Prince to seek terms with the Empire as arbitrator. Later, Greek fire is also recorded being used on land, for example against the Norman invasion of the Balkans in the early 12th century.

The Byzantines were not above using Greek fire on their fellow Christians: it is recorded as being used by the Constantinopolitan navy against Pisan opportunist pirates, who raided the Greek islands during the First Crusade. As well, it was deployed in the context of internal political strife: for example, the loyal Imperial navy routed the Thematic (provincial) navies with the use of Greek fire during their rebellion in 727 CE, and it was also deployed during the serious revolt led by Thomas the Slav in 821. However, by the end of the 12^th century, Greek fire appears to have been lost: the Angeloi Emperors (1185-1204) make no recorded use of it against their many internal enemies, and it makes no appearance during the Constantinopolitan resistance to the farcical Fourth Crusade. After the conquest of Byzantium by the Crusaders and Venetians, Greek fire disappears from the historical record.

Was Greek Fire Real?

As with any investigation of a historical phenomenon as captivating and attractive as Greek fire, we have to start with a solemn assessment of our sources and their biases. As to, ‘was Greek fire real’, there can be little doubt that Greek fire existed. It is referenced consistently by Byzantine historians, military leaders who themselves deployed it, and by foreign observers (sometimes even those on receiving end, in the case of the Rus’ Primary Chronicle). However, beyond the fact that the Byzantine navy had apparently exclusive access to the use of an effective incendiary anti-ship weapon, we’re left to make educated guesses as to the genesis, use and even the possible chemical makeup of Greek fire.

How Did It Work?

From a study of the sources, we can identify a common set of characteristics for Greek fire, which remain more or less consistent over the four centuries of its use. Sources agree that it is a liquid projected from a siphon, as opposed to a projectile thrown from a catapult. But most intriguingly are the persistent observations that it burned on water, perhaps even being ignited by contact with water. Either way, most observers comment that it was extremely difficult to extinguish by usual methods, requiring quenching with sand (another method used by Muslim armadas was strong vinegar). This seemingly miraculous set of properties baffled contemporaries, and forces modern comparisons with napalm and white phosphorus.

Many long hours of fruitless study have gone into attempting to reconstruct, and even reproduce, Greek fire. As yet, nobody has sufficiently recreated an effective flame-throwing weapon from the technology and materials available to Medieval Byzantines – which is an enormous testament to the ingenuity of Constantinople’s chemists. Based on contemporary descriptions, Greek fire was a complete weapons system, comprising of a number of different carefully-designed components: a vessel in which to heat the Greek fire, a pressurizing system, a nozzle to direct the jet of flame, and the composition of the substance itself.

Perhaps naturally, the latter component has been the subject of most attention. Tantalizingly (and against the advice of her august progenitor Constantine Porphyrogenitus), the historian and princess Anna Komnene gives what is widely regarded as a partial recipe for Greek fire:

This fire is made by the following arts: From the pine and certain such evergreen trees, inflammable resin is collected. This is rubbed with sulfur and put into tubes of reed, and is blown by men using it with violent and continuous breath. Then in this manner it meets the fire on the tip and catches light and falls like a fiery whirlwind on the faces of the enemies.

Anna relates this within the context of the Sicilian-Norman invasion of the Western Balkans, where the defenders of the Byzantine garrison town of Dyrrhachium used Greek fire upon the invaders. Whilst her description is infuriatingly (perhaps deliberately?) vague and is clearly missing some key components, this has given practical archaeologists (ie. pyromaniacs with an excuse) something to go on. Gunpowder or saltpetre is fairly easy to rule out as a primary ingredient: it was not synthesized in the West for centuries, and was originally the province of Islamic scholars in contact with India and China rather than the Byzantines. Quicklime was a substance that was already used sporadically in naval warfare, and it reacts powerfully with water – but practical experiments failed to produce anything like the destructive power reported by eyewitnesses.

The scholarly consensus has coalesced around a form of refined pitch or even petroleum. In his writings, Emperor Constantine VII Porphyrogenitus specifically highlights known strategic reserves of crude oil within the Empire – and although he does not explicitly link these reserves with Greek fire, this might provide a clue to its composition. Naptha-based chemical weaponry, usually grenade-like hand weapons, were developed within the Islamic world in the following centuries, demonstrating that crude oil-based weapons were possible in the period. Combined with Anna Komnene’s suggestion of pine resins and other flammable ingredients, this might well create a thick, sticky, oily substance that would float upon water, and would be very difficult to extinguish.

The other critical component is Greek fire’s delivery system. Multiple sources, including, amongst things, a Viking saga, attest that the siphōn included a furnace with a copper vessel which likely heated the Greek fire to make it volatile. As well as the siphon-head itself, hand-pumps and bellows also appear in the literature. In 1977, Byzantinists John Haldon and Maurice Byrne attempted to draw all of these together to produce a theoretical model for use in creating a working reproduction of a siphōn. Overcoming many problems, for example with the pressurization of the copper vessel, Haldon eventually produced a working example for the 2002 TV series ‘Machines Time Forgot’ – it had an effective range of 15 metres, and produced flames that were measures at over 1,000°C. Could such a design be turned against a ship so that it ‘burned up in a short time so that all of it became white ashes’, as the Viking saga says?

Greek Fire Mythbusters

To fully appreciate the significance of Greek fire, we have to burst a few misconceptions, peddled by credulous Victorian historians and sometimes uncritically reproduced by pop history and sensationalized modern media. Firstly, the supposed inventor of Greek fire, the chemist Kallinikos, is only attested by sources many years later. Theophanes, the Byzantine historian, was working from primary sources which are lost to us – but he was writing at least almost a century-and-a-half after the first Arab siege, in the 810s CE. The story of the individual genius chemist is attractive, but probably not accurate. Roman military doctrine had included earlier chemical weaponry that predated Greek fire – even as far back as the 5^th century BCE, flame-breathing siege equipment was recorded in Greece (or at least imagined by historians), and by the 6^th century CE, Byzantine Emperors were using sulphur as an anti-ship weapon. It seems likely that Greek fire, rather than being a sky-blue invention by a single genius, was a significant but incremental improvement on these earlier weapons, likely by the Imperial alchemists working in Constantinople.

Even if we conceive of Greek fire as an incremental improvement, this does little to diminish its clear importance at a critical time of need for the Byzantine Empire. The late 7^th century was a time of severe political crisis for the Eastern Romans, with the new Islamic Caliphate expelling the Byzantine authorities from Egypt, North Africa and the Levant, and there was a very real possibility that the Caliphs would not stop there. But the invention of Greek fire gave the Byzantines a critical advantage at sea. Whilst the Theodosian land walls were nearly impossible to storm, Constantinople was vulnerable to being cut off if the Bosporus and the Sea of Marmara were blockaded – but the Imperial fleet of oar-driven dromōns armed with vital Greek fire meant that hostile occupying fleets could be driven away or burned en-masse. Clearly, Greek fire alone was not solely responsible for the survival of the Empire: domestic political fragmentation within the Islamic world in the form of the First Fitna gave the Byzantines a lifeline, after which the ability for the Caliphate to project its power outward in a systematic fashion declined (especially so after the rise of the Buyids and the Seljuk Turks). But we should equally not underestimate its impact either: Greek fire provided a qualitative superiority in a critical theater of conflict at a time when the Byzantines needed everything they could get.

A Secret Well Kept

Another fascinating aspect of Greek fire is that it is an early example of a successfully-kept state secret. No other people of the Eastern Mediterranean managed to successfully reproduce Greek fire, at least for hundreds of years after its invention. Byzantine writers explain this in multiple different ways: for example, Constantine VII Porphyrogenitus states that the secret of Greek fire was delivered directly from the heavens by an angel sent to Constantine the Great, the Eastern Roman Empire’s first Christian Emperor, retelling of how those who divulged the secret were struck by holy lightning. George Kedrenos, writing in the 1050s CE (four centuries after the first use of Greek fire) records that the secret of the fire was kept by the descendants of the mythic chemist Kallinikos, a sort of secret society of Byzantine alchemists called the Lampros (‘radiant ones’).

Neither of these satisfy we sober-minded historians, who tend to look at the relative capabilities of the other Eastern Mediterranean states who encountered Greek fire. Even though several of them seem to have captured either an intact marine-fire dromōn (as the Muslims did in 827 CE) or even a number of siphōns and a quantity of the substance itself (as the Bulgars did in 816 CE), none could effectively bring it to the field. The answer is probably a much less narratively satisfying combination of factors: the recipe was probably well compartmentalized (ie. that no one person knew the whole of its production and use), that the siphōnarioi who operated the weapon were highly-trained and could not be easily replicated, that the machinery required to operate the weapon was produced by highly specialist Constantinopolitan craftsmen whose workshops could not be reproduced elsewhere, and so on.

As well, Greek fire was a critical political tool, giving the Emperors an extra shock weapon in their arsenal against internal opponents. This is inherently linked to its secrecy, which Emperor Constantine VII Porphyrogenitus appears to have been very aware of. His work De Administrando Imperio (‘On the Government of the Empire’) takes the form of an extended letter to his son Romanus, comprising a comprehensive manual of statecraft. He devotes many dozens of lines to his fire-and-brimstone condemnation of those who would suggest divulging the secret of Greek fire. Beneath the layers of Christian imagery, we can see a solid piece of political advice to his son: Greek fire has saved both the Empire, and the Emperor, from external and internal threats; do not risk giving away the ace up our sleeve. The revolts in which loyal Imperial navies used Greek fire against their internal opponents reveals another dimension of this weapon: that it was specifically Constantinopolitan, always in the employ of the Emperors who controlled the capital, rather than ‘the Greeks’ more generally. Constantine himself deployed Greek fire on multiple occasions, such as the destruction of the Arab fleets based at Crete in 957 CE.

Never To Be Told

So – we now understand what Greek fire was – and what it was not. Was Greek fire real? Yes, it definitely was. But why did it disappear? The critical event which seems to mark the end of the era of Greek fire was the reign of the Angeloi Emperors at the close of the 12^th century. Coming on the back of the fragmentation introduced by the Komnenian restoration, the reign of the Angeloi was a parlous period for the Byzantine state: marked by spectacular levels of corruption, the military affairs of the Byzantine state were badly neglected, and in particular the navy was sold off, according to one contemporary observer, ‘down to the very nails’. If, as seems likely, Greek fire was a largely naval affair, this naval neglect would have badly effected the continuity of Greek fire, and in this environment it would have been much more likely for the compartmentalized, highly-regulated secret to be lost. After the Latins were finally expelled from Constantinople almost sixty years later, the Byzantine restoration was a fractured mess, with large swathes of the Empire never fully reintegrated, and constantly losing ground to the rising Ottoman Sultanate. If Byzantium depended on strategic oil reserves to produce Greek fire, then it was even less likely to have access to them in the late 13^th century and beyond. Thus, the secret slipped into obscurity, and became only a myth, surviving only in the apocalyptic fury of Constantine Porphyrogenitus, and the quiet secrecy of the Lampros.