It's interesting how many people hear the word " Byzantium" and immediately ask about Greek Fire, and you can do a search of AskHistorians and find a number of previous posts. There do not seem to be really hard answers for a lot of questions about it: whether it contained saltpeter , whether it had quicklime. JR Partington, in his History of Greek Fire and Gunpowder, ransacked and carefully considered a lot of chemical possibilities and rejected the idea of a gunpowder weapon. It does seem to have had oil, resins of some sort ( like pine rosin), and needed a delivery system.

Like the composition question, why it was lost has only some possible explanations. I'm not sure what Byzantine scholars are doing with the question right now, but one pretty good explanation was proposed by Alex Roland in 1992. He placed Greek Fire into the world of craft production. In those times, of course, everything from making barrels to salting fish was done as a craft. Crafts often have trade secrets. Those can be recipes- like the Chinese formula for porcelain- or methods, like the process for making Wootz steel, or a combination of the two- like the Cremona violin makers' varnish. By their very nature, trade secrets are proprietary, held by only a few privileged individuals, and as the family was a basic structure for craft producers in this pre-industrial period, often they would be known within a family. The recipe for Greek Fire could have been controlled fairly easily: it could have been manufactured in one secure facility, put into jars and distributed. Similarly, the process for wielding it could have been somewhat secret. An apprentice in the later pre-modern world would be taught the "art and mystery" of a craft, and that would include not only things that could be described- like the proper proportions of a tenon to the mortise for a joint in a wooden beam- but skills that have to be demonstrated and then learned, like the feel of when an ash boat rib has been steamed long enough and is ready to be bent and nailed into place. Greek Fire seems to have required a complex delivery system: it was heated belowdecks by a crew and then pumped up to be wielded by another crewman, handling something like a swiveling pipe with a nozzle. Making all that work could have been pretty tricky. Pressure vessels would have to be strong enough to take both heat and pressure, pumps would have to be able to do the same; some typical sealing materials, like leather, would not have worked well in boiling hot oil and rosin, and the risks of having some of the flaming stuff leak out into their own wooden boat would be pretty high, require a lot of care and thought. Maybe this is why when the Bulgars captured some Greek Fire in 814 they couldn't actually put it to use.

There were, according to one account, only two families who knew about it: the family of the emperor, and one called Lampros. If there was a lot of art and mystery to making and using it, it's tempting to assume that, whatever the emperor's family might know, they would have other important things to be doing and the Lampros family would be the ones actually making it and training themselves and their children to use it. But, whichever, a small group had control of the art and mystery of Greek Fire for almost 500 years. During that time there were four major upheavals in Byzantium: emperors were murdered, factions warred with other factions. Roland suggests that it was during one or more of these upheavals that the art and mystery of Greek Fire was lost. There's some obvious merit in this- when an emperor is being murdered he likely doesn't have time to pass on the art and mystery of much of anything, and a family that knew about Greek Fire might have been powerful enough to get caught in some similar bloody factional dispute and similarly extinguished. Maybe Greek Fire could have become dangerous to know about.

Roland's explanation is appealing but possibly too neat and tidy: it relies perhaps a bit too much on the ability of humans to keep a secret for 500 years. Also perhaps it relies too much on the stereotype of Byzantium as being a violent, conspiracy-ridden.... well, byzantine place, and maybe some actual Byzantium scholars would take issue with this. And Roland admitted there are also other possible explanations. Greek Fire was good for defense, especially defending Constantinople, but not as good for attacking. Byzantium would begin to face more and more attacks on land, from Arabs and Turks, and Greek Fire was not as effective against armies. Also, weapons systems generally provoke development of defenses. If you know an enemy is going to approach you with a fire-shooting hose, you're likely to think about fire-proofing your boat. If you have 500 years to work on fire-proofing, you might come up with some very good stuff. And, unlike Greek Fire, fire-proofing is likely to be pretty dull, like wet cowhides on wood frames. Nothing that anyone wants to write about. Just like in computer games, there's a lot of attention given to describing , say, the 104mm Howitzer but not the composition of the concrete in the bunker it's shelling. Maybe when the enemies of Byzantium found ways to defend themselves from Greek Fire, and then stopped attacking by boat, the Byzantines just stopped trying to go to the immense trouble of using it.

Alex Roland: Greek Fire and the Defense of Byzantium 678-1204 Technology and Culture, 1992 v.33 No.4