A bone white plain stretches before a silent city, and the winter sun never warms it. What look like stones are not stones at all. They are skulls, stacked in tidy rows along the road. This is not a battlefield but the aftermath of a siege. To understand medieval warfare, you have to understand sieges, because more people died from sieges than from open field battles. Today, we will walk through the deadliest medieval sieges, what made them so lethal, how they worked, and what they changed in politics, technology, and culture. First, a quick grounding. The medieval world did not revolve around weekly pitched battles. Fortified places controlled trade, taxes, and legitimacy. Castles and walled cities were vaults that held power. If you wanted a throne, a port, or a harvest, you needed walls to open. Armies therefore spent more time sitting and starving enemies than charging them. That choice shaped casualties. Slow hunger, waterborne disease, and breakdown of sanitation killed more defenders and civilians than swords did. The deadliest sieges were not the most dramatic assaults. They were the longest and the filthiest. Let us set out a simple framework. A siege had five overlapping phases: blockade, reduction, assault, consolidation, and relief or surrender. Blockade cut the food. Reduction tore down defenses with mining or engines. Assault tried to breach at a chosen moment. Consolidation meant securing the breach and suppressing pockets of resistance. Relief or surrender decided whether the city fell or survived. At every step, disease lurked. Understanding the phases helps us map why some sieges spiral into catastrophe.

Now to the cases. We will cover several sieges across regions and centuries that together show the full spectrum of lethality. We will look at Antioch during the First Crusade, Acre during the Third Crusade, the Mongol sack of Nishapur, the siege of Baghdad in the thirteenth century, the long fall of Xiangyang in Song China, the siege of Caffa during the Black Death, the endless pressure on Constantinople culminating in its fall to the Ottomans, and we will end with what all of these teach about technology and policy. Antioch, in the late eleventh century, shows the math of starvation. Crusader armies reached the city in autumn, after a brutal march. The walls were strong. The countryside was stripped of grain. The army besieging the city began to starve even as it tried to starve the defenders. Horses died by the thousands. Men ate weeds and boiled leather. Contemporary accounts describe soldiers trading armor for handfuls of barley. Antioch flipped the script because both sides suffered siege conditions. A relief army from Mosul attacked the crusader camp. The siege became a cycle of blockades within blockades. When the crusaders finally entered the city through treachery, they in turn were besieged by a second relief force. Antioch’s lethality came from duration and disease. Chronicles speak of tens of thousands dead in and around the city. Antioch teaches that sieges kill attackers as well as defenders, and that poor logistics magnify the toll before any wall falls. Acre a century later shows how coalition politics extend the agony. The Third Crusade brought fleets, engineers, and money. Yet the siege lasted almost two years. Naval blockades fought over a few miles of coast. Counterfortifications sprang up like mushrooms. Siege towers burned and were rebuilt. Catapults hurled stones day and night, but food remained the core weapon. The city’s wells fouled. Refugees packed the streets. Meanwhile, the besieging army camped in marshy ground where latrines overflowed. Typhus and dysentery rolled through tents. One observer counted more burials than meals. When Acre surrendered, chroniclers put the dead well above fifty thousand, though medieval numbers are often inflated. Even if corrected downward, Acre stands among the deadliest episodes because armies on both sides lost heavily to disease while the civilian population collapsed inside the walls. The lesson is clear. Machines of siege matter, but water quality determines survival. Nishapur in the early thirteenth century illustrates a different path to mass death. The Mongol campaign used speed, terror, and calculated reprisals to force quick surrenders. When a city resisted or killed a Mongol envoy, retribution followed. After a Mongol prince died in fighting near Nishapur, the city was targeted. Once taken, the Mongol command ordered a complete massacre. Entire neighborhoods were executed in organized fashion. Mongol practice used unit tallies to count bodies. This was not starvation death over months. It was a post capture killing executed in days. Exact numbers are impossible to verify. Some Persian sources claim hundreds of thousands. Modern estimates are more modest yet still enormous. Nishapur is a reminder that the deadliest medieval sieges sometimes produced the highest daily kill rates after walls fell, not during the encirclement. Baghdad in the year twelve fifty eight marks a civilizational break. The Mongols under Hulagu surrounded the Abbasid capital with siege engines unmatched in the region. Chinese and Persian engineers assembled counterweight trebuchets that threw stones heavy enough to batter brick ramparts. The Tigris was controlled by boat bridges and anchored chains. The caliph refused realistic terms and the city was stormed after less than two weeks of bombardment. What made Baghdad’s fall so lethal was not time but population density and deliberate destruction. Libraries and hospitals burned. Workshops and bazaars collapsed in fires. House to house killing spiked. Recent scholarship argues for death tolls in the high tens of thousands to a few hundred thousand when you include those who fled and never returned. Baghdad shows that a short siege can still be the deadliest event in a century if the victors choose annihilation and if fire spreads through a tightly packed urban core. If Baghdad shows how engineering can hasten a deadly storm, Xiangyang in southern Song China shows how engineering can extend slow strangulation to its endpoint. The twin fortified cities of Xiangyang and Fancheng guarded the Han River. For years, Song forces used river control and fortified islands to block invaders. The Mongol Yuan command turned the river into the battleground. They built a chain of forts to cut boat traffic and introduced traction and counterweight trebuchets supplied by engineers from the Islamic world. Giant arrows wrapped with incendiaries fell inside the walls. But the decisive factor was a complete blockade. Grain could not reach the garrison. The cities resorted to eating horses and then rats. Contemporary accounts mention cannibalism. When the cities surrendered after several years, the population was wasted by hunger. The death toll included not only combat losses but slow deaths across the region as villages lost trade and harvests. Xiangyang teaches that control of water routes worth more than any single machine. Meanwhile, in the Black Sea, Caffa on the Crimean coast shows a siege whose legacy did more damage than its direct casualties. During the mid fourteenth century, the Golden Horde besieged this Genoese trading enclave. An outbreak of plague erupted in the besieging camp. One chronicle claims that plague corpses were hurled over the walls with siege engines. Whether that detail is accurate or not, the disease did spread inside the city. Refugees and merchants then fled by sea, carrying plague bacteria westward. Within a few years the Black Death ravaged the Mediterranean and beyond. The siege of Caffa matters because it shows how a localized military event can become a pandemic vector. The dead inside Caffa were far fewer than those who died later, yet the siege stands as a hinge moment linking military pressure and microbial catastrophe. Constantinople offers a long window into siege culture. The city faced dozens of sieges across a millennium. Its defenses were unrivaled. The land walls stood in three structured lines with towers and a moat. The sea walls protected harbors with chains. Most besiegers failed because they could not maintain a full blockade and because defenders used the city’s wealth to buy time. And yet even failed sieges killed many people in the countryside and suburbs. The final fall in the mid fifteenth century blended old methods with new technology. The Ottomans prepared months in advance, built roads, stockpiled grain, and ferried cannon. They cut the Bosporus with a new fortress to stop any relief fleet. Their cannon battered key towers. Engineers filled moats and moved ships over land to threaten the inner harbor. Christian defenses were brave and disciplined but undermanned. When the breach opened, the sack followed rules common to the era. The city suffered looting, enslavement, and killings within a set window allowed by custom. The exact death toll is debated. What matters for our purpose is that the siege compressed a wide zone of slow attrition and then released the spring in a single day. For the population inside, years of decline ended in a final convulsion.

Now that we have toured case studies, step back and break apart the mechanics of lethality in medieval sieges. Four forces dominate the death count. First is logistics. Grain and clean water decide timelines. If a city lacks reserve stores or if an army cannot secure supply lines, bodies fall before a single spear breaks. Second is disease. Dysentery, typhus, plague, and malaria exploit crowding, malnutrition, and poor sanitation. Third is the policy of no quarter or limited quarter. When leaders order massacre or enslavement, the curve of death spikes in hours. Fourth is technology that changes the balance between defense and offense. New engines and gunpowder create short sieges; river chains and blockading fleets enable long ones. The deadliest outcomes tend to combine at least two of these forces. Caffa combines disease with flight. Baghdad combines engines with massacre. Xiangyang combines blockade with water control. Understanding siege engines helps explain choices on both sides. In the early medieval period, traction trebuchets used teams of men to haul ropes and fling stones. They were effective but limited in range and force. Counterweight trebuchets, introduced from the Islamic world and perfected in Europe and East Asia, used a falling weight to throw heavier projectiles with far greater accuracy. These machines could chip at curtain walls and threaten towers. Battering rams protected by wheeled sheds tried to break gates. Siege towers offered height for archers to suppress defenders. Miners worked underground to collapse walls by burning wooden props that held up tunnels. Defenders built hoardings to drop stones and fire on attackers, dug countermines, and built inner walls for defense in depth. By the fourteenth and fifteenth centuries, gunpowder appeared in bombards and culverins. Early cannon were unreliable and slow but they frightened defenders and shattered specific points in walls, forcing concentrated defense. None of these tools kills as many people as sewage does, but they shape whether a siege ends in days, months, or years. Naval power multiplies siege effects. Ports like Acre, Lisbon, and Constantinople show what happens when a city can be fed by sea. A partial blockade leads to stalemate, while a full blockade drives collapse. That is why the Ottomans built forts on both sides of the Bosporus, why Richard the Lionheart fought naval skirmishes at Acre, and why Song generals fought over river islands at Xiangyang. Control of a waterway is like holding the keys to a silo. It fixes the timeline. Medieval commanders understood this, even if their chronicles emphasize heroics more than cargo manifests. Civilians shape outcomes in ways often ignored. As food dwindles, city authorities must ration, police hoarding, and maintain trust. If confidence collapses, riots break out. Factions form. Some press for negotiations, others for hard defense. Defectors can open gates. Antioch fell in part due to betrayal by a tower commander. Acre’s end involved bargaining over prisoners and promises that were later broken. In many sieges, population pressure makes rulers cruel. Punishments harden. Deserters are executed. In a few cases, as at Xiangyang, desperate people turn to cannibalism. Civilians also work on the walls, carry water, and maintain fire watch. Siege is total war, but without factories. Labor decides whether engines burn or keep firing. Negotiation sits beside violence in almost every siege. Terms can mean the difference between survival and massacre. Islamic law, Christian custom, and Mongol policy all had frameworks for surrender. A city that opened its gates before a breach might be granted safety, tax obligations, or relocation without slaughter. If a city forced attackers to take it by storm, the victors often claimed rights to plunder and to kill resisters for a set period. That convention created a terrible calculus. Leaders had to judge whether the enemy would keep terms, and whether their own soldiers would accept mercy. The deadliest episodes often involved broken promises, as with Acre, where the execution of captured garrisons inflamed both sides and encouraged harsher measures later. Studying these patterns shows how law and honor matter not as decoration but as determinants of body counts. Let us compare time scales. Long sieges are attritional furnaces. They kill slowly but they sweep broadly. Short sieges with sack and massacre kill quickly but leave a different trauma. For a city like Xiangyang, the cumulative dead include surrounding villages, boat crews, and distant markets strained by collapsed trade. For a city like Baghdad, most of the dead died inside the walls within a week. Long sieges degrade infrastructure and cause famines that ripple for years. Short sacks destroy libraries and leadership networks that never regrow in the same shape. Measuring lethality is not only about counting bodies. It is also about counting lost years of recovery. Numbers are always contested. Medieval chroniclers wrote for patrons and moral aims. Victories were inflated to show divine favor. Atrocities were magnified to incite revenge. When a source claims hundreds of thousands dead, we must compare plausible city size, grain capacity, and burial logistics. Even with sober corrections, the deadliest sieges stand out. Baghdad’s toll likely ranks near the top for a single event. Xiangyang’s broader famine toll is immense. The composite deaths from the two year siege of Acre and its associated clashes match or exceed the casualties of many campaigns. Nishapur’s mass execution, even at the lowest estimate, remains one of the darkest episodes of punitive siege warfare. Siege warfare also transformed technology and institutions. Castle design evolved from high thin walls to thicker, lower profiles with angled bastions as cannon matured near the end of the medieval period. That change, which blossomed in the early modern trace italienne, began in response to the experiences at places like Constantinople and the growing effectiveness of bombards. Logistics institutions strengthened. Grain magazines, road maintenance, and river fleets received investment from rulers who learned that supply feeds victory. Professional engineering corps emerged. Specialists in mining, bridge building, and artillery moved from city to city, selling expertise. Warfare became more bureaucratic because siege demanded plans that lasted longer than a single summer.

There are ethical legacies too. The memory of massacre at Nishapur, of destruction at Baghdad, and of the sack at Constantinople echo through culture and identity. Legal traditions wrestled with the question of civilians in walled cities. Some jurists argued for protection if defenders were disciplined. Others sanctioned the sword once a breach was forced. The distance between those views could be measured in tens of thousands of lives. Over time, ransom and prisoner exchange became tools to limit slaughter by placing value on captives. Yet when religious zeal or revenge politics ruled, restraint vanished. Understanding these moral swings helps explain why some sieges ended with negotiated orders and others ended in flames. One surprising theme is how often sieges failed quietly. For every famous sack, many encirclements ended in stalemate or negotiated withdrawal. Disease inside the besieging camp, as at Caffa, or the approach of a relief army forced attackers to lift the siege. These near misses still killed. They left burned fields and ruined villages. But they did not produce the concentrated catastrophe of a city taken by storm. Success for defenders did not always mean safety. Populations might be evacuated or taxed into ruin to fund garrisons. The absence of a massacre does not mean the absence of a tragedy. We should also note regional differences. In Europe, fragmented polities and smaller urban centers meant many sieges were short affairs against castles or market towns. In the Islamic world, richer urban networks and river systems raised the stakes, so sieges like Baghdad’s had civilization scale impact. In East Asia, riverine warfare and large administrative cities meant blockades could starve entire prefectures. The Mongol Empire, stretching across these regions, integrated technologies and tactics, moving engineers and ideas along steppe routes. That cross pollination explains both the speed of some conquests and the severity of their sieges. So which was the deadliest medieval siege. If you define deadliest as the single event with the highest immediate death toll inside a city over a short window, Baghdad in twelve fifty eight is a leading candidate, with estimated deaths possibly reaching into the hundreds of thousands when including killings, drownings, and those who perished in fires and rubble. If you define deadliest as a prolonged siege whose total deaths include defenders, besiegers, civilians, and the knock on famine in the region, Xiangyang likely rivals any European case, though numbers spread across years. If you define deadliest as the siege with the largest indirect mortality through secondary effects, Caffa during the Black Death stands alone, as its breakout helped carry plague into the Mediterranean world. There is no single answer because lethality depends on the measure. Recognizing that helps avoid poor comparisons and invites clearer thinking about what siege does to societies. Take away practical lessons that help you evaluate accounts. When you read about a siege, ask four questions. How secure were the food and water supplies on each side. What was the disease environment. What were the stated surrender terms and the commanders’ reputations for keeping them. What technologies or geographic features controlled time, such as rivers, straits, or mountain passes. With those answers, you can forecast whether death will come slowly by attrition or fall suddenly after a breach. Now think about the people who made these decisions. A general weighing whether to assault a breach at dawn had to consider not just courage but the number of days of grain left in the wagons, the temperature of the latrines, and whether a rival prince might arrive with a relief army. A city council deliberating whether to open gates had to reckon with the loyalty of mercenaries, the patience of guilds, and the fate of their neighbors who surrendered last year. These are calculations we can reconstruct from sources, and they help strip away the mystique. Siege is logistics, law, and psychology layered over stone. Finally, think about memory. Survivors of Antioch told stories about holy relics and visions because that helped make sense of starvation. Survivors of Baghdad remembered smoke and silence and the color of the river. Survivors of Constantinople remembered the sound of cannon and the moment the chain fell. Memory exaggerates, but it also tells you what mattered to the people there. For them, the deadliest siege was the one that took their street, their book, their well. Our job, as learners, is to connect those human details to the larger forces we have mapped. To close, here is the short version you can carry forward. Medieval warfare was siege warfare. The most lethal sieges combined bad sanitation, broken supply lines, deliberate brutality, and decisive technology. Antioch shows attrition grinding both sides. Acre shows coalition complexity magnifying disease. Nishapur shows punitive massacre. Baghdad shows short siege, high death, and civilizational shock. Xiangyang shows river control turning siege into regional famine. Caffa shows siege as a spark for global plague. Constantinople shows the endgame of medieval walls under gunpowder pressure. When you encounter a new siege in your reading, apply the framework. You will see why some cities starved for years and why others burned in a week. You will also see why, for civilians within the walls, the deadliest siege was the one they could not escape.