Collections: Why Archers Didn’t Volley Fire
This week we’re looking at a specific visual motif common in TV and film: the arrow volley. You know the scene: the general readies his archers, he orders them to ‘draw!’ and then holds up his hand with that ‘wait for it’ gesture and then shouts ‘loose!’ (or worse yet, ‘fire!’) and all of the archers release at once, producing a giant cloud of arrows. And then those arrows hit the enemy, with whole ranks collapsing and wounded soldiers falling over everywhere.
From Alexander (2004) showing the Battle of Gaugamela. This bit is amazing because Darius III **silently gestures** and all of his archers draw their bows (also why are they kneeling? They’re shooting at a high angle! There’s no need to kneel!) and then at another **silent gesture** which they **cannot see** because Darius III is **behind them** , they all release at once.
**And every part of that scene is wrong**.
Now the thing that, in the last couple of decades, _everyone_ has realized is wrong (I suspect some early Lindybeige videos had something to do with how widespread this notion is), is that you don’t tell archers to ‘fire’ because their weapons _don’t involve any fire_. But the solution in film has been to keep the arrow _volleys_ – that is, the coordinated all-at-once shooting – and simply change the order to ‘release’ or ‘loose.’ Which isn’t actually any better!
Archers didn’t engage in coordinated all-at-once shooting (called ‘volley fire’), they did not shoot in volleys because there wouldn’t be any point to do so. Indeed, part of the reason there was such confusion over what a general is supposed to shout instead of ‘fire!’ is that historical tactical manuals don’t generally have commands for coordinated bow shooting _because armies didn’t do coordinated bow shooting_. **Instead, archers generated a ‘hail’ or ‘rain’ (those are the typical metaphors) of arrows as each archer shot in their own best time**.
More to the point, they _could not_ shoot in volleys. And even if they _had_ shot in volleys, those volleys wouldn’t produce _anything_ like the impact we regularly see in film or TV. **So this week, we’re going to walk through those considerations** : briefly looking at what volley fire is for and why archers both wouldn’t and couldn’t do it, before taking a longer look at the problem of **lethality** in massed arrow fire.
But first, if you want to help support this project you can do so on Patreon! I don’t promise to use your money to buy myself more arms and armor, but I also don’t promise _not_ to do that. And if you want updates whenever a new post appears, you can click below for email updates or follow me on Twitter (@BretDevereaux) and Bluesky (@bretdevereaux.bsky.social) and (less frequently) Mastodon (@bretdevereaux@historians.social) for updates when posts go live and my general musings; I have largely shifted over to Bluesky (I maintain some _de minimis_ presence on Twitter), given that it has become a much better place for historical discussion than Twitter.
##
Email Address
Subscribe!
## What Is Volley Fire For?
**We want to start by understanding what volley fire _is_ and what it is _for_**. Put simply, ‘volley fire’ is the tactic of having a whole bunch of soldiers with ranged weapons (typically guns) fire in coordinated groups: sometimes with the entire unit all firing at once or with specific sub-components of the unit firing in coordinated fashion, as with the ‘counter-march.’ In both cases, the problem that volley fire is trying to overcome is **slow weapon reload times** : this is a solution for _slow-firing_ but _powerful_ ranged weapons. That has generally meant firearms, historically, but we do actually see volley fire drill with crossbows in China from a very early period as well (but, interestingly, there’s no evidence I am aware of that volley fire was ever done with crossbows in Europe – when Europeans decide to do volley fire with firearms, it seems to have been an entirely new idea).1
From The Two Towers (2002). Théoden here has been having his archers **hold their shots** for going on a couple of minutes at this point, so while volley fire with arrows is stupid, he’s probably right that he’s only going to give the enemy one volley before his archers collapse from exhaustion.
Volley fire can cover for the slow reload rate of guns or crossbows in two ways. The first are volley fire drills designed to ensure a continuous curtain of fire; the most famous of these is the ‘counter-march,’ a drill where arquebuses or muskets are deployed several ranks deep (as many as six). The front rank fires a volley (that is, they all fire together) and then rush to the back of their file to begin reloading, allowing the next rank to fire, and so on. By the time the last rank has fired, the whole formation has moved backwards slightly (thus ‘counter’ march) and the first rank has finished reloading and is ready to fire. The problem this is solving is the danger of an enemy, especially cavalry, crossing the entire effective range of the weapon in the long gap between shots. This, by the by, was the volley fire tactic that was being used in China with crossbows before gunpowder; I don’t know that anyone ever did volley-and-charge with crossbows, which lack the lethality of muskets.
From The Fellowship of the Rings (2001). It’s a little hard to see in a still, but what these archers are doing is that each rank releases and then knees while the next rank shoots, which is wholly unnecessary, as you can see from the angles the bows are being held. Its also counter-productive: kneeling and standing again take longer to do than to just knock a new arrow and fire! This is a volley fire drill that is *slower* than just shooting normally!
**The other classic use is volley-and-charge**. Because firearms are _very_ lethal but slow to reload, it could be very effective to march in close order right up to an enemy, dump a single volley by the entire unit into them to cause mass casualties and confusion and then immediately charge with pikes or bayonets to try to capitalize on the enemy being demoralized and confused. You can see variations on this tactic in things like the 17th century Highland Charge or the contemporary Swedish Gå–På (“go on”). By charging rather than waiting to reload, the attacker could take advantage of the high lethality of firearms without suffering the drawback of long reload times.
From House of the Dragon (2022-present). Again, hard to see here, but in this scene, you can hear a voice shouting ‘loose!’ to coordinate shots and then two ranks of archers are exchanging positions at the wall, essentially a two-rank counter-march with bows, again made pointless by the fact that an archer can shoot faster standing still than these rank exchanges.
This is also not generally how we hear of gunpowder-based troops firing from a parapet like this: more often what we hear is that each file has a single shooter and several men behind him reloading muskets and handing them forward.
Crucially, note that volley-and-charge works because it compresses a _lot_ of lethality into a very short time, which I suspect is why we don’t see it with bows or crossbows (but _do_ see it with javelins, which may have shorter range and far fewer projectiles, but seem to have had higher lethality per projectile). As we’re going to see in a moment, the lethality of bows or crossbows against armored, shielded infantry – even in close order – was pretty low at any given moment and needed to add up over an extended period of shooting. By contrast, muskets were powerful enough to defeat most armor and thus to disable or kill basically anyone they hit, limited of course by reload time: with a reload time of as much as 30 seconds for earlier matchlocks, a line of musketeers might only be able to fire a few times at an advancing infantry unit (which might take two or three minutes to walk through effective range) and given the limited accuracy of smoothbore muskets, only the last shots would hit at a high level. By contrast, a unit doing volley-and-charge is compressing probably close to 50% of the lethality of sustained shooting, devastating moment and then _immediately_ charging.
Putting _**that much**_ lethality into a singular instant was valuable from a morale perspective and of course it enabled a unit to quick march through the enemy’s effective range, stopping only briefly to fire and charge, limiting losses from steady enemy fire. But as we’re going to see, the lethality of bows (and, to a significant extent, crossbows) was _**much lower**_ and so couldn’t be effectively compressed into that single, devastating, confusing moment.2
## Why They Wouldn’t and Why They Couldn’t
But as you’ve hopefully noted, these tactics are built around firearms with their long reload times: good soldiers might be able to reload a matchlock musket in 20-30 seconds or so. **But traditional bows do not have this limitation** : **a good archer can put six or more arrows into the air in a minute** (although doing so will exhaust the archer quite quickly), so there simply isn’t some large 30-second fire gap to cover over with these tactics. As a result **volley fire doesn’t offer any _advantages_ for traditional bow-users**.
And so, as far as we can tell, organized volleys with bows weren’t done. We _do_ have evidence in China for volley fire _with crossbows_ , but of course crossbows, particularly more powerful ones, have all of the same reload-time problems that firearms do, so it is no shock to see the same tactics emerge. But historians have searched the ancient and medieval sources for any hint of volley fire with bows and have come up wanting. Now, I should caution here that this is a topic where if you are reading sources _in translation_ you are likely to be fooled: **many translators will use the word ‘volley’ to describe things happening in the original Greek or Latin or Old French or what have you that are not volley fire** , for the same reason that filmmakers keep putting archer volley fire in their movies: volley fire is a big part of how we _imagine_ warfare. But as hard as it is to prove a negative, I will note that I have never seen a clear instance of volley fire with bows in an original text and so far as I can tell, no other military historians have either. **And we _have_ been looking**.
Of course the other reason we can be reasonably sure that ancient or medieval armies using traditional bows did not engage in volley fire is **that they couldn’t**. You will note in those movie scenes, that the commander invariably gives the order to ‘draw’ and then waits for the right moment before shouting ‘release!’ (or worse yet ‘fire!’). The thing is: how much energy does it take to hold that bow at ready? The key question here is the bow’s ‘draw’ or ‘pullback’ which is generally expressed in the pounds of force necessary to draw and hold the bow at full draw. Most prop bows have _extremely low_ pulls to enable actors to manipulate them very easily; if you look closely, you can often _see this_ because the bowstrings are under such little tension that they visibly sway and wobble as the bow is moved. This also helps a film production because it means that an arrow coming off of such a bow isn’t going to be moving all that fast and so is a lot less dangerous and easier to make ‘safe.’
But obviously actual bows are _supposed_ to be dangerous.
And here folks will say, “ok, that’s prop bows, but I hold a hunting bow at full draw while lining up a shot all the time.” But there are two considerations here. The first is that many modern hunting bows are _compound bows_ (note: compound, not composite), which is to say they use lever and pulley systems with wheels (‘cams’) which enable the energy at each stage of the bow’s draw to be controlled and are typically designed so that the energy necessary for the final bit of draw (that is, holding the bow at full draw) is relatively low. As a result, the strength required to hold a compound bow at full draw for an extended period is actually _lower_ that what would be implied by its raw pullback.
**But also the pullbacks of hunting bows are _much_ lower than those of war bows**. Modern hunting bows generally feature pullback weights around 40-60lbs (going higher for compound bows but still generally topping out around 75lbs and typically being much less) and shoot lighter, thinner arrows than war bows. And that should make a fair degree of sense: _deer cannot shoot back and do not generally wear armor_. **The military archer, by contrast, needs a _lot_ of lethality and a _lot_ of range because he is shooting at someone with armor and weapons who means to shoot back** (or run up and stab him), although as we’ll see, even with _extremely powerful bows_ the ability of war archers to inflict lots of casualties is pretty limited against properly equipped enemies. If your hunting bow mortally wounds a deer but does not disable it, that’s not ideal but the deer is going to run _away_ , not charge at you spear in hand.3
**As a result, the pullback weights of war bows tend to be higher**. How much higher? We’ve actually run through this evidence before: at least in Afroeurasia, as far as I can tell, **80lbs pullback is about as light as a war bow will usually get. Draw weights anywhere from 100lbs to as high as 170lbs** (see Strickland and Hardy, _The Great Warbow_(2005) for details) are known for the highest end bows like the English longbow and Steppe recurve bows. Which is to say that the pullback weight range of ‘old world’ war bows exceed _at their lowest end_ the heaviest common draw weights of hunting bows and keep going up dramatically from there. **The typical war bow was more than _twice_ as powerful as the typical modern hunting bow.** These war bows shot with enough force that they required specialized arrows with thicker, more robust construction to withstand the amount of energy being imparted.
**Which neatly answers why no one had their archers hold their bows at draw to synchronize fire: you’d exhaust your archers _very quickly_**. Instead, war bow firing techniques tend to emphasize getting the arrow off of the string as quickly as possible: the bow is leveled on the target as the string is drawn and released basically immediately. Remember back to our statistic that a good archer can put around 6 arrows in the air in a minute? Well, even the best archer can’t do that _for very long_. I often see folks asking about how many arrows an archer could carry, seemingly imagining archers shooting at their maximum rate for prolonged periods (like they do in video games), but if you imagine pumping a 150lbs weight as fast as you can, I think you’ll immediately recognize that you aren’t going to be able to keep that up for more than a minute or two (more on this as well in Strickland and Hardy, _The Great Warbow_(2005), by the by). **Holding the bow at draw for any length of time is going to accelerate that exhaustion and thus _lower_ the rate at which shots are made and the time that rate can be maintained**.
**So the reason we have no evidence for archer volley fire is because they didn’t do it and they didn’t do it because it doesn’t solve a problem that exists with bows** (whose rate of shot is fast enough not to require volley tactics)**but it _does_ cause all sorts of new problems** (exhausting your archers).
But there’s a second related problem to these scenes: **arrow lethality**.
## Modeling Arrow Lethality
Because when these arrow volleys arrive, the result is usually devastating, with large numbers of men falling all over the place (often being shot straight through their heavy armor).
From Alexander (2004). It’s hard to tell in a still image which actors are in the process of falling down, so I’ve put red ‘x’s on each actor that seems to me to be falling. The scene after this cuts to closer shots where we see even higher casualty rates.
But how lethal were arrow barrages? Well, the short answer is that we don’t know and it must have varied considerably. Teasing out the specific lethality of one part of an engagement from others is difficult even with modern warfare; for pre-modern warfare, we are often lucky to even have reliable estimates of total casualties in a battle, much less specific estimates of casualties caused by a specific source or weapon. Still, we have more than a few solid indications that the _lethality_ of barrages of arrows, in some cases even over extended periods, could be quite low, which isn’t to say such weapons were _ineffective_.
From Lord of the Rings: The Two Towers (2002), and here I’m not going to try to count the falling extras, but it is quite a lot of them, a few entire ranks at least.
We can start with ‘modeling,’ thinking through the question as a thought experiment (since we haven’t the expensive computer hardware and expertise to actually simulate it).4 Especially at long range, our archers are not shooting at individual enemies, but rather firing _en masse_ into a large body of infantry, so we can assume shots are probably distributed fairly evenly in the target area. That’s already actually significant because as we discussed before, even in close-order infantry formations, there’s usually quite a lot of empty horizontal space (file width) where an arrow is simply going to hit…no one.
Depending on the way the men in the target infantry formation are facing and the formation, in most fighting formations, upwards of 50% of the total horizontal space simply doesn’t contain and humans to hit and arrows plunging into that space are going to hit nothing but the ground. Now the vertical space is trickier: there’s going to be a lot of empty space between the ranks as well, though we are almost never informed about how much. One exception is the Macedonian _sarisa_ phalanx, where we’re told (Polyb. 18.29) that the _sarisa_ of the fifth rank extends two cubits beyond the first rank, which lets us calculate roughly a 90cm rank interval. Other formations might have been tighter or looser, of course. But the implication here is that an arrow shot on a flat trajectory (so at very close range) _at least_ half of the target area is entirely empty space; for an arrow shot in a high arc, as much as 75% of the target area might be. And of course in this estimation, we’ve been treating our soldiers like they are large rectangular prisms (our army of gelatinous cubes will be very effective), but of course actual humans aren’t going to physical occupy a lot of the space we’re even giving them here (note the silhouettes below). **So the majority of arrows are simply _going to miss_**.
**But of course then our target infantrymen are also not unprotected**. Let’s assume here an average infantryman who is roughly 170cm in height (5ft 7in, a touch on the tall side, but not unreasonable for pre-modern agrarian soldiers). The first thing he is likely to have protecting him is a shield. For the purpose of our arrows killing or disabling our infantryman, a decent shield is essentially perfect protection in the area it covers: even very light shields can ‘catch’ arrows effectively (and indeed, this is what very thin hide or wicker shields are for). The one risk we face is the arrow punching through the shield into the shield arm, which could certainly happen, but many shields have reinforced metal bosses over where they are gripped, making this less likely. But as we discussed with shield walls, **shields often cover quite a lot of the body** ; shields could be _quite big_. So let’s draw that out with some example shields, to scale with a human silhouette (again, 170cm tall) and see how much of this relatively big fellow (by pre-modern standards) typical shields covered:
Note that the human silhouette shape is via Wikimedia Commons.
**What you can immediately see is that just about any shield is going to massively reduce the target area of the body _even if it isn’t moved_**. _All_ of these shields are large enough to cover the entire trunk of the body, protecting all of the vital organs in the torso. Assuming our infantryman has crouched down a little and put his shoulder into his shield (and kept his weapon hand behind it), our archer has lost upwards of three-quarters of his target area (even higher for very large shields like the Roman _scutum_). Worse yet, the target area that remains is _mostly legs_ where arrow strikes, while painful, are a lot less likely to be lethal and may not even be disabling.
And of course these soldiers **can move their shields** , angling them up if the arrows are plunging downward or crouching behind the shield if they’re arriving on flat trajectories. Moreover **arrows at range move slowly enough to be actively blocked and dodged** , to the point that we know that ‘arrow dodging’ was a martial skill of some import in cultures that engaged in small-scale bow exchanges as part of ‘first system‘ warfare.5 Of course, if the incoming hail of arrows is dense enough, soldiers might be unwilling to put their heads up to try to spot incoming and block (at Agincourt we’re told the French soldiers angled their helmets into the arrow-rain, for instance), but infantry under lighter ‘fire’ might actively move their shield to block specific incoming arrows.
**And then _behind that shield_ our infantryman is also probably wearing some kind of armor!** Now a full plate harness is going to provide only _extremely_ few points of vulnerability, but to give our archers a more favorable case, let’s stay in the ancient world and consider two ‘edge’ cases from the Hellenistic period: a mailed Roman legionary (the most heavily armored infantryman of the period) and a Gallic warrior (one of the less armored infantrymen of the period). By picking soldiers this early, we’ve given our archers a bit of a hand: these fellows don’t have fully enclosed helmets, or significant arm protection; later medieval combatants, particularly with wealth, would have been _much better protected_ , with things like aventails to cover the neck and fuller protections for arms and legs. The Roman has a mail _lorica hamata_ , a Montefortino-type helmet (with cheek-flaps protecting much of the face) and greaves, while our Gaul has just the helmet and probably some thickened textile body protection. The coverage might look like this (please forgive my very rough efforts to draw out irregular shapes):
Once again, the human silhouette shape is via Wikimedia Commons.
Now as we’ve discussed, armor protection against arrows isn’t necessarily a binary. Armor often gets discussed as if arrows either _always_ defeat it or _never_ do and really only one of those is correct: arrows will not defeat good iron or steel plate armor at effectively any range. But for other forms of armor, the range and the power of the bow matter a lot. I’m going to summarize my previous estimates here (but I sure do wish we had more long-range bow-penetration testing!): at relatively long range (c. 200m) even powerful bows might struggle to reach the target with enough impact energy to penetrate mail and relatively weak war bows – which are still bows with _80lbs pullback_ (so our weak war bow is roughly 50% _more powerful_ as a typical hunting bow) – may struggle to even penetrate a good textile defense with a solid hit. Even at moderate ranges (c. 100m), mail will probably sometimes defeat even the most powerful bows (but sometimes it will fail) and even a gambeson provides a degree of protection from the weakest (again, _still 80lbs pullback_ bows).6
What that means for our Roman legionary up there the good news is that very few arrows are going to accomplish much; the situation is worse for our Gaul, but actually not _much worse_. For the Roman legionary, he has upwards of 85% of his body covered by his _giant_ shield. Should an arrow get around that shield somehow, to hit anything vital (except his face) it has to contend with his mail. Now powerful war bows, especially at short range _can absolutely_ defeat mail, but not every shot is going to be the most powerful bow shooting a point-blank range shot hitting dead on and for the rest, a decent chunk of them are going to fail to split the mail rings or else expend so much energy doing so that they don’t penetrate lethally deep through the thick textile padding (the _subarmalis_) beneath the mail. Meanwhile, his lower legs below the shield are covered with solid bronze greaves which will almost always deflect an incoming arrow (they’re both solid metal, but also _curved_ so an arrow is likely to glance off). His head and neck remain the big point of vulnerability, but something like three quarters of that space is covered by his helmet and his cheek-guards: an arrow slamming into a solid, 1.5kg bronze helmet is going to be _unpleasant_ , but the arrow isn’t _usually_ going to penetrate (though the impact may daze or even knock out the soldier).7
And if we start stacking these ‘filters’ for our arrows, we see the lethality of our barrage drops _very fast_ against infantry. Maybe two-third to three quarters of our arrows just miss entirely, hitting the ground, shot long over the whole formation and so on. Of the remainder, another three-quarters at least (probably an even higher proportion, to be honest) are striking shields. Of the remainder, we might suppose another three-quarters or so are striking helmets or other fairly solid armor like greaves: these _hurt_ , but probably won’t kill or disable. Of the remainder, a portion – probably a small portion, because of those big shields – are being defeated by body armor that they could, under ideal circumstances, defeat. And of the remainder that actually penetrate a human on the other side, maybe another two-thirds are doing so in the arms, feet or lower legs, many of them with glancing hits: painful, but not immediately fatal and in some cases potentially not even disabling.
After all of those filters, we’re down to an estimated arrow lethality rate hovering 0.5-1%, meaning each arrow shot has something like a 1-in-100 or 1-in-200 chance to kill or disable an enemy.8 To put that in perspective with the images above: Aragorn’s book-inaccurate Elf allies (about five hundred of them) could all shoot over the whole approach (probably about a minute) and kill or disable about 25 Uruk-hai out of that host of _ten thousand_.9
Of course they wouldn’t be firing in volleys and numbers would matter. But we can extend our model a bit. Let’s assume an equal sized force of heavy infantry, advancing at the quick step (so a march, not a charge) against an equal sized force of archers. Bow shot is about 200m, which a quick march will cross in about 2-and-a-quarter minutes (quick step is 120 steps per minute, 75cm covered per step, roughly). Each archer can loose six arrows a minute, so each infantryman has, on average, 13.5 arrows to deal with. His chance of being killed or disabled by one of those arrows over the course of marching into contact (assuming our 0.5% arrow lethality) is thus about 6.75%. And that is under _very_ favorable assumptions for our archers: our infantry doesn’t break into a charge, has no screening forces, the archers can shoot at maximum effective range, don’t tire out their arms and can all shoot effectively for the entire period (no return shots, no being blocked by friendly troops, etc). In practice, we should probably also impose a pretty sharp lethality ramp for these arrows: our 0.5% lethality figure is based on arrows loosed at pretty close range on flat trajectories, but of course the earliest shots in this scenario would be at much longer range, with less power and accuracy and so _**much less lethal**_ ; our 6.75% figure is thus something of a _maximum_. A 6.75% ideal disable rate is not going to stop the determined advance of heavy infantry: that infantry is going to march right on into contact and if those archers don’t have their own heavy infantry to meet it, they are going to be put to flight very quickly.
## The Model and the Metal
Now if all we had was modeling, this sort of analysis would be shaky, because we’re making so many simplifying assumptions. But of course we now want to compare our model with _actual battles_ to see if it seems like it is describing their mechanics accurately. At the Battle of Marathon (490 BC), a force of 10,000 Athenian and Plataean hoplites advanced over open ground into contact with a larger force (perhaps roughly double) of Persian soldiers, most of whom were likely archers, given how the Achaemenid army fought: the Athenian-Plataean army charged into contact and routed their enemy with just 192 KIA; many of these losses moreover were not from arrows, as our best source, Herodotus, is clear that the hardest fighting was in contact at the ships.10 At Issus (333BC), Alexander orders a quick approach for his infantry, worried about the large numbers of Persian archers (Arr _. Anab_. 2.10.3), but the Macedonians reached the Persian line and in the whole battle reportedly sustained only 150 killed, 4,500 wounded (Curt. 3.11.27). At the Siege of Nicaea (1097 AD) the relief army of Kilij Arslan, composed primarily of Turkish horse archers – some of the finest and most dangerous archers around – attempted to move the crusader shield wall but was unable to do so despite a prolonged effort (he eventually gets pulled into contact with heavy crusader cavalry and is quite soundly defeated).
And then, of course, there is Agincourt (1415 AD). On the one hand, Agincourt is held up as the great example of the victorious power of the English longbow. On the _other hand_ , both the initial French cavalry charge _and_ the subsequent French infantry advance were able to cross a muddy, open field into contact with the English force.11 **Agincourt reflects, in many ways, an ideal battle for the English longbow** : the enemy was forced to advance the full range of the weapon, without cover, over difficult ground and did so in distinct ‘waves’ (the French army was deployed in three successive lines), on a battlefield where the forests ‘canalized’ (funneled into a narrow space) the French advance and secured the English flanks. **And yet under these conditions the French infantry were able to cross the terrain in good order and attempt to breach the English line**. Of course, despite outnumbering the English, the French infantry attack was too weakened by the arrows to overcome the English men at arms and archers in contact and so the English won a great victory.
Via Wikipedia, a map of the English and French positions at Agincourt, showing the confined nature of the battle-space, which greatly aided the English. Note that there is a _**much better**_ much of the battle in Livingston, _Agincourt: Battle of the Scarred King_ (2023).
But the nature of that victory is actually quite telling: even in ideal circumstances, with one of the most powerful bows in history (and a body of experienced archers to wield them) the English could not simply ‘mow down’ the incoming infantry attack slogging forward. But at the same time,**the continuous rain of arrows created the conditions for the English to win** in the press of melee despite being outnumbered. The Roman historian Livy has these phrases that always jump to mind in these situations, describing men or armies – often still very much _alive_ – as _fessus vulneribus_ or _vulneribus confectus_ , “tired/worn-out by wounds” (Livy 1.25.11; 22.49.5; 24.26.14). After all, an arrow that gives a shallow cut glancing off an arm or bangs off a helmet or other piece of armor or slams into a shield isn’t going to kill you and probably isn’t immediately disabling, _but it does hurt_ and the added impact of cuts and bruises is going to contribute to exhaustion (and arrows stuck in a shield make it harder to wield), slowly but steadily diminishing the fighting capability of the recipient.
**That is how I would understand the failure of the French infantry advance at Agincourt**. It isn’t that the longbows _killed_ them all, but that they injured, exhausted, confused and disconcerted the advancing infantry, so that by the time the French reached the fresh, close-ordered and prepared ranks of the English, they were at a substantial disadvantage in the close combat.
Via Wikipedia, a 15th century drawing of the Battle of Crecy (1346) from Jean Froissart’s Chronicles. This is, of course, a meaningfully later illustration, but I want to draw attention to a few things: first, neither the archers nor the crossbowmen are shooting in volleys – we can see men in both groups in different stages of shooting. Second, several of the crossbowmen are injured, with arrows in their legs, but **continuing to shoot** , because those wounds, while doubtless painful, haven’t disabled them. Of course we need to read such drawings with a healthy dose of skepticism, but I think this is closer to the reality than most modern movie scenes.
Now since I have brought up Agincourt, we also want to talk about _cavalry_. Because so far, we’ve been focused on _infantry_ facing massed archery. But note that at battles like Crécy (1346) and Agincourt (1415), the French also try cavalry charges and in both battles, these are _very_ roughly repulsed.12 That may seem strange because in strategy games and the like, cavalry is the _solution_ to archers, able to close the distance and defeat them quickly.
But actual battles are more complicated. On the one hand, cavalry is _faster_ : even heavy cavalry can cut the time spent crossing the ‘beaten zone’ of bowshot from around 2.5 minutes to just 1 minute. On the _other_ hand, horses are _big_ and react poorly to being wounded: a solid arrow hit on a horse is very likely to disable both horse and rider. And while light or archer cavalry might limit exposure to mass arrow fire by attacking in looser formation, as we’ve discussed, European heavy horse generally engages in _very_ tight lines of armored men and horses in order to maximize the fear and power of their impact. Unsurprisingly then, we see from antiquity forward, efforts to armor or protect horses, called ‘barding’: defenses of thick textile, scale, lamellar, and even plate are known in various periods, though of course the more armor placed on the horse, the larger and stronger it needs to be and the slower it moves. Nevertheless, the size and shape of a horse makes it harder to armor than a human and you simply cannot achieve a level of protection for a horse that is going to match a heavy infantryman on the ground, especially if the latter has a large shield.
Finally, the other thing about cavalry is that they weren’t as numerous. The cavalry charge at Agincourt had in it only 800 horsemen, for instance.13 But horses are big and cavalry cannot be packed in a deep formation, for reasons we’ve discussed, so the cavalry would still take up a fair bit of space on the battlefield, meaning that they would draw shots from a _lot_ of archers, potentially overwhelming the advantage of covering the space more rapidly. Michael Livingston, _op cit_ , does his own modeled simulation of the longbow impact on the French cavalry charge, with a lethality ramp from 0.25% to 2% over the charge and estimates that well over half of the riders wouldn’t have made it to the English lines.14 With so many archers firing at so few horsemen, the imbalance quickly produces catastrophe, although it is worth noting that even at this point the French cavalry charge _did reach the English line_ , albeit without the numbers or the morale impact to overcome it, with French knights being pulled off of their horses _within_ the English infantry formation, having presumably slammed through in their initial impact.
## Conclusions
One of the challenges in understanding pre-modern warfare is in navigating between the extremes of ‘wonder weapons’ and ‘useless’ weapons. **If bows were so powerful that they could mow down heavy infantry or invalidate cavalry, no one would have fought any other way**. We know that, of course, because eventually a technology emerges – firearms – which _was_ so lethal that it steadily pushed every other way of fighting off of the battlefield, save for a bit of light cavalry. Bows and crossbows existed for far longer and didn’t have this effect, because they weren’t that powerful: **they simply lacked the tremendous lethality of firearms**. The very strongest war bows might deliver at _most_ around 130 joules of impact energy, slicing and piercing through a target. By contrast even relatively early (16th century, for instance) _muskets_ could deliver one to two _thousand joules_ of impact energy, with a projectile that didn’t neatly slice or pierce the target (it didn’t need too), but smashed through, shattering bone and shredding issue over a much larger area.
**At the same time, bows and crossbows obviously weren’t useless**. Of course for nomadic steppe-based armies, they were the primary weapon and rapidly maneuvering horse archers could use bows to devastating effect (in part because unlike foot archers, they could repeatedly caracole into that higher lethality zone at very short range). For agrarian armies, archers and other ‘missile’ troops could screen heavy infantry or cavalry, harass enemies and under the right circumstances degrade an enemy force quite heavily, even if they couldn’t simply ‘mow down’ advancing infantry. To counter this, more sophisticated armies might advance their close-order heavy infantry with screening forces of light infantry, often with looser spacing (thus lowering the incoming arrow ‘hit rate’ even further). The Roman legion of the Middle Republic had a built-in screening force, the _velites_ , while we see the French, particularly at Crécy, attempting (and failing) to use their crossbows in this way. Those screening forces existed in part because harassing ‘fire’ from missile troops, while it might not turn back the advance of a legion, could significantly hamper it and so it was worth tasking a significant portion of the army to preventing that (and harassing the enemy in turn).
Of course TV and filmmakers are not thinking in these terms, but instead deploying – often without much thought – a set of visual tropes for battles which all have their origins in warfare in the gunpowder period. Directors love, for instance, having characters hold each other at bow or crossbow point, something that makes sense with modern firearms, but not with bows or crossbows (if you had to hold someone at weapon-point in the pre-gunpowder world, you used a sword or a spear).
The visual film ‘language’ for ranged engagements, in turn is very clearly drawn from warfare in the 1700s and 1800s. I suspect we can actually be a lot more specific, with the touchstones here being the American Revolutionary War and the American Civil War. Film as a genre, after all, emerged and was in its early days substantially shaped in an American context and much of filmic language remains dominated by Hollywood and in the United States, reenactments of ARW and ACW battles are quite common and for many movie-makers would be the primary way of engaging with any kind of warfare before the emergence of the genre of film itself in the early 1900s. This, of course, introduces some of its own problems _even for the warfare of the 1700s and 1800s_ , as reenactments tend to recreate parade-ground and field manual maneuvers and impose them on battles that were probably quite a bit more fluid and disorganized, but that’s a question for other scholars, I think, to unpack.15
But that mental model of warfare imposes both a physical logic and a dramatic logic on to battle scenes set in pre-gunpowder societies which simply do not belong there: the most obvious being the hero-commander dramatically giving the order to ‘fire’ at the key moment, something that calls back to the mythology around “Don’t fire until you see the whites of their eyes,” but which is inappropriate for bows and crossbows, which – among other things – we know often began slow shooting right at maximum range.
As with our discussion of “The Battlefield After the Battle,” I think there’s an opportunity here for filmmakers to break with that tradition and attempt to show the view a meticulously reconstructed battle and reap the dramatic benefits of how interesting and alien that would be. But until then, I suppose, I will have to suffer through more films showing archers doing volley fire drills, while kings shout for the men to ‘fire!’ their bows.
### Share this:
* Click to share on X (Opens in new window) X
* Click to share on Facebook (Opens in new window) Facebook
*
1. On drill and in particular, counter-march volley fire with crossbows, see Andrade, _The Gunpowder Age_ (2016), 149-160.
2. It also didn’t generate a smokescreen to help with the final rushing charge, whereas a musket-and-bayonet unit might benefit significantly from firing and then charging through and out of its own obscuring smoke into a terrified and confused enemy.
3. And for animals that might do so, there’s a reason that for hunting something like a boar there were specialized _spears_ to deal with an angry charging one.
4. This, it seems to me, should be possible with modern technology, to simulate each arrow’s physics reasonably accurately, but history research is almost never so well funded as to be able to do that kind of work and right now the federal funding for history research, the National Endowment for the Humanities, is facing cuts and possible extinction, rather than being expanded.
5. On arrow dodging in a Native North American context, see Lee, “The Military Revolution of Native North America: Firearms, Forts and Polities” in _Empires and Indigenes_ , (2011), 58, fn. 34.
6. That said, even for a fellow with full plate protection, there are points of vulnerability: not every component is as thick as the breastplate, but the big worry are the eye-slits and breaths (breathing holes) in the visor. These are generally small enough that an arrow can’t get through whole, but an arrow striking one might shatter, sending deadly or debilitating sharp splinters through. Note, for instance, these experiments by Tod Todeschini.
7. Roman helmets, like medieval helmets, were worn with padded textile liners, which would absorb some of the impact, but an 50-80 joule head impact is still going to hurt quite a lot!
8. 100 * 0.25 (miss) * 0.25 (shields) * 0.25 (helmets, greaves) * 0.33 (non-disabling hits (arms, legs, feet) = 0.52%
9. Of course the effectiveness of bows in sieges is that attackers looking to set ladders or scale walls are going to need to be vulnerable for a lot longer and a well-defined fortification is going to enable ‘enfilade’ fire (shots coming from the sides, via projecting towers), all of which means the attackers have to sit in that higher lethality point-blank-range zone for a lot longer.
10. Hdt. 6.114-117. I have seen online many times the figure of 11 dead for the Plataeans cited to Hdt 6.117, but the figure is not there in the passage and I do not know where it is from.
11. On the battle, a good and readable primer is M. Livingston, _Agincourt: Battle of the Scarred King_ (2023).
12. I want to note, because we’re trying to see how archers work when everything is _going well_ , we’re focusing on English victories, but English longbowmen did not always win their battles either: sometimes the French infantry and cavalry _were_ able to close the gap and win, as for instance at Formigny (1450). For the longbowmen to succeed, they needed quite a lot to go right for them.
13. Livingston, _op cit_ , 253
14. I will note that Livingston breaks this simulation into ‘volleys,’ but I don’t think he means they’re actually firing in volleys (they don’t seem to have been, the sources describe ‘clouds’ and ‘hail’ of arrows, implying continuous shooting), it’s just a useful set of mathematical divisions to break his lethality ramp into.
15. But consider reading Spring, _With Zeal and With Bayonets Only: The British Army on Campaign in North America, 1775-1783_ (2008). I also have with me, but haven’t yet read, but have very high hopes for A. S. Burns, _Infantry in Battle: 1733 – 1783_ (2025), so we might return to this topic at a later date.
### Like this:
Like Loading...
