The Arrow in King Harold's Eye: The Legend That Just Won't Die

By Michael Livingston

The Battle of Hastings is almost assuredly the most famous event in English history: amid dates popularly known in the West, the year 1066 is rivaled only by 1492.

Today, the artifact most associated with William of Normandy’s 14 October 1066 defeat of Harold Godwinson is the Bayeux Tapestry, and the most recognizable image in this nearly 70-meters-long embroidered cloth is undoubtedly that of the supposed death of Harold, the last Anglo-Saxon king of England: a man, clad in mail and a standard helm, armed with a shield and spear, standing with an arrow protruding from his face. For most people, 1066 and all that comes down to this one iconic moment: King Harold taking it in the eye.

As a scene, it is a moment of awful fate, now indelibly etched in our popular imagination. It is a moment of high drama and horrible tragedy.

It is also a moment that did not happen.

This present study is hardly the first of its kind to make this claim, though several points of evidence in its favor are presented here for the first time. Most notable among recent arguments against the veracity of this vision of the death of King Harold is that of Martin Foys, who in 2009 penned a detailed refutation with the magnificent title “Pulling the Arrow Out” (and whose translations for certain primary sources are used herein). Still, despite such efforts, the legend of the king’s death-by-arrow persists. It is, one might say, a legend that just won’t die.

But it should. First, there is very good reason to doubt that Harold was actually killed by an arrow in the eye. Our earliest sources, as we will see, make very different claims about his fate. Second, beyond that bit of historical sleuthing, there is this: the man being struck with the arrow in the Bayeux Tapestry — the popular authority for viewing Harold’s death this way — almost assuredly wasn’t being struck with an arrow before nineteenth-century restorers made it so. And third, even if this unnamed figure was being struck with an arrow in the original form of the tapestry — which he is not, but even so — he is most certainly not Harold Godwinson.

The Missing Arrow

Pinning down the identity — not to mention the date — of our earliest known account of the Battle of Hastings is a surprisingly difficult task. Depending on how one dates the Anglo-Saxon Chronicle — a frightful literary-historical knot — it may be that this record is nearly contemporaneous. Even if so, however, it is of little use in the present problem; its most detailed version of the battle (MS E) says only that King Harold was killed. It does not say how:

Ond Harold com northan ond him with feaht ear than the his here com eall, ond thaer he feoll ond his twaegen gebrothra Gyrth ond Leofwine” [And Harold came from the north and fought with him [William] before all the army had come, and there he fell and so did his two brothers Gyrth and Leofwine].

Another candidate for the honor of being the earliest account is the 835-line poem Carmen de Hastingae Proelio [Song of the Battle of Hastings], a Latin work that exists in a single manuscript copy of the early twelfth century (Bibliothèque royale de Belgique MS 10615-729, fols 227v-230v). Some researchers (largely following a 1978 article by R.H.C. Davis) believe that the poem is of twelfth-century origin. That said, other scholars studying the Carmen (including its editors, Catherine Morton and Hope Muntz) have argued that its origin lies earlier, in the eleventh century. They identify the work with an otherwise lost poem by Bishop Guy of Amiens, which would neatly date the Carmen’s composition between Christmas 1066, when the victorious William the Conqueror was crowned king of England, and Easter 1067, when the poem was probably presented to William while he was in Normandy to further the cause of Guy’s family.

The Carmen is not, to be sure, the finest of Latin poems. It appears to have been written in some haste. What the Carmen lacks in its technical artistry, however, it makes up for with its remarkable imagery. Its description of the death of Harold, for instance, is graphic in detail as William, Eustace of Boulogne, Hugh of Pontheiu, and Giffard all bear down upon him:

Regis ad exicium quatuor arma ferunt. / Ast alii plures; aliis sunt hi meliores. / Si quis in hoc dubitat, action uera probat: / Per nimias cedes nam bellica iura tenentes / Heraldum cogunt pergere carnis iter. / Per clipeum primus dissoluens cuspide pectus, / Effuso madidat sanguinis imbre solum; / Tegmine sub galee caput amputat ense secundus; / Et telo uentris tertius exta rigat; / Abscidit coxam quartus; procul egit ademptam: / Taliter occisum terra cadauer habet.

[These four bore arms to kill the king. Others indeed were there; but these were better than the rest. If anyone doubts this, what they did proves it true, for in accordance with the rules of war they compelled Harold by many blows to go the way of all flesh. The first of the four, piercing the king’s shield and chest with his lance, drenched the ground with a gushing stream of blood. The second with his sword cut off his head below the protection of his helm. The third liquified his entrails with his spear. And the fourth cut off his thigh and carried it some distance away. The earth held the body they had in these ways destroyed.]

Harold, in other words, is hacked to death. The Carmen makes no mention at all of an arrow striking the Anglo-Saxon king in the eye — or, for that matter, anywhere else.

As it happens, our other contender for the earliest documentary source, the Gesta Normannorum Ducem [Deeds of the Norman Leaders] of William of Jumièges, is similarly arrow-less. Written in the late 1060s, it simply reports that “Heraldus etiam ipse in primo militum congressu occubuit vulneribus letaliter comfossus” [Harold himself was slain, pierced with mortal wounds, during the first assault]. Several medieval writers considered this source authentic enough to recycle into their own accounts, including the twelfth-century Orderic Vitalis.

Slightly more detail is available from William of Poitiers, in his Gesta Guillelmi [The Deeds of William], which was written in the 1070s (perhaps as early as 1071): “At the close of the day,” he writes, “the king himself, his brothers, and the leading men of the kingdom had been killed.” Later, he describes how William “returned to the battlefield,” and he found that “the ground was covered with corpses for a vast distance, stained with blood: they were the flower of the nobility and youth of the English. Beside the king, two of his brothers were found: he himself, stripped of all marks of his rank, was recognized not by his face, but by certain signs.” Like the poet behind the Carmen, William of Poitiers appears to point to a bloody and merciless death-by-melee for Harold.

Much the same is in the majority of our accounts. Writing in the first quarter of the twelfth century, John of Worcester says of Harold’s demise only that “Heu, ipsemet cecidit crepusculi tempore” [he himself fell, alas, at dusk]. The local chronicle of Battle Abbey, which was established by William on the very grounds of his victory, states that “Regeque suo belli fortuito ictu prostrato, per diversa dissiliens latebras querit” [When their king was laid low by a chance blow, the army broke up and fled in different directions to find hiding-places]. And, in a particularly fascinating account, the Vita Haroldi (written ca. 1205), which eventually tells a tale of Harold surviving the battle, specifically rebukes the arrow-in-the-head report of William of Malmesbury, which we will examine below. This rumor, the anonymous writer says, is a malicious attack on truth.

The Bayeux Tapestry

Despite the fact that not one of these early accounts accepts the familiar arrow-in-the-eye image, it is clearly there today, in glorious artistry, embroidered on the Bayeux Tapestry, which is usually attributed to the oversight of Bishop Odo around the year 1077. In Scene 57, one of the last sections of the work, the tapestry provides the now-famous moment of a spearman being struck with a golden arrow in the eye. Standing beside the man is a standard-bearer holding aloft the dragon of the kings of Wessex, and beside him, in turn, is another standing spearman and another Wessex standard-bearer — this one dead. The previous panel, we might note, carries the label “Et ceciderunt qui erant cum Haroldo” [And those who were with Harold have fallen].

Returning to the man with the golden arrow, we can observe that behind him a Norman on horseback is in the process of hacking at another fallen man, who has dropped his axe. Above the striking scene of this set of men are the words “Hic Harold rex interfectus est” [Here King Harold had been killed]. And so, it has long been assumed, the spearman with the golden arrow in his eye is Harold.

Intriguingly, however, we know that the Bayeux Tapestry has undergone several bouts of restoration. This section in particular was much deteriorated. As scholars N. P. Brooks and H. E. Walker summarize it:

The scene of the death of HaroId is not in its present form the work of the designer of the Tapestry, but of a nineteenth-century restorer. Early in the nineteenth century the ends of the Tapestry were badly damaged by the winch on which it was wound and unwound. In 1842 the Tapestry was restored in machinespun wools whose dyes are distinct from the original medieval embroidery. Enough of the inscription of the scene of the killing of Harold is in original wools to show that it has been correctly restored; but the figures have been almost entirely renewed. Only the head and shoulders of the arrow-in-the-eye figure is original; the arrow and most of the body are restoration. The Norman horseman chopping down the dying king is also mostly in modern wools; whilst of the falling figure only some of the head (but not the moustache) and a little of the mail have survived intact.

We are most fortunate that sketches were made of the embroidery in the eighteenth century, before this major restoration effort. What they reveal, among other things, is that the arrow in Harold’s head may not have always been an arrow. Around 1729, the French historian Bernard de Montfaucon commissioned Antoine Benoît to make a detailed sketch of the tapestry as it then appeared. This sketch was then used as the source for two separate published engravings: one by Montfaucon in 1730, and one by Antoine Lancelot in 1733. Together, these are our earliest surviving depictions or descriptions of this pivotal scene in the embroidery — and not one of them has the spearman in question receiving an arrow in the eye. The spearman has his hand raised, as he still does today, but it is not an arrow that he holds. It is, fittingly, a spear or, at best, a long arrow shaft without fletchings of any kind.

Another clue that the current depiction of the golden arrow is not original can be found from a close look at the arrow itself: its fletchings are distinctly different from the three arrows that are stuck in the man’s shield (these others are in the 18th-century engravings). In fact, the fletchings are different from not just this set of arrows, but every other set of arrows in the tapestry. Even more telling, in order to get the arrow to fit the angle of the man’s hand and still come close to hitting the man in the eye (it really seems to be striking his helmet), the shaft of the arrow is quite unnaturally bent.

In 1816, due to increased interest in the embroidery, Charles Stothard was commissioned by the London Archaeological Society to make a color copy of the Bayeux Tapestry. Published between 1819 and 1823, his reproduction included his own hypothetical reconstructions of parts of the work that were damaged or missing. It is here that an arrow first appears.

Unlike Benoît and the earlier depicters of the scene, Stothard makes clear his belief that this unfortunate spearman (and the axe-man behind him, too) was Harold. In his report on the tapestry, he describes “the death of Harold, who appears first fighting by his standard-bearer, afterwards where he is struck by the arrow in his eye, and lastly where he has fallen, and the soldier is represented wounding him in the thigh.”

Frustratingly, it is unclear whether the arrow is Stothard’s own invention or his accurate rendering of restorations done between the mid-18th-century engravings and his own day. The Bayeux Tapestry had undergone some repair in the late 18th century, for instance, due to damage it suffered in the Napoleonic period. For all we know, the arrow could have been added at any point between the time of Montfaucon and that of Stothard.

It is an arguably trivial but fascinating aside that, despite this uncertainty, in 1907 Charles Dawson accused Stothard of manipulating the record when it came to the sudden appearance of the arrow. In 1979, the aforementioned Brooks and Walker then took Dawson to task, observing that he could not himself be trusted since he was, after all, the perpetrator of one of the most famed modern forgeries: the Piltdown Man. Dawson is indeed a suspicious character (the Piltdown Man is only a drop in the bucket of his misdeeds), but he is nevertheless correct to observe that, in the end, the only physical evidence we have that there ought to be an arrow in this poor spearman’s face is our trust that Stothard was right to draw it there. In their attack on Dawson, Brooks and Walker contend there is “no reason to doubt that Stothard’s engravings were an honest and accurate work,” but this may be casting stones in a glass house. Stothard, we now know, not only poured hot wax on parts of the tapestry to make plaster casts for himself, but he physically cut at least two pieces from this magnificent work of art and stole them. (Stothard’s wife, Eliza, was long thought to be the vandal, but recent work by Michael Lewis has definitively pointed to the artist himself.)

The Bayeux Tapestry underwent a series of major restorations in 1842, which Brooks and Walker (among many scholars) claim “were directly based on Stothard’s ‘reconstructed’ engravings.” Whether or not this is so (though reported again and again in the scholarship, I have found no official statements that Stothard’s work was utilized in such a guiding way), there is no question that by the middle of the nineteenth century, the Bayeux Tapestry has shown what we have today: a golden arrow, bending most unnaturally into the face of a most unfortunate Anglo-Saxon spearman.

Identifying Harold

If we take away that hardly-straight-as-an-arrow arrow — if we replace it with what appears to have been a spear originally — what reason do we have to think that this poor fellow is King Harold?

There is the label on the tapestry itself, of course: “Hic Harold rex interfectus est” [Here King Harold had been killed]. But this statement doesn’t mean that the spearman is Harold. To the contrary, the label could just as easily refer to the falling axe-man: the words, after all, are awkwardly smashed together — leaving the spearman well behind — in order to line up with the head of this second man, who is being hacked to death in exactly the way that the early sources describe happened to Harold, and who is using the very weapon that Harold’s loyal housecarls (and perhaps Harold himself) rather famously used in battle. Even more, a look at Benoît’s early sketch shows that much of the word ‘interfectus’ was at that time missing. As a result, M. K. Lawson has recently argued that it originally read “Hic Harold rex in terra iactus est” [Here King Harold has been thrown to the ground], which would precisely fit the depiction of the axe-man.

Some scholars, noting this indicative location of the labeling — whatever it once read — have argued that the spearman and the axe-man are both Harold. In support of this position, the modern viewer can still today see, without magnification, a line of holes — almost assuredly from removed stitching — that could well be the remnants of an arrow striking the axe-man. Quality photographs of these marks from both the front and the back of the Tapestry appear in the work of David Bernstein, who regards them as evidence of an original arrow here. Why and by whom this arrow was removed — if it was indeed an arrow — he cannot say, but there is a still greater problem with any attempt to regard these holes as original to the work. Not one of those studying this important scene in the 18th and 19th centuries — including Stothard, who we know believed that this man was Harold — even hint at an awareness of such an arrow or any marks thereof, despite their obviousness to the naked eye. Without any further evidence coming to light, we surely must conclude that the marks are the work of an over-zealous “restorer” who added the arrow. It would stand to reason, after all, that if both men were thought to be Harold, the fallen one ought to have an arrow just like the standing one. It was thus added, only to be taken away when someone, realizing its absence from the prior visualizations, removed it.

Plus, there is essentially no chance that the spearman and the axe-man are the same man. Aside from the simplicity of the fact that they wield different arms — the shield-bearing spearman has no axe and the axe-man has neither spears nor shield — there is a long list of discrepancies. They aren’t wearing the same clothes. They are both wearing mail, but the leggings beneath this are different. It’s hard to conceive that those involved in the embroidery would know the two men are the same and then utterly fail to dress them the same even when they are beside each other. The same is true for their helms, which are entirely different colors. Their sword scabbards, too, don’t match: the spearman’s is a simple black, while the axe-man’s is a far more regal-seeming gold.

And there is this: the spearman is wielding his weapon right-handed, which would be proper for any common member of the Anglo-Saxon army. After all, their primary battle tactic of forming a working shield wall — which we know they executed at Hastings — required such uniformity. As it happens, though, when we look closely at the axe-man, it appears that he wields his weapons left-handed: his golden scabbard is hanging off his right hip. He would surely be an odd-man out amid the common men on the shield wall! Adding to the intrigue of this fallen left-handed man, it is worth noting that several earlier scenes in the embroidery indicate that Harold is left-handed. He is shown this way when William presents arms to Harold after their joint defeat of Conan, for instance. And when noblemen, after the death of King Edward, offer Harold the crown of England and an axe to symbolize royal authority, Harold is greeting them with — yes — an axe in his left hand. Whether such depictions reflect reality, we cannot know. The least we can say, however, is that the Bayeux Tapestry is invested in depicting Harold — the future last Anglo-Saxon king of England — as a lefty, just like the axe-man.

In sum, whether he ought to have an arrow in his face is somewhat immaterial to the more important conclusion that the standing spearman isn’t Harold. The fallen axe-man is.

The Arrow in the Eye

That said, the idea that Harold took it in the eye has a long history. What could be the earliest reference to Harold being struck by an arrow in the eye is the account of Amatus, a monk in the southern Italian abbey of Montecassino, whose History of the Normans records that William “gouged out his [Harold’s] eye with an arrow” (46). This source is often dated to the early 1080s, though in truth the original Latin text is now lost. It exists only in a fourteenth-century translation into French that may or may not be wholly accurate. More reliable in terms of its textual history is a poem written in the first years of the twelfth century for William’s daughter by Abbot Baudri of Bourgueil. Harold, he writes, was struck and killed by an arrow — though where he was struck, and to what effect, goes very much unsaid.

William of Malmesbury, in his Chronicle of the Kings of England (1118), writes that Harold “fell, from having his brain pierced with an arrow … receiving the fatal arrow from a distance, he yielded to death. One of the soldiers with a sword gashed his thigh, as he lay prostrate; for which shameful and cowardly action he was branded with ignominy by William, and expelled from the army.” This still isn’t an arrow to the eye, but it is, we might say, getting close. Then, sometime after 1123 (likely 1130-40), Henry of Huntingdon’s Historia Anglorum tells a similar tale that — at last! — specifies the blow to the eye: “et ipso in oculo ictus corruit” [and he himself was pierced in the eye], Henry says, after which point the wounded king was overrun by knights.

Around 1155 we get the most detailed story of all in Wace’s Roman de Brut, which attempts to combine the many earlier traditions (he is particularly drawn to William of Malmesbury) by having Harold struck by an arrow early, then hacked to death later. His account makes for grand theater (which by its nature might indicate a drift toward ahistorical rhetorical elaboration), and as such is worth seeing in full:

Then it was that an arrow, that had been thus shot upwards, struck Harold above his right eye, and put it out. In his agony he drew the arrow and threw it away, breaking it with his hands and the pain to his head was so great, that he leaned upon his shield. So the English were wont to say, and still say to the French, that the arrow was well shot which was so sent up against their king; and that the archer won them great glory, who thus put out Harold’s eye. … And now the Normans had pressed on so far, that at last they reached the standard. There, Harold had remained, defending himself to the utmost; but he was sorely wounded in his eye by the arrow, and suffered grievous pain from the blow. An armed man came in the throng of the battle, and struck him on the ventaille of his helmet, and beat him to the ground; and as he sought to recover himself, a knight beat him down again, striking him on the thick of his thigh, down to the bone…. The standard was beaten down, the golden gonfanon was taken, and Harold and the best of his friends were slain; but there was so much eagerness, and throng of so many around, seeking to kill him, that I know not who it was that slew him…. I do not tell, and I do not indeed know, for I was not there to see, and have not heard say, who it was that smote down king Harold, nor by what weapon he was wounded; but this I know, that he was found among the dead. His great force availed him nothing; amidst the slain he was found slain also.

Though clearly a minority tradition among the literary sources, these late stories are so similar to the depiction of the Bayeux Tapestry as it now stands — an arrow in the eye of a man, then a horseman slashing him as he falls — that it’s no surprise that many scholars have assumed that William of Malmesbury (and thus the writers like Henry and Wace who followed him) received his information from the famed embroidery. Brooks and Walker, for instance, are certain that “William had seen the Tapestry itself at Bayeux, or just conceivably had used at Canterbury either the Tapestry designer’s sketches or perhaps even the account that the designer followed.” Yet while it is tempting to imagine that the current tapestry’s closeness to William’s account is evidence that he got his details from it, the obverse can be just as equally true: the closeness could mean that a restorer of the tapestry had read William. Indeed, since we have seen that the spearman and the axe-man cannot both be Harold, and that the arrow in the eye of the former is almost assuredly a mistaken effort at restoration, the latter position seems far more likely to be true.

Still, Baudri and William (who seem to be independent traditions) clearly received their death-by-arrowshot stories from somewhere. And popular though it clearly was, there are several reasons to be suspicious of the veracity of this tradition — even aside from the fact that our death-by-melee stories occur closer to the historical event. Chris Dennis, for instance, argues that William’s court might have had good reason to begin circulating (or at least supporting) stories that an arrow killed the former king:

If Harold had been killed by a fateful arrow, his death could be directly attributed to the will of God. Duke William and the Normans were exonerated from any blame, and the Anglo-Saxons could be appeased to some extent by the portrayal of their king as a hero, since he could only be killed at a distance by a chance arrow.

Perhaps this is so, though it assumes that William and his court at some level acknowledged Harold’s position as a rightful king and were thus worried about his killing — when, to the contrary, there is precious little evidence that they viewed him as anything but a usurper who by all rights deserved to be hacked to pieces. If anything, it could be that the symbolism of the arrow was meant to underscore a fittingly ignoble end: he wasn’t killed by a noble, but by a random, run-of-the-mill low-class archer. (Though, as to that, William Rufus was killed by an arrow while hunting, and no one seemed to see this as anything but an unfortunate accident.)

Setting aside political conspiracy theory, the most troubling issue with the tradition that King Harold died from an arrowshot is its eyebrow-raising echo to the contemporary death of a homophonous king. On 25 September 1066 — less than three weeks before his death at Hastings — Harold Godwinson defeated Harald Hardrada at the Battle of Stamford Bridge. Early in that battle, Harald was killed when an arrow struck through his throat. It is highly probable (and arguably inevitable) that the deaths of these two kings Har(a/o)ld so close to one another would quickly elicit contamination in the historical record. And since Harald’s time of death is consistently reported in independent records, it would seem far more likely that his death by arrow was transferred to Harold than the other way around. That an arrow was ever sent Harold’s way, in other words, might be simple confusion with the death of Harald.

We are left with a relatively simple choice. Either (1) Harold and Harald were both improbably struck down by an arrow (so Baudri et al), or (2) Harold was not struck down by an arrow at all (so the earliest sources). Probability, one might say, favors the second choice. Baudri and William (or whatever source or sources they had) have confused kings and thence confounded a millennium of scholars. More than that, they set the stage for someone, somewhere between the 1730s and Stothard’s trip to Bayeux in 1816, to add a corresponding arrow to the Bayeux Tapestry. Whether the mistaken restorer was Stothard or someone before him we do not have the facts to say. To that matter, however, it is intriguing to note that Rev. John Sharpe’s English translation of William of Malmesbury was published to some fanfare one year before the artist journeyed to Bayeux, and that the language of Stothard’s aforementioned description of the scene in the tapestry very much points to him having a pre-conceived tale of the king’s death in mind and is matching up the images to it. That tale, as it happens, fits that of Malmesbury quite perfectly.

Harold Godwinson died at Hastings. Legends of his miraculous survival notwithstanding, this much seems fact. Our concern shouldn’t be whether he truly took it in the eye — as I hope we have seen, he almost assuredly did not — but why and how such a story took aim upon him. Anything else is missing the point.

Michael Livingston teaches at The Citadel and is the author of numerous books on medieval history as well as fiction novels. You can learn more about Michael on his website, or follow him on Twitter @medievalguy