The Battle of Stamford Bridge ... at Holtby

1066 was a hectic year. The Battle of Hastings gets most of the press, but the major events were all linked. William could not have taken the English Crown if Edward the Confessor had not died when he did, or if London and Winchester had chosen to resist, or if Harald Hardrada had not invaded Northumbria. Arguably then, Hardrada’s invasion which culminated in the Battle of Stamford Bridge was effectively a phase of the Norman Conquest, just as momentous to English history as the Battle of Hastings.

Figure 1: Vale of York geography in 1066. Y= York, S=Stamford Bridge, T=Tadcaster, B=Brough,
white dotted line=glacial moraine ridges, black lines=Indication of main Roman roads

By tradition, the Battle of Stamford Bridge was fought at Battle Flat, 850m east of the current bridge at Stamford Bridge in Yorkshire. There is no physical supporting evidence. Burne says that “swords and other battle debris” get ploughed up from “time to time” at Battle Flat. Confused perhaps. As far as we know, no battle related finds have ever been recorded anywhere. The western side of the traditional battlefield was excavated during a housing development in 1998. They made no medieval finds, military or otherwise. It was not that they had corroded away either, because older Romano-British finds were unearthed.

In our opinion, the battle was fought somewhere else. For one thing, Battle Flat is an unlikely battlefield: flat, featureless and boggy. Norse commanders prided themselves on their ability to choose favourable battlefields. It would defy 500 years of Norse military practice if Hardrada chose to defend Battle Flat. For another, it is in the wrong direction from the bridge. The Norse fleet must have been moored south of Stamford Bridge, for a long list of reasons we will return to below. If the main Norse army was attacked at Stamford Bridge or if they retreated to Stamford Bridge, they would surely have headed south, towards their ships, safer ground and reinforcements, whereas Battle Flat was east of the traditional bridge location.

In this blog we will review the evidence that the battle happened at Battle Flat then explain where we think the battle is more likely to have been fought and why. If you would rather read it as an ebook, it is available on Amazon, ASIN B08CM5MRN4. Alternatively, you can download it as a PDF by clicking this link.

Before starting, we have a couple of notes about terms. Harald is the Norse spelling of Harold, making for potential confusion between King Harold Godwinson of England and King Harald Sigurdsson of Norway. We will generally refer to Sigurdsson by his epithet 'Hardrada', meaning 'hard ruler'. The Norse sagas, on occasion, refer to both kings as Harald. We have edited their names in translations below to prevent confusion. For the sake of brevity, we will generally refer to the belligerents as English and Norse. They would be horrified. There was a large Flemish cohort among the Norwegian invaders. The English would have thought of themselves as Angles and Danes.

It is helpful to be familiar with the geography. Figure 1 is a topology heat map of the Vale of York as it would have looked in 1066. Most of the action took place in the central belt between York (Y) and Stamford Bridge (S). It was boggy fenland in the 11th century. Even today - after a 5m fall in sea level, a latticework of relatively modern drainage ditches and the installation of modern pump defences - it still floods after heavy rains, disastrously so as recently as 2000 and 2015. It is not much better in the summer. We tried to walk the Minster Way from York to Beverley in relatively dry 2008 but had to give up before reaching Stamford Bridge because the claggy ground made it like walking with 5kg weights attached to our feet.

The only practical way for 11th century armies to get around was on Roman roads and some ancient ridgeways. The approximate routes of the Roman roads, based on the work of Ivan Margary, are shown as black lines on Figure 1 and Figure 2. We will refine this below. There were two relevant stretches of glacial moraine (white dotted lines on Figure 1) that probably carried roads or ridgeways. The more northerly arced east and north from Fulford to Hutton via Holtby. Some refer to it as the York ridge. We prefer the name Fulford ridge because the part west of the Ouse is irrelevant here. The more southerly arced east and north from Escrick to High Catton via Wheldrake. It had a 1km break just east of Wheldrake where the River Derwent flowed through.

Lower Vale of York, now known as the Humberhead Levels, had no land routes. It was a tidal saltmarsh that was yet to be inned. A few islands stood proud of the marsh, most notably those big enough to be inhabited and farmed at Riccall, Bubwith, Cottingley and Barmby in the Marsh.

The Traditional Narrative

In the absence of physical evidence, the traditional narrative had to be based on circumstantial evidence, including 11th century geography (Figure 2), medieval warfare tactics and, above all, contemporary accounts.

Figure 2: Vale of York geography in 1066. Y= York, S=Stamford Bridge, R=Riccall, T=Tadcaster, H=Holtby, K=Kexby, U=Bubwith, M=Barmby-in-the-Marsh, B=Brough, W=Wheldrake, SD=Sutton upon Derwent, EC=East Cottingwith, E=Escrick, Marsh=Humberhead Levels, black lines=Indication of major Roman roads

Seven contemporary accounts contain credible unique information about the battle. They are all short, just a few paragraphs at most. Three are English, four are Norse. The English accounts seem to corroborate each other. The Norse accounts seem to corroborate each other. Unfortunately, they do not always agree, most notably insofar as the Norse accounts imply that the main battle was several miles west of the River Derwent whereas the English accounts suggest that it was at Stamford Bridge or east of the Derwent.

The Norse accounts are in the 'King's sagas'. English historians do not trust them. Professor Freeman once lamented they were: “hardly more worthy of belief than a battle-piece in the Iliad”. He is right up to a point. They were written 160 years after the battle, primarily to promote Norse culture with only an incidental interest in historical accuracy. Everyone accepts that they contain a lot of exaggeration and poetic license. Hence, the traditional narrative is based on the more prosaic English accounts. But they are not much better. They have been so severely abridged that their meaning is obscured. One of them does give the impression that the main battlefield was east of the river near Battle Flat, but that might only be because the narrative has been distorted through its abridgement.

There is more consensus about the battle prelude. On the 18th September, Harald Sigurdsson, king of the Norwegians, and Tostig Godwinson, exiled younger brother of English King Harold, brought a Norwegian and Flemish army to the mouth of the Humber. The fleet was 300 to 350 ships carrying roughly 10,000 troops plus a thousand or so followers. They moored at Riccall (R on Figure 2) on the River Ouse. Two days later, they made their way north to attack York. They met and defeated a Northumbrian army at Fulford (F). A truce was agreed in which the invaders agreed not to sack York, as long as the townsfolk provided them with food and fighters. The Norse army withdrew.

King Harold resolved to expel the invaders. He brought the English army from London, entering the battle theatre on the Roman road shown in the southwest corner of Figure 2. He mustered with his naval forces at Tadcaster (T). On the day of battle the main Norse army was at Stamford Bridge (S) heading towards York. King Harold attacked from the direction of York.

English tradition is based on three contemporary accounts written within 60 years of the battle. The earliest, probably written within a decade of the battle and therefore the most trusted, is the Anglo-Saxon Chronicle. This is what the Worcester recension (ASC-D) has to say in the Ingram translation:

In the midst of this came Harold, king of the English, with all his army, on the Sunday, to Tadcaster; where he collected his fleet. Thence he proceeded on Monday throughout York. But Harald, king of Norway, and Earl Tosty, with their forces, were gone from their ships beyond York to steinford brygge; for it was given them to understand that hostages would be brought to them there from all the shire. Thither came Harold, king of the English, unawares against them beyond the bridge ; and they closed together there, and continued long in the day fighting very severely. There was slain Harald the Fair-haired, king of Norway, and Earl Tosty, and a multitude of people with them, both of Norwegians and English, and the Norwegians that were left fled from the English, who slew them hotly behind; until some came to their ships, some were drowned, some burned to death, and thus variously destroyed; so that there was little left: and the English gained possession of the field.

A later folio has been added to one of the ASC manuscripts which says:

But there was one of the Norwegians who withstood the English folk, so that they could not pass over the bridge, nor complete the victory. An Englishman aimed at him, with a javelin, but it availed nothing. Then came another under the bridge, who pierced him terribly inwards under the coat of mail. And Harold, king of the English, then came over the bridge, followed by his army ; and there they made a great slaughter, both of the Norwegians and of the Flemings.

John of Worcester’s account appeared in 'Chronicon ex Chronicis', dating from the early 12th century:

Harold Harfaager, king of Norway, brother of St. Olave the king, suddenly arrived at the mouth of the river Tyne, with a powerful fleet of more than five hundred great ships. Earl Tosti joined him with his fleet, as they had before agreed, and they made all sail into the Humber ; and then ascending the river against the current, landed their troops at a place called Richale. As soon as king Harold received this news, he marched with all expedition towards Northumbria ; but, before the king's arrival, the two brothers, earls Edwin and Morcar, at the head of a large army, fought a battle with the Norwegians on the northern bank of the river Ouse, near York, on the eve of the feast of St. Matthew the Apostle [20th September], being Wednesday ; and their first onset was so furious that numbers of the enemy fell before it. But, after a long struggle, the English, unable to withstand the attack of the Norwegians, fled with great loss, and many more of them were drowned in the river than slain in the fight. The Norwegians remained in possession of the field of death ; and, having taken one hundred and fifty hostages from York, and leaving there one hundred and fifty hostages of their own, returned to their ships. However, on the fifth day afterwards [Monday], the seventh of the calends of October [25th September], Harold, king of England, having reached York, with many thousand well-armed troops, encountered the Norwegians at a place called Stanford-brycge, and put to the sword King Harald and Earl Tosti, with the greatest part of their army ;  and, although the battle was severely contested, gained a complete victory. Notwithstanding, he allowed Harald's son Olaf, and Paul, earl of Orkney, who had been left with part of the army to guard the ships, to return to their own country, with twenty ships and the relics of the defeated army ; having first received from them hostages and their oaths.

Henry of Huntingdon’s account appeared in 'Historia Anglorum', which also dates to the early 12th century:

Then they joined their forces and came up the Humber, as far as York, near which they were encountered by the Earls Edwin and Morcar ; the place where the battle was fought is still shown on the south side of the city. Here Hardrada, king of Norway, and Tostig, his ally, gained the day. When this intelligence reached Harold, king of England, he advanced with a powerful army, and came up with the invaders at Steinfordesbrige. The battle was desperately fought, the armies being engaged from daybreak to noonday, when, after fierce attacks on both sides, the Norwegians were forced to give way before the superior numbers of the English, but retreated in good order. Being driven across the river, the living trampling on the corpses of the slain, they resolutely made a fresh stand. Here a single Norwegian, whose name ought to have been preserved, took post on a bridge, and hewing down more than forty of the English with a battle axe, his country's weapon, stayed the advance of the whole English army till the ninth hour. At last someone came under the bridge in a boat, and thrust a spear into him, through the chinks of the flooring. The English having gained a passage, King Harold and Tostig were slain ; and their whole army were either slaughtered, or, being taken prisoners, were burnt.

All three of these accounts are pretty unhelpful about the main battle, which they abridge into a couple of sentences that can be paraphrased, “the fighting was intense, lasting most of the day with many casualties on both sides, including fatal injuries to Hardrada and Tostig, before the English won”.

As for the battlefield location, English historians (and Canadian DeVries) put their trust in the ASC-D statement that the Norse army was at Stamford Bridge on the morning of the day of battle and that the battle happened “beyond the bridge”. They assume that this bridge was the one at Stamford Bridge and that "beyond" is from King Harold's perspective. If so, the main battle was near to and east of Stamford Bridge, which is to say, somewhere near Battle Flat. John of Worcester just says that the battle was at Stamford Bridge, which is unhelpful but not contradictory.

Huntingdon says that the initial encounter was at Stamford Bridge and that the Norse army was then driven over the river. The English were coming from York, to their west. Therefore, he is saying that the initial encounter was west of the Derwent, that the Norse were driven over the bridge, then made a "fresh stand" east of the river. English historians interpret this to mean that the initial encounter west of the Derwent was too insignificant to be reported by the ASC-D. They think the main battle happened where the Norse army made their "fresh stand" east of the river. It is a plausible explanation that unifies the three English accounts, and which has satisfied historians for centuries.

Frank McLynn (Figure 3) and Kelly DeVries (Figure 4) have created diagrams to show the traditional troop deployments. McLynn's is based on Huntingdon's account, with the initial Norse shield wall depicted as an arc west of the river. DeVries's is based on the ASC-D account, with no initial encounter west of the river.

Figure 3: Frank McLynn, battle scenario
Figure 4: Kelly DeVries, battle scenario

There are differences in how the various studies describe the action east of the river. McLynn’s diagram shows the Norse shield wall at about 800m from the river, which agrees with the Ordnance Survey battlefield marker, whereas DeVries shows it at about 400m. McLynn’s text, however, says it is about 300m from the river. Burne’s diagram shows it at about 600m. Barrett and Leadman show it initially on the east riverbank, then retreating back to about 300m from the river.

As DeVries points out, the only factor that ties all the studies together is the place name Battle Flat. In the absence of archaeological evidence that hints at a battlefield elsewhere, it would be career threatening for professional historians to argue that the battle was fought anywhere other than Battle Flat.

This is analogous to the Battle of Hastings, where the traditional narrative is based on seven contemporary accounts which say or imply that the battle was fought at Battle Abbey. No matter how unlikely it is as a battlefield, it would be career threatening for professional historians to argue against a Battle Abbey battlefield when there is no archaeological evidence of it having happened elsewhere.

We are not historians, so going against the grain does not bother us. We intend to use much the same arguments against Battle Flat that we use against Battle Abbey.

Debunking the Battle Flat battlefield

The historical studies fall into two camps. One group think the battle happened roughly 400m from the river (red rectangle on Figure 5), the other that it happened roughly 800m from the river (magenta and blue rectangles). The difference might seem trivial and unimportant. We think not. The reason is the slightly raised ground now in the Battleflats Way housing development, shown as a yellowish patch east of the bridge on Figure 5.

Figure 5: Battle Flat relief heatmap with traditional battlefields - Magenta=McLynn et al, Blue=DeVries et al, Red=Barrett et al

Studies that favour the battlefield near the river assume that the Norse fell back to occupy rising ground. Studies that favour the battlefield away from the river assume that the battle was fought at Battle Flat. Effectively, each reason is a prop for the other, but they are mutually exclusive. Battle Flat, as its name suggests, is flat. If the Norse fell back to occupy rising ground, the battle was not fought at Battle Flat. If the battle was fought at Battle Flat, the Norse did not fall back to occupy rising ground. Remove these props and the only evidence (albeit entirely circumstantial) militates against both battlefields.

Rising ground first. The higher ground - proposed as the battlefield by DeVries, McLynn's text and Barrett (red rectangle on Figure 5) - is only five metres above the riverbank. It is enough to help Hardrada and Tostig with command, but the slope is too shallow and the height difference too small to have had a significant impact on combat. Also, the higher ground is too small. Heimskringla's description of the Norse shield wall is enclosed, long and thin, perhaps only three deep. In order for all the Norse troops to be on rising ground at Stamford Bridge, an enclosed shield wall would have been some 800m long and 8 deep, which is neither long or thin. Given the choice of occupying this higher ground or retaining control of the bridge and riverbank, Hardrada would surely have taken the riverbank.

This is especially pertinent to the Norse narrative, which says that a third of the Norse army were left behind to guard the fleet on the day of battle. They were summoned as soon as Hardrada realises his danger but did not arrive until it was too late. There is no obvious reason why Heimskringla would invent this story. It does not reflect well on Hardrada's leadership qualities.

Yet, if the Norse army was at Stamford Bridge and the fleet guard were coming to the rescue, they would surely have done everything they could to defend the bridge (by tradition at the location shown in red on Figure 5). After all, if one Norseman could hold up the entire English army at the bridge, as ASC-D and Huntingdon state, a determined bridge guard could hold up the English army all day. If they had defended the bridge, there would have been plenty of time for the fleet guard to arrive and turn the course of the battle.

Even if the Norse narrative is wrong about the fleet guard, it is difficult to believe that they would voluntarily relinquish the bridge and riverbank. They might have fallen back 100m, to goad the English army over the bridge, where they could be shield charged to death in the river. But no further. Falling back any further would not only abandon the invaders' greatest military advantage, it would risk disaster because moving an army on boggy ground is dangerous. Stumbles are inevitable. Rather like a crash in the middle of the peloton at the Tour de France, each stumble presents the risk of a mass flattening and/or a mass crush as troops trip over each other. Once down, it would be tough to get back up, especially in bare feet or slick soled sandals, let alone get enough purchase to fight.

Leadman reports the widespread belief that the Norse fell back from the riverbank due to the shock that their bridge-blocking hero had succumbed. Bunkum. In a battle between 10000 and 1, it should not come as much of a shock when the one is defeated. And, even if it did, these were battle hardened Vikings for whom victory would gift the most valuable kingdom in Europe. There is no way the Norse would voluntarily fall back. There is no way the English could have driven them back. The English would obviously have crossed the Derwent upstream and tried to trap them on the riverbank. The most obvious reason that none of this happened is that the Norse were not at Stamford Bridge when they were attacked.

Battle Flat has more fundamental problems as the battlefield. The Norse army would have had to retreat over the raised ground to get to Battle Flat. They would then have had to fall back over an extra 400m of boggy ground, with its inherent dangers. Their stand would be at the one place that allowed the English army to occupy rising ground during the battle. It would not only help the English repel any counterattacks, but it would give the English commanders a good view of the battle while the Norse commanders were half blind.

Leadman suggests a combination of the two battlefields. He reckons that the battle started with the Norse shield wall on the rising ground 300m from the river, but that they got pushed backwards to Battle Flat, where they made their ‘fresh stand’. This could explain the pragmatic issues we just mention, but it does not match any of the contemporary accounts and it seem unlikely. It assumes that the Norse shield wall was a line, whereas Heimskringla says that it was enclosed. It assumes that the English then chose not to ride around the open ends of the Norse shield wall to attack the Norse commanders. It assumes that the Norse put up more of a fight on the flat than they did on the rising ground.

If, no matter how unlikely, the Norse army fell back enough to allow the English across the Derwent, Hardrada’s best option would have been to make a stand on a Roman road. There is uncertainty about the route of these roads (see below), but at least one must have passed through Stamford Bridge. Not only would the road guarantee better ground than the enemy, but it would probably have had ditches on both sides that provided readymade fortification. If the Norse army made a disciplined defence on a Roman road, the English stood virtually no chance of breaking through before the fleet guard arrived. We cannot envisage any scenario in which Hardrada would choose to defend Battle Flat, or its adjacent higher ground, if he could instead make a stand on a Roman road, especially considering the Norse army would have had to leave or cross the road to get to Battle Flat or the higher ground.

The English had hundreds of horsemen, the Norse only had a handful. If the armies faced each other in parallel lines, as tradition dictates and as depicted on Figure 3 and Figure 4, the English horsemen would have ridden around the open ends of the Norse line to attack the weakly defended Norse commanders.

Moreover, as we mention at the start, both the proposed battlefields at Stamford Bridge are in the wrong direction. Hardrada must have moored south of Stamford Bridge, for reasons we will return to momentarily. Therefore, the fleet guard reinforcements must have been coming from the south. The terrain was easier to defend in that direction too. For all these reasons, if the Norse army retreated from the riverbank at Stamford Bridge, they would have headed south towards their ships, rather than east towards Battle Flat.

Another nail in the coffin is that archaeologists now think that the medieval bridge at Stamford Bridge was midway between modern Stamford Bridge and Low Catton (shown in white on Figure 5), 1.5km south of its traditional position. We will run through the evidence for this shortly. From this revised bridge location, it was twice as far to Battle Flat and an extra 22 degrees in the wrong direction. If it is unlikely that the Norse army would head east to Battle Flat from the traditional bridge location, it is exponentially less likely that they would head ENE over double the claggy terrain from the revised location. And, if they did, they would have kept going to Pasture Hill (PH) rather than veer north and then stop shy of the higher ground.

The real clincher for us is that High Catton (HC), the best defensive location east of the river, was only 1km from the revised bridge location (2km from the traditional location). It is at the northern end of the Escrick glacial moraine ridge, only 35m high but just 50m wide along the defendable ridge crest. It lies between two boggy streams. The ridge sides would have been steep, slippery and backing into the streams. The ridge crest was almost level but so narrow that Hardrada could have reinforced the contact zones 20 men deep. RU3303 - listed in the NAA report here - confirms that there was a Roman road from RR810 to High Catton. We cannot envisage a scenario in which the Norse army would choose to defend Battle Flat or anywhere else at Stamford Bridge, if they could more easily have defended High Catton.

Burne acknowledges that Battle Flat is an unpromising choice of battlefield. He reckons the Norse did not care. His thinks that their primary concern was not to be disadvantaged by the battle terrain. He says they: “considered themselves pre-eminent, and their one thought was to form this shield-wall on a fair battleground ”. We have never heard this conjecture before or elsewhere. In our limited knowledge of Viking combat, they always looked to defend the most advantageous terrain. Indeed, they took a pride in choosing the most advantageous battlefields. It was mentioned in Harald Hardrada's obituary as one of his greatest assets. We are convinced that he would have looked for the most advantageous battlefield terrain here too.

Finally, we think that Huntingdon's initial encounter west of the Derwent is unlikely to have been on the riverbank. It would be an idiotic place to defend at the best of times, with nowhere to go apart from gloop, river and a bottleneck bridge. Worse, David Harrison reckons there was a bridge or ford over the Derwent at Howsham, less than five miles upstream. If the Norse army was initially protecting the west riverbank, as shown in McLynn's diagram, the English would have delayed their attack until horsemen had crossed the river and blocked the eastern side of the bridge. In this scenario, the Norse army would be trapped on a boggy riverbank in a hopeless position. If Huntingdon is right that the initial encounter was west of the Derwent - and the Norse accounts agree - it was some considerable distance from the river. If it was some distance from the river, it is unlikely to have been a minor skirmish.

DeVries suggests that the topography might have changed significantly since medieval times. There is no evidence of it. After all, Battle Flat would not have got its name if it were not flat. Perhaps there has been some levelling for the modern housing developments, but it cannot have been much because a Roman road ran right across the middle of it (Figure 8). And, anyway, they did not have the manpower, tools or incentive to move huge quantities of earth before the 19th century. It has not changed significantly since the advent of maps. DeVries tries to make a case that Battle Flat was defensible, but he does say that its main selling point as the battlefield is its name.

The most plausible reasons for a battlefield east of Stamford Bridge are that the English corralled the Norse army onto bad ground or the Norsemen fled to boggy ground in order to escape English horsemen and this was where the English infantry caught up with them. Neither scenario fits the descriptions in any of the contemporary battle accounts.

It would come as no surprise if some battle related archaeology did turn up at Battle Flat eventually. The Norse probably fled in all directions after the battle. Some of them may well have been caught and killed at Battle Flat. Perhaps that is how it got its name. Otherwise, in our opinion, Battle Flat is a red herring. It is not mentioned on Saxton, Blaeu or any other pre-19th century maps. Its name might have been invented by a Georgian estate agent.

The real location of the 11th century bridge at Stamford Bridge

All the major historical studies assume that the bridge at Stamford Bridge was at its traditional location, to the northwest of the modern town, roughly 100m upstream of the current bridge (where McLynn shows it to be on Figure 3). But these studies predate several important archaeological surveys.

H.G. Ramm reports that York Sub-Aqua Club performed an underwater survey of the Derwent in 1964, looking for evidence of bridges that predate the current one. They found nothing. SYO had no more success when they excavated the riverbanks in 2000. Indeed, there is no evidence that there was ever a bridge at Stamford Bridge before the 12th century. During the same excavation SYO did, however, find plentiful evidence of a Roman settlement, Roman road crossing and bridge some 1500m downstream, near Low Catton.

Figure 6: Relabelled SYO diagram showing Roman road excavations - Copyright © SYO

Lawton reports that SYO found five sections of Roman road on both sides of the river (D to H on Figure 6). He proposed that the settlement was the minor Roman town of Derventio. His proposal is now orthodox, accepted by Historic England (listing 1416328) among many. Roman roads were built mainly for military purposes. From the direction of the road east of the river, it seems likely to be Margary’s RR810, between the Roman garrisons of York and Bridlington. We will return to SYO's northern road.

If Derventio spanned the river, it must have had a ford or bridge. There are a couple of logistical reasons why Derventio is a more likely Derwent bridging point for RR810 than the traditional bridge location northwest of Stamford Bridge: 1) Roman engineers were famous for their straight roads and Derventio is in line between York and Garrowby, the only gap through the Yorkshire Wolds to Bridlington; and 2) A crossing point at Derventio cuts the distance from York to the Roman garrison of Petuaria (Brough) by 3km compared to the traditional bridge location.

Figure 7 - Royal Commission aerial photo showing RR810 Roman road heading east  and RR81a heading northeast

Lawton’s Roman road at Derventio was corroborated by the rediscovery of a 1970s B&W aerial photograph (Figure 7, facing east), which shows the road and buildings as cropmarks 200m east of the Derwent. This road points directly at Garrowby Hill, which makes it likely to have been a section of RR810.

Figure 8: NAA pipeline excavations near Stamford Bridge, showing Roman roads as orange lines

Lawton’s findings were further corroborated and expanded in 2005 by a major excavation for a proposed new water pipeline undertaken by the Northern Archaeological Association. Details are recorded here and summarised in Figure 8, Roman roads are shown as orange lines. Stamford Bridge town is the built-up area. The Derwent is shown in pale green. RR810 is shown crossing the Derwent at Roman Derventio near the south of the diagram (25). RR2e from Brough is the dotted orange line coming from the bottom right of the diagram. The road shown intersecting with RR810 at Stamford Bridge Junior Football ground (10) is from the Escrick moraine ridge at High Catton. RR81a to Malton is shown branching off northeast through Stamford Bridge town.

The Roman Roads Research Association (RRRA), who coordinate references on Britain’s Roman roads, have recently reviewed all the existing evidence and conducted some new studies of their own. They agree that RR810 crossed the Derwent at former Derventio and stretched away east to Garrowby and Bridlington.

The evidence for Derventio bridge looks convincing to us. We will assume henceforth that this is the bridge being referred to in the battle accounts. It is 1.5km south of where had been previously assumed, and south of modern Stamford Bridge rather than to its north. This then is the bridge over which the Norse army headed to York and this is the bridge over which they retreated. Lawton pointed out 20 years ago that the revised bridge location would have a profound impact on the battle narrative.

Stamford Bridge has no listing in Domesday. The land on which modern Stamford Bridge developed was part of the major Saxon manor of ‘Cattune’ (from which Low and High Catton get their names). ‘Place names of the East Riding and York’ describes a place named Flawith in the parish of Stamford Bridge. A H Smith, its editor, interprets this to mean ‘flag ford’. There is a bed of flaggy Keuper sandstone to the north of modern Stamford Bridge. As Ramm says, it would have been ideal for the solid bed of a ford. Ramm suggests that Flawith and Stamford were cognates, both referring to this ford.

Yet, the ASC, whose 1066 entry would have been recorded before Domesday was compiled, refers to a place named ‘steinford brygge’. This must refer to Derventio bridge. So, the bridge was at steinford, a cognate of ‘stamford’, both meaning ‘stoney ford’. We believe that there were two fords, one flaggy to the north of modern Stamford Bridge, the other stoney to the southwest of modern Stamford Bridge, where the Saxons built a bridge at Roman Derventio. Both fords were in the Saxon manor of Cattune which encompassed modern Low Catton, High Catton and Stamford Bridge. The main Saxon era settlement was at Low Catton on the same ground as Roman Derventio. Modern Stamford Bridge was probably just a few houses beside Flagwith ford in Saxon times. It was still not a settlement when Domesday was compiled in the 1080s. We suspect that the main settlement moved north to modern Stamford Bridge when the wooden bridge collapsed, probably sometime in the 12th century.

There is obvious room for confusion. The contemporary account references to ‘Stamford bridge’ and cognates, referred to the ‘bridge at stam ford’ rather than to a settlement named ‘Stamford Bridge’. Henceforth in this paper, ‘Stamford Bridge’ means the eastern side of Derventio bridge, unless we prefix it ‘modern’ to mean ‘modern Stamford Bridge’.

One other minor point is that Derventio was previously thought to be modern Malton – on Figure 10, for example - so the Roman town at Malton is now known as Delgovicia or Derventio Brigantium.

Battlefield location according to Norse saga accounts

Norse tradition about the Battle of Stamford Bridge is based on accounts in the Norse 'King's sagas'. There are 23 of them in all, four of which contain original information about Hardrada's invasion of Northumbria. These four are known as Heimskringla, Fagrskinna, Morkinskinna and Orkneyinga. We will focus on Heimskringla which has the most original detail as well as a synthesis of nearly all the original details in the others. It was written by Snorri Sturluson around 1230. The whole text can be reviewed here, with the invasion account starting at page 83.

We are more positive about the Norse sagas than proper historians. As with Wace’s much maligned account of the Battle of Hastings, we think that one just has to be selective about which bits to trust, and for much the same reasons. Norse sagas were created to glorify Norse kings and Norse culture. Of course, they exaggerate and contain a lot of poetic license. Doubtless, they big-up Hardrada's successes and downplay his failures. But they had no incentive to invent incidental detail or to move the location of major events. We are inclined to believe details that do not glorify Norse heroes or Norse culture.

As far as we know, historians have always tried to fit saga events into a narrative based on the English accounts, rejecting those saga events that are inconvenient or that do not fit. We think that a more coherent narrative can be built by fitting English account events into Heimskringla, in which case nothing needs to be rejected.

This is Laing's translation of what Heimskringla says about the initial encounter:

Now as they came near the castle a great army seemed coming against them, and they saw a cloud of dust as from horses' feet, and under it shining shields and bright armour. The king halted his people, and called to him Earl Toste, and asked him what army this could be. The Earl replied that he thought it most likely to be a hostile army, but possibly it might be some of his relations who were seeking for mercy and friendship, in order to obtain certain peace and safety from the king. Then the king said, "We must all halt, to discover what kind of a force this is." They did so; and the nearer this force came the greater it appeared, and their shining arms were to the sight like glancing ice.

Heimskringla had previously explained that Hardrada was due to attend a meeting in York castle "early on Monday morning, where he would name new officers to rule over the town, give out laws and bestow fiefs". This Monday was the day of battle. Thus, Heimskringla is saying that the Norse army was approaching York early in the morning. They must have been approaching from the east because they eventually retreat east across the Derwent. In case it is not obvious, if they had been approaching York from any other direction, they would have had to retreat an implausible 12km or more in order to retreat across the Derwent.

The Norse army must have been at least two miles west of Stamford Bridge when they first spot the enemy. That is where the Fulford glacial moraine ridge - shown in red and yellow on Figure 9 – crosses between Stamford Bridge (S) and York (Y). It is not hard to work out. If the Norse army was anywhere east of the ridge, they could not see the English army on the other side. Hardrada asks Tostig about the nature of the approaching army. The only place he would have had a good enough view to know it was an approaching army would have been on the moraine ridge.

Historians have another explanation. They think Hardrada was at modern Stamford Bridge or High Catton when he sees the English army crest the moraine ridge, which would corroborate the traditional interpretation of the English accounts. But this, and any other position east of the ridge, would contradict the rest of Heimskringla's engagement account.

So, if Hardrada was anywhere east of the Derwent, he would see the English army cresting the ridge one rank at a time, whereas Heimskringla has him seeing the entire approaching army. It would have been 15 minutes or more until he realised it was an army cresting the ridge, whereas Heimskringla suggests that he immediately knows it is an army. He would not know that his army was outnumbered until more than half the English army had crested the ridge, by which time the English vanguard would be on top of him. Then there is the dust cloud. If Hardrada was east of the ridge and the wind was from the north, west or not at all, he would see a dust cloud before seeing any men. If the wind was from any other direction, he would first see the English vanguard with no dust.

Figure 9: Battle theatre topography heatmap

The Norse army must have been on the moraine ridge to fit Heimskringla’s engagement description. If so, the English army must have been some miles distant because Tostig cannot work out whether the approaching army is friend or foe. Hardrada orders the Norse to hold their position: “We must all halt to find what kind of a force this is". When the gap has closed, Heimskringla reports a discussion between Hardrada and Tostig:

Then said King Harald, "Let us now fall upon some good sensible counsel; for it is not to be concealed that this is an hostile army and the king himself without doubt is here". Then said the earl, "The first counsel is to turn about as fast as we can to our ships to get our men and our weapons, and then we will make a defence according to our ability; or otherwise let our ships defend us, for there these horsemen have no power over us". Then King Harald said, "I have another counsel. Put three of our best horses under three of our briskest lads and let them ride with all speed to tell our people to come quickly to our relief. The Englishmen shall have a hard fray of it before we give ourselves up for lost." The earl said the king must order in this, as in all things, as he thought best; adding, at the same time, it was by no means his wish to fly. Then King Harald ordered his banner Land-ravager to be set up; and Frirek was the name of him who bore the banner.

We interpret this to mean that nothing happens between Hardrada’s discussion with Tostig and planting the banner around which the shield wall is formed. If so and we are right that Hardrada was on the moraine ridge, then the battlefield was on the moraine ridge too, at least two miles west of modern Stamford Bridge. Calculating where on the ridge depends on where it intersected with the Roman road between Derventio and York. We will return to this below. 

Huntingdon does not contradict Heimskringla's account. He says that "the battle was desperately fought" and then that the invaders "retreated in good order" and were "driven across the river". This could easily be saying that the battle was desperately fought on the moraine ridge and that the invaders retreated across the Derventio bridge.

Common sense supports Heimskringla's account too. The moraine ridge is the only militarily advantageous place between York and Stamford Bridge. If Hardrada could choose any location west of the river to defend, it would be a raised area on the Fulford moraine ridge.

But if Heimskringla and Huntingdon are right, then ASC-D, Chronicon ex Chronicis and part of Historia Anglorum are either wrong or have been misinterpreted. We will come back to ASC-D. Huntingdon says that Harold: "came up with the invaders at Steinfordesbrige", John of Worcester that Harold: "encountered the Norwegians at a place called Stanford-brycge". Both are inaccurate if the main battle was on the Fulford moraine ridge. But, as we say in our blogs about the Battle of Hastings, place names are more woolly in sparsely populated places. 11th century Vale of York only had a hundred or so families outside York. We guess that they are trying to say that 'Stanford bridge' was the closest place to the battlefield that readers would recognise. This is exactly the case for the Battle of Worcester, which took place 2km south of Worcester, and the Battle of Naseby, which took place 2km north of Naseby, and many others.

ASC-D, at the very least, is ambiguous. Historians focus on its statement that the encounter was “beyond the bridge”. Earlier it had said that the Norse army was at steinford brygge on the day of battle. The implication then is that the main battlefield was beyond ‘steinford brygge’ from King Harold’s perspective. Harold was in York, which would mean the battlefield was east of the Derwent. Yet the passage about the Norwegian giant contradicts it.

ASC-D explains that Hardrada and Tostig were killed, and that the surviving invaders fled. Then it says: “But there was one of the Norwegians who withstood the English folk, so that they could not pass over the bridge, nor complete the victory." In other words, the Norse army had capitulated before crossing the bridge, which means that the main battle was west of the Derwent.

Something must be wrong with one or other of the ASC-D statements. The section about the Norse giant holding the bridge has been inserted into the narrative at a later time in another hand, but it is unambiguous. We guess that its “beyond the bridge" meant something other than what it would mean today.

ASC-D previously says that Hardrada and Tostig were “beyond York”. This cannot mean 'beyond' in the modern sense relative to King Harold, because he was in York. The ASC has a specific use of the term 'beyond the sea' to mean in Continental Europe. All the other occurrences of 'beyond' in the Anglo-Saxon Chronicle, in our opinion, mean 'be yonder'; i.e. 'some distance from'. So when it says Hardrada and Tostig were 'beyond York', we think it means that they were 'some distance from York'. 'Beyond the bridge' might therefore mean 'some distance from the bridge', without an implication of whether they were east or west of it.

Alternatively, perhaps ASC-D was not referring to the Derwent bridge. The English army would have had to cross a bridge over the Foss outside York (marked with a red X and Y on Figure 10). Under normal circumstances, this would have been a busier and better known bridge in the 11th century. So perhaps ASC-D is saying that the battlefield was somewhere east of the Foss, which is not terribly helpful, but at least it does not contradict the sagas or Huntingdon.

A third possibility is that Harold was not the relative origin. Modern ‘beyond’ is always relative to something. Thus, Edinburgh is 'beyond Hadrian’s Wall’ relative to London but not relative to Inverness or Glasgow. British historians assume the ASC is written from the King’s perspective. But perhaps it was relative to somewhere or something else, Worcester, maybe, since that is where the D recension was written. If that relative origin was west or south of Stamford Bridge, ASC-D could be saying that the main battlefield was west of the Derwent.

In summary, Heimskringla suggests that the battle happened on the Fulford moraine ridge. We think that there are simple interpretations of Chronicon ex Chronicis and Historia Anglorum that agree, and a plausible interpretation of ASC-D that concurs.

Roman roads between York and the Derwent

We think the battlefield was adjacent to the intersection of the Fulford moraine ridge and a Roman road between York and the Derwent. It is not as specific as it sounds. There were at least two Roman roads heading east out of York, perhaps three. It is not obvious which the armies were on and their routes are uncertain anyway.

There were three important Roman era destinations east of York: Malton, Bridlington and Brough. Malton is on the Derwent, 17 miles northeast of York. The other two are on the far side of the Derwent from York. Bridlington is on the coast, 37 miles ENE from York. Brough is down near the mouth of the Humber, 23 miles south of Stamford Bridge. Contemporary Roman records list paved roads to all of them. Their Margary numbers are RR800, RR810 and RR2e respectively.

Figure 10: RCHME diagram of York Roman roads

RCHME have the traditional picture of the roads and destinations around of York (Figure 10). The dashed lines indicate possible routes, which shows it is mostly guesswork. The only solid line is a road leaving York’s northeast gate then crossing the Foss. There is a suggested road heading east to Brough on the Fulford moraine ridge. The western end of the road was confirmed by later excavations. There were three destinations and only two roads heading east. One of them must have branched. RCHME show the tentative split on the north road, with one branch going to modern Stamford Bridge, the other to Malton.

As far as we know, the RCHME diagram is just an educated guess. It is partially refuted by Lawton’s revised location for the bridge, by excavations by SYO, NAA and RRRA, and by recent cropmark discoveries (shown in Figures 6, 7 and 8 and collated in Figure 13).

Figure 11: Roman road at Stamford Bridge shown on Ordnance Survey map

The south branch of RCHME’s northern road is labelled “From STAMFORD BRIDGE”. By tradition, it goes from York’s northeast gate to Gate Hemsley (pink dotted line on Figure 13), then continues along the modern A166 to cross the Derwent 100m upstream of the current bridge at modern Stamford Bridge. Ordnance Survey maps mark this section as ‘ROMAN ROAD’ and show it continuing east of the Derwent to Garrowby and then Bridlington (Figure 11). This then is the traditional route of Margary’s RR810 between York and Bridlington.

As far as we know, the traditional route of RR810 was based on three anchors. One was the road known to have left York’s north-eastern gate heading east. The second was the traditional location of the bridge at Stamford Bridge. The third was a place named Street Farm some 2km ENE of Stamford Bridge. ‘Street’ traditionally refers to medieval roads built on the route of Roman roads. In the 1950s, the route seemed to have been confirmed by two new sections of Roman road with roughly the right orientations. One was at Apple Tree Farm 3km ENE of York’s north-eastern gate. The other was at the McKechnie factory, on the Derwent’s west bank, near the traditional location of the bridge at Stamford Bridge.

Lawton’s discovery that the bridge was south of modern Stamford Bridge rather than in its traditional location does not refute the traditional route of RR810. As we say above, there might have been a ford – known as Flawith we think – at the traditional location of the bridge.

Figure 12 : RRRA diagram showing Roman roads east of York - strong red dotted lines show probable routes, light red dotted lines show possible route, uncoloured dotted lines show previously proposed routes that have been rejected - Copyright © RRRA

The Roman Roads Research Association (RRRA) disagree with the orthodox route of RR810. They have reviewed the likely Roman road routes east of York in the light of the revised bridge location and post-Margary excavations and recent cropmark discoveries, especially those either side of the A64. Their proposed map of Roman roads in the Yorkshire Ridings is shown in Figure 12.

RRRA’s map shows RR810 as straight between Apple Tree Farm and Derventio. Roman roads are famously straight. They note that the A64 cropmarks have exactly the right location and orientation to be a section of this road. The sections of road exposed at either end have roughly the right orientation. RRRA show the southern Roman road running along the crest of the Fulford moraine ridge to Holtby where it joined RR810. It might seem odd that there were two roads from the same source to a minor destination, but RRRA have an explanation.

RRRA discovered that the southern road is half the width and several centuries later than their RR810. York to Bridlington was an important Roman military route used before the southern road was made. RRRA suggest that the southern road was probably made as an alternative route for when RR810 flooded. This makes some sense. The water meadows and fen to the east of the Foss are prone to flooding now. They probably flooded every year in medieval times. The southern road was a flood-proof 25m above the flood plain.

There are two problems with RRRA’s map: 1) It ignores the sections of Roman road that have been found between Dunnington and Derventio; and 2) It ignores the section of road found at the McKechnie factory.

In the case of the former, we can only imagine that it is an oversight. Figure 13 shows all the sections of road that have been found in the vicinity as red dots. The four sections of road found between Dunnington (D) and Derventio (between C and S) make it virtually certain that they were joined by a road. This does not mean that the southern road from York did not continue along the ridge crest to Holtby, as RRRA suggest. It might well have done. It just means that if it did, it also branched off the ridge at Dunnington on its way to Derventio.

In the case of the latter, SYO excavated a 1km section between Apple Tree Farm and Gate Hemsley (mid-way along the magenta dots A to G on Figure 13 and shown on Figure 6), on what was believed to be the route of the road to the McKechnie factory, without finding any evidence of a road. Maybe this persuaded RRRA that the section of road found at the McKechnie factory was part of some other road network, perhaps related to the fort they show (as a blue square) on Figure 12. But it might be instead that the road from Apple Tree Farm to the McKechnie factory was not straight.

Figure 13: Likely and possible routes of Roman roads between York and Stamford Bridge

The route of the RR810 is important because it is the most likely road that the English army would have used when they left York. If the southern road out of York went to Deventio, it is plausible that RR810, as tradition suggests, did indeed go to the McKechnie factory before crossing the Derwent on its way to Bridlington. It seems unlikely that it took the traditional route straight to Gate Hemsley, because that route has been excavated with no finds. Also, the section that has been excavated at Apple Tree Farm points to Holtby and Derventio. But perhaps this is just a coincidence. RR810 might have gone straight to Holtby, then turned north on the moraine ridge to Gate Hemsley before following the A166 to the McKechnie factory. Or it might have doglegged off a section of road heading east from Apple Tree Farm, avoiding the A166.

In the previous edition of this book, based on some cropmarks we found near Holtby (Figure 14 and Figure 15), we speculated that the latter was most likely. HER (1174922), based on a desk study, lists these cropmarks as probably post-medieval enclosures. Or, they might be from a Roman road. If so, the cropmarks they extrapolate east to the McKechnie factory and west to RRRA’s route for the RR810 (at 53.973716, 1.009466, 450m southeast of the Red House Cattery). We thought it possible then that there was a dogleg or branch on RRRA’s route for the RR810 that went to the McKechnie factory, but a York Archaeological Trust excavation seems to disprove it.

Figure 14: Google Earth Roman road cropmark near Holtby
Figure 15: Ordnance Survey Roman road cropmark near Holtby

YAT excavated around Newsham House on the northern outskirts of Holtby in 2018. Trench 3 (trenches shown in light blue on Figure 16) should have been right over a straight Roman road from the elbow of Holtby Lane to the McKechnie factory, but it showed no archaeology at all. Perhaps the road terminated at the nearby Holtby timber circle.

Figure 16: York Archaeological Trust excavation at Newsham House, Holtby

That brings us back to the section of Roman road between Hemsley House and Gate Hemsley that is labelled on Ordnance Survey maps. It is based on county boundaries rather than archaeology, but they seldom get these things wrong. At the eastern end it extrapolates to the McKechnie factory. At the western end it extrapolates to the lowest point of the Fulford moraine ridge. If it was a Roman road, it looks like it passed through the gap roughly midway between Wandales and Hungry Hill. The drainage channels run roughly southwest to northeast, parallel to the moraine ridge. If we were selecting a route to York, we would run it parallel to and between Carr Goit and Rudcarr Lane.

Time to rewind. We have no reason to doubt that RR810 went straight from Apple Tree Farm (A) to Derventio, as RRRA suggest. It is likely that a southern Roman road followed the Fulford moraine crest to Dunnington (Du on Figure 17), where it doglegged or branched northeast to Derventio (D). A third road passed through the McKechnie factory. Assuming Ordnance Survey have got it right, that road also passed through Gate Hemsley. There are two possibilities: the northern road branched off RR810 from somewhere near Holtby Grange, or it branched off the southern road at Dunnington. We previously thought the first more likely, but an October 2020 walk around Rudcarr Lane persuaded us that it would probably have flooded in most years. We now think it more likely that the northern branch connected to the Fulford moraine ridgeway at Gate Hemsley. 

Figure 17: Roman roads east of York: From Apple Tree Farm (A) to Derventio (D), red; From Apple Tree Farm To McKechnie factory, magenta; From Dunnington (Du) to Derventio, black

If this is right, three Roman roads headed east from York across the Derwent. It sounds unlikely but we think there is a credible explanation. RRRA’s route for RR810 between Apple Tree Farm and Derventio is wide and straight, typical of important Roman roads. They reckon it was the earliest of the roads east of York and was therefore probably built for military use. However, it had an 8% slope as it crossed Fulford moraine ridge which would have been marginal for horse drawn carts carrying heavy freight. It also had to cross the Foss and two miles of boggy fen. It may well have been flood prone. The southern road along the Fulford moraine crest had neither of these issues. It is at least 5m above the flood plain and had no slopes greater than 2.5%. We think it was built primarily for freight between York and Brough, although it may well have been used by non-freight traffic when RR810 flooded. Historic England list the likely presence of two Roman marching camps just north of the McKechnie factory. It seems likely that the northern branch serviced these camps.

Our main interest here is in Hardrada’s location on the moraine ridge crest when he spotted the English army. The English army must have been on a paved road because they were making clouds of dust. They are likely to have been on a paved road anyway because men and horses would have risked injury and exhaustion crossing the gloopy fens between York and the Derwent. If we are right about the road routes, there are three possibilities, depending on which road the Norse army was on: northwest of Dunnington, or on the southern outskirts of Holtby (H) or midway between Holtby and Warthill (W).

We think the Norse army was at Holtby because there are issues with Dunnington and the Warthill gap. Heimskringla makes it sound like Hardrada and Tostig were looking down on the English army coming towards them, able to see its size, strength and composition while still too far distant to know whether it was friend or foe. It is unlikely that the Norse army was on the northern branch because they were heading to York from Derventio bridge. Even if the armies were heading towards each other on the northern branch, Hardrada would not have been able to see the English army in any detail because Warthill gap is only 5m above the road from York. If, on the other hand, the armies were heading towards each other on the southern road, the Norse would not be looking down on the English because the ridge crest is fairly level.

Even if Heimskringla invented the conversation between Hardrada and Tostig about their view of the English army, it seems likely that both armies were on the RR810 between Apple Tree Farm to Derventio anyway. The Norse army was heading to York from Derventio bridge. The English army was heading to Derventio bridge from York. They were not expecting to encounter each other on the road. Both, then, were likely to have chosen to use RR810 because it was the widest, shortest and best surfaced route to their destination. 

Riccall & the Norse fleet

A coherent battle narrative requires an assumption about where the Norse fleet was moored on the eve of battle. That location defined where the Norse army was likely to have been when they spot the enemy, how long it would take to summon the fleet guard and how long it would take to flee to their ships.

These latter points would have been uppermost in Hardrada's mind when he rejected Tostig's recommendation to flee for the ships. They were crucial to the outcome of the battle, because Heimskringla says that a third of the invading army was left at their ships on the day of battle. Hardrada would have known that he would probably be victorious if the fleet guard arrived soon enough and still vigorous, or would otherwise probably lose.

John of Worcester says that the Norse fleet was moored at Riccall before the Battle of Fulford. This would make sense. It was an island in those days, the last riverside settlement before York that was not visible from York, it being hidden by the Escrick moraine ridge. It had a name, which means it was farmed. It probably had livestock, grain stores and a well. It sounds like an adequate overnight camp and a reasonable launchpad for an attack on York. Riccall was an island, so the Norse fleet must have rowed upstream for the Battle of Fulford. By tradition, they returned to Riccall after their victory and stayed there until the Battle of Stamford Bridge. This last part makes no sense.

The map below is a repeat of Figure 2, showing the Vale of York geography in 1066. Riccall is labelled R. No source says that the Norse fleet was at Riccall on the day of the Battle of Stamford Bridge. On the other hand, nor do they say that the fleet moved from Riccall before the battle, so most historians assume that they were still there. If so, the Norse army spent the night before the battle at Riccall, the main army left from Riccall, the fleet guard came to the rescue from Riccall, and battle survivors fled towards Riccall.

Figure 2: Vale of York geography in 1066. Y= York, S=Stamford Bridge, R=Riccall, T=Tadcaster, H=Holtby, K=Kexby, U=Bubwith, M=Barmby-in-the-Marsh, B=Brough, W=Wheldrake, SD=Sutton upon Derwent, EC=East Cottingwith, E=Escrick, Marsh=Humberhead Levels, black lines=Indication of major Roman roads

There are a number of reasons why it is unlikely that the Norse fleet was at Riccall on the morning of the day of battle:

  1. ASC-D says that Harold collects his fleet at Tadcaster. In order for the fleet to get to Tadcaster, they would have to pass Riccall. The Norse army is hardly likely to have let them pass unhindered and, if they did, they would be forewarned that Harold was on his way.
  2. Heimskringla says that Marshall Styrkar escapes the battle on a horse to notify Hardrada’s son Olaf that the battle is lost. Olaf was guarding the ships. Styrkar meets a waggoner on the way. He: “…swung his sword, and struck him on the neck, so that his head came off. He then took the skin-coat, sprang on his horse, and rode down to the strand”. He must have been riding along a ridgeway when he met the waggoner, because he later rode down. He rode down to a coastal or estuarine strand where the ships were moored. It could not have been Riccall because it was an island in a tidal saltmarsh. It was not overlooked by a ridgeway to ride down from, and it would not have had a strand anyway.
  3. Styrkar was cold because of the wind not because he was wet, so it sounds as if he did not cross any significant waterways on route. Heimskringla says that Hardrada sent his three swiftest riders on his three swiftest horses to summon the fleet guard. He would have wanted the lightest riders and lightest horses if they went by fen or bog, so they must have ridden on roads and ridgeways. But the only roads and/or ridgeways from Stamford Bridge to Riccall had to cross the deep and boggy Derwent at Wheldrake.
  4. Even if there was a ferry over the Derwent, Styrkar and the messengers could not have ridden to Riccall because it was an island. They would have needed to hail the fleet guard from the nearest land, using semaphore or trumpet perhaps. The nearest land was Escrick, 5km away. It would not have been impossible to hail over 5km, but it is difficult to believe that Hardrada would gamble the fate of his campaign on it.
  5. Heimskringla says that: “In every division he allowed two men to land, and one to remain behind”. But the land party could not have gone anywhere if they disembarked onto Riccall because it was an island. For this same reason, the Norse fleet could not have been moored at Barmby (M), East Cottingwith (EC) or any other island.
  6. The Norse ships could have taken the landing party to somewhere with a land link to Stamford Bridge on the morning of battle – even though Heimskringla clearly implies that they did not – but they could not get to battlefield in time to comply with Huntingdon’s statement that the battle started in the early morning. Day started at dawn, 05:26 on the 25th September. They ate breakfast. They would then need to row 20km downstream to Barmby on the Marsh (M), 21km upstream to Sutton upon Derwent (S), disembark, muster, and march 14km to Stamford Bridge. The tides would roughly cancel each other out. Roskilde Ship Museum told us that longships could row at 2-3 knots over sustained periods, so that is 10 hours of rowing and three hours of marching. Even without eating and mustering time, they could not get to Stamford Bridge before mid-afternoon, and we think the battle was two miles further. There might seem to be a shorter route up the Ouse and Stillingfleet Beck to Wheldrake, but the beck is only one metre wide, too narrow to row, and it would then have needed a 1km bog crossing and a Derwent crossing to get to Stamford Bridge. They could have rowed to Fulford then run to Stamford Bridge in time to arrive before midday, but then they would have seen Harold’s army approaching York from Tadcaster.
  7. Heimskringla says that Hardrada sent his three swiftest riders on his three swiftest horses to summon the fleet guard from Riccall. But, as we say above, it contradicts Heimskringla which says that Styrkar gets to the fleet without crossing a major river, and it would be risky because of the need to hail the fleet guard from several kilometres. It would also have been too slow. The shortest route from Stamford Bridge was 11km down the Escrick moraine ridge, across the Derwent to Wheldrake and then another 7km to Escrick. It would have taken at least 90 minutes, if a ferry was ready and waiting to cross the Derwent. Having received their summons, the fleet guard could not possibly have arrived in time to affect the battle. Working back from the Norwegian giant blocking the bridge until 15:00, the main battle ended no later than 14:00. Heimskringla says that the fleet guard arrived beforehand, nearly turning the course of the battle. They could not have arrived any later than 13:30. The messengers could not have been dispatched before the initial sighting, which could not have been any earlier than 08:00, so they could not have arrived near Riccall before 09:30, giving the fleet guard four hours to get to the battlefield. They could not have gone via Sutton upon Derwent, which would have taken over 13 hours, as we say above, but they had a quicker route. They could have rowed 17km upstream to Fulford, then run 10km east along the Fulford moraine ridge. Without allowing time any time to dress for battle or time to decide route to take or to muster, if they caught a perfect tide and rowed like the clappers and ran like the wind and avoided getting caught in the open by English horsemen, they might have got to a moraine ridge battlefield by 13:30. But, once again, it is difficult to believe that Hardrada would have gambled the fate of his campaign on it.
  8. The Fulford truce committed York’s townsfolk would provide the Norse army with food. Assuming they lost 1,000 men at Fulford, they would have needed perhaps 400 sheep/pigs cum 40 cattle a day to eat, plus at least 16,000 litres of fresh water. Riccall was an island farm surrounded by saltmarsh. They could not have kept thousands of men hydrated with a couple of buckets but nor could the townsfolk of York have ferried thousands of litres of fresh water, hundreds of sheep/ cattle to Riccall every day.
  9. The English army marched north on RR28c. It ran along a 40m ridge on the approach to Tadcaster the day before battle. Their dust cloud would have been visible to anyone with good eyesight at Riccall, just 9 miles away. It would have been even more visible the following morning as they marched from Tadcaster to York on a Roman road just 7 miles away. Yet it is clear from the sources and subsequent events that the Norse did not know that Harold was in the vicinity with a powerful army until it was too late.

For all these reasons, we are convinced that the Norse army moved away from Riccall on the day after the Battle of Fulford, and that their new camp was linked to Derventio bridge by Roman roads and/or ridgeways.

The Derwent was probably navigable up to Stamford Bridge at high tide. It might therefore seem feasible that the Norse fleet moored at or near Stamford Bridge on the eve of battle. Heimskringla suggests not. Indeed, it suggests that the Norse fleet moored at least an hour’s march south of Stamford Bridge, because if they were any nearer:

  1. The Norse army would have arrived at York in time to ambush the English army as they crossed the Ouse
  2. It would have been pointless for Hardrada to make a fuss about choosing the fastest horses and the fastest riders to carry the summons, if the fleet was only a ten minute ride away
  3. The fleet guard would have arrived at the battlefield before the main battle started
  4. The fleet guard would not have been so exhausted by their run to the battlefield that some of them died of fatigue
  5. Most of the battle survivors would have made it to their ships and escaped.

It would not have been possible for the Norse fleet to moor up near Derventio and Stamford Bridge anyway. Both sides of the river were too marshy. On the east bank heading south from Stamford Bridge to Sutton upon Derwent, the places are named Bell Ings, Throwmire, The Ings and Northland Ings, all of which are Norse words associated with marshland. There was no land link to Stamford Bridge from the west bank, but it was even more marshy with places named The Ings, Thackmere Ings, Thack Mire, North Ings, Haghill Leas Ings and Ing Marsh. There was nowhere with a land link to Stamford Bridge from south of Wheldrake either, because Humberhead Levels just had a few islands standing proud of the marsh.

By a process of elimination, the Norse fleet must have moored on the east bank between Sutton upon Derwent and Wheldrake (magenta line on Figure 18), where the Fulford moraine ridge abuts the river. It might have been Hobson's choice, but it could have been a lot worse. The route to Derventio bridge would have been solid dry ridgeway and Roman road all the way. And the riverbank would have had a shingle strand (because the ridge was made of glacial moraine), on which the invaders could build a solid dry camp.

Figure 18: Norse fleet mooring location near Wheldrake shown in magenta, with RRRA Roman roads

According to Huntingdon, the Norse fleet had roughly 350 ships. They would have needed roughly 3km of riverbank to moor. The Derwent flowed around the south of the moraine ridge in those days (it has rerouted since) and a tributary ran up east of the ridge. Each of the three sides would have had roughly 1km of mooring space, ideal for the Norse fleet. Indeed, it was the only 3km stretch of contiguous solid riverbank between Stamford Bridge and Brough.

Heimskringla says that the messenger carrying news of defeat "rode down to the strand". This implies that he was riding on elevated ground. The local whose coat he takes was riding in a horse and cart. This implies he was on a road or ridgeway. There can only be a strand where the river passes solid ground. In this vicinity, Heimskringla can only be describing a route along the Escrick moraine ridge and down to the strand on the east bank between Sutton upon Derwent and Wheldrake.

Hardrada would have wanted to camp at a farming settlement, ideally with access to a fresh water well or stream, livestock and grain stores. Sutton means 'southern farm', so it was a farm in Anglo-Saxon times. The only other farms with river frontage between Catton and Wheldrake were Kexby and Elvington, both of which were west bank islands with no land links.

Timing

It might be possible to substantiate, or refute, our theories about the battlefield and mooring place by checking the timings described in the contemporary accounts. There are only two.

Historia Anglorum, in the standard Forester translation, says: “the armies were engaged from daybreak until noonday”. At that point, the Norse army is forced to retreat over the river. This is where Huntingdon says that a Norwegian giant on Stamford Bridge: “stayed the advance of the whole English army till the ninth hour”. Times were calculated from sunrise in those days. Sunrise on the 25th September was just before 06:00. So the giant is overcome around 15:00. Something must be wrong.

The only way that the armies could have been engaged from daybreak is if the English army camped at York and the Norse army camped at Stamford Bridge. We explain above why we are sure that the Norse fleet was not moored anywhere within an hour's march of Stamford Bridge on battle eve. If the Norse army camped an hour or more away from Stamford Bridge, either Huntingdon is mistaken or there is something wrong with Forester's translation. We think the latter.

We reckon there is an ambiguous Latin statement in Historia Anglorum, namely “namque a summo mane usque ad meridiem”. Forester’s translation “from daybreak until noonday” is viable, but not definitive. We translate it to mean: “for most of the morning and afternoon”, so Huntingdon is saying: “the armies were engaged for most of the morning and afternoon".

It is good to replace Forester's translation of Huntingdon's statement about the length of the battle because it made no sense. Medieval melee battles seldom lasted more than an hour. This one probably had a couple of breaks, but it still seems implausible that the melee could have lasted a combined six hours to noon. He says that the Norwegian giant holds up the English army until 15:00. It seems even more implausible that one person, notwithstanding he was a big Viking on a narrow bridge, could have held up the English army for nearly three hours.

In order to verify the timings against our revised Huntingdon translation, an assumption is needed about where the armies were at dawn. We conclude above that the Norse army was on the Derwent east bank south of Sutton upon Derwent. ASC-D, the earliest and best informed of the English contemporary accounts, says that the English camped at Tadcaster. Unfortunately, it is contradicted by Heimskringla, which says that they camped inside the walls of York. Moreover, that account looks too detailed to have been invented:

That same evening, after sunset, King Harold Godwinson came from the south to the castle with a numerous army, and rode into the city with the good-will and consent of the people of the castle. All the gates and walls were beset so that the Northmen could receive no intelligence, and the army remained all night in the town.

Heimskringla cannot be saying what it implies. Secrecy was crucial to King Harold's plan. Tostig was related to some of the Northumbrian barons. He would have had spies in the city and in the nearby countryside. It is difficult to believe that they would not have seen the English army arrive. Heimskringla says that Harold stations guards to prevent spies getting out, but it would have been too late. They would have seen and heard the army in the field or would have escaped the city before the exits were blocked. Harold could not run the risk. If word got to Tostig, the invaders could no longer be caught by surprise, and they might lay siege to the city or set an ambush or return to their ships, none of which were in Harold's interest.

Harold probably did go to York on the eve of battle. He needed intelligence from the city elders, who knew about the Norse strength and location because they had been supplying the Norse army with food. They knew of Hardrada's plan to come to York on Monday morning. Harold needed to ensure that his army would be allowed through York's gates, because the only quick and safe route from west of the Ouse to east of the Foss in medieval times, was to pass through York. And Harold needed to press any surviving local militia into his service.

We guess that Heimskringla is trying to say that Harold arrived at York in the evening with a contingent of men to guard himself and the city exits, but not enough men to make Tostig's spies suspicious. Heimskringla describes the English army as 'immense' on the day of battle whereas it describes Harold's accompaniment to York as only 'numerous'. We guess he brought perhaps 100 men, all of whom stayed all night.

If the main English army did not camp at York, there is no reason to dispute ASC-D's statement that they camped at Tadcaster. Indeed, Tadcaster is exactly where they would be expected to camp, because it is the only place that the footbound army could muster with the Harold's elite Huscarls who would arrive by ship.

Most of the morning at that time of year means that the armies engaged before 09:00.  Assuming both armies left their camps soon after dawn, they would each have marched around 10 miles by 09:00. Ten miles from Tadcaster via York would take the English army to the Foss river crossing (X or Y on Figure 10). Ten miles from Wheldrake via High Catton would take the Norse army to the Fulford ridge. They would engage a little before 09:00, with the proviso that Huntingdon's "engaged" referred to when they first sighted each other rather than when the battle started. This does not seem unreasonable.

Our Huntingdon translation and proposed camp locations also seem to fit the river blockade by a Norwegian giant. He is overcome at around 15:00. There would have been three miles between the armies if they were where we describe the initial sighting. The English had to close the gap, then Harold made a peace offer, withdrew and issued his battle orders. It must have taken at least 90 minutes. If the armies sighted each other around 09:00, the battle started no earlier than 10:30. Then a melee, the death of Hardrada, more negotiation and another melee. It seems about right for a 150 minute gap to the arrival of the fleet guard, which was no earlier than 13:00. They nearly turn the battle but run out of puff. It probably took 30 minutes. The Norse army therefore did not capitulate before 13:30. They shuffled millipede-like two miles to the river, which cannot have taken much less than an hour. This means the giant only held out for a plausible 30 minutes, or less.

Revised Battle Narrative

Soon after dawn on Monday 25th September, the English army leaves Tadcaster heading for Stamford Bridge via York. At the same time, two-thirds of the Norse army leave their ships heading for York via Stamford Bridge. The Norse fleet was moored on the east bank of the Derwent between Wheldrake and Sutton upon Derwent. They march up the Fulford moraine ridge and cross the Derwent at Derventio.

By 09:00, both armies have marched around 10 miles. They are heading towards each other on the same Roman road between York and the Derwent river, most probably the straight road between York and Derventio. The armies spot each other as Hardrada crests the Fulford moraine ridge. Hardrada is unsure whether the other army is hostile. He orders his men to halt. The English army closes the gap. Realising that it is the English army led by Harold, Hardrada sends his fastest messengers to summon the fleet guard from his ships.

Figure 19: Army movements leading up to the initial encounter on the day of battle with the English in cyan and the Norse in magenta

Hardrada arranges his men in an enclosed shield wall (described in the Revised Battlefield section below) on a nearby raised section of the ridge. King Harold approaches in person to negotiate a truce. Hardrada and Tostig reject his offer. Heimskringla picks up the story.

Now the battle began. The Englishmen made a hot assault upon the Northmen, who sustained it bravely. It was no easy matter for the English to ride against the Northmen on account of their spears; therefore they rode in a circle around them. And the fight at first was but loose and light, as long as the Northmen kept their order of battle; for although the English rode hard against the Northmen, they gave way again immediately, as they could do nothing against them. Now when the Northmen thought they perceived that the enemy were making but weak assaults, they set after them, and would drive them into flight; but when they had broken their shield-rampart the Englishmen rode up from all sides, and threw arrows and spears on them. Now when King Harald Sigurdson saw this, he went into the fray where the greatest crash of weapons was, and there was a sharp conflict, in which many people fell on both sides. King Harald then was in a rage, and ran out in front of the array, and hewed down with both hands; so that neither helmet nor armour could withstand him, and all who were nearest gave way before him. It was then very near with the English that they had taken to flight. King Harald Sigurdson was hit by an arrow in the windpipe, and that was his death-wound. He fell, and all who had advanced with him, except those who retired with the banner.

In summary, after a long period of posturing with no combat, some Norsemen chase out of the shield wall to attack the English. The English surround them. There is fierce hand to hand combat. Hardrada joins the fray on foot. He gets too close to the English archers who shoot him fatally through the throat. Heimskringla continues:

There was afterwards the warmest conflict, and Earl Toste had taken charge of the king's banner. They began on both sides to form their array again, and for a long time there was a pause in fighting. But before the battle began again Harald Godwinson offered his brother, Earl Toste, peace, and also quarter to the Northmen who were still alive; but the Northmen called out, all of them together, that they would rather fall, one across the other, than accept of quarter from the Englishmen. Then each side set up a war-shout, and the battle began again.

In other words, Tostig takes command of the Norse army and reforms the shield wall. Harold offers another truce. Tostig rejects it again. The English attack again. Just as they are getting the upper hand, the fleet guard arrive under the command of Eystein Orre, fiancé of Hardrada’s daughter.

Eystein Orre came up at this moment from the ships with the men who followed him, and all were clad in armour. Then Eystein got King Harald's banner Land-ravager; and now was, for the third time, one of the sharpest of conflicts, in which many Englishmen fell, and they were near to taking flight. This conflict is called Orre's storm. Eystein and his men had hastened so fast from the ships that they were quite exhausted, and scarcely fit to fight before they came into the battle; but afterwards they became so furious, that they did not guard themselves with their shields as long as they could stand upright. At last they threw off their coats of ringmail, and then the Englishmen could easily lay their blows at them; and many fell from weariness, and died without a wound. Thus almost all the chief men fell among the Norway people.

For a while the English buckle under the onslaught, but the fleet guard are exhausted after their jog. They discard their armour to concentrate their energy on swinging axes. It makes them vulnerable. Heimskringla says that all the 'chief men' were killed in this phase of the conflict. It does not mention Tostig’s fate, but one of the other sagas says that he: “fell there with glory and a good reputation”. Regardless, the English are victorious in the main battle.

Revised Battlefield Location

Where then should archaeologists look for evidence of the battlefield? There are two clues, both in Heimskringla. One is that Hardrada does not move far from where he first sees the English army. The other is the shape of the shield wall. Heimskringla explains:

Then King Harald arranged his army, and made the line of battle long, but not deep. He bent both wings of it back, so that they met together; and formed a wide ring equally thick all round, shield to shield, both in the front and rear ranks. The king himself and his retinue were within the circle; and there was the banner, and a body of chosen men. Earl Toste, with his retinue, was at another place, and had a different banner. The army was arranged in this way, because the king knew that horsemen were accustomed to ride forwards with great vigour, but to turn back immediately. Now the king ordered that his own and the earl's attendants should ride forwards where it was most required. "And our bowmen," said he, "shall be near to us; and they who stand in the first rank shall set the spear-shaft on the ground, and the spear-point against the horseman's breast, if he rides at them; and those who stand in the second rank shall set the spear-point against the horse's breast."

The shield wall was enclosed, which implies they were on open ground. Heimskringla says that the shield wall was equally thick all the way around, which implies the slope was the same all the way around. The inner rank form a shield wall, which was presumably to prevent English archers shooting over the facing shield wall to hit the rear of the opposite wall. If so, the shield wall must have been an eccentric ellipse or stadium (i.e. sausage) shape. English archers would have used the same bows as their opponents. Some Viking bows had draw weights of 80lbs plus. Pulled to the chest, they would have had a range of around 100m. The Norse shield wall was therefore probably no more than 75m across the short side.

“Not deep?” It had to be at least three deep because Heimskringla describes the deployment of three ranks. Assuming the main Norse army - i.e. excluding the fleet guard - had 6,000 men of which 1,000 were archers, barons and their entourages, the shield wall would have been roughly 700m on the long side.

Just to reiterate, we think the battlefield was on a raised section of the moraine ridge near where it was crossed by a Roman road. We think it more likely that the Norse army was on the northern road near Holtby (X on Figure 17), but we cannot discount the possibility they were on the southern road near Dunnington (Y). If they were on the southern road, they had the choice between Mill Hill (b) or Mill Field (c), to the north. If they were on the northern road, they might have defended Mill Hill or Mill Field to the south or Mirk Hills (a) to the north. Only three candidates. All three are adjacent to Holtby, so we feel justified to say the battlefield was at Holtby. Which best fits Heimskringla’s description?

Figure 20: Likely battlefield location

We did not consider Mill Field in the first edition of this book. If the Norse army was on the southern Roman road, it is the most likely battlefield location, perhaps doglegging north to encompass the peak of Mill Hill. We think it more likely that the Norse army was on the northern road, in which case, they could choose between Mill Hill to the south or Mirk Hills to the north. There is little to separate them. Mill Hill is a little higher and a little steeper, but Mirk Hills has more unform slopes, which better fits Heimskringla’s statement that the shield wall was the same depth all the way around. On balance, we would prefer to defend Mirk Hills, but there is almost nothing between the two candidates.

Revised Flight

The main battlefield was probably scavenged. Heimskringla, ASC-D and William of Malmesbury all say that the English chased and slew the Norsemen as they fled from the battlefield. They presumably dumped their weapons and armour, in order to run faster. There may be a better chance of finding isolated remnants of the Norsemen as they fled.

Henry of Huntingdon picks up the story on the main battlefield:

the Norwegians were forced to give way before the superior numbers of the English, but retreated in good order. Being driven across the river, the living trampling on the corpses of the slain, they resolutely made a fresh stand.

The Norse and Flemish survivors evidently retreated across the bridge at Derventio. We imagine this retreat was like a human millipede - a military column eight to ten abreast with shields facing outwards - crawling slowly towards the river and then across the bridge. The trampling means that the English were attacking the head of the millipede as it shuffled along.

Having made it to the bridge, the Norwegian giant is deployed. Henry of Huntingdon explains:

Here a single Norwegian, whose name ought to have been preserved, took post on a bridge, and hewing down more than forty of the English with a battle axe, his country's weapon, stayed the advance of the whole English army till the ninth hour.

The Norwegian giant was clearly posted to buy time for the other survivors to do something. He could not be expected to hold out for more than 10 minutes - regardless how long he actually held out - which would limit the possibilities. This is one reason why historians interpret Huntingdon’s statement that “they resolutely made a fresh stand”  to mean that they formed a new shield wall at Battle Flat, which was roughly 10 minutes from the traditional bridge location (20 minutes from the revised location) at Stamford Bridge.

We do not believe that the Norse would deliberately defend Battle Flat under any circumstances, for a host of reasons we list above. If they wanted to make a stand to buy time, they would obviously have done so at the bridge.

The question is whether all the Norse survivors participated in the stand or whether it was just a bridge guard posted to allow the rest to flee for their ships. From what William of Malmesbury says about what happened after the giant was skewered, it sounds like the latter:

The army immediately passing over [the bridge] without opposition, destroyed the dispersed and fleeing Norwegians.

Heimskringla does not mention a bridge guard or the giant:

This [the death of the Norse leaders at the battlefield] happened towards evening; and then it went, as one might expect, that all had not the same fate, for many fled, and were lucky enough to escape in various ways; and darkness fell before the slaughter was altogether ended.

ASC-D:

… the Norwegians that were left fled from the English, who slew them hotly behind; until some came to their ships, some were drowned, some burned to death, and thus variously destroyed; so that there was little left.

So, the Norse survivors fled towards their ships. We explain earlier why we think their ships were moored between Sutton upon Derwent and Wheldrake (SD and W on Figure 16). Some Norsemen might have fled along the riverbank. If it were us, we would have run roughly along the 10m contour, where English horses would get stuck in the mud but humans might not. The rest would have retraced their steps, fleeing along the Roman road to High Catton (H), then south down the Escrick moraine ridge.

Figure 21: White dots showing where archaeology might be found

If we are right above, Norse battle related artifacts might be found east of the Derwent. We guess that the best places to look will be near the moraine ridge crest and along the 10m contour line between the moraine ridge and the Derwent, shown as white dots on Figure 21.

Loose Ends

The battle scenario outlined above is simple, logical and militarily sensible. It ties together all the source accounts into a coherent narrative, just leaving what might seem to be an anachronism and a contradiction. In addition, there are three interesting military puzzles and English Heritage’s criticisms of Heimskringla.

The Anachronism

According to Laing’s translation, Sturluson (Heimskringla’s author) repeatedly says that the English army had a large contingent of archers and cavalry, and that they started hostilities with a cavalry attack. Historians think that cavalry and archers were only introduced to English warfare after the Norman Conquest. Freeman and others use this apparent anachronism as evidence that the sagas are untrustworthy. If the sagas are wrong, our proposed battle scenario is probably wrong too.

Battle of Hastings accounts make it clear that English horseback troops at the time were mounted infantry rather than cavalry. The difference is not just semantics. Mounted infantry used their horses to get around the battlefield and to outflank their enemy but dismounted to fight with sword or axe. Knights fought on horseback with lances. Knights and their horses were trained to give and take commands by knees and heels. Cavalry horses were bred to be strong and trained to ignore the noise of battle and the smell of blood. Historians are probably right that the English had not developed these skills at the time of the battle.

We think the sagas are not trying to say that the English had a lot of knights. The original Icelandic Heimskringla script says: “Haraldur konungur Guðinason var þar kominn með her óvígjan, bæði riddara og fótgangandi men”, which (we are told) is most likely to translate as: “King Harold Godwinson had come with a hostile army of both horsemen and infantry”. Laing’s translation is viable, but the original is ambiguous: ‘riddara’ can mean knights or horsemen. It is generally interpreted to mean knights in recent times, because it is the Icelandic name for the chess piece, but chess terms post-date Heimskringla. If it meant horsemen, it is not anachronistic. Even if Sturluson did mean knights, he might have misunderstood the meaning of the oral saga, because he transcribed it more than a hundred years after the battle, by which time horseback troops in England were likely to be knights.

As for archers, many historians assume that there were few bowmen in England at this time because none are mentioned in the contemporary accounts of the Battle of Hastings and only one is depicted on the Bayeux Tapestry.

We think the English army at Stamford Bridge probably did have a significant number of archers. Harold pressed the surviving Northumbrian militia. Many people in Northumbria were of Viking descent. They would have continued to use Viking-style bows, for hunting if nothing else. Indeed, it looks like the practice had spread to Mercia, Wales and parts of Essex because, for instance, “The Battle of Malden”, a 10th century poem, mentions English bowmen several times. Most of Harold’s army came from these bow-wielding regions. Perhaps the reason there were no English bowmen at Hastings is that the practice had not made it as far south as Wessex, Sussex, London or Kent.

The Contradiction

Forester’s translation of Historia Anglorum says:

At last someone came under the bridge in a boat, and thrust a spear into him, through the chinks of the flooring. The English having gained a passage, King Harald and Tostig were slain; and their whole army were either slaughtered, or, being taken prisoners, were burnt.

It seems to imply that Hardrada was killed after the English cross the Derwent. Heimskringla and ASC-D say that Hardrada was killed at the main battlefield. If that is what they are saying and they are all right, the main battlefield was east of the river. We think the main battlefield was west of the river. If we are right, one or other of those sources is either wrong or has been misinterpreted.

Historia Anglorum, or its translation, look culpable. It contradicts the ASC D account, which carries the same narrative only with Harald and Tostig being slain at the main battlefield before the surviving Norsemen retreated over the bridge. Huntingdon was writing 55 years after the event. It is possible that he worked from a faulty contemporary account that has subsequently been lost. We suspect that Forester’s translation is more to blame. His term ‘having’ and his semicolon make it sound like Hardrada and Tostig were killed as a consequence of the English gaining passage across the bridge, which means they died on the east side of the river. We disagree with the translation.

The Latin original says: “Transeuntes igitur Angli, Haroldum et Tostig occiderant, et totam Norwagensium aciem vel armis straverunt vel igne deprehensos combusserunt”. Forester’s translation of ‘occiderant’ to ‘were killed’ is viable but it is pluperfect, so it would normally be translated as ‘had fallen’. We translate this passage as: “The English gained passage, Harald and Tostig had fallen, and the entire Norwegian army was either slaughtered or burned.” It looks to us like a list of events to conclude the narrative, with no implication that Hardrada or Tostig died as a consequence of the English gaining passage over the bridge. If so, the main battlefield could be west of the river without contradicting any sources.

Military Puzzle 1

Huntingdon says that the Norse make a ‘fresh stand’ after being forced over the river. It is important because it is the only evidence in all the contemporary accounts that there might have been any organised fighting east of the river. It is one of the main justifications for the main battle having been fought at Battle Flat. But why then is it not mentioned in any of the other accounts?

We think the ‘fresh stand’ is being taken out of context. The whole passage reads: “Being driven across the river, the living trampling on the corpses of the slain, they resolutely made a fresh stand. Here a single Norwegian, whose name ought to have been preserved, took post on a bridge”. Once the giant had been overcome, Huntingdon explains that the English swarmed across the bridge to chase and slaughter those fleeing. It is clear to us that Huntingdon’s ‘fresh stand’ refers specifically to the Norse attempt to hold the English at the bridge and that he does not say or imply that there was any organised fighting east of the river.

Military Puzzle 2

If, as Heimskringla says, Hardrada quickly realises that he is outnumbered and outgunned, why did he reject Tostig’s advice to flee for the boats?

Heimskringla starts its obituary for Hardrada by saying: “King Harald never fled from battle …”. It sounds like he was always itching for a fight, no matter what the odds. But this is not what it means. It is trying to say that having engaged in a battle, he never fled. Reading back through Heimskringla it is clear that he assiduously avoided battle if he thought he might lose. The rest of the sentence says:

… but often tried cunning ways to escape when faced with great superiority of forces.

Hardrada was faced by superiority of forces at Stamford Bridge. His only chance of survival, apart from fleeing, came from incompetent English command or the timely arrival of his fleet guard without exhausting themselves. Neither were likely. Moreover, Hardrada had brought an entire generation of Norwegian warriors. If his army was badly defeated, Norway would be left powerless and undefended. One would have thought he would be especially circumspect considering the consequences of defeat, which makes it all the more odd that he rejected Tostig’s advice.

The obituary ends:

All the men who followed King Harald in battle or skirmish said that when he stood in great danger, or anything came suddenly upon him, he always took that course which all afterwards saw gave the best hope of a fortunate issue.

We think this was the case at Stamford Bridge. In our opinion, Hardrada worked out that he had a greater chance of surviving the battle at Holtby than of successfully fleeing, because the latter was fraught with danger too.

It is 3km from Derventio to Holtby. The English army could not have been more than 3km west of Holtby when Hardrada acknowledges what it is. The English horsemen could beat the footbound Norsemen to the bridge and blockade it. Even if the Norse could get there first, it might have taken an hour for them to stream single file over the bridge, during which time those still on the west bank would be trapped and vulnerable. Meanwhile the English horsemen could cross the Derwent upstream at Howsham and get back on the other side, to catch the fleeing Norsemen in the open. Moreover, Hardrada could not be sure that the ships were safe. Harold might have dispatched part of his army to overcome the fleet guard and hole the ships, in which case the Norse army would get trapped opposite Wheldrake on adverse terrain.

If, as we think, the Norse army was at Mill Hill or Mirk Hills, they were at the best defensive position west of the Derwent. Hardrada may well have reasoned ‘better the devil you know’. He says as much when he explains to Tostig how difficult the English will find the battle. He sends messengers to summon the fleet guard. If the English had not defeated the fleet guard, they should arrive in time to turn the battle. Otherwise, a retreat to the fleet would have been disastrous.

Hardrada’s judgement would have been influenced by his faith. He would have believed to his bones in Valhalla and Tiwaz, the magical spear that Odin wielded to control battles. If he chose to fight, he would have expected that Odin would use Tiwaz to deliver a Norse victory. Fighting valiantly, at the very least, would guarantee him a place in Valhalla. Fleeing could not be influenced by Odin or Tiwaz. If he chose to flee, his survival and his place in Valhalla would depend on factors outside of his control. If we were Hardrada, we would have chosen to summon the fleet guard and fight too.

Christian Tostig, on the other hand, would want to flee. He would not have believed in Valhalla or Tiwaz, and he had led an immoral life. His inclination would have been to fight another day on more favourable terms, then to buy salvation by building monasteries when he gained power.

Military Puzzle 3

When the Norse army retreat across the river, why did they not defend the riverbank until dusk, then flee for their ships under the cover of darkness? Considering that a single Viking warrior held the entire English army at bay, during which time he killed 40 Englishmen, even a diminished Norse army could easily have held the riverbank until dusk.

The most common explanation is that the Norse were too demoralised to fight once their champion had been defeated. Clearly ridiculous. We guess the issue was that the ships were lightly defended. The Norse survivors could not tarry by the river bank for fear that English horsemen would cross the Derwent upstream at Flawith (i.e. the traditional location of the bridge at Stamford Bridge) or further upstream at Howsham, then ride south down the Fulford moraine ridge to sabotage the ships.

It was 7½ miles from Derventio to the riverbank opposite Wheldrake, where we think the fleet was moored. The fittest Norsemen could jog cross country at roughly 6mph without armour. They could get to the fleet in perhaps 75 minutes. The English horsemen could trot at roughly 10mph. They could get to the fleet in perhaps 45 minutes. Whoever was commanding the Norse army would have calculated that the fittest of the Norse survivors could make their escape if they had at least a 30-minute head-start. Huntingdon’s account seems to say that they stationed a bridge guard, including the Viking champion, to buy time. Presumably, their orders were to block the bridge for at least 30 minutes.

The English accounts say or imply that all the invaders were killed or captured. Heimskringla says that some survivors from the battle did escape, although these might have been the lucky few that Harold allowed to return home. By the sound of it, the bridge guard did not quite hold out long enough.

English Heritage on Heimskringla

Our analysis above relies heavily on Heimskringla’s account. Historians do not trust it. Leadman says that it is too poetic to be credible. It certainly has errors. It says, for instance, that Morcar died at the Battle of Fulford. It certainly exaggerates the achievements and heroic deeds of its main protagonists, but it had no incentive to invent place names, troop composition or enemy manoeuvres. Thomas Carlyle said that the sagas were among the greatest works of history ever written. We are inclined to agree.

In our opinion, all the apparent inconsistencies between Heimskringla and the traditional narrative are caused by historians working back to front: they assume the traditional narrative is correct and list the many contradictions in Heimskringla as faults. But the traditional narrative is not fact. It is conjecture based on interpretation of enigmatic accounts. In our opinion, almost every part of the traditional narrative is wrong. What we try to show above is that there are alternative interpretations of the English contemporary accounts that match Heimskringla, thereby establishing a more credible narrative.

English Heritage (EH) go a step further than their predecessors by listing the most extreme examples of Heimskringla’s faults and inadequacies. We thought it might be useful to counter-comment. Note that EH spell Sturluson’s name as Sturlasson and that they use the A.H. Binns’s 1968 Heimskringla translation. We will switch to that for this section. This is what it says about the days before the battle:

Now King Harald began his expedition to conquer York, and the army lay at Stamford Bridge, and because the king had won such a great victory against great chieftains and superior force, everyone was afraid, and did not think there was any hope of withstanding him. The townsfolk decided to send an offer to Harald, yielding themselves and the town to him, and on the Sunday King Harald went with his troop and all the army to York, and had a meeting outside the town, and all the great men agreed to submit to King Harald and gave him hostages, rich men’s sons whom Tostig could pick out for the king, because he knew who were the worthiest men in the town. They went back in the evening to the ships with an undisputed success and were contented. It was decided that there should be a meeting in the morning in the town, when King Harald was to appoint governors of the place and give them offices and areas.

At dawn on the day of battle:

And on the Monday when King Harald Sigurdarson had eaten, and all his army had breakfasted, he ordered the signal for landing to be blown. He divided up the army who was to go or stay behind, had two men from each troop ashore whilst one remained, so he had two parts of the army. Earl Tostig got ready to go ashore with his troop with King Harald. But Olaf the king's son stayed behind to look after the ships, so did Eystein Moorcock, son of Porbrand Arnason, who was the best and dearest to King Harald of all his nobles. King Harald had promised him his daughter Maria when they got back to Norway. The weather was very hot and sunny, and they left their mailshirts behind and went ashore with shields and helmets and spears and wore their swords and many had bows and arrows.

The engagement:

They were very happy, with no thought of any attack, and when they were getting near the town they saw [a great cloud of dust and under it] bright shields and shining mail. They saw that a great army was riding towards them and King Harald straight away halted his army, had Earl Tostig called to him, and asked him what army that might be that was coming towards them. The earl answered that it was most likely enemy, but it might be, he said, some friends of ours, who may wish to join us with friendship and offer us their help and loyalty. The king spoke, 'We must await quietly this army which is coming'. They did so and the army got bigger and bigger as it got nearer and they saw it clearly, and it was like looking at an ice-field. Then said the earl, 'My lord, let us take some shrewd plan; it is not to be hidden that those are enemies, and the king himself must be with such an army'. Then said King Harald, 'What is your advice?' Earl Tostig answered, 'The first thing to do is to turn back as quickly as possible to our ships for our men and our armour and then after offer such battle as we can; but another plan would be to take to our ships, and then the cavalry cannot overcome us'. The king said, 'We shall do something else. Put our fastest horses under three bold fellows: let them ride as fast as they can and tell the Norwegians of the danger; they will come straight away to help us. The English must sooner expect more fight from us than flight, and we shall fight bitterly a good time before we acknowledge we are beaten.' The earl said, 'You must decide in this, my lord, as in everything, and I was no more eager to flee than anyone else'.

EH: “Heimskringla states that Tostig and Hardrada waited with their army at Stamford Bridge. However, due to an understandable confusion concerning local topography, Sturlasson is less precise in identifying Stamford Bridge as the actual site of the battle.
MB: Heimskringla does not say or imply that Tostig and Hardrada waited with their army at Stamford Bridge on the day of battle. It does say that the Norse army and their commanders were at Stamford Bridge several days before the battle, preparing to sack York. It is irrelevant to the battle narrative because it also says that they went back to their ships later that day and that they slept on their boats on battle-eve. Heimskringla does not say or imply - imprecisely or otherwise - that Stamford Bridge was the actual site of the battle, not least because it implies the battlefield was elsewhere.

EH: “Sturlasson was ignorant of the topography of both the battlefield and East Yorkshire, seeming to believe that the geography of Hardrada’s campaign was condensed into a very limited area with Stamford Bridge, Fulford and Riccall all located close to the walls of York.
MB: Fulford was close to the walls of York. Heimskringla does not say or imply that Stamford Bridge was near the walls of York. It never mentions Riccall. EH are trying to match Heimskringla to their interpretation of the English accounts. They think that Sturlasson is compressing the geography because they assume that Heimskringla’s references to the location of the fleet and the location of the battle referred to Riccall and Battle Flat. Heimskringla says nothing of the sort. Its statements that the approaching army was too far away to identify whether it was hostile, that Hardrada orders the fastest riders be put on the fastest horses to summon the fleet guard, and that the fleet guard arrives so exhausted that some die from weariness, implies to us that Sturlasson knew perfectly well how far it was between York, the fleet and the battlefield.

EH: “It is also apparent that Sturlasson could have confused events that occurred at the Battle of Hastings with those of Stamford Bridge.
MB: The Battle of Hastings and the Battle of Stamford Bridge both featured incomplete losing armies and shield walls that were breached when troops ran out to chase retreating attackers. However, there are only two ways that a determined shield wall can be defeated: 1) By killing or barging aside everyone in a section of the wall; 2) For everyone in a section of the wall to leave their position. The fact that the English won means that they successfully employed one or other of these tactics. The latter would be quicker and less casualty prone. We guess that attackers always tried to goad parts of a shield wall out of place and that ill-discipline in shield-walls was a common problem with less experienced troops.

EH: “The Heimskringla is confused as to whether the Viking army is making its way towards York or Stamford Bridge early on the morning of Monday 25 September 1066.
MB: Heimskringla says that on Sunday: “It was decided that there should be a meeting in the morning in the town”. Then on Monday: “They were very happy, with no thought of any attack, and when they were getting near the town they saw a great cloud of dust and under it bright shields and shining mail.” It is clearly saying that the Norse army was heading for York and getting near. Again, EH try to match Heimskringla to their interpretation of the contemporary accounts. They wrongly assume that the battle was fought at Stamford Bridge, so wrongly assume that the Norse army was heading for Stamford Bridge. Heimskringla is not confused. It says that the Norse army was heading for York and means it.

EH: “According to the Heimskringla, Hardrada and Tostig had barely reached Stamford Bridge when they became aware of Harold’s approach”.
MB: Again, Heimskringla says nothing of the sort. It does not say or imply that the Norse army was at Stamford Bridge on the day of battle. It says: “when they were getting near the town they saw a great cloud of dust and under it bright shields and shining mail.” It is implying that the Norse army was on the Fulford moraine ridge when they spot the English leaving York. Again, EH try to match Heimskringla to their faulty interpretation of the English contemporary accounts. They think that the initial engagement was on the west bank of the Derwent, so they wrongly assume the Norse army had only just got across the bridge at Stamford Bridge when they become aware of the English army.

EH: “The Heimskringla states that as the Viking army neared Stamford Bridge it became aware of Harold’s army approaching from the west along the north bank of the Derwent.
MB: Heimskringla does not say or imply that the Norse army was nearing Stamford Bridge on the day of battle. It does not say or imply that the English army approached Stamford Bridge along the north bank of the Derwent. EH try to match Heimskringla to their faulty interpretation of the English accounts and their faulty interpretation of the contemporary geography. Heimskringla only says that Hardrada sees the English army as the Norse army approaches the town/city/castle. This is only possible if the Norse army was on the Fulford moraine ridge or west of it.

EH: “For the detail of the fighting on Battle Flats we are forced to rely upon the Heimskringla with all its anachronisms”.
MB: Heimskringla provides the only detailed account of the fighting. It is therefore true that everyone has to rely on it of this part of the narrative, but Heimskringla does not say or imply that any fighting happened at Battle Flat or Stamford Bridge. Again, EH try to match Heimskringla to their faulty interpretation of the English accounts. We presume the anachronisms to which EH refer are the presence of English ‘cavalry’ and archers, insofar as most historians think that they were not present in English armies until the 12th century. We have mentioned these above. In short, we think the ‘cavalry’ is a mistranslation of ‘horsemen’ and that the archers were there.

Sources

Contemporary sources

Anglo-Saxon Chronicle (of which three versions covered the invasion, known as C, D and E); reasonably contemporary with events
John of Worcester; Chronicon ex Chronicis; c1125
Henry of Huntingdon; Historia Anglorum; c1125
William of Malmesbury; Gesta regum anglorum; c1135
Fagrskinna; c1200; Translation by Alison Finlay, 2004
Snorri Sturluson; Heimskringla; c1200; Translation by Samuel Laing (Vol III), 1844

Modern sources

Auden, G A; The Strategy of Harold Hardrada in the Invasion of 1066; 1927; Journal of the Society for Army Historical Research, VI
Charles R.B. Barrett; Battles and Battlefields in England; 1896; A.D. Innes & Co
F.W. Brooks; The Battle of Stamford Bridge; 1956; East Yorkshire Local History Series (No. 6)
Alfred H. Burne; The Battlefields of England; 2005, Pen and Sword Books
Kelly DeVries; The Norwegian Invasion of England in 1066; 1999; The Boydell Press
English Heritage; English Heritage Battlefield Report: Stamford Bridge 1066; 1995; English Heritage
Edward A. Freeman; The History of the Norman Conquest, Vol III; 1864; Clarendon Press
Historic England listing: https://historicengland.org.uk/listing/the-list/list-entry/1416328
I.G. Lawton; Roman Roads Around Stamford Bridge; 1997; SYO
A.D.H. Leadman; The Battle of Stamford Bridge; YAJ Volume 11 (1890)
Ivan D. Margary; Roman Roads of Britain; 1973; John Baker
Frank McLynn; 1066 The Year of the Three Battles; 1999; BCA
Northern Archaeological Associates: Stamford Bridge Water Pipeline; 2005
H.G. Ramm; A Romano-British kiln, and the Roman road at Stamford Bridge; YAJ Volume 38 (1955)
H.G. Ramm; The Derwent Crossing at Stamford Bridge, The; YAJ Volume 41 (1966)