The Vikings left two kinds of trace in Britain: buried silver, and the names of the places they farmed. Do the two maps agree? And where is there silver without names, or names without silver?
In May 1840, workmen mending the embankment of the River Ribble at Cuerdale, near Preston, found a corroded lead chest holding more than 8,600 pieces of silver: coins, ingots, jewellery and chopped-up bullion, 32 to 40 kg of it. By the archaeologist Richard Hall's reckoning it is four times larger than any other Viking hoard from Britain or Ireland. Within 20 km, on the same map, are places called Kirkham, Ribby and Westby.
Those are the two things the Vikings left behind on the ground. Silver, buried and never collected. And names: the farms and villages ending in -by, -thorpe, -thwaite and -toft, that historians have long read as marks of Scandinavian settlement. Both are used to draw the map of Viking Britain. I wanted to know whether they draw the same map.
So I put 19 Viking hoards from the Wikipedia lists of hoards in Great Britain against every settlement on the Ordnance Survey's gazetteer, 43,274 of them, of which 1,360 carry a Norse element. Mostly the two maps agree, strongly. But there are some striking gaps, and the biggest gap is a whole county.
Silver discs are the Viking hoards, sized by how many pieces the list gives. Rust dots are settlements with a Norse element in their name. Switch on the density layer to see where the names crowd together, the hoards of other periods for comparison, and the approximate line of the Danelaw. Tap a hoard for its story.
To compare a hoard with the names around it, I drew a 15 km circle round each find place and counted the settlements inside: what share of them have a Norse name? Across Britain as a whole the answer is small. A typical place, picked at random from the gazetteer, has 3.1% Norse names in its circle, and more than half of all places have none at all.
Around the 17 Viking hoards I could place to village level in Britain, the average is 10.0%: 3.2 times as high. In 20,000 random draws of 17 places, a result that high came up with a probability of 0.002. 10 of the 17 hoards sit in the most Norse quarter of Britain, and the median hoard has a Norse-named place 5.8 km away, against 20.0 km for a typical settlement.
A fair objection: maybe hoards of any date are found in the north and east, where farming, building and metal detecting turn up buried metal. So I ran the same test on the 57 other located hoards in the Great Britain list, from all other periods (most of them medieval or later). Their circles average 4.7% Norse names, only 1.5 times a random place. The Viking hoards are clearly more Norse than that (a one-sided rank test gives p = 0.016). The result holds with 10 km circles (3.3 times) and 25 km circles (2.7 times), within England alone (2.9 times), and with a broader set of names that adds -holme and -garth (3.0 times).
One comparison is weaker. The 17 hoards that the lists date to the Viking Age but file under Anglo-Saxon or Pictish hoards sit in less Norse country (5.2%), but with so few hoards on each side the gap could be chance (p = 0.15).
Where the Vikings buried silver, they, or people like them, also named farms. The two maps broadly agree.
In the late ninth century Alfred of Wessex and the Viking leader Guthrum agreed a frontier: up the Thames, up the Lea, to Bedford, up the Ouse and along Watling Street. Land north and east of it became the Danelaw. The line on the map is my approximate tracing of that description, then a straight line from Wall, near Lichfield, to Chester, as in my earlier study of landscape words.
The Norse names are almost all on the Danelaw side: 96.5%, against 49.0% of all English settlements. The Viking hoards follow: twelve of the fourteen in England. The two that do not are the Watlington Hoard in Oxfordshire and the Eye Hoard in Herefordshire, both dated by the list to the late 9th century and both holding coins of Alfred and of Ceolwulf II of Mercia. They look less like settlers' savings than like silver carried through a frontier zone in a time of war.
Five of the 17 Viking hoards have no Norse-named settlement at all within 15 km:
Their stories fit. Ashdon in Essex, Watlington and Eye are all dated to the late 9th century, the years of the great Viking armies, when silver could move with soldiers rather than with settlers. Talnotrie is in Galloway, where my element list finds few Norse names: its nearest is Sorbie, 25 km away. Bryn Maelgwyn, near Deganwy on the north Welsh coast, held 204 pennies of Cnut and dates to the early 11th century, a generation or more after every other hoard in the set. These look like hoards on the move, not hoards at home.
Now turn it round. Of all 1,360 Norse-named settlements, 58% are more than 50 km from any Viking hoard in the lists. And one county stands out. Lincolnshire has 294 Norse-named settlements, 32% of all its settlements and 23% of all of England's. It has no Viking hoard in the lists, and no hoard of any kind dated 793 to 1066. The typical Lincolnshire settlement is 103 km from the nearest one.
The ten 10 km squares with the most Norse names in Britain are all in Lincolnshire and Leicestershire, places like Spilsby, Skendleby and Revesby. Every one of them is between 85 and 123 km from the nearest Viking hoard in the lists.
I do not think this means the Lincolnshire Vikings had no silver. It means that the lists, and perhaps the finds themselves, are shaped by other things: where great hoards were buried in a crisis and never dug up, where the ground has been disturbed, where finds were reported and became famous enough for a Wikipedia article. The Wikipedia lists are a register of notable hoards, not of every coin found. A full count from the Portable Antiquities Scheme or the Treasure reports could look different, and that is the obvious next test. But on the evidence of the notable hoards, the densest Norse-named landscape in Britain is a silver blank.
Vikings arrived by ship, so I expected hoards near the sea or big rivers. They are not, on this evidence. The median Viking hoard is 40.5 km from the open coast and 20.1 km from tidal water (the sea or a tidal river), further than a typical settlement (24.0 and 14.9 km). For rivers at least 50 km long the hoards are no closer than ordinary places either (5.1 km against 5.6 km). A handful are right at the water's edge, Cuerdale on the Ribble among them, but the group as a whole lies inland, many of them in Yorkshire.
Every story about Viking silver in Britain comes back to Cuerdale. The hoard was found on 15 May 1840 on the south bank of a bend in the Ribble, packed in a lead box. It held more than 8,600 items and weighed 32 to 40 kg, mostly bullion. Its coins were mostly from the Viking kingdoms of eastern England, with Wessex pennies and foreign coins, and it was buried between about 903 and 910, soon after the Vikings were expelled from Dublin in 902. One idea is that it was a war chest for Irish Norse exiles planning to retake Dublin, at a time when the Ribble valley was a route between the Irish Sea and York; there are other theories.
Cuerdale alone is 80% of all the pieces counted across the 16 Viking hoards for which a number is given, and about 13 times the Vale of York Hoard, the next biggest.
The surprise is the names. Cuerdale is only 0.7 km from the Ribble and 1.5 km from tidal water, and 105 km in a straight line from York. But of 275 settlements within 15 km, only two have Norse names; within 25 km there are eight, out of 667:
Lancashire as a whole has 13 Norse-named settlements out of 772, or 1.7%. My element list is deliberately narrow, counting only settlement names with the clearest elements, so it will miss some Norse names here as everywhere; but the contrast with Lincolnshire is still stark. The richest hoard sits in country with few names; the richest names sit in country with no hoards. That is what you would expect if Cuerdale was silver passing through, on a river road, rather than the wealth of the farms around it. Lancashire's other Viking hoards, at Silverdale and Warton, lie 4.9 km apart on the edge of Morecambe Bay, where the names are thicker.
Show Cuerdale and the Ribble on the map
The hoards are a short, selective list. 19 Viking hoards in Great Britain, 17 placed to village level, is not many, and they are the notable ones: hoards with enough history or publicity to be listed on Wikipedia. Of these, nine were found in 2000 or later. The Ireland list has only one Viking hoard, so I map the Irish hoards but test only Britain, where the place-name data are.
Hoard points are approximate. Most are the village-level points given in the lists. Watlington's point is the town, because the find spot has not been published. The Storr Rock Hoard is placed only roughly on Skye, from its name, and is left out of the tests. The Galloway Hoard's find spot is withheld, so it is not mapped.
Names are modern. OS Open Names is today's map. A -by name says a place has a Norse element now, not that Vikings founded it, and some names are later coinages or borrowed words. I counted only settlements, only elements that my cited sources give as Old Norse, and only where they are not likely to be false friends: kirk only in England, since Scots kirk means church too (Falkirk); -ster and -bister only in the Northern Isles and the Highland area; -by in Scotland only where Norse names are common. A sample check still finds doubtful cases, such as Darby End in the West Midlands. The broad set, adding -holme and -garth that can also be English or Welsh, gives the same result.
Agreement is not cause. Silver and names both cluster in the north and east because that is where Scandinavians had power, but also because of where land has been ploughed, built on and searched. The comparison with hoards of other periods helps with that, but cannot remove it.
Hoards. From the Wikipedia lists of hoards in Great Britain and in Ireland (CC BY-SA 4.0), I took every row in the lists' "Viking hoards" tables (20 hoards, after merging the list's Herefordshire Hoard row, which is the Eye Hoard). As a second, wider set I added every other row whose sortable date in the list is 793 to 1066 (19 more). Each hoard is placed at the coordinate given in its list row; the few without one were placed from the stated find place in OS Open Names, or from Wikidata (Talnotrie). Each hoard's source is recorded in the data. I checked the 7 hoards that also appear in Wikidata: the largest gap between the two coordinates is 0.88 km. Counts of coins and items are taken only where the list states a number. Baseline: the 57 other located hoards in the Great Britain list.
Names. Every OS Open Names populated place (43,274 cities, towns, villages, hamlets, suburbs and other settlements in Britain), split into words, tested for whole-word matches to -by, -thorpe, -thwaite, -toft, kirk-, -scale, and in the far north -ster, -bister and -bost. Meanings come only from cited sources: Wikipedia's List of generic forms in place names in the British Isles and Toponymy of England, the Survey of English Place-Names entry for Scaleby, and the Shetland Amenity Trust's note on -bister.
Tests. For each point, the share of settlements within 10, 15 and 25 km that carry a Norse element. Random baseline: 20,000 draws of the same number of settlements. Rank tests compare the Viking hoards with the other-period hoards. Distances to coast, tidal water and major rivers use Natural Earth land and OS Open Rivers.