In the winter of 1069 to 1070 William the Conqueror's army laid waste the north of England. Sixteen years later Domesday Book wrote one word against hundreds of places: vastum, waste. I mapped every one, asked where it fell and what it meant, and looked for its scar on the map of today.
Orderic Vitalis, a monk writing about fifty years after the event, said William had the crops, herds and food of the north burned so that rebels would have nothing to live on, and that more than 100,000 people died of the famine that followed. Historians still argue over how far to believe him. The one record made close to the time is Domesday Book, compiled in 1086. For each holding it gives a value before 1066 and a value in 1086, and against many in the north it writes a single word: waste.
I took the Hull Domesday Project's coding of the whole survey, 20,195 entries, and followed that word. Of the 1,345 entries the Hull coders flag as waste, 79% are in six counties: Yorkshire, Cheshire, Derbyshire, Nottinghamshire, Staffordshire and Shropshire. I asked four questions: where the waste was, what the word means, whether it clusters where an army would have marched, and whether the places it touched are smaller or poorer now. This study builds on three earlier ones: The Conquest in Castles, Is Domesday Destiny? and The Villages That Vanished.
Each dot is a place Domesday names, from the Midlands to the Tees. Ember red means an entry for the place calls it waste and gives it no value in 1086. Amber means the entry uses the word but the place had a value again by 1086, often because the land "was waste" when its new lord received it. A red ring marks a place named only as a member of a larger estate that the entry calls waste, like the 62 vills of Preston. Switch the layers to see value in 1086 against 1066, what stands on each spot today, and the hexagon summary.
Waste is almost absent from the south and east. Outside the six northern and north midland counties and Herefordshire, only 2.1% of places have an entry calling them waste, and the recorded value of those counties rose: in 1086 it stood at 110% of 1066. In Yorkshire, 37% of the 1,302 places with an entry of their own are called waste, and 50% of all 1,786 Yorkshire places are named in a waste entry once the member vills of great estates are counted. Yorkshire's recorded value fell to 35% of its 1066 level, or 32% leaving out the city of York.
Cheshire, harried in the same winter, tops the English counties (only its Welsh borderland ranks higher): 61% of its places carry the word. Yet its value fell less than Yorkshire's, to 42%, and only 14% of its places were still waste with no value in 1086. That gap is the first clue that "waste" does not always mean the same thing.
For every county except Yorkshire, the source carries the English translation of each entry, so I could read how the word is used. I sorted each waste entry by its wording: land that is waste in 1086, land that was waste (usually "when he acquired it", or "he found it waste") but had recovered, land waste before 1066, and land partly waste.
The counties fall into two groups. In Derbyshire, Nottinghamshire and Staffordshire, 95% of the 166 waste entries say the land is waste now, in 1086. In Cheshire and Shropshire, 70% of 384 waste entries look backwards: the land was waste when the Normans took it, around 1070, and had a value again by 1086. Cheshire also has 30 entries, most in the Welsh borderland, that call land waste even before 1066.
Yorkshire has no text in the source, but its figures speak. Of its 518 waste entries, 83% give no value at all for 1086. Where I can check the same rule against text elsewhere, waste entries with no 1086 value say the land is waste now in 80% of 334 cases. So Yorkshire's waste is overwhelmingly waste in 1086 itself, sixteen years after the harrying.
Historians have long disagreed about what Domesday's waste records. The plain reading is destruction by William's army in 1069 to 1070. Others point to later violence, including Scottish raids into the north in the 1070s; to new Norman lords merging and reshaping estates, so that a holding lost its own value when its people and ploughs were moved to a neighbouring manor; and to the possibility that the word sometimes meant only that the commissioners could put no value on a holding. Paul Dalton questioned whether William had the time, men and weather to turn the north into a desert. For the wider debate see Wightman (1975), Palmer (1998), Dalton (2002) and Thomas (2008), listed below; the overview I relied on is the English Wikipedia article on the Harrying of the North. My data cannot settle which reading is right for any one place. What it can show is where the word falls, and whether that pattern fits a marching army.
Yorkshire was divided into wapentakes, the northern word for hundreds. Waste is not spread evenly between them. In Hunthow, 95% of places are named in a waste entry; in Burton, 76%; in Pickering Lythe, 76%. Recorded value collapsed hardest in Burton (to 2% of 1066), Hunthow (5%) and Pickering Lythe (8%).
If waste were the footprint of an army, you might expect it thickest around York, where William spent Christmas 1069, along the Roman roads his troops would have used, and in the rich river valleys where the food was. I measured, for each of the 1,301 Yorkshire places with an entry of its own (leaving out the city of York), its distance from York, its height, the lie of the land around it, and its distance to the nearest Roman road in Itiner-e and the nearest named river.
Distance from York makes almost no difference. The median waste place is 44 km from York, and so is the median place without waste (44 km). Nor do Roman roads: waste places sit a median 3.9 km from one, the others 4.0 km. What does matter is height. Below 25 m, 27% of places are waste; above 200 m, 59%. Places above 150 m are 15% of the total but 24% of the waste.
Taken together, each 100 m of height multiplies the odds of waste by 2.1, other things equal, while distance from York and from Roman roads have no clear effect. Places farther from a named river are a little more likely to be waste. The same holds with the broader measure that counts member vills, except that there places nearer a Roman road are somewhat more often waste (odds ratio 0.84 per 5 km farther away).
The money tells a slightly different story. Near York, value held up better. Within 15 km of the city, rural places were worth 65% of their 1066 value in 1086; between 15 and 60 km, about a third; beyond 60 km, 23% or less. In Cheshire, waste rises steadily away from Chester, from 41% within 10 km to 79% beyond 30 km.
The waste of 1086 is not a trail of burnt villages along the roads to York. It is a map of the hills and the margins.
That fits more than one story. An army living off the land would strip the valleys, but valleys also recover first, as survivors and new lords put the best land back to work. Upland holdings that were marginal before 1066 might never have been resettled by 1086, or might simply have been folded into the lowland manors that owned them. My data cannot tell burning from neglect; it can say that, by 1086, waste had retreated to the higher, poorer ground.
To ask whether waste in 1086 left a mark that lasts, I took the 1,445 places in the six northern and north midland counties that have their own entry and a 1066 value, and compared the 323 that were waste with no value in 1086 against the 1,122 with no waste at all. Each comparison is made within the same county and among places of similar 1066 value, and also adjusts for that value directly. (309 places whose waste entries had a 1086 value are left out of both groups.)
The scar is faint. Waste places are no less likely to lie in a built-up area today (difference -1.4 percentage points, from -8.1 to +5.3), and their parishes are about as populous. Household income in their neighbourhoods is a little lower, by 2.5% (confidence interval 0.5% to 4.4%), a small gap that could easily come from something other than 1069. Where waste places are built up, they sit in bigger towns, not smaller ones: their built-up areas are 61% larger. That is mostly the West Riding: 31% of the built-up waste places are in its southern wapentakes, against 19% of the others, and the woollen towns of Leeds and Bradford grew over land Domesday called waste.
Did waste places become deserted villages? In Yorkshire, 15% of places named in a waste entry are in Beresford's gazetteer of lost villages, against 14% of the rest. In the like-for-like comparison the gap is +4.3 percentage points, with a confidence interval that includes zero. In all, 195 places named in waste entries later became deserted villages, 136 of them in Yorkshire. In that companion study most dated desertions fall after 1348, in the centuries of plague, sheep and parks.
Lancashire did not exist as a county in 1086. North of the Ribble, Amounderness and the lands beyond it to Furness appear at the end of Yorkshire; between the Ribble and the Mersey, the six hundreds of south Lancashire are an appendix to Cheshire. Neither part is described like the rest of England.
All these vills belong to Preston; and there are three churches. In sixteen of these vills there are but few inhabitants, but how many there are is not known. The rest are waste.Domesday Book, Amounderness, in John Porter's translation (History of the Fylde of Lancashire, 1876); punctuation simplified
Amounderness is the most famous waste in the north-west. Domesday names Preston and 61 vills belonging to it, says sixteen had a few people and the rest were waste, and gives no value at all. The Hull coding marks the whole entry as waste, so on my map all its located vills are ringed, which overstates the waste by those sixteen unnamed inhabited vills. Further north, in Lonsdale, Furness and Cartmel, the entries list only tax assessments: they are not called waste, but they record no people or values either. Of 162 located places north of the Ribble, 34% are named in a waste entry.
South of the Ribble the picture is different. Of 56 entries for the land between the Ribble and the Mersey, only 3 use the word: Altcar "was waste", Salford hundred had "12 carucates of waste land", and part of Leyland "is partly waste". But Blackburn hundred tells its own story. Its lords had settled men with 11½ ploughs and given them three years free of dues, "therefore not now valued". That is a landscape being restocked, not one being burned. These entries give values for 1066 holding by holding, but for 1086 mostly one sum for each hundred, so I do not compare values here.
Clitheroe itself is not named in Domesday Book. Within 15 km of the town, the survey names 38 places, 18 of them in waste entries: 13 of the 29 in Craven, then part of Yorkshire, and all 5 in Amounderness. None of the 4 places of Blackburn hundred, Whalley and Pendleton among them, was called waste. Bolton by Bowland, Rimington, Nappa and Horton were waste; Grindleton, Slaidburn, Waddington and Mitton were not. Craven as a whole has 36% of its 123 places in waste entries. Today 12 of the 38 lie in a built-up area, from Longridge and Whalley to Barnoldswick; the rest are farms and hamlets, as most of them have probably always been.
Domesday. Entries and places come from the Hull Domesday Project's coded Domesday via the realm-domesday repository. I used the 20,195 entries of the manor and urban kinds, leaving out Yorkshire's summary and claims sections, which repeat holdings. "Waste" is the Hull flag on an entry. A place is "waste (own entry)" if an entry that describes it is flagged, and "named in a waste entry" if it is also listed as a member vill of a flagged entry. The castles study counted Yorkshire's summary entries as well, which is why it gave a lower Yorkshire share.
Values. Values are in pence. An entry with a 1066 value is paired with its 1086 value; a waste entry with no 1086 value counts as worth nothing in 1086 (225 Yorkshire entries); other entries without a 1086 value are left out (85 in Yorkshire). Counting only entries with both values, Yorkshire stood at 39% of 1066. An entry's value is shared equally among the places it describes.
Text. For counties with text I sorted waste entries by their wording with simple patterns ("is waste", "was waste", "found it waste", "waste before 1066", "partly waste"). A few entries fit none and are counted as other.
Setting. Height, relief (the range of height within 1 km), distance to the nearest named river (OS Open Rivers) and to the nearest Roman road (Itiner-e, with a datum correction) come from the companion study of lost villages, which computed them for every located Domesday place. Distance from York is to the Domesday place of York.
Today. Each place is matched to the nearest ONS Built-up Area 2022 whose edge, drawn as a circle of the same area round its centre, is within 1 km. Built-up area populations, from summing Census 2021 LSOA counts by the ONS best-fit lookup, exist only for larger areas, so I use area as the size. Income is the ONS estimate for the MSOA containing the place's LSOA. Parish residents come from the companion Domesday Destiny study. The long-shadow comparisons are linear regressions with county-by-1066-value-tercile fixed effects plus log 1066 value, with robust standard errors.
Known problems. The Hull flag cannot tell "waste now" from "was waste" in Yorkshire; I use the missing 1086 value as a stand-in. Member-vill lists inflate waste where only part of an estate was waste, as at Preston. Places are located to 1 km. Today's measures describe neighbourhoods, not the medieval vill, and nine centuries of other history lie between. None of this proves cause: it describes patterns.