Britain is full of places named after the spot where people crossed a river: Oxford, Bradford and Stratford waded it; Cambridge, Tonbridge and Pontypridd built over it. Which kind of place is richer today, in total and per household? And did having a bridge help a place grow?
Before there were bridges there were fords. A ford is a shallow place where people, carts and cattle could wade across, and Old English named hundreds of settlements after them. A bridge needed money, timber or stone, and someone to keep it up. So I wondered: if a place was named for its bridge, did that bridge make it?
I took every populated place on the Ordnance Survey's open gazetteer, 43,274 of them, and sorted their names into fords (Oxford, Garforth, Welsh rhyd), bridges (Tonbridge, Bridge of Allan, Welsh pont) and everything else. Then I set them against the 2021 census and the Office for National Statistics' newest estimates of household income for small areas.
The short answer surprised me. Fords win in total, by a lot, and the typical ford place is a little richer per household too. The bridge places are younger names, mostly missing from Domesday Book, and having a bridge in the name does not seem to have made a place bigger.
Each dot is a place whose name says ford or bridge. Ochre for fords, blue for bridges. Dots are sized by the population of the place's own built-up area in 2021, where it has one; small hamlets and villages get the smallest dot. Tap a dot for its income, its Domesday record and the rule that classified it.
Two patterns jump out. England is ford country: it has 1,034 ford places and 434 bridge places. In Wales and Scotland it is the other way round, with 110 bridges to 50 fords in Wales (most of them Welsh pont names) and 99 to 45 in Scotland, where 15 places have names of the "Bridge of" kind. And the bridges sit right on the water: half of all bridge places are within 170 m of a named river on the OS river map, against 360 m for fords and 700 m for everywhere else. That is a good sign the names mean what they say.
To count people I need a settlement with edges. The ONS draws "built-up areas" around every town and village, and I matched each place name to the built-up area of the same name nearby, then summed its 2021 census population. In England and Wales 144 ford places and 69 bridge places have a built-up area with a population of their own.
Ford towns hold 2,290,193 people, 3.8% of everyone in England and Wales. Bridge towns hold 867,692, or 1.5%. That is 2.6 times as many people, and 2.6 times as much household income. Part of that is just numbers: there are 2.1 ford towns for every bridge town. But the fords also have the giants.
Bradford alone, with 331,757 people, outnumbers the top four bridge towns put together (309,791). Ford country also includes Oxford, Watford, Chelmsford, Salford, Bedford, Guildford and Stafford. On the bridge side, Cambridge (149,510) towers over the rest; the next biggest, Stourbridge, has 58,064.
Total size is one question. Wealth per head is another. The ONS small-area income estimates give the mean total income of households in each neighbourhood (an MSOA; the median one has 7,956 residents). For a town I take the population-weighted average of its neighbourhoods. The ONS does not publish income per person at this scale, so household income is the closest measure.
The typical ford town is £6,980 a year richer per household than the typical bridge town. Weight each town by its population, so that every resident counts once, and the race is a dead heat: £55,537 for fords and £55,664 for bridges. That tie is Cambridge's doing. Leave Cambridge out and the bridge figure falls to £52,499; leave Bradford out of the fords and theirs rises to £57,738.
Built-up areas leave out most hamlets and small villages, so I also looked at all 1,084 ford places and 544 bridge places in England and Wales, each given the income of its own town or of the nearest neighbourhood. The picture is the same. The median ford place has households on £55,028; the median bridge place £49,534; the median of everywhere else £51,623.
Big towns and small villages differ in income, so a fair test compares places of the same size. Within every size band but the very largest, ford places come out ahead of bridge places.
Much of England's wealth sits in London, the South East and the East of England. If fords crowd into the south-east and bridges into the north and Wales, the gap could be geography rather than anything about the names. And partly it is: 31% of ford places are in those three regions, against 22% of bridge places. Wales alone holds 20% of the bridge places in England and Wales, and its incomes are among the lowest of any region.
Within regions the fords still lead in 8 of 10, though in two of them only by a whisker. The bridge places come out ahead only in London (with just 8 of them) and the North East. In the West Midlands and Wales the two are level, within 1%.
Putting it all together, I fitted a simple model of household income (on a log scale) for every named place, with the name type, the region and the size band. The number that matters is the gap between bridge places and ford places once those are held equal. I call it the bridge premium. It is negative.
Raw, bridge places have household incomes 8.5% lower than ford places. Holding size equal changes almost nothing. Holding region equal shrinks the gap to 3.7%, so more than half of the raw gap is the south-east effect. With both, it is 3.6% (95% interval 2.1% to 5.2%). It survives dropping the names I am unsure of, dropping two-word names such as "Hebden Bridge" or "Ford End", and keeping England only.
One measure tells a different story. On income after tax and housing costs, adjusted for household size, bridge towns are 1.8% ahead of ford towns, with an interval that comfortably includes zero. Some of the ford lead is gross income that goes on housing.
Fords outnumber bridges, hold more people and, like for like, slightly richer households. If there is a premium in a crossing name, it belongs to the ford.
The second question is the harder one. A bridge name tells us there was a bridge when the name was coined. Did that bridge help the place grow? I tried three ways of looking, and none of them can prove cause and effect.
I looked up every English place in Domesday Book, the survey of 1086. Of the 500 ford-named towns and villages in England, 57% are there. Of the 156 bridge-named ones, only 21%. The average for all other towns and villages is 51%.
Domesday households point the same way. Ford places recorded in 1086 had a median of 22 households; bridge places 13. And 14 ford places had burgesses, the townsmen of a borough, including Hereford, Hertford, Stafford, Thetford and Bradford on Avon. Of the bridge places, only Cambridge did. Most bridge names are later: many surely replaced older names once a bridge was built, and the old name is lost to this kind of matching.
Among towns and villages with a built-up area of their own, bridge places are no bigger than ford places once region is held equal: the estimate is 10% larger, with an interval from 23% smaller to 55% larger. That is no evidence either way. A similar share of each has a built-up area at all (13.3% of ford places, 12.9% of bridge places). For the 1,761 English towns and villages recorded in Domesday, a bridge name adds an estimated 119% to size today once 1086 households and region are allowed for, but the evidence is weak (p = 0.10) and rests on a handful of towns: Cambridge, Bridgwater, Trowbridge, Tonbridge and a few more.
I hoped to ask the direct question: does a place with a medieval bridge do better? Wikidata lists 11,038 bridges in the United Kingdom, but only 1,403 have a date, and only 26 of those are before 1540. They are the famous ones: Elvet and Framwellgate bridges in Durham, the Old Dee Bridge in Chester, the Wye Bridge in Hereford, the Monnow Bridge. The 9 towns with one within a kilometre have a median population of 33,176, against 2,793 for the rest. But that mostly shows which bridges get written about, not what bridges do. Only 4 bridge places and 2 ford places have one. I would not lean on it.
Crossings sit on routes. Bridge places are a little more often within a kilometre of a Roman road in the Itiner-e atlas (21%) than ford places (19%) or other places (17%). Adding the distance to a Roman road to the income model leaves bridge places 3.4% behind ford places. The best check of the road data is the "street fords": all 19 Stratford and Stretford places lie within a kilometre of a mapped Roman road, as their name ("ford on a Roman road") says they should.
Scotland has no small-area income estimates to match the English and Welsh ones, and its built-up areas have no census population in the files I used, so it sits outside the main tests. Its 99 bridge places outnumber its 45 fords. Wikidata gives a population for 22 of the bridges and 17 of the fords, summing to 105,626 and 26,494 people; Coatbridge alone accounts for 42% of the bridge total. Wikidata populations come from mixed years, so treat these as rough.
Names. I took the 43,274 populated places in OS Open Names (cities, towns, villages, hamlets, suburban areas and other settlements). A name is a ford if its last main word ends in -ford, -fords or -forth (the northern spelling, as in Garforth); a bridge if it ends in -bridge, -bridges, -brig or -brigg. Words after "on", "upon", "in", "le" and similar are ignored, so Bradford on Avon is a ford. "Bridge of" and "Brig o'" names are bridges. In Wales, names that begin Pont or Pen-y-bont are bridges (86) and names that begin Rhyd are fords (35); Gaelic names beginning Àth are fords. A short list of names with the crossing word first (Bridgend, Brigg, Brighouse, Fordham, Fordwich) is included. Where the last word decides, Stamford Bridge and Fordingbridge are bridges.
Exceptions. I keep a list of 13 checked names, each with a source: Telford (named after the engineer in 1968), Gifford (a family name) and Milford Haven (Old Norse fjord) are taken out; Pontefract (Latin "broken bridge"), Brigg and Brighouse are added; Cambridge stays a bridge (Old English Grantebrycge, "the bridge over the Granta"). Bridgwater, Bridgnorth, Seaforth, Forth and two Welsh names whose Welsh form contains ffordd ("road") are flagged as uncertain (6 names); dropping them changes nothing. I did not check every one of the 1,129 plus 643 names by hand, and a few will be wrong.
Size. Each place is matched to the ONS Built-up Area 2022 of the same name whose centre is within 3 km (or 2.5 times its radius, for big towns), one place per built-up area. Its population is the sum of the Census 2021 residents of the small areas (LSOAs) whose best-fit built-up area it is. Very small built-up areas have no small area of their own and get no figure.
Wealth. ONS income estimates for small areas, financial year ending 2023: mean total annual household income for each MSOA. A town gets the population-weighted average of the MSOAs whose best-fit built-up area it is; any other place gets the MSOA whose population-weighted centre is nearest. "Share of income" weights each resident by their area's mean household income, which assumes similar household sizes. The models use log income with region (the nine English regions and Wales) and size band as controls.
History. Domesday places come from the Hull Domesday Project data via the realm-domesday repository; a place counts as recorded if a Domesday place within 3 km has the same name (Domesday's editorial brackets removed) or the same nearest modern name. Bridges come from Wikidata (class "bridge" and its kinds) and OpenStreetMap; only Wikidata start dates are used for the medieval test. Roman roads are the Itiner-e atlas (2025), with the datum correction from my Roman roads project. Rivers are OS Open Rivers.
Limits. This is correlation, not causation. A name records a crossing at the time it was coined, which may be long before or after the place grew. Wikidata's bridge dates cover a small and famous minority. Income is for households, not people, and for neighbourhoods, not for the exact place; a hamlet takes the income of the area around it. London is out of the totals. Scotland and Northern Ireland have no comparable income data here.