Roads to RomeENRO
Every mapped Roman road as a line that thickens towards Rome, on a relief map of the Mediterranean.

Drawing 299,171 km of roads…

  1. 299,171 km of Roman road, on a map published in 2025

    Do all roads lead to Rome? The proverb, tested on every mapped Roman road

    Walk out of Rome along every road at once, 30 km a day. After 87 days the land runs out, with 57% of the roads behind you. The rest lies across water: most of it across 4.4 km.

  2. Day 0 to day 20

    Leave Rome by every road at once#

    A line lights up when a walker from Rome would have reached it at 30 km a day, close to the 20 Roman miles that Vegetius sets for a summer march at the military step.

    Every line is a road as 20 researchers drew it for Itiner-e, a map of the empire’s roads published in 2025. A line is thicker where more road lies beyond it: its width counts the kilometres whose shortest way to Rome passes there. The walk is a calculation run over all the roads at once, not anyone’s journey.

    On day 20 the walk reaches Milan, 602 km out. It reached Brindisi, on the Adriatic, on day 18.

  3. Day 20 to day 50

    Over the Alps, to the Rhine and the Danube#

    Lyon on day 32, Paris on day 46, Cologne on day 48. Past Aquileia the roads turn east along the Danube and the Sava.

    The straight-line distance from Rome to Cologne is 1,091 km. The road takes 1,443.

  4. Day 50 to day 87

    Where a walker runs out of land#

    57%of mapped Roman road can be walked to Rome without crossing water

    Porolissum, the northern end of Roman Dacia, on day 61. Tomis on the Black Sea on day 74. Byzantium on day 77. Cádiz on day 81.

    The last road that can be reached without crossing water ends near Cape Malea, at the tip of the Peloponnese, on day 87: 2,599 km from Rome.

    By then 171,539 km are lit. That is 57% of all the Roman road on the map, and every kilometre of it is in Europe.

  5. Back to day 77

    Asia and Africa hang on 4.4 km of water#

    Ten days before the land ran out, the walk had reached Byzantium. There the road stops at the Bosphorus. The map carries on across it by a white line marked “Byzantium ferry”, 4.4 km long. A second one crosses the Dardanelles at Abydos. Either alone joins the continents; without both, Asia and Africa are a separate piece.

    They are not roads. The map’s authors call them “two artificial road connections … at known ancient ferries”, added so that the continents could be studied as one network.

    Everything drawn in orange from here on is joined to Rome only through those two lines: 109,893 km of road, 37% of the total.

  6. Day 87 to day 150

    Across Anatolia and down the Levant#

    Ankara, then Antioch on day 112, Jerusalem on day 133, Alexandria on day 150.

    The walker has now walked for five months and 4,506 km. In a straight line Alexandria is 1,952 km from Rome.

  7. Day 150 to day 296

    Round the whole sea, back to within sight of Spain#

    8,872 kmby road from Rome to Ceuta, the farthest point of the network

    Cyrene on day 179, Leptis Magna on day 210, Carthage on day 239.

    Carthage is 590 km from Rome in a straight line and 7,155 km by road, 12 times as far. The 7,155 km measure the road network, not anyone’s journey: a map of roads alone leaves out the sea that held this empire together. Stanford’s ORBIS models the roads together with sea lanes and rivers.

    The road ends on day 296 at Ceuta, 8,872 km from Rome and 28 km across the strait from the Rock of Gibraltar, which the walk passed on day 79.

    The last stretch is the weakest line on the map: between Melilla and Tangier the coast road is marked “hypothetical”. Leave such lines out and the walk ends near Melilla on day 280, with Roman Morocco one more island. The Barrington Atlas says that east of Volubilis land communication with the neighbouring province “always remained uncertain and probably sporadic”.

  8. What never lights

    The roads that no road joins to Rome#

    17,731 km of road, 5.9% of the total, are on islands: Britain with 11,585 km, then Sicily (1,802), Sardinia (1,647), Cyprus (1,368), Crete (819) and four smaller pieces.

    Britain’s roads end 37 km from the nearest road in Gaul. Sicily’s end 3.3 km from the toe of Italy.

  9. The proverb, measured

    Do all roads lead to Rome?#

    57% · 37% · 5.9%on foot · behind a ferry · on islands

    The saying is medieval. Its earliest known written form, from about 1175, is Alain de Lille’s: “mille viae ducunt homines per saecula Romam”, a thousand roads lead men through the ages to Rome. Tested on the ancient roads:

    57% of them do, if the traveller refuses every boat.

    94% do, if one ferry of 4.4 km is allowed (the map has two).

    The remaining 5.9%, nearly all of it on islands, need a ship. By length of road, a little over half of the empire on this map can be reached from Rome on foot.

  10. The catch

    The same roads lead to Byzantium too#

    Roads run both ways. Whatever can reach Rome can reach any other town on the same network, so a walk that starts at Byzantium, Antioch, Carthage or Milan lights exactly the same 94% of the roads. Here they are drawn towards Byzantium.

    What a town’s place in the network changes is how far away the roads are:

    StartRoads reachedMean distance to the roadsFarthest
    Rome94.1%2,781 km8,872 km
    Milan94.1%2,487 km8,643 km
    Byzantium94.1%2,395 km6,576 km
    Antioch94.1%2,778 km5,502 km
    Carthage94.1%5,446 km9,095 km

    The mean is taken over every kilometre of joined road, not over junctions. By that measure 43% of the junctions joined to Rome, carrying 43% of the road, are closer to the rest of the network than Rome is. The junction closest of all, the diamond, lies in the Morava valley in Serbia, 37 km from Niš, the Roman Naissus: 2,297 km on average.

    So reaching Rome from every road cannot show that Rome was the centre. What the roads can show is how far the empire lay from Rome, and by which road it came in.

  11. Arriving

    Almost everything enters Rome from the north#

    Each road has the colour of the road by which its shortest way enters Rome. Of all the road joined to the city, 97% arrives from the north: 62% by the Via Flaminia and 35% by the road from Veii, which the map labels Via Triumphalis. Put the edge of the city at 1 km and the two count as one road, the Flaminia, with 97%. These are shares of road length, each kilometre counted along its shortest way to Rome; they say nothing about travellers or goods.

    The Via Labicana brings in 2.3%: most of southern Italy. The Appian Way, the most famous road of all, brings in 0.2%, or 569 km; on this map the inland road is the shorter way from the south.

  12. Why

    South of Rome the roads end at the sea#

    South of the city lie the heel and the toe of Italy, and then the sea, where the roads stop. A walker can go no further, so by land the East, Egypt and Africa come in by the Flaminia: through the Balkans and down the Adriatic.

    Gaul and Spain arrive by the road from Veii; the Danube, the Balkans and everything beyond them by the Flaminia; the roads of the Rhine are divided between the two.

  13. What the map’s authors found

    Rome is the end of the roads, not their crossroads#

    In September 2026 the team behind the map asked a different question: which roads lie on the quickest way between any two places? Their answer: “major cities like Antioch, Ancyra, Byzantium and Corinth are better located than Rome as intermediaries”. The Italian peninsula, they write, is “itself an appendix to the highest betweenness path”, that is, to the route most of those quickest ways pass along.

    The two results do not contradict each other; they differ in two ways. Their measure counts the ways between all pairs of places and weighs them by walking time, slower on slopes. The walk from Rome counts only the ways to Rome, in kilometres. A city at the end of a peninsula can be reached from every road and still lie on the way between few other places.

  14. How sure the lines are

    For nine kilometres in ten the road is known and its line is approximate#

    Only 2.7% of the length is “certain”: well documented and drawn to within 50 m in the mountains and 200 m on the plains. 89.8% is “conjectured”: an identified road drawn with less accuracy, in many cases along a modern road. 7.4% is “hypothetical”, dashed here: a road identified but not located.

    Nor were all these roads in use at one time. The map gathers roads used at some point in the Roman period, to about AD 400, across the empire at its widest, around AD 150.

    Leaving out every hypothetical line changes the answer little: 93.3% of what remains is still joined to Rome.

  15. A test from antiquity

    The Romans wrote their own distances down#

    The Peutinger map is a medieval copy of a Roman road map, with a distance in miles written along most legs between two stations. Its western end is lost; there the figures come from the Antonine Itinerary, a Roman list of routes. Between pairs of towns drawn with a symbol there are 356 stretches where every leg has a figure and the modern map has a road to measure: 34,800 km in all.

    On 113 of them (32%), in green, the written miles and the measured road differ by a tenth at most. Shuffle the figures among all the stretches and 7% still agree; shuffle them only among stretches of similar length and 23% do. The gap is nine percentage points. None of 20,000 shuffles among stretches of similar length produced an agreement rate this high.

    The list does better than the map: 22 of 44 stretches agree in the west (50%), 91 of 312 on the Peutinger map itself (29%). The trouble may lie more with the figures copied onto the map than with the modern lines, or the roads of the west may simply be better mapped.

    Of the 199 stretches that fit neither miles nor leagues, the written figure is the smaller one on 70%. On 40 it is smaller than the straight line between the two towns, which no road can be: a figure was lost or miscopied, or a place is put in the wrong spot today.

  16. A unit shows up on its own

    North of Lyon the figures fit as leagues far more often than as miles#

    Blue marks the stretches where the written figure matches the road if each unit is 2.22 km, the Gallic league of one and a half Roman miles. Inside Gaul north of Lyon, 15 of 33 stretches fit as leagues and 2 as miles. Elsewhere 9% fit as leagues.

    The map says so itself. Beside Lyon it carries the words “Lugduno caput Galliarum usque hic legas”, which Richard Talbert renders as “Lugdunum, capital of the Gallic [provinces]; up to here, leagues.”

    On the 674 single legs where the written miles and the measured road roughly agree, the measured roads are on average 1.7% longer than the written miles.

  17. Britain

    11,585 km of road, and none of it leads to Rome#

    Britain is the largest piece of the network that stands apart: 3.9% of all Roman road. Its roads end on the Kent coast, 37 km from the nearest road in Gaul.

    Across that gap, Roman London is 1,664 km of road from Rome, about 55 days on foot. York is 1,970 km and Carlisle, at the western end of Hadrian’s Wall, 2,148 km (from the ancient sites; the search box measures from the point that marks today’s town).

    The Antonine Itinerary, a Roman list of routes, opens its British section with the crossing from Boulogne to Richborough and gives it as 450 stadia.

  18. The road from your town to Rome

    Type a town between Scotland and the Nile. The map finds the nearest mapped Roman road and follows it to Rome.

    No town chosen yet.

    Start the walk from

    Colour the roads by

    Drag the map with the mouse, or move it with these:

How it was measured

The roads are the 14,769 segments of the Itiner-e map, 299,171 km in all, as published on Zenodo in September 2025. The two ferry lines are the last two rows of that file, named “Byzantium ferry” and “Abydos ferry” and described as “temporary”, so the published total includes their 8.8 km; the 14 pieces reported with the dataset are those of the file without them. Two segment ends within 30 m of each other count as one junction; the result is the same at half a metre and at 100 m. That gives 13 separate pieces, the number the map’s authors report: the one that holds Rome, nine islands and three fragments under a kilometre long in the Netherlands. Without the two ferries there are 14; either ferry alone is enough to join Europe to Asia. The lengths add up, to within a kilometre of rounding: 171,539 km on foot, 109,893 behind the ferries, 17,731 on islands and in the three fragments, and 8.8 km for the two ferry lines themselves.

Rome is every junction within 3 km of the Forum: a convention that puts the start near the edge of the ancient city, not a historical boundary. With 1 km or 5 km the farthest point is the same; with 1 km the two northern roads count as one. Without the hypothetical lines the Flaminia brings in 60% and the road from Veii 37%. The distance from any point is the shortest path over mapped segments. A segment’s width is the length of road whose shortest path to Rome runs over it. A town’s mean distance to the roads is the average, over every kilometre of road joined to it, of the shortest road distance to that town; it does not depend on how densely the junctions were drawn.

The walker covers 30 km every day and never rests; day counts are rounded to the nearest day. Vegetius sets the military step at 20 Roman miles, 29.6 km, in five summer hours.

The islands

For a town on an island, the island’s roads are linked, by the shortest straight line over water, to the nearest road that already reaches Rome; the length of that line is stated but not counted in the kilometres of road or the days on foot. The rule was chosen because it can be applied everywhere. It is not a record of where ships sailed.

The Roman test

The figures come from OmnesViae, which follows Talbert’s edition of the Peutinger map, takes Britain, Spain and Morocco from the Antonine Itinerary, and places most stations after the Pleiades gazetteer. A stretch runs between two towns drawn with a symbol, through stations in single file, every leg with a figure, none over water, both ends within 1.5 km of a mapped road. The measured length is the shortest mapped path between the two ends. “Agrees” means within 10% of 1.4815 km per mile, or of 2.222 km per league. The tolerance and the box that stands for Gaul were fixed before the count; Gaul is a rectangle north of Lyon (5° W to 8.3° E, 45.6° to 51.6° N) within today’s France, Belgium, Luxembourg, Germany, the Netherlands and Switzerland.

At other tolerances: within 5%, 16% of stretches agree, against 3% when all figures are shuffled and 12% when they are shuffled among stretches of similar length (eight groups by measured length; 95 shuffles in 100 give 14% or less). Within 10%: 32%, 7% and 23% (26%). Within 20%: 51%, 14% and 41% (44%). Of 20,000 length-matched shuffles, 169 reached the observed share at 5%, 0 at 10% and 0 at 20%. 8 pairs of towns appear twice, where the source gives two routes with different figures; counting only the first route of each pair, 107 of 348 agree. Italy and Turkey supply a third of the stretches. “Roughly agree”, for the single legs, means within 20%.

The test bears on three things at once: the copied figure, the identification of the town, and the modern line. It cannot say which of them fails on a given stretch; a figure smaller than the straight line clears only the modern line.

Distances in the chapters are measured from the ancient site (Londinium, Napoca, Tomis); a town typed into the search box is measured from the point GeoNames gives for today’s town. The two can differ by a few kilometres.

Where the map stops

A road that has not been found on the ground or inferred from other sources is not on the map, so “no road” here means “no mapped road”. The authors left out Armenia and Mesopotamia.

Neither the sea nor the rivers are on this map, though Romans travelled and shipped goods on both. The 7,155 km to Carthage measure the road network, not anyone’s journey.

Towns are today’s, placed by GeoNames at one point each. A town “stands on a Roman road” here when that point is within 2 km of a mapped line.

The 356 stretches, written against measured

Written figures (miles, or leagues in Gaul) and place names as transcribed by OmnesViae from the Peutinger map and, in the west, the Antonine Itinerary; measured lengths on the Itiner-e map. Longest first.

Written against measured
Stretch on the map (today)WrittenMeasured, in milesMeasured, kmVerdict
Tarracone – Carthago Partaria (Tarragona – Cartagena)360348.5516agrees in miles
Camarata – Sufasar315284.5422agrees in miles
Actanicopoli – Megara (Nicopolis – Megara)168255.7379fits only as leagues, outside Gaul
Ocelodorum – Emerita (Zamora – Mérida)225233.1345agrees in miles
Mastrvm – Sinope (Amasra – Sinop)69212.1314written figure below the straight line
Comana Capadocia – Samosata (Şar – Samsat)60211.5313written figure below the straight line
Aquis Terebellicis – Virovesca (Dax – Briviesca)208209.7311agrees in miles
Augusta Bracaria – Asturica (Braga – Astorga)247207.8308written figure larger
Agripina – Dvrocortoro (Köln – Reims)121204.4303written figure below the straight line
Chidvm – Patras (Reşadiye – Gelemiş Kalkan)147192.5285written figure smaller
Emerita – Toleton (Mérida – Toledo)111188.3279written figure smaller
Miletvm – Chidvm (Balat – Reşadiye)83178.1264written figure smaller
Clampeia – Salerno (Amantea – Salerno)110177.1262fits only as leagues, outside Gaul
Actanicopoli – Avlona (Nicopolis – Vlorë)171173.7257agrees in miles
Aqvas labodes – Siracvsis (Sciacca – Siracusa)150173.6257written figure smaller
Sertica – Stopis (Sofia – Palikura Negotino)189173.3257agrees in miles
Sinope – Amasia (Sinop – Amasya)149173.2257written figure smaller
Petavione – Mvrsa maior (Ptuj – Osijek)156173.1256written figure smaller
Actanicopoli – Larissa (Nicopolis – Larisa)70171.1254written figure below the straight line
Caesaraugusta – Virovesca (Zaragoza – Briviesca)161169.2251agrees in miles
Segodvni – Aqvis Segete (Rodez – Moingt)107163.1242fits only as leagues, outside Gaul
Antiochia pisidia – Sidi (Yalvaç – Side Manavgat)80157.9234written figure smaller
Thermis – Messana (Termini Imerese – Messina)131157.4233written figure smaller
Polemonio – Trapezvnte (Bolaman – Trabzon)125154.4229written figure smaller
Ossonoba – Ebora (Faro – Évora)214146.5217written figure larger

Sources