A network analysis of 299,171 km of Roman roads finds Antioch, Ancyra, Byzantium and Corinth were better-placed overland crossroads than Rome, while 35 of 45 provincial capitals ranked in their province’s top tenth on at least one centrality measure. By sea, the picture differs.
A team led by Adam Pažout of the Universitat Autònoma de Barcelona and Tom Brughmans of Aarhus University modelled the Roman road network as it stood around AD 150, using the Itiner-e atlas of 299,171 km of road.12 The study was published in Nature Communications on 15 September.3 The roads were reduced to 10,516 junctions and 14,901 links, and travel time on each link was estimated from its slope using Tobler’s hiking function, a standard walking-speed formula.1
The team then measured betweenness: how many of the quickest routes between all other points pass through a given road or place. Across the whole empire, the heaviest path ran along the Mediterranean coasts, through Lepcis Magna, Alexandria, Antioch, Byzantium and on through the Balkans towards Gaul.1 The authors conclude that Antioch, Ancyra (Ankara), Byzantium (Istanbul) and Corinth were better located than Rome as intermediaries for overland movement. Rome was central within the Italian peninsula, but the peninsula hangs off the main path, and Rome was not on Italy’s highest-betweenness road.1
Byzantium’s position rests partly on a modelling choice. The network joins Europe to Asia only through two ferry links the authors added, at the Bosporus and the Dardanelles, which makes the city the land bridge between the two halves.1
The firmest measured result concerns the empire’s 45 provincial capitals. Scored against each province’s own network, 35 had at least one of three measures in the top tenth: betweenness, closeness (how quickly a place is reached from everywhere else) or degree (how many roads meet there). Fourteen were in the top fifth on all three.1
By construction only a tenth of positions can sit in any one measure’s top tenth, so a location with no advantage would clear that bar on at least one of three measures at most 30% of the time. Among capitals the figure was 78%. That arithmetic is Noodle News’s; the authors report the counts.
Of the ten capitals that missed the top tenth on every measure, the authors identify eight as major river or sea ports, among them Carthage, Alexandria and Lugdunum (Lyon), whose links ran by water.1 They place Rome among the capitals well located for both water and land travel. Estimated city population, by contrast, generally did not track road centrality.1
Rome’s exact rank. The paper’s main text gives no numeric scores for Rome or the four cities it names; city-level empire-wide scores sit in a supplementary dataset that Noodle News could not retrieve. This piece reports the comparison as the authors state it.1
Where the roads ran. Only 2.7% of the mapped length is located with certainty; 89.8% is conjectured, placed with lower spatial accuracy.2 The authors report that their road-level betweenness results held when 20% of uncertain segments were removed at random.1
When. Every road and town is treated as in use at once. Not all existed together, so the model best reflects the 2nd to 4th centuries AD.1