My map of Roman roads said Carthage was 210 days from Rome. Readers rightly pointed out that nobody walked there: they took a ship. So here is the empire again, with the sea added, sailing speeds taken from ancient voyage reports, and the summer winds that made some passages fast and others painfully slow.
In the first part of this project I took Itiner-e, the new open map of the roads of the Roman Empire, and asked how many days it took to walk from Rome to anywhere. The answers were long: 81 days to London, 180 to Alexandria, and 210 to Carthage, because a road-only traveller to Carthage has to go through Spain and round the whole of North Africa. The roads were never meant to do that job alone. The Mediterranean was the empire's main highway, and for most long journeys a ship was quicker, cheaper and, in summer, safer than the road.
So this map adds the sea. It joins 211 ancient ports to the road network and links them across open water, with sailing speeds taken from Lionel Casson's classic collection of ancient voyage reports, and a simple model of the summer etesian winds, which carried ships quickly from Italy to Egypt and held them back on the way home. In winter, when the sea was largely closed, the map falls back to the roads.
Roads: Itiner-e static version 2024 (de Soto, Pažout, Brughmans et al. 2025), CC BY 4.0. Ports: Pleiades (CC BY 3.0) and the harbours of the ancient voyage reports. Sailing speeds: Casson (1951), calibrated to his Table 1. Click anywhere on land or sea for the fastest time and the route, split into road and sea legs. Times are modelled, not recorded.
With ships allowed, Carthage comes 36 times closer: 5.8 days in summer, a short road to Ostia, a day waiting for a ship and about five days at sea, against 210 days on foot round North Africa. Alexandria falls from 180 days to 12, Leptis Magna from 231 to 9, and Cádiz from 116 to 13.
Days from Rome: road only against road and sea
Grey dot: roads only (and winter). Coloured dot: fastest by road and sea in summer, with the range from the sensitivity tests. Switch how the land legs are travelled. Click a city to see its route.
The gains are smallest where the road was already direct. London falls from 81 days to 30, by ship from Ostia round Spain and up the Atlantic coast to the south coast of Britain. Constantinople falls from 75 to 17, and it is the most uncertain number on this page. The etesians blow down the Aegean from the north, straight into the bows of a ship bound for the Hellespont. Casson's own estimates put the voyage up from Alexandria at 17 to 20 days, and the model's ship, sailing at the speeds that fit his estimates best, spends 16 days at sea from Ostia. With the slower speeds of the alternative model in section 03 it would take 36. For an imperial courier, who could ride 80 km a day, the ship saves less: 16 days against 22 by road.
By road alone, Carthage was 210 days from Rome. With a ship and a fair wind it was 6.
In 1951 the historian Lionel Casson gathered every ancient report he could find of a voyage whose length and duration were both stated: Pliny's record passages from Ostia, Paul's journey to Puteoli in Acts, the letters of Synesius, the life of Porphyry of Gaza. With favourable winds his 24 voyages averaged 2.7 to 6.2 knots, with a median of 4.7; against the wind his 10 voyages made 1.5 to 3.3 knots, median 2.0. His summary: 4 to 6 knots over open water with the wind behind, less than 2 to 2½ against it.
The wind decides which of those applies. From about mid-May to mid-September the etesians blow down the Aegean from the north, and in the southern Aegean and around Crete from the north-west. Casson notes that in summer there was no direct sailing from Alexandria to Rome at all; his estimates put the outward voyage at 10 to 13 days and the return at 53 to 73. The grain ships went home the long way, north-east to Cyprus and west along the coast of Asia Minor and Crete.
1. In the Aegean the summer wind blows from the north, and in the central and eastern Mediterranean (south of 38.6° N and east of 15.5° E) from the north-west. A ship running straight downwind there makes 5.2 knots (231 km a day), and speed falls away as the heading turns into the wind, to 1.0 knots dead into it between Sicily and Crete and 1.6 in the Aegean and the Levant. 2. Where my sources name no prevailing summer wind (the western Mediterranean, the Adriatic, the Black Sea and the Atlantic), a ship makes 3.7 knots. 3. These speeds are calibrated to Casson's Table 1, his 57 estimated sailing times between named ports: his median speeds, scaled up by 1.11, with the two speeds into the wind chosen to fit the same table. All of them sit inside the ranges he gives: 4 to 6 knots with the wind, 3 to 4 along the coasts, less than 2 to 2½ against it. 4. Boarding a ship costs 1 day of waiting; landing costs nothing extra. 5. From November to March the sea was considered too dangerous and largely closed, so the winter map is roads only.
I also built a second, slower version, which I keep as a sensitivity test. It uses Casson's median speeds unchanged (4.7 knots downwind, 3.3 with no prevailing wind) and a single speed into the wind, 0.4 knots, fitted so that the voyage from Alexandria to Rome takes 63 days, the middle of his 53 to 73.
Casson's Table 1 gives his estimated sailing times for 57 pairs of ports, each a range of days, from Rome to Carthage (2 to 4) to Alexandria to Massilia (58 to 78). I asked both versions of the model, sea only, for the same passages. The calibrated model lands inside Casson's range for 14 of the 57 (25%) and within a quarter of it for 36 (63%); its median time is 1.01 times the middle of his range. The slower model lands inside for 12 (21%) and within a quarter for 25 (44%), and its median passage takes 1.38 times as long as his estimate. It is slow almost everywhere except on the one voyage it was fitted to.
That is why the map uses the calibrated model. Its one clear failure is the long voyage home from the east: from Alexandria or Rhodes back to Italy or Gaul it takes about half the time Casson estimates (the median of the 4 such passages is 0.46 times his figure; Alexandria to Rome takes 30 days against his 53 to 73). The slower model gets those right. But this map measures journeys from Rome, which run mostly with the wind, and on those the calibrated model is the better fit: Ostia to Alexandria takes 11.0 days, inside Casson's 10 to 13. Every city figure on this page also comes with the slower model's answer in its range.
On the 34 actual voyages in Casson's other tables, the calibrated model's median time is 1.01 times the reported one, and 68% of voyages are within half as much again either way; the slower model's median is 1.49 times. Ancient writers recorded fast passages, and the model gives every ship the same steady wind all summer, so a single voyage can easily be quicker or slower than the model.
The model against Casson
By road alone, only 3.9% of the land of the provinces was within 30 days on foot from Rome. Add the sea in summer and 59% is. Within a fortnight the share rises from 1.4% to 10.8%; within a week, from 0.5% to 1.2%. For a courier riding the roads and taking ship where it helped, 92% of the land was within a month.
Share of the empire's land within a given time of Rome
Provinces of about AD 117 (DARE), land within 50 km of a mapped road. Hover a bar.
Of the 919 major ancient places on the network, 94% are quicker to reach from Rome in summer with a ship than without. The biggest gains are along the coast of Africa, which the road network reaches only by going round: Carthage is 37 times closer, and Hadrumetum (Sousse), Hippo Regius (Annaba) and the other ports of Africa Proconsularis and Numidia 32 to 33 times. At the other end, 54 places gain nothing at all: they are in central Italy, all within 14 days of Rome by road, where no ship can beat the walk.
The places a ship brought closest
Major ancient places, by how many times quicker they are to reach in summer with the sea than by road alone (on foot). Click a row to see the route.
Roads. Exactly as in the roads study: the Itiner-e network of 14,769 sections, 22.5 km a day on foot, 45 by cart and 80 by courier (the middle of the ranges in Wikipedia's article on Roman roads), slowed on slopes by the dataset's passability field, with four short sea crossings at one day each.
Ports. Every Pleiades place typed as a port, harbour or anchorage with a location (135 kept), plus 76 harbours named in the ancient voyage reports and the grain trade (Ostia, Puteoli, Byzantium, Ephesus and others), each within 25 km of the sea grid and 15 km of a road.
Sea. A grid of 15 km cells over open water (Natural Earth land as the obstacle) covering the Mediterranean, the Black Sea and the Atlantic coast to Britain; ships move between neighbouring cells, and the Bosporus, the Dardanelles, the Kerch Strait and the Strait of Messina are joined by hand. The Red Sea and the rivers are left out: my sources give no sailing or towing speeds for them that I could check. The speeds and wind rules are in the box above; every one has been varied in the sensitivity tests, and the coloured ranges in the city chart show the spread.
What this cannot say. The model has one summer wind pattern, no calms, storms, currents or night stops, and treats every port as equally able to take you anywhere. Real ships were not timetabled: you waited for a ship going your way, and the ancient record is biased to fast and memorable voyages. The calibrated speeds make the voyage home from the east too quick, so any journey that ends in Italy from Egypt or Syria should be read with the slower model's figures in mind.
ORBIS. Stanford's ORBIS model of the Roman world (Scheidel and Meeks) was built for exactly this question and goes much further: monthly wind and wave conditions, two kinds of ship, hundreds of ports and sea lanes, rivers and a dozen kinds of road traveller. I have not used or copied its numbers. This map is simpler and fully open: one season's winds, speeds you can check against Casson's own estimates, and the far more detailed Itiner-e road network underneath. The broad story is the same as ORBIS tells: the sea shrank the empire, but unevenly, and the winds made Rome much closer to Egypt than Egypt was to Rome.