Ancient intermodal hubs & networks

According to Wikipedia, a hub is the central part of a wheel that connects the axle to the wheel itself. Many expressions use the term for a literal or figurative central structure connecting to a periphery. A transport hub is a place where cargo is exchanged between transport modes. With the growth of containerisation, intermodal freight transport has become more efficient.

Today, there are several major nodal points for maritime traffic which are related to the network of main streams of traffic:

  • consumer goods transported in containers from China, Korea and Japan to Europe via the Suez Canal and to the US west coast via the Pacific Ocean;
  • energy such as oil, Liquid Natural Gas (LNG) transported in bulk from the Middle East, to China, Korea and Japan and many other countries;
  • other raw materials such as coal and iron ore are also transported in bulk from Africa, Australia and South America to many countries.

The major nodal points, now called “hubs”, are therefore located in Europe (Rotterdam, Hamburg), in USA (Los Angeles, San Francisco, New York, New Orleans), in Asia, (Shanghai, Hong Kong, Busan, Yokahama, Singapore)[1].

Alexandria was the “greatest emporium of the world”, acc. to Strabo[2]: Goods were imported from, and exported to, South Arabia, East Africa and India[3], and paid for with gold and silver bullion; they were taxed at 25% by the Roman state, thereby providing a substantial part of its total income:

  • Some goods, such as perfumes and dyed silk, were transformed and manufactured in Alexandria, thereby adding great value to the imported goods;
  • Goods were exported to Rome and other cities of the empire: not only exotic spices and goods from beyond the Red Sea, but also vast quantities of grain produced in Egypt.

Alexandria was a hub of the Roman economy. Additional nodes of a large-mesh Roman trade network might be located at Gades (Baetica, for garum, salted fish, olive oil) and at Carthago (Proconsular Africa, for wheat and olive oil). This coarse network shows 3 lines converging on Rome. The question is whether finer-mesh networks might be added to the coarse one by including nodal points of smaller importance[4].

Let’s elaborate on this with an analysis of our database on ancient ports: we know of around 4200 ancient coastal settlements, out of which around 1500 are explicitly mentioned as ports by ancient authors:

Ancient ports mentioned explicitly by at least one ancient author.
This map has no pretention of accuracy; it just intends to show concentrations of ports; more accurate locations are available on Google Earth maps shown on:

A corpus of 75 ancient authors from 500 BC to 500 AD (with a few exceptions like Homer, Procopius) has been analysed, searching for the word ‘port’ in the 19th c. French translations available on the web (mainly Each author is counted only once for each port, even if he mentioned the port several times in several books or chapters.

Obviously, various reasons motivated ancient authors to mention these ports: historical (military, naval), commercial (trade, emporia), geographical (description of land and peoples) or sailors following the coasts. In the picture above, trips like those of Arrian on the Black Sea or the Stadiasmus can nearly be distinguished.
Furthermore, ancient authors may sometimes have been somewhat egocentric when describing only their own part of the world, like Pausanias in Greece, which may have led to ‘zooming’ effects in some areas.
Conversely, some areas were not much mentioned by ancient authors, like Hispania, Lusitania, Gaul and it cannot be said if that is because there were no ports (which is surely untrue) or because these somewhat remote areas were of lesser interest to ancient Greek and Roman authors.
Anyway, a concentration of ports mentioned by ancient autghors can be seen around the Aegean Sea.

Nearly 1000 ports are mentioned by only one ancient author,

Nearly 300 are mentioned by two ancient authors,

81 are mentioned by five or more ancient authors. These places are listed below in a clockwise ranking around the Mediterranean, with the number of authors mentioning it in brackets:

  • Hibernia (Isle of Ireland) (5)
  • Gades (5)
  • Carthago Nova (5)
  • Massalia (5)
  • Monoeci (Monaco) (5)
  • Portus Pisanus (6)
  • Aithalia (Isle of Elba) (5)
  • Portus Augusti Ostiensis (over 10) and Ostia (7)
  • Antium (5)
  • Caiete (5)
  • Misenum (6) and Puteoli (5)
  • Rhegium (7)
  • Zankle (Messina) (9)
  • Syracuse (over 10)
  • Crotone (5)
  • Lilybaion (5)
  • Tarentum (7)
  • Hydruntum (Otrante) (6)
  • Brindes (9)
  • Corcyra (8) and Casiope (5) (Isle of Corfu)
  • Glykys Limen (5)
  • Nisea (5)
  • Kytlene (5)
  • Pylos (5)
  • Gytheion (7)
  • Skandeia (Isle of Kythera) (7)
  • Nauplia-Argos (5)
  • Lechaion-Corinth (10) and Sicyon (8)
  • Kenchreai-Corinth (over 10)
  • Salamis (Isle of Salamis) (6)
  • Piraeus (over 10)
  • Phaleron (7) and Munychia (5)
  • Aegina (6)
  • Aulis (6), Chalkis (5) and Eretria (5) (Isle of Evia)
  • Thasos (Isle of Thasos) (5)
  • Abydos (10) and Sestos (6)
  • Byzantium (6)
  • Portus Symbolorum (Crimea) (5)
  • Sindicos (Anapa) (5)
  • Sinop (6) and Armene (5)
  • Calpe (5)
  • Cyzikos (5)
  • Sigeion (5)
  • Delos (Isle of Delos) (10)
  • Naxos (Isle of Naxos) (5)
  • Tenedos (Isle of Tenedos) (8) and Troy (6)
  • Mytilene (Isle of Lesbos) (over 10)
  • Phokeia (6)
  • Elaia (5)
  • Chios (Isle of Chios) (over 10)
  • Ephesus (10)
  • Pythagoreion (isle of Samos) (over 10)
  • Miletos (9)
  • Kos (Isle of Kos) (5)
  • Knidos (7)
  • Rhodes (over 10)
  • Kaunos (5)
  • Patara (8)
  • Korikos (Kizkalesi) (5)
  • Phaselis (5)
  • Paphos (Isle of Cyprus) (5)
  • Salamis (Isle ofCyprus) (5)
  • Sidon (6)
  • Tyr (6)
  • Alexandria (over 10)
  • Paretonius (5),
  • Menelaus (5),
  • Neapolis-Leptis (5)
  • Cercenna (6)
  • Carthago (8) and Utica (5)
  • Melite (Isle of Malta) (6)

The listed places are shown on the map below as ‘local hubs’ together with the four ‘main hubs’. Local hubs are fairly concentrated in an area between Rome and Rhodes covering the southern part of Italy, Greece, the Aegean Sea and Asia Minor. It cannot be denied that this area was the most active area both for trade and for naval operations during a millennium from the 5th c. BC to the 5th c. AD.

Note that no time frame was defined, hence Greek hubs of the 5th c. BC are mixed with imperial Roman hubs. Had we restricted the time frame to e.g. the 6th to 4th c. BC, we would have seen Piraeus (over 10), Emporion (Spain) (4), and Naucratis (Egypt) (1) as main hubs. Had we taken the 3rd and 2nd c. BC, we would have mentioned Delos (10).

Proposal for hubs in the trade networks in the Roman Mediterranean Sea: Black dots are main hubs: Rome, Alexandria, Carthage, Gades; Red dots are regional hubs; Green dots are local hubs.

Based on ancient sources and modern archaeological research, a series of ‘regional hubs’ has been defined in a more intuitive way, including Carthago Nova (5), Tarraco (3), Narbo (3), Arelate (3), Puteoli (5), Aquileia (3), Athens (over 10), Byzantium (6), the Tanaïs river area (2), Nicomedia (1)[5],  Ephesus (10), Rhodes (over 10), Attaleia (1)[6], Antioch ad Orontem/Seleucia Pieria (3) and perhaps Gaza (4), if it was more than a place of transit such as Myos Hormos (4) and Berenike (4).

In addition to Indian places such as Muziris (Pattanam, north of Cochin), a lesser known place such as Omana (possibly located at al-Dur, ed-Dur, in Umm al-Quwain Emirate) should be mentioned here too in order not to under-estimate ancient traffic in the Gulf.

This pattern of imbricated networks could be refined almost indefinitely as each local hub may have its own trade with its hinterland and other nearby small ports. Like a fractal that exhibits a repeating pattern displayed at every scale.

Appendix: Detailed results of the database analysis



[1] DUCRUET, C., (2015), “Inside the pond: an analysis of Northeast Asia’s long-term maritime dynamics”, International Journal of Maritime Affairs and Fisheries, Korea Maritime Institute, 2015, 7 (2), pp.25-40.

[2] STRABO, Geogr. 17, 1, 13

[3] ARNAUD, P., (2015c), “ La batellerie de fret nilotique d’après la documentation papyrologique (300 avant J.-C.-400 après J.-C.) ”, in La batellerie égyptienne, Archéologie, histoire, ethnographie, éd. P. Pomey, Centre d’Etudes Alexandrines, 34 – 2015.: Kerkouros-type ships were sailing and rowing southward on the Nile in winter time, at least during the Hellenistic period.

[4] BRAUDEL, F., (1949), “La Méditerranée et le monde méditerranéen à l’époque de Philippe II”, éd. Armand Colin, Paris, (533 p) : he distinguishes various basins: “La Méditerranée n’est pas une mer, mais une succession de plaines liquides communiquant entre elles par des portes plus ou moins larges.” Each basin is the result of human cultures superimposed upon physical constraints, with continuous changes always going on. See also ARNAUD, P., (2005), “Les routes de la navigation antique”, éd. Errance, Paris, (248 p).

[5] Byzantion and Nicomedia were both ancient Greek cities, but they were on each side of the Bosphorus, on different continents. Nicomedia was a major Roman city in the 2nd and 3rd c. AD, while Byzantium was reconstructing after Septimus Severus’ destructions in 195 AD and finally heading for becoming a capital city when renamed Constantinopolis in 330 AD.

[6] Pergé was part of the Roman Empire since 188 BC and was the capital city of Pamphylia. It had its own river port some 16 km from the sea, but the seaport of Attaleia could be used when the coast was free of pirates.