Last week I wrote about my big data map of Roman Urban culture, for which I used architecture and inscriptions, and which, I argued, returned an image of the Roman world that was, to some extent, rooted in historical reality – there are biases and distortions, but the overall picture seems reliable to me to the extent that regions where a lot of places have returned relevant evidence in antiquity also were rather densely urbanized and relatively wealthy, whereas the regions that are empty on the map – with one notable exception – also appear to have been relatively empty in antiquity. The exception, as ever, is Egypt.
By relatively empty I do not necessarily mean desolate or poor. Take, for example, Tripolitania, which is almost completely empty on the map (see below), whereas it is well known to have been a major exporter of agricultural products, particularly olive oil and grain – the remains of farmsteads and ocasionally very large oil press sites are well known. Yet for some reason, this expressed itself in only three urban centres: Sabratha, Oea, and Lepcis Magna. The same is true for Cyrenaica, which only developed a few cities. Both regions have been intensively explored, but they markedly contrast with Africa Proconsularis and Numidia further to the west, where the concentration of cities is incomparably higher. Apparently, something prevented cities from developing in great numbers in Tripolitania and Cyrenaica.
It makes sense to briefly look at how the evidence was spread over the Roman world. Looking at the overall map, there are six clearly defined concentrations of evidence in the Roman world. Four large clusters immediately catch the eye: Africa, Italy, Greece and South-Western Asia Minor. Smaller clusters can be identified in southern Spain, and in the Roman Near East, around the Syrian Decapolis. These clusters are real in the sense that these six regions behave measurably differently from the rest of the Roman world. Generously counted, these clusters comprise around 10% of the total surface of the Roman world, but they include almost half (48%) of all places where relevant evidence has been found, and an even higher proportion of the actual evidence (though the latter may be a less reliable figure). Nevertheless, some well-known and well-preserved Roman cities do not belong to one of the clusters. For instance, this is true for Tarraco, Augusta Emerita (Merida) and Volubilis.
This evokes a number of questions, the first of which is what this actually means: does this mean that these regions really were wealthier and more densely populated than the rest of the Roman Empire, or does it mean that, for some reason, they simply developed a denser network of (smaller) cities? Both may be true, and it necessitates a closer look at individual regions to decide. There certainly were regions that were mostly empty – the western part of Africa seems an example, as is true for what is now South-West France and North-West Spain – and, indeed, Eastern Anatolia.
Other regions may have been rather densely settled, but simply had relatively few cities. Cisalpine Gaul – present-day North Italy – is a key example. Here, cities cluster along the roads that ran close to the hills, leaving the central part of the Po valley mostly empty, though we know it was intensively cultivated (and centuriated) by the Romans. To some extent, a similar pattern can be found in Baetica in southern Spain, where the settlements are equally clustered along two lines indicating the lower end of the mountain ranges surrounding the Guadalquivir basin. Compared to Cisalpine Gaul, however, the landscape is less flat, and much more fragmented, which may have led to the emergence of more, but smaller cities, and thus, to a denser pattern on the map. To some extent, therefore, the map defies a simplistic reading.
Still, it is clear that the regions where higher concentrations of evidence can be found – including the six clusters, but also regions like Cisalpine Gaul, Cilicia in eastern Turkey and the lower Rhone Valley – share some important characteristics that set them apart from most regions where less evidence has been found. They all include decent amounts of fertile land not only able to support a reasonable local population but also producing considerable amounts of marketable surpluses. Moreover, they all have a rather good internal viability by road and, sometimes, by river. All six clusters emerged around fertile coastal plains and, particularly, river valleys, where transport is relatively cheap and uncomplicated. Furthermore, though these clusters by no means were concentrated on the coast, they all were extremely well-connected to the Mediterranean networks of trade and transport.
In other words, the picture suggests that both agriculture and accessibility played a role in urbanization and prosperity. This is to some extent confirmed if you look at this map from a different angle. While almost all major rivers that mouthed in the Mediterranean developed higher concentrations of evidence (notable exceptions: Nile and Ebro), the reverse is true for almost all rivers that mouthed in the Atlantic (not counting the Guadalquivir) or even in the Black Sea – even if some of these river systems watered large amounts of rather fertile land. In the Roman world, fertile land alone was not enough – it was access to the wealthy Mediterranean consumer markets that counted.