<--Retour à la liste des gagnants / Return to Essays 2000-2001

Castra et urbs romana:
An Examination of the Common Features of Roman Settlements in Italy and the Empire
and a System to aid in the Discovery of their Origins.  


Anders Bell
(Concordia University)

Archaeological evidence from cities throughout the Roman Empire has shown similarities between certain sites such as Exeter (Isca Dumnoniorum) in Britain and Timgad (Thamugadi) in North Africa.  Despite the distance separating them, both of these cities share a grid-like system of streets that cross each other at right angles. This orthogonal design betrays the military origins of these sites, based on the castra of the Roman army that will be discussed below.  Other Roman cities such as Pompeii on the Italian mainland and of course Rome itself, show a much less organised development and do not seem to have any form of strict urban planning. This is again due to the origins of these cities.  In this essay I will begin by focusing on the cities of Exeter, Timgad, and Pompeii.  I will discuss, through a process of examination and comparison, how these cities differ in their design and what the main causes of these differences are (i.e. the origins of their respective development and the systems that were used (if any) in their construction).  Secondly, I will discuss whether these three cities can be shown to be examples of four types of Roman settlement.  I will do this by comparing these models to several other Roman cities around the empire.  In conclusion, I hope to show that it is possible to learn both the origins and reasons for the original construction of many Roman cities and even the identity of their original inhabitants by examining the urban planning of their respective sites.

        Many Roman cities developed from the military castra that preceded them. These castra consisted of a rectilinear tract of land that was traversed by two main roads.  One ran north to south, the other east to west and they would cross each other at the centre.  The street running from north to south was labelled the cardo; that running from east to west was called the decumanus.  The rest of the streets were laid out in an orthogonal grid pattern that left insulae as starting points for buildings.  If this camp developed into a town, then the forum was placed on one of the corners of the intersection between the two main streets. This orthogonal design was most likely adopted from the Etruscan design used at Marzabotto, but this in turn may have been influenced by the Greek grid-pattern of Hippodamus (as it is of a different design that replaced the cardo with three cardines).  This is a reasonable idea as the Etruscans were in contact with the Greek colonies of southern Italy. However, the Romans continued the sacred Etruscan ritual used at the founding of cities, which certainly points to a great deal of infuence.  This ritual consisted of the choosing of a sacred spot where a priest would place a groma, which was then use to determine the path of the decumanus.  This was achieved by noting the place where the sun rose in order to know exactly where east was.  The cardo was then calculated running at a right angle from the decumanus at the groma.  Then streets were added (cardines and decumani) at an equal distance from the centre up to the boundary point of the settlement.  The settlementís four gates were then placed at the points where the main streets arrived at the boundary.  This was the basic structure of the castrae used by the Roman military.  As the Roman Empire grew it was used as the template for new cities throughout the colonies.  Its simple structure allowed it to be set up in every territory without difficulty.

        A good example of a city that developed from a military camp is Exeter in Britain.  This Roman camp was built after Flavius Vespasianusí military campaign in the west of Britain.  It was well situated on the east side of the River Exe, which provided it with natural fortification of both the river and the cliffs on either side.  As can be seen from the plan, the fortress was constructed with a cardo and decumanus with three cardines and two decumani.  The town that developed from the castra shows more sub-divisions with two cardines added, and also the addition of the forum and basilica at the centre of the town.  The town was most probably founded at the site at the time of the Roman armyís departure c. AD 80-85.  However, as can be seen, the town utilised the street layout of the castra within the confines of the old camp fortifications.  This limitation of size is contrasted by the city of Timgad in Algeria, where the conditions for the settlement and the reason for its construction were different from the situation at Exeter.

        The city of Timgad differs in that primarily it was not constructed as a military fortress.  The town was built around AD 100 as a veteran colony for ex-members of the Third Legion that had been stationed not far away in Lambeisis.  As can be seen from the plan the original part of the town (not including the suburban constructions to the north, west and south) resembles the rectilinear form of the castra.  The decumanus is visible, but the cardo appears to have been either displaced towards the western gate, or replaced with cardines.  Though this was not a fortress but a military colony, it still needed to be easily defended; therefore the form of the castra was the most suitable.  The city was constructed by the Third Legion, which also accounts for the military planning.  The sheer size of Timgad is the largest difference from the settlement at Exeter.  This may be evidence for a lack of native habitation in the area before the town was built.  At Exeter, there is some evidence for previous non-Roman settlement at the site, but it is not that of a town, and the tribal peoples of the area more probably had their habitation concentrated in the nearby hill-forts. Timgadís situation also explains its size.  It is set upon a desert plain, therefore without the natural features such as hills that could interfere with the military design of its layout.

        Timgad represents a fine example of a town that was built for one express purpose, that of a colonia.  It was built quickly with the style of a castra, but unlike towns that developed from an actual camp it was built with public amenities such as a forum and a theatre.  Despite the differences between these two sites, their similarities in origin dictate the form of their urban planning.  Their military background gives them both an organised and practical layout.  This is even more noticeable when these towns are compared to Roman cities such as Pompeii, whose origins have spawned a different kind of layout, seemingly without order.

        Pompeii dates to at least the sixth century BC (The date of the Doric Temple that preceded the Triangular Forum).  The settlement was originally Oscan, a native Italian tribe, but the city at its end in AD 79 was constructed in a rectilinear fashion that shows a great deal of Greek influence.  This was more than likely due to the proximity of Greek colonies such as Cumae and Neapolis.  The former was settled in the eighth century and the latter the sixth century BC.  The grid-like plan of the city was probably influenced by the work of the Greek city planner Hippodamus, the probable instigator of the Roman military style examined above.  However the cityís proximity to Rome meant that it was often subject to restructuring at the hands of its governor and several emperors.  As can be seen from the plan, the cityís oblong shape does not resemble the previous examples of Exeter or Timgad.  The eastern part containing the amphitheatre shows some resemblance to the form of the castra but this is contrasted with the non-orthogonal south west of the city.  Pompeii is therefore an example of a city that grew over time with its population, and like Rome itself, was not subject to a rigid orthogonal plan.  Rome too, with its Etruscan beginnings, was, according to Grimal, built with an orderly layout.  However, the quickly rising population and also the famous seven hills meant that orthogonal urban planning was not a realistic policy for the capital.

        These three towns/cities show three basic types of Roman settlement.  Exeter represents the oppida that developed from the castra.  It was a town that adopted a vacated military installation and grew outwards from it.  Timgad, although similar in design to Exeter, is associated with the colonia or strategic settlement of veterans from the army.  It was built expressly for its occupants, and its urban planning was an extension of the process in use by the Roman army.  Pompeii was a city that was occupied before the Romans achieved their domination of the Italian peninsula.  It was built using a Greek model (perhaps a major influence on later Roman urbanism) based on the ideas of Hippodamus.  Its irregular plan is the result of the increasing needs of the population, as at Rome.  It is now necessary to discuss some other Roman cities in order to show how these three types may be applied as standards to other Roman settlements across the empire.  For the purpose of this paper, these three types will be given labels for convenience.  Type A are those towns/cities whose founding was based on an earlier Roman castra such as at Exeter.  Type B refers to settlements that began as colonia as in the example of Timgad.  Finally, type C will be used to label cities that were not dependent on any single system of urban planning such as at Pompeii.

        In Britain as in other countries that were uncivilised before their contact with the Roman invaders, there was a propensity for the re-use of castrae in urban development.  Examples based on the same development as at Exeter can be seen at the cities of Lincoln (Lindum), where the original fortifications were conserved and Cirencester (Corinium), which was abandoned by the Roman army c. AD 49.  However sites of the same origin can also be found in other parts of the empireís frontiers: for example Augst (Augusta Raurica) in Germany, Vienna (Vindobona) in Austria and Budapest (Aquincum) in Hungary.  All these sites can be labelled type A. It is important to note that all these sites were extremely important in Roman military campaigns.  The sites in Britain were fundamental towards the conquest of the Celts, and those on the continent in the subjugation of the Dacians and the guarding of the frontier against the ever-threatening Germans.  These places share their origins as military castrae and are therefore different from the veteran coloniae similar to Timgad.

        The cities of type B are normally known as coloniae.  They are found throughout the former empire.  They were set up as a method of both showing Roman influence and power in the new territories, and as a way to deal with the ever-increasing problem of land for ex-military personnel.  Among this type of settlement are numbered several in North Africa, along with Timgad.  This may have been due to the vastness of the land available, but was also due to the continuing threat posed by African tribes.  A good example apart from Timgad itself is Djemila (Cuicul) which was constructed under Nerva c AD 96-98.  The town is not as castramented as Timgad.  This is due to the fact that Timgad is more an exception than the rule.  Surveyors were not always able to position their settlement on an ideal flat plain.  Factors such as landscape, roads and small pre-existing settlements were often incorporated.  Djemila is an example of this, as it was built on the road from Sicca Veneria to Cirta, and is also constructed on a hillside.  Other examples include Aosta (Augusta Praetoria) in Northern Italy and Zader (Zadar) in Dalmatia. There is a second category of colonia that also falls into type B; this is the planned town, constructed for civilian colonists.  A good example of this type can again be found in Roman Britain at Silchester (Calleva Atrebatum), and another at Zaragoza (Caesaraugusta) in Spain.  These planned towns were built for their population, and their construction represents the best examples of Roman urbanism due to their normal establishment on sites without previous settlement.  These towns differ a great deal with type C due to this exact factor.

        Towns and cities in type C are generally of a pre-Roman nature and are of the largest number.  Their roots were in the culture that began each settlement, and the Romans were merely tenants who could not always change a great deal of their urban layout.  Pompeii, as was discussed above, has its roots with the Oscans and the Greek traders of Magna Graecia and the Hellenistic world.  The other cities that fall into this category come from the civilisations that predated the Romans, but that also fell under their dominion.  Hellenistic cities such as Alexandria and Cyrene were all absorbed.  Cyrene fell under Roman control in 96 BC.  Here the Romans extended the harbour and destroyed then rebuilt a great deal of the old city.  In Africa, the Punic town of Carthage (Carthago) after its destruction in 146 BC became a colonia just before the death of Julius Caesar in 44 BC.  These cities became conglomerations of Roman and pre-Roman design.  This leads to urban plans that contain both castramented areas and older less structured street patterns.  The clearest example of this phenomenon can be seen on the plans from Pompeii (see above) which show the difference between the newer Roman building projects in the east and the older part of the city around the Triangular Forum.

        As can be seen, types A, B and C have clear differences between the urban plan of each of them.  Type A consists of a small and organised central area that has been left over from a castra.  Type B has been shown to have a larger planned layout, which though built in the military style similar to the castra, includes amenities at its conception such as a forum or a theatre.  Type C is of course difficult to quantify. Roman cities of this type have several different origins due to their construction by various cultural groups.  These groups, such as the Hellenistic Greeks, preceded the Romans and had already constructed large functional cities.  When the Romans arrived, there was little that needed to be re-planned, especially in the case of cities that had been designed by, or were based on the ideas of Hippodamus.  However, even in these Roman cities of non-Roman origin which possess an orthogonal layout there are still major differences distinguishing them from the Roman model.  First of all, the planning of type C cities is not based around a settlement within a rectilinear fortification wall such as those based on or descended from the castra.  Secondly type C does not possess the (usually central) crossing main thoroughfares of the cardo and decumanus that are closely associated with the other two types.  Therefore, I wish to propose that this classification system allows the archaeologist to surmise the origin and even the original inhabitants of a Roman settlement.  This proposition assumes that artefacts of Roman manufacture have been found on-site (otherwise type C could not be proven conclusively, although this does not affect the identification of types A and B).  To show how this identification process would work it is necessary to depict a hypothetical situation.  For the purpose of this exercise, a site ìXî has been located and its plan is visible from an aerial photo.  From this photo, it would then be possible to identify the town/city type from answers to the following questions.  1) Is there a rectangular or square fortification/boundary line visible in the plan of X?  If the answer is negative then it can be assumed immediately that this settlement is of type C. T his result then shows that Xís origins were probably non-Roman, and signs of other more ancient peoples should be saught.  If the answer is positive then the response of the next question will help to differentiate between types A and B.  2) Does the fortification/boundary line around X contain 24 or less insulae or divided sections?  If the answer to this question is positive, then it seems extremely likely that X can be labelled as type A.  This shows that the settlement developed from a Roman castra and that the original inhabitants were soldiers.  If the answer is negative then it must be of type B.  This is due to the enormous size of the settlements of type B (around 140 insulae at Timgad for example).  A large contrast with the castramented and small centre of type A, which precluded the development that grew outwards from the military core.  Type B can of course be subdivided into military and civilian colonies, however for the purpose of this paper they are grouped as one.  Therefore in order to find out who were the original inhabitants more investigation needs to take place.

        From the above results it is shown that these two simple questions prove how simple the main differences between these three types are. There are of course variations that have perhaps not been accounted for here. However, even with this margin of error, I believe that the above system of identification is a valid method (providing that there is air reconnaissance material available for study) for beginning the process of distinction at the discovery of a Roman settlement.


Cornell, T. & Matthews, J., Atlas of the Roman World (Oxford 1982)
Crummy, P. ìThe Origins of Some Major Romano-British Townsî Britannia 13 (1982) 125-134.
Colchester: The Roman Fortress and the Development of the Coloniaî  Britannia 8 (1977) 65-105.
Grant, M., Cities of Vesuvius: Pompeii and Herculaneum (New York 1971)
Grimal, P., Roman Cities (trans. M. Woloch) (Ottawa 1979)
Hermanson, G., Ostia: Aspects of Roman City Life (Alberta 1981)
Kraeling, Carl H., City Invincible: An Oriental Institute Symposium 190-223.
Morris, A.E.J., History of Urban Form Before the Industrial Revolutions (New York 1994)
Scullard, H.H., The Etruscan Cities and Rome (New York 1967)
Ward-Perkins, J.B., Cities of Ancient Greece and Italy: Planning in Classical Antiquity (New York 1974)
Webster, G., Fortress into City: The Consolidation of Roman Britain in the First Century AD  (London 1988)


<--Retour à la liste des gagnants / Return to Essays 2000-2001
Fermer ]