The centuriation of CCAA: a research proposal


In principle, the land of a Roman colonia of the first century AD (1) was surveyed by centuriation. It was a legal requirement to do so. The agrimensores tell us that it was the normal practice. The purpose was to define and register, in a cadastre, the individual holdings of the retired soldier colonists.

In reality the picture is more complex.

Once recognised, these facts change the ways in which we might look for centuriations, and the sort of evidence which we expect to see.

A lost centuriation

The first century AD was a period of vigorous land development within the expanding Roman empire. Many cadastres were set up. Centuriations were everywhere.

Some territories which already had centuriations were resurveyed, as in southern and eastern France (Béziers 'A' and south of Dijon). Others had their first centuriation - including southern Tunisia (24-30), Damascus (3) and Szombathely (Colonia Claudia Savaria) in Hungary.

In this picture of widespread strenuous activity, managed by central authority at the highest level, the north western part of the empire looks out of place. Many think that Britain (4) was untouched by centuriation, nor is it easily to be found in northern France, Belgium, the Netherlands, or the Rhineland.

The case of Köln is particularly striking. Colonia Claudia Ara Agrippinensium (CCAA) was, as its modern name suggests, the colony in its province. By rights, it should have a centuriation; so where is it?

Others have asked the same question.

Neither of these proposals has been accepted and it seems that, up until now, all attempts to find the centuriation of CCAA have failed.

A happy chance discovery?

When direct assaults upon a problem fail, indirect approaches may succeed. In this case an investigation of the reality of the supposed
centuriation of South Limburg (Netherlands) may have lead me to stumble, almost by accident, upon a centuriation of Köln. Further work could show whether or not there is any substance in this idea, thus demonstrating how serendipity can play its part in research.

The South Limburg centuriation

Edelman and Eeuwens (1959) proposed the South Limburg centuriation, but in the intervening half century the idea has not been widely accepted. Now, more recent data on site location shows (Peterson 1996) that it may well exist and extend much further than originally thought, in the direction of CCAA.

My interest in the South Limburg centuriation was stimulated by earlier work on the equally controversial centuriation of South Norfolk 'A'. The systems are similar, in that they are both

Nevertheless, there are some differences.

Because the Limburg data are so voluminous (nearly 500 sites), they produce very significant results from Kolmogorov-Smirnov tests of their distribution with respect to the limites of the centuriation. The Roman sites are not always located solely according to the suitability of the natural environment. They also tend to be sited near the limites, as is normal in other centuriations.

Since, to judge from the statistics, the centuriation probably exists, and since modern national boundaries may not limit it, it is natural to ask if it might extend outside South Limburg.

This question does not seem to have a positive answer, except in one case. The system seems to have left no obvious trace further north in the Netherlands, west of the Maas, or very far south of the southern Belgian border. However, in Germany the case is different.

The South Limburg centuriation in Germany

In the area between Limburg and Köln a systematic search for traces of the centuriation has not yet been attempted, but there are a number of localities (see general map above) whose features seem to suggest that such a study could be fruitful.


Here, immediately across the German border, the possible traces of the centuriation are just as visible as they are in the Netherlands.

Many existing roads and other ways (thick red lines) and other boundaries (thin red lines) have the orientation of the projected centuriation grid (brown lines). In some areas (such as Nieder and Ober forstbach) most of the longest oriented traces are on or close to limites.

Roman roads around Aachen, after Horn (1987), are shown by blue lines. One of these is continued by the (former) modern main road towards Jülich, the purple line. This road approximates a limes in this section.


Further east again, in the area of brown coal extraction near Hambach, some of the excavated Roman farms have a possibly significant relationship with the extended centuriation, as shown here. Their relationships with the grid speak for themselves, and the apparent rectification of the course of the stream which may have been a water source for Hambach 512 is also interesting. However, it should be noted that the situation in the wider area is complex, since villas nearer the main Roman Jülich-Köln road (the double line on the above map) are not organised in this way. Gaitzsch (pers. comm.) thinks that uniform land planning over the whole area is unlikely.


Going south east, at Vetweiß, there is another area (not illustrated) in which modern features coincide with the projected centuriation. Here also it may have been preserved in existing landscape features.

Near Vetweiß, at Soller, structures in an area of Roman pottery production could be linked to the centuriation. Most of the kilns and buildings have its orientation, despite the appearance on the 1:25,000 map that the natural features and modern boundaries have an orientation that is different.

Yet another interesting feature of this area is the existing Düren - Zülpich road. It seems to be linked to the grid by closely approximating a series of oblique segments passing through corners, as shown on the left. The angles of each segment to the grid are, reading from the top (north): 1:3, 1:4, 1:3, 1:2. Such constructions are typically Roman. They seem to suggest that the road is originally of that period.


At Köln there are also some possible links between the hypothetical centuriation and Roman topography, see the above figure. The eastern part of the city's decumanus is very nearly coincident with a line joining opposite corners of a square of the centuriation (brown lines). Hence the east-west streets are generally at 45 degrees to the centuriation (but note that for north-south streets this is not so, because they are not at right angles). Such a 1:1 relationship is, again, typical of planning in centuriations.

There are five major Roman cemetaries around Köln (the fifth is along the western main road, just outside the border of the figure). If I interpret Hellenkemper (in Horn 1987 p. 486-8) correctly, there are existing churches in three of these areas which developed from late Roman memorial chapels. These churches are rather close to theoretical limites .

The cemetaries are discussed by Riedel (in Horn 1987 p. 493-7). He points out how 4 of them are related to main roads (including the road to the west). This is the normal arrangement of cemetaries associated with a Roman town.

The exception is the cemetary to the north-west, which is, according to Hellenkemper, the oldest. There is no known road leading to it. Despite this Riedel (Horn 1987 p. 495) suggests that such a road "forming the backbone of the cemetary" must have left the Roman town in the north-west direction. Quite so!


This research should be carried forward to the point where we can judge whether or not the South Limburg centuriation could be part of a centuriation of CCAA, for if the answer were "yes" we would have a great aid to understanding the process of development of the city and its territory.

One approach would be to use all the available German data on Roman sites within the area that is likely to have been the territory of CCAA. As with South Limburg the Kolmogorov-Smirnov test could be used to measure the degree of association between them and the theoretical grid. Other, less indirect approaches could also be tried, such as comprehensive studies of data on Roman field boundaries in the area.

This is not necessarily a big task. What is needed is collaboration and sharing of information and techniques. I am naturally especially interested in German contacts, but I would welcome contributions from anyone who finds this puzzle as intriguing as I do.

John Peterson

28 August 1997

(slightly amended 9 March 2005)

Last updated on 1 September 2013 by John Peterson



1 Those settlements which acquired the title colonia in a later period are not relevant here. The title was honorific and hence there was not necessarily any new allotment of land holdings.

2 Nor need the territories of each town be disjoint. The cadastre of Orange B proves that the holdings of the local tribe, who had their capital at Augusta Tricastinorum, were intermingled with those of the colony, all within the same centuriation. This shows that we cannot define the territories of ancient settlements by Thiessen polygons.

3 The existence of the centuriation of Damascus is not in doubt, but a date before the start of the first century AD is possible, see Dodinet et al. (1990)

4 For Britain , I suggest that this anomaly may not be real. Others have struggled, sometimes in a most unconvincing fashion, to find reasons for it. In particular they have felt the need to explain why Roman approaches to land administration around the island's three perfectly ordinary first century coloniae at Colchester, Gloucester, and Lincoln should be so different from what would normally be expected.

5 Serendipity the faculty of making happy chance finds. [Serendip, a former name for Ceylon. Horace Walpole coined the word (1754) from the title of the fairy-tale 'The Three Princes of Serendip', whose heroes 'were always making discoveries, by accidents ......, of things they were not in quest of'.] Chambers Twentieth Century Dictionary (1972)


Dodinet, M., Leblanc, J., Vallat, J.-P, and Villeneuve, F. 1990. Le paysage antique en Syrie : l'exemple de Damas. Syria 67 : 339-367.

Edelman, C. H., and Eeuwens, B. E. P. 1959. Sporen van een Romeinse landindeling in Zuid-Limburg. Berichten Rijksdienst Oudheidkundig Bodemonderzoet 9 : 49-56.

Gaitzsch, W. 1986. Grundformen römischer Landsiedlungen im Westen der CCAA. Bonner Jahbücher 186 : 397-427.

Hauptmann, C. 1915. Die Erdvermessung der Römer. Raum des heutigen westlichen Kriegsschauplatzes bis zur Rheingrenze. Rhenenia-Verlag, Bonn.

Horn, H. G., ed. 1987. Die Römer in Nordrhein-Westfalen. Theiss, Stuttgart.

Klinkenberg, J. 1936. Die Stadtanlage der römischen Köln und die Limitation des Ubierlandes. Bonner Jahrbuch 140: 259-298.

Lambert, A. M. 1971. The Making of the Dutch Landscape. Seminar Press, London.

Peterson, J. W. M. 1996. A computer model of Roman landscape in South Limburg In Interfacing the Past: Computer Aplications and Quantitative Methods in Archaeology 1995 Analecta Praehistorica Leidensia 28.1, Kamermans, H. and Fennema, K. (eds.), 185-194. Leiden, University of Leiden.