Roman Legionary Fortresses


This website serves to augment the book Handbook to Roman Legionary Fortresses (Bishop 2012), but should nevertheless be able to stand alone as a source of information, further reading, and links to do with fortresses. It is an evolving work and, whilst its first stage is now complete, further enhancements are planned. Just as mapping each individual site was an integral part of the process of writing the book (few things are quite so revealing about the quality of scholarship on a site as when the published plans are unclear as to an exact location), so the production of the website has helped channel my thoughts about what else can be done with the data beyond bare and boring statistical exercises. Each page has been marked-up by hand (using the ever-reliable Bluefish) to a template designed and – contantly – tweaked by me.

Types of fortresses

Within the book, I saw fit to divide sites into three categories: proto-fortresses, early fortresses, and late fortresses. A few words on these distinctions may be appropriate here.


Proto-fortresses are those sites that I saw as leading towards what we perceive to be the legionary fortresses of the Principate. Of course, throughout what follows, there is no guarantee that the Romans saw things in this way. They could not know where developments in castrametation in the late 1st century BC would lead.

I included the Castrum at Ostia as my first proto-fortress, following it with the Spanish bases of Càceres el-Viejo and those around Numantia (Castillejo, Renieblas, and Dehesilla). These established the idea of maintaining long-term garrisons in one place, a notion which was taken further in the Augustan period, with sites like Marktbreit, Oberaden, Haltern, and Anreppen.

Early fortresses

Early fortresses are the stereotypical legionary bases that might be expected when the term is mentioned. They are, contrary to popular opinion, quite a diverse bunch of sites. They resemble each other in some respects, but to find two that are even vaguely alike is quite unusual.

The earliest of these reflected their heritage by being constituted as double-legionary fortresses. The best known of these is Vetera I, but other examples existed at Nijmegen–Batavodorum and Mainz–Mogontiacum. These began to be replaced by single-legionary fortresses, like Colchester–Camulodunum, Exeter–Isca, and Gloucester–Glevum, until the point where Neuss–Novaesium is now recognised (whether rightly or wrongly) as the archetypal legionary fortress. Connected with the development of these, even though it was not strictly a legionary fortresses, was the Tiberian, brick-faced-concrete-built Castra Praetoria.

There is at least one incomplete legionary base (Inchtuthil) within this group, the construction of which happened to coincide with a move from timber to stone construction amongst the British fortresses (York–Eboracum, Caerleon–Isca, and Chester–Deva all receiving stone walls by the beginning of the 2nd century AD).

The 2nd century AD saw new fortresses at Ločica–Celeia and Albing, one possibly following the other, but still following the by-now established plan. Severus’ foundation of the base of legio II Parthica at Albano Laziale–Castra Albana and subsequently at Afamia–Apamea saw the continuation of this tradition.

Late fortresses

Late fortresses developed from those standard, earlier fortresses. They seem to reflect the way in which the army was evolving, with a trend to legions being split into smaller sub-units. However, a degree of caution is advisable. It is now accepted that the Castra Praetoria had two-storey barracks and some late fortresses may also have enjoyed this facility. This would render traditional calculations of unit size slightly problematical.

Alongside de novo constructions like El-Lejjun–Betthorus (-um?) and Udruh–Adrou, there were conversions of existing structures, such as the Egyptian temple at Luxor (al-Uqsor)–Thebae, a corner of the city at Tadmor–Palmyra, or the curious adaptation of one side of the Principate-era fortress at Budapest II–Aquincum. Most of these were smaller than the earlier fortresses, but some, like Köln-Deutz–Divitia, Mautern–Favianis, or Schlögen–Ioviacum might even be termed micro-fortresses, a new form derived from the fragmentation of once-centralised units.

Vexillation fortresses

Vexillation fortresses are sites that were larger than a standard fort (or castellum) and smaller than a legionary fortress, designed to contain a force made up of more than one type of unit. A conscious decision was made to omit vexillation fortresses from the book, but this may be reviewed at a future date with regard to this website. Moreover, as is shown in the book, more than one type of unit could in fact be found in a ‘normal’ legionary fortress. There is thus something of a grey area between large vexillation fortresses and small legionary fortresses and a good example of this is the controversial base at Rottweil–Arae Flavia I, omitted from the main text of the book as a legionary fortress. One point of view sees the Rottweil site as the legionary base of legio XI Claudia before it moved to Windisch–Vindonissa (Franke 1997), whilst another sees it as a large vexillation fortress (Fellmann 2000, 128–9), pointing out the proximity of a smaller and later vexillation fortress at Rottweil–Arae Flavia III. Ultimately, the distinction is a semantic one which may not be diagnosable through archaeology.


Some fortresses, like Ptuj and Xanten–Vetera II, have been removed by migrating rivers, others (like Sadak–Satala) flooded by dams. The majority of the remainder are now urbanised to a greater of lesser degree, but a significant minority still lie in open country. Paradoxically, the threat to the latter from agriculture, quarrying, and other forms of development can be greater for these than sites buried deep within tell-like modern cities.

Plenty of fortresses still remain to be identified. The bases at al-Rifniah–Raphanaea and Lajjun–Legio have only recently be found, whilst Ra's al-'Ayn–Resaina, Tayibeh–Oresa, and Mehin–Danaba remain, for the time being, just possibilities. There may even be some that are unknown and unsuspected, Rumsfeldian ‘unknown unknowns’; only time will tell.


The site bibliographies are largely those included within the print volume, augmented by new publications as they become available. Whilst every attempt has been made to be comprehensive in coverage, there will inevitably be omissions. If you know of anything that should be included, please contact me at the email address below to let me know.


Bishop, M.C. 2012: Handbook to Roman Legionary Fortresses, Barnsley

Fellmann, R. 2000: ‘Die 11. Legion Claudia Pia Fidelis’ in Le Bohec, Y. and Wolff, C. (eds), Les légions de Rome sous le Haut-Empire. Actes du IIème congrès de Lyon (17–19 septembre 1998) sur l’armée romaine, Paris, 127–31

Franke, R. 1997: ‘Ein Lager der legio XI Claudia in Arae Flaviae/Rottweil und die Besetzung des oberen Neckargebietes’, Gesellschaft Pro Vindonissa, Jahresbericht 25–32

Additional information for inclusion may be sent to
(please put Introduction in the subject line; all other emails may be treated as spam and deleted).

Page last updated 12th December 2013