Reverse Engineering Roman Cavalry

1. Context.

In the precious post I was quite hard, if not downright disrespectful about some of strange conceptions about how cavalry horses might be stabled. So it’s time to put some cards on the table, and explain how horses and built environments are meant to interact – from a more traditional perspective, after all, cavalry barracks have existed and continue to exist since ancient times, and horse care in a variety of contexts is hardly a forgotten art.

My work for Philip Crummy in Colchester on Britain's first legionary fortress taught me to respect the accuracy of Roman Military surveyors. An important point also was apparent; Romans favoured stone foundations, but significant parts of the superstructure were built from timber for ease and speed of construction.  In terms of Hadrian’s Wall, we have very poor data sets, but the Roman army was very systematic in its architecture, so the Fort chosen is North of the Wall at Birrens in South West Scotland.

Theoretical modelling is a way of testing and refining ideas about how buildings and built environment might work. In this context, given the number of unknowns, modelling can only give you a range of answers and help delineate the problem. Having tried out various options against the available data, what is presented is a best fit model, directed to a specific aspect of a problem, in this case how many horses can you house in a cavalry stable, and what does this tell us about the size of cavalry units.

There is certain fetish about deploying Latin vocabulary which is not always helpful, however, I have to confess a prejudice to the use of Cohort sized units [600] and centuries [100]. There are a variety of terms deployed notably, equase, decurion, turma, ala, quingenaria, and Milliaria.  While the primary objective is to develop a model of a cavalry barrack, and from this delaminate the scale of units, relating this to Latin vocabulary would be seen as an academic objective. 

2. Roman Forts Basics

It is generally understood that the layout of camps mirrors the layout of the Roman marching camp, which in turn reflects the structure of the army, with individual 8 men tents [contubernium] formed up in centuries, and cohorts. The disposition of cavalry may reflect picket lines of horses, which have to laid out with same aims in mind, namely keeping them securely organised for ease of care, deployment and preventing them from kicking each other.  

In the Roman Army class is important in the distribution of Space, notable for the space allocated to the Roman commanders house & bathing. Forts are similar in plan being built up from the similar components, principally, a single Headquarters building, Commanders house [+bath], Hospital, and  varied numbers of; Infantry barracks, Workshops / Stores , Granaries, and Stables. 

 3. Hadrian's Wall & Birrens

This exercise is try get a sense of Hadrian’s design for his Wall in terms of cavalry, principally because they represent be an ideal supervised the emperor, which is a great context for archaeology. In reality, the evidence from this Broad Wall phase is fragmentary, and most of the forts were built or rebuilt after the initial phase. The best preserved cavalry fort in the North is at Birrens, part of the Antonine Frontier built after Hadrian’s Wall was initially abandoned.

Much of the Plan was recovered, but has a few problems, particularly the lack of a commanders house, some ambiguity about building A2, and a sense that the fort is a bit light on storage / workshops. There is an annex to the East which probably accounts for this. The site appears to have been tarted up at some stage with rebuilding or refacing the gable ends of buildings.

 4. Cavalry; Human Consideration & assumptions

As I understand it, in this period Auxiliary Cavalrymen were high status non-Roman troops from horse owning elites elsewhere in the empire. It is not known whether the status of cavalry was reflected in their accommodation. They were an expensive resource intensive asset, and presumably would be equipped the best horses available in the area where the unit was raised.

My guess / assumption is that there is a 1:1 ratio of support troops to cavalry troopers. Apart from Historical best practice; the following considerations may apply:

• Need to deploy the whole unit en masse and quickly.
• Armoured men with weapons are not suited to the preparation or care of horses.
• Cavalry extends fort’s curtain wall which requires additional troops to defend.

Cavalry have to be part of a predominantly infantry unit.

5. Horses General Consideration & assumption

Keeping large numbers of horses undercover in a compact area on a long term basis requires a specially adapted built environment. Horses prefer high airy spaces, with underlying floor surfaces that are dry, firm, but not too hard, and graded for drainage.

Stables

A building with access to water supply and suitably drained, with individual Space for each horse, allowing for; Feeding, Watering, Mucking Out, Grooming and Tacking up.

• Storage for tack and other kit.
• Wider and taller entrances. 
• Space for mounting and forming up.

Other buildings;

• Sheds or lofts for Forage / Hay / Bedding
• Granaries for Barley & Oats
• Buildings for: Blacksmith / Farrier; Equine vet; Sadler
• Ideally, a covered space for training, assembly and practice.  

 6. General model; The Roman fort at Birrens

The model includes a consideration the whole fort as a context for the cavalry stables.

 The fort is regularly laid out in Roman feet shown at @0.296m; however, it is not necessary to best guess precisely what is going on in the minds of the surveyors and builders, for the following reasons:

• I do not wish to push the accuracy planning historical and modern.
• Precisely what is being measured, including the thickness of the component is not clear.
• The dataset is for the measurements of foundations not roofs
• Allocation of external spaces like roadways is not implicit in the design of individual buildings
• It is only relevant to the number of bays in the stables and their capacity. 

Looked at as a whole the fort is a nominal 600 x 400 ' centred on the HQ building, each set of stables occupies a plot about 150' square, roughly the same as the 6 double infantry blocks.

Including Granaries G3 & G4, and  building A3, the stables account for 40% of the forts area.

The use of standard measurements and components greatly simplifies construction and is evidenced by the uniformity of the building types. The use of standard roof trusses, which in turn is reflected in standard bays, is most clear on the buttressed buildings.

The 150’ long barracks are presumably about 10 – 12 bays long, [11 being not uncommon on the Wall]. Some produced evidence for partitions, perhaps at 60, 120, feet, while others produced no evidence. It cannot be cannot presumed there is only one type of barrack represented.

The end on nature of the Stables allows them to be packed closer together than side on buildings like the barracks. Double Walled barracks show the dedication to the single truss and the avoidance of valleys. Built back to back they can make use of the terraced sloping sites to improve drainage.

It is worth noting that between Granary G1 and building A1 there appears to be a corn dryer, which suggest that crop processing is taking place.  In which case A1 may be a barn for processing and storage of harvested grain, indicating some degree of local arable production by the garrison.  For a cavalry fort the need for meadow and grazing, and more specialised grains implies a more complex agricultural hinterland.

7.  Roman Stable theoretical model

Based on the buildings at Birrens, the model has two sets of stalls with a central passage, and serviced by feeding passages. There is a loft for hay or bedding overhead.

The capacity of buildings are dependent

• The choice of bay size
• The number of occupied bays.
• The number of horses per bay

The model assumes 4 horses per bay, utilising 10 - 13 bays per building as shown in the table below. [2]  The buildings are 150’ long; allow for the thickness of partitions, 10 – 12 bays at 15 – 12.5 foot intervals work well, which offers a range of values for the number of horses of 30 – 50.

Once the models covered central passage is accepted, it is simply a matter of how many horses can fit on the two rows to either side. In this respect a five foot stall is probably too narrow, and 6 foot stall generous. Assumptions also have to be made about the provision for spare horses and even whether officers had more stable space.

8]. Conclusions

Within the parameters set out the model stable has a capacity between 30 – 50 horses or 180 – 300 horses for the whole fort / unit. Offering three probable scenarios for the six buildings 

• A]. 2 x 100 man units
• B]. 3 x 100 man units
• C]. 2 or 3 x units of intermediate strength. 

Solution A; 200 horses

Each stable would hold the horses of a troupe of 30 + 3 officers, a century being 3 such troupes; 90 troopers + officers.  The six blocks are a nominal 200 horses for 200 cavalry men. 

Two units fits with the 2 Extra Granaries for feed,{G3/4} each with six bays, which together the 2 Cohort Granaries, {G1/2} and the axial symmetry of the fort layout, all suggesting we are have two units. 

At Birrens there is potentially 16 spaces, representing a surfeit of potential infantry barracks with room for 1200 in the 6 double blocks alone, with another 4 blocks spare; two are in prime position next the central area with access from both sides.  Various combinations seem possible, based on 2 cohorts +/- cavalry.

One way of accounting for this space is to assume the cavalry have more generous quarters, perhaps  3 blocks per unit mirroring the layout of the stables.  Giving 6 blocks for cavalrymen, 2 for support troops, and 6 for an infantry cohort, making 14 units in total.   Thus, detailed understanding of the garrison is dependent on assumptions about the ratio of horses to cavalrymen and support troops, as well as how their status might be reflected in the allocation of living space.

The minimum view is that we have two centuries, comprising 200 horses, with 200 cavalrymen & 200 support troops.  With the addition of an infantry Cohort of 600 this would create a garrison / unit size of 1000.

Solution B 300 horses & Solution C 200 -300 horses.

To get 50 or more horses per stable would require narrow stalls and leave little options for spare horses.  It seems too much of a push to get to 300 horses comfortably onto the site.  However, within the modelling it is possible to conceive of an understrength 300 man unit.

Thus, all the above does not preclude a unit comprising any vaguely sensible number between 200 – 300; a unit of 240 horse could be 3 units of 80 would fit, and make supporters of the 80 man century very happy.  

After Inscription from Birrens  dedicated by II cohort of Tungrians {Rib I 2110} [3]

The modelling pushes an engineering perspective, which has its own logic and drive for uniformity, which in the context of the Roman Army is not inappropriate.  Notwithstanding the issues of  how cavalrymen were accommodated, the ratio of support troops and infantry to horses, the model helps delimit the scale of cavalry units, particularly emphasising the space required compared with regular infantry units.

Happily, historically it is thought that the fort was occupied by II cohort of Tungrians a 1000 strong unit mixed unit.

 Notes.

 [1] Traditionally, the front ranks of a nominal army line is formed from the younger fitter soldiers, with the more senior and more experienced towards the back.  Support troops are the more senior soldiers who are not as physically fit, but are highly trained and well motivated. They represent the specialist in medicine, leather work, armourers and administrators, who do not routinely leave the fort; on campaign they travel with the baggage.  The eight men leather tents which each unit used, formed the basis of the layout of the camp, but would require a mule to carry, and a second to carry the collective equipment of the unit. The baggage is the most valuable part of the army group in the field

[2] The table below is based on the following

1. .Bay width / stall width in Roman foot of 0.296m
2.  Stall is the Stall width – 0.5‘
3.  Assumes 4 horses per bay
4.  Assumes two bays not in use or spare

[3] For the Emperor Caesar Titus Aelius Hadrianus Antoninus Augustus Pius, pontifex maximus, in the twenty-first year of tribunician power, four times consul, the Second Cohort of Tungrians, a thousand strong, part-mounted, publicly praised, (set this up) under Julius Verus, emperor’s propraetorian legate.