A Trajanic Embroidered Shoulder Panel from Luguvallium
(reconstructed by Robin Brown)

Shortly after the emperor Trajan came to the throne, in AD 98, an area just to the south of the Flavian fort at Luguvallium (modern Carlisle) was probably occupied by the enterprising craftsmen who followed the army and supplied soldiers with manufacture and repair facilities. The buildings here were made mainly of reused beams from earlier constructions. The oldest building was wattle and daub on a timber frame, and about ten years old. The two rooms were variously used for accommodation and possibly as a workshop. Outside in an adjacent compound, fragments of armour, cavalry equipment, and carpentry tools. There was probably a butchers' and an armourers' shop nearby.

Fig. 1 The surviving roman remains

In this area, together with other leather fragments, two pieces of embroidered leather were discovered. (Fig. 1) These have been interpreted by Susan Winterbottom of the Carlisle Archaeological Unit, as coming from the shoulder piece of a garment of some kind.

It is not possible to tell whether the item is military, nor whether it has legionary or auxilliary connections. Nonetheless it is a reasonable hypothesis since there were a number of items which were of military origin found in the same place.

This fragment suggests an authentic and practical solution to the discomfort caused to modern re-enactors when wearing heavy loricae, as well as being an ideal way to smarten up our appearance when in "undress kit", i.e. without a lorica.

The first stage of the reconstruction consisted of the manufacture of a plain piece of leather identical in size to the larger of the two original finds. (The second fragment is a long narrow strip which overlays the main piece along the top edge on Fig. 1). Using this, I was able to experiment with the hypothesis that this was a shoulder panel.

Given the proximity of certain cavalry finds, I wished first to discount the possibility that it was part of an elaborately decorated roman saddle, due to the superficial resemblance to a horn cover. I quickly discovered after comparison with the saddle components, that the straight edges of this find cast doubt on this possibility. Perhaps the most convincing demonstration that this is indeed a shoulder panel comes when it is laid on your shoulder. The angle of the straight edges follows the seams on a modern shirt almost exactly.

Fig. 2 Peter Connoly's reconstruction of a roman saddle. Note the curve edge on the bottom.

Having established the plausibility of the leather panels as shoulder panels, the next step was to reconstruct the original design on the panel, using the surviving stitch marks. I have interpreted the design working on the assumption that all the stitching on the decorated panel served to attach a piece of leather. The best solution was to create a decorative leather shape underlain by a contrasting oblong panel. Given that the (modern) function of the "tunic" is to pad the shoulders, there was a requirement to produce a panel for the reverse of the shoulder. Not wishing to produce something totally hypothetical, I have used the same shape as that used on the front, to give a pleasing effect. This does allow some manufacturing shortcuts to be taken in reconstruction, as a single piece of leather can be used for each shoulder piece. Given the limited availability of primary evidence, I consider this departure from the evidence is acceptable for a practical item. I decided to reconstruct the remainder of the "tunic" in leather, although there is no reason why some kind of cloth could not have been used.
The only remaining problem is the resulting broadness of the "shoulders" of a tunic made with replica leather panels. These appear to be far too long and extend beyond the shoulders of even the largest of modern re-enactors.

Fig. 3 Leather panel takes the weight of the lorica segmentata	An experiment carried out during an event at Corbridge, Northumberland in 1998, provided a possible explanation for this design feature. These apparently excessively broad shoulders provide excellent protection when worn beneath a lorica segmentata.
Fig. 4 The final reconstruction: here a legionary optio is depicted in undress uniform, 98 AD	The close fit of the leather to the neck, and its extension beyond the end of the shoulder helped to prevent the leather straps and metal plates from leaving their usual imprints in the soldier’s flesh. It was perhaps appropriate that this discovery was made at the location of one of the most important find of this type of armour.

Unfortunately the leather over tunic does not appear to suit the wearing of lorica hamata, unless the lorica were to be specially altered to suit the enlarged shoulder area. I expect that the same will be true for lorica squamata. In both cases, there may be a reduced level of flexibility which is already inherent in lorica segmentata which may take wearers of other types of loricae some getting used to. Whether the garment has the same beneficial effect will need to be the subject of future experiments.
As far as durability is concerned, the "tunic" pictured below has been worn under a lorica segmentata at every show since, for nearly 5 years (at least 50 days wear), and shows no sign of deterioration.