A Roman Republican Shield

by L Whiteley

For my practical project I chose to reproduce a Roman Legionary shield, a Scutum. After looking at the various shield types used by the Roman Army throughout its history, as well as considering the evidence available for each one, I decided I would attempt to make a shield from the Republican period.

     The history of the scutum can be traced back to the 8th Century BC. In Italy at around this time the traditional circular, “hoplite” shield began to be replaced by an oval shaped shield. This transition was no doubt due to changes in the style of fighting, for the defence aspect of military technology always develops to counter advances or changes in the methods of attack.

       The oval design quickly spread across Europe and was adopted by the first Roman legionaries. The Roman version also incorporated a central reinforcing rib, the spina, an element taken from earlier types of circular shield. This type of shield is described by Polybius (VI,23) and is represented on various monuments (see figs.1 & 2). However the best example comes from Kasr-el-Harit in Egypt (see fig.3). Although it cannot be accurately dated it is likely that the design was used throughout the Republican period (see Feugere, 1993, 76).

The Kasr-el-Harit scutum provided the best evidence for methods of construction and so was my primary source when building the shield.

Description of Kasr-el-Harit Shield

-         120x64cm

-         Three layers of thin laths glued together giving it a curved form.

-         Covered in layer of felt sown around its circumference.

-         Any colour and decoration has disappeared.

-         Reinforced by a solid spina in three sections.

-         The central boss has been hollowed out.

-         Horizontal handgrip enabling shield to be moved in any direction.

-         No metal components except for nails to secure the spina.

-         No metal umbo or metal edging at the head or foot of the shield.

    The shield was basically made from three layers of wood. “The method of construction can best be described as plywood” (Stephenson and Dixon, 2003, 36.). Two layers of horizontal wooden ribs were used to sandwich a layer of vertical strips in between (see fig.4).

    The Kasr-el-Harit shield was made of strips of birch wood. I was able to purchase 1.5mm birch plywood. Three layers of this material would give the shield a depth of 4.5mm, comparable with the 5mm depth of the Duro Europas shield (see fig.5). The shields found at Doncaster and Kasr-el-Harit both had a depth of 10mm. The reason I did not aim to make my shield thicker was partly due to the cost of thicker plywood and also because I felt a thicker ply would be much more difficult to work with, particularly when trying to attain a curve in the shield.

     I marked out the shape of the shield on the plywood board using the measurements of the Kasr-el-Harit shield, 64x120cm. I then divided it into horizontal strips approximately 50mm wide (the horizontal strips of the Kasr-el-Harit shield were 25-50mm wide). I repeated this process in order to get the two layers of horizontal ribs. I then marked out a third shield this time dividing it into vertical strips 70mm wide (the vertical strips of the Kasr-el-Harit shield were 60-100mm wide). I numbered all the strips so that after I had cut them out I would be able to reconstruct the shield accurately.

    The Romans would obviously have cut these strips straight from a larger body of wood like a felled tree. Unlike me they did not have access to large boards of processed wood which were already big enough from which to cut the shape of a whole scutum.

    These strips of wood from which the shield was built would have been sliced from a piece of timber using a tool with a sharp cutting edge. The Roman hatchet knife would have been ideal (see fig.6) and the legionary’s sword or dagger would have worked just as well for repairing or building scutums whilst in the field (see fig.7). I simply used a pen knife which turned out to be very hard work and very time consuming. Even though the Romans had saws they would not have been applicable for cutting such thin pieces of wood.   

    There is very little evidence on the details of shield construction, such as how the curvature of the shield was achieved. Polybius does not describe it, referring simple to “two layers of wood fastened together.” Soaking the wood in water would make it more malleable and according to Treloar  “plywood could be readily moulded by heat to the right shape.” The Roman re-enactment group Legio XX glue their shields together then tie them round a large tree to dry which results in the shield retaining a curve. This method seems quite plausible for shields built in the manner described by Polybius but would not be possible for shields constructed of thin strips for obvious reasons. A scutum press (see fig. 8) seems much more reasonable and it is likely such devices were used, however to make one would constitute another project so I decided to use water treatment. Bearing the grain in mind I lined a bucket with the horizontal wooden strips and filled it with water.  I left them for a day in order to saturate the wood. I then emptied the water and left the wood to dry. As I intended when dry the ribs had retained a curve moulded by the circular shape of the bucket. With some crude bending and adjustment I managed to get the strips to retain a curve approximately 3” in depth. 3” is the depth proposed by Treloar for Republican shields, he claims “The monuments cannot be made to give the precise evidence of a modern air-photo, but they do indicate that in the Republican period the curvature was comparatively shallow, while by the time of Trajan it appears to form a semicircle.” See fig.9.  

     According to Polybius the Romans used “bull hides glue” to construct their shields. I opted for a more readily available adhesive, wood glue. Although not completely authentic, I was still using glue, the same technology as was originally employed.

      To glue the strips of wood together proved a time consuming process. To retain the curve in the wood without a special press meant only one vertical strip could be glued to the first layer of horizontal ribs at a time. I would then have to wait twenty four hours to ensure the glue was thoroughly dry before attaching the next piece. Fig. 10. Shows how pressure was applied to the part being glued without flattening the overall shape of the shield. Obviously a purpose built press like that illustrated in fig. 8. would have greatly accelerated the process.

   With all the vertical lengths of wood attached to the first layer of wooden ribs the shield became sturdy and took on a recognisable form for the first time. The top layer of horizontal ribs could now be glued in place, completing the three layer structure. These were held in place while the glue dried with an assortment of clamps ensuring adhesion across the curved body of the shield.

    The core of the shield was now complete. I now set about creating a handle, boss and spina. All the surviving examples of Roman shields I have looked at featured a horizontal handgrip which would have been grasped in an overhand fashion “which can be seen on many monuments” (Connolly,  233). The shields from Kasr-el-Harit and Dura Europas both had wooden grips, although a grip made of bone has also been found.

      To make the hand grip I cut out two semi-circles in the centre of the shield leaving a horizontal band of wood which would serve as the handle. I used a pen knife to cut out the semi-circles, being careful to ensure they were big enough to accommodate my hand whilst clenched round the grip. However I could not make this opening too big for it had to be covered by the boss or else I would be left with a shield with holes in it.

     Making the boss was probably the greatest challenge in the process of building the shield. On the Kasr-el-Harit shield the boss was made of wood and “had been hollowed out” (Feugere, 1993, 77). I managed to obtain a piece of timber with the length and width to cover the handgrip. It also had the depth to accommodate my hand whilst remaining suitably dense to fulfil its purpose to protect the hand and as a weapon. Like me the Romans would have started with a block of wood and then crafted it and hollowed it out. This task would have been undertaken with a hammer and chisel (see fig.11). Using a hammer and chisel proved very time consuming and there was the constant risk of the wood splitting which would force me to start again. After many hours I managed to create a piece of wood that resembled a boss and that my hand would fit into. Unfortunately its shape was not as refined as the boss on the Kasr-el-Harit shield, but I was a novice, plus I suspect the Romans would have used a plane to achieve such a defined and tapered finish (see fig.12).

     The spina of the Kasr-el-Harit shield was composed of three parts (see fig.13), the central boss and two lengths of tapering wood that extended to each end and served to reinforce the shield. After making the boss I simply used two pieces of triangular wooden baton as the other components of the spina. I then glued and nailed all three pieces to the main body of the shield just as had been done with the Kasr-el-Harit shield. The nails used to attach the spina were the only metal parts used on the Kasr-el-Harit shield. 

     The wooden core of the Kasr-el-Harit shield was covered in lambs felt. The Dura Europas shield was covered in thin leather and Polybius tells us that shields were covered in canvas and then calve skin. Clearly what was required was a tough, durable material that would encase the shield and protect its interior structure, holding the shield together and thus extending its lifespan. Animal hides were obviously the most readily available and suitable material.

     Unfortunately I was limited by financial resources and was unable to purchase a hide large enough to cover the whole shield. Instead I used leather from old leather jackets. Due to the size of the jackets I could not cut the leather in large enough pieces to make it worth stitching it through the wood as the Romans did (see fig.14), so I simply glued it into place.

   I also added some leather to the interior of the boss in order to make it more comfortable on my knuckles when I held the shield. Although I had no direct evidence for this there can be little doubt that such alterations occurred as soldiers adjusted their equipment to suit their needs. Obviously practicality and uniformity still had priority as it does in any disciplined army, for example, we hear of a soldier being reprimanded for having a shield that was too big.

    On top of the leather covering the Dura Europas shield had a layer of linen and the shields Polybius described had a second layer of calve skin. I used a durable red material (see fig.14) that I stitched round the circumference of the shield with a strong linen thread in the Roman fashion. It is not clear from the evidence whether this material would have been dyed or painted but either option is plausible. It is however known that the motifs on the shields were painted as opposed to embroidered. These designs seem to have functioned as a form of unit identification. Tacitus relates to us a story in which two Roman soldiers recover the shields of two slain enemies and hiding behind them manage to infiltrate enemy lines. Although this story is from AD69 there is no reason to assume the same could not have applied in Republican times. The Roman army was extremely conservative, for example the breaking of camp was described in exactly the same way by both Polybius and Josephus despite the length of time that had elapsed between the two works.         

     I chose to decorate my shield with the winged design, a motif that has several examples in Roman artwork, for example it appears on Trajan’s column. Although the wings are often accompanied by vertical, conical spines and lightening bolts I chose to use just the wings, inspired by ancient military historian Peter Connolly’s illustration of a Republican legionary who had a scutum of this design (see fig.15).  To apply the design to the shield I created a stencil and then used it to reproduce four identical wings in each quarter of the shield (in pencil initially). To paint it I used modern acrylic fabric paint, using Naples yellow for the feathers and a dark brown for the outline. It is now known that the Dura Europas shields were in fact painted with casein, a milk derivative but unfortunately I was prohibited financially from obtaining such paints, this I consider a minor lapse in authenticity for the method was essentially the same.

    I now had to add a rim; Polybius tells us that “The upper and lower edges are bound with iron to protect the shield both from the cutting strokes of swords and from wear when resting on the ground.” However the Kasr-el-Harit shield which had been my primary source had no metal rim or boss, instead it had a raw-hide rim stitched through the wood as was the case with the shields found at Dura Europas. The absence of metal components is easily explainable as “minor alterations could have been made during the two or three centuries the shield was in service” (feugere,1993,77). For rawhide I again used the logical substitute of an old leather jacket, cutting it into thin strips that would neatly bind the edges of the shield.

     I had drilled the necessary double row of holes around the edge of the shield (with a hand-drill), carefully measuring a gap of 2cm between each in the hope this would make stitching through the wood easier. However they were already covered by the top layer of red material and I did not want to risk tearing this by probing blindly for the holes. Several tests conducted on scraps of plywood bound in leather and the red material did not boost my confidence so I opted to glue on the edging and clamp it in place whilst it dried. I feel that this was the only time I really strayed from the method of construction that the evidence dictated, but I was not prepared to risk the outcome of the shield as a whole for this relatively minor part of the construction.

© L Whiteley

[As will be obvious from the text, Leon's project was well illustrated. However, the images were not digital and it has not been possible (so far) to digitise them. TR]