19/08/2026
𝗥𝗲𝗰𝗼𝗻𝘀𝗶𝗱𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀 𝗼𝗻 𝘁𝗵𝗲 𝘀𝗰𝘂𝘁𝗮 𝗮𝗻𝗱 𝘁𝗵𝗲𝗶𝗿 𝗲𝘅𝘁𝗲𝗻𝘀𝗶𝗼𝗻
Recently, a student preparing his BA thesis contacted me to inquire about the weight of Roman shields. Much of the literature, especially the older scholarship, mentions surprisingly high weights of around 10 kg for the large shields of the Republican period. This figure has persisted over time and has influenced numerous works, often leading to exaggerated or erroneous conclusions. In contrast to this tradition, some more recent researchers have reduced the estimate to between 5 and 7 kg, although in many cases the materials used in the calculations are not specified, making it difficult to confirm their reliability.
It was in the context of this debate that my interlocutor pointed me to an article I had glanced at years ago and which I have now reread carefully, along with another that formed part of the same publication. The author examines a broad sample of Germanic and Roman umbos and observes that the fl**ge of the umbos rarely forms a 90° angle; on the contrary, it typically exceeds 5° of inclination, sometimes reaching as much as 125°. For his analysis, he discarded those umbos with evident deformations and selected only those that preserved at least three-quarters of the fl**ge, thus ensuring the reliability of the measurements.
However, upon examining his work closely, I find two main objections. The first is that the article does not specify the proportion of Roman umbos analyzed relative to Germanic ones, nor the ratio between square/rectangular and round umbos. This aspect is critical, since it determines whether the curved scuta presented -apparently- a certain lenticularity in addition to the curvature we already know, or whether, on the contrary, the sample of square/rectangular umbos with this characteristic was too limited to draw general conclusions. The second objection is more substantive: although the hypotheses proposed to dismiss the idea that umbos with inclined fl**ges were associated with flat boards have greater solidity than the mere opposite assumption, the author takes for granted that an inclined fl**ge necessarily implies a lenticular board. This equivalence is not without nuances.
The existence of lenticular shields in Roman contexts appears to be demonstrated by archaeological remains, and medieval iconography undeniably shows shields of this type. Their existence, therefore, is beyond doubt. The relevant question, rather, is whether the curved scuta shared that lenticularity or whether, on the contrary, they were curved exclusively on a single axis. This is where shield no. 15 from Dura Europos may offer some answers.
Let me recall that this shield is not truly rectangular: its upper and lower sides present two subtle arcs, so that the height at the central axis is 106 cm, while at the lateral edges it reduces to 102 cm. Following the author's hypothesis and assuming that this lenticularity was lost due to its fragmentary and buried state, converting the shield's profile into a perfect rectangle would require postulating the existence of an arc on the vertical axis, such that the height would be uniformly 102 cm across its entire width. The maximum preserved depth of the shield is 231 mm. If we accept the author's hypothesis, the resulting lenticularity would be approximately half of that depth, that is, about 115 mm on each face. This figure is not negligible and would have notable implications for the total weight and mechanical behavior of the shield, thus meriting a more detailed treatment than the author affords it.
Finally, one of the umbos analyzed by the author is the square umbo from the River Tyne. As he himself notes, this umbo exhibits a lenticularity on its vertical axis. In my own experience making umbos, I have observed that this lenticularity arises naturally from the tensions in the metal sheet during the process of forming the dome. Nevertheless, I am aware that my craft practice does not rival that of professional smiths dedicated to this matter, so I do not rule out that my inability to eliminate or avoid that deformation may simply reflect a lesser technical skill. This subjective factor obliges me to be cautious when extrapolating my empirical observations to the author's conclusions. I expressed my doubts to the author via email, but received no response.
On the other hand, the number of documented square- or rectangular- based umbos is truly scarce, with only 16 identified specimens, whereas the sample of round-fl**ged umbos is much more extensive, representing approximately 80% of the total. Roman art inevitably leads us to associate these rectangular umbos with the plywood scuta of the legions, but this association requires some nuance.
On the one hand, several of these square- or rectangular-based umbos have been found in auxiliary camps, which could indicate the use of rectangular shields -whose use we know from shield covers remains- that could be flat, or could be the same as those used by legionaries, or could even have been employed on flat shields of other geometries. Specifically, three of these rectangular umbos associated with auxiliary contexts appear to have flat bases, although it cannot be ruled out that they were not originally so.
On the other hand, another series of square or rectangular umbos are linked to legionary camps, while others lack the archaeological context necessary to determine their attribution.
In summary, although I remain skeptical of the hypothesis of lenticularity applied to rectangular scuta -a possibility that, in my view, still lacks sufficient empirical support- I do consider the association of square and rectangular umbos with both legionary and auxiliary contexts to be plausible. This contextual ambivalence opens, nevertheless, several interpretive avenues worth considering: either the rectangular shield was not as predominant among legionaries as has traditionally been assumed, or its use was not exclusive to them, and it could have been employed by auxiliary troops as well. It is even possible that legionaries, at least in certain periods or regions, resorted to other shield typologies, so that the relationship between umbo type and military unit is less direct than classical iconography suggests. In any case, the scarcity of the sample and the lack of archaeological context for many specimens preclude, for now, drawing definitive conclusions.
Let me know what you think.
Source:
1. Die Schildbewaffnung römischer Legionäre im 2. Jahrhundert n. Chr
2. Neue Gedanken zur Rekonstruktion römischer Schilde
Image: My hypothetical reconstruction in 3D render.
