The Roman legionary is often reduced to one familiar scheme: a rectangular or curved scutum, the pilum, a short gladius and segmented armour. That image is useful for the early Principate, but it does not describe the entire history of Roman heavy infantry. Between the third century BCE and the fourth century CE the shield, polearm, sword and use of greaves all changed.
This article treats equipment as a system rather than as isolated objects. When the shield changes, the role of the polearm, the length of the sword and the visual image of the soldier often change with it. Atypical images of legionaries should therefore not automatically be dismissed as artistic mistakes: some reflect a different period, a different military role, a provincial setting or a deliberate use of older artistic language.
The charts below are a working model, not final statistics. A strict publication would require a full catalogue of monuments: date, region, monument type, the figure's status, shield type, polearm, sword, greaves and confidence level of attribution. Here the charts provide an analytical frame for showing the direction of change and separating the normal pattern of a period from outliers.
It is especially important to separate source types. Reliefs and tombstones show visual convention and social image; archaeological finds show the object itself; late manuscripts and shield lists preserve a different layer of tradition. These categories should not be merged without attention to genre.
The early and middle Republican tradition did not begin with the familiar rectangular scutum. Visual and material evidence includes oval and elongated forms connected with manipular infantry and the broader Italic environment. In the late Republic and early Principate the scutum + pilum + gladius complex becomes stronger, and in the first and second centuries CE the large curved or rectangular shield becomes the most recognizable form.
From the second and third centuries the picture changes again. Visual sources more often show oval, round and lenticular shields, and the late Roman army no longer reproduces the classic Trajanic legionary image. A lenticular shield in a Principate context therefore needs separate attribution: it may belong to an auxiliary, cavalryman, praetorian, special role or a problematic artistic context.
The main line of development runs from spear tradition to the dominance of the pilum and then to the late Roman return of spears and javelins. Early Roman infantry belonged to a world in which the spear was a normal weapon for a heavy fighter. In the Republican and early imperial legion the pilum becomes the central marker: it links the throw before contact with subsequent shield-and-sword fighting.
The later Empire shifts the emphasis back toward the lancea, spiculum and related polearms. This is not a simple return to archaism but a different tactical complex: a longer sword, an oval shield, greater importance of distance pressure and more varied battlefield roles.
Sword length changed for structural reasons as well. The Republican gladius Hispaniensis was longer than later forms and fits the transitional environment of the second and first centuries BCE. In the early Principate the shorter Mainz/Pompeii type becomes prominent, working well with the large scutum and close fighting.
From the late second and third centuries the long spatha gradually moves beyond cavalry use and becomes an important part of late Roman infantry equipment. Blade length, however, cannot be measured reliably from images alone: archaeological finds are primary here, while images help explain suspension, status and visual language.
Greaves show why ordinary legionaries, centurions, cavalrymen and special images must be separated. In the early Italic and Republican environment leg protection fits naturally with spear-based fighting. In the classic Principate greaves are less common among ordinary legionaries, but they retain meaning for centurions and in special scenes as a status or functional feature.
In the second and third centuries images of a single greave reappear, often on the leg protected by the shield. This is not a simple return to the early Republic: the feature must be interpreted together with the date of the monument, the figure's role and the type of scene.
Outlying images should neither be deleted from the corpus nor merged mechanically with the main pattern. They should be coded separately. An archaic shield or spear may belong to an older artistic language; an oval or lenticular shield in a classic Principate context may belong to auxiliaries, cavalry, praetorians or a provincial tradition; a funerary monument may preserve a veteran's biographical image rather than current field equipment.
The practical rule is simple: if a figure stands next to legionaries, that alone does not prove that he is an ordinary legionary. In statistics the inscription, monument context, findspot, date, troop type and neighbouring figures all matter. Otherwise one striking relief can artificially break a pattern supported by a larger series of monuments.
The final index is not an independent historical fact. It is a calculated visualization based on the same features as the charts above: the share of curved/rectangular scuta, the share of pila, the closeness of the sword to the short gladius and the rarity of regular greaves among ordinary infantrymen. The phalanx background and manipular transition are therefore not included in the line: this version does not contain equivalent coded series for them.
The peak naturally falls in the first and early second centuries CE, when scutum, pilum, short gladius and the lack of regular paired greaves among ordinary soldiers combine into the most recognizable complex. After the second century the index falls not because the soldier becomes worse, but because the late Roman infantryman is typologically different.




Interested in Ancient Rome beyond reading? Join Legio X Fretensis or explore our reenactment directions.