Roman Plate Armour

M.C. BISHOP ILLUSTRATED BY GIUSEPPE RAVA

Elite • 247

Roman Plate Armour

M.C. BISHOP ILLUSTRATED BY GIUSEPPE RAVA

Series editors Martin Windrow & Nick Reynolds
CONTENTS
INTRODUCTION 4
Regal Period n Republican Period n Terminology

KALKRIESE-TYPE LORICA SEGMENTATA 11

History n Description n Variants

CORBRIDGE-TYPE LORICA SEGMENTATA 16

History n Description n Variants

NEWSTEAD-TYPE LORICA SEGMENTATA 24

History n Description n Variants

OTHER FORMS OF PLATE ARMOUR 31

Hybrid forms n Armguards n Muscled cuirasses n Greaves n Horse armour

MANUFACTURE AND DECORATION 48

Workshops n Manufacture n Decoration n Maintenance

PLATE ARMOUR IN USE 55

Combat n Carriage n Versatility n Who wore what? n Legacy

BIBLIOGRAPHY 62
Ancient sources n Modern sources

INDEX 64
ROMAN PLATE ARMOUR
INTRODUCTION
Roman plate armour is one of the most easily recognizable cultural identifiers
of any of the peoples in the ancient world. Whether it be the ubiquitous lorica
segmentata of the ordinary troops, or the muscled cuirasses of senior officers,
the modern observer obediently follows an iconographic agenda set nearly
2,000 years ago on a monument in the centre of Trajan’s Forum in Rome.
Trajan’s Column was constructed using the not inconsiderable proceeds
of the Emperor Trajan’s (r. ad 98–117) two wars of ad 101–02 and ad 105–
06 in Dacia (closely approximating to modern Romania) and completed
in ad 113. Its inscription merely highlighted the fact that it marked the
depth of the excavations undertaken to complete Trajan’s Forum, but the
sculpted helical frieze running anti-clockwise up its shaft left no doubt that
its real purpose was to commemorate and glorify Trajan’s trans-Danubian
conquests. Although it was long thought that ‘field sketches’ may have
been used as visual aids, it is now believed that the metropolitan sculptors
made reference to what they saw around them in Rome, where they would
of course have been familiar with the sight of troops, whether they were
the Praetorian Guard, the equites singulares Augusti, or passing bodies of
provincial legionaries and auxiliaries. It is as well to remember, however, that
just because they could observe armour, it did not necessarily mean that they
understood what they were seeing.
In order to make its visual message simple to onlookers (who could
not only view it from ground level but also from surrounding galleries in
the forum), types of troops in the opposing armies were stereotyped into
particular groups. Roman citizen troops (both legionaries and Praetorians)
were depicted wearing articulated plate cuirasses (the famed lorica
segmentata), while most non-citizen auxiliary troops wore mail to distinguish
them from their social superiors. Senior officers, usually Trajan himself and
his accompanying advisors, were shown wearing Hellenistic-style muscled
cuirasses, probably made of metal plate (see p.49 for a discussion of the
material used). Inevitably, because large numbers of citizen troops are
shown on the frieze, lorica segmentata has seeped into the Western cultural
perception of what a Roman soldier looked like, assisted by works such as
those of Piranesi or Just Lips (aka Iustus Lipsius) that referenced Trajan’s
Column. This was undoubtedly helped by the willingness of medieval and
early modern artists, followed by Hollywood, to adopt the convenient visual
shorthand it offered when producing Roman ‘sword and sandal’ epics.

4
 Citizen soldiers depicted on
Trajan’s Column working in
their lorica segmentata. From a
cast in the Museo della Civiltà
Romana. (Photo © M.C. Bishop)

The prominence of the Trajan’s Column reliefs is in many ways enhanced

by the comparative rarity of depictions of lorica segmentata in provincial
art. It may be glimpsed in one or two places, such as one of the pedestal
reliefs from the legionary headquarters building in Mainz (Germany) or,
possibly, on a relief from Saintes (France), but it is surprisingly absent
from the canon of tombstone reliefs from the 1st century ad. The bulk
of the surviving representations are on metropolitan sculpture from the
city of Rome itself, wholly or partly influenced by the helical frieze of
Trajan’s Column. Some slightly earlier, large-scale reliefs believed to be
from the Temple of the Gens Flavia on the Quirinal Hill in Rome include
a soldier wearing segmentata similar to that shown on Trajan’s Column,
with narrow, bipartite, overlapping chest plates with rivets at the end of
the shoulderguards. Contemporary with Trajan’s Column was the Great
Trajanic Frieze, which may also have originated in Trajan’s Forum but
which was ransacked mercilessly to adorn later monuments, most notably
the Arch of Constantine next to the Colosseum. The sculptors of the
Great Trajanic Frieze were clearly less constrained than those of Trajan’s
Column, since they depicted citizen troops wearing mail and scale as
well as segmental armour. Trajan’s Column had a profound influence on
other sculpture too, and not just the derivative Marcus Column erected
to commemorate Marcus Aurelius’ Marcomannic Wars (ad 166–80), as
numerous other metropolitan reliefs seem to have used it as a source in
preference to accurate observation of real armour.
From this it should be clear that the reliefs on the helical frieze of Trajan’s
Column are not a near-photographic record of the Dacian campaigns or the
equipment used on them, but have to be interpreted with great care in order
not to misunderstand what is being shown. For that reason, archaeological

5
 evidence has been vital in the accurate reconstruction of lorica segmentata
in particular. Thus it was only at the end of the 19th and beginning of the
20th centuries, when serious archaeological excavations on Roman military
sites began to recover examples of this type of armour – notably on the
legionary fortress at Carnuntum (near Bad Deutsch-Altenburg, Austria) and
the fort of Newstead (Scotland) – that progress could be made. Carnuntum
was excavated by an Austro-Hungarian artillery officer, Maximillian von
Groller-Mildensee (usually known by modern scholars just as Groller) and
that work uncovered a rampart-back building (possibly an armamentarium)
full of Roman military equipment of various types (Groller 1901). One room
included numerous fragments of lorica segmentata. Indeed, there was so much
ferrous corrosion product that voids in it preserved the outline of shelving
upon which the armour and other material had evidently been stored. Groller
devoted part of his report to these finds and included an analysis of how he
thought the segmental body armour worked. Influenced by the reliefs on the
helical frieze of Trajan’s Column, he decided that it had been fixed to a leather
jerkin of some kind in order to make it work in the way illustrated on Trajan’s
Column (Groller 1901: 98). These attempts to understand articulated plate
cuirasses using the reliefs as their starting point were inevitably doomed to
dismal failure, however. Even so, scholars persisted in using the reliefs on the
helical frieze of Trajan’s Column as a guide to how the armour worked, usually
with unfortunate results (Webster 1960).
It was not until the discovery of elements of a number of segmental
cuirasses at Corbridge (England) in 1964 that a complete, working
reconstruction of two principal variants of lorica segmentata became feasible.
It was undoubtedly helped by the fact that the armourer involved – Henry
Groller’s sketches of (above) Russell Robinson – was a specialist on articulated oriental armour. The
lorica segmentata on Trajan’s Corbridge discovery also made it possible for Robinson to reinterpret the
Column and (below) how
he thought the pieces he
Newstead fragments, using his newly acquired understanding of the structure
found at Carnuntum should of this type of armour. The beginning of the 21st century saw a number of
be reconstructed. (Author’s finds that elucidated the workings of the articulated plate armguard (manica)
Collection) in much the same way.
Unfortunately, no such key finds have as yet been made for Roman muscled
cuirasses. Although there is a considerable amount of representational
evidence in the form of reliefs, statues, and even wall paintings, these are
almost invariably over-simplified in much the same way as the reliefs on
Trajan’s Column. Here, the problems encountered reconstructing lorica
segmentata from sculptural evidence alone should caution against taking
what little can be derived too literally or dogmatically.

Regal Period
We know next to nothing about early Roman armour. This bald and rather
negative statement can to some extent be mitigated by pointing out that quite
a lot is known about contemporary armour from elsewhere on the Italian
peninsula (Burns 2005). A panoply of arms and armour from a burial at
Lanuvium, about 30km south-east of Rome and dating to the first half of the
5th century bc, represents the equipment of an early Roman foe and includes
a copper-alloy muscled cuirass and helmet. It is not unreasonable to assume
that the Romans were extremely familiar with, and probably used, such
defences. This does not alter the fact that there is no surviving Roman plate
armour from the early period, however, and there is no way of knowing if it
differed substantially from that of the Romans’ neighbours.

6
 A warrior’s panoply, including
a muscled cuirass, from a
burial at Lanuvium, some
30km south-east of Rome,
and dating to the first half of
the 5th century bc. (Ursus/
Wikimedia/Public Domain)

Republican Period
The earliest surviving examples of plate body armour that can be attributed
to the Roman Army of the Republican period were the circular pectoralia
found in the 2nd century bc camps around the hill town of Numantia
(Spain). A pectorale (often now referred to as a kardiophylax, following
the Greek historian Polybios, who described it, but who has sometimes
been misinterpreted as saying it was square) was a copper-alloy disc with
concentric corrugations that was worn, as its name implied, over the heart
of an infantryman. As such it was definitely at the economical end of the
armour compromise. The discs were surrounded by a number of small holes,
used to attach strap fittings (some of which still survive) to hold them in
place, but the rest of the apertures may have been employed to rivet some
sort of padded textile or leather backing in place, similar to that found on
Italian triple-disc breastplates (Burns 2005: 57).
Polybios, himself an eyewitness to the mid-2nd-century bc Roman Army
in action, described the use of this type of armour:

The common soldiers wear in addition a breastplate of copper alloy

a span [228mm] in diameter, which they place in front of the chest
and call the heart-protector (kardiophylax), this completing their
accoutrements; but those who are rated above ten thousand drachmas

7
RIGHT
A copper-alloy pectorale or
kardiophylax from Room P in
Camp II at Castillejo, the Roman
base near Numantia (Spain),
retaining two of its attachment
plates to which straps were
fixed. (Photo © Römisch-
Germanisches Zentralmuseum,
Mainz/S. Steidl)

FAR RIGHT
Engraving from Iustus Lipsius’
(Just Lips’) volume De Militia
Romana showing the first
published use of the term
loricae segmentatae. (Author’s
Collection)
wear instead of this a cuirass (thorax) of mail. The principes and triarii
are armed in the same manner except that instead of the pila (hyssos)
the triarii carry long spears (doru). (Polybios, Histories 6.23.14)

Here Polybios introduced an additional factor into the armour equation:

cost. A small disc of copper alloy took far less time to produce – and
therefore cost less – than a coat of mail. A disc found in the so-called Camp
of Marcellus at Numantia measured 170mm in diameter, while another
example was 175mm.
The technological leap from simple plate armour like this to the
overlapping articulated plates of lorica segmentata is difficult to explain
on the currently available evidence. Articulated armguards were certainly
known in the Hellenistic period (see p.34), but the suspicion that there may
be a Late Republican origin for full cuirasses is perhaps fuelled by pieces
of the Kalkriese variant of lorica segmentata from Dangstetten (Germany),
datable to 9 bc at the latest (see p.11).

Terminology
The Romans termed all forms of body armour ‘lorica’, qualifying the noun
(when it suited them) with an adjective, ‘hamata’ for mail and ‘squamata’
for scale, but what they called segmental armour remains a mystery. The
term ‘lorica segmentata’ was coined during the Renaissance when scholars
began to study monuments – particularly Trajan’s Column – in some
detail. The earliest attested use of this neologism is in Just Lips’ 1596
work De Militia Romana, which was a commentary on the Histories of

THE PRINCIPAL TYPES OF LORICA SEGMENTATA from the earliest to the latest form of this type of armour
A The three main types of lorica segmentata cuirass can be actually was.
seen here, both in their component parts and complete. All Assuming six girth hoops for each set of armour, the
three types share certain common elements: they are component count for the various types differs slightly, with
constructed as four units: two upper and two lower, with the Kalkriese type having 30 ferrous plates and 160 copper-
hinged shoulderguards and collar plates, as well as the alloy components, while the Corbridge type had 38 and 172
overlapping girth hoops with the lowest pair of hoops left respectively, and the Newstead type 34 and 174. These
unfastened. The Kalkriese (1), Corbridge types A (2) and B/C numbers were only ever approximate (and standardized on
(3) and Newstead (4) types are shown alongside a speculative seven girth hoops apiece) once modifications and repairs,
reconstruction of the hybrid Alba Iulia type (5). Such a side- such as attaching two fittings with one rivet, are taken into
by-side comparison demonstrates how natural the evolution consideration.

8
 1
2
3

4
5

9
The terminology used in Polybios and was heavily influenced by the reliefs on the helical frieze of
describing the components of Trajan’s Column. Just Lips used the term (both in his text and to caption
lorica segmentata (from Bishop
2002). (Drawing © M.C. Bishop)
an illustration) without explanation, however, suggesting that he was not
the first to do so:

Lorica segmentata
Finally, on another subject, what was frequently found in the time
of Statius and Silius? The body was encircled with pieces of iron,
arranged like bands one over the other. I confess that I have not read
of this anywhere: but on the column of Trajan this type [is used]
constantly and almost exclusively for Roman soldiers. (Lipsius 1630,
132 – Liber III Dialog. vi, tr. author)

The true origins of the term probably lie soon after the study of Trajan’s
Column began in the early 16th century with Jacopo Ripanda being lowered
down from its balcony to sketch the reliefs.
The Greeks used the word θώραξ (thorax) to mean any sort of cuirass, but
when the Romans transliterated and borrowed it (as the Latin substantive
thorax), they seem exclusively to have meant a muscled cuirass. The coinage
‘lorica musculata’ is modern and was invented (so far as it is possible to
be sure) by re-enactors. It would probably have been incomprehensible to
Romans, however, since the Latin words for mouse, mussel and muscle all
share a common etymology.

10
KALKRIESE-TYPE
LORICA SEGMENTATA
In 2018, after many years of both smaller and larger components being discovered
at Kalkriese (Germany), the most complete find ever of a single set of segmental
body armour was made during a new campaign of excavations (Crossland 2020).
This site has long been associated with the clades Variana, when the Roman
commander Quinctilius Varus lost three legions during a rebellion led by native
German troops serving as his auxiliaries (McNally 2011). As such, the date of
that disaster, ad 9, has inevitably been attached to the site, although the story was
undoubtedly complicated by looting by the German tribes and a subsequent visit
in ad 15 by a force led by Germanicus (Tacitus, Ann. 1.61–62).

History
The first form of Roman segmental body armour for which evidence has
been found is the Kalkriese-type lorica segmentata (Bishop 2002: 23–29).
Earlier examples exist than those found at Kalkriese, however. Once it was
realized that the Kalkriese finds belonged to this early form, it was possible
to identify pieces of the same type from the base at Dangstetten, which was
constructed around 15 bc and abandoned by 9 bc, two decades before the
earliest possible date for the Kalkriese material. The Dangstetten fragments
clearly hint that segmented body armour was in use before 9 bc, but the date
of its introduction is, for the time being, unknown, although it was clearly
prior to 9 bc and possibly even before 15 bc.
The Kalkriese form of lorica segmentata remained in use until sometime
after the Claudian invasion of Britain in ad 43. Characteristic components
have been identified from a few Roman military sites in southern Britain,
including Chichester and Waddon Hill (both in England) (Thomas 2003: 63),
both of these associated with the campaigning of legio II Augusta under its
commander, Flavius Vespasianus. None are known from Flavian sites in the
north of Britain, however, so the Kalkriese type seems to have been completely
phased out by the second half of the 1st century ad. Another possible late

The first Kalkriese-type

lorica segmentata upper
shoulderguard component,
excavated at Kalkriese itself,
showing the inward-curving,
almost hooked end of the
plate, edged with copper-alloy
piping. (Photo © Varusschlacht
im Osnabrücker Land)

11
 example of this type may be a top backplate from Rißtissen (Germany) re-
BELOW
used as part of a Corbridge-type cuirass. Fittings or plates from this early
Kalkriese-type lorica form of cuirass are also known from the legionary base at Vindonissa
segmentata fittings from (Windisch), Kaiseraugst and Oberwinterthur (all in Switzerland), Dangstetten
Kalkriese (1–2, 6–11), and Kempten (Germany), Nijmegen (Netherlands), Magdalensberg (Austria),
Strasbourg (3 & 5), Chichester
(4), Vindonissa (12) and Hod
Astorga and Iruña (both in Spain) and Novi Banovci (Serbia). Finds from
Hill (13). (Drawings © M.C. Strasbourg (France) are especially interesting, given that it was the base for
Bishop) legio II Augusta before it moved to Britain in ad 43.

12
Description OPPOSITE
There were four basic units that made up segmental body armour and this Left top backplate, mid-
collar plate and breastplate
was to remain true of all the subsequent main forms. There was a shoulder from the complete lorica
section in left and right halves and a torso section, also in left and right halves. segmentata found during
Each of these four elements was riveted to its own internal leather harness. the 2018 excavations at
A series of ferrous collar plates – breastplate, mid-collar plate and top Kalkriese. (Photo Hermann
Pentermann, © Varusschlacht
backplate – surrounded both sides of the neck and were joined together by
im Osnabrücker Land)
copper-alloy hinges riveted to the plates at the points where they overlapped
slightly. The inner edges of these plates, next to the wearer, were turned out.
The backplate was shorter than the breastplate, so two additional plates were
hung below it, joined by internal leather straps, to provide the same depth of
protection at the back as at the front. The leather straps were riveted to the
interior faces of the plates by means of copper-alloy rivets which passed through
the plate and leather and which were then peened over (sometimes using internal

The first Kalkriese-type lorica

segmentata breastplate,
showing how the mineralized
leather straps of the horizontal
and vertical fasteners were
riveted directly to the ferrous
plate. Whether these are the
original fittings or a repair is
unclear. (Photo © Varusschlacht
im Osnabrücker Land)

13
 roves). Above these, and attached by means of internal leather straps with
rivets, were three upper shoulderguards: front, middle and back. As with the
collar plates, these overlapped slightly and were joined together by hinges. The
front and back upper shoulderguards curved inwards towards the centre of the
thorax. The centre plate broadened and curved downwards over the shoulder
slightly. These were the only forms of shoulder protection and the upper arms
were left unprotected, just as they were with mail or scale cuirasses. The left
and right upper halves were joined together using single buckles on each of the
breastplates and backplates. There is clear evidence from finds from Kalkriese
itself that some of the copper-alloy fittings were tinned or silvered.
The abdomen was protected by a series of overlapping ferrous bands,
now usually known as girth or girdle hoops. The number of these – usually
between five and seven – seems to have varied not as a distinguishing mark of
different variants as was once thought (Robinson 1975: 177), but rather to
provide the best fit for the individual. These bands overlapped top to bottom
and were joined together by means of three internal straps for each half, with
straps riveted to the plates at the front, back and side. The two halves were
joined together by means of a series of external buckles, both front and back,
set against the lower edge of each plate.
The top and bottom sections of the cuirass were fastened together
with external buckles, one for each breastplate and the lowest of the three
backplates on either side.
Many of the ferrous plates that made up a Kalkriese cuirass were edged
with copper-alloy piping in much the same way that contemporary Roman
helmets were. Although it provided a decorative touch to the armour,
this was in fact a simple way of avoiding having to finish the edges of the
plates too carefully, since it was far less time-consuming just to apply such
piping than to finish the edges. The piping was placed in areas which could
come into contact with (and possibly damage) the wearer’s underlying
garments. The copper rivets which attached the internal leather straps to
the ferrous plates on the upper (but not the lower) sections might also have
been supplied with decorative washers, once again matching those used on
contemporary helmets.
No actual organic leather has survived on any of the recovered
examples, but examples of mineralized leather straps have been found.
When large amounts of ferrous material are deposited in the right ground
conditions, the corrosion products seep into the cells of any attached
organic material over time and thus ‘fossilize’ them as they rot in a process
known as mineralization.

SEGMENTAL ARMOUR IN USE AT KALKRIESE strap fittings. A third is a hybrid, however, with some of the
B Although lorica segmentata was not the only form of body hinged fittings replaced by straps riveted directly to the
armour used during the sporadic fighting with German armour plate. Yet another is mostly of the first kind but with
warriors in ad 9, it had become important to the legions of one upper unit taken from the second. Cuirasses may
Germany even before Varus led three of them beyond the ultimately have acquired additions by their owners in the form
Rhine. Several different variants can be seen at the same time of additional plates over the tops of the arms. This sort of field
in this scene of an attack on a Roman temporary camp during modification would eventually become incorporated into the
the long march through hostile territory, as well as the more later Corbridge type. The method of fastening the girth hoops
familiar mail (which covers more of the lower body than the would also ultimately change, and it is possible that the
segmental cuirasses). Two basic forms can be seen here, but original hinged buckles would have been removed and the
there is also clear evidence of repairs and even cannibalization. rivet holes used to lace plates together with leather thongs
The first form uses simple rectangular fittings, while the (many of the soldiers can be seen with spares tied to their
second has more elaborately decorated hinged buckles and cuirasses for just such eventualities).

14
15
 Variants
An interesting range of variants of the Kalkriese form can be identified
(types A and B – Bishop 2002: 23). One of the first pieces found, a left-
hand breastplate, had both horizontal and vertical fastening straps riveted
directly onto the ferrous plate with large, flat-headed rivets. Whether this
was a crude repair to a piece that originally had copper-alloy fittings, or
whether it pre-dated such embellishments is impossible to tell. The complete
set of armour found in 2018 used simple hinged, rectangular buckle plates
to attach buckles to their respective armour plates and rectangular hinges
to join the shoulderguards and collar plates. More elaborate hinged buckle
plates with serrated edges have also been found at that same site, however,
and it is these, together with the characteristic sub-lobate hinges that joined
collar and shoulderguard plates together, that have been found in Britain.
It is tempting to see a development from simple to complex fittings, but
in reality all may have been in use at the same time and these variations
Plan of the Corbridge Hoard may just reflect the work of different armourers or even the handiwork
made during the excavation of different legions. Equally, shoulder hinges seem to have presented a
in 1964, with north towards problem from the very beginning: the number of rivets was increased from
the top. The lid (in grey) was
originally thought to be
three on each hinge half to four (and was to be increased again with the
floorboards overlying the chest. Corbridge type).
(Photo © M.C. Bishop)

CORBRIDGE-TYPE
LORICA SEGMENTATA
A full understanding of how lorica
segmentata functioned only became possible
after a find during training excavations at
the Roman site of Corbridge, just 2km
south of Hadrian’s Wall (Robinson 1975:
174–80, Figs 178 & 180). In 1964, the
remains of a wooden box packed full with
a wide range of items was discovered and
lifted in a block for subsequent examination
in a museum laboratory. The finds from
the Corbridge Hoard chest included six
sets of lower halves, and six of upper
halves, of lorica segmentata, seemingly
collapsed for ease of storage. It was this
find that enabled Charles Daniels and
Henry Russell Robinson finally to solve the
mystery of how segmental body armour
was put together (Daniels in Allason-Jones
& Bishop 1988: 97–100).
The Corbridge type incorporated
a number of major changes over the
Kalkriese type, all of which seem to have
been improvements based on experience.
First, the method of fastening the girth
hoops, using buckles and straps (with the

16
hinged buckles and straps contributing up
to 48 moving parts), was simplified into
the new pairs of static tie loops. Second,
the method of attaching the upper and
lower assemblies changed at the rear
from one external to double internal
buckle-and-strap pairs. Third, the lesser
shoulderguards were added to protect
the tops of the shoulders, the upper
shoulderguards being accordingly reduced
in width and straightened out. Fourth, the
collar plates, including the breastplate and
backplates, were reduced in width. Fifth,
the number of rivets attaching hinges
to a plate on the upper shoulderguard
was increased to five for each half,
leading to the characteristic lobate shape
that is nowadays so familiar for lorica
segmentata. Finally, all copper-alloy
piping on the edges of plates was done
away with, the edges being rolled instead,
while the corners of plates were clipped.

History
The earliest finds of Corbridge-type
lorica segmentata have been made at
Magdalensberg (Austria) and date to the
early Claudian period, when the Roman
Army seems to have abandoned the site
(Thomas 2003: 91). Its use was certainly
common by the time of the invasion
of Britain, some of the earliest British
finds coming from the Roman military
occupation of the native dyke system
at Colchester Sheepen (England) (e.g.
Thomas 2003: 82). It then became the
principal form of this armour throughout
the rest of the 1st century ad and into the 2nd. Excluding the Corbridge Right-hand lorica segmentata
Hoard material, major finds of armour of this variant include a set of breastplate from the Bank
of England site, next to the
girth hoops from St Albans (England), most of an upper assembly from Walbrook in London, showing
Plantation Place in London (along with a cut-down and extremely well damage received to the plate,
preserved breastplate from the Bank of England site), much of the cuirass which has subsequently been
of a legionary thought to have been killed during the attack on Gamla scrapped and cut down. (Photo
© M.C. Bishop)
(Israel), as well as a collection of components from Rißtissen which, had
Robinson seen them when they were first found, might have helped him
solve the puzzle of segmental armour much earlier. Significant assemblages
of segmentata plates and fittings are also known from Richborough
(England), Caernarfon (Wales) and Svishtov (Novae, Bulgaria).
Finds of the Corbridge form are the most common and widespread,
and the available dating evidence indicates that it continued in use until
at least the latter part of the 3rd century ad in some places (Aurrecoechea
2003/04: 52).

17
 The Corbridge Hoard
In 1964, a training excavation at the Roman site of Corbridge in Northumberland (England),
just 2km south of Hadrian’s Wall, made a spectacular discovery while examining the remains
of a 2nd-century ad fort: what became known as the Corbridge Hoard. The remains of an
iron-bound wooden chest containing a wide variety of items were located and carefully
lifted in a block by the excavation supervisors (a process which had to be completed under
car headlights) for later examination in laboratory conditions. The bulk of the material in the
box had been iron, although all that remained were corrosion products. The process of
mineralization (see p.14) had preserved a range of organic materials, including wood,
leather, textile, bone, feathers, and even scraps of papyrus. The items in the chest had been
carefully packed and it is thought that it was intended for departing troops to take it away
with them and that, for whatever reason, it ended up being left behind and buried,
presumably to deny the raw materials to any enemy.
Among the contents, and carefully wrapped in cloth, were the components of the lorica
segmentata which would enable the director of the excavations, Charles Daniels, and the
armourer Henry Russell Robinson, to reconstruct for the first time exactly how this type of
armour had functioned – even using cardboard templates cut out of breakfast cereal
packets in order to understand the finer points. The armour, which had been collapsed and
compressed to take up as little room as possible when packed, included six upper (three left
and three right) and six lower sub-units (again, three left and three right). These very
roughly corresponded to enough sub-units to make two Corbridge type A/B and one type
C cuirasses, although if they had been put together they would have looked rather
mismatched. All of the sub-units showed signs of damage and repair, sometimes repeatedly
carried out.
The limited available evidence indicated that the wooden chest was buried in the first
half of the 2nd century and probably towards the end of the reign of the Emperor Hadrian
(r. ad 117–38).

Description
As with the Kalkriese types, the Corbridge form of segmental body armour
consisted of two upper and two lower halves, each riveted to its own internal
leather harness.
Once again, a ferrous breastplate, mid-collar plate and top backplate
encompassed both sides of the neck, but were now joined together by
elaborate lobate copper-alloy hinges, which were riveted to the plates at
OPPOSITE the point of overlap with five rivets on each half. Those edges adjacent to
Corbridge-type lorica the wearer were turned out once again but, in the absence of copper-alloy
segmentata fittings, including piping, also rolled inwards to form a rounded edge. The backplate was
lobate hinges from Sheepen
still shorter than the breastplate and had two additional plates beneath
(1), Rheingönheim (2), Chester
(3), Hofheim (4) and Oberstimm it, once more joined by internal leather straps. Above the collar plates,
(5); decorated washers from and attached to them with internal leather straps, there were three upper
Silchester (6), Rheingönheim shoulderguards at the front, middle and back, which also overlapped
(7), Longthorpe (8), Chichester
slightly and were joined together by lobate hinges. The shape of these
(9) and Chester (10); hinged
buckles and hinged strap shoulderguards was very different from that of their predecessors,
fittings from Sheepen (11), however. The middle plate was narrower and the front and back plates
Carnuntum (12), Chichester were straight-sided, not curving as on the Kalkriese form. In addition,
(13), Oberstimm (14 & 16), four new lesser shoulderguards were added on either side, attached to
The Lunt (15), Rheingönheim
(17, 20 & 22), Broxtowe (18),
the upper shoulderguards by internal straps, providing protection for
Aislingen (19) and Vindonissa the upper arms. The left and right upper halves continued to be joined
(21); and tie loops from Hod together using single hinged buckles on the breastplate and each of
Hill (23 & 25), Carnuntum the backplates.
(24), Rißtissen (26), The Lunt
(27), Rheingönheim (28) and
The abdomen was again protected by a series of overlapping girth hoops,
Corbridge (29). (Drawings between five and seven in number. These bands also overlapped top to
© M.C. Bishop) bottom and were joined together by means of three internal straps for each

18
19
 half, with straps riveted to the plates at the front, back and each side. The
two halves were now fastened together by means of a series of tie loops,
both front and back, which were riveted to the underlying ferrous plates. As
with the Kalkriese form, these tie loops were set flush with the lower edge
of each plate but, unlike the earlier type, the lowest two hoops were left
without fasteners. The other fasteners were used to secure the lower sub-
units with knotted leather laces and vastly reduced the number of moving
parts employed over the earlier type.
The top and bottom sections of the cuirass were still fastened together
with strap-and-buckle combinations in the Corbridge type A variant, one
external on each breastplate, but now with two internal ferrous examples
on each of the lowest of the three backplates on either side. The B/C variant
introduced a new means of attaching the top half of the cuirass to the lower:
a metal hook-and-eye system (copper alloy on the B, ferrous on the C) with
two at the front and four at the back.
None of the ferrous plates of a Corbridge cuirass were edged with
copper-alloy piping, suggesting that time was now taken to finish the edges
of the plates, but in reality these were often just rolled, turned out, or both.
The corners of ferrous plates tended to be clipped to help blunt the point.
The copper rivets attaching the internal leather straps to the ferrous plates
invariably included decorative washers of a similar type to those found on
contemporary helmets.
The mineralization of the organic components allowed a complete
reconstruction of the leathering regime that acted as a flexible matrix for
the armour. It was clear how copper-alloy rivets passed through ferrous
plates from the front, through the leather, and were then peened over,
sometimes (but not always) through square or rectangular roves on the
inside. Repairs were evident and in one place two separate internal leather
straps on the three backplates of an upper section had been replaced by
one large patch.

Variants
It was Robinson who devised the A/B/C naming system for the Corbridge
form of the armour. The type A variant was characterized by external,
hinged, copper-alloy buckles joining the upper and lower sections at the
front, accompanied by internal ferrous buckles at the back. The upper

CLEANING CORBRIDGE-TYPE ARMOUR

C Part of a Roman legionary’s life was always spent in strapping and oil the leather to keep it supple and resilient;
maintaining his arms and armour, whether in his quarters any that was damaged had to be replaced. Rather than
in a legionary fortress, or out on campaign. Sitting on the work on a complete cuirass, each soldier breaks his armour
veranda of the barracks at Isca Silurum affords these men down into its four constituent units which he rests upon his
of legio II Augusta the chance to spend some time looking knee, one piece at a time.
after their kit after a recent patrol. A legionary had to be This is not the only routine maintenance being
fastidious in how he cleaned both the metal and leather undertaken, since helmets had to be cleaned just as carefully,
components, since moisture – whether from rain or sweat swords sharpened with a hone, and boots checked for any
– was the enemy of an effective set of armour. He not only broken straps or missing hobnails – and it is clear from the
had to remove corrosion from the surface of the ferrous archaeological record that Roman soldiers shed hobnails in
plates as well as from the brass fittings, but also clean alarming amounts. A painstaking inspection of the
carefully around the vulnerable points where those fittings contubernium leather tent now, in barracks, could save the
made contact with the ferrous plate. That was not the end need for later unwelcome repairs in the field. Any items in
of the job, however, as it was then necessary to check every need of more serious repair would be set aside for later
rivet holding the metal plates to the internal leather attention from the legionary workshop.

20
21
 shoulderguards were of a uniform width from front to back. The B/C variants
differed in having broader middle upper shoulderguards, which swelled to
a point. They also used a different means of attaching the upper and lower
sections: instead of vulnerable buckle-and-strap junctions they employed
simpler, external, hook-and-eye fittings. Robinson distinguished type B
from C by means of these fittings, the former made of copper alloy and the
latter ferrous, but in reality the truly significant difference lay between the
strap-fastened A and hook-and-eye B/C, particularly in light of subsequent
developments with the Newstead type.
One unusual adaptation was found on the remains of a cuirass from
Gamla (Stiebel 2014: 67), for which the internal leathering joining the
three backplates on each side was replaced (on one side only) by a system
of sliding rivets of a kind that would not be seen again until its use in
medieval plate armour. This is the only known example of this variant,
so it may just have been a local variation (like the large leather patch,
see p.20) rather than a separate type; it was certainly a replacement for
the original leather straps since their rivets were still in situ. An example
of an upper shoulderguard found at Chichester, and associated with the
early campaigns of legio II Augusta soon after the invasion of Britain
in ad 43, demonstrated the use of double rivets to secure the internal
leathers, while an upper backplate from the same site had oval roves for
the leathering rivets.

A reconstructed set of
Corbridge-type lorica
segmentata broken down
into its four component units
with the girth hoops stacked
inside the shoulder sections
in much the same way as in
the Corbridge Hoard. (Photo
© M.C. Bishop)

22
View of the internal leathering
of a reconstructed Corbridge-
type lorica segmentata shoulder
section with the rivets peened
over square roves. (Photo
© M.C. Bishop)

Interior face of Corbridge-type

lorica segmentata backplates
from Gamla showing internal
leathers replaced by sliding
rivets. (Drawing © M.C. Bishop)

23
 NEWSTEAD-TYPE
LORICA SEGMENTATA
This, the third major form of articulated cuirass, is named from the site in
lowland Scotland where a significant collection of parts was found at the
beginning of the 20th century (Curle 1911: 156–58, Fig. 11 & Pl. XXII).
Although early attempts were made to reconstruct this form of armour, it was
not until the Corbridge discovery was made that it was finally understood
(Robinson 1975: 180–82, Fig. 181). Even then, some details were missing,
and it only became clear from subsequent finds that this form also used
lobate hinges, albeit much larger ones than earlier examples (some examples
were already known from the Carnuntum find but were not recognized as
having come from a different form).
Just as the Corbridge type incorporated a number of improvements
over its predecessor, so the makers of the Newstead form sought to address
issues which the Corbridge type had in turn introduced. Chief among
these (although not one Robinson recognized, because the Newstead find
had no such pieces) was the increase in size of the lobate hinges on the
upper shoulderguards and collar plates, presumably to counter the twisting
moment between the plates by making the hinge itself broader. The second
major improvement was the replacement of the triple backplates with one
large plate, corresponding to the breastplate (which was also increased in
size compared with its predecessor). Third, a new method of fastening both
breastplates and the new backplates was introduced. Finally, the troublesome
nature of the girth hoop fastenings required attention again, so the pairs of

The Newstead well

Like most headquarters buildings (principia) in Roman forts, the stone fort of Trimontium at
Newstead included a well in its courtyard. It is often suggested that this had a ritual purpose,
so it is unsurprising that the fill of such wells is sometimes viewed as having a special
significance by archaeologists. Excavated in September 1905, the well at Newstead
(designated Pit I by its excavator, James Curle) was 7.8m deep, and 6.1m in diameter at the top
and almost 2m at the bottom. It contained a variety of interesting items in its fill (Curle 1911:
47–48, 116–17), leaving little room for doubt that it had been deliberately backfilled at the
time of abandonment.
Near the top was a large deposit of rubble (including a fragment of an inscription),
presumably designed to seal off the material that lay beneath. There was more stone
throughout the fill of the well. At a depth of 1.5m, some fragments of jewellery. At 3.66m, a
stone altar to Jupiter and a coin of Hadrian were found, with animal bones and leather,
including shoes, at 4.27m. At 5.49m, there was a stone moulding, along with sherds of
amphorae and Samian ware. At 6.4m there was an iron bar, while at 6.7m, there were two
human skulls, brass scale armour, and more sherds of amphorae and Samian ware.
At the bottom (and so the first material to be thrown in), besides the eponymous lorica
segmentata, there was the upper stone of a quern, two iron knives, a linchpin, the staves and
bottom of an oak bucket, the iron rim of a large bucket, a large block of sandstone bearing
a relief of a boar (the emblem of legio XX Valeria Victrix), five iron arrowheads, some mail
armour, an iron shield boss, fragments of brass, a coin of Vespasian or Titus, a holdfast of
iron, a fragment of wall plaster, and amphora sherds.
The presence of arms and armour in the fill of the principia well has been used to argue
that the rooms surrounding the courtyard served as a magazine (armamentarium), but since
the other material accompanying the weaponry must have been derived from other areas of
the fort, this is unconvincing.

24
 Diagrammatic section of
Newstead Pit I, the principia
courtyard well, showing
the location of the lorica
segmentata. (Drawing ©
M.C. Bishop)

tie loops were abandoned and yet another system was introduced, this time
incorporating cast copper-alloy loops passing through the opposing plates
and thereby minimizing movement between the halves. To judge from the
numbers of cast loops that appear in the archaeological record, this last
innovation brought its own problems.

History
Although the find from Newstead itself can be dated to soon after the middle
of the 2nd century ad (the site was abandoned by around ad 180), there
is good reason to suspect that this type of armour was being introduced
towards the end of the 1st century ad. The Carlisle Millennium backplate
came from a Hadrianic deposit within the Roman fort at Carlisle (England),
but a large lobate hinge from the timber legionary fortress level of Ulpia
Traiana Sarmizegetusa (Romania) is probably Trajanic (Băeştean & Barbu
2015: Pl. IV.1). It became the dominant form of lorica segmentata by the
second half of the 2nd century ad, a fact attested by finds of this form from
Eining (Germany), Newstead itself, León (Spain), Carlisle (in two strikingly

25
The components of the
original Newstead-type lorica
segmentata found in the
principia well (Pit I) at Newstead
itself. (Drawings © M.C. Bishop)

different contexts), Caerleon (Wales), and at Stillfried (Austria), as well as

from the former cavalry fort (later regarrisoned with infantry) at Aalen
(Germany). It continued in use in some areas into the 4th century ad, as the
discoveries from Carlisle (Caruana 1993) and León (Aurrecoechea 2003/04:
52–53) have revealed.

NEWSTEAD ARMOUR AT THE TYPE SITE segmentata fragments were found. The armour was obviously
D In the latter part of the 2nd century ad, the fort of Trimontium in need of repair and was probably being kept for recycling or
at Newstead was abandoned by the Romans, sometime after cannibalization purposes. In this reconstruction, the process
the retreat from the Antonine Wall further to the north. The of filling the well has just begun, while in the background the
process of dismantling a fort was particularly thorough and demolition of the building is already under way. Since not all
quite distinct from any damage an enemy attack might cause, of the material found in the well would have belonged in the
with buildings systematically dismantled and any resources principia, it must have been brought from nearby and that will
that might be of use to an opponent rendered inaccessible. In have affected the order in which items were discarded.
the case of Newstead, this involved dumping material into pits Legionaries are using the wicker baskets they normally
and wells, including the well in the corner of the courtyard of employ for removing soil from excavations, such as when
the headquarters building, which was where the lorica ditch digging, to carry the material that requires disposal.

26
27
ABOVE LEFT Description
A Newstead-type backplate The Newstead type of lorica segmentata followed its predecessors
dating to the Hadrianic period
from the Carlisle Millennium
by retaining the two upper and two lower halves. As with the other
excavations. (Photo © Pascal forms, each of these elements was riveted to an internal leather harness.
Lemaire) Substantial differences were introduced in the detail of the cuirass,
however.
ABOVE RIGHT There were ferrous breastplates and mid-collar plates as before, but
Detail of a reconstructed now the earlier three backplates were replaced by a single, large backplate
Newstead-type cuirass showing
the backplates fastened using
that now matched the breastplate. The three collar plates were still joined
two turnkeys and a vertical together by copper-alloy lobate hinges, but they were now much larger
cotter pin, attached to the than before. As before, there were three upper shoulderguards at the
turned neck by a thong. (Photo front, middle and back, which overlapped and were joined together by
© Arik Greenberg)
the new, larger form of lobate hinge. These shoulderguards were situated
above the collar plates and were once again attached to them with internal
leather straps. Although no upper shoulderguards have as yet been found,
the new large hinges supply a minimum width for them and the simplest
reconstruction sees the larger Corbridge plates (with a central point)
continued. Four lesser shoulderguards were added on either side, as before,
and were attached to the upper shoulderguards by internal straps. A major
change was the manner in which the left and right upper halves were joined
together. This was now accomplished by means of one turnkey on the front
plate and two on the rear.
The abdomen was again protected by a series of girth hoops, between
five and seven in number. These bands also overlapped top to bottom and
were joined together by means of three internal straps for each half, with
straps riveted to the plates at the front, back and sides. The means by which

28
the two halves were fastened together changed again, now using cast rings Newstead lorica segmentata
attached through the ferrous plate with shanks which were peened over to fittings, including lobate hinges
from Carlisle (1), Carnuntum (2),
the rear, with one of the plates having an aperture to allow the lower loop to León (3) and Sarmizegetusa (4);
pass through (presumably still fastened with knotted leather laces, although vertical fasteners from Carlisle
split pins on a thong would also have worked). To judge from the number (5) and Iža (6 & 7); girth hoops
of examples found in the archaeological record, the tie rings seem to have from León (8), Iža (9 & 13) and
Carnuntum (14); and girth
been very vulnerable and prone to becoming detached. The ends of the plates
hoop tie rings from Carlisle
with the apertures were either faced with copper-alloy sheathing riveted to (10), Newstead (11) and Iža
the front or had a smaller surround to the opening of this metal. Overall, (12). (Drawings © M.C. Bishop)
this new structure made for a more rigid hoop than had previously been

29
Reconstructed Newstead-type
cuirass with the front open
showing the larger lobate
hinges, vertical fasteners, the
turnkey-and-slot horizontal
fastener on the breastplate,
and the loop-and-slot girth
plate fasteners. (Photo © Arik
Greenberg)

the case and mounting the fastening centrally was probably intended as an
The Stillfried lorica segmentata improvement over the Kalkriese and Corbridge fastenings on the lower edge.
girth hoops on display in
the site museum. (Wolfgang
The top and bottom sections of the cuirass were once again fastened
Sauber/Wikimedia/ together with a similar metal hook-and-eye system to that found on the
CC BY-SA 4.0) Corbridge B/C variant, with one on either side of the breastplates and two on
the backplates. One unusual detail of the Stillfried
cuirass fragments was an additional pair of ferrous
hooks mounted externally at the side and pointing
upwards, possibly to support the wearer’s belt,
although this is purely speculation.
Unlike the Corbridge form of the cuirass, some
use was made of copper-alloy piping, perhaps to
reduce the time taken to finish the plate edges.
The clipped corners of the ferrous plates of the
Corbridge type were now replaced by simply
rounding them. The copper rivets attaching the
internal leather straps to the ferrous plates could
include decorative washers, but this was not
always the case.

Variants
Robinson believed that all Newstead cuirasses
had fewer girth hoops than their Corbridge
predecessors, but the find from Stillfried makes
this assumption unlikely. Fragmentary girth hoops
from Zugmantel on the German limes do indeed
display a deeper lower plate, however, so this may
be an indication of a separate variant from the
Stillfried finds, unless it is part of a hybrid cuirass
(see p.31).

30
OTHER FORMS OF PLATE ARMOUR
Two additional forms of articulated plate armour have been suggested from
iconographic evidence and these may be termed ‘hybrid’ forms. Articulated
plates were also used as limb defences, most commonly armguards. Single-
or double-plate greaves were known throughout the Roman period, while
head and breast protection for horses also used metal plates.

Hybrid forms
The principal behind articulated plate armour seems to have been applied to
other forms of armour to form composite hybrids of the various technologies.
A sculpted relief of a soldier from Alba Iulia (Romania), with the curved,
rectangular shield characteristic of a legionary, is of particular interest here
(Bishop 2002: 62–66). He is clearly a soldier rather than a gladiator (the Relief from the legionary
tip of his sword scabbard and its chape are visible: gladiators did not use fortress of Apulum at Alba
scabbards) and is depicted wearing a form of armour comprising metal Iulia depicting a soldier with a
curved, rectangular shield; an
bands around his abdomen, along with scale on his shoulders and the kind articulated armguard on his
of breastplates regularly found fastening later mail and scale shirts. The piece sword arm; a hybrid form of
probably dates to the later 2nd or early 3rd centuries ad, since the sword is cuirass with segmental bands
worn on the his left hip. He also wears a segmented armguard on his sword around his torso; and scale
armour over his shoulders,
arm (see p.34). This hybrid cuirass raises a very interesting question: would which is fastened with
it necessarily be recognized in the archaeological record from its components breastplates. (Photo © J.C.N.
if they were not found still attached (Bishop 2002: 62–65)? Coulston)
Assuming this to be an at-least-partially accurate representation of a
genuine form of cuirass, rather than a misunderstanding on
the part of the sculptor, it begs the question of why this
amalgamation of two different forms of armour might
have been thought necessary. It may have been a
genuine attempt to do away with the problems of the
shoulder assembly, but it is also possible that it is in
fact a pragmatic cannibalization of available parts to
produce a functioning piece of body armour when
resources were limited: a one-off that was not found
anywhere else.
The Alba Iulia example is not the only potential
hybrid cuirass depicted on sculpture. A relief from
Arlon (Belgium) shows cavalrymen in action. They
appear to have been depicted wearing mail on the
torso but – in a reversal of the Alba Iulia sculpture
– this has been interpreted as having plate upper
shoulderguards like those of lorica segmentata
(Bishop 2002: 72–73). The problem here is that the
sculpture is sufficiently ambiguous that while one
observer might see plate shoulderguards, another
might identify the standard folded shoulder doubling
found on other cavalry mail cuirasses. The question
of whether plate armour at this point might have
been more or less effective than simple mail doubling
is clearly open to debate, but having only one
ambiguous representative source is, at the very least,
problematic. Comparison with a second relief from

31
What appears to be a Kalkriese-
type top backplate re-used in
a Corbridge-type cuirass from
Rißtissen. (Drawing © M.C.
Bishop)

Arlon, together with one from Liège (Belgium) and a number of cavalry
tombstone reliefs, all of which show standard shoulder doubling, does
little to assist the argument for segmental shoulderguards attached to
mail cuirasses (Bishop 2002: 72).
In both cases, an actual example of a hybrid cuirass would help clarify
the uncertainties outlined here and it is not impossible that such a find
may one day occur. There is, however, a form of hybrid cuirass that is well
attested and incorporates a small amount of plate armour: later mail and
scale cuirasses. During the 2nd century ad, at a time of great innovation
in Roman military equipment, both mail and scale armour are known
with central pairs of breastplates. For a long time dismissed as ‘parade
armour’, it is now realized that it was practical, saw everyday use, and just
happened to be decorated; and the Alba Iulia relief is, coincidentally, the
only depiction of such plates. The plates had the same basic L-shaped form,
produced by removing a curving section from a rectangle to accommodate
the neck. They were reinforced by careful use of corrugation, in the form
of a raised border and internal panel divisions, as well as by the embossed
decoration itself. The standard mythological motifs appeared (see p.51)
but they sometimes also incorporated unit identifiers, specifically when
they had belonged to legionaries. This type of breastplate was, obviously,
used in pairs, nicely demonstrated by a right-hand plate from Mušov

THE ALBA IULIA TYPE IN 3RD-CENTURY DACIA

E Legionaries face up to attacking Goths outside the porta including mail, scale, both Corbridge and Newstead types of
praetoria of the fortress of Apulum in order to give a lorica segmentata, and a hybrid Alba Iulia cuirass, with a scale
returning patrol time to get to safety. Commanded by the top, and girth hoops and shoulderguards harvested from
senior centurion, the primus pilus, who wears a muscled previously damaged segmental armour. The returning
cuirass in recognition of his newly achieved equestrian troops mostly wear mail and scale but the centurion leading
status, they wear a variety of different forms of armour, them has a set of pristine Newstead armour.

32
33
 (Czech Republic), inscribed LEG[io] X, above a representation of a bull,
the totemic animal of legio X Gemina, together with a left-hand plate from
Orgovány (Hungary) bearing the inscription GEM[ina], likewise above a
depiction of a bull. Another (incomplete) right-hand plate from Carnuntum
bore images of an eagle and a figure of Mars with the full title LEG[io]
XIIII GEM[ina] for the legion based in that fortress. The missing lower
register probably included the legionary totem, the capricorn. These pairs
of plates were fastened by overlapping them and then passing a turnkey
attached to the lower plate through a slot in the upper plate. The rotated
turnkeys could then be held fast by a long cotter pin or a pair of split pins
on a thong. Examples of these turnkeys are quite common individually but
some have also been found still attached to a lower plate, as on an example
from the Carlisle Millennium excavations (Bishop in Howard-Davis 2009:
691, Fig. 338). There does not seem to have been a preference as to which
plate, left or right, should be the lower or upper.
There was also a form of single, symmetrical plate attached to mail and
scale, often flared outwards towards the base, and similarly decorated with
embossed mythological scenes. These have much smaller neck cut-outs at
the top and, as with the much shallower neck openings on backplates of
Newstead-type lorica segmentata, they may have been attached to the rear
of a cuirass. Such a plate comes from Ritopek (Serbia) and is covered in
decorative motifs, including two centurial standards, a vexillum, Mars, and
busts thought to depict various martial personifications (such as a genius
Depiction of an armguard on
the tombstone of Sex. Valerius
legionis and virtus). A piece belonging to a mail shirt from Bertoldsheim
Genialis. (Drawing © M.C. (Germany) is of this shape but is actually formed from one narrow and one
Bishop) broad plate and served as a closure.

Armguards
Laminated arm defences were known in the Hellenistic period, with an
example excavated from Ai-Khanoum (Afghanistan) (Bishop 2002: 18),
although the true origins of this form of limb defence, as with body armour,
may lie even earlier with steppe peoples.
In the Roman period, the intimate relationship between legionary
and gladiatorial equipment is readily apparent in the form and use of the
articulated armguard (manica). Armguards provided protection for the
sword arm, which was normally held close to the side of the body, but when
striking was extended and vulnerable. Although representational evidence,
especially reliefs, can be ambiguous and difficult to date accurately, wall
paintings make it clear that two types of arm defence were used. The first
was a padded fabric form, worn on both arms and legs, and generally
shown as light in colour. The second was a metallic defence, normally
of overlapping metal plates. Murals found at Pompeii (Italy) make it
clear from the use of colour and the subtle treatment of the reflection
of light by the artist that this latter form was in use by gladiators before
the destruction of that town in ad 79. While mail and scale manicae are
attested in the arena, however, they do not seem to have been adopted by
the Roman Army.
The convenient terminus ante quem provided by the eruption of
Vesuvius is unusual and most gladiatorial depictions are difficult to date
precisely. Legionary tombstones, on the other hand, are often easier to place
within a time range. Thus it is that two legionary tombstones from Mainz,
belonging to Sex. Valerius Severus and G. Annius Salutus, depict a range

34
of equipment in relief around the inscriptions. Legionary movements are
fairly well understood, thus it is possible to determine that these pieces date
to between ad 43 and ad 70, when legio XXII Primigenia was based at
Mainz. Although it might be objected that the tombstones actually show
gladiatorial equipment, the presence of a pilum makes this unlikely. This
confirms that manicae saw use in at least one unit in the Roman Army during
the second half of the 1st century ad. This is interesting, not least because it
has in the past been suggested that armguards were first introduced during
Trajan’s Dacian Wars at the beginning of the 2nd century ad specifically to
counter the fearsome, scythe-like falx used by Dacian warriors. Armguards
are depicted on the sword arms of legionaries on the sculpted panels or
metopes of the Tropaeum Traiani at Adamclisi in Romania (ad 108–09).
The reliefs on the helical frieze of Trajan’s Column, on the other hand, do
not show armguards in use. The latest depiction of an articulated armguard
is on the Arch of Severus at Leptis Magna (Libya). This probably dates to
c.ad 203 and shows armguards in use with lorica segmentata (J. Coulston,
pers. comm.). Simplified diagram illustrating
Fragments of articulated armguards were found at the beginning of the the internal leathering of an
armguard, with the leathers
20th century at both the legionary fortress at Carnuntum (ferrous) and
attached to the interior of each
at Newstead (copper alloy), although they were not correctly identified lame with a rivet. The larger
at the time of discovery. Some extremely significant discoveries have circles are apertures for the
been made in more recent times, particularly in the deposit in the fort at attachment of a removable,
padded lining. (Drawing © M.C.
Carlisle (see p.38).
Bishop)

Carlisle Millennium ferrous

Armguard A showing both
exterior (left) and interior (right)
views with cross-sections. The
lowest four or five plates (down
to the wrist) were missing when
deposited. (Drawings © M.C.
Bishop)

35
The copper-alloy plates from
the Newstead armguard,
found in the fort headquarters
building. (Drawing © M.C.
Bishop)

Located within what appeared to have been a workshop, elements of

three ferrous arm defences were excavated among an anaerobic deposit
that contained a wide range of organic material, as well as the three arm
defences, still articulated when they were discarded. These finds confirmed
that the plates on such a defence overlapped upwards, from wrist to
shoulder, unlike articulated body armour, which overlapped downwards.

ARMGUARDS IN COMBAT IN DACIA

F Roman legionaries in hand-to-hand combat with Dacians laminated armguard lies along the top of his arm to protect it
outside the walls of the Dacian hilltop capital, Sarmizegetusa against downward blows. Similarly, the greaves worn on the
Regia, during the Second Dacian War (ad 105–06). Wearing leg are a vital defence against a low swing from a two-handed
mail and scale body armour, as shown on the metopes of the falx, the lethal weapon favoured by the Dacians. Damage is
Tropaeum Traiani at Adamclisi, and lorica segmentata as visible on both body and limb armour, some of it quite severe,
depicted on Trajan’s Column, they are all equipped with whereas a tribune to the rear of the legionaries, wearing a
laminated arm defences on their sword arms and greaves on well-polished muscled cuirass of copper alloy, is slightly safer
both shins. It is clear how, when a legionary reaches forward from the blows of their opponents. Some of the legionaries
to strike an opponent and exposes his sword arm, the could have worn of the latest Newstead-type cuirasses.

36
37
 The Carlisle manicae (Bishop in Howard-Davis 2009: 694–700) had
apparently been articulated on leather straps fastened to the plates with
rivets and roves but, despite the excellent state of preservation, none of
the leather straps survived. Arguably the most important revelation from
the Carlisle finds was the extent to which all of the armguards had been
repaired, in some cases multiple times (see p.54).
Other finds have been made in the legionary base of León (ferrous,
Reconstruction of the end of the 3rd century ad: Aurrecoechea et al. 2008) and in what was
Newstead copper-alloy
armguard produced by Peter
apparently an auxiliary fort at Till-Steincheshof in Germany (copper alloy,
Connolly for the Trimontium late 1st or early 2nd century ad: Brüggler et al. 2012). A near-intact ferrous
Trust. (Photo © M.C. Bishop) example from the legionary base at Sarmizegetusa Regia (Grădiștea de
Munte, Romania) remains unpublished. All
of these examples confirm that the metal
plates overlapped upwards from the wrist
to the shoulder. Further fragments of plates,
readily identifiable as deriving from manicae,
are known from Roman military sites at
Corbridge, Richborough and Eining. Overall,
there was little difference between the earliest
and latest surviving examples, although
the Carlisle armour had clipped corners,
while those of the León piece had rounded
ones, much like the plates on contemporary
segmental cuirasses.
The archaeological evidence from
these various finds has provided detailed
information on the structure of this type
of arm defence. The same method of
construction was used regardless of the metal
employed. The top plate was always the
largest (45mm wide on one Carlisle example)
and incorporated some means of attaching
the defence – a hook or a ring – either to
the body armour or to some sort of harness
passed around the upper body of the wearer.
This was essential to stop the defence rotating
on the arm when in use (Aurrecoechea et
al. 2008: 261). The upper edge was also
turned outwards and at the same time
rolled inwards, just as on some plates from
Corbridge-type cuirasses. The strips of metal
(often referred to as ‘lames’) below the top
plate ranged in width between 28mm and
36mm (28–34mm on one Carlisle manica,
31–36mm on another) although they were
not always consistent in width, some varying
by as much as 3mm over their length. The
actual lengths of the lames varied in a more
uniform manner: the shortest plates were
found near the wrist, getting progressively
longer the higher up the arm they were
situated. The terminal plate at the wrist end

38
was the shortest – just 100mm long,
tapering to 70mm, on the León
example (Aurrecoechea et al. 2008:
259, Fig. 5) – and flattest, since it
just had to cover the back of the
hand. These plates were wrapped
around the arm by bending to form
a horseshoe-shaped section, but they
did not completely enclose the arm,
thus avoiding the need for a couter
plate at the elbow. Rivets (both
copper-alloy and ferrous on the
Carlisle examples) along the upper
edge of each plate served to attach
them to the internal leather straps
upon which they were articulated.
A minimum of three straps was
employed (one at the centre and one
towards either end), but this was
often increased to four on the upper
arm, where the plates were longer.
The Carlisle plates were found in
associations that showed that they
were still articulated when deposited
although, despite the deposits
preserving tanned leather extremely
well, the armguard internal leathers
themselves were no longer present
except as small patches of mineral-
preserved matter around the rivets.
Similarly, no substantial traces
of lining nor means of fastening
the defence around the arm were
found (although they could have
been stripped away, leaving just the
plates joined to the internal straps).
This indicates that, as with lorica
segmentata (see p.49), the internal
leathers were probably not made of tanned leather. Forearm section of a ferrous
The upward-overlapping plates of the defence were designed to deflect armguard, still articulated,
excavated from the legionary
blows towards the crook of the arm, where the plates would concertina fortress at León, and dated
when the arm flexed, thereby enhancing local protection. The armguard towards the end of the
thus sat along the top of the forearm and front of the upper arm when the 3rd century ad. The smaller
wearer was holding a sword. plates over the wrist clearly
illustrate how laminated
Besides the lames themselves, there would be a large plate at the
armguard plates overlapped
top of the arm from which the other plates were ultimately suspended. upwards towards the top of the
At least one of the Carlisle examples incorporated metal tie loops and arm. (Drawing © M.C. Bishop)
rings which may have been related to the means of attachment of the
armguard. All surviving examples have additional small holes on the sides
and along the top of the large upper plate, indicating that the defence had
a lining attached; and some examples have been found with fragments of
leather attached.

39
 Muscled cuirasses
For the Romans, the definitive ‘uniform’ of the officer classes – the
senators and equestrians who were appointed to command military
units, provincial armies, and even rule the empire itself – was an
idealized imitation of Hellenistic equipment. Moreover, there has to be
a suspicion that it was intended to imitate the equipment of one man in
particular: Alexander the Great of Macedon (r. 336–323 bc). For this,
the plate cuirass was a central component, but our understanding of this
form of armour is hampered by an absence of any complete examples
available for study. There are many examples of cuirassed statues, both
in stone and cast bronze, and arguably no finer example than the statue
of Augustus from the Villa of Livia at Prima Porta (Italy). These provide
many tantalizing details, but an excavated cuirass of the Republic or
Principate has yet to be recovered.
A remarkable example of a Hellenistic iron cuirass was excavated from
a tomb at Vergina (Greece) and is often attributed to Philip II of Macedon
(r. 359–336 bc), the father of Alexander the Great. Rather than a muscled
cuirass, however, this was an imitation in iron of a Classical Greek linen
BELOW LEFT
cuirass or linothorax. A representation of Alexander wearing a metal cuirass
Detail of the mosaic from the (to judge from the attempted depiction of reflections on the lower part of
House of the Faun in Pompeii his torso) is to be found on a mosaic from Pompeii depicting the battle of
depicting Alexander the Great Issus (333 bc) and generally agreed to be a faithful copy of a painting by
at the battle of Issus with a
gorgoneion in the centre of his
Aristides of Thebes or Philoxenus of Eretria. Alexander’s armour features a
breastplate. (Berthold Werner/ gorgoneion, a depiction of the head of the gorgon Medusa, in the centre of
Wikimedia/Public Domain) the breast. This served an apotropaic purpose to protect the wearer and it is
a common detail to be found on emperor statues. His cuirass also features
BELOW RIGHT shoulder pieces like those on Philip’s iron corselet. It is noteworthy that
Copper-alloy Etruscan or Alexander is not shown wearing a muscled cuirass, since there is no attempt
Greek muscled cuirass of the
4th century bc from Ruvon
to imitate idealized pectoral or rectus abdominis muscles.
(Italy). (© The Trustees of the An actual example of a Hellenistic ferrous muscled cuirass, probably
British Museum) dating to the early 3rd century bc, was excavated from a tomb at Prodromi

40
(Greece) and is much more obviously the model for the cuirasses depicted ABOVE LEFT
on emperor statues. Equipped with shoulder tabs, which were tied to the Statue of Augustus from the
Villa of Livia at Prima Porta near
chest by means of rings on the gilded nipples, it lacked the gorgoneion so Rome, wearing an elaborately
beloved of emperor statues. It also seems that the front and rear halves decorated muscled cuirass
were attached by means of ties through rings fixed below the armpits on highlighting the return of one
either side. of Crassus’ lost eagles. (Till
Niermann/Wikimedia/Public
All cuirassed statues depicted emperors wearing this piece of armour
Domain)
during the Roman period, rather than a segmental, mail or scale cuirass
(Stemmer 1978). The two forms were a low-abdomen, shaped form for men
ABOVE RIGHT
on foot, while those on horseback had higher, horizontal lower edges to Relief generally thought
enable them to sit more comfortably (Bergemann 1990). As such, emperors to come from the Arch of
were attempting to depict themselves as part of the officer class, rather than Claudius in Rome and now in
as ordinary soldiers. The high quality of such depictions, whether in stone the Louvre in Paris. Although
heavily restored (note the
or bronze, means that the main components of the muscled cuirass are anachronistic moustache), it
known. It consisted primarily of the front and back plates, joined on either may depict members of the
side beneath the armpit apertures. The front plate was usually decorated Praetorian Guard, at least two
with a gorgoneion between the emphasized pectorals: an apotropaic of whom appear to be officers
wearing muscled cuirasses.
representation in relief of the gorgon Medusa’s head, famed for turning (Christophe Jacquand/
any who beheld it to stone. The shoulders of the front plate were also Wikimedia/CC BY-SA 4.0)
ornamented with faux doubling, similar to that found on real mail and
scale cuirasses, tied down to rings on the emphasized nipples. Since no
muscled cuirasses survive, however, it is not possible to tell if these were
hinged to the front or rear plate, although the latter might make more sense.
The means of securing the front and rear plates is inevitably simplified on
statues, although a bow on the visible side of the Prima Porta Augustus
suggests they may have been tied together. The cuirass was finished off

41
Funerary tondo in the museum
at Graz, originally from
Seggauberg (Austria), depicting
an officer wearing a muscled
cuirass and holding a sword
in his left hand and (possibly)
a vitis in his right, suggesting
he might be a centurion. (IKAI/
Wikimedia/CC BY-SA 2.5)

with a sash tied around the body at the junction between the thorax and
abdomen, and invariably worn with pteryges (strips terminating in tassels)
at the shoulders and around the midriff. Although it was once thought that
the pteryges were attached to the armour, it is now generally accepted that
they belonged to a garment worn beneath the armour and over the tunic
which is possibly to be equated with the thoracomachus of the literary and
sub-literary sources (see p.57).
It is important to remember that the number of muscled cuirasses in
circulation at any one time was unlikely to have been large. With only
around nine officers per legion likely to have worn one (legatus, praefectus
castrorum, the six tribuni and perhaps the primus pilus, who had attained
equestrian status) and presumably just one (the praefectus or tribunus)
in every auxiliary unit, the chances of even one surviving are slim. An
Antonine funerary tondo from Seggauberg (Austria) depicts an officer
(possibly a centurion, to judge from the vitis (the vine wood staff of office)
he is holding) wearing a muscled cuirass, which might have increased the
number if this practice was widespread at that time. It is also not just a
case of numbers: the differential for survival between mail and segmental
body armour (see p.51) is a known problem, but the factors affecting
large items like a muscled cuirass, with only two major components, are
completely unknown.

42
 Depiction of a muscled cuirass
with pteryges, spear and
vexillum on a relief from the
Temple of Hadrian in Rome.
(José Luiz Bernardes Ribeiro/
Wikimedia/CC BY-SA 4.0)

Greaves
Defences for the lower leg, in the form of greaves (ocreae), were employed
sporadically by the Romans throughout their history. Their use meant that,
when a legionary was in the customary ‘at the ready’ pose, with the left leg
advanced, shield held in front just below eye height to protect the torso,
helmet to protect the head, the last vulnerable exposed part of his body – the
left shin below the lower edge of the shield – was protected by a greave. The
use of only one greave was to be found in both infantry and gladiators for
Plain, undecorated copper-
the same reason: protecting the right shin was less important than the left. alloy greave (unprovenanced).
Polybios recorded that the hastati ‘have two pila, a copper-alloy helmet, and Height 357mm. (Photo P. Gross
a greave’ (Polybios, Histories 6.23.8). © Arachne)
No greaves survive from the Republican era, but the components of a
device excavated from the fortress at Cáceres el Viejo (Spain), generally held
to be the Castra Caecilia of Q. Caecilius Metellus Pius, dating to around
78 bc, were interpreted as a greave press for producing these items from
sheet metal (Mutz 1987).
Under the early Principate, we find that depictions of centurions on
gravestones regularly included greaves that covered the knee. The greaves of
the centurion T. Calidius Severus from Carnuntum were depicted with human
faces on the knees, echoing a classical tradition found on some Greek examples
and also perpetuated on Roman cavalry greaves with knee protection. Those
of M. Favonius Facilis, a centurion of legio XX who died at Colchester
(England) within a decade of the invasion of Britain, were plain, but the greaves
of Q. Sertorius Festus from Verona (Italy) incorporated a decorative, vegetal
design, presumably embossed into the metal. Other centurions’ gravestones,
e.g. an unattributed fragment from Burnum (Croatia) or M. Pompeius Aspro
from Labico, near Rome, choose to only display the attributes of the man,
so instead of a full-figure relief of the deceased, distinctive elements such as a
transverse helmet crest, the vitis and greaves are shown, clearly implying that,
at this time, greaves were the mark of a centurion.
The situation had changed by the beginning of the 2nd century ad,
and possibly even earlier than that. The Adamclisi metopes unequivocally

43
 illustrate legionary troops wearing pairs of greaves. One interpretation
of this might be that encounters with the scythe-like falx used by Dacian
warriors had inspired the Romans to adopt lower leg defences, along
with the laminated armguard, for enhanced protection from this weapon.
That, however, overlooks the fact that an earlier (pre-Flavian) tombstone
of C. Annius Salutus (CIL XIII, 6953) from Mainz has a decorative frieze
surrounding the inscription which depicts, among a variety of pieces of
legionary equipment, greaves and an armguard.
Actual finds of greaves from the first half of the 2nd century ad have
also been made. The same excavations at Carlisle that produced a deposit
containing armguards (see p.35) also contained a greave (Bishop in Howard-
Davis 2009: 700), which was Hadrianic in date. Possibly contemporary
is an assemblage of items reputed to have come from a cave in Hebron
(Palestine), which included an iron Weisenau-type infantry helmet, a mail
shirt and two iron greaves, all tentatively dated to the time of the Bar
Kochba Jewish uprising of ad 132–36 under Hadrian (Weinberg 1979).
Unlike hoplite greaves, which were shaped to fit the shin and calf without
any form of fastening, Roman greaves only ever covered the front part of
the shin and required a fastening, usually in the form of three ring-and-
tie attachments near the top, middle and bottom, to hold them in place.
Rather than being half-round in section, these greaves were often formed
Plain, undecorated iron greave with a central ridge, presumably to provide extra strength and enhanced
from Carlisle dating to the resilience to blows to the front of the shin, but it may also have made them
Hadrianic period. Height
314mm. (Drawing © M.C.
easier to produce.
Bishop) As with any form of plate armour, some form of lining was essential for
both comfort and blunt-force shock absorption (see p.57), and at least two
possible leather greave linings were found during excavations in the midden
outside the fortress at Vindonissa.
This form of infantry greave continued into the 3rd century ad,
examples coming from a hoard of equipment found at Künzing
(Germany) that was probably buried at the time of the Alemannic
invasions of ad 233 or 259/60 (Garbsch 1978: 47). This period also
saw the introduction of cavalry greaves designed for use in the cavalry
training and display exercises known as the hippika gymnasia. These
greaves were both elaborately decorated, with embossed mythological
elements, and structurally different to earlier forms, since the knee and
shin covers were now separate and joined by a hinge. Probably dating to
around the same time as the Künzing find, another hoard of equipment
found near Straubing (Germany), the fort (named Sorviodurum) of a part-
mounted cohort (as yet unidentified) on the Danube, included six cavalry
greaves with knee protectors (Garbsch 1978: 48–49) – Roman infantry
greaves did not need to cover the knee because of the length of the
shield, but the knees of cavalrymen on horseback were more vulnerable.
Ownership inscriptions showed that two of the Straubing greaves had
belonged to men in the turma of Moronus. They were decorated with a
variety of embossed motifs (see p.51). Other decorated greaves are known
from frontier sites, but none in the same concentration that was found
at Straubing.
OPPOSITE By the time of the Dominate, greaves had largely passed out of fashion
Decorated cavalry greaves from
Straubing. Height 499mm (left
once more, and the Late Roman writer Vegetius only describes them in
and right). (Wolfgang Sauber/ the context of their past use by his idealized antiqua legio (Vegetius De Re
Wikimedia/CC BY-SA 3.0) Militari 1.20, 2.15).

44
45
 Horse armour
The Romans did not use plate horse armour in combat but they did find a
place for it in the hippika gymnasia as protective headgear (chamfrons or
shaffrons) for the horses. These are mentioned by Arrian in his description
of those cavalry exercises:

The horses are carefully protected with chamfrons (prometopidia).

On the other hand, they do not need side armour, because the javelins
used in those exercises do not have iron tips, so they could injure the
horses’ eyes, but hardly their flanks, especially since these are largely
protected by a saddle cloth. (Arrian, Techne Taktike 34.8, tr. MCB)

Chamfrons (Garbsch 1978: 13–14, 85–88, Taf. 44–48) were by no

Copper-alloy eyeguard that means a Roman invention, and they were depicted on Hellenistic friezes
would originally have been (often subsequently mimicked in Roman reliefs). In the 1st and early 2nd
attached to a leather chamfron
to protect a horse’s eye
centuries ad, leather chamfrons were used, since they only had to provide
during the hippika gymnasia. protection against soft-tipped practice weapons. Formed from two layers of
(Photo © National Museum of leather (one goatskin, the other cow, both skin outwards), it has been argued
Antiquities, Leiden) that, since these were of double-thickness leather, they would have been made
of toughened leather, formed over a last and treated with heat (Dobson 2018:
7). Although these leather chamfrons were decorated with patterns of studs

A tripartite 3rd-century ad
copper-alloy chamfron with
large central plate and side
plates from the Straubing
Hoard. The central plate depicts
a bust of Mars beneath an
eagle, while the side plates are
decorated with lunate designs
below the integral eyeguards.
An inscription showed that
this belonged to Tertius in
the turma commanded by
Ianuarius. (Wolfgang Sauber/
Wikimedia/CC BY-SA 3.0)

46
 A tripartite copper-alloy
chamfron with a smaller,
hexagonal, central plate
and side plates from the
Straubing Hoard, dating to
the 3rd century ad. Both the
central plate and the eyeguards
depict the head of Ganymede,
the eyeguards being pierced
to enable the horse to see
through them. An inscription
recorded that this was the
property of the duplicarius
Primitivus. (Wolfgang Sauber/
Wikimedia/CC BY-SA 3.0)

and small appliqués, they featured perforated copper-alloy eyeguards for the
horse. The development of a metal form during the 2nd century ad allowed
far greater range in embossed decoration. Without the survival of some of
the leather chamfrons, however, these copper-alloy eyeguards would be the
only evidence for such defences.
Most of the upper part of a one-piece copper-alloy chamfron decorated
with a lion’s head is known from the legionary fortress at Neuss (Germany)
(Garbsch 1978: Taf. 44), but many more-complete and later examples are
known from frontier forts around the empire. These originally consisted
of a small plate that fitted between the eyes of the horse and which was
hinged to perforated eyeguards similar to those used with leather chamfrons.
The central plates became larger and were matched by larger side pieces
incorporating perforated eyeguards. These provided much more scope
for the sort of decorative detail involving deities and heroes beloved on
3rd century ad Roman cavalry. As with the chamfrons made of leather, they
were not intended for use in battle, but rather served to protect the horse’s
face during the vigorous dummy missile volleys of the hippika gymnasia.
The Straubing Hoard produced a group of five, three-part, metal
chamfrons. Each one consisted of a gently tapering frontal that rested along
the nose of the horse, with a hinged plate on either side that incorporated
pierced hemispheres to cover the eyes of the horse while still permitting it
some degree of vision. Each chamfron has some means of attachment to
the head of the animal, some in the form of integral loops, others as free-
moving rings attached to the periphery. Presumably these were attached to
the regular bridle used for the horse, although given the unusual nature of
hippika gymnasia equipment – specifically designed not for combat but for
this unusual and colourful combination of training and display – it cannot be
ruled out that they had their own specialized bridle used just for this purpose
with dedicated attachment points for a chamfron.
The question of whether copper-alloy saddle horns could be regarded as
a form of plate armour (protecting the vulnerable parts of both the saddle
and rider), or whether they were merely a part of the structure of the saddle,
remains moot.

47
 MANUFACTURE AND DECORATION

Workshops
During the Principate the bulk of production of arms and armour seems
to have rested with the Army itself. A papyrus from Egypt records work
undertaken by legionaries, auxiliaries, civilians and even slaves in a legionary
workshop (fabrica). Listing the products over two days, one item (lamnae
levisatae or ‘light plates’) could well be construed as components for lorica
segmentata, but this is of course speculative. A fragment of text by the jurist
(and Praetorian Prefect) Tarrutienus Paternus preserved in Justinian’s great
Digest of Roman law recorded the range of specialists carried on staff by the
legions, including coppersmiths and blacksmiths (Digest 50.6.7). Under the
Dominate, however, there were specific fabricae, spread across both cities
Copper-alloy stamp from and fortresses in the eastern and western halves of the empire and tasked
Oulton used for producing with producing various types of military equipment (Notitia Dignitatum,
embossed rosettes employed Or. 11; Occ. 9).
on both Corbridge-type lorica
segmentata and Weisenau/
Imperial-Gallic helmets. Manufacture
(Drawing © M.C. Bishop) There is no doubt that, to some extent, Roman plate armour had to be
constructed for the individual. When Robinson made his first reconstruction of
a Corbridge cuirass, basing the dimensions on those of the excavated material,
he found that ‘it dug-in even on my daughter’s slender neck’ (Daniels in Allason-
Jones & Bishop 1988: 99). There could be no ‘one size fits all’ approach to
Roman plate armour in general and lorica segmentata in particular.
One of the more unusual aspects of the plates used to form lorica
segmentata is that they were technologically extremely sophisticated.
Metallographic examination of surviving uncorroded plates (of which there
are very few) shows them to have a harder exterior and softer interior that
resembles modern mild steel. Given the available technology, however, and
the fact that it might be anticipated that the plates would have been formed
by working billets of iron, they are surprisingly free of the slag inclusions
that might be expected. Instead, the plates that have been examined have a
degree of purity suggestive of having been produced from a molten state –
something generally believed not to have been possible until the invention
of the Bessemer converter in the 19th century, although a similar process
had in fact existed in China since around the 11th century ad. Moreover,
the uniform thickness of the sheet metal (Sim & Kaminski 2012: 49–50),
together with marks on the surface, points towards some sort of rolling
process being used in the formation of the steel sheet (Sim & Kaminski
2012: 137). The thickness of the uncorroded London (Bank of England)
breastplate is just 1mm but, as has been observed (Sim & Kaminski 2012:
138), Roman steel plate was every bit as effective as High Medieval plate
twice the thickness; and this quality was available to every soldier, not just
an elite few.
By contrast, the production and assembly of sets of fittings for lorica
segmentata can seem quite crude at times, so much so that one suspects
that semi-skilled labour was being used for the process: unsurprising, if
legionary workshops were using legionaries, auxiliaries and civilians. The
copper-alloy components were cut from thin brass sheet which, like the
steel sheet of the armour, shows signs of having been rolled in order to

48
 Two upper shoulderguard
plates of lorica segmentata from
the Corbridge Hoard that have
been overlapped more than
normal then riveted together,
although the redundant lobate
hinges have not been removed.
The purple-grey colouration of
the decorated rosette washers
has been confirmed by pXRF
as being derived from silver.
(Photo © M.C. Bishop)

achieve its thinness. Riveted and hinged fittings were usually produced by
doubling over the sheet before it was riveted in place, although fittings are
sometimes found where only the end with the hinge is doubled over and
the rest of the body is made from single-thickness brass sheet. Decorative
rosette washers for rivets were stamped from brass sheet before being cut
out, and stamped-but-uncut examples have been found at military sites
including Rheingönheim (Germany). A stamp for producing these is also
known from Oulton (England).
The brass fittings were normally fitted to the ferrous plates by means of
rivets with a very high copper content, which was reflected in their colour:
orange in contrast to the golden yellow of the brass. The high proportion
of copper made the rivets softer and thus easier to peen over and secure
in place.
It was noted earlier that, while the Carlisle deposit contained considerable
amounts of tanned leather, no internal leathering survived on any of the
armguards, despite their still being articulated. It is also true that no lorica
segmentata components have been found with organic remains of leather still
attached, only mineral-preserved straps. It is not unreasonable to conclude,
therefore, that a form of untanned leather (such as rawhide) was used for the
internal straps of both cuirasses and armguards. This must have been oiled in
some way to keep it supple – if allowed to dry out the leather could become
extremely hard; a property the Romans exploited when using it for shield
edging in the 3rd century ad (Bishop 2020: 22).
It has on occasion been argued that both muscled cuirasses and lorica
segmentata may have been formed from thick moulded or hardened
leather, but there is no plausible evidence that this was the case. Sculpture
cannot in and of itself assist with such an identification, while both finds
of metal body armour in Greece and Italy, alongside colour depictions,
suggest metal was the favoured material for muscled cuirasses. The
absence of actual finds of leather muscled cuirasses in a Roman context
is matched by a similar lack of metal examples, but given the likely
limited numbers of such cuirasses in use at any one time, this is hardly
surprising. By contrast, the fact that components (and even complete

49
 sets) of ferrous-plate segmental cuirasses survive in considerable numbers
heavily mitigates against the use of leather for this type of armour, not
least because organic armour could never have been as effective as steel at
the same thickness and weight (so would have to have been heavier and
bulkier to come anywhere near matching its defensive qualities), quite
apart from the additional demands it would have placed upon Rome’s
livestock resources.

Decoration
So far as is known, there was nothing by way of decoration on the
ferrous plates of lorica segmentata, but the appliqué fittings were another
matter altogether. Made from orichalcum brass (the same alloy used in
some coins), these looked almost indistinguishable from gold, providing
they did not become tarnished. Whether it was the embossed rosettes
that acted as washers for leathering rivets or the lobate hinges used to
join neighbouring plates, there was a standard grammar of ornament
associated with this type of armour and at least in part shared with
helmet decoration under the Principate. Leathering rivet washers, for
instance, could be embossed with rosette designs or concentric circles,
while some examples incorporated red enamel on the rivet heads. Some
of the Kalkriese fittings were tinned or silvered and pXRF analysis has
shown that even the Corbridge Hoard armour included silver (or silvered)
decorated washers on some upper shoulderguards (Dr F. McIntosh, pers.
comm.). Rivet holes on fittings were often adorned with concentric incised
rings, possibly a product of the tool used to punch the hole through the
sheet metal. The plain rectangular hinges of the shoulder units on early
Kalkriese forms of lorica segmentata quickly evolved into sub-lobate
hinges and, ultimately, into lobate hinges on the Corbridge type. There
was no need for such an elaborate form for these fittings, which had to
be cut out from brass sheet, since all that was needed was a hinged plate
that could be attached to two ferrous plates with five rivets on each
half: the shape was purely decorative. Robinson (1975: 177) believed that
such fittings began elaborate and eventually became cruder with time,
but the elaborate openwork lobate hinges of the Newstead type disprove
this hypothesis. Moreover, while the use of copper-alloy piping on the
Kalkriese and Newstead types was undoubtedly primarily functional, in
that it could conceal poorly finished edges that might snag on the wearer’s
clothes, there can be no denying that it had visual appeal too, and this
was enhanced in some cases with additional decorative details such as
the embossed beading of the León examples (Aurrecoechea & Muñoz
Villarejo 2001/02: 20). As noted earlier, some copper-alloy fittings on the
Kalkriese form could be tinned or silvered (see p.14).
It is not known how the exterior of segmental body armour was presented
and whether it was highly polished or just kept clean. There is a limited
amount of evidence to suggest that shiny armour was thought to intimidate
an opponent, Vegetius noting: ‘For the brilliance of equipment terrifies
enemies. Who can believe that a soldier is warlike, if, through negligence,
his arms are disfigured by filth and rust?’ (Vegetius, DRM 2.14).
It was thought to be a mark of how decadent city-based, 2nd-century ad
eastern troops had become when Lucius Verus’ general, Pontius Laelianus, ‘a
man of character and a disciplinarian of the old school, in some cases ripped
up their cuirasses (loricae) with his fingertips’ (Fronto, Letters to Lucius Verus

50
19). Regardless of whether this refers to mail, scale or plate armour, it
underlines the importance of well-maintained armour to the Romans
(see below).
At the same time, little evidence survives in the archaeological
record for the original condition of the interior of lorica segmentata
and whether any forge blackening (with, for example, beeswax
or olive oil) was engaged in, principally for the purposes of
preventing corrosion.
Other forms of plate armour (in copper alloy, at least) made great
use of embossing. Decorated greaves are regularly depicted in Roman
representational art and finds of what are thought to be such greaves
from cavalry sports armour are elaborately embossed and chased
to show common mythological figures and deities. Examples from
Straubing included Mars, Hercules and Minerva, along with eagles
and dolphins (common military decorative motifs at this time).
The chamfrons from Straubing (see p. 47) also – unsurprisingly –
exhibit a very similar grammar of ornament to the greaves, including
embossed representations of Mars, Minerva and the Dioscuri (Castor
and Pollux), along with eagles and snakes. The hemispherical eyeguards
incorporated into the chamfrons were mostly pierced with a pattern of
triangles, although one piece had eyeguards in the form of the pierced
head of the Gorgon, with the eyes of the figure similarly pierced,
meaning the horse was also looking through the Gorgon’s eyes!
It is unclear whether this type of embossed decoration was
usually produced by chasing from the rear against a former – a
labour-intensive operation – or by some sort of stamping process,
which was much closer to a form of mass production. Stamping was used The early 3rd-century ad
for smaller items like belt plates, lorica segmentata decorative washers (see tombstone of Marcus Aurelius
Alexys from Sparta, possibly
p.49), and even for larger items such as decorated sword scabbards (Bishop showing the girth hoops of
2016: 16). The discovery of what have been identified as components of a a lorica segmentata. (George
Republican-era greave press during excavations at Cáceres el Viejo, if they E. Koronaios/Wikimedia/
have been correctly interpreted, would seem to support the possibility that CC BY-SA 4.0)
embossed greaves could have been produced by stamping copper-alloy sheet
(see p.44). If that was indeed the case, then it would not be surprising if the
same was true of chamfron components.

Maintenance
In the archaeological record, pieces of segmental plate armour are far more
common than scale or mail. Comparison with the representational record,
however, suggests that this is not because it was the most common form
of armour in use. Both sculpted, figural tombstones and state propaganda
monuments depict scale and mail in use. The problem seems to have been
that segmentata was much more prone to damage than either mail or scale.
Careful examination of finds like those from the Corbridge Hoard usually
reveals evidence of damage and repairs (both competent and botched).
In contrast, no complete examples of a muscled cuirass from the Late
Republic or the Principate have survived, despite the fact that they seem to
have been the preferred form of body armour for equestrian and senatorial
Army officers. In fact, there are more fragments of cast bronze statues
depicting the cuirasses of members of the imperial family.
These observations reveal fundamental differences between these two
main forms of plate armour in use by the Roman Army. If a metal muscled

51
 cuirass – with few major components – was damaged, the entire object had
to be repaired; if a piece fell off of a segmental cuirass, however, the sheer
number of components became advantageous, since damage to one item
was unlikely to be as crucial to the whole defence as those pieces of muscled
cuirass. Roman re-enactors will frequently carry with them spare lengths of
leather lace with the specific purpose of being able to manage running repairs
to their armour if something should break.
Quite apart from the complexity and relative fragility of its structure,
every version of lorica segmentata had a fundamental flaw: the combination
of ferrous and copper-alloy components made it vulnerable to bi-metallic
corrosion, particularly if it was not carefully dried and cleaned after use.
Indeed, some of the points of contact – underneath brass fittings riveted to
ferrous plate, for example – could not be cleaned without de-riveting and
removal, which was impractical under normal circumstances and thus made
Two collar plates of lorica these points of weakness.
segmentata from the Corbridge The majority of lorica segmentata components that have been found
Hoard, twisted out of alignment
and with one original hinge
have been in obvious need of repair. That was, presumably, the reason they
half (bottom) and a larger had been set aside and ultimately ended up deposited in the archaeological
replacement (top), which has record: to be repaired, cannibalized or scrapped. Many of the finds also show
been riveted into place with signs of having already been repaired, typically manifested with fittings that
four larger dome-headed rivets,
do not match the original design or cruder riveting. Original rivets tended
one of which has been secured
centrally through a triangular to have small, hemispherical heads, whereas repairs often featured non-
cut-out instead of in either of matching, flat, asymmetrical heads, giving every sign of having been a hasty
the two available holes. (Photo and (viewed with a mildly critical eye) inept, if functional, solution. Similarly,
© M.C. Bishop)
the detailed cutting that was necessary to form the fittings, when executed
by skilled craftsmen, was a long way from the frankly inept and misshapen
examples that have been found at some sites. Some fittings clearly originated
with other, ‘donor’ cuirasses, which is one of the possible explanations for
the armour that was found among the contents of the chest containing the
Corbridge Hoard, since one of the type B/C upper sets seems to have been
caught in the process of having its upper shoulderguard assembly removed
and added to one of the type A shoulder sets which lacked such plates
(Allason-Jones & Bishop 1988: 100). This was cannibalization in action,
sacrificing components from a damaged cuirass to keep another one fit for
service; armies have always done this and probably always will. The Roman
fort in post-Boudican London produced more evidence of cannibalization,
with an articulated set of lorica segmentata shoulder and collar plates, one
component of which was missing (presumably removed), the remainder
having been discarded as if of no further use. One of the valuable products of
the discovery of the Corbridge Hoard is the revelation that Roman soldiers
did not seem to have cared about the look of such repairs, which serves to

REPAIRING ARMOUR AT LEÓN armguards. In some cases, this involves taking one of the
G The scene is a rectangular rampart-back building behind the quarter-units of the armour to pieces, de-riveting the internal
southern defences of the legionary fortress at León in leathers or the brass fittings in order to replace them. The
Hispania. The original 1st-century ad stone wall has been craftsmen, specialists in their tasks, are assisted by semi-skilled
strengthened and raised by the addition of a new wall in front workers detailed from century strength to perform tasks like
of (and incorporating) it and this can be seen above the cutting out sheet metal or leather, acting as strikers around
original wall. the anvil, or just sweeping the workshop clean. There are also
Through large, full-height openings with shutters pulled slaves present to assist with the running of the facility
back, craftsman are hard at work repairing both Newstead and (carrying fuel or water and raw materials), along with civilian
Corbridge types of lorica segmentata, as well as segmental contractors carting in supplies.

52
53
 remind us that modern concepts of military ‘uniformity’ were not necessarily
observed in the same way in antiquity.
Often, repairs did not require major replacements, but rather minor
‘tweaks’. The substitution of a copper-alloy fitting was a common solution
on lorica segmentata. The Corbridge Hoard sets of armour include examples
of replaced lobate hinges, as well as a former hinged joint on an upper
shoulderguard repaired by simply riveting the plates together (thereby
coincidentally demonstrating that the hinged joints on upper shoulderguards
were far from essential). On Newstead-type cuirasses, a common solution
to problems with missing tie rings or damaged fasteners was to use a strip
of copper alloy bent around a circular-sectioned former as a replacement;
this was used instead of tie loops on the girth hoops from Newstead and
Zugmantel, as well as on the vertical fastener from Eining.
Repairs were not just confined to the metal components, however. The
Upper backplate of lorica leathering was vulnerable to mechanical damage from repeated movement,
segmentata from the Corbridge particularly from plates rubbing or even cutting worn straps; from being
Hoard (with rosette washer
under the leathering rivet)
soaked (whether by rain or perspiration) and then dried; while the riveting
incorrectly overlapped by the points were inevitably going to be points of stress. One of the sets of
middle backplate below it. backplates from the Corbridge Hoard had had the two parallel leathers
(Photo © M.C. Bishop) replaced by one large patch, ironically anticipating Groller’s leather-based
reconstruction. Re-leathering would be far less obvious when one strap was
simply replaced with a newer one, although the variety of mixed decorative
washers on some sets of armour may be indicators of repairs in this way.
The possibility that some older forms of segmental cuirass were upgraded
when new forms appeared should also be considered. There is, for example,
a mid-collar plate from the excavations at Chichester which has the lobate
hinges of the Corbridge type but the broader dimensions of the Kalkriese
form (Thomas 2003: 122, Fig. 79, 15).
The fittings of lorica segmentata are among the most common finds of
armour to come from Roman military sites. This raises the question of just
how common its use was: was it ubiquitous from the 1st century ad onwards,
as the finds seem to suggest, or does the fragility of these fittings skew the
finds spectrum in favour of segmental and against mail and scale armour?
There is no way of being certain, but the images on the Adamclisi metopes
and on gravestones seem to imply that mail and scale might have been just
as prevalent among legionaries as lorica segmentata. Trajan’s Column has
done its job too well.
Examination of the sets of armguards found in the Carlisle excavations
revealed a great deal about the repairs that had been undertaken during their
lifetimes and showed them to have been just as vulnerable as segmental body
armour (Bishop in Howard-Davis 2009: 694–700). Since some of them used
dissimilar rivets to attach the internal leathering to the plates, it was possible
to identify one manica that had been formed by joining sections from two
different units, one of them originally being provided with copper-alloy rivets,
the other with iron examples. This joining of two different assemblies was
also apparent from a realignment of the internal leathering regime. Other
signs of repairs include multiple rivet holes (and even rivets) at one location,
as well as a number of examples of plates simply riveted together so that they
no longer articulated. Because the elbow was the only region that demanded
flexion in the defence, this was at least feasible for the upper and lower arm.
Asking why so many repairs were necessary for both segmental limb and
body armour prompts some interesting observations. Riveting together plates

54
that were originally articulated (evidenced for lorica segmentata and manica
components) hints at an element of haste in their repair, while cannibalization
– using elements of one defence to repair another – is indicative of an urgent
need that could not be met by the manufacture of new items. The fact
that such repairs were necessary on items that had presumably entered the
archaeological record because they had been put aside for repair only serves
to underline the pressure under which the Roman Army sometimes found
itself, the reasons for which will be examined next.

PLATE ARMOUR IN USE

Understanding how all Roman armour was used has been greatly aided
by the plethora of reconstructions based on archaeological evidence that
followed on from Robinson’s pioneering work, although, as with all
experimental archaeology, they can only ever show what was possible, not
what actually happened. That said, the overall fragility of segmental body
armour in comparison to mail seems beyond doubt, and the archaeological
evidence clearly indicates that the evolution of segmental armour was guided Copper-alloy figurine from a
by pragmatic responses to genuine problems that arose from its use. It is not shrine at Versigny interpreted
as depicting a crupellarius
for the modern commentator to decry lorica segmentata as a flawed form gladiator. (Photo © Musée
of defence when the Romans seemed quite happy to rely on it in battle for Jeanne d’Aboville/La Fère)
more than 300 years.

Combat
Unsurprisingly, perhaps, no unequivocal accounts survive of plate armour
being employed in combat by the Roman Army. There is, however, a rather
interesting description of the uprising of Sacrovir (together with Florus) in
Gaul in ad 21:

His followers amounted to forty thousand; one-fifth armed on

the legionary model; the rest with boar-spears, knives, and other
implements of the hunting-field. To these he added a contingent of
slaves, destined to be gladiators and encased in the continuous shell of
iron usual in the country: the so-called crupellarii – who, if too weighty
to inflict wounds, were impregnably fortified against receiving them.
(Tacitus, Ann. 3.43)

A copper-alloy statuette from Versigny (France) has been identified

as depicting one of these crupellarii (Picard 1980). If this is a correct
identification, then it is significant in that it depicts overlapping plate armour
similar to lorica segmentata with manicae and thigh guards. Once the Roman
forces of G. Silius confronted the rebel forces, the legionaries adopted some
ingenious solutions to deal with these crupellarii: ‘... in front, the iron-clad
men offered a brief impediment, as their plating was proof against pilum and
sword. However, the legionaries seized their axes and pickaxes (dolabrae)
and hacked at armour and flesh as if demolishing a wall: others overturned
the inert masses with poles or forks, and left them lying like corpses unable
to get up again’ (Tacitus, Ann. 3.46).
This was the first time that legionaries in the western empire had
encountered a foe wearing plate armour and experienced its efficacy

55
 against their traditional armament: the pilum and short sword. This in turn
provides an important insight into how formidable legionaries equipped
with plate armour must have seemed to their opponents. Nevertheless, all
armour is a compromise between protection and mobility, and legionary
lorica segmentata only really offered such protection to the torso, leaving
plenty of vulnerable areas exposed. The crupellarii sacrificed mobility for
increased protection, as Tacitus makes clear, although his observation
about felled gladiators unable to get up again may just be his idea of
an amusing aside or a literary topos, as used elsewhere when armoured
Sarmatian cavalrymen are brought down by Roman troops and unable to
get up (Tacitus, Histories 1.79).
In fact, the bulk of the evidence for the use of plate armour in combat
comes from archaeological remains of the artefacts themselves. It has been
argued that one of the explanations for the ubiquity of segmental armour
fragments in the archaeological record is related to its vulnerability to
damage for whatever reason. Indeed, traces of damage are rife among
surviving pieces of lorica segmentata and manica; some of it clearly is
the result of combat while other examples may also originate with enemy
BELOW LEFT action but lack obvious signs to confirm it. The best example of what is
Interior view of an over-
compressed set of Corbridge
probably combat attrition is a damaged Corbridge type A lorica segmentata
type B/C girth hoops from breastplate from London which sustained a penetrating blow at some point
the Corbridge Hoard with the before deposition. The fact that it was most likely being worn at the time is
remains of the internal leathers indicated by the limited amount of overall distortion to the plate beyond the
and the pairs of copper-alloy
rivets attaching them to each
actual penetration damage. This suggests that the body of the wearer and
ferrous strip. (Photo © M.C. any garments worn between his body and the armour plate were supporting
Bishop) the plate as it was struck.
Ergonomic design was key to lorica segmentata, because (unlike mail
BELOW RIGHT and scale) every care was taken to deflect blows in much the same way that
A reconstructed set of it was with Roman helmets. The upper and lesser shoulderguards sought
Corbridge-type lorica to deflect blows outwards and downwards in much the same way that the
segmentata girth hoops
showing how they stand
girth hoops did. Additionally, the upper shoulderguards served to provide
naturally on slightly worn additional shock absorption and reinforcement for the shoulder region, in
internal leathers with a degree much the same way as shoulder doubling in mail or scale cuirasses. This was
of sag and demonstrating how partly as a result of the overlapping plates being supported on the internal
much they would have to be
compressed to match the girth
leathering, but also one of the possible functions of the otherwise enigmatic
hoop sets in the Corbridge hinges between plates: to introduce some ‘give’ and thereby dissipate the
Hoard. (Photo © M.C. Bishop) energy of a blow.

56
 Re-enactors wearing
Corbridge-type loricae
segmentatae showing the
same misaligned backplates
evidenced by archaeological
finds. (Photo © M.C. Bishop)

Carriage
One problem with Robinson’s reconstruction of lorica segmentata was
readily apparent from the way that it sat upon the human frame. There was
an unavoidable ‘sag’ in the shoulder sections which caused the breastplates
to overlap at an angle, producing an awkward gap between the breastplates,
although the placement of the fastenings upon them implied that this should
not in fact have been the case. Most suggestions to counter this problem,
including the possibility that an additional ‘centre plate’ had somehow been
missed (Daniels in Allason-Jones & Bishop 1988: 99), bore no fruit, but
there was nevertheless a simple solution in both the ancient literary sources
and in the study of medieval and post-medieval armour: padding. All armour
required some sort of lining or undergarment, such as an arming doublet or
aketon, to help spread the force of a blow against the carapace of the armour.
Such garments seem to be described in both the Late Roman De Rebus
Militaris by an unknown author, in which it was called the thoracomachus
(Bishop 1995), and in a similarly anonymous Byzantine military manual
(Dennis 1985: 55). It did not need to be particularly thick over most of the
body to be effective, but such a garment offered the ideal place to incorporate
some padding on the shoulders to counter the natural slope of the trapezius
muscles which was causing the sag observed in reconstruction segmental
cuirasses. Such a garment was also the obvious means of attaching the
pteryges visible with muscled cuirasses and lorica segmentata and, indeed,
examples occasionally appear in representational art, such as on a relief now
in the Baths of Diocetian in Rome (Robinson 1975: Fig. 158). These strips,
variously shown as both rigid and flexible, may have been formed from
double-thickness tanned leather, which would have afforded some limited
protection to those parts of the body left uncovered by armour, such as the
tops of the shoulders or the buttocks and lower torso. The pteryges were
usually in two overlapping layers and, when worn with the muscled cuirass,
the top layer was sometimes folded over a waist belt, as with the officer
depicted on the Louvre Praetorians relief. Pteryges can also be seen being

57
For storage or transport,
segmental body armour could
be disassembled into its four
constituent components (two
upper and two lower halves)
and partially collapsed. Over-
enthusiastic compression
could, however, lead to damage
to the internal leathering
regime and potentially to
the metal fittings too. It was
in this form that the pieces
of armour were found in the
chest containing the Corbridge
Hoard: crammed together to
make them as small as possible.
These Corbridge-type lorica
segmentata girth hoops from
Rißtissen demonstrate the
same sort of over-compression
found in the Corbridge
Hoard, indicating that the set worn with 2nd-century segmental armour on the Marcus Column, on the
was forcibly compressed for panels of Marcus Aurelius on the Arch of Constantine, and on a copper-alloy
storage. Examination of the
mineral-preserved remains of
figurine now in the British Museum in London.
the armour in the Corbridge The implication of using external vertical strap fasteners on the front of
Hoard revealed that it had the Kalkriese and Corbridge type A cuirasses, but internal buckles at the
been wrapped in cloth prior rear, is that the cuirass was designed to be put on with the back fastened and
to deposition, presumably to
help protect it. (Drawing © M.C.
the front unfastened, so that the wearer could put on and take off his own
Bishop) armour without the need for assistance.
It is difficult to believe that that Corbridge Hoard contained three left and
three right upper units (two A, one B/C in each case), as well as the same
number (and type) of lower units. This would imply – if indeed these were
deposited as complete sets – that the Romans were not overly fussy about
whether cuirass halves matched perfectly; and, ignoring typological niceties,
that there may have been both B and C fittings together on the same cuirass!

Versatility
A key advantage to lorica segmentata was that it allowed for the full six
degrees of freedom in bodily movement. It was by no means the only form of
armour to do this (mail was just as flexible, although heavier), but this meant
that it was possible to march, work and fight while wearing the cuirass.
The sculptors on the helical frieze of Trajan’s Column showed citizen troops
performing all of these tasks, including the construction of fortifications,
while wearing segmental body armour. A detail on that frieze (Scene XXVI)
depicts a legionary wading across a river. His body armour, which is very
clearly lorica segmentata, is resting in the hollow of his shield, which he
holds above his head (Bishop 2020: 58 Fig.). Although these are all extremely
enticing images, it is unclear to what extent they represent artistic licence on

SEGMENTATA STORED AT CARNUNTUM – can be seen in various states of disassembly on the sets of
H The armamentarium built into the rampart immediately behind shelves where they have been stored. In the background, other
the western wall of the legionary fortress at Carnuntum near types of armour can be seen stored on shelves, including mail
Bad Deutsch-Altenburg was found to contain a wide range of and scale cuirasses. There are also intact armguards hanging
weaponry when it was excavated. In this reconstruction, two from the shelves, as well as components from such defences
different types of lorica segmentata – Corbridge and Newstead once again awaiting repair or cannibalization.

58
59
 the part of the metropolitan sculptors as opposed to observation of actual
practices undertaken by troops in the field. Indeed, the notion that the scenes
shown on the reliefs were based on accurate ‘field sketches’ has been called
into question in recent years.
Segmentata was lighter than scale or mail – reconstructions typically
weigh between 5kg and 9kg – which meant it was less fatiguing to wear, and
arguably offered better protection than the other two main types of body
armour in certain circumstances. Other advantages include its diminished
requirements for raw materials, its ease of manufacture and the time taken
to manufacture it (Sim & Kaminiski 2012: 137–38). Since the Roman state
was mining the raw materials, harvesting scrap for re-use or re-forging,
and fabricating and mending the armour, the cost was in terms of time and
manpower, rather than money. This is why the haste of some repairs (see
p.52) is significant.

Who wore what?

If nothing else, the overly simplified categories of troops represented on
Trajan’s Column should warn against using the sculptures of that monument
as a literal representation of what the various components of the Roman
Army may have looked like. Segmental cuirasses were reserved for citizen
troops belonging to the Praetorian Guard or the legions. This is confirmed
by the Adamclisi metopes, on which citizen soldiers are only shown wearing
mail or scale cuirasses.
Archaeological evidence has been used to try to counter the impression
given by the reliefs of Trajan’s Column and suggest that it was not just
legionary infantry who used lorica segmentata. Fittings are known from a
number of smaller Roman military sites from the early Principate of a type
unlike the fortresses large enough to accommodate one or more legions.
Although inscriptions do not survive from early timber forts, it has been
argued by analogy with later installations that demonstrably held auxiliary
troops, such as those along Hadrian’s Wall, that it must also have been the
case for these earlier sites. As a hypothesis, this is crucially flawed, because
legionary detachments are well known to have been outposted in smaller
forts (e.g. Tacitus, Ann. 3.74).
Additionally, forts have produced this material in the province of Raetia
(now part of modern-day Germany) which initially had no legion as part of
its forces. Some scholars have concluded that here auxiliaries also used the
lorica segmentata. The discovery of a helmet in the river near the Roman
fortlet of Burlafingen seriously undermines this interpretation, however, since
it bore the ownership inscription of a soldier in legio XVI Gallica, thought to
have been based at Mainz in Germania Superior, at the time of the loss of the
helmet. In reality, legions based in provinces under commanders appointed
by the emperor frequently deployed into neighbouring provinces as required
by circumstance. This also happened more than once with Judaea, well before
the Jewish uprising of ad 66–70. There is thus no reason why legionary
equipment – in this case lorica segmentata – should not be found in provinces
without a permanent legionary establishment.
So it seems that Trajan’s Column may in some respects have been correct
in attributing segmental body armour to legionaries and praetorians, but this
has to be qualified by noting that there is substantial evidence (both sculptural
and archaeological) that these troops also used scale and mail armour. In
other words, the sculptors of Trajan’s Column were deliberately selective

60
in their decision to indicate citizen troops by means of lorica
segmentata, together with shield types and military standards.

Legacy
Segmental plate armour was known long before the Romans
– the set of Mycenaean bronze armour from Dendra (Greece)
comes to mind – and it was of course widely utilized in the
medieval and post-medieval period. There is no need to seek
a connection between these instances, however. Articulated
plate armour was such an obvious solution to enhanced bodily
protection for soldiers that it is unsurprising that it was invented
more than once. The lack of scholarly interest in Trajan’s
Column and similar monuments before the Renaissance ensured
that articulated medieval plate armour – even those with sliding
rivets – owed nothing to the Romans.
The absence of a direct line of transmission of the idea of
segmental armour does not mean that Roman lorica segmentata
was not influential. That legacy was artistic, however, and
not military. From the earliest sketches made of Trajan’s
Column, via Hollywood epics, to the latest television dramas,
the identification of Roman soldiers with this type of armour
is absolute and the two are difficult to prise apart. Whether
depicting the Republican period, the Principate or the Dominate,
Spartacus or Constantine, Hollywood and the visual media have
frequently reverted to that agenda set by Trajan’s Column, so
effective has it been at transmitting its message.
The love of classical imagery so typical of Renaissance art
has also meant that the individual, heroic officer or emperor in
a muscled cuirass has become just as powerful an image as the
massed ranks of segmentata-clad legionaries shuffling across the
silver screen. A statue of King James II in Trafalgar Square in
London is a perfect example of the adoption of this persona in comparatively Copper-alloy figurine of a
recent history: wearing an elaborately decorated muscled cuirass (rather legionary wearing lorica
segmentata, now in the British
confusingly rendered with a scale texture) adorned with pteryges at both the Museum (acquired from
arms and waist, he is the epitome of what has recently come to be known as the collection of Alessandro
‘classical reception’. Similarly, the figurehead of the Royal Yacht HMS Royal Castellani in 1867). The style
George depicting King George IV in a Roman cuirass (now in the collection of beard suggests that it
dates to the latter half of the
of the National Maritime Museum in Greenwich, London) continued this
2nd century ad. (© The Trustees
tradition into the 19th century. of the British Museum)
The popularity of Roman re-enactment and living history in general
have guaranteed that Henry Russell Robinson’s legacy, as the person who
(along with Charles Daniels) pieced together and began to understand the
original find of cuirasses from Corbridge, before going on to produce his
own fully functional replicas, would enhance (if not replace) the original first
examinations of Trajan’s Column. Today, Roman re-enactment groups are
numerous and worldwide, but the armour they wear owes more to Robinson
than it does to Trajan’s Column. Some re-enactors make their own cuirasses,
but many buy from commercial vendors (often producing the components on
the Indian subcontinent, using the workshop skills of the craftsmen there).
This in turn means that it is now possible to buy a more-or-less accurate
set of lorica segmentata online, so the legacy of this form of plate armour
is assured.

61
BIBLIOGRAPHY

Ancient sources
AE – L’Année Epigraphique. Available from Polybios, Histories. 1922–27 Loeb edition, trans.
Epigraphik-Datenbank Clauss/Slaby at https:// W.R. Paton. Available at https://tinyurl.
tinyurl.com/ryexz6k com/298bsf
Arrian, Techne Taktike, eds R. Hercher and A. RIB – Roman Inscriptions of Britain. Available at
Eberhard. Available at https://tinyurl.com/ https://tinyurl.com/rmxz8q9
vkn8fmbt Tab. Vind. – Tabulae Vindolandenses. Available
CIL – Corpus Inscriptionum Latinarum. from Vindolanda Tablets Online at https://
Epigraphik-Datenbank Clauss/Slaby. Available tinyurl.com/m7gnl5
at https://tinyurl.com/ryexz6k
Tacitus, Agricola. 1876, trans. A.J. Church &
Justinian, Digest. 1932 Central Trust Company
W.J. Brodribb. Available at https://tinyurl.com/
edition, trans. S.P. Scott. Available at https://
v2oqbvp
tinyurl.com/ybc66j2j
Livy, History of Rome. 1912, trans. Revd Canon Tacitus, Annals and Histories. 1925–37 Loeb
Roberts. Available at https://tinyurl.com/ edition, trans. J. Jackson. Available at https://
wnnnm7k tinyurl.com/8jpv49
Notitia Dignitatum. Available at https://tinyurl. Varro, On the Latin Language. 1938 Loeb edition,
com/uvz3z4e trans. R.G. Kent. Available at https://tinyurl.
Pliny the Elder, Natural History. 1855–57 Bohn com/rhade6v
edition, trans. J. Bostock & H.T. Riley. Vegetius, De Re Militari. 1885 Lang. Available at
Available at https://tinyurl.com/cvmp3bu https://tinyurl.com/ybvq5ypu

Modern sources
Allason-Jones, L. and Bishop, M.C. (1988). Bishop, M.C. (1995). ‘Aketon, thoracomachus,
Excavations at Roman Corbridge: the Hoard, and lorica segmentata’, Exercitus 3:1: 1–3.
HBMCE Archaeological Report No. 7. Bishop, M.C. (2002). Lorica Segmentata I. A
London: English Heritage. Handbook of Articulated Roman Plate
Aurrecoechea, J. (2003/04). ‘New perspectives on the Armour, JRMES Monograph 1. Chirnside:
evolution and chronology of Roman segmental The Armatura Press.
armour (lorica segmentata and manica), based Bishop, M.C. (2015). ‘The Eining lorica
on Hispanic finds’, JRMES 14/15: 49–55. segmentata’, in P. Henrich, Ch. Miks, J.
Aurrecoechea, J., Fernández Ibáñez, C., García Obmann & M. Wieland, eds, NON SOLUM
Marcos, V. & Morillo, Á. (2008). ‘Un protector ... SED ETIAM. Festschrift für Thomas Fischer
laminado de brazo (manica) procedente del zum 65. Geburtstag. Rahden: Leidorf: pp.
campamento de la Legio VII Gemina en León’, 31–40.
Archivo Español de Arqueología 81: 255–64. Bishop, M.C. (2016). The Gladius: The Roman
Aurrecoechea, J. & Muñoz Villarejo, F. (2001/02). Short Sword. Weapon 51. Oxford: Osprey
‘A legionary workshop of the 3rd century AD Publishing.
specialising in loricae segmentatae from the Bishop, M.C. (2020). Roman Shields. Elite 234.
Roman fortress in León (Spain)’, JRMES Oxford: Osprey Publishing.
12/13: 15–28. Bishop, M.C. & Coulston, J.C.N. (2006). Roman
Băeştean, G. & Barbu, M. (2015). ‘Ulpia Traiana Military Equipment from the Punic Wars to
Sarmizegetusa. About the wooden phase the Fall of Rome, 2nd Edition. Oxford: Oxbow
structures on the Insula 3’, Sargetia. Acta Books.
Musei Devensis (S.N.) 6: 181–92. Brüggler, M., Dirsch, C., Drechsler, M., Schwab,
Bergemann, J. (1990). Römische Reiterstatuen. R. & Willer, F. (2012). ‘Ein römischer
Ehrendenkmäler im öffentlichen Bereich. Schienenarmschutzaus dem Auxiliarlager
Mainz: von Zabern. Till-Steincheshof und die Messingherstellung

62
 in der römischen Kaiserzeit’, Bonner Robinson, H.R. (1974). ‘Problems in
Jahrbücher 212: 121–52. reconstructing Roman armour’, in E. Birley, B.
Burns, M. (2005). The Cultural and Military Dobson & M.G. Jarrett, eds, Roman Frontier
Significance of the South Italic Warrior’s Studies 1969. Cardiff: University of Wales:
Panoply from the 5th to the 3rd Centuries BC. pp. 24–35.
Unpublished PhD thesis, University College Robinson, H.R. (1975). The Armour of Imperial
London. Available at https://core.ac.uk/ Rome. London: Arms & Armour Press.
display/78075788 (accessed 27.6.21). Sim, D. & Kaminski, J. (2012). Roman Imperial
Caruana, I. (1993). ‘A third-century lorica Armour: The Production of Early Imperial
segmentata back-plate from Carlisle’, Arma 5: Military Armour. Oxford: Oxbow Books.
15–18. Stemmer, K. (1978). Untersuchungen zur
Crossland, D. (2020). Oldest Roman body armour Typologie, Chronologie und Ikonographie der
discovered in Germany, The Times 26.9.20. Panzerstatuen. Berlin: Mann.
Available at https://www.thetimes.co.uk/ Stiebel, G.D. (2014). ‘Military equipment’, in
article/oldest-roman-body-armour-discovered- Syon, D., Gamla III. The Shmarya Gutmann
in-germany-p6lsv5jq6 (accessed 15.6.21). Excavations 1976–1989. Finds and Studies,
Curle, J. (1911). A Roman Frontier Post and its Part 1. IAA Reports 56. Jerusalem: IAA:
People. The Fort at Newstead. Glasgow: pp. 57–107.
Maclehose. Syon, D. (2014). Gamla III. The Shmarya
Dennis, G.T. (1985). Three Byzantine Military Gutmann Excavations 1976–1989. Finds and
Treatises. Washington, DC: Dumbarton Oaks, Studies, Part 1. IAA Reports 56. Jerusalem:
Research Library and Collection. IAA.
Dobson, C. (2018). As Tough as Old Boots: Thomas, M.D. (2003). Lorica Segmentata II. A
Essays on the Manufacture and History of Catalogue of Finds. JRMES Monograph 2.
Hardened-Leather Armour. Florence: Dobson. Chirnside: The Armatura Press.
Gabelmann, H. (1973). ‘Römische Grabmonumente Webster, G. (1960). ‘A note on the Roman cuirass
mit Reiterkampfszenen in Rheingebiet’, Bonner (lorica segmentata)’, Journal of the Arms and
Jahrbücher 173: 132–200. Armour Society 3: 194–97.
Garbsch, J. (1978). Römische Paraderüstungen. Weinberg, S.S. (1979). ‘A hoard of Roman armor’,
Munich: Beck. Antike Kunst 22: 82–86.
Groller, M. von (1901). ‘Römische Waffen’, Der
Römischer Limes in Österreich 2: 85–132.
Howard-Davis, C. (2009). The Carlisle
Millennium Project: Excavations in Carlisle,
1998–2001, Volume 2, Lancaster imprints 15,
Lancaster: Oxford Archaeology North.
Lipsius, J. (1630). De Militia Romana Libri
Quinque, Commentarius Ad Polybium.
Antwerp: Officina Plantiniana.
McNally, M. (2011). Teutoburg Forest AD 9: The
destruction of Varus and his legions. Campaign
228. Oxford: Osprey.
Mutz, A. (1987). ‘Die Deutung eines Eisenfundes
aus dem spätrepublikanischen Legionslager
Cáceres el Viejo (Spanien)’, Jahresberichte aus
Augst und Kaiseraugst 7: 323–30.
Picard, G.C. (1980). ‘Cruppellarii’, in H. Le
Bonniec and G. Vallet (eds), Mélanges de
littérature et d’épigraphie latines, d’histoire
ancienne et d’archéologie. Hommages à la Copper-alloy armguard, still articulated, excavated from the
mémoire de Pierre Wuilleumier. Paris: Les auxiliary fort at Till-Steincheshof and dating to the 2nd century ad.
Belles Lettres: pp. 277–80. (Drawing © M.C. Bishop)

63
INDEX
References to illustrations are shown in fittings 14, B(15), 17, 20, C(21), 32, E(33), rings 38, 41, 39, 47
bold. Plates are shown with page locators in 48–49, 50, 52 G(53), 54, 58, 60 rivets 5, 8, 12, 13, 13, 14, 16, 17, 17, 18, 20,
parentheses. France, finds in 5, 12, 12 C(21), 22, 23, 28, 30, 34, 38, 39, 49, 49, 50,
52, G(53), 52, 54–55, 54, 56
Afghanistan, finds in 34 Germany, finds in: Aislingen 19; Aalen Robinson, Henry R. 6, 17, 18, 20, 22, 24, 30,
armguards 6, 8, 31, 31, 34, 34, 49 26; Bertoldsheim 34; Burlafingen 60; 48, 50, 55, 57, 61
articulation, means of 38, 54 Dangstetten 8, 11, 12; Eining 25, 38, 54; Romania, finds in 8, A5(9), 25, 29, 31, 32,
attachment/fastening, means of 38–39 Hofheim 19; Kempten 12; Kalkriese see E(33), 35, 36, 38, 43–44, 54, 60
components/structure 35–36, 35, 36, 38–39, cuirasses; segmental armour; Künzing 44; rosette washers 48, 49, 49, 50, 54, 57
38, 39, 54–55, 63 Mainz 5, 34–35; Neuss 47; Obsrtimms 19; roves 14, 20, 22, 23, 38
repairs/maintenance 52, G(53), 54–55, 56, Rheingönheim 19, 49; Riβtissen 12, 17, 19,
58, H(59) 32, 57; Straubing 44, 45, 46, 47, 47, 51; Scotland, finds in (Newstead) see cuirasses;
wearing of 34–35, 36, F(37), 39, 44, 54, 55 Till-Steincheshof 38, 63; Zugmantel 30, 54 segmental armour
Austria, finds in 12, 17, 26, 29, 30, 42, 42: girth hoops 8, A1–5(9), 14, B(15), 16–17, 18, segmental armour 6
Carnuntum 6, 19, 24, 29, 34, 35, 58, 20, 22, 24–25, 27, 28–30, 29, 30, 32, E(33), collapse/compression/storage of 16, 18, 52,
H(59) 51, 54, 56, 56, 57 G(53), 57, 58, H(59)
auxiliaries, armour of 4, 11, 42, 48, 60 gladiators, armour worn 31, 34, 35, 43, 55, components/units 8, A1–4(9), 10, 12, 13–14,
55, 56 16–17, 18, 20, 22, 22, 24, 25, 28–30, 30,
Belgium, finds in 31, 32 greaves 36, F(37), 43–44, 43, 44, 45, 45, 51 32, 34, 48, 48, 49, 50, 52, G(53), 54, 56,
brass, use of 20, C(21), 24, 48–49, 50, 52, Greece, finds in 40–41, 49, 61 56, 57, 57, 58
G(53) Groller-Mildensee, Maximillian von 6, 54 decoration 14, 30, 48, 49, 50, 51
buckles A1–4(9), 14, B(15), 16–17, 18, 19, 20, design improvements 24–25, 55
22, 51, 58 helmets 6, 7, 8, 14, B(15), 20, C(21), 36, F(37), fasteners/fastenings 14, B(15), 20, 28–30, 52,
Bulgaria, finds in 17 41, 43, 44, 48, 50, 51, 57, 60, 61 G(53), 56, 57, 58
hinges 13, 14, B(15), 16–17, 44, 49, 52, 56: fittings 12, 14, B(15), 16–17, 19, 29, 30, 32,
centurions, armour of 32, E(33), 36, F(37), lobate A1–4(9), 18, 19, 24, 25, 28, 28, 29, E(33), 48–49, 50, 52, G(53), 54, 56, 58
42, 43 30, 49, 50, 54, 57; sub-lobate 16, 50 forms/types: Alba Iulia (‘hybrid’) 8, A5(9),
citizen troops, armour of 4, 5, 5, 58, 60, 61 hook-and-eye fittings 22, 30, 38 31–32, 34; Corbridge 6, 8, A2–3(9),12,
copper, use of 14, 30, 49 horse armour 46–47, 46, 47, 51 14, 16–18, 16, 19, 20, C(21), 22, 22, 23,
copper alloy, use of 5, 7, 8, A1–5(9), 11, 13, Hungary, finds in 32 24, 28, 30, 32, 32, E(33), 38, 48, 48, 49,
14, 16, 17, 18, 20, 22, 25, 28, 29, 30, 35, 50, 51, 52, 52, G(53), 54, 54, 56, 56, 57,
36, 36, F(37), 38, 38, 39, 43, 43, 46, 47, 47, Israel, finds in 17, 22, 23 58, H(59), 61; Kalkriese 8, A1(9), 11–14,
48–49, 48, 50, 51, 52, 54, 56, 63 Italy, finds in 6, 7, 40, 40, 41, 49 11, 12, 13, B(15), 16–17, 18, 20, 30, 50,
cotter pins 28, 34 54, 58; Newstead 6, 8, A4(9), 22, 24–26,
Croatia, finds in 43 laces 14, B(15), 20, 29 25, 26, 26, D(27), 28–30, 28, 29, 30, 32,
cuirasses 10, 12, 18, 28, 32 lames 38, 39 E(33), 34, 35, 36, F(37), 38, 50, 52, G(53),
components/features 6, 20, C(21), 28, 30, 41, leather, use of 7, A1–4(9), 13, 14, B(15), 18, 54, 58, H(59)
49, 52, 54 20, C(21), 22, 28, 29, 30, 34, 35, 38, 39, 44, manufacture/production 48–50
decoration of 30, 41, 41, 42 46–47, 49, 50, 52, G(53), 54, 56, 56, 57, 57 origins/longevity of 8, 10, 11
fastening/joining 20, 30, 41–42, 41 legionaries, armour of 4, 17, 20, C(21), 26, reconstructions/replicas 6, 18, 30, 38, 48, 55,
forms/types 41: Corbridge 12, 16, 17, 32; D(27), 31, 32, E(33), 34, 36, F(37), 43, 44, 57, 57, 61
‘hybrid’ 14, B(15), 31, 31, 32, E(33); mail/ 48, 54, 55, 56, 58, 60, 61, 61 repair/maintenance 20, C(21), 50–51, 52,
scale 8, 14, 41, 56, 58, H(59), 60; muscled G(53), 54–55, 56, 60
4, 6, 7, 10, 32, E(33), 36, F(37), 40–42, 41, mail armour 4, 5, 8, 14, B(15), 24, 31, 31, 32, wearing of 4, 5, 8, 10, 14, B(15), 26, 32,
42, 43, 48, 49, 51–52, 57, 61; Newstead E(33), 34, 36, F(37), 41, 42, 44, 51, 54, 55, E(33), 40, 54, 55–56, 58, 60–61, 61
30, 30, 36, F(37), 54 56, 58, H(59), 60 Serbia, finds in 12, 34
manufacture/production 49 mineralization 13, 14, 18, 20, 39, 49, 57 shoulder doubling 31–32, 41, 56
reconstructions/replicas 30, 61 shoulder sections 22, 23, 57
repairs/maintenance 20, C(21), 52, 54 organic armour 49–50 shoulderguards 5, 8, A1–4(9), 10, 11, 14, 16,
wearing of 4, 14, B(15), 17, 32, E(33), 36, 17, 18, 20, 22, 24, 28, 31, 32, E(33), 49, 50,
F(37), 40, 40, 41, 41, 42, 42, 51, 57, 58, padding/lining, use of 7, 34, 39, 44, 57 52, 54, 56
60, 61 Palestine, finds in 44 silver, use of 50
Curle, James 24 ‘parade armour’ 32 silvering/tinning 14, 32, E(33), 36, F(37), 50
Czech Republic, finds in 32 pectoralia/kardiophylax 7, 8, 8 Slovakia, finds in 29
plates 8, A1–5(9), 20, C(21), 50, 56 Spain, finds in 7, 8, 8, 12, 43, 51: León 25, 29,
Daniels, Charles 18, 61 attachment/joining 13, 14, B(15), 18, 20, 28, 38, 39, 39, 50, 52, G(53)
29, 34, 50 sports armour (cavalry) 51
England, finds in: Broxtowe 19; Carlisle corners/edges, working of A1–4(9), 11, 14, stamping/stamps 48, 49, 51
24–25, 28, 29, 34, 34, 35–36, 38, 39, 44, 17, 18, 20, 30, 38, 50 strap-and-buckle combinations 20
44, 49, 53; Chester 19; Chichester 11, 12, types of: back 10, 12, 13, 17, 18, 20, 22, strap fittings (hinged) 7, 16–17, 19
22, 54; Colchester 17, 19; Corbridge see 23, 24, 25, 28, 30, 32, 34, 41, 54, 54, 57; straps (leather) 8, A1–4(9), 13, 13, 14, 14,
cuirasses; segmental armour; Hod Hill 12, breast 7, 12, 13, 13, 16, 17, 17, 18, 20, B(15), 16, 18, 20, 20, C(21), 22, 22, 23, 28,
17, 19; London 17, 48; 52, 56; Longthorpe 24, 28, 30, 31, 31, 32, 34, 40, 48, 56, 57; 30, 38, 39, 49, 54, 57
19; Oulton 48, 49; Richborough 17, 38; collar 8, A1–4(9), 10, 13, 14, 16, 17, 24, Switzerland, finds in 12, 12, 19, 44
Silchester 19; St Albans 17; The Lunt 19; 28, 52, 52; front/chest 5, 18, 41; mid-
Waddon Hill 11 collar 10, 12, 13, 18, 28, 54; shoulder/ thongs (leather) 14, B(15), 28, 29, 34
shoulderguard 49, 52 tie loops 17, 19, 20, 25, 29, 39, 54
fasteners/fastenings A1–4(9), 13, 14, B(15), Praetorian Guard, armour of 4, 41, 60 tie rings 29, 29, 54
16–17, 17, 18, 20, 24–25, 28, 29–30, 29, 30, pteryges 42, 43, 57–58, 61 turnkeys 28, 28, 30, 34
54, 57, 57, 58
ferrous/iron, use of 8, A1–4(9), 20, C(21), 22, repair/maintenance work 14, B(15), 20, C(21), Wales, finds in 17, 26
28, 29, 30, 35, 36, 38, 39, 40–41, 44, 44, 49, 26, 31, 32, E(33), 51–52, G(53), 54–55, 58, washers 19, 20, 30, 50, 51, 54
50, 52, 54 H(59), 60 workshops 20, 48, 49, 52, G(53)

64
OSPREY PUBLISHING Acknowledgements
Bloomsbury Publishing Plc
Kemp House, Chawley Park, Cumnor Hill, Oxford OX2 9PH, UK I am grateful to Dr Christian Miks of the Römisch-Germanisches
29 Earlsfort Terrace, Dublin 2, Ireland Zentralmuseum for his help in obtaining the photograph of the Numantia
1385 Broadway, 5th Floor, New York, NY 10018, USA pectorale. Dr Stefan Burmeister of Varusschlacht im Osnabrücker Land
E-mail: info@ospreypublishing.com GmbH once again helped with illustrations, including some from the recent
www.ospreypublishing.com new find from Kalkriese. Pascal Lemaire very kindly allowed me to include
his image of the Carlisle backplate. Dr Arik Greenberg was good enough to
OSPREY is a trademark of Osprey Publishing Ltd supply images of his reconstructions, while Drs Fraser Hunter, Holger von
Grawert and Ross Cowan also helped with the sourcing of images. As ever,
First published in Great Britain in 2022 my friend and colleague Dr Jon Coulston allowed me to dip into his vast
This electronic edition published in 2022 by Bloomsbury Publishing Plc collection of images, for which I thank him, almost as much as for agreeing
to read and comment upon an early draft of this text. I am also grateful to
© Osprey Publishing Ltd, 2022 Mariel Hennequin of the Musée Jeanne d’Aboville for images of and
information about the Versigny figurine. Last, but most definitely not least,
All rights reserved. No part of this publication may be reproduced or thanks are also due to Dr Frances McIntosh of English Heritage’s Corbridge
transmitted in any form or by any means, electronic or mechanical, Roman Site and Museum for arranging for some of the Corbridge Hoard
including photocopying, recording, or any information storage or retrieval armour to be examined with a pXRF machine to confirm a long-harboured
system, without prior permission in writing from the publishers. suspicion of mine.

A catalogue record for this book is available from the British Library.
Artist’s note
ISBN: PB 9781472851871; eBook 9781472851864; Readers may care to note that the original paintings from which the colour
ePDF 9781472851840; XML 9781472851857 plates in this book were prepared are available for private sale. All
reproduction copyright whatsoever is retained by the publishers. All
Index by Rob Munro enquiries should be addressed to:
Typeset by PDQ Digital Media Solutions, Bungay, UK
info@g-rava.it
Osprey Publishing supports the Woodland Trust, the UK’s leading woodland
conservation charity. The publishers regret that they can enter into no correspondence upon
this matter.
To find out more about our authors and books visit
www.ospreypublishing.com. Here you will find extracts, author Title-page photograph: Detail of lorica segmentata fittings depicted on
interviews, details of forthcoming events and the option to sign up for Trajan’s Column. From a cast in the Museo della Civiltà Romana. (Photo
our newsletter. © M.C. Bishop)