The Imperial Roman Principate Army


The professionalization of the Roman army after Marius’ reforms led directly to the use and abuse of consular power by individual generals seeking to usurp the power of the Senate. Consequently the last five decades of the Republic were characterized by two important features: the jostling for power and status by a number of dynamic political players, and the calamitous civil wars generated by their personal, be it selfish or altruistic, ambitions. It was the last of these republican warlords who was to emerge victorious as the first Roman emperor under the new name of Augustus. Officially he was addressed as princeps, that is the first citizen of the state, and his reign was the beginning of the Principate.

The army of the Principate established by Augustus drew heavily on the nomenclature and traditions of the dead Republic. But it was new. He decided to meet all the military needs of the empire from a standing, professional army, so that there was no general need to raise any levies through conscription (dilectus), which in actual fact he did on only two occasions, namely following the military crises in Pannonia (AD 6) and Germania (AD 9). Military service was now a lifetime’s occupation and career, and pay and service conditions were established that took account of the categories of soldier in the army: the praetorians (cohortes praetoriae), the citizen soldiers of the legions (legiones), and the non-citizens of the auxiliaries (auxilia). Enlistment was not for the duration of a particular conflict, but for twenty-five years (sixteen for the praetorians), and men were sometimes retained even longer. At the end of service there was a fixed reward, on the implementation of which the soldier could rely. The loyalty of the new army was to the emperor, as commander-in-chief, and neither to the Senate nor the Roman people.

Cassius Dio, writing of the events of 29 BC, reports two speeches made before Augustus by his counsellors, Marcus Vipsanius Agrippa and Caius Maecenas, in which the best way of securing the continuation of the Roman state and defence of its empire was discussed. Agrippa apparently advocated the retention of the traditional system (by which men would be conscripted to serve short periods, and then released into civilian life). Maecenas, on the other hand, argued for ‘a standing army (stratiôtas athanatous in Cassius Dio’s Greek) to be recruited from the citizen body [i.e. legiones], the allies [i.e. auxilia] and the subject nations’, and despite Agrippa’s contention that such an army could form a threat to the security of the empire, carried the day.

Dialogues were a convention of ancient historiography, and these speeches need not be judged the true record of a real debate between the two. In part at least they reflect the political situation of Cassius Dio’s own time and were aimed at a contemporary emperor, perhaps that psychopathic fratricide and builder of the eponymous baths in Rome, Caracalla (r. AD 211–217). Nevertheless, in 13 BC, after he had returned from Gaul, Augustus ordained that terms of service in the legions should in future be fixed at sixteen years, to be followed by a four or five-year period ‘under the flag’, sub vexillo, to be rewarded by a fixed cash gratuity, though this could be commuted to a plot of land, measuring 200 iugera (c. 50 ha), in a veteran-colony in the provinces.

However, the scheme did not work, for in AD 5 discontent was rife and in the following year major army reforms were carried out by the emperor. The fundamental problem was that veterans were discontent with sub vexillo, which apparently entitled them to lighter duties after their sixteen-year stint. But no government, ancient or modern, is noted for keeping its promises. So some alterations were made to the conditions of service. The number of years that the new recruit had to serve under arms was raised to twenty years, with a further period (not specified, but probably at least five years) in reserve. The cash gratuity was now fixed at 12,000 sestertii (3,000 denarii) for an ordinary ranker, a lump sum the equivalent of more than thirteen years’ pay.

Seemingly as part of this same package, but recorded by Cassius Dio under the following year (AD 6), Augustus masterminded the creation of a military treasury (aerarium militare). Its function was to arrange the payment of bounties to soldiers. Augustus opened the account with a large gift of money from his own funds, some 170 million sestertii according to his own testimony, but in the longer term the treasury’s revenues were to come from two new taxes imposed from this time onwards on Roman citizens: a 5 per cent tax on inheritances and a 1 per cent tax on auction sales in Rome. The introduction of these taxes caused uproar, but taxation was preferable to displacement, acrimony and ruin, which had been the consequences of land settlement programmes of the civil war years. Augustus thus shifted a part of the cost of the empire’s defence from his own purse to the citizenry at large. But the wages of serving soldiers (225 denarii per annum for an ordinary ranker) continued to be paid by the imperial purse; Augustus could brook no interference, or divided loyalties there. The management of the army, particularly its pay and benefits, were from the start one of what Tacitus calls ‘the secrets of ruling’. Power was protected and preserved by two things, soldiers and money. And so the security and survival of the emperor and his empire was now the sole responsibility of the emperor and his soldiers.

The legions had been the source of Augustus’ power. However, serious mutinies broke out in Pannonia and Germania in AD 14 partly because the legionaries were worried about their conditions of service after the death of Augustus, so closely had he become associated with their emoluments. But there was obviously significant discontent with low rates of pay, especially in contrast to the praetorians, long service, and unsuitable land allocations. Here Tacitus takes up the story:

Finally Percennius had acquired a team of helpers ready for mutiny. Then he made something like a public speech. ‘Why’, he asked, ‘obey, like slaves, a few commanders of centuries, fewer still of cohorts? You will never be brave enough to demand better conditions if you are not prepared to petition – or even threaten – an emperor who is new and still faltering.

Percennius, a common soldier, was the ringleader of the mutineers in Pannonia, then garrisoned by three legions (VIII Augusta, VIIII Hispana, XV Apollinaris) based in a camp near Emona (Ljubljana). Once the mutiny was crushed, he was to be hunted down and executed for his troubles.

These mutinies clearly showed the danger of having too many legions (there were four involved in the Germania mutiny) in the same camp. Also, living in tents, even during the summer months, on the Rhine and Danube frontiers must have been miserable to say the least. The bleakness of life under canvas is the subject of a telling passage of Tertullian: ‘No soldier comes with frolics to battle nor does he go to the front from his bedroom but from tents that are light and small, where there is every kind of hardship, inconvenience, and discomfort.’

As mentioned above, in the time of Augustus the annual rate of pay for a legionary was 900 sestertii (225 denarii), Percennius’ piddling ‘ten asses a day’. But Percennius’ complaint was all in vain, the basic rate remaining so until Domitianus, who increased the pay by one-third, that is, to 1,200 sestertii (300 denarii) a year. Wages were paid in three annual instalments, the first payment being made on the occasion of the annual new year parade when the troops renewed their oath to the emperor. Official deductions were made for food and fodder (for the mule belonging to the mess-group, contubernium). In addition, each soldier had to pay for his own clothing, equipment and weapons, but these items were purchased back by the army from the soldier or his heir when he retired or died. These were the official charges. As we know, Tacitus records that one of the complaints of the mutineers was that they had to pay sweeteners to venal centurions in order to gain exemption from fatigues. Another complaint was that time-expired soldiers were being fobbed off with grants of land in lieu of the gratuity of 12,000 sestertii, and these plots tended to be either waterlogged or rock-strewn.


A great body of information on the unit size and organization of the Principate army has been amassed by the patient work of several generations of scholars. The literary sources are often obscure or contradictory on the details of unit structures, but we are fortunate in that much information has been derived from epigraphic, numismatic and papyrological record as well as that of archaeology. Here contemporary evidence, if not overabundant, is explicit and reliable. As a result a fairly coherent picture of the army’s structure has emerged and what follows, then, is the briefest of sketches of the army as it existed in Neronian times.

As an instrument of war the Principate army presented a powerful picture, and there is certainly little about it that a modern infantry soldier would fail to recognize. The professional standing force of a modern size, conscription, military training, institutionalized discipline, weapons factories, administrative and combat staffs, military maps, roads, logistics systems, military hospitals, intelligence services, communications, strategy and tactics, efficient killing technologies, siege machines, rank structures, scheduled promotions, permanent records, personnel files, uniforms, regular pay, and even military pension schemes – to name but a few – had already become part of every day, military life.


Men had a thousand reasons for joining the army, but mainly they were escaping from poor local conditions or looking for what they hoped would provide a regular source of food and income. Unsurprisingly, therefore, the army seems to have been most attractive as a definite career to the poorest citizens. For such men, long underfed and ill-clothed, the legions offered a roof over their head, food in their bellies, and a regular income in coin. Basic military pay was not the road to riches, but there was always the chance of bounties and donatives, and the certainty of a discharge bonus, a rich contrast to civilian unemployment. Army pay certainly did not depend on the weather, taxation, rent, interest payments or fluctuating prices. Overall, a soldier’s life was more secure than that of an itinerant labourer (an unpaid labourer would starve; an unpaid soldier still ate), and he enjoyed a superior status too.

Of course, we must remember the harsher side of such a career. The rewards of army life may have been greater, but so were the risks. A soldier ran the risk of being killed or crippled by battle or disease, but also on an everyday basis was subject to the army’s brutal discipline. And then there was maltreatment, which did not include the routine harshness or the standard Spartan quality of military life. The dividing line between discipline and maltreatment was crossed when officers treated their men with unnecessary severity, when they paid no attention to their welfare, and when they expected fear rather than respect from their men. Such officers firmly believed that you got more out of men by using brutality, than by treating them with patience tempered by firmness. Most of us are familiar with the martinet centurion Give-me-Another, nicknamed because of his habit of beating a soldier’s back until his gnarled vitis – the twisted vine-stick that was his badge of rank – snapped and then shouting for a second and a third.

Such a bully and a beast was common in the army, the general assumption being that soldiers had to be treated roughly so as to toughen them up for fighting, yet to many people in the empire who struggled to survive at subsistence level, the well-fed soldier with his ordered existence in his well-built and clean camp must have seemed comfortably off. Soldiers also shared a comradeship with their fellow soldiers, which was often warm and comforting. And so the legions became permanent units with their own numbers and titles and many were to remain in existence for centuries to come.

Most prominent in the life of the empire was the army, the organization of which began, and almost ended, with the legion. Yet from Augustus onwards the emperor commanded no more than twenty-five legions in total (twenty-eight before the Varian disaster of AD 9), which seems paltry considering the extent of the empire. Legions were probably in the order of 5,000 men strong (all ranks) and composed of Roman citizens, though sickness and death could quickly pare away at this figure. Legionaries were mostly volunteers, drawn initially from Italy (especially the north), but increasingly from the provinces. As the first century progressed, many recruits in the west were coming from the Iberian provinces, Gallia Narbonensis, and Noricum, and in the east from the Greek cities of Macedonia and Asia. Thus, by the end of the century the number of Italians serving in the legions was small. Statistics based on nomenclature and the origins of individuals show that of all the legionaries serving in the period from Augustus to Caius Caligula, some 65 per cent were Italians, while in the period from Claudius to Nero this figure was 48.7 per cent, dropping even further to 21.4 per cent in the period from Vespasianus to Trajan. Thereafter, the contribution of Italians to the manpower of the legions (but not of the Praetorian Guard naturally) was negligible. It must be emphasized, however, that these statistics represent all legionaries in the empire. In reality, there was a dichotomy in recruitment patterns between the western and eastern provinces, with legions in the west drawing upon Gaul, Iberia, and northern Italy, while those stationed in the east very quickly harnessed the local resources of manpower.

Ordinarily a legion consisted of ten cohorts (cohortes), with six centuries (centuriae) of eighty men in each cohort, apart from the first cohort (cohors prima), which from AD 70 or thereabouts was double strength, that is five centuries of 160 men. Commanded by a centurion (centurio) and his second in command (optio),4 a standard-size century (centuria) was divided into ten eight-man subunits (contubernia), each contubernium, mess-group, sharing a tent on campaign and pair of rooms in a barrack block, eating, sleeping and fighting together. Much like small units in today’s regular armies, this state of affairs tended to foster a tight bond between ‘messmates’. There would have been a strong esprit de corps among men built upon the deep concern each had for everyone. In the pressure cooker environment of small combat units where soldiers are forced into close contact with one another, they worked together, they fought together, they shared discomfort and death and victory. This was man-to-man friendship, a gutsy bond. A spirit of military brotherhood would explain why many soldiers (milites) preferred to serve their entire military career in the ranks despite the opportunities for secondment to specialized tasks and for promotion. Nonetheless, a soldier (miles) who performed a special function was excused fatigues, which made him an immunis, although he did not receive any extra pay.

Finally, there was a small force of 120 horsemen (equites legionis) recruited from among the legionaries themselves. These equites acted as messengers, escorts and scouts, and were allocated to specific centuries rather than belonging to a formation of their own. Thus, the inscription on a tombstone from Chester-Deva describes an eques of legio II Adiutrix pia fidelis as belonging to the centuria of Petronius Fidus. Citizen cavalry had probably disappeared after Marius’ reforms, and certainly was not in evidence in Caesar’s legions. However, apart from a distinct reference to 120 cavalry of the legion in Josephus, the equites seem to have been revived as part of the Augustan reforms.


When territory was added to the empire, a garrison had to be put together to serve in its defence. New legions were sometimes raised, but normally these green units were not themselves intended for service in the new province. So when an invasion and permanent occupation of Britannia became a hard possibility under Caius Caligula, two new legions, XV Primigenia and XXII Primigenia, were formed in advance. Their intended rôle was as replacements for experienced legions earmarked to join the invasion force: XV Primigenia to release legio XX from Neuss-Novaesium, and XXII Primigenia to release XIIII Gemina from Mainz-Mogontiacum. The invasion force that eventually sailed for southern Britannia in the summer of AD 43 consisted of XX and XIIII Gemina, along with II Augusta, which had been at Strasbourg-Argentoratum, this camp was now left vacant, and VIIII Hispana from Sisak-Siscia in Pannonia, which may have accompanied the outgoing legate governor, Aulus Plautius, on his journey to take up his new post as the expeditionary commander. It must be said, however, that only II Augusta and XX are actually attested as taking part in the invasion itself, though all four legions are recorded very early in Britannia.

Nevertheless, transfers of legions to different parts of the empire could leave long stretches of frontier virtually undefended, and wholesale transfers became unpopular as legions acquired local links. An extreme case must be that of II Augusta. Part of the invasion army of AD 43, its legatus legionis at the time was in fact the future emperor Vespasianus, this legion was to be stationed in the province for the whole time Britannia was part of the empire. An inscription from near Alexandria, dated AD 194, is of particular interest to us as it records the names of forty-six veterans of legio II Traiana fortis who had just received their honourable discharge and had begun their military service in AD 168. Of the forty-one whose origins are mentioned, thirty-two came from Egypt itself and twenty-four of these state the camp as their place of birth, or more precisely origo castris, ‘of the camp’. It is likely that most of them were illegitimate sons born to soldiers from local women living in the nearby canabae legonis, that is, the extramural civilian settlement associated with the garrison. So it seems that many recruits were the sons of serving soldiers or veterans, and in time these soldiers’ sons became a fertile source of recruits, particularly so as soldiers’ sons did not have to make a major adjustment from a civilian to a military world. With bastard sons following their soldier fathers into the army, the custom developed of sending not an entire legion to deal with emergencies, but detachments drawn from the various legions of a province. As we have seen, in the year AD 69 legionary detachments played a major rôle in the formation of the Vitellian and Flavian armies.

Detachments from legions operating independently or with other detachments were known as vexillationes, named from the square flag, vexillum, which identified them. Until the creation of field armies in the late empire, these vexillationes were the method of providing temporary reinforcements to frontier armies for major campaigns. And so it was that Cnaeus Domitius Corbulo received a vexillatio from legio X Fretensis, then stationed at the Euphrates crossing at Zeugma, during his operations in Armenia. Later he was to take three vexillationes of a thousand men (i.e. two cohorts) from each of his three Syrian legions (III Gallica, VI Ferrata, X Fretensis) to the succour of Caesennius Paetus, whose army was retreating posthaste out of Armenia. Likewise, despite the disaster to legio VIIII Hispana during the Boudican rebellion, no new legion was despatched to Britannia, but a vexillatio of 2,000 legionaries gathered from the Rhine legions.


Under Augustus the rather heterogeneous collection of auxiliary units, auxilia, serving Rome was completely reorganized and given regular status within the new standing army. Trained to the same standards of discipline as the legions, the men were long-service professionals like the legionaries and served in units that were equally permanent. Recruited from a wide range of warlike peoples who lived just within or on the periphery of Roman control, with Gauls, Thracians and Germans in heavy preponderance, the auxilia were freeborn non-citizens (peregrini) who, at least from the time of Claudius, received full Roman citizenship on honourable discharge after completion of their twenty-five years under arms.

Tacitus tells us that the Batavi, on the lower Rhine, paid no taxes at all, but ‘reserved for battle, they are like weapons and armour, only to be used in war’. The Batavi made capital stuff for a soldier, and from Tacitus we hear of eight cohortes and one ala, nearly 5,000 warriors from the tiny region of Batavia serving Rome at any one time. He also remarks of a cohors Sugambrorum under Tiberius, as ‘savage as the enemy in its chanting and clashing of arms’, although fighting far from its Germanic homeland in Thrace. Further information concerning these tribal levies comes from Tacitus’ account of the ruinous civil war. In April AD 69, when Vitellius marched triumphantly into Rome as its new emperor, his army also included thirty-four cohortes ‘grouped according to nationality and type of equipment’.

Take the members of cohors II Tungrorum for instance, who had been originally raised from among the Tungri who inhabited the northeastern fringes of the Arduenna Silva (Ardennes Forest) in Gallia Belgica. Under the Iulio-Claudian emperors it was quite common for such units to be stationed in or near the province where they were first raised. However, the events of the AD 69, with the mutiny of a large proportion of the auxilia serving on the Rhine, would lead to a change in this policy. After that date, though the Roman high command did not abandon local recruiting, it did stop the practice of keeping units with so continuous an ethnic identity close to their homelands.

As expected, by the late first century, units were being kept up to strength by supplements from the province where they were now serving or areas adjacent to it. Such units retained their ethnic identities and names, even if they enlisted new recruits from where they were stationed. The epitaph of Sextus Valerius Genialis tells us that he was a trooper in ala I Thracum, and his three-part name indicated he was a Roman citizen. But it adds that he was a ‘Frisian tribesman’. So, Genialis came from the lower Rhine, served in a Thracian cavalry unit stationed in Britannia and styled himself a Roman. So after the military anarchy of AD 69, auxiliary cohorts were plausibly made up of a great diversity of individuals of all kinds of nationalities. Nonetheless, despite such conflicting backgrounds and cultures, the Roman military system forged these foreign cohorts into cohesive, aggressive fighting units.

Auxiliary cohorts were either 480 strong (quingenaria, ‘five hundred strong’) or, from around AD 70, 800 strong (milliaria, ‘one-thousand strong’). Known as cohortes peditata, these infantry units had six centuries with eighty soldiers to each if they were quingenaria, or if milliaria had ten centuries of eighty soldiers each. As in the legions, a centurion and an optio commanded a century, which was likewise divided in to ten contubernia.

Now to turn to matters concerning mounted auxilia. Cavalry units known as alae (‘wings’, it originally denoted the Latin-Italian allies, the socii, posted on the flanks of a consular army of the Republic) are thought to have consisted of sixteen turmae, each with thirty troopers commanded by a decurio and his second-in-command the duplicarius, if they were quingenaria (512 total), or if milliaria twenty-four turmae (768 total). The later units were rare; Britannia, to cite a single example, had only one in its garrison. Drawn from peoples nurtured in the saddle – Gauls, Germans, Iberians and Thracians were preferred – every horseman of the alae was well mounted, knew how to ride, and was strong enough and skilful enough to make lethal use of his long straight sword, the spatha. The alae provided a fighting arm in which the Romans were not so adept.

Additionally there were mixed foot/horse units, the cohortes equitatae. Their organization is less clear, but usually assumed, following Hyginus, to have six centuries of eighty men and four turmae of thirty troopers if cohors equitata quingenaria (608 total), or ten centuries of eighty men and eight turmae of thirty troopers if cohors equitata milliaria (1,056 total). An inscription, dated to the reign of Tiberius, mentions a praefectus cohortis Ubiorum peditum et equitum, ‘prefect of a cohort of Ubii, foot and horse’, which is probably the earliest example of this type of unit. It may be worth noting here that this Tiberian unit was recruited from the Ubii, a Germanic tribe distinguished for its loyalty to Rome. In Gaul Caesar had employed Germanic horse warriors who could fight in conjunction with foot warriors, operating in pairs.

Organized, disciplined and well trained, the pride of the Roman cavalry were obviously the horsemen of the alae, but more numerous were the horsemen of the cohortes equitatae. Having served for some time as infantrymen before being upgraded and trained as cavalrymen, these troopers were not as highly paid, or as well mounted as their brothers of the alae, but they performed much of the day-today patrolling, policing and escort duties.


In addition, as in earlier times, there were specialists fulfilling roles in which Roman citizens, better utilized as legionaries, were traditionally unskilled. The best-known of these specialists were archers from Syria and slingers from the Baleares, weapon preferences that were solidly rooted in cultural, social and economic differences.

Among the Romans the bow seems never to have been held in much favour, though after the time of Marius it was introduced by Cretans serving Rome. During our period, however, archers were being recruited from amongst experienced people of the eastern provinces. Like slingers, it is possible they were equipped as regular auxiliaries rather than their exotic appearance on Trajan’s Column would indicate (e.g. scene lxx depicts them with high cheekbones and aquiline noses, wearing voluminous flowing skirts that swing round their ankles). Certainly first-century tombstones show archers in the usual off-duty uniform of tunic with sword and dagger belts, cinguli, crossed ‘cowboy’ fashion.

Also likely is the possibility that individual soldiers within any given unit acquired the necessary ability to use bows, rather than simply relying on specialist units. In his military treatise, among other matters, Vegetius includes a recommendation that at least a quarter of all recruits should be trained as bowmen. Yet, despite his sound advice here, Vegetius says it is the self-bow that will be used in training soldiers in the art of archery. It is assumed, therefore, that the standard of archery was obviously not expected to be the same as that provided by specialists units such as cohors I Hamiorum Sagittariorum, who were trained and experienced in precisely the sort of warfare in which the Principate army was decidedly deficient.


The Roman army is seen as rather standardized in terms of its equipment, especially its armour. This view, however, is rather misleading and we would probably be closer to the truth envisaging a situation where, as long as the soldier had fully functional equipment, then the army did not mind what type it was. I am not suggesting that Roman soldiers went into action looking like the crew of a pirate ship, but as long as their equipment did the job it was required to do and was well maintained, the pattern was of secondary importance. Besides, as with all professional, state-sponsored armies, improvements in equipment took place relatively slowly, necessitating the continued use of material that was of considerable age, even if certain older items, helmets in particular, were relegated to inferior grades of soldier. It may be said with truth of Roman arms that as long as a piece remained in serviceable condition, it continued to be used – generally speaking, given a minimum amount of daily care, weapons and equipment have a long-life span. Thus, many soldiers would not have had top-of-the-line helmets and equipment. All the same, without going into too much detail, the Principate army was made up of a high proportion of armoured fighting men, men who were also as a rule individually more heavily armed and armoured than their adversaries, than any other.

It should be said at this juncture that the Romans were not at all resistant to technological innovation on cultural grounds because the innovation in question originated with a hated enemy; this was a luxury they fully understood that could not be afforded. Indeed, the Romans were adept at cultural borrowing, though they left their own unique stamp on each borrowed idea (and institution) in moulding it to their own purposes.

Finally, when we write of such things, we should not dissociate the weapon from the man who wields it, nor the load from the man who carries it. The legionary, like all professional infantry soldiers before. Finally, when we write of such things, we should not dissociate the weapon from the man who wields it, nor the load from the man who carries it. The legionary, like all professional infantry soldiers before his day and after, was grossly overloaded – alarmingly so according to some accounts. In the evening of his life, Cicero wrote of ‘the toil, the great toil, of the march: the load of more than half a month’s provisions, the load of any and everything that might be required, the load of the stake for entrenchment’. Normally, perhaps, a legionary carried rations for three days, not the two weeks to which Cicero refers (a brilliant orator, yes, a practical soldier, no). However, it has been estimated that the legionary was burdened with equipment weighing as much as thirty-five kilograms if not more. As Edward Gibbon justly says, this weight ‘would oppress the delicacy of a modern soldier’.

Helmet (galea)

Roman helmets, of Celtic inspiration, were made of iron or copper alloy (both bronze and brass are known). Bronze was a more expensive metal, but cheaper to work into a helmet: whereas iron helmets could only be beaten into shape, bronze ones were often ‘spun’ on a revolving former (a shaped piece of wood or stone) from annealed bronze sheets.

Whatever the material or type (viz. Coolus, Imperial Gallic, Imperial Italic), however, the main features were the skull-shaped bowl, a large neck guard to provide protection from blows to the neck and shoulders, cheek pieces to protect the sides of the face – these were hinged so they could move freely – and a brow-guard, which defended against downward blows to the face. The helmet invariably left the face and ears exposed, since the soldier needed to see and hear to understand and follow battlefield commands. Soldiers often punched or scratched their names and those of their centurions onto their helmets to prevent mistaken ownership or indeed theft.

Unlike infantry helmets, cavalry helmets had extensions of their cheek pieces to cover the ears. Often shaped as simulated ears themselves, the extra protection to the face was clearly considered to be more important than some loss of hearing. Also the neck guard was very deep, reaching down to close to the shoulders, but not wide, since this would have made the rider likely to break his neck if he fell from his horse. The cavalry helmet, therefore, protected equally well against blows to the side and the back of the head, vital in a cavalry mêlée when the two sides soon become intermingled.

Hobnailed boots (caligae)

The standard form of military footwear for all troop types, caligae consisted of a fretwork upper, a thin insole and a thicker outer sole. The twenty-millimetre thick outer sole was made up of several layers of cow or ox leather glued together and studded with conical iron hobnails – evidence from Kalkriese, the probable site of the Varian disaster in AD 9, suggests 120 per boot. Weighing a little under a kilogram, the one-piece upper was sewn up at the heel and laced up the centre of the foot and onto the top of the ankle with a leather thong, the open fretwork providing excellent ventilation that would reduce the possibility of blisters. It also permitted the wearer to wade through shallow water, because, unlike closed footwear that would become waterlogged, they dried quickly on the march.

There is that well-known story from Josephus relating to the tragic death of a centurion during the siege of Jerusalem in AD 70. Of Bithynian origin, centurion Iulianus was bravely pursuing the enemy across the inner court of the Temple when the hobnails of his caligae skidded on the smooth marble paving and he crashed to the deck, only to be surrounded and eventually despatched by those he had audaciously chased. This was a possible danger of course, but in most circumstances the hobnails served to provide the wearer with better traction. Not only that, they also served to reinforce the caligae, and to allow the wearer to inflict harm by stomping. More to the point, the actual nailing pattern on the sole was arranged ergonomically and anticipated modern training-shoe soles designed to optimize the transferral of weight between the different parts of the foot when placed on the ground. Experiments with modern reconstructions have demonstrated that, if properly fitted of course, the caliga is an excellent form of marching footwear, and can last for hundreds of kilometres. Much like all soldier’s equipment past and present, caligae would have needed daily care and attention, such as the replacement of worn or lost hobnails or the cleaning and buffing of the fretwork upper.

If there is one thing of importance to an infantryman, it is his feet. Tacitus represents the Flavians during the civil war as demanding clavarium, or ‘nail money’, that is, money to buy hobnails for their boots. At the time of the claim the Flavian legionaries were making their long march to Rome. Suetonius seems to be alluding to a similar claim when, referring to marines (classiarii), who had regularly to march from Ostia or Puteoli to Rome, they demanded of Vespasianus calciarium, or ‘boot-money’. The emperor, famed for his sardonic wit (and measures), is said to have suggested to them they make the journey barefoot, a practice that was still being continued in Suetonius’ day. Soldiers were charged, amongst other things, for their rations, clothing and boots by stoppages debited to their pay accounts and these demands must have been for extra pay to offset the expense of boot repairs.

With such footwear fully laden Roman soldiers could ‘yomp’ for miles, provided of course that they adapted themselves to local conditions. In cold weather, for instance, caligae could be stuffed with wool or fur, while one piece of sculptural evidence depicting a parade of praetorians, namely the Cancellaria Relief, suggests that thick woollen socks (undones), toeless and open at the heel, could also be worn with the boot. We also have that well-known writing tablet from Chesterholm-Vindolanda:

I have sent (?) you… pair of socks (undones). From Sattua two pairs of sandals (soleae) and two pairs of underpants (subligares), two pairs of sandals… Greet… ndes, Elips, Iu…, … enus, Tetricus and all your messmates (contubernales) with whom I hope that you live in the greatest of good fortune.

Written in colloquial Latin, this letter was evidently sent to a soldier serving at Chesterholm-Vindolanda as the author refers to his contubernium, one of which, Elpis, bears a Greek name (literally Hope). The recipient was probably a member of one of three auxilia units known to have been stationed here at the end of the first century, namely cohors III Batavorum, cohors VIIII Batavorum, or cohors I Tungrorum. The mention of socks (and underpants) strongly suggests the writer is a close relative or friend whose concern for the recipient’s material comfort led him or her to send a parcel from home. Even if socks and underpants were not standard issue, provision of additional clothing of this kind is abundantly paralleled in the papyri from Egypt. What is more, evidence of this sort promotes the commonsense idea that Roman soldiers adapted to local conditions, however extreme. Finally, it is interesting to note that the Chesterholm-Vindolanda texts show us how swiftly auxiliary soldiers acquired literate habits and how proficient they became in Latin, a language that was not their native tongue but their Roman military service forced them to acquire. As a point of comparison, Charlemagne was distinguished among his peers because he could write his name; however, there is no evidence that he could read at all.

As a final point, cavalrymen could attach spurs, of iron or bronze, to their caligae. Prick spurs had been worn by Celts of the La Tène period, being evident in late Iron Age Gaulish and central European graves and settlements, and were known in Greece from the late fifth century BC onwards. They were sporadically used by Roman cavalry throughout the empire, especially by troopers on the Rhine and Danube frontiers.

Body armour (lorica)

The Romans employed three main types of body armour: mail (lorica hamata), scale (lorica squamata), and segmented (lorica segmentata) All body armour would have been worn over some kind of padded garment and not directly on top of the tunic. Apart from making the wearer more comfortable, this extra layer complemented the protective values of each type of armour, and helped to absorb the shock of any blow striking the armour. The anonymous author of De rebus bellicis, an amateur military theoretician writing in the late fourth century, describes the virtues of such a garment:

The ancients [i.e. the Romans], among the many things, which… they devised for use in war, prescribed also the thoracomachus to counteract the weight and friction of armour… This type of garment is made of thick sheep’s wool felt to the measure… of the upper part of the human frame…

The author himself probably coined the term thoracomachus (cf. Greek thorâx, breastplate), which seems to be a padded garment of linen stuffed with wool. One illustration on Trajan’s Column (scene cxxviii) depicts two dismounted troopers on sentry duty outside a headquarters who appear to have removed their mail-shirts to expose the padded garment.

Lorica hamata: Mail was normally made of iron rings, on average about a millimetre thick and three to nine millimetres in external diameter. Each ring was connected to four others, each one passing through the two rings directly above and two directly below – one riveted ring being inter-linked with four punched rings. The adoption of lorica hamata by the Romans stems from their having borrowed the idea from the Celts, among whom it had been in use at least since the third century BC, albeit reserved for use by aristocratic warriors only. This early mail was made of alternate lines of punched and butted rings. The Romans replaced the butted rings with much stronger riveted rings, one riveted ring linking four punched rings.

The wearer’s shoulders were reinforced with ‘doubling’, of which there were two types. One had comparatively narrow shoulder ‘straps’, and a second pattern, probably derived from earlier Celtic patterns, in a form of a shoulder-cape. The second type required no backing leather, being simply drawn around the wearer’s shoulder girdle and fastened with S-shaped breast-hooks, which allowed the shoulder-cape to move more easily. The shoulder-cape is indicated on numerous grave markers belonging to cavalrymen, which also show the mail-shirt split at the hips to enable the rider to sit a horse.

Although mail had two very considerable drawbacks – it was extremely laborious to make, and while it afforded complete freedom of movement to the wearer, it was very heavy, weighing anywhere between 10 and 15 kilograms, depending on the length and number of rings (at least 30,000) – such armour was popular. A mail-shirt was flexible and essentially shapeless, fitting more closely to the wearer’s body than other types of armour. In this respect it was comfortable, whilst the wearing of a belt helped to spread its considerable weight, which would otherwise be carried entirely by the shoulders. Mail offered reasonable protection, but could be penetrated by a strong thrust or an arrow fired at effective range. It was also vulnerable to bludgeoning weapons such as clubs or maces. Worn underneath the lorica hamata was undoubtedly some form of leather or padded protection, without which mail was relatively useless, to dissipate further the force of a blow.

Lorica squamata: Scale armour was made of small plates (squamae), one to five centimetres in length, of copper alloy (occasionally tinned), or occasionally of iron, wired to their neighbours horizontally and then sewn in overlapping rows to linen or leather backing. Each row was arranged to overlap the one below by a third to a half the height of the scales, enough to cover the vulnerable stitching. The scales themselves were thin, and the main strength of this protection came from the overlap of scale to scale, which helped to spread the force of a blow.

A serious deficiency lies in the fact that such defences could be quite readily pierced by an upward thrust of sword or spear; a hazardous aspect of which many cavalrymen must have been acutely aware when engaging infantry. This weakness was overcome, certainly by the second century, when a new form of semi-rigid cuirass was introduced where each scale, of a relatively large dimension, was wired to its vertical, as well as its horizontal, neighbours.

Scale could be made by virtually anyone, requiring patience rather than craftsmanship, and was very simple to repair. Though scale was inferior to mail, being neither as strong nor as flexible, it was similarly used throughout our period and proved particularly popular with horsemen and officers as this type of armour, especially if tinned, could be polished to a high sheen. Apart from those to cavalry, most of the funerary monuments that depict scale armour belong to centurions.

Lorica segmentata: This was the famous laminated armour that features so prominently on the spiral relief of Trajan’s Column. Concerning its origins, one theory suggests that it was inspired by gladiatorial armour, since these fighters are known to have worn a form of articulated protection for the limbs. Part of a lorica segmentata was found at Kalkriese, the probable site of the Varian disaster in AD 9, making this the earliest known example of this type of armour. It is last seen (in a carving) sometime around AD 230 in the time of Severus Alexander – a good run, but of only perhaps two-and-half centuries as opposed to the six centuries many book illustrations and epic films (and, unfortunately, television documentaries) would have us believe.

The armour consisted of some forty overlapping, horizontal curved bands of iron articulated by internal straps. It was hinged at the back, and fastened with buckles, hooks and laces at the front. As the bands overlapped it allowed the wearer to bend his body, the bands sliding one over another. The armour was strengthened with back and front plates below the neck, and a pair of curved shoulder pieces. In addition, the legionary would wear a metal studded apron hung from a wide leather belt (cingulum), which protected the belly and groin. Round the neck was worn a woollen scarf (focale), knotted in front, to prevent the metal plates from chafing the skin.

Superior to mail with regard to ease of manufacture and preservation, but most particularly in view of its weight, this could be as little as 5.5 kilograms, depending on the thickness of the plate used. It was also more resistant to much heavier blows than mail, preventing serious bruising and providing better protection against a sharp pointed weapon or an arrow. Its main weakness lay in the fact that it provided no protection to the wearer’s arms and thighs. Also full-scale, working reconstructions of lorica segmentata have shown that the multiplicity of copper-alloy buckles, hinges and hooks and leather straps, which gave freedom of movement, were surprisingly frail. It may have been effective against attacking blows or in impressing the enemy, but with its many maintenance problems we can understand why lorica segmentata never became standard equipment in the Principate army.

Body shield (scutum)

Legionaries carried a large dished shield (scutum), which had been oval in the republican period but was now rectangular in shape. Besides making it less burdensome, the shortening of the scutum at top and bottom was probably due to the introduction into the army of new combat techniques, such as the famous Roman ‘tortoise’ (testudo), a mobile formation entirely protected by a roof and walls of overlapping and interlocking scuta. On the other hand, auxiliaries, infantrymen and horsemen alike, carried a flat shield (clipeus), with a variety of shapes (oval, hexagonal, rectangular) recorded.

Shields, scuta and clipi equally, were large to give their bearer good protection. To be light enough to be held continually in battle, however, shield-boards were usually constructed of double or triple thickness plywood made up of thin strips of birch or plane wood held together with glue. The middle layer was laid at right angles to the front and back layers. Covered both sides with canvas and rawhide, they were edged with copper-alloy binding and had a central iron or copper-alloy boss (umbo), a bowl-shaped protrusion covering a horizontal handgrip and wide enough to clear the fist of the bearer.

An overall central grip was probably adopted from the Celts, who certainly used such an arrangement at an early date. More specifically, however, such a grip was to encourage parrying with the scutum and not, as mentioned earlier in this book, the gladius. When the scutum was used offensively, the horizontal handgrip allowed for a more solid punch to be delivered as the fist was held in the correct position to throw a ‘boxing jab’. It also meant that the legionary’s elbow was not over extended as the blow was delivered. Such an arrangement, however, provided less balance to the scutum as a whole. If, for instance, the scutum were struck above or below its central plane, it would have been very difficult to prevent the shield from pivoting upon its axis. The resulting movement would have thus created an opening in the legionary’s defence. The vertical handgrip, on the other hand, was more reliable in terms of blocking blows. The vertical orientation of the hand would have enabled the legionary to keep the scutum as straight as possible, even if the scutum were struck above or below the shield-boss. Of course, a less solid punch can be delivered with the fist in a vertical position as compared to the horizontal position.

When not in use shields were protected from the elements by leather shield-covers; plywood can easily double in weight if soaked with rain. Oiled to keep it both pliant and water resistant, the cover was tightened round the rim of the shield by a drawstring. It was usual for it to have some form of decoration, usually pierced leather appliqué-work stitched on, identifying the bearer’s unit. A cavalryman had the luxury of carrying his shield obliquely against the horse’s flank, slung from the two side horns of the saddle and sometimes under the saddlecloth.

So much for the defensive equipment employed by the Romans. Now we must turn our attention to their weaponry. Throughout our chosen period, the main weapons in use were the bow, the sling, the javelin, the spear, the sword, and the dagger. The first three were missile weapons, and the rest were used in the close-quarter mêlée. We will examine each, in the Roman context of course, in turn.

Shafted weapons

Few ancient civilizations eschewed the spear in their arsenals, and in its simplest form, a spear was nothing more than a stout wooden stick with a sharpened and hardened business end, the latter being achieved by revolving the tip in a gentle flame till it was charred to hardness. Beyond that, it was a long, straight wooden shaft tipped with a metallic spearhead, although the sharpened stick with the hardened tip could be quite an effective killing weapon too. Ash wood (as frequently mentioned in the heroic verses of Homer and Virgil) was the most frequently chosen because it naturally grows straight and cleaves easily. Moreover, it is both tough and elastic, which means it has the capacity to absorb repeated shocks without communicating them to the handler’s hand and of withstanding a good hard knock without splintering. These properties combined made it a good choice for crafting a spear.

Heavy javelin (pilum)

Since the mid-third century BC the pilum had been employed by legionaries in battle as a short-range shock weapon; it had a maximum range of thirty metres or thereabouts, although probably it was discharged within fifteen metres of the enemy for maximum effect. The maximum range of any thrown or shot weapon is irrelevant. Man is capable of running a sub-four minute mile, but this does not mean the average man is so able. Similarly, it may be of academic interest to know that a legionary was capable of launching a pilum more than thirty metres. What is of greater interest is the fact that the enemy had to be within twenty metres before the legionary had any real chance of scoring a hit.

By our period the pilum had a pyramidal iron head on a long, untempered iron-shank, some sixty to ninety centimetres in length, fastened to a one-piece wooden shaft, which was generally of ash. The head was designed to puncture shield and armour, the long iron shank passing through the hole made by the head.

Once the weapon had struck home, or even if it missed and hit the ground, the soft-iron shank tended to buckle and bend under the weight of the shaft. With its aerodynamic qualities destroyed, it could not be effectively thrown back, while if it lodged in a shield, it became extremely difficult to remove. Put simply, the pilum would penetrate either flesh or become useless to the enemy. Modern experiments have shown that a pilum, thrown from a distance of five metres, could pierce thirty millimetres of pine wood or twenty millimetres of plywood.

Continuing the practice of the late Republic, there were two fixing methods at the start of our period, the double-riveted tang and the simple socket reinforced by an iron collet. With regards to the tanged pilum, however, there is iconographical evidence, such as the Cancellaria Relief and the Adamklissi Monument, to suggest that a bulbous lead weight was now added under the pyramid-shaped wooden block fixing the shank to the shaft. Presumably this development was to enhance the penetrative capabilities of the pilum by concentrating even more power behind its small head, but, of course, the increase in weight would have meant a reduction in range.

Light spear (lancea)

Auxiliary foot and horse used a light spear (lancea) as opposed to the pilum. Approximately 1.8 metres in length, it was capable of being thrown further than a pilum, though obviously with less effect against armoured targets, or retained in the hand to thrust overarm as shown in the cavalry tombstones of the period.

Even though such funerary carvings usually depict troopers carrying either two lancae or grooms (calones) behind them holding spares, Josephus claims Vespasianus’ eastern cavalry carried a quiver containing three or more darts with heads as large as light spears. He does not say specifically where the quiver was positioned but presumably it was attached to the saddle. Arrian confirms this in his description of an equestrian exercise in which horsemen were expected to throw as many as fifteen, or, in exceptional cases twenty light spears, in one run. Presumably infantrymen carried more than one lancea; a low-cut relief recovered from the site of the fortress at Mainz-Mogontiacum depicts an auxiliary infantryman brandishing one in his right hand with two more held behind his clipeus. Analysis of the remains of wooden shafts shows that ash and hazel were commonly used.

The simple term ‘spearhead’, however, embraces a great range of shapes and sizes, complete with socket ferrules (either welded in a complete circle or split-sided) to enable them to be mounted on the shaft and often secured with one or two rivets and/or binding. Head lengths can vary from six to eight centimetres all the way up to forty centimetres, whilst socket lengths range from seven to thirteen centimetres in length. For use as a stabbing weapon, practical experience tells us that the width of the blade was important, for a wide blade actually prevents the spearhead from being inserted into the body of an enemy too far, thus enabling the spear to be recovered quickly, ready for further use. In our period of study the most common designs were angular blades, with a diamond cross section, and leaf-shaped blades, with a biconvex cross section.

Edged weapons

The modern infantryman is no longer armed exclusively with the rifle and bayonet. His tactical repertoire has expanded to include hand grenades, pistols, daggers, and sharpened entrenching tools. If he is a specialist, he also masters a machine gun, a light mortar, or an anti-armour weapon. The Roman legionary, on the other hand, was a specialist with one weapon only, the gladius.

Short sword (gladius)

It would be a pointless sort of exercise to debate the relative importance of the different swords of the fighting forces throughout the period of Rome’s domination of the Mediterranean world, whether it was single or double edged, short or long, heavy or light, straight or curved, and so forth, which played the most vital rôle, but certain it is that, as Rome expanded, the gladius was making an increasing contribution to the winning of empire. Against the unarmoured formations frequently met with by Roman legions it was a most potent weapon.

Of all the weapons man has seen fit to invent, the spear was probably the one most universally used during our period of study. Not so, however, with the Roman legionary, who was primarily a swordsman trained in the science and skill of a very singular sword. It goes without saying that the use of weapons has transformed man’s capacity to inflict violence. The capacity of man-as-animal to deliver violence is very limited, as a psychologist appropriately named Gunn describes vividly:

It is extremely difficult for one naked unarmed man to kill another such man. He has to resort to strangulation, or to punching him hard enough to knock him over so that he may gash open an important vessel with his teeth, or break his head open by dashing it against the ground … Weapons magnify the aggressiveness of a creature many times.

It had been back in the third century BC when the Romans had adopted a long-pointed, double-edged Iberian weapon, which they called the gladius Hispaniensis (‘Iberian sword’), though the earliest specimens date to the turn of the first century BC. In our chosen period the gladius was employed not only by legionaries, but by auxiliary infantrymen too.

Based on gladii found at Pompeii and on several sites along the Rhine and Danube frontiers, Ulbert has been able to show that there were two models of gladius, the one succeeding the other. First was the long-pointed ‘Mainz’ type, whose blade alone could measure sixty-nine centimetres in length and six centimetres in width, and is well-evidenced in the period from Augustus to Caius Caligula. The ‘Pompeii’ type followed this, a short-pointed type that replaced it, probably during the early part of Claudius’ principate. This pattern was shorter than its predecessor, being between forty-two and fifty-five centimetres long, with a straighter blade 4.2 to 5.5 centimetres wide and short triangular point. Whereas the ‘Mainz’ type weighed between 1.2 and 1.6 kilograms, the ‘Pompeii’ type was lighter, weighing about a kilogram. The blade of both types was a fine piece of blister steel with a sharp point and honed down razor-sharp edges and was designed to puncture armour. It had a comfortable bone handgrip grooved to fit the fingers, and a large spherical pommel, usually of wood or ivory, to help with counterbalance. In the hands of the trained soldier it was a formidable instrument of destruction.

This sword lacked a typical cross hilt, which tells us it was not a parrying weapon. There is little doubt, therefore, that the gladius was designed primarily as a short-range thrusting weapon, that is, the intention was to deliver the fatal blow by a thrust. Accordingly, the sword blade was straight and the handle short, while the main design feature was the point. The short length of the weapon allowed for ease of control, while its comparative lightness enhanced the swordsman’s reaction time. It was weighted more closely to the actual hilt than a standard sword so as to discourage the slash and to encourage the thrust. This arrangement also helped to increase the speed of the thrust. In addition, the oversized pommel increased the weight at the back end of the sword, thus increasing the power behind the thrust. Finally, the oval-shaped cross-guard just above the handgrip allowed the swordsman to apply maximum pressure to the thrust, and this, combined with a contoured grip to fit the hand snugly, increased the potency of the thrust. The handgrip itself was some 7.5 to 10 centimetres in length, that is, slightly longer than the width of a hand. This arrangement allowed for better efficiency and control in weapon handling.

Unusually, legionaries and auxiliaries carried their sword on the right-hand side suspended by the cingulum worn around the waist. The wearing of the sword on the right side goes back to the Iberians, and before them, to the Celts. The sword was the weapon of the high-status warrior, and to carry one was to display a symbol of rank and prestige. It was probably for cultural reasons alone, therefore, that the Celts carried their beloved weapon, the long slashing sword, on the right side. Customarily a sword, when not in the right hand, was worn on the left, the side covered by the shield, which meant the weapon was hidden from view. However, the Roman soldier wore his sword on the right hand not for any cultural reason. As opposed to a scabbard-slide, the four-ring suspension system on the scabbard enabled him to draw his weapon quickly with the right hand, an asset in close-quarter combat. Inverting the fighting fist to grasp the hilt and pushing the pommel forward drew the gladius with ease. More to the point, it also meant the sword arm was not exposed when doing so.

For the Roman soldier, the gladius fitted his temperament perfectly. It was a close-quarter weapon and the Roman was never one to stand around, engaging in distant firefights, but rather he sought every opportunity to engage his foe in a mêlée. Surely another sword type would never take the place of the Queen of Death, the gladius.

Long sword (spatha)

Cavalrymen, on the other hand, used a longer, narrower double-edged sword, the spatha, which followed Celtic types, with a blade length from 64.5 to 91.5 centimetres and width from four to six centimetres. The middle section of the blade was virtually parallel-edged, but tapered into a rounded point. It was intended primarily as a slashing weapon for use on horseback, and it usually was so employed, though the point could also be used for thrusting if need be.

In spite of its length, the spatha was worn on the right side of the body, as Josephus says, and numerous cavalry tombstones confirm, suspended from a waist belt or baldric whose length could be adjusted by a row of metal buttons. At the turn of the second century, however, the spatha started to be worn on the left side, although not exclusively so.

Military dagger (pugio)

The pugio – a short, edged, stabbing weapon – was the ultimate weapon of last resort. However, it was probably more often employed in the day-to-day tasks of living on campaign. Carried on the left-hand side and suspended on the same cingulum that carried the sword (though two separate belts crossed ‘cowboy’ style was a dashing alternate), the pugio was slightly waisted in a leaf-shape and some 20 to 25.4 centimetres long. The choice of a leaf-shaped blade resulted in a heavy weapon, to add momentum to the thrust. Like the gladius, the pugio was borrowed from the Iberians and then developed; it even had the four-ring suspension system on the scabbard, characteristic of the gladius.

The pugio, worn by legionaries and auxiliaries alike, was obviously a cherished object. Soldiers seldom wasted time on aesthetics. Their equipment, whether it belonged to an individual or to the state, remained functional, the pila, gladii, scuta – all such equipage were primarily utilitarian. Not so, it seems, the pugionis. The highly decorative nature of Roman daggers of our period, and particularly their sheaths, suggests that even common soldiers were prepared to spend considerable sums of money on what could be classified as true works of art. Though remaining an effective fighting weapon, the pugio was plainly an outward display of its wearer’s power.

Missile weapons


Whereas bows of self-wood construction were made exclusively of wood in one or more pieces, the composite bow had an elaborate network of sinew cables on its back to provide the cast, with a horn belly to provide the recovery of the bow. The whole was built up on a wooden core with bone strips stiffening both the grip and the extremities (the ‘ears’), and then covered with skin or leather and moisture-proof lacquers. The back of the bow was that side away from the archer, which was stretched in stringing and stretched even more when the bow was drawn. The belly was the side nearest the archer when drawn, which was compressed by stringing and compressed even more by drawing.

The design of a composite bow thus took full advantage of the mechanical properties of the materials used in its construction: sinew has great tensile strength, while horn has compressive strength. When released the horn belly acted like a coil, returning instantly to its original position. Sinew, on the other hand, naturally contracted after being stretched. This method of construction and the materials employed thus allowed the bow to impart a greater degree of force to the arrow when fired, compared with the wooden self-bow of the same draw weight. This latter characteristic enabled the archer to choose between two tactical options, depending on the need of the moment. He either could deliver a lightweight projectile over a distance twice what a wooden self-bow could shoot, or deliver a projectile of greater weight or force at short range when the capacity to pierce armour or to thoroughly disable an opponent was needed.

The actual range and performance of the composite bow is open to debate, and a number of varied figures have been suggested. Vegetius recommended a practice range of 600 Roman feet (c. 177 metres), while later Islamic works expected an archer to display consistent accuracy at 69 metres. Modern research tends to place an accurate range up to 50 to 60 metres, with an effective range extended at least 160 to 175 metres, and a maximum range at between 350 and 450 metres. Range is as much dependent upon the man as the bow and coupled with this was bow quality – the better made the composite bow; the more tailored to the individual archer’s height, draw weight and length, the better the performance.

That is not to say that the composite bow was the ‘ultimate weapon’. Draw weights of 150 pounds or more taxed even a very strong man and a high rate of fire could not be sustained indefinitely. Unlike firearms or weapons such as the crossbow, which store chemical or potential energy and, by releasing this, propel their missile, a bow converts the bodily strength of the firer into a force propelling the arrow. This skill is much harder to learn than the use of a crossbow or a firearm such as an arquebus, which only has to be loaded and aimed to be effective. Indeed, where a few hours sufficed to teach an ordinary fellow to draw and aim a crossbow, many years of constant practice and a whole way of life were needed to master the use of the composite bow. Archery did not consist of a series of mechanical movements that anyone could be trained to execute, but of a whole complex cultural, social and economic relationship.

Other externalities to consider are those of the accuracy and the effectiveness. The target’s size and rate of movement, as well as the skill of the individual archer, governed accuracy. In actuality the archer would have practised shooting at a stationary target. A trained archer, when fresh, could get off six aimed shots in a minute with ease. Perhaps the best modern estimate of his long-range accuracy comes from W.F. Paterson, himself an experienced archer. Paterson concludes that a good shot on a calm day could be expected to hit a target the size of a man on horseback at 280 yards about once every four shots. At a hundred metres such an archer would have been able to pick his man. At fifty he would have been sure of death.

However, whilst this level of accuracy – the ability to pick-off individuals – would have been useful in the archer’s role as a skirmisher his main task was to stand behind his own infantry and shoot, indirectly, at a unit rather than an individual. In this case accuracy was more concerned with all of the arrows arriving on the target area at the same time, than with individual marksmanship. Thus, an effective shot was not necessarily an accurate one, and in pitched battle a murderous barrage was possible even without aiming.34 When Titus formed a fourth rank of archers behind three of legionaries, these men could have had little or no opportunity of seeing their target. They would have been firing blind, over the heads of the men in front, hoping to drop their arrows down from a high trajectory onto the general area occupied by the enemy and spread death and injuries amongst their ranks.


The arrow penetrates because of the concentration of its kinetic energy behind a narrow cutting edge. The types of arrowheads used by the Principate army reflected these different penetrative requirements. Bodkin heads or flat-bladed types were used for armour piercing, be the target shield or body armour, acting on the same principles as the pilum. On the other hand, trilobate, triple-or quadruple-vaned were more effective against unarmoured horses and men. A bow relies on sending an arrow deep into the body of its victim. As the arrow passes through it severs blood vessels and major arteries. Since arrows ordinarily kill by bleeding, projectiles that could open up a large wound on an unarmoured opponent were useful. A third category was designed to carry incendiary material, such as a naphtha-soaked ball of twine with wood splints. Five such arrowheads were recovered from the fort at Bar Hill, Antonine Wall. They have three projecting wings that form an open iron framework capable of holding inflammable material.

In terms of arrow technology heads were mainly tanged, although the Romans also employed socketed types as well, and were glued and bound to the shafts. The shafts were made from wood, cane or reed. The latter is one of the best materials, having a combination of lightness, rigidity and elasticity ideal for shafts. Reeds are also already well adapted to their aerodynamic role as shafts by their need while growing to maintain an evenly round profile to reduce wind drag, as well as by having the elasticity and strength to bend and return to upright position. This latter adaptation is critical as shafts need to be able to bend round the bow during shooting, and flex so the tail swings clear of the bow before swinging back into line as the arrow travels along the line it was aimed at the moment of release. Where cane or reed was used the arrowhead was first attached to a wooden pile, which was then glued onto a cane or reed shaft. The piles reduced the risk of the cane or reed splitting on impact, which would, if it occurred, reduce the arrow’s penetrative power. The remains of cane or reed arrows from Doura-Europos show that the surface of the shaft was roughened and the fletching glued onto this surface. The lower portions of three shafts actually bear painted markings, and it has been suggested that the arrows were marked to identify their owners or to denote a matching set.

Because human ribs are horizontal and the bow held vertically, the head of a war arrow is often perpendicular to the plane of the notch, so that the point may pass more easily between the ribs. For the same reason, the plane of the head matches that of the notch in hunting arrows. Modern experiments have shown that socketed arrows tended to break behind the socket itself. Whether this action was deliberately intended or simply an accidental by-product of manufacture is hard to ascertain. Tanged arrowheads proved less likely to break on impact, the sinew binding required to hold the head in place limiting their degree of penetration. These field trials were conducted against all known forms of Roman armour. Not surprisingly, lorica hamata proved to be the easiest to penetrate, followed by the lorica segmentata and finally the lorica segmentata, in which none of the arrowheads penetrated to a depth sufficient to cause a fatal wound even at a range of seven metres. Somewhat surprisingly the wooden shield, especially if covered with leather, almost provided as much defence.


Slingers normally served as a complement to archers; the sling not only outranged the bow but a slinger was also capable of carrying a larger supply of ammunition than an archer. Slingshots were not only small stones or pebbles, but also of lead, acorn or almond shaped, and weighing some eighteen to thirty-five grams. Unique to the Graeco-Roman world, these leaden bullets, the most effective of slingshots, could be cast bearing inscriptions, such as symbols – a thunderbolt emblem was popular – or a short phrase, usually only a few letters. Some of these may be an invocation to the bullet itself or invective aimed at the recipient – AVALE, ‘swallow this’. Doubtless the last had a double meaning as the Latin term for lead sling bullet is glans (pl. glandes), which literally means ‘acorn’ but can also refer to the head of the penis. Other inscriptions, usually cut by hand in longhand writing rather than impressed into the two-part moulds from which the bullets were made, bear witness to the soldier’s jesting filth.

The sling itself, as deadly as it was simple, was made of a length of non-elastic material such as rush, flax, hemp or wool. It comprised a small cradle to house the bullet, and two braided cords, one ending in a knot and the other with a loop. In all techniques of slinging, the slingshot was placed in the centre of the cradle, the loop was secured to the second finger of the throwing hand and the knot between the thumb and forefinger of the same hand. In the Mediterranean world the common casting technique was the horizontal whirl, as described by Vegetius. The sling was held with the throwing arm raised above the shoulder. The slingshot was placed in the cradle, and the hand and sling allowed to drop backwards over the head, and whirled around the head clockwise (from the slinger’s point of view) a number of times (Vegetius only allows a single whirl around the head) until sufficient momentum was built up. The knot was then released perpendicular to the body so the slingshot flew out straight ahead, its range being related to the angle of discharge, the length of the whirling cords and the amount of kinetic energy imparted by the thrower.

The potential range of the sling has been a the subject of considerable debate and speculation. The problem lies with the testimony of the soldier-scholar Xenophon, who claimed that the leaden missiles of the Rhodian slingers carried ‘no less than twice as far as those from Persian slings’, the latter slinging a stone ‘as large as the hand can hold’. The more conservative estimates are around the 200 metres mark, while 350 metres, 400 metres, and even 500 metres, have been suggested as possibilities for the maximum range of a leaden bullet. On the practical side, Korfmann observed for himself Turkish shepherds, males who grew up in a sling-using milieu, sling ordinary pebbles. In five out of eleven trials the pebbles reached 200 metres, and three of the best casts were between 230 and 240 metres.

Fast-moving slingshot could not be seen in flight and did not need to penetrate armour to be horrifically effective. Vegetius writes that slingshot was more effective than arrows against armoured soldiers in armour, since they did not need to penetrate the armour in order to ‘inflict a wound that is still lethal’. A blow from a slingshot on a helmet, for instance, could be enough to give the wearer concussion.

Extra slings were normally carried, and those not employed were normally tied round the head or the belly. Slingshot was carried in a bag slung over the shoulder.