Praetorian guardsmen, on campaign; early 1st Century AD.
(31 BC–AD 14)
Octavian turned the praetorians into a regularized, privileged and organized part of his power base as Augustus, Rome’s first emperor. The Praetorian Guard enjoyed advantageous pay and conditions, but by dispersing them around Rome and Italy Augustus avoided creating the impression that he ruled at the head of a military dictatorship. Augustus focused far more publicly on his constitutional position and playing down the fact that he had come to power, and remained in power, as the result of naked force. He needed the army and the Guard, and they needed him. This interdependence established a dynamic that would have significant consequences in the centuries to come. The essential post of the praetorian prefect was created, but this involved putting a man, or men, in charge of a potentially very dangerous force. Augustus’ attempts to create a dynasty also raised the question of whether the Guard was loyal to the office or to the person of the emperor.
During the civil war after Caesar’s death in 44 BC an important principle and concept had been established: a Roman general would equip himself with a bodyguard drawn from his most effective, experienced and reliable troops. Now that he had supreme power, Octavian moved fast to make himself a permanent institution. After the Battle of Actium in 31 BC, Octavian was unchallenged. The permanent Praetorian Guard which he created at this point, made up from both his praetorians and Antony’s, emerged before the end of his reign as a far more regularized and coordinated force than anything that had gone before. The praetorians swore an oath of allegiance to Augustus, something that would be transferred to other emperors. In time this would present some praetorians, ordered to carry out killings on behalf of the emperor or confronted by an emperor such as Nero whose behaviour became intolerable, with serious challenges to their loyalty.
Octavian had reorganized the army after the war was over in 31 BC. At that stage he controlled around sixty legions, comprising his own army and what was left of the forces controlled by Antony’s faction, as well as praetorians. Later, as Augustus, Octavian would claim to have demobilized approximately 60 per cent (300,000) of the half million troops he controlled after Actium, dispersing these veterans throughout new colonies or back to their own home towns. Dio states that by AD 5 Augustus had twenty-three or twenty-five legions, as well as ten thousand praetorians in ten cohorts. Writing about the situation in 23, Tacitus gives twenty-five legions, though in the interim three legions had been lost in Germany in AD 9, and nine praetorian cohorts (size unspecified). These praetorian cohorts seem to have been part-mounted with the equites praetoriani, perhaps in a proportion of four to one in favour of the infantry component.
Unfortunately we possess no piece of evidence in any form that tells us the configuration of the Guard when Augustus founded it, in terms either of total size or the organization of cohorts. All we have to go on is Dio’s description of the armed forces in AD 5, and Tacitus’ summary of the disposition of the same in AD 23, both of which were written long after the event, and sporadic pieces of evidence that crop up in various contexts across the first century AD. Dio states that the Guard in AD 5 was made up of ten thousand men in ten cohorts, Tacitus that in AD 23 it consisted of nine cohorts, but without specifying either the total size of the Guard or the individual cohorts. Obviously, these figures are incompatible, which means either that one is wrong or that the arrangements had changed since the date of foundation and probably continued to change. The men recruited to the Guard were predominantly aged around eighteen to twenty years at the start of their service, a similar age to that of around two-thirds of legionaries, and drawn from Italian Roman citizens, by far and away the most convenient source. The nature of their individual origins is less certain, for example whether they came from poorer families or were drawn from the better-off. The distinction is unlikely to have made much practical difference once they had been absorbed into the Guard.
A number of modern authorities have made different assumptions about the size of the cohorts, based on the available evidence, and come to a variety of conclusions. One of these is that Dio was actually referring to the configuration of the Guard in his own time in the early 200s, and was therefore mistaken in suggesting that this was how the Guard was organized in Augustus’ time. This position is largely responsible for inferring that the nine cohorts referred to by Tacitus was the correct number for the Augustan era. It has also contributed to the assumption by some that these cohorts were quingenary in size, with a nominal 480 men each. This would have made the Augustan Guard similar in size to, but smaller than, a legion. Together with the equestrian tribunes who commanded each cohort, the centurions, optiones and standard-bearers, the total in each cohort would have been around five hundred. It is, however, important to stress that nine cohorts is a retrospective estimate based on the figures provided by Tacitus for the army in the year 23 under Tiberius, some fifty years or more after the foundation of the Guard by Augustus. We do not know for certain how many cohorts were involved when the Guard was founded. We also do not know if the numbers changed or, crucially, whether those cohorts were quingenary or milliary (one thousand strong) at this date. Quingenary cohorts are never specifically attested for the Praetorian Guard, now or at any other date, though quingenary cohorts were normal for the legions at the time. The praetorians, however, were not legionaries so there is no particular reason, in the absence of any evidence for praetorian quingenary cohorts, to insist that this was how the Guard was organized.
In fact, there is good reason to support the idea that milliary cohorts are not only possible but even likely for this early date, though there can be no certainty about this. Dio’s reference to there being ten thousand praetorians in ten cohorts must imply milliary cohorts, making the Guard close to the size of two legions. The possibilities are that this was the size of the Guard when it was founded, or the size to which it had grown by AD 5, or was a retrospective reference by Dio. If both Dio and Tacitus were correct this would have to mean that the Guard had been reduced by one cohort between 5 and 23. The number of cohorts, however, may also have varied in ways that have gone unrecorded, rendering futile efforts to reconcile the disparate evidence. There were at least nine cohorts on the evidence of tombstones at Aquileia, which specify cohorts up to and including the VIIII (IX) cohort under Augustus, while other evidence suggests an expansion to twelve cohorts by the 30s or 40s. In 76, one of the earliest praetorian discharge diplomas known explicitly refers to nine praetorian cohorts, though this surely reflects the reorganization of the Praetorian Guard after the civil war, during which Vitellius had increased its numbers to sixteen cohorts. Whatever reason there might have been for using nine cohorts under Augustus or at any other time, as opposed to another number, is unknown to us, especially as a graffito from Pompeii, which must predate August 79, refers to a tenth cohort. It is quite clear that arrangements could and did change.
The idea that Dio was mistaken is a convenient, but unsatisfactory, way of refuting one piece of evidence to resolve the discrepancy between Dio and Tacitus. When discussing the division of Rome into fourteen districts for the purposes of firefighting in 7 BC, Dio makes it clear with his phrase ‘this is also the present arrangement’ that he was fully aware that circumstances might have differed between his time and what had prevailed two centuries earlier. There is no reason therefore to assume as a matter of course that he was wrong about the size of the Guard in AD 5. Since the Guard at that date was not accommodated in a single location, and was widely dispersed, none of the normal methods (such as the size of a fort) by which a theoretical number could be estimated applies.
The figure of ten thousand given by Dio is also the same number given by Appian for Octavian’s praetorians in 43 BC. Given the lack of precision in our sources, and the difficulty of reconciling all the disparate pieces of information available to us, it is possible that Octavian did indeed create a Praetorian Guard on the scale described by Dio, based on ten milliary cohorts. After all, it would have been useful to reward especially loyal soldiers towards the end of their careers in the aftermath of the war. Moreover, it would be some time before Octavian could feel completely secure, so a large Praetorian Guard might have seemed a good idea to begin with. The conspiracy led by Fannius Caepio in 22 BC showed that almost a decade after Actium, Augustus could still face danger at home just as Caesar had. The only other early reference to milliary cohorts in the Guard is by Philo for the year 40, who clearly mentions a thousand-strong cohort. In other words, the only specific evidence we have for the size of early imperial praetorian cohorts comes from Dio and Philo; both suggest the cohorts were milliary, at least for the times they were writing about.
The layout of the Castra Praetoria, built in Tiberius’ reign by 23 when the Guard was all based in Rome, allows for a very substantial garrison, perhaps in excess of fifteen thousand. This certainly does not conflict with the idea that Dio was indeed right about the size of Augustus’ Guard in AD 5, and that the cohorts were milliary from the outset. Even if the Guard was smaller under Augustus, the availability of the Castra Praetoria would have transformed the extent to which praetorians could be accommodated in Rome, making an enlargement of the Guard logistically feasible from at least thereon.
All these speculative possibilities really serve to show is that we simply do not have enough precise information to know how the size of the Guard changed over time from its inception, if indeed it did change, and to resolve the contradictions in the evidence. We do not have to assume that Dio was wrong about the size of the Guard under Augustus, or that his information conflicts with that of Tacitus. It is possible that the cohorts were military sized from the beginning, and that the numbers of cohorts fluctuated in ways that we are now unable to pin down with certainty, as indeed has been acknowledged by some authorities. The latter seems certain, the former no better than the balance of probability. As an aside, it is worth noting that the individual cohorts of the Guard did not possess at this date any form of honorific or loyalist name commemorating particular acts or campaigns as the legions did. There was thus no overtly prestigious or privileged praetorian cohort.
There is also the question of the cohors speculatorum, ‘the cohort of scouts’, the subunit of the Guard associated with intelligence and spying, first attested in Antony’s forces at Actium. Augustus is even recorded as having remained on social terms with a former speculator on his staff. They were still in existence in 68, when they were described by Tacitus as forming a special bodyguard chosen by height, and presumably remained a permanent feature. They did not necessarily serve in separate cohorts, even if they had done so under Antony.
The Guard was, in any relative sense, a tiny part of the Roman world’s armed forces but a very definitely privileged one, regardless of how it was organized. In 27 BC Augustus saw to it that the senate passed a decree authorizing the Guard to receive ‘double’ the pay of the ordinary legionaries. Predictably enough this does not tie up with other evidence for later dates, which suggests that the ratio was modified to 3.33:1. In AD 14 a praetorian was paid two denarii per day, according to Tacitus, whereas a legionary received 10 asses (0.625 of a denarius) per day. The latter figure was a slight approximation, equating to 228 denarii per year. The actual annual rate for a legionary was 225 denarii per year, rising to 300 under Domitian (81–96), so the day rate given by Tacitus is obviously a rounded figure for the sake of simplicity. Soldiers in the Roman army were in theory paid three times a year so the figure needed to be divisible into thirds, which of course 225 and 300 are. Based on this, the differential recorded in AD 14 would probably have to mean that by then a praetorian was perhaps paid 720 denarii per year (that is, 3 × 240), equal to 1.97 notional denarii per day, in other words the ‘two denarii’ cited by Tacitus, who had no means even to express the term ‘1.97’. Another, more likely, possibility is that the praetorians were paid 750 denarii per year (3 × 250 denarii), equivalent to 2.05 denarii per day. Either approximation would fit with Tacitus and fulfil the need to be divisible by three. Moreover, 250 denarii is equal to 1,000 sestertii, the amount awarded praetorians in Augustus’ will, just as legionaries were awarded 75 denarii (300 sestertii). It is of course entirely possible that the praetorians’ ‘double’ pay in 27 BC had been raised to a higher rate by AD 14, though we have no evidence to prove this, other than the discrepancy discussed here.
In 13 BC Augustus fixed the length of service for a praetorian at twelve years and a legionary’s at sixteen, modifying this in AD 5 to sixteen and twenty years respectively. These should be seen as minimum figures rather than fixed points at which discharge automatically took place. The gratuity received at discharge for praetorians and legionaries seems to have been set at a related ratio to that for pay. Praetorians received 20,000 sestertii but a legionary received 60 per cent of that amount – 12,000 sestertii. This is equivalent to 1.67:1 or approximately half of the 3.33:1 ratio for pay, which is unlikely to be a coincidence. The amounts left to praetorians and legionaries in Augustus’ will were also based on the ratio of 3.33:1. The slight differences are easily explained by the need to have rounded figures divisible by three for the purposes of annual pay, and rounded figures for the single distribution of retirement gratuities and imperial bequests. The idea that the ratios are inconsistent therefore simply does not hold up, since it is quite clear they were very closely related and were rounded for convenience.