The list shows that this corps d’élite was not representative of all the cavalry of the Empire; the proportion of orientals is far too low. It was still upon the Western provinces that the emperors of the second century relied, and from these, therefore, that the guard was recruited. As regards these provinces, the composition of the Equites Singulares reflects fairly accurately the relative importance of each as a recruiting-ground for the auxiliary cavalry of this period, and the change which has taken place in the military situation since the days of Augustus is at once apparent. The Galli243 and Hispani, who were then the flower of the imperial cavalry, and continued to give their names to nearly half the cavalry regiments in the service, have entirely vanished. Speaking generally, in fact, the inland provinces no longer contribute, and the recruiting-areas have contracted to the purely frontier districts. The relative importance of these, too, has altered since the beginning of the first century. The tribes on the Lower Rhine are still well represented, but the contingent of the German provinces is entirely surpassed by that of Pannonia. If we assume that the honour of serving in the Guards was bestowed upon the natives of each province in proportion to the size of the contingent which they supplied to the cavalry of the line, this increased importance of the Pannonians follows naturally upon the universal adoption of local recruiting for the frontier armies; since not only had the balance of military power now definitely shifted from the Rhine to the Danube,244 but local conditions required an exceptionally high proportion of mounted men.245

The preceding survey of the evidence has purposely omitted that dealing with the oriental regiments, which seemed, on account of its exceptional character, to merit a special discussion. In Pannonia and Dacia we find three such regiments, the Ala I Augusta Ituraeorum and the Cohors I Hemesenorum in Pannonia Inferior, and the Cohors I Augusta Ituraeorum, stationed in Pannonia during the first century, and transferred to Dacia at the time of the creation of the province.246 Thanks to the recent work of Hungarian archaeologists, the second of these corps, the Cohors I Hemesenorum Miliaria Equitata Civium Romanorum Sagittariorum, to give it its full title, is perhaps better known to us than any other auxiliary regiment.247 Probably enrolled in the Roman army at the beginning of the second century, at the time of the annexation of the small client kingdom from which its name was derived, this regiment had been transferred to Pannonia by the beginning of the reign of Antoninus Pius, and certainly remained in the province until 240.248 Throughout this period it seems to have been stationed at Intercisa, where upwards of fifty inscriptions, chiefly sepulchral, have now been discovered, the majority of the latter belonging, as the frequent use of the name Aurelius shows, to the end of the second and the beginning of the third century. Of the five soldiers whose birthplace is mentioned on their tombstones, three came from Hemesa itself, one from Samosata, and one from Arethusa,249 and the owner of a diploma which dates from between 138 and 146 came from Syria.250 It is clear, then, that during its whole stay in Pannonia this regiment was not recruited, like the majority of the auxilia, from the neighbouring district, but was constantly in receipt of fresh drafts from the province in which it was originally raised. As further proof of the tenacity with which this connexion was maintained, we find that at the end of the reign of Severus the soldiers dedicated a temple to their deus patrius Sol Aelagabalus.251 An examination of the evidence dealing with the other oriental regiments leads to the same result. Soldiers discharged in 98 and 110 from the Cohors I Augusta Ituraeorum and the Ala of the same name, give Cyrrhus in Syria and Ituraea as their places of origin, and another Ituraean is mentioned on a sepulchral inscription of the latter regiment in Pannonia.252 Oriental regiments, and in particular oriental archers, appear also on other provinces253, and although there is a lack of dated evidence we can hardly doubt that a rule which was maintained along the whole Danube frontier also held good elsewhere.254

The reason for the adoption of these exceptional methods in the recruiting of the oriental auxilia was probably the purely military one that good archers were born in Syria, and could not be made elsewhere, but the consequent presence in every frontier army of an oriental element, holding firmly to its own customs and religious beliefs, was a fact of more than military significance. In particular it must be remembered that the enfranchised children of these oriental auxiliaries were qualified and readily accepted for service in the legions. In the inscriptions of the Cohors I Hemesenorum we have abundant evidence of this process, which gave oriental ideas an opportunity for wider penetration.255

The British regiments, too, show a tendency to keep up national recruiting, although the evidence on this point is still scanty.256 In this case the explanation is probably to be found in the intractable nature of the tribes of North Britain, which made it appear undesirable to use the contingents which they supplied near at home. Certainly all the British regiments seem to have been sent abroad, and only one soldier of British origin is found in the province.257 It would appear, in fact, that the army of Britain was maintained largely by drafts from the Rhine area, but the evidence at present is an insufficient basis for any general conclusions. In Dacia, too, as was natural, the auxilia raised immediately after the conquest were transferred elsewhere. Here, however, there does not seem to have been the same objection to local recruiting for the troops stationed in the province, and the practice was certainly, as has been shown, in force during the second century.

The Numeri.

It is clear, then, that in the second century the cohorts and alae of the Augustan system, with certain definite and limited exceptions, were recruited locally from the provinces in which they were stationed, without any general attempt to justify the ethnical titles which they still bore. At the same moment, however, as this principle seems definitely established, there begin to appear on inscriptions certain regiments of a new type which stand outside the prevailing system. These regiments are given the name numerus, which does not seem to have had any very precise meaning, and for which it is difficult to find an English equivalent. They also bore tribal titles, a list of which is given below, together with the names of the provinces in which numeri of each tribe occur.258

Brittones. At least ten numeri in Germania Superior. The earliest inscriptions date from the reign of Pius, and the theory is generally accepted that these Brittones were newly conquered tribesmen from the district lying between the two frontier walls, deported after the campaigns of Lollius Urbicus. These numeri also bear secondary titles derived from the names of the districts in which they were stationed, e.g. Murrenses, clearly connected with the river Murr.259

Germani. Dacia.260

Palmyreni. Africa,261 Dacia,262 and probably Britain.263

Mauri. Dacia (as a vexillatio from Mauretania Caesariensis).264 Frequently also in the two Mauretanias.

Raeti Gaesati (i.e. Raetians armed with the gaesum, a kind of heavy spear). Britain.265

Syri. Dacia,266 Mauretania,267 Moesia Inferior,268 and possibly Britain.269

The titles show that these numeri are closely connected with the five nationes, Cantabri, Gaesati,270 Palmyreni, Daci, and Brittones, who form part of the army described by Hyginus, and are distinguished from the cohorts and alae of the regular auxilia. In fact the term numerus is not invariably found on inscriptions, and for a corps which simply described itself as ‘Syri sagittarii,’271 natio may easily have seemed a convenient term. It is also clear from Hyginus that what distinguished the numeri from the older formations was their looser organization and more barbarous character, and their titles show that they were drawn from the outermost borders of the Empire, or the most uncivilized districts within it.

Evidence about the character of these troops is scanty; being the least civilized part of the army they were not very prone to indulge in inscriptions. We do not even know the size of a numerus, or, indeed, if these regiments had any definite size at all. The nationes of Hyginus range from 500 to 900,272 but the low rank of the praepositus numeri below the praefectus cohortis,273 and the smallness of the accommodation arranged for these regiments in German forts, suggest 200 to 300 as a more probable figure. Smaller they can hardly have been, since they are divided into centuries and turmae.274

As regards recruiting, the Palmyrenes, at any rate, obtained fresh drafts, not from the province in which they were stationed, but from their native home. The numerus Palmyrenorum, which was stationed at El-Kantara in Africa, has left inscriptions covering the whole period from the middle of the second to the middle of the third century, which show clearly what pains were taken to preserve the original character of the regiment.275 Similar inscriptions of a numerus Palmyrenorum in Dacia give evidence of the same principle.276 Unfortunately, we cannot tell whether this was the case with all the numeri, or whether the Palmyreni occupied in this respect the same peculiar position as the oriental regiments among the regular auxilia.277

But whatever the later practice, the original intention of the imperial government in raising this new class of troops seems to be clear. The local recruiting of the regular auxilia presupposes a rapid progress of ‘Romanization’ among the provincials and the disappearance of all such national feeling as had caused the mutiny of the German and Pannonian auxilia in the first century. From the military point of view, however, this advance in culture, although it facilitated the raising of recruits, was by no means an unmixed blessing. The old levies of wild tribesmen, schooled by centuries of local warfare, who strove to preserve in the Roman service their local reputation, had qualities which were lacking to the regiments of civilized Latin-speaking provincials, in which national methods of fighting278 had been replaced by a uniform training and clan feeling by esprit de corps. It was to provide a leaven of the old spirit that the numeri were raised from the wildest of the border tribes, and not only encouraged to fight after the manner of their fathers, but even permitted to continue the use of their native tongues.279

The first experiment of this kind was made by Trajan, when he brought over Lusius Quietus and his Mauri gentiles for the Dacian war,280 but it was probably Hadrian who made the numeri a regular part of the military system. It is in his reign that they first appear on inscriptions, and it is to the numeri that we must refer the passage in Arrian’s Tactica, in which the emperor is praised for encouraging his troops to keep up their national methods of fighting, and even their national war-cries.281 The tribes to whom he refers are Κελτοί (by which Germans are probably meant),282 Dacians, and Raetians. For the first and third there is epigraphical evidence, and the last two appear also among the nationes of Hyginus. It appears, then, that Hadrian was not, as is sometimes stated, the originator of the system of local recruiting; rather he found it already in existence, and sought to correct its defects by utilizing again in the service of the Empire the clan spirit of uncivilized tribes, which had often proved so useful in the past.

The Praefecti.

The previous discussion of the methods by which the auxilia were recruited has dealt only with the private soldiers, and, as a natural corollary, with such officers as were promoted from the ranks. To the position of praefectus, however, the private soldier could not normally aspire, and he attained it, if at all, only under exceptionally favourable circumstances.283 Normally, the commanding officers of the auxiliary regiments were drawn from an entirely different social stratum to the men, and although the method of their appointment varied and the area from which they were drawn shifted its boundaries at different periods, these changes did not follow the same lines as those which we have been tracing in connexion with the recruiting of the rank and file.

The auxiliary commands are familiar to all students of the Roman Empire from inscriptions of men who went through the equestrian career, the first stage of which was formed by the posts of praefectus cohortis, tribunus legionis, and praefectus alae. It has, however, been pointed out by von Domaszewski284 that this system was not established until the middle of the first century. Under Augustus and Tiberius, not only was the relative rank of these posts still undetermined, but they were filled in many cases not by young men beginning the equestrian cursus, but by veteran centurions from the legions, especially the primipili. We have noticed this system in the army of Caesar, so that here, as elsewhere, Augustus was continuing a republican practice. The following inscriptions,285 which are both of early date, give typical careers of this character.

This system is heartily commended by von Domaszewski, on the ground that the auxilia were thus commanded by more skilful officers, and were more under Roman (i.e. Italian) control than was the case in the second century.286 The assertion, however, seems far too sweeping, since by no means all the auxiliary regiments were commanded by men of this class; there were, on the contrary, many praefecti at this period who came neither from Italy nor even the more Romanized provinces. The Histories of Tacitus show clearly that at the end of the pre-Flavian period a number of auxiliary regiments, particularly those drawn from the more independent border tribes, were commanded by their own chiefs. This practice had not sprung up during the reign of Nero, but was a natural consequence of the development of these corps from contingents supplied by states nominally ‘in alliance’ with Rome.287 The eight Batavian cohorts who play such an important rôle in the rebellion of 69 were so commanded,288 as well as an ala of the same tribe.289 Iulius Civilis himself was a praefectus cohortis,290 and two Treveri, Alpinius Montanus and Iulius Classicus, commanded a cohort and ala respectively.291 All these officers, as their names show, had doubtless received the franchise, but they were employed in their capacity as tribal chiefs, not as Roman citizens, and are to be distinguished from the praefecti, who were drawn at this period from the Romanized districts of Spain and from Gallia Narbonensis. It is chiefly as commanders of cohorts that officers of this type appear, since many of the alae dated back, as we have seen, to the period of the civil wars, and had long lost their original character as tribal regiments. This explains the fact that among Italian officers of this period, the title praefectus alae or praefectus equitum appears far more frequently than praefectus cohortis, although the cohorts were of course more numerous than the alae.

During the first half of the first century, therefore, we have a system which differs widely from that revealed by the equestrian cursus honorum. The establishment of the equestrian monopoly of the auxiliary commands was, in fact, only completed by a series of reforms carried out during the period which began with the administrative activity of Claudius, and ended with the reorganization of the army by Vespasian after the disastrous Civil War of 69.

The first of these changes was that the praefecti ceased to be drawn as before from among the veteran centurions of the legions. Early in the reign of Claudius the post of praefectus alae disappeared from the career of the primipili, who were promoted henceforward to the tribunate of one of the cohorts of Household Troops at Rome.292 Centurions of lower rank were still advanced to the command of cohorts both in this and the succeeding periods, but such cases are very rare.293 A trace of the old connexion between the legionary officers and the militia equestris still survives, however, in the regular use of a centurion as praepositus cohortis—that is to say, as temporary commander in case of the death or absence of a praefectus.294 The numeri, too, were often placed in charge of an ex-centurion bearing this title, an arrangement which was probably called for by the intractable character of these barbarian irregulars.295

The employment of tribal chiefs as praefecti also became less frequent, as the auxiliary regiments, transferred from one province to another, and recruited from different nationalities, gradually lost their original character. The mutinous officers of the Rhine army, who were doubtless cashiered by Vespasian, were probably the last examples of praefecti drawn from this class.

Lastly, the respective rank of the different posts in the militia equestris was finally determined; and the order praefectus cohortis—tribunus legionis—praefectus alae is hardly ever varied after 70, except that the tribunate of a cohors miliaria sometimes appears in the second place.296

The result of these changes was that henceforward the auxiliary officers were practically all of one type, men of equestrian rank entering upon what was now the accustomed cursus honorum of their class. That this system was not universally adopted at an earlier date is not surprising. The equestrian praefecti were young men directly appointed by the emperor, without any previous military training; before the auxilia could be entrusted to their charge a certain advance in civilization and tractability had to be made by the provincials, and the veteran centurions and tribal chiefs of the Augustan system were more fitted to deal with the men who composed the auxiliary levies during the first hundred years of the Empire. As it was, these regiments contained in the second century far fewer representatives of the governing class than the native corps in our own Indian army. With the exception of the praefectus, who himself was not necessarily an Italian, the officers—that is, the centurions and decurions—were practically all, as we have seen, promoted from the ranks. But to the Roman Empire, in which rulers and ruled, never separated by any deep racial or religious gulf, were gradually made closer akin by the bond of a common civilization, our rule in India affords in this respect no real parallel.

The majority of these praefecti were, at the beginning of this period, of Italian origin, taken from the leading families of the country towns, the class which formed, under the rule of the Flavian emperors who were themselves sprung from it, the backbone of the Roman bureaucracy. The Romanization of the Western provinces led to an increasing proportion of men from the provincial municipia being admitted into the imperial service, but until the reign of Marcus the Italian element still predominated. The praefecti mentioned on five diplomata from Pannonia Superior and two from Dacia dated 133, 138, 136-46, 148, 149, 157, and 145-61 were natives of Sassina, Bovianum, Faventia, Suessa, Rome, Hispellum, and Picenum.297

The accession of Septimius Severus possibly accelerated the speed at which the provincial element was increasing, but there is not, as has been suggested by von Domaszewski, any sign of a violent and wholesale exclusion of Italians from this branch of the service. This point may be illustrated by the following inscriptions of Italian praefecti, which can be dated after 193:

The earliest provincial praefecti came from the thoroughly Romanized districts of Spain and Gallia Narbonensis, natives of which appear even in the pre-Flavian period. These were followed in the course of the second century by representatives of nearly all the Western provinces; Africa in particular sent praefecti from its many flourishing towns to almost every frontier during the latter half of the second century, and the accession of the African Septimius Severus at its close possibly gave his fellow countrymen a specially favoured position in the succeeding period.

Only in Britain and Gallia Lugdunensis do the Celtic chiefs seem to have made no attempt to maintain in the second century the military position held by their fathers in the first. It is hardly likely that their absence from the lists of praefecti298 is due to deliberate exclusion on the part of the imperial government. It was more probably a voluntary abstinence, due largely to the fact that these military commands were now regarded merely as an introduction to the civil service, not as a career in themselves. The Celtic nobles were not uninfluenced by the tendencies of the age. But although they might speak and read Latin with ease, decorate their homes with the material products of Roman civilization, and employ Greek rhetoricians to tutor their children, these country gentlemen living in the midst of their estates preserved a very different outlook to that of the leading townsmen of the municipalities of Africa or even Narbonensis. The Celt retained his martial qualities down to the last days of the Empire in the West, but seems to have found little that was congenial to him in the prospect of forming part of that great administrative machine, in the perfection of which almost every other province in the Empire took its share.299

The Eastern provinces of the Empire occupy, as usual, a somewhat exceptional position. As in the West, these provinces began to contribute praefecti in some numbers towards the end of the first century. A certain C. Julius Demosthenes of Oinoanda went through the ‘militia equestris’ in the reign of Trajan, and his son, Julius Antonius, followed in his footsteps in the succeeding generation.300 To a citizen of Caria, L. Aburnius of Alabanda, probably, as his name shows, the descendant of one of the families of veterans settled by Augustus in the south of Asia Minor,301 the wars of Trajan presented opportunity for a military career of considerable variety and distinction. This officer was successively praefectus fabrum, tribunus legionis III Augustae, praefectus cohortis III Augustae Thracum equitatae, praefectus cohortis III Thracum Syriacae equitatae, praepositus302 cohortis I Ulpiae Petraeorum, praepositus annonae302 on the Euphrates during the Parthian War, tribunus legionis VI Ferratae, during his tenure of which post he was decorated by Trajan, and praefectus alae I Ulpiae singularium.303 These cases are not isolated, and it is clear that a military career was open to the Greeks and the more or less Hellenized orientals who constituted the equestrian class in the Eastern provinces. But while the praefecti from the Western provinces were sent indiscriminately to every frontier, the majority of those drawn from the East seem, during the first two centuries, to have been confined to service with the Eastern commands. Aburnius, for example, only left the East once for service with the Legio III Augusta in Africa, and his son’s career seems to have been similarly localized.304 We should perhaps add Moesia Inferior to the list of provinces in which Eastern praefecti appear frequently during the second century, since those mentioned on the diplomata for 134 and 138 came from Palmyra and Side respectively.305 But Moesia Inferior was reckoned in other respects as coming within the Hellenic sphere of influence. These restrictions are probably due to the low estimate which was placed throughout the first two centuries on the military qualities of Greeks and orientals, in spite of the value of the latter as archers.306 But we may also see evidence of the unbridgeable gulf which still existed between the two halves of the Empire, and of the reluctance of the Hellene to embark upon a career in what he considered to be the barbarous provinces of the West. It is only with the advent of the semi-orientalized dynasty of the Severi that praefecti drawn from the Eastern provinces appear in any numbers on the Western frontiers.

In all this the course of events is what one would naturally expect. The spread of uniform culture throughout the Western provinces of the Empire, the prosperity of the ubiquitous municipalities which were its material expression, and the general extension of the franchise which accompanied this development, involved a steady increase in the class qualified and eager for the equestrian career. The admission of these provincial equites to the posts for which they were qualified followed automatically without special encouragement from any particular emperor,307 and the diverse origins of the praefecti at the beginning of the third century are one of the best proofs that can be adduced of the prosperity and civilization of the provinces at this period. It is impossible to follow von Domaszewski in concluding from the evidence that at this date the inhabitants of Italy and the more civilized areas in the provinces were deliberately excluded from the militia equestris, and that the auxiliary regiments were given into the hands of barbarians.308 The army was indeed beginning to suffer from the introduction of a barbarian element, but it is not among the officers of the auxilia that this element is most noticeable. The following list of praefecti, who can be dated to the first half of the third century, does not bear out the accusation:

vii. 344. Britain.309 Aemilius Crispinus natus in provincia Africa de Tusdro (dated 242).

viii. 2766. Britain. P. Furius Rusticus. Lambaesis. Severus or later. Britannia Inferior is mentioned.

xiii. 6658. Germania Superior. Sentius Gemellus. Berytus. Date probably 249.

xiii. 7441. Germania Superior. Flavius Antiochanus. Caesarea. Date 191 or 211.

I. G. R. R. i. 10. Raetia. T. Porcius Porcianus. Massilia. 3rd century.

iii. 1193. Dacia. C. Julius Corinthianus. Theveste. circa 200.

C. I. Gr. 3497. Dacia. T. Claudius Alfenus. Asia. circa 200-210.

These men cannot fairly be called barbarians. Massilia of course speaks for itself, but in Theveste, Thysdrus, and Lambaesis Roman culture was no new thing at the beginning of the third century. The same may be said of Caesarea, if the capital of Mauretania be meant. Berytus, too, was a colony famous for its Roman character, and Asia was not a province notorious for its barbarism. The increased oriental element, which is certainly noticeable among the auxilia of this period, although not to the same extent as in other branches of the service, is a more significant fact. But however undesirable one may consider the influence of oriental religions and ideals to have been, the conflict cannot be called one between civilization and barbarism. The real matter at issue is the wisdom of the imperial government in utilizing the material which the spread of culture and prosperity provided, and substituting for the old hegemony of Italy a governing class drawn from all parts of the Empire. It is true that this policy was a failure, and that the Empire organized on this basis did not succeed in erecting defences strong enough to resist the external pressure brought to bear upon them in the third and fourth centuries. But if it was a failure it was not necessarily a mistake.310 It is more than doubtful whether a narrower policy which rigidly maintained the supremacy of the Italians and denied to the majority of the provincials all share in the administration would have been more successful: it is certain that, had the progress of civilization lacked the stimulus which the hope of political power supplied, the after-effects of the Roman Empire in Europe would have been less.