honoured for his loyal co-operation against the African rebel Tacfarinas.[15] Along the eastern frontier, kingdoms of this type, the wreckage of the old Hellenistic system, were more numerous and played a more important part. Thus Antiochus III of Commagene, Agrippa II, Sohaemus of Emesa, and Malchus of Damascus contributed 15,000 men to the army which Vespasian led into Palestine in the spring of 67.[16] Even the more autonomous city states seem to have retained a militia which was occasionally made use of. So late as the reign of Hadrian, in the army which Arrian led against the Alani, we find a contingent from the ‘free’ city of Trapezus, which is reckoned among the σύμμαχοι as opposed to the regular imperial troops.[17] A similar freedom from the direct control of Roman officers was permitted to the chiefs of some of the border tribes, who were allowed to lead their own clansmen to battle. To this type of militia belong the _tumultuariae catervae Germanorum cis Rhenum colentium_, including the Batavians under their _dux_ Chariovalda, who serve in the campaigns of Germanicus,[18] and the levies of the Dalmatian clans who started the rebellion of 6.[19] Last come the permanently embodied regiments raised from the subject communities, the auxilia properly so called, who form the subject of this treatise. Here, probably more than in any other department of the military system of the Empire, we can trace the results of Augustus’s own activity. Regiments of this kind had, as we have seen, existed under the Republic, but they had probably been few in number and the incidence of the levy had been uneven and capricious. Under Augustus not only was the number of regiments largely increased—we hear of no less than fourteen alae and seventy cohorts taking part in the Pannonian War of 6-9[20]—but the inscriptions show us that, with the exception of Greece, always the spoiled child of Roman sentiment, every quarter of the Empire contributed its quota. Details respecting the incidence of this levy on different provinces, and the methods of organization and recruiting, will be found in later sections. It will be sufficient to say here that while the subject communities had probably more reason than any other class to complain of the military demands of the state, the burden was at least more equitably distributed than under the Republic and the total contribution required, in most cases at any rate, not excessive. It is natural to suppose that the fixing of the quota supplied by each community was connected with the drawing up of the census, which placed the taxation of the Empire for the first time on an organized basis, and it seems probable that more evidence might show a reciprocal variation between the two forms of contribution required. We know, for instance, that the Batavians were altogether excused from the payment of tribute on account of the size and value of their contingent, and this case was probably not exceptional.[21]