It is notoriously difficult, in all or nearly all military history, to disentangle the conscious purposes and previous plans of the commanders from the results of pure chance or the improvisation of the moment; probably few wars have ever worked out entirely according to plan, and even an accurate analysis of the chief factors in the success of any particular campaign would, to those who really knew the facts, appear as an over-simplified diagram. Yet there is a correct use of diagrams, and we need not be ashamed of trying to pick out the true strategical significance from even the most complex military events, provided we are honest with ourselves. In attempting such a task with regard to the Jugurthine War, we must take very special care to notice the limitations that surround us : our results can be, at best, hypothetically sound, since our authorities are neither complete nor notably trustworthy in such important matters as personal impartiality and chronological detail. We have no contemporary source. Sallust is notoriously partisan, on the political if not also on the military side; he is, moreover, open to the suspicion, to put it mildly, of choosing for detailed description the more picturesque rather than the more vitally important episodes; and comparison with Orosius (who almost certainly derives from Livy) suggests that one battle, if not more, may have been altogether misrepresented.

The following stones, like no. 230 published in Part IV of this article (J.R.S. XVII, 1927, p. 49 ff.), were seen by W. M. Calder and C. W. M. Cox in the course of a hurried visit to the Upper Tembris Valley in July, 1924. The MS. of Part V was completed in March, 1926; in the revision of the proof-sheets we have been privileged to use the notes, impressions and photographs made by C. W. M. Cox and A. Cameron in the course of a systematic survey of the Tembris Valley monuments in May, 1926. For the notes on nos. 253 and 254 Calder and Cox are jointly responsible, for those on nos. 232 and 233 Cox, and for the remainder Calder. The letters C. Co. indicate the copyist.

When literary records are few and faulty, the movements of legions win a certain importance as historical evidence, not least for the wars of Domitian. Whole legions were not displaced for purposes trivial or temporary; it might almost be said that in Flavian times, when a legion leaves a province, it leaves it for good. The aim of the present enquiry is, therefore, to narrow as far as may be the margin of dangerous error that fringes the subject and to propose a theory of the number and position of the permanent Danube garrisons in the interval of peace and stability that followed the wars of Domitian and preceded those of Trajan. The choice of the latter topic may be excused by the variety of the views expressed by historians, and by the fact that both the monumental work of Ritterling and a recent English book seem rather to have avoided the problem.

Amid all the disputes and controversies that have centred on the Roman army of the Empire since Mommsen's day, one important chapter in that army's history has remained curiously free from question; however many legions Augustus retained, whenever this legion was raised or that destroyed, all writers are in agreement that Vespasian, early in his principate, cashiered some of the mutinous legions of the Rhine armies.

The great majority of writers are agreed that the legions so dismissed were I, IIII Macedonica, XV Primigenia, and XVI Gallica; though Hardy, following Grotefend and Schilling, suggested that XV might have survived even into Trajan's principate: to-day, however, no one maintains this view, which was the outcome of a failure to account for Vespasian's apparent reduction of the number of legions.

In the account which I gave in this Journal of the excavations carried out in 1925 at and near the east corner of the legionary fortress, I explained that the earliest defence there had consisted of a double ditch and a clay rampart, which appeared to have been laid out in the time of Petillius Cerialis (c. A.D. 71–74); that a stone wall subsequently built along the outer margin of this clay rampart probably dated from the reign of Trajan (about 108–9); that at some time this wall was destroyed, and that a second wall, built immediately inside the line of the wall which it replaced, belonged to the period of Commodus or Septimius Severus. Those conclusions were drawn from the evidence of trenches cut on either side of the corner. Of the corner itself enough was uncovered to prove that the tower here, as was to be expected from the date of the adjacent work, was of the internal type usual in the first three centuries. The complete excavation of the corner was the main task of 1926.