The circumstances surrounding one of the most grievous disgraces which ever befel a Roman army are familiar. In 321 B.C. the so called Second Samnite War was at its height. The Roman consuls, while at Calatia with their armies, received news that the enemy forces were in Apulia under the new Samnite general, Gavius Pontius. Pushing forward, either in the hope of winning a decisive victory on the Apulian plains, or in order to relieve the pressure on the allied city of Luceria, they advanced into a valley which had but two exits: on trying to get out of this defile, they found both entrances blocked. Doubtless the army made attempts to fight its way out of the encircling Samnite ring, but they proved unavailing. To avert death by starvation, the consuls surrendered. Pontius, being in a position to dictate his own terms, demanded that Roman garrisons should be withdrawn from territory which the Samnites regarded as theirs, while the Romans were not to re-open the war: in addition he took six hundred Roman knights as hostages, disarmed the Roman army and sent it ‘under the yoke.’

The dating of coarse pottery on Roman sites in the first and second centuries has been fairly well established, but there is still much uncertainty in the chronology of coarse ware during the third and fourth centuries. Hence it may be helpful to describe the pottery found in a Roman well which would appear, from the evidence to be adduced, to have remained open only during the third century. In my plan of Margidunum published in the Journal of Roman Studies, xiii, Pl. VIII, the position of this well, marked R, is shown to be situated halfway between the Via Quintana and the southern rampart. It lies just outside the eastern wall of a small building which may have been used as a stable, at any rate in the earlier occupation of the camp, for in the interior I found a pit (J) which had been filled up in the early part of the second century. The lower two feet contained brushwood, straw and horse-dung quite unaltered, owing to the pit having been hermetically sealed by clay. Doubtless this pit was closed when the camp was dismantled and the Fosse Way driven arbitrarily across it, over the old camp-roads and the foundations of the razed buildings.

Romano-British archaeologists are now generally agreed in supposing that the final abandonment of Hadrian's Wall occurred about the time of the usurpation of Maximus in 383. This theory was first put forward by Dr. H. H. E. Craster in 1914, and it is supported by the results of the Birdoswald Excavations of 1929. It rests mainly on negative evidence, but it is evidence which was formidable in 1914 and is still more formidable to-day, since no coin later than 383 has ever been reported from the Wall—except the one which is the subject of this paper. Moreover it becomes evident that the reason for the cessation is historical and not numismatic, when it is realised that at other sites, some quite close to the Wall, the coin-series extends beyond 383. Romano-British archaeologists, therefore, impressed by the weight of the evidence, have assumed that the one exception to Dr. Craster's rule, the coin of Arcadius from Heddon-on-the-Wall, is a stray, the presence of which does not imply a garrison on the Wall at the time of its issue. Dr. Schultz, however, writing in this Journal, seems to imply that it is not a stray, but a testimony to the maintenance of a garrison on the Wall down to the fifth century.

Text I (Vergil). Found by Petrie at Hawara eighty years ago, P(apyrus) Hawara 24 is now in the collection at University College, London. It has writing on both sides. On one side is a line of Latin verse repeated over and over; on the other side is other Latin, also repeated over and over. The first editor, A. H. Sayce, whose publication was merely two lines in a check-list, recognized, however, that the line on one side is Vergil, Aeneid 2, 601:

Text I Non tibi Tyndaridis facies [invisa Lacaenae]

The last two words are torn away and are wholly missing, but all of the rest is plain. Certainly both sides were written for practice. Scholars commonly felt that exercises of this sort have little interest, and so the papyrus had received only a few mentions (bibliography with complete references infra).

At Margidunum (Notts), halfway between Leicester and Lincoln, I have been carrying on excavations for several years, through the kindness of the owner, Mr. F. W. Dobson, J.P., in one of the three fields of the site, the only field at present available. Although much remains to be done, my results serve already to substantiate the surmise that this camp formed a link in the chain of Claudian frontier-posts established by P. Ostorius Scapula in A.D. 47, between the Severn and the Trent, when (according to Bradley's generally accepted emendation of Tacitus, Annals, xii, 31, 32) detrahere arma suspectis cunctaque cis Trisantonam et Sabrinam fluvios cohibere parat. Nothing was known of this camp beyond the name Margidunum, which occurs twice in the Antonine Itineraries (Iter 6 from London to Lincoln, and Iter 8 from York to London). It is about 8 acres in extent, and is rhomboidal in outline (plan, plate VIII), recalling in this respect the Claudian camp of Hofheim. The Fosse Way, which now traverses it diagonally, was not the original Roman road. In those days, undrained marshes protected the camp on the south, and the approach from this direction was by a short causeway (bordered by ditches), which can still be traced from a natural ridge that comes near the SE. angle of the camp. Thence the main road of the camp ran diagonally across it to NW., and from the NW. angle another well-marked causeway leaves the camp and can be traced for a considerable distance to the NE.; it is still so conspicuous that it is known locally as the ‘hump.’ The entrances to the camp have not yet been explored, for the fields in which they occur are unfortunately not available for excavation.

The original purpose served by this curious panel of mosaic and marble (plate I), which is decorated with scenes from the life of Achilles, has not yet been determined, nor has any successful attempt been made either to date it or to assign it definitely to any school or centre of art. The mosaic which occupies the centre and corners is probably rightly attributed to the Cosmati, who made the ‘ambon’ in S. Maria in Ara Coeli, of which this panel formed a part. The circular band of marble, decorated in relief, is generally regarded as the top of a puteal or well-head although, as has been observed, it is hardly consistent with the practical sense of the Greeks and Romans to cover with decoration a surface of marble so liable to rough usage. More probably the band formed the border of a large shallow bowl, a possibility not considered heretofore owing to the difficulty of close examination arising out of its post-classical use.

Tentative suggestions which conform to the conditions of a scientific hypothesis may play a useful part in furthering the slow and unending process of rediscovering antiquity. Proof or disproof of them must naturally await the discovery of fresh evidence, and such evidence may never come; but if it does, its appearance is doubly welcome when it furnishes proof. It is nearly forty years since my travels in Anatolia took me into the angle of north-western Galatia, lying between north-eastern Phrygia and the southern boundary of Bithynia, which ran along the watershed between the middle Sangarius and its chief tributary the Tembrogius or Tembris. The southern side of this watershed, immediately to the west of the junction of the two rivers, is an infertile and thinly populated district, but farther west, towards the Phrygian border, the mountain slopes open out into a broad plain which runs down to the Tembrogius; near its head are the two villages of Upper (or Great) and Lower Igde-agatch. Westward of this plain and separated from it only by a low ridge is a pleasant valley, watered by an affluent of the Tembrogius.

The two passages have often been compared, but with no agreed result. For instance, H. Last, CQ XVII, 1923, 153 f., and K. Btichner, Lateinische Literatur und Sprache, Bern, 1951, 81 f., give the priority to the invectiva; A. Klotz, Berl. philologische Wochenschrift, XLIII, 1923, 261 f., and O. Seel, Klio, Beiheft XLVII, 1943, 73 ff., to the epistula; R. v. Pöhlmann, Sitzungsb. der K. Bayer. Akad., 1904, 27, and G. Carlsson, Eine Denkschrift an Cäsar über den Staat, Lund, 1936, 57, think that the passages may have a common source; B. Edmar, Studien zu den Epistulae ad Caesarem senem, Lund, 1931, 105 f. leaves the solution open. Progress can hardly be made without fresh evidence.