AJRCH.EOLOGY.] ROME 813 down to the valley of the Coelian ; at other places the east side is precipitous. The buildings of the emperors have, however, made such radical changes in the form of the ground that any degree of certainty on this point is impossible. This third gate, according to Varro, who mentions (L.L., v. 164-5) all three, was called Janu- alis. 1 With regard to the last gate there is some doubt whether he is right in its position; Macrobius (Satur., i. 9) places it on the slopes of the Viminal. 2 Remains of Period of Kings (753-509 B.C.). The most important remains of the regal period are the great wall, principally the work of Servius Tullius, by which he included within one circuit the separately fortified hills which were then in- habited, the Palatine, Capitoline, Aventine, Quirinal, and Ccelian, and added two more, the Esquiline and Viminal. These seven hills formed the Septimontium 3 (Varro, L. L., iv. 41, vi. 24). Of even earlier date was the fort (Liv., i. 33) on the top of the Jani- culum, connected by walls with the Pons Sublicius, some massive foundations of which still exist, though now buried near the church of S. Pietro in Montorio. Existing remains show that the great wall is of more than one date ; part is probably earlier than Servius, and may be remains of the wall which, according to Livy (i. 36, 38) and Dionysius (iii. 37), was planned and in part carried out by Tarquinius Priscus ; it would seem impossible that a work of such gigantic magnitude could have been begun and completed in the lifetime of one man. It should, however, be remembered that a complete circuit of new wall was not required in this undertaking ; each, probably, of the five hills first inhabited had its own fortifica- tions, and these were utilized in the line of the new wall. The space thus included was divided by Servius for political, military, .ind religious purposes into four regiones (Varro, L. L., v. 46-54) (1) the Suburana, including the Ccelian, and probably most of its adjacent valleys, with the Subura, the Carinse, and part of the Esquiline the derivation of Suburana from sub urbe is from Junius, quoted by Varro ; (2) Esquilina, the main part of the Esquiline, including the Oppius and Cispius (es-quil-iae, dwellings outside, cf. in-quil-inus) ; (3) Collina, the Viminal and Quirinal, which were called " colles " in contradistinction to the other five hills, which were called "montes" ; (4) Palatina, the Palatine and its adjacent spurs, the Velia and Germalus. It will be observed that these four regiones do not include the Aventine, the Capitol, and some of their adjacent valleys, an omission for which it is not easy to account. 4 Excavations of the last fifteen years have done much to deter- mine the line of the Servian wall (see Plate VI. ), especially the great works undertaken in laying out a new quarter of the city on the Quirinal, Esquiline, and Viminal, which have laid bare and then mostly destroyed long lines of wall, especially along the agger. P>eginning from the Tiber, which the Servian wall touched at a point near the present Ponte Rotto, and separating the Forum Olitorium (outside) from the Forum Boarium (inside), it ran in a straight line to the Capitoline Hill, the two crests of which, the Capitolium and the Arx, with the intermediate valley the Asylum, were surrounded by an earlier fortification, set (Dionys., ix. 68) eiri oty . . . Kal Tr^rpcus diroTijUots. In this short space there were three gates, (1) the Porta Flumentana next the river (see Cic., Ad Att., vii. 3 ; Liv., xxxv. 19, 21) ; (2) Porta Triumphalis, site un- known and usually only mentioned in connexion with triumphal processions (see Cic., In Pis., 23 ; Joseph., Bell. Jud., vii. 5, 4), it was probably not used except on the occasion of triumphs ; (3) Porta Carmentalis, close to the Capitolium. 5 From the Capitoline Hill the wall passed to the Quirinal along a spur of elevated ground, afterwards completely cut away by Trajan. Close to the Capitol was the gate afterwards called the Porta Ratumena, whence issued the Via Lata (Plut., In Publ., 13; Plin., H.N., yiii. 42). Remains of the wall and foundations of the gate exist in Via di Marforio, .Nos. 81 C and 81 E. After passing Trajan's forum, the first remains of the walls are on the slope of the Quirinal in the Colonna gardens. Near the foot of the Quirinal was the Porta Fontinalis (Liv., xxxv. 10). A piece of the wall has been exposed in the new Via Nazionale, and also an archway under the Palazzo Antonelli, which has been thought to be the Porta Fontinalis. This arch is, however, only 6 feet 6 inches wide and (to the springing of the arch) 5 feet high, which seems too small for one of the principal gates. The Porta Sanqualis (see Festus, ed. Miiller, p. 345) was also on the Quirinal, probably on the slope between the Trevi fountain and the royal palace. Its position is indicated by the 1 See Dionys., ii. 65 ; Plut., Rom., ix. ; Dion Cass., various fragments ; Liv., i. 7, 9 ; Sol., i. 17 ; Festus (ed. Miiller), pp. 258, 262. 2 See Becker, De Muris et Portis Romx, Leipsic, 1842; also "La fondazione di Roma," in Ball. Comm. Arch., ix., 1881. 3 At an early date the terra "Septimontium" had a different meaning (see Plut., Qutes. Rom., 69 ; Burn, Rome, p. 37). 4 Becker suggests (Handbuch, i. 386) that the Capitoline Hill was excluded on account of its sacred character, while the Aventine was not yet thickly populated, and the Janiculum was only occupied by the fort ; see also Ann. lust., 1861, p. 61. 5 See Sol., i. 13 ; Liv., ii. 49, xxiv. 47, xxv. 7; Ascon., Ad Cic. in Toga, p. 60, Orell. existence of some tombs which give the line of the road. On the north-west of the Quiriual was the Porta Salutaris (Festus, p. 327 ; Liv., ix. 43), probably near the "Quattro Fontane." In the Barberini palace gardens, and especially in those of the Villa Bar- berini (Horti Sallustiani), extensive remains of the wall have been recently exposed and destroyed, which was also the fate of that fine piece of wall that passed under the new office of finance, with the Porta Collina, which was not on the line of the present road, but about 50 yards to the south (see Dionys., ix. 68 ; Strabo, v. 3). Thus far in its course from the Capitol the wall skirted the slopes of hills, which.were once much more abrupt than they are now ; but from the Porta Collina to the Porta Esquilina it crossed a large tract of level ground ; and here, by the construction of his great agger, Servius gained the strength which elsewhere was given by the natural formation of the hills. The whole line of this agger has been recently traced and mostly destroyed. About the middle of it the Porta Viminalis was found in 1872 ; it stood, as Strabo says, virb /z^cry r$ x^art, and from it led a road which passed through the Porta Chiusa (ancient name unknown) in Aurelian's wall. Foundations of the Porta Esquilina were found in 1876 close behind the arch of Gallienus. The further course of the wall across the valley of the Colosseum, with its Porta Querquetulana . and Porta Ccelimontana, probably a little beyond, is the least known part of the circuit. Hence the wall skirts the slopes of the Ccelian to the valley along which the Via Appia passed through the Porta Capena, near the church of S. Gregorio. Its line along the Aventine is fairly distinct, and near S. Balbina and in the Vigna Torlonia are two of the best-preserved pieces, the former 11 courses high (22 feet), the latter 25 (50 feet). Under the Aventine it appears to have touched the river near the existing foundations supposed to be those of the Pons Sublicius. The Porta Trigemina was close by the bank. Hence to our starting-point the river formed the defence of the city, with its massive quay wall, the KOTJ &KTT) of Plutarch (Rom., 20). A fragmentary passage of Varro (L. L., v. 163) mentions two other gates, Nsevia and Rauduscula, "the bronze gate," but their positions are unknown. The site of the Porta Kavalis is also very doubtful ; it was probably not in the Servian wall. The wall is built of blocks of tufa, usually the softer kinds, but Its con- varying according to its position, as in most cases the stone used struc- was that quarried on the spot. In some places a good deal pftion. peperino is used. The blocks average from 23 to 24 inches in thickness roughly 2 Roman feet and are laid in alternate courses of headers and stretchers. The best preserved piece of wall that on the Aventine in the Vigna Torlonia has one complete arch and the starting of another ; their sills are about 34 feet from the ground outside, and probably level with the ground inside (see fig. 8). FIG. 8. Part of Servian wall on Aventine. These arches, though built of the harder stone, are not later inser- tions, but are contemporary with the wall itself. The blocks, both beds and vertical joints, are very carefully worked and set in mortar; in most cases they are bevelled round the joints. Some blocks are of great length, d/wi&oi i0oi Dionysius calls them. At this place the wall is backed with a thick mass of concrete ; the use of the arched openings is doubtful : they may have been embrasures for catapults. The discoveries of recent years have shown the cor- Agger of rectness of the description given by Dionysius (ix. 68) of the great Servras. agger, with its wall and foss, begun by the earlier kings and com- pleted by the last Tarquin. Fig. 9 shows it in section ; the earth taken from the foss (which was 100 feet wide at the bottom by 30 deep) was heaped up to form the agger, and was kept in its position by a lofty retaining wall on the front and a lower one behind. The outer wall was in places strengthened with massive buttresses closely set, or with towers ; in other places it had no projections. The back wall, the position of which shows the thickness of the
