possible to suppose that a citadel presenting such natural advantages
was not occupied. The Romans, too, probably made use of it, though the
lower quarters were mainly occupied in imperial times. A. Taramelli
(_Notizie degli Scavi_, 1905, 41 seq.) rightly points out that the
nucleus of the Roman _municipium_ is probably represented by the present
quarter of the Marina, in which the streets intersect at right angles
and Roman remains are frequently found in the subsoil. An inscription
found some way to the north towards the amphitheatre speaks of paving in
the squares and streets, and of drains constructed under Domitian in
A.D. 83 (F. Vivanet in _Notizie degli Scavi_, 1897, 279). The
amphitheatre occupies a natural depression in the rock just below the
acropolis, and open towards the sea with a fine view. Its axes are
95-1/2 and 79 yds., and it is in the main cut in the rock, though some
parts of it are built with concrete. Below it, to the south, are
considerable remains of ancient reservoirs for rain-water, upon which
the city entirely depended. This nucleus extended both to the east and
to the west; in the former direction it ran some way inland, on the east
of the castle hill. Here were the _ambulationes_ or public promenades
constructed by the pro-consul Q. Caecilius Metellus before A.D. 6
(_Corp. Inscrip. Lat._ x., Berlin, 1883, No. 7581). Here also, not far
from the shore, the remains of Roman baths, with a fine coloured mosaic
pavement, representing deities riding on marine monsters, were found in
1907. To the east was the necropolis of Bonaria, where both Punic and
Roman tombs exist, and where, on the site of the present cemetery,
Christian catacombs have been discovered (F. Vivanet in _Notizie degli
Scavi_, 1892, 183 seq.; G. Pinza in _Nuovo Bullettino di Archeologia
Cristiana_, 1901, 61 seq.). But the western quarter seems to have been
far more important; it extended along the lagoon of S. Gilla (which lies
to the north-west of the town, and which until the middle ages was an
open bay) and on the lower slopes of the hill which rises above it. The
chief discoveries which have been made are noted by Taramelli (loc.
cit.) and include some important buildings, of which a large Roman house
(or group of houses) is the only one now visible (G. Spano in _Notizie
degli Scavi_, 1876, 148, 173; 1877, 285; 1880, 105, 405). Beyond this
quarter begins an extensive Roman necropolis extending along the edge of
the hill north-e