still be traced for nearly a mile in two parallel lines, with the
transverse wall which unites them. Certainly all these indications
seem to point to the existence at this spot of a great provincial city
of the Empire, and to make Mr. Evans' conjecture more probable than
that of M. Lenormant, who identified the ruins at Roccella with those
of Castra Hannibalis, the seaport of Scylacium. It would seem
probable, if Mr. Evans' theory be correct, that the city may have been
removed to its present site in the early middle ages, in order to
guard it against the incursions of the Saracens.

[Sidenote: The Vivarian Monastery.]

II. As to the situation of the _Vivarian Monastery_ Mr. Evans comes to
nearly the same conclusion as M. Lenormant. Both place it on the
promontory of Squillace (eastward of Staletti), and, as Mr. Evans
observes, 'only such a position can be reconciled, on the one hand,
with the presence of an abundant stream and rich Campagna, on the
other with the neighbourhood of caves and grottoes on the sea-shore.'
But while M. Lenormant places it at a place called Coscia, almost
immediately to the north of and under Staletti, Mr. Evans pleads for
the site now occupied by the Church of S. Maria del Mare, on the cliff
top, very near the sea, and about three kilometres south of Staletti.
This church is itself of later date than Cassiodorus, and probably
formed part of the work of restoration undertaken by Nicephorus Phocas
in the Tenth Century; but there are signs of its having formerly
joined on to a monastery, and some of the work about it looks as if
materials taken from the Cassiodorian edifice had been used in the
work of reconstruction.

[Sidenote: The Fons Arethusae.]