¹ Paget, R. F., In the Footsteps of Orpheus (London: Robert Hale, 1967Google Scholar).

² We know little about the harbour-facilities at Ischia, where according to Livy (8, 22), the Greeks had made a settlement before they moved over to colonise the better-endowed position of Cumae. The island is very subject to volcanic activity. Strabo (5, 4, 9) mentions several eruptions of Mt. Epomeo in Greek times, one in the early period, another about the time of Hiero of Syracuse that was alluded to by Pindar, (in Pyth. I, 33Google Scholar), and a third said by Timaeus the historian to have occurred ‘shortly before his own time’, i.e. about 350 B.C.; and Pliny, (NH 2, 203Google Scholar) tells of another—perhaps identical with the one Timaeus knew of—that destroyed a town and created a lake, which is actually the modern harbour of Porto d'Ischia. Apart from Porto d'Ischia the island is not specially well provided with protected anchorages; but the eruptions that occurred before and after ancient times make the position hard to recover. See further Buchner, G. and Rittmann, A., Origine e passato dell'isola d'Ischia (Naples, 1947)Google Scholar; Buchner-Niola, D., L'Isola d'Ischia, studio geografico (Naples 1965)Google Scholar.

³ Mahan, Captain A. T., The Influence of Sea-Power in History (London 1892Google Scholar).

⁴ See Strabo 5, 4, 4, who calls it a ὕλη θαμνώδης.

⁵ Thus Boardman, J., The Greeks Overseas (London 1964) 183Google Scholar, to take a recent example.

⁶ In Homer's, Iliad XIV, 30–36Google Scholar. But in view of their objective—Troy itself—the Greeks may have had little choice.

⁷ Compare Strabo's words (5, 4, 4): ‘according to some Cumae is named after the waves (kumata), for the neighbouring shore is surfy and exposed to the wind’.

⁸ The anchors were found by John D. Lewis and Robert Love, Jr., in 1965 off the Galli islands, on the south side of the Sorrento peninsula: they are now in the museum at Salerno. For the size of vessels see esp. Casson, L., The Ancient Mariners (London 1959), esp. pp. 126 ff., 174Google Scholar f.

⁹ Dion. Hal. 3, 44; Strabo 5, 3, 5. Cp. R. Meiggs, Roman Ostia 51 ff.

¹⁰ ‘Redrawing the Coastline of Southern Italy’, Illustrated London News, Jan. 18, 1964. But his views have been challenged on scientific grounds: see Flemming, N. C. and Burton, A. N., ‘Marine Levels in South Italy’, Nature cciii (1964), 1060 ff.CrossRefGoogle Scholar; Schläger, H., Röm. Mitt. lxxi (1964), 245 ffGoogle Scholar.

¹¹ For other examples, see McBurney, C. B. M., The Stone Age of Northern Africa (London 1960), 112Google Scholar; F. E. Zeuner, Dating the Past ⁴ (1958), 127 ff.; and the works of A. C. Blanc quoted by these writers.

¹² See Maiuri, A., I Campi Flegrei ³ (1958), 24 f.Google Scholar

¹³ As was shown by Professor Blanc for Monte Circeo, Zeuner o.c., note II, 224. True, we cannot always be sure what is due to geological movements and what to bradyseism of a local and volcanic nature. We may note that in Ischia there are indications of considerable changes of level in historic times, D. Buchner-Niola (o. c., note 2), 9 ff.

¹⁴ The destructive power of the sea in local gales can be terrifying. In November 1966 the storm which caused the flooding at Florence washed away the isthmus connecting Punta Pennata (the northern arm of Portus Misenus) to the mainland, leaving a channel 100 hundred yards wide and 12 feet deep. Before this I had observed that the isthmus was 100 yards wide and composed of sand-dune; but it had the appearance of having been a permanency for many centuries. It is easy to picture the effects of a storm on an area already reduced to near sea-level by bradyseism.

My geological knowledge of these phenomena is derived from some research I did on the structure of the Warwickshire, Birmingham (S. Staffs) and Kent Coalfields. These resulted in papers in 1936 for the Institution of Mining Engineers and the Coal Trade Review on the Birmingham coalfield, and some unpublished reports on the structure of the North Kent coal seams made to the Chislet Colliery. These researches dealt with the formation of lagoons and the effects of faulting in the coalfields.

¹⁵ The city walls in the sixth century, so far as they have been traced, did not reach as far as the amphitheatre (see Gabrici, E., Mon. Ant. xxii, 1913, IIGoogle Scholar f.). But some early pottery has certainly been found in the neighbourhood and near the Villa Vergiliana, see McKay, A. G., Naples and Campania (1962), 63Google Scholar.

¹⁶ Ant. Rom. 7, ch. 3–8.

¹⁷ In the book cited above, note 1.

¹⁸ The role played by the Cumaean fleet is admitted by the scholiast on Pind, ., Pyth., I, 71Google Scholar and Diod. II, 51; and for the helmet dedicated by Hieron at Olympia, and now in the British Museum, see Tod, GHI no. 22. Other indications of a navy are to be seen in Dion. Hal. 7, 3, 4 and 7, 7, 4 (a mention of dockyards).

¹⁹ Cp. A. G. McKay, o.c. note 15, 66 ff.

²⁰ Strabo 5, 4, 4; Diod. 12, 76 (under 421 B.C.); Livy 4, 44 (under 420 B.C.).

²¹ Strabo 5, 4, 4: εἰσὶ δὲ καὶ κητεῖαι παρ᾿ αὐτοῖς ἄρισται. The word κῆτος is used of tunny-fish, see L-S-J, s.v.

²² Pliny, , HN 19, 11Google Scholar; cp. Cato, Agr. 135; Grattius, Cyneg. 35.

²³ For the attempts on Naples, see Livy 23, 1–15; for Puteoli, 24, 13, 6. For the assault on Cumae, 23, 35–37; again, Livy stresses Hannibal's anxiety for a port (36, 6—‘quamquam haud modice cupiebat ut, quia Neapolim non potuerat, Cumas saltem maritimam urbem haberet, tamen …’). The consul Gracchus also had a naval force under his command at Cumae (Livy 23, 38, 3), but we are not told where it was based.

²⁴ Brown, F. E., ‘Cosa I: History and Topography’, MAAR XX (1951), 89 ffGoogle Scholar.

²⁵ The scouring-channel and sluices in the port of Sidon are referred to by Achilles Tatius, Leucippe and Clitophon I ff.), and the remains have been carefully studied by Poidebard, A. and Lauffray, J., Sidon, aménagements antiques du port de Saida (Beyrouth 1951), 51 ffGoogle Scholar. When the port of Ephesus silted up in the second century B.C. (cp. Livy 37, 14, 5), Attalus II of Pergamum attempted to increase the flow by a narrow channel (Strabo 14, 1, 24). The large port of Seleucia Pieria on the Orontes had acute silting problems; a rock-cut tunnel was designed to produce a scouring action, evidently in Flavian, times (CIL III, 6702Google Scholar; cf. Paus. 8, 29, 3 and Libanius, , Or. xi, 159Google Scholar; for other sources see P–W II A, 1190 ff.; and Poidebard and Lauffray, o.c., 31 ff.). Ancient engineers often elsewhere would employ a scouring current to meet the danger of silting: e.g. Ostia (Plut., , Caes. 58, 1Google Scholar), Ravenna (Strabo 5, 1, 5; Pliny, , HN 3, 119Google Scholar; Jordanes, , Getica 29, 147 ffGoogle Scholar.), Caesarea in Palestine (Jos., , Ant. J. 15, 331–41Google Scholar); and for Cnidus, Alexandria and Apollonia see Lehmann-Hartleben, K., Antike Hafenanlagen (Klio Beiheft XIV, 1923Google Scholar).

²⁶ Appian, , BC 4, 84–86Google Scholar; Dio 48, 17–29, 2; Livy, Per. 123 for the events to Philippi. From there to the peace of Misenum: Appian, , BC 5, 18, 25, 67–74Google Scholar; Dio 48, 19, 3–20, 30–31, 36–38; Livy, Per. 127; Florus 4, 8; Velleius 2, 72–73, 77; Plut., Ant. 32. From Misenum to the appointment of Agrippa: Appian, , BC 5, 77–96Google Scholar; Dio 48, 45–49; Livy, Per. 128; Velleius 2, 79; Plut., Ant. 35.

²⁷ Strabo 5, 4, 4.

²⁸ Dio 48, 48, 6; cp., however, Appian, BC 5, 97. Rhegium and Vibo remained loyal to Octavian, as he promised them that they would not be made veteran colonies: Appian, , BC 4, 86Google Scholar.

²⁹ Appian, , BC 5, 78, 80 ff.Google Scholar; Dio 48, 46 ff.

³⁰ Appian, , BC 5, 89Google Scholar.

³¹ Appian, , BC 5, 93, 95Google Scholar; Dio 48, 49, 2; 49, 1.

³² Dio 48, 49; cp. Appian, , BC 5, 92Google Scholar.

³³ Suetonius, Div. Aug. 16.

³⁴ Dio 48, 51, 5, which makes clear that the decking, as well as the training of the rowers, was done at Portus Julius (ἐπειδἠ τάχιστα τὰς ἔσπλους ἐξεποίησε τάς τε ναῦς καὶ τούς ἑρέτας ἤθροισε, καὶ τὰς μὲν κατέφραττε, τοὺς δὲ ἐπ᾿ ἰκρίων ἐρέττειν ἤσκει).

³⁵ For this part of the struggle, see Appian, , BC 5, 97–122Google Scholar; Dio 49, 1–11, 1.

³⁶ Appian, , BC 5, 98–99Google Scholar; Dio 49, 1.

³⁷ Dio 49, 1; Appian, , BC 5, 106, 118–119Google Scholar.

³⁸ Appian curiously enough has nothing to say about Agrippa's preparations. See, however, Dio 48, 49–51; Suetonius, Div. Aug. 16; Velleius 2, 79; Floras 4, 8; Vergil, , Georgics 2, 161 ff.Google Scholar, with comm. ad loc.; Cassiodorus, Chron. under the year 717 a.u.c.; Pliny, , HN 36, 125Google Scholar; Strabo 5, 4, 4–7.

³⁹ Strabo 5, 4, 5.

⁴⁰ For the oyster-beds, started here c. 100 B.C., see Val. Max. ix, 1, 1; Pliny, , HN 9, 168Google Scholar.

⁴¹ Strabo 5, 4, 6.

⁴² The storm is attested by Servius' commentary on Verg., , Georg. 2, 162Google Scholar ( = ‘Servius Auctus’, ed. Thilo iii, 235): ‘Agrippa in secundo vitae suae dixit excogitasse se ut ex Lucrino lacu portum faceret. Verum huius operis gloria Augusto cessit. Indignatum ideo dixit, quia quo tempore in Lucrinum lacum mare immissum est, deinde, terra effossa inter ipsum Lucrinum et Avernum, contigit ut duo lacus miscerentur, tanta tempestas orta est ut prodigii loco habita sit et nuntiatum sit, simulacrum Averni sudasse; propter quod pontifices ib, piacularia sacra fecerunt’.

⁴³ See Maiuri, o.c., note 12, 154 f. for remains of the navalia; a drawing of the remains may be found in A. D'Arrigo, Natura e Tecnica nel Meszogiorno (1956), 624. There is early imperial masonry, not easily datable, under the bath-building known as the ‘Temple of Apollo’, see Maiuri, o.c., 158 f.

⁴⁴ Sestieri, P. C., Paestum ⁴ (1958), 34Google Scholar.

⁴⁵ Strabo 5, 4, 5 and 7 appears to say clearly that this tunnel and that from Naples to Puteoli were built by the same man, see Beloch, Campanien ² 83 f., and doubts (e.g. Mommsen's in CIL X) seem misplaced. Strabo adds that Cccceius, by building tunnels, was following ‘an ancestral tradition of the region’, which must mean that he lived in Campania. His full name was L. Cocceius Auctus, and he was a freedman of L. Cocceius and C. Postumius Pollio (another well-known architect). He has left his signature on the temple of Augustus at Puteoli (CIL X, 1614), which is the modern Duomo, and on part of a large pediment from Cumae, (CIL X, 3707Google Scholar), which may be part of a building associated with the tunnel or the road that passed through it.

⁴⁶ The cape is a well-known hazard, cp. Tac., , Ann. 15, 46Google Scholar; and according to legend Misenus was drowned there, Verg., , Aen. 6, 174Google Scholar; Strabo 1, 2, 18.

⁴⁷ I take it as certain, with A. Maiuri, o.c., note 12, 135 f., that the tunnel from Avernus to Cumae and that driven beneath the acropolis are part of the same scheme. Both had a similar system of lighting by shafts.

⁴⁸ The opus reticulatum, apparently the only evidence for the date of these works, was observed by myself and confirmed by Lewis. By Seneca's time, the site was occupied by the villa of Servilius Vatia (see Ep. 55 for an interesting ancient description); and perhaps the name Torregaveta is derived from ‘Torre di Vatia’ (as suggested by McKay, o.c., note 15, 4).

The Roman masonry is mostly covered by modern repairs, but reticulate masonry can be seen in the south wall of the canal near the point where it enters the Lake Fusaro. From here, some yards further towards the sea, several sluice-gate recesses can be seen in the tufa walls on both sides of the canal (Pl. XIII, 2), and are very similar to those at Cosa (Brown, o.c., note 23, 94 f.). The tunnel that pierces the hill of Torregaveta is some 10 m. high and 3 m. wide; it has some brick vaulting, probably also of Roman date.

The seaward entrance to the tunnel, which is inaccessible except by boat, was studied by Lewis in April, 1968, to whom this note owes much. The deep water is fed into the tunnel over a lip (now some 2 m. below the surface), before passing down the gently sloping floor of the tunnel. On the south side of the entrance a spur of tufa projects into the sea; beyond this, some 4–5 m. further out, two masonry jetties, now just awash, were angled to deflect the waves like training walls towards the spur and the tunnel entrance, and to make the most of the frequent westerly winds. Another feature paralleled at Cosa is the skylight shaft cut in the roof of the tunnel at a point some 6 m. from the entrance; its purpose was apparently to reduce the shock of waves as they broke in the tunnel entrance in very rough weather.

⁴⁹ Compare Maiuri, o.c., note 12, 123 ff. But his identification of this very defensive-looking tunnel with the Sibyl's Cave remains a mere guess, and has its difficulties. I need not discuss these here, but see now Napoli, M., Atti IV Convegno di Studi sulla Magna Grecia 105 ff.Google Scholar

⁵⁰ The major buildings of the Roman town date mainly from the late Republic and Augustan periods, with some additions in the later empire—see Johannowsky, W., Enc. Arte Antica, s.v. Cuma (1959)Google Scholar; but neither they nor the inscriptions throw any light on the fate of the port. Belisarius placed garrisons in Naples and Cumae in A.D. 536; to him both sites were simply strongpoints and easily defended. Recaptured by Totila, Cumae was once more besieged by Narses in A.D. 552–3. See Bury, J. B., History of the Later Roman Empire ii, 180, 331, 271–77Google Scholar; A. Maiuri, Par. d. Pass. 1949, 41 ff.

⁵¹ Cp. Tac., , Ann. 4, 5Google Scholar; Suetonius, Div. Aug. 49; C. Starr, The Roman Imperial Navy 13 ff.

⁵² The site of the ‘Admiralty House’ can be fairly clearly established from the letters of Pliny the younger. In this area, in February 1968, a shrine of the local Augustales (?) was discovered, quite by accident, with two statues representing Vespasian and Titus: more recently still, an adjoining chamber containing fragments of an equestrian statue of Nerva has come to light. These interesting finds are to be published by the Italian authorities: in the meantime, see the report by Professor A. De Franciscis in the issue of Il Mattino for 5 July, 1968.