¹ The survey and excavation took place for six weeks in August and September in conjunction with Professor Mario Torelli's 1977 season of land excavations in the sanctuary at Gravisca. I am indebted to Professor Torelli both for his invitation to carry out this work and for arranging it under his permit. The project was conducted as an extension of the Tuscan Port Survey under the auspices of the American Academy in Rome with the kind cooperation of the Tuscan Port Survey Director, Dr. Anna Marguerite McCann, to whom I am also deeply grateful for help and support. Special thanks are due to expedition members Cynthia Orr, Nan Bray, Richard Swete and Tom Gross for their invaluable and energetic assistance throughout the long weeks of diving; to Signor Ezio Alessi and his survey team, who generously volunteered their time and equipment to help us map the site, and to Signori A. Pontani, G. Ciurluini and C. Borzacchi and Signora M. Slaska for their logistical support at Gravisca; to Dr. John Oleson, Dr. Elizabeth Will and Mr. J. Richard Steffy for their advice on the research; and to the countless other individuals who helped with various phases of the project. Funding was provided by a grant from Texas A&M University, diving and excavation equipment was contributed by the American Academy in Rome, and additional support was given by the Institute of Nautical Archaeology. This paper is a condensed version of my Masters Thesis presented to Texas A&M University in May 1978.

² For founding of the Roman colony see Livy xl. 29. 1 and Velleius Paterculus i. 15. 3. For Etruscan Gravisca see Silius Italicus viii. 475 and Virgil, Aeneid x. 134.

³ Torelli, M., ‘Il Santuario di Hera a Gravisca’, La Parola del Passato 136 (1971), 44–67Google Scholar; idemet al., ‘Gravisca—Scavi nello citta etrusca e romana. Campagne 1969 e 1970’, Notizie degli Scavi Series 8, 25, 1 (1971), 195–299.

⁴ Torelli, M., ‘Il Santuario Greco di Gravisca’, La Parola del Passato 177 (1977), 398–458Google Scholar. For the implications of this discovery on Greek-Etruscan relations, see Torelli, M., ‘Greek artisans and Etruria: A problem concerning the relationship between two cultures’, Archaeological News 5 (1976), 134–8Google Scholar.

⁵ Evidence for further commercial activity in the Roman period can be found in Pliny's reference to Gravisca's wine (NH xiv. 67-8) and coral (NH xxxii. 21) in the first century A.D.

⁶ Following the Maritime Itinerary of Antonius, E. Westphal, ‘Topografia dei contorni di Tarquinia e Vulci’, AnnIst (1830), 32, placed the site of Gravisca on the right bank of the Marta. Dennis, G., The Cities and Cemeteries of Etruria vol. I ((London, 1848), 393Google Scholar, placed the port and colony 2 mi. upstream in accordance with what he believed were the remains of an ancient quay, sewer and causeway. Dasti, L., Notizie Storiche Archeologiche di Tarquinia e Corneto (Tarquinia, 1910), 27–8Google Scholar, incorporated into this view the description by Rutulius Namantianus (De redito suo i. 282–4) of Gravisca's scattered house tops north of the Mignone, and placed it back at the mouth of the Marta on the left bank. On the other hand, Canina, L., L'Antica Etruria Marittima vol. II (Rome, 1846), 46Google Scholar; AnnInst (1847), 92, note 1, used Strabo's (v. 218) figure of 180 stadia for the distance from Gravisca to Pyrgi to suggest that the site lay at the mouth of the Mignone River.

⁷ Although Pallottino, M., MonAnt 36, 1 (1937), 579–80Google Scholar, suggested that the port might lie at the mouth of the Marta, he also wrote of great blocks of limestone lying underneath the jetty at the entrance to the salt marsh which he believed were part of an ‘etrusco-romana cloaca … certamente di un avanzo dell'antico porto interno, scavato ad imitazione di un “cothon” greco-punico probabilmente nella prima metà del II secolo av. Cr'. We found no sign of such construction, however, and the jetty itself is most likely the remains of the nineteenth century construction by Pope Pius VII to drain the salt marsh for salt production. See Zeri, A., ‘I Porti del litorale Romano’, Monographia storica dei porti dell'antichita nella penisola Italiana (Rome, 1905), 238Google Scholar.

⁸ A. Pasqui, Notizie degli Scavi (1885), 519, note 3.

⁹ Quilici, L., ‘La Via Aurelia da Roma a Forum Aureli de Luca’, Quaderni dell'Istituto di topographia antica dell'università di Roma, IV (1968), 107–20Google Scholar. According to the plan on p. 109 (fig. 251), Quilic projected a curvilinear mole extending c. 500 m. out from the promontory.

¹⁰ Schmiedt, G., ‘I Porti Etruschi’, Atlante aereofotografico delle sedi umane in Italia (Florence, 1970), 121–2Google Scholar. The aerial photographs shown in table CXXX, fig. 3, did not reveal any traces of under-water structures.

¹¹ For this interpretation see Fòti, G., ‘The Principal Cities of Southern Etruria and their Special Characteristics’, in CIBA Foundation Symposium on Medical Biology and Etruscan Origins (London, 1959), 14Google Scholar; Hencken, H., Tarquinia, Villanovans, and Early Etruscans vol. I (Cambridge, 1968), 17Google Scholar; and, most recently, Grant, M., The Etruscans (NY, 1980), 131Google Scholar.

¹² A. Zeri, op. cit. (in note 7), 237–8; C. de Cesaris, ‘Il Porto di Corneto’, Bollettino delle attivita nell'anno 1976 (1977), 50–7.

¹³ A description of this bastion from the sixteenth century was given by Guglielmotti, P. A., Storia delle fortificazione nella spaggia romana. Risarcite ed accresciute dal 1560 al 1570 (Rome, 1880), 507Google Scholar. Its function as ‘hotel, barracks, stable, customs house, hospital, prison and warehouse’ was mentioned, but no reference was given for the date of its original founding.

¹⁴ See Kretschemer, K., Die Italienischen portolane des Mittelalters (Berlin, 1909), 486, 597Google Scholar.

¹⁵ For discussion of coastal silting and sea-level change along the Tyrrhenian coast since Etruscan times see Schmiedt, G., Il Livello antica del Mar Tirreno. Testimonianze dei resti archeologiche (Florence, 1972)Google Scholar, and Bruno, V., ‘The Mystery of the Etruscan Coastline’, Archaeology 26 (1973), 198–212Google Scholar. For Gravisca in particular see below note 41 and related discussion.

¹⁶ Apart from the few limited investigations of Porto Clementino conducted from shore by Pasqui, Quilici and Schmiedt and one brief series of dives on the site by A. M. McCann's Tuscan Port Survey team in 1968, no surveys or underwater investigations had been attempted at Porto Clementino before 1977.

¹⁷ Poseidonia Oceania overruns all the rocky or sandy substrata along the Italian coast and southern France, growing in depths from the surface to about 40 m. Because it can cover pure sandy areas as well as rocks, it is often difficult at first glance to tell which areas are relative to the interpretation of an ancient breakwater. For discussion of this and other marine features see Nesteroff, W. D., ‘Geological Aspects of Marine Sites’, Underwater Archaeology—A Nascent Discipline (Paris, 1972), 175–83Google Scholar.

¹⁸ The dotted line indicates the extent of the area mapped, although reconnaissance dives were conducted beyond these areas searching for signs of man-made remains.

¹⁹ Clementino', Porto, Guida d'Italia del Touring Club Italiano Lazio (Milan, 1964), 125Google Scholar. Local Italians report that this was done by the Germans to prevent the Allies from landing at Porto Clementino and using it as a marine base.

²⁰ This tool was first perfected by J. Shaw at Kenchreai and is described in Shaw, J., ‘Shallow-water Excavations at Kenchreai’, AJA 71 (1967), 230CrossRefGoogle Scholar, and idem, ‘Shallow-water Excavations at Kenchreai: II’, AJA 74 (1970), 179–80.

²¹ This opening was also observed by Quilici in his projection of the ancient mole, op. cit. (in note 9), 109.

²² It should be noted that in all the sand trenches excavated, depths were restricted to about 1 m. at the deepest, as we were unable to use the large steel caissons which would normally have kept the sand from caving in. For use of these caissons in other port excavations, see McCann, A. M. and Lewis, J. D., ‘The Ancient Port of Cosa’, Archaeology 23, 3 (1970), 204–6Google Scholar, and McCann, et al. , ‘Under-water Excavations at the Etruscan Port of Populonia’, JFA 4 (1977), 283Google Scholar.

²³ This sample is of interest in that early exploitation and shipment of iron ore from Elba is associated with the Etruscans. See McCann, A. M., op. cit., and ‘Survey of the Etruscan Port of Populonia’, Muse 5 (1971), 20Google Scholar.

²⁴ See catalogue in Section V. Due to limitations of time, only identifiable necks, bases, rims and handles were collected and catalogued.

²⁵ For Cosa see McCann and Lewis, op. cit. (in note 22), 204; McCann, A. M., ‘The Harbor and Fishery Remains at Cosa, Italy’, JFA 6, 4 (1979), 391–411Google Scholar; and the forthcoming publication by A. M. McCann et al., The Portus Cosanus: an Early Roman Port and Fishery, shortly to be published by Princeton Press. For Populonia: McCann et al., op. cit. (in notes 22 and 23); Pyrgi: Oleson, J., ‘Underwater Survey and Excavation in the Port of Pyrgi (Santa Severa)’, JFA 4, 3 (1977), 297–308Google Scholar.

²⁶ See description of Cosa's breakwater in Lewis, J. D., ‘Cosa: An Early Roman Harbor’, Proceedings of the Twenty-Third Symposium of the Colston Research Society (London, 1973), 233–58Google Scholar.

²⁷ Cf. Lewis, op. cit., 239. The parallel here is closest to Cosa, for at Pyrgi the rocks are piled in jetties rather than separated mounds.

²⁸ McCann, op. cit. (in note 25), 396.

²⁹ For Pyrgi see Oleson, op. cit. (in note 25), 304. Civitavecchia's reef is described in Pliny, Epistulae vi. 31.

³⁰ Lewis, op. cit. (in note 26), 241. On sand base foundations for rubble mounds in general see Minikin, R. R., Winds, Waves, and Maritime Structures (London, 1950), 90Google Scholar. Note that recent geological studies at Cosa have revealed the possibility that the rocks on the large breakwater closest to shore may lie on a natural bedrock tongue. See forthcoming publication by A. M. McCann, et al., op. cit. (in note 24).

³¹ The natural volcanic shoals at Pyrgi were visible underwater at several points around the bottom edge of the moles where the layer of tumbled rocks ended. See Oleson, op. cit. (in note 25), 304.

³² Dubois, C., ‘Observations sur un passage de Vitruve’, MélRome 22 (1902), 442Google Scholar. For more information on the actual construction technique see Lewis, op. cit., 241, and Shore Protection, Planning and Design. U.S. Army Corps of Engineers. Coastal Research Centre: Technical Report No. 4 (Washington D.C., 1966), 345.

³³ N. Flemming, ‘Archaeological Evidence for Eustatic Change of Sea Level and Earth Movements in the Western Mediterranean During the Last 2,000 Years’, Geological Society of America Special Paper 109 (1969), 109.

³⁴ A Similar wind pattern is generally thought to have existed in antiquity, cf. Rougé, J., Recherches sur l'organisation du commerce maritime en Méditerranée sous l'empire Romain (Paris, 1966), 34–5Google Scholar.

³⁵ Lewis, op. cit., 240.

³⁶ For further explanation of this design at Cosa see McCann, op. cit. (in note 25), 396, and Lewis, op. cit., 237–42. A scouring channel was also evident between the shoreline and the two jetties found at Pyrgi, cf. Oleson, op. cit., 307, although this cannot be used as a parallel with Gravisca.

³⁷ This wind driven surface current is a result of the natural 45-degree deflection of the current to the right of the wind direction caused by the Coriolis effect.

³⁸ Current velocity measurements taken at this point using a small drogue showed an average of 0·6 knots.

³⁹ A more thorough penetration through the grass bed than was possible with our equipment would, of course, be necessary to substantiate this suggestion.

⁴⁰ Shepard, F. P., Submarine Geology (NY, 1963), 169Google Scholar, and personal communication Captain T. K. Treadwell, Jr., Department of Oceanography, Texas A&M University.

⁴¹ As yet there has been no agreement on the sea-level change at Gravisca, where figures range from a 0·50–1·50 m. rise since Etruscan times. Although a recent publication by Frau, B., Gravisca. Il porto antico di Tarquinia e le sue fortificazioni (Gruppo Archeologico Romano, 1980)Google Scholar, states the rise has been 1·40–1·50 m. since the Etruscan period, G. Schmiedt, op. cit. (in note 15), 316, estimated a 1 m. rise since Roman times for the Tyrrhenian coast in general, while recent data obtained from the Marta and Mignone rivers and several points south by Pirazzoli, P. A., ‘Sea Level Variations in the Northwest Mediterranean During Roman Times’, Science 194 (1976), 519–21CrossRefGoogle ScholarPubMed, shows a rise of only 50 cm. since A.D. 0. Compare these with the figures of 2·0–2·5 m. since Etruscan times for Populonia, op. cit. (in note 22), 288, and 1 m. since Roman times for Cosa, op. cit. (in note 25), 396, both of which are north of Porto Clementino. Archaeological evidence at Pyrgi c. 25 km. to the south suggests a change of only 0·50–1·0 m. since Roman times, cf. Oleson, op. cit., 304.

⁴² McCann, op. cit. (in note 25); see also Will, E. L., ‘The Sestius Amphoras: A Reappraisal’, JFA 6, 3 (1979), 339–50Google Scholar.

⁴³ I am grateful to Dr. E. L. Will for help in dating these sherds.

⁴⁴ Moretti, M., La Tomba della nave (Milan, 1961)Google Scholar, and Casson, L., ‘The Earliest Two-masted Ship’, Archaeology 16 (1963), 108–11Google Scholar.

⁴⁵ Cf. Populonia, McCann, op. cit. (in notes 22 and 23).

⁴⁶ Boëthius, A. and Ward-Perkins, J. B., Etruscan and Roman Architecture (Baltimore, 1970), 80, 91Google Scholar. For the most famous examples, which come from Veii, see Ward-Perkins, J. B., Veii. The Historical Topography of the Ancient City, PBSR 29 (1961), 47–51Google Scholar.

⁴⁷ A recent study of interest here, cf. B. Frau, op. cit. (in note 41), which came to my attention as this article was being completed, describes what the author believes are Etruscan harbour installations buried several hundred metres inland at Porto Clementino. Unfortunately datable evidence for his conclusions has yet to be revealed.

⁴⁸ For historical references to Etruscan maritime activity and naval achievements see Diodorus Siculus V. 9, v. 19, V. 40 and xi. 51; Dionysius of Halicarnassus i. 11. 1, 1. 25. 1; Strabo v. 2. 2, v. 3, 5, vi. 2. 10, x. 4. 9; Herodotus i. 166–7, vi. 17; Thucydides vi. 104 and Pliny, , NH vii. 209Google Scholar.

⁴⁹ E. L. Will, personal communication. All comments by Dr. Will incorporated into this catalogue are based upon her kind examination of the drawings and photographs.

⁵⁰ A parallel with Punic amphora types was suggested by E. L. Will. Cf. Cintas, P., Ceramique Punique (Tunis, 1950)Google Scholar, pls. 16 and 17, but these differ slightly in shape and are of a different texture and colour.

⁵¹ E. L. Will, personal communication.

⁵² For Kyrenia ship see Swiney, H. W. and Katzev, M. L., ‘The Kyrenia Shipwreck: A fourth-century B.C. Greek Merchant Ship’, Proceedings of the Twenty-Third Symposium of the Colston Research Society (London, 1973), 349Google Scholar. For Brown's Ferry see Albright, A. B. and Steffy, J. Richard, ‘The Brown's Ferry vessel, South Carolina. Preliminary report’, IJNA 8, 2 (1979), 130–2CrossRefGoogle Scholar.

⁵³ All Mediterranean ships so far excavated dating between the fifth century B.C. and the fourth century A.D. have had their hull planks edge-joined with pegs locking mortise-and-tenon joints. Hull', , IJNA 5, 2 (1976), 122–3Google Scholar.

See F. H. van Doorninck, Jr., ‘The 4th Century Wreck at Yassi Ada. An Interim Report on the the exposed and fragmented state of this hull left little room for interpretation of the ship. However, approximate measurements and sketches of the wood were made.

The two sections of the keel lay several metres apart, but appeared to be from the same ship. They were approximately 2·13 m. long and were sided 12·70 cm. and moulded 20·32 cm. The keel was rabbeted, with notches in the top of varying sizes, averaging 21 cm. wide by 2·54 cm. high, and spaced c. 17·78 cm. apart. One piece of the keel had no extant frames, while on the other, remains of eight floor timbers were found in situ on top of the keel extending to the east side. Measurements recorded for the frames were 5 m. long, sided c. 45 cm. and moulded 33 cm. Although these pieces were observed on the bottom as eight individual pieces, it is possible that they actually represented four frames, each consisting of an adjoining frame and futtock. In either case, the size and spacing of the framing indicates heavy construction and contradicts the lightness of the keel, which is relatively small. Perhaps a heavy keelson made up for the minimal structural contribution of this keel, but unfortunately no evidence for a keelson was found. No cargo was found from this ship either, but we were told by local Italians that Venetian pottery has been brought up from the ship in the past.