Bou Ferrer, 1st century AD
Country: Spain
Place: Off the beaches of La Vila Joiosa (Alicante), at a depth of 25 m
Coordinates (approximate): 38.50, -0.22
Type: Large Roman merchant ship (amphora carrier)
Identified: No
Dated: Mid-1st century AD, by comparison of the amphorae with those of the Sud Lavezzi 2 and Lavezzi 1 wrecks (c. AD 30) and the Tour Sainte Marie wreck (mid-1st century AD).
Beam: Unknown.
Keel Length: Unknown.
Length Overall: Estimated at 27–30 m.
Number of Masts: Unknown.
History of the wreck
The discovery of the Roman wreck Bou Ferrer was reported to the authorities more than a decade ago by two sport divers whose surnames gave the underwater site its name (Juan Fuertes et al. 2008: 270). The type of seabed where it lies was unattractive for diving (cold, murky water without fauna), which preserved the wreck intact until it was located. A review of the short history of European underwater archaeology shows that all well-preserved and easily accessible wrecks suffered, to a greater or lesser degree, from divers who looted their cargoes, to the point that it is now unlikely that an intact wreck like the Bou Ferrer will be found. Unfortunately, the looting of this wreck began too, first of a few amphorae and then in a systematic and organised way. Drawing on previous experience, the heritage authority of the Comunidad Valenciana set up a project to protect the site from looting, installing a structure over the cargo that met the following criteria: protection, reversibility, no impact on the site and easy installation at a depth of 25 m1 (Juan Fuertes et al. 2008: 271–272). After the work carried out within the European ANSER project (Anciennes Routes Maritimes Méditerranéennes, 2003 and 2004), the lines of research on the site were defined (Juan Fuertes et al. 2011).
The Dirección General de Patrimonio Cultural Valenciano y Museos de la Generalitat Valenciana began the project of excavation, research, museum presentation and public outreach of the Bou Ferrer site with a first underwater excavation campaign in the summer of 2006, directed by Cibecchini and De Juan. This consisted of a 3 × 3 m test trench in a central part of the site, from which the data could be extrapolated (Figure 2). The second campaign, in 2007, covered an area of 100 m², which included the 2006 test trench, so as to have one sector more advanced than the rest of the excavation (Figure 1). In September 2012 a further campaign, co-directed by De Juan and Cibecchini, was carried out to obtain the first data on the construction of the ship.
As to the possible port of origin of the ship, the area of Cádiz is the most plausible hypothesis: 55% of the production of the main amphorae for transporting fish sauces in the Imperial period came from there (Etienne and Mayet 2005: 45); some fifty of the workshops identified in this area made amphorae of the Dressel 7-11 family, four of them specifically Dressel 11; and in numerous centres an earlier production of Late Punic amphorae,2 such as the Mañá C2, is attested. It cannot be ruled out that much of the cargo had different origins, owing to the storage of products in a main port, the only kind that could shelter large ships and stow them; from this point of view the port of Gades seems to meet all the requirements.
The discovery by DRASSM of some twenty wrecks in the Strait of Bonifacio, with large cargoes from Baetica (Liou 1990 and 2000), suggests that the most likely destination of these ships was Rome. Large ships with Baetican cargoes (e.g. Arles IV, Long 1998) have also been found near the mouth of the Rhône, so this possible destination should not be dismissed. The most logical route would have run from the area of the Strait of Gibraltar, keeping slightly to the coast of North Africa or southern Iberia, then leaving the coast to head across the open sea towards the Balearics, and then towards Bonifacio, losing sight of land for much of the voyage. According to Pliny (H.N. 9, 3–4), the voyage from Gades to the port of Ostia took seven days, crossing the western Mediterranean through the Ibericum, Balearicum, Sardoum and Tyrrhenum seas (probably as a record performance). To reconcile distance, speed and time, open-sea navigation is necessary, so this route would have passed far from the place where the Bou Ferrer sank. Could a ship drift, for instance 50–70 nautical miles, from the Balearics to La Vila Joiosa because of a technical problem or a storm? The probability is very high.
The presence of Dressel 11 amphorae identical to those of the Bou Ferrer among the archaeological material now stored and displayed in the forum of Pompeii prompts reflection on the final destination of these amphorae, which contained an exquisite product, fish sauce, perhaps related to the Latin garum described in Apicius’ recipes in De re coquinaria. This sauce was intended for the wealthiest and not for the plebs. The presence of the elite of towns like Pompeii, together with the many sumptuous villas around Neapolis in the 1st century AD, such as that of Poppaea at Oplontis (Torre Annunziata) or those of Arianna and San Marco (Stabiae), buried by Vesuvius in AD 79 together with Pompeii, may well explain the presence of these amphorae in a quantity that places them second at Pompeii after the Campanian Dressel 2-4 wine amphorae. This does not mean that the final destination of the Bou Ferrer was unequivocally the port of Naples, since the product could have been redistributed from the port of Ostia itself; but the presence of these containers at Pompeii, the proximity of numerous large villas of the Roman political elite, and the fact that refined fish sauces were not a product for the lower classes of Rome, as cheap wine or wheat might be, could be sufficient reasons to consider this possibility.
Description of the site
The wreck lies at a depth of 25 m off the beaches of La Vila Joiosa. At the time of its discovery its surface level measured 24 × 8 m. The 2007 excavation covered 100 m² of the site (Figure 1); the area of the 2006 test trench was re-excavated in 2012, exposing about 9 m² of the hull (Figure 4).
As regards Teredo navalis, everything suggests that, since the wreck is heeled over to one side (cross-section of frame C-123, Figure 5), all the timber above a certain level has been eaten by wood borers. On the other hand, this suggests that on the side not yet investigated, east of the keel (Q), the state of preservation will be exceptional.
Figure 1. Plan of the 2007 excavation. The colours indicate the stowage layers of the cargo: beige, layer 2; green, layer 3; brown, layer 4. The 2006 test trench is in the north-west corner (drawing: C. de Juan Fuertes).
Cargo
The 2007 campaign allowed the recording and recovery of all the amphorae of layer no. 2 in the area excavated, exposing the necks of the amphorae of layer no. 3. To clarify the sequence of layers of the ship’s cargo, work resumed in the 2006 test trench, where two layers of amphorae could be clearly distinguished in situ, with a lower one still to be excavated. Three lead ingots were recorded, aligned parallel to the ship’s centreline, lying on planking.
A total of 274 amphorae were recovered. So far, all of them belong to the Dressel 7-11 family, although four main morphological types have been distinguished (Cibecchini et al. 2011). No archaeometric study has yet been carried out, but basic observations and descriptions of the fabrics have been made, and four different fabrics were distinguished on the basis of colour, texture and type of inclusions. The working hypothesis, within the limits of the research done so far, is that the amphorae made with these fabrics are different productions from the same area or production centre.
The four morphological types of amphorae on the wreck have been called BF1, BF2, BF3 and BF4, the first two being the most common in the excavated area (Figure 3). So far, of a total of 248 amphorae, 58 have been identified as type BF1, 136 as type BF2, and 54 as “hybrids” or difficult to distinguish. The two main amphora types of the Bou Ferrer, BF1 and BF2, have their best parallels in the Sud Lavezzi 2 and Tour Sainte Marie wrecks. The few amphorae of type BF4 (= Dressel 9) found so far also have parallels in the Sud Lavezzi 2 and in the Lavezzi 1, both wrecked in the Strait of Bonifacio. While these two wrecks are well dated to around AD 30, the Tour Sainte Marie has a much broader date, in the mid-1st century, because of the presence of Beltrán IIB amphorae in its cargo, so far not attested on the Bou Ferrer.
Analyses carried out by Piques (Lattes, CNRS-UMR 5140) on the contents of the Bou Ferrer amphorae show mainly anchovy (Engraulis encrasicolus), mackerel (Scomber sp.) and horse mackerel (Trachurus sp.). Although the species cannot yet be quantified, it can be stated that the contents of the amphorae are a fish-derived product and not simply salted fish.
In 2012, four lead ingots were found aligned in position (Figure 4), and a fifth, to the north, lay upside down against a sixth ingot. The four in position were raised; some of them bear the incised marks IMP. GER AVG.
As preliminary conclusions on the cargo, after the excavation of 100 m² of the site: the cargo consisted of Dressel 11 amphorae made in workshops in the area of Cádiz; it was probably stacked in four layers; the amphorae were stowed in an ordered staggered (quincunx) grid; their contents were fine fish sauces; there was a secondary cargo of lead ingots of Domergue type I; and the total cargo is estimated at about 2,500 amphorae.
Figure 2. First layer of the amphora cargo in the area of the 2006 test trench (photo: C. de Juan Fuertes).
Figure 3. Main amphora types on the Bou Ferrer wreck (drawing: Cibecchini, DRASSM).
Ballast
Not reported.
Anchors
Not reported.
Guns
Not applicable.
Iron concretions
Not reported.
Hull Remains
For the 2012 campaign the 2006 test trench was chosen, where only the first layer of the amphora stowage remained to be removed, in order to reach the lead ingots known from the 2007 campaign and the hull of the ship. Since a long time had passed since the previous intervention, the trench was completely filled with silt and had to be re-excavated, although this was mechanical work, since all the amphorae had already been drawn and only had to be raised. After this laborious operation, of little archaeological interest, about 9 m² of the ship’s hull were exposed (Figure 4).
The excavation of all the timber within the limits of the trench then began, revealing an area to the west of the site where the frames had no preserved ceiling. To the east, by contrast, several ceiling planks were preserved (see below).
Figure 4. View of the test trench after removal of the first layer of amphorae and the beginning of the excavation of the spaces between the frames (photo: C. de Juan Fuertes).
Figure 5. Plan and sections of the ship’s hull in the “2006 test trench” (drawing: C. de Juan Fuertes).
Keel, keelson and mast step
To the east of the trench, right at the limit of the intervention, a timber of considerable width, about 50 cm, had its upper face completely eaten by Teredo navalis. The working hypothesis was that it might be the ship’s mast step, which would place the keel directly beneath it. A 10 cm wide section of it was therefore cut, in order to interpret the piece, take samples for wood identification3 and transfer the section to ARQUA for conservation. The result was entirely positive: the cut allowed the decayed timber to be identified as the ship’s mast step or keelson, and its seating system, characteristic of the Republican architectural family,4 to be understood, while also allowing a detailed observation of the keel, the garboard strakes and the floor timbers on which the mast step rested (Figures 6 and 7). The lines of the hull along the centreline are those of a flat-bottomed ship, which relates directly to the palaeo-landscape of the area where the cargo came from, as with the flat bottom of the Titan wreck.
Inside the hull there is no difference in height between the keel and the garboard strakes; they form a plane in which a slight deadrise angle of the planking seems to be visible (Figure 8).
Figure 6. North side face of the cross-section of the mast step of the Bou Ferrer. The growth rings show that it was cut from a large trunk, more than 50 cm in diameter (photo: C. de Juan Fuertes).
Figure 7. Underside of the fragment of mast step lying to the east. 1: all the edges of the heel fitting between the floor timbers were bevelled, giving it the side view of a half-moon; 2: the right angle formed between the heel fitting between the floor timbers and the underside of the mast step resting on the top of the floor timber; 3: a groove worn by friction from the plank to the east, on which rests another row of lead ingots not excavated in the 2012 campaign (photo: C. de Juan Fuertes).
Figure 8. The space between frames C-123 and C-124, showing the width of the keel, marked by two white lines, and the central limber hole of rectangular section with a rounded top (photo: C. de Juan Fuertes).
Frames
The framing appears extremely robust, but it must be remembered that the area excavated lies right on the ship’s centreline; visually, it can be compared with the Sud Perduto 2 wreck, which carried a Baetican cargo (Pomey et al. 1997: 154). The room-and-space (malla) is tight, 12 cm on average, for frames 25 cm moulded in the longitudinal section along the third plank of the hull at the west of the trench (Figures 5 and 10). Two of them were found to be consecutive floor timbers, but no pattern of arrangement with half-frames can be established. The frames are fastened to the hull with circular wooden treenails of 1.8–2 cm diameter, reinforced, judging from an area to the west where the frames had completely disappeared, by a series of iron nails driven blind (a punta perdida) from the outside of the hull, located along the central longitudinal axis of the frame and spaced about 10 cm apart. The floor timbers seen in the cut through the mast step faithfully follow the flat lines of the hull and have a central limber hole about 7 cm wide and 5 cm high (Figure 5), recalling those documented in the Ouest Giraglia 2 wreck (Figure 9).
Comparisons with the Albufereta 1 are inevitable, since the two wrecks share the same chronology, the same area of sinking (south of Cap de la Nao, in the present province of Alicante) and the same geographical origin of the goods carried (Baetica), although the two ships represent two very different forms of trade. As far as construction is concerned, the comparison can for now only concern the technique used to fasten the frames to the hull. This is not especially distinctive: wooden treenails fasten the framing to the hull, and this fastening is reinforced by blind iron nails along the frames at an average spacing of about 10 cm, well attested in the Albufereta 1 and observed in one frame of the Bou Ferrer. This has led the author to relate the two wrecks.
It is surprising that the dolia wreck of Cap Corse, the Giraglia 1 (AD 20), has an identical technique for fastening the framing. This is significant because the construction of dolia ships is specific in certain respects.5 The most likely explanation is that these ships had a particular building process owing to the fixed containers they carried, and since the dolia were made in Campania, their construction may have been carried out near the mouth of the river Garigliano at Minturnae. This is attested by the volcanic inclusions in their fabrics and the planta pedis stamps of the Piranii family, so the shipyard where these ships were built would have been on the coast of that Italian region. This argument leads the author to propose, with due caution, that the Bou Ferrer and Albufereta wrecks may have been built in a region where this frame-fastening technique was common in the customary knowledge of the fabri navales, pointing to southern Latium or Campania as a possible area of construction, on the basis of the comparison with the Giraglia 1.
Figure 9. Idealised section of the shape of frame 123 of the Bou Ferrer (drawing: C. de Juan Fuertes).
Figure 10. View to the north of frames C-120 to C-124, in the area of shallowest burial, where they have been affected by Teredo navalis. The robustness of the framing can be seen (photo: C. de Juan Fuertes).
Planking
The hull planking strakes in the area studied are 22 to 29 cm wide (Figure 5); there are as yet no data on their thickness. No information is available so far on the mortise-and-tenon joints between them, other than that the locking pegs have a constant diameter of 1.4 cm, and some of them had not been properly trimmed. The pairs of pegs are about 18 cm apart (Figure 11). The pegs observed in the joint between the keel and the garboard strakes, by contrast, were 2 cm in diameter.
Figure 11. Diagram of the average spacing of the locking pegs in the plank joints of the Bou Ferrer wreck (drawing: C. de Juan Fuertes).
Ceiling, thick stuff, and wales
To the east of the trench several ceiling planks were preserved, heavily damaged on their upper faces by Teredo navalis, to the point that in some of them it was impossible to determine their thickness accurately. The thickest (7 cm) flanked a plank on which lay the row of lead ingots already recorded in the previous campaign, and was fastened to the framing by some iron nails of square section, 1 cm across, the impression of the (missing) nail head in the wood reaching 4 cm in diameter. The other planks to the west of the trench had such degraded upper faces that their thickness could not be assessed. Next to the mast step, ceiling strakes (vagras) carried the aligned lead ingots, and a stringer (palmejar) 19 cm wide and 7 cm thick was recorded (Figure 12).
Figure 12. View of the south-east part of the test trench after removal of the lead ingots. 1: mast step; 2: ceiling strakes next to the mast step, on which the lead ingots were aligned; 3: stringer 19 cm wide and 7 cm thick (photo: C. de Juan Fuertes).
Caulking
Not reported.
Size and scantlings
Calculations, intended only as a guide, suggest that the length of the Bou Ferrer was between 27 m (171 t burden) and 30 m (235 t burden). With these values, a cargo estimated at between 162 and 215 t of amphorae, plus some minerals, could have been carried safely. With a length close to 30 m, the Bou Ferrer should be considered a large ship for its period; its cargo is estimated at about 2,500 amphorae.
Table 1. Scantling of the timber remains of the Bou Ferrer Shipwreck
| Timber | Sided [cm] | Molded [cm] |
| Keel | – | – |
| Keelson | – | – |
| Maststep Mortise | c. 50 – | – – |
| Buttresses | – | – |
| Floor timbers | – | 25 |
| Futtocks | – | – |
| Room-and-space At central frames | 12 (avg.) | |
| Planking | 22–29 | – |
| Ceiling Stringer | – 19 | 7 (max.) 7 |
| Main mast Ø Chock | – – | – – |
Frame moulded dimension taken in the longitudinal section along the third plank at the west of the trench. Planking: sided = width, moulded = thickness.
Wood
Not reported; samples from the mast step are awaiting analysis. The mast step was cut from a trunk more than 50 cm in diameter.
Bibliography
Cibecchini, F. et al. 2011. [Full reference to be completed.]
Etienne, R. and Mayet, F. 2005. [Full reference to be completed.]
Juan Fuertes, C. de, Cibecchini, F. and Vento, E. 2008. Intervención arqueológica subacuática en el pecio Bou Ferrer. Resultados preliminares campaña 2006. In J. Pérez and G. Berlanga (eds), Actas de las V Jornadas de Arqueología Subacuática, 269–277. Valencia.
Juan Fuertes, C. de et al. 2011. [Full reference to be completed.]
Lagóstena 1996. [Full reference to be completed.]
Liou, B. 1990. Le commerce de la Bétique au Ier siècle de notre ère. Notes sur l’épave Lavezzi 1 (Bonifacio, Corse du Sud). Archaeonautica 10: 125–155.
Liou, B. 2000. Las ánforas béticas en el mar. Les épaves de Méditerranée à cargaison d’amphores de Bétique. In Congreso Internacional Ex Baetica Amphorae. Conservas, aceite y vino de la Bética en el Imperio Romano (Sevilla–Écija 1998), vol. 1, 1061–1111. Écija.
Long, L. 1998. L’archéologie sous-marine à grande profondeur: fiction ou réalité. In G. Volpe (ed.), Archeologia subacquea. Come opera l’archeologo sott’acqua. Storie dalle acque. VIII Ciclo di Lezioni sulla Ricerca applicata in Archeologia (Certosa di Pontignano 1996). Florence: All’Insegna del Giglio.
Pomey, P. (ed.), Gianfrotta, P. A., Nieto, X. and Tchernia, A. 1997. La navigation dans l’Antiquité. Aix-en-Provence: Édisud.
Notes
1 The protection method was designed by A. Fernández, J. A. Moya and C. de Juan. The project was carried out by C. de Juan, J. A. Moya and Subatech S.L.
2 Other production centres of this type of amphora are located in the area of Málaga and in present-day Morocco (cf. Lagóstena 1996: 146–148).
3 Awaiting analysis.
4 So, for the time being, the change in the seating of the mast step to two parallel keelsons, as seen in the Saint Gervais 3, has not yet been introduced.
5 As shown by the comparative study of the Ouest Giraglia 2 and Ladispoli wrecks.