Marsala (Punic ship) and Sister Ship, mid-3rd century BC
Country: Italy
Place: Marsala, western Sicily, near the island of Mozia, at a depth of 6 m
Coordinates (approximate): 37.86, 12.45
Type: Punic warship (navis longa)
Identified: No
Dated: Mid-3rd century BC; related to the First Punic War.
Beam: Estimated at 4.80 m.
Keel Length: Unknown.
Length Overall: Estimated at about 35 m.
Number of Masts: Unknown; no mast step was found.
History of the wreck
The town of Marsala lies on the west coast of Sicily, a short distance from the island of Mozia. In 1969 harbour works (dredging) brought to light a large quantity of potsherds of Punic date, as well as timber structures related to shipwrecks. Frost carried out archaeological excavations at the site between 1971 and 1974, documenting two practically identical wrecks, the Marsala and the Sister Ship,1 of military type, sunk at a depth of 6 m and 70 m apart (Frost 1972, 1976, 1978, 1993). They are two sinkings of the mid-3rd century BC, related to the First Punic War. The proximity of two Punic wrecks of identical characteristics led Frost to assume that they were part of the same military action.
The interest of the construction of these military wrecks lies in knowing their general building characteristics, since they are among the few archaeological examples of their period, together with the Tour Fondue wreck, while the Marsala and the Sister Ship are the only examples of warships known to date. In the author’s opinion, their attribution to state shipbuilding is of great importance, since their cross-section (and therefore their lines) and the techniques observed in their assembly may bring together the best of Punic shipbuilding knowledge of the time.
All this must have had consequences for the speed with which this type of vessel could be built (Basch 1998: 35), as Polybius describes (I, 20.13) when he tells how in 260 BC the Romans were able to build very rapidly replicas of a Carthaginian ship captured in the First Punic War.2 This may be the fleet to which Pliny refers when he says that one was built in 60 days (H.N. 16.192), something repeated in 242 BC when Rome built 200 quinqueremes, all copied from a Carthaginian ship whose qualities were much admired (Polybius I, 59.8).
Description of the site
The two wrecks lie at a depth of 6 m, 70 m apart. Of the first, the Marsala, the keel and one side (10 × 3 m) are preserved (Figure 1).
Figure 1. The Marsala, with the sternpost at the top of the figure (after Frost 1976).
Cargo
There was no cargo associated with these ships (only a small number of amphorae), apart from equipment belonging to the crew, animal bones related to food, and stone ballast.
Ballast
Stone ballast.
Anchors
Not reported.
Guns
Not applicable. A possible ram, shaped like an elephant’s tusk (Pinus pinaster, with bronze impressions of the rostrum), was found on the twin wreck, the Sister Ship (Figure 2).3
Iron concretions
Not reported. Bronze nails fastened the lead sheathing.
Hull Remains
The Marsala shows classic construction with punicana coagmenta (Figure 1). Its length was estimated at about 35 m and its beam at 4.80 m, so its proportions could fit perfectly with those of a navis longa or galley. There is no evidence of oars, which may have floated away during the sinking. The absence of a mast step and of cargo (only a small number of amphorae), the presence of loose pieces for supporting a deck, and a possible ram found on the Sister Ship suggest an interpretation as warships, questioned only by McGrail (2001: 153).
Figure 2. Structure of the ram, based on measurements taken underwater on the Sister Ship (after Frost 1976: 267).
According to a series of marks (rings) painted on planking strakes nos. 10 and 11 and on the diagonal (s-shaped) scarfs,4 it could be concluded that groups of planks had been joined together before being fitted to the hull. This is one of the most interesting questions in the construction of the Punic ships of Marsala, since the study of the letters and marks in the Phoenician-Punic alphabet on the keel and strakes (painted or incised), made by at least five individuals (Frost 1993), showed that the ship had been built “in series”, surely from “templates”, a question related to state shipbuilding (Basch 1998: 22), to the organisation of work in the shipyards and to the military organisation of harbour spaces. The length, width, thickness and position of the mortises for the keel and each of the planking strakes were predefined, and the author believes that, in the same way that work was organised to produce the backbone and hull in series, the same may have been done for the frames. Only in this way could the desired lines be obtained, with a large lateral plane created by the prominent keel and the high length-to-beam ratio.
The ship was begun with the keel and built up asymmetrically: first the starboard side up to the eleventh strake, in groups of planks previously joined together, then the other side, and finally the frames were fitted in positions predetermined by marks on the keel.
Keel, keelson and mast step
The keel (Pinus sylvestris) has notched rabbets with a marked recess for the joint with the garboard strakes (Figure 3). The joint with the sternpost (Acer sp.) was made by a trait de Jupiter (rayo de Júpiter) scarf in which the locking key entered through the side face (Figure 3). No mast step was found.
Figure 3. Detail of the plan, with a side view of the keel at the trait de Jupiter scarf joining it to the sternpost and a section showing the notched rabbet for the garboard strake (after Frost 1976: fig. 10).
Frames
In the central part of the hull the framing shows an alternation of floor timbers and half-frames, with an average room-and-space (malla) of 25 cm. The floor timbers (Quercus sp.) were fitted after the eleventh strake had been assembled. Deliberately, none of them is in direct contact with the keel; a space is always left, perhaps for the circulation of bilge water, since limber holes are recorded in only some floor timbers (Figure 5). Where timber suitable for making the piece in one was not available, the lower heel of the floor timber was joined to it by a large tongue.
The frames were fastened to the hull by nails that used wooden treenails as guides or plugs, the points being bent over, always towards the keel. A fixing putty (non-resinous) was recorded along the edges where they met the hull, the only known example of its use.
In the wreck, two longitudinal pieces of timber 51.5 cm long were found with a technique of internal lashings for fastening them to the hull (Figure 4). This consists of a series of lashings passing through the pieces which, once tied, are locked by pegs, as documented in the frames of a group of wrecks known mainly between the French coast and northern Catalonia: Tour Fondue (c. 250–225 BC) (Dangréaux 1996: 52; Dangréaux 1997: 90); Roche Fouras (150–100 BC) (Liou 1975: 597; Pomey 1996: 57); Port la Nautique (2nd century–70 BC) (Falguera and Jézégou 2000: 42–44); Cavalière (c. 100 BC) (Charlin et al. 1978: 89; Pomey 1996: 56–57); Dramont C (second quarter of the 1st century BC) (Pomey et al. 1992: 42–44; Joncheray 1994: 24, 28); Plane 1 (c. 50 BC) (Liou and Pomey 1985: 556–557; Pomey 1995: 57); Cap Béar 3 (third quarter of the 1st century BC) (Pomey et al. 1987–1988: 33); Cap del Vol (late 1st century BC–early 1st century AD) (Nieto and Foerster 1980: 163–177; Foerster 1980: 250; Nieto 1982: 165–168; Liou 1987: 272); Barthélémy B (second quarter of the 1st century AD) (Lopez 1996: 59; Dangréaux 1997: 90); Saintes Maries 24 (AD 40–75) (Long et al. 2004: 68–71; Marlier 2005: 340); Saintes Maries 2 (third quarter of the 1st century AD) (Long 2003: 53; Long et al. 2004: 71; Marlier 2005: 328); Baie de l’Amitié (second half of the 1st century AD) (Jézégou 2001: 400–401; Wicha 2002: 43–44); Perduto 1 (1st century AD) (Sciallano and Liou 1985: 145); and Port Vendres 3 (mid-2nd century AD) (Jézégou 2007: 41).
The Punic ship of Marsala has a fastening of this type of lashing, as does the repair of a frame in the Jeaume-Garde B (late 2nd century BC) (Pomey 1993: 44). However, the earliest evidence known to date is not in a Punic or Italian cultural context, but dates to the end of the 6th century BC in a Massaliote context, in the Jules Verne 9, where a piece of wood without structural connection to the hull was found with the classic traces of this type of internal lashing along the longitudinal axis of the piece (Marlier 2005: 171). The simplicity of this technique allows a theory of its general spread across the western Mediterranean.
Figure 4. Views of a fragment of timber 51.5 cm long showing the lashings, in groups of two holes (after Frost 1976: 261).
Figure 5. Drawings of the side faces/sections of the floor timbers, showing angular limber holes in nos. 15 and 17 and an added heel in nos. 19 and 21 (after Frost 1976: fig. 156).
Planking
The garboard strakes were joined to the keel by mortises, tenons and pegs (all of Quercus sp.) (Figure 6), together with widely spaced nails. They were not merely sawn but practically sculpted, like the second strakes, thus creating the angles needed to obtain a pinched (wine-glass-shaped) hull section, rather than by longitudinally twisting planks of rectangular section. According to Frost, this would be related to the characteristics of a warship with a ram, which required a strong longitudinal structure.
The planking strakes, of pine (Pinus sylvestris and Pinus nigra), were 3 cm thick and, although laid edge to edge, they are not all of rectangular section, as is usual in ancient ship construction: a group of them are slightly trapezoidal at the waterline and at the ends, which gave the hull the visual appearance of clinker construction, perhaps an Atlantic influence from northern Europe (Frost 1999), also documented in some other wrecks such as Chrétienne A (Pomey 1997: 103). Its function would have been to prevent the hull from throwing up too much water when under way, so as to keep the deck as dry as possible (Frost 1976: 247). For this reason a wale was placed just above these strakes (Figure 7).
Figure 6. Three examples of tenons, the first belonging to a scarf joint, hence its third hole. The smaller tenons are found in the bottom of the ship, hence the difference in size between the second and third (after Frost 1976: 90).
Ceiling, thick stuff, and wales
A wale was placed just above the trapezoidal strakes at the waterline (Figure 7). Loose pieces for supporting a deck were found.
Figure 7. Wales without connection to the remains. On the right of the figure (A) is the theoretical section of the spray deflectors placed under the wale to protect the deck from spray (after Frost 1976: 256).
Caulking
The whole ship below the waterline was sheathed with lead sheets fastened by bronze nails; notably, a fabric impregnated with pine resin was placed between the lead and the wood to improve the protection of the hull.
Size and scantlings
The ship’s length was estimated at about 35 m and its beam at 4.80 m, proportions matching those of a navis longa or galley, with a prominent keel and a high length-to-beam ratio giving a large lateral plane. The planking is 3 cm thick and the room-and-space 25 cm.
Table 1. Scantling of the timber remains of the Marsala Shipwreck
| Timber | Sided [cm] | Molded [cm] |
| Keel | – | – |
| Keelson | – | – |
| Maststep Mortise | – – | – – |
| Buttresses | – | – |
| Floor timbers | – | – |
| Futtocks | – | – |
| Room-and-space At central frames | 25 (avg.) | |
| Planking | – | 3 |
| Ceiling | – | – |
| Main mast Ø Chock | – – | – – |
No other timber dimensions are reported.
Wood
Keel: Scots pine (Pinus sylvestris). Sternpost: maple (Acer sp.). Planking: Scots pine (Pinus sylvestris) and black pine (Pinus nigra). Floor timbers, tenons and pegs: oak (Quercus sp.). Ram (Sister Ship): maritime pine (Pinus pinaster).
Bibliography
Basch, L. 1998. Construction privée, construction d’État. In E. Rieth (ed.), Concevoir et construire les navires. De la trière au picoteux. Technologies, Idéologies, Pratiques 13.1: 15–48. Ramonville Saint-Agne: Érès.
Frost, H. 1972. The discovery of a Punic ship. The International Journal of Nautical Archaeology 1: 113–117.
Frost, H. 1978. Seul bateau de guerre connu de l’Antiquité méditerranéenne: le navire punique de Marsala. Les Dossiers de l’Archéologie 29: 53–61.
Frost, H. 1993. Notion de construction. Les marques peintes de l’épave punique de Marsala. Les Dossiers d’Archéologie 183: 52–57.
Frost, H. 1999. Simulated clinkers in the third century BC Mediterranean. Archaeonautica 14: 161–163.
Frost, H. et al. 1981. Lilybaeum (Marsala). The Punic Ship: Final Excavation Report. Notizie degli Scavi di Antichità, suppl. 30 (1976). Rome: Accademia Nazionale dei Lincei.
McGrail, S. 2001. Boats of the World: From the Stone Age to Medieval Times. Oxford: Oxford University Press.
Pomey, P. 1997. Dendrochronologie et dendromorphologie des épaves antiques de Méditerranée. Bilan scientifique du DRASSM 1996, 102–103. Paris.
[Further references cited in the list of lashed-frame wrecks to be completed from the source bibliography.]
Notes
1 The Marsala was the one excavated, being the larger; the second wreck, in which only a test trench was dug at the bow, was named the Sister Ship.
2 Just as the history of Rome goes back to the 7th century BC with the reign of Tarquinius Priscus, Roman shipbuilding has an ancient origin; it did not begin with the First Punic War (264–241 BC). It could be said that military and state shipbuilding was updated through the “espionage and copying” of the enemy’s weapon. In the 3rd century BC Polybius (III, 22–26) tells how by the end of the 6th century BC Roman merchant ships had already extended their own routes to Sicily and North Africa. The construction of these ships, unknown for the time being, was presumably similar to Etruscan construction.
3 Although there are doubts as to whether this is really an armed ram rather than a cutwater (cf. McGrail 2001: 153), in the author’s opinion the type of fastening to the hull of the central piece, which has not survived, and the bronze impressions of the rostrum support the hypothesis of a ram.
4 The scarfs joining the planking strakes on the wreck fall into two groups: bevelled ones for the joints below the waterline (the bottom) and diagonal (pico de flauta) ones for those above it. In fact, examples of wrecks with diagonal scarfs are very few: Jules Verne 7, Ma’agan Mikhael, Binissafúller, Marsala and Chrétienne C.