Ouest Giraglia 2, 1st century AD
Country: France
Place: West of the island of Giraglia, Cap Corse (Haute-Corse), at a depth of 35 m
Coordinates (approximate): 43.03, 9.39
Type: Roman merchant ship carrying wine in bulk (dolia ship)
Identified: No
Dated: 1st century AD, based on the Dressel 2-4 amphorae; about 50 years later than the Ladispoli wreck.
Beam: Unknown.
Keel Length: Unknown; 7.10 m preserved.
Length Overall: Unknown; estimated at no more than 20 m.
Number of Masts: One, probably stepped well forward.
History of the wreck
The Ouest Giraglia 2 wreck lies at a depth of 35 m near the island that gives it its name, off Cap Corse (Corsica). It carried wine in dolia together with Dressel 2-4 amphorae of the 1st century, and a substantial portion of the ship’s hull was preserved under the sand. It was excavated in 2010, 2011 and 2012 by DRASSM under the direction of Cibecchini, so the data on the cargo are still preliminary. The study of the ship’s construction was the responsibility of Marlier and the author, together with a large team of specialists (Figure 1).
The scientific justification for the intervention was that the Giraglia 2 wreck could help towards a better understanding of the system of maritime trade in wine shipped in bulk, mainly during the 1st century. One of the working hypotheses with which research on the site began was that the fixed containers (dolia) and the maritime trade system1 of wine in bulk might be associated with ships built specifically for this type of transport, which would open the way to the possible existence of an architectural system/signature not considered until now.
The idea of a ship construction geared to a specific cargo, very different from the usual one, is easy to justify, even without numerical data. The ratio between the weight of the container and the volume carried per m³ of hold is far more favourable for stowage in dolia than in amphorae; but, equally, the total weight (containers plus contents per m³) is not distributed in the same way over the hull. Hundreds of amphora toes, like a grid, distribute the forces evenly, whereas dolia concentrate all their weight on very specific points (e.g. 11 in the Grand Ribaud D) in the central part of the hull (Figure 2). If to this is added the argument that dolia ships would have had to sail to the wine production centres themselves, entering fluvio-maritime or lagoon waters and therefore crossing the sandbars of river mouths, spits and lagoons, their construction should have been adapted to the needs of this model of trade.
The work of excavation, recording and study of the Giraglia 2 followed established methodology (Pomey and Rieth 2005). In 2011 the central part of the hull was excavated, where the presence of a dolium had prevented recording in the previous campaign (Figure 1), as was the area to the north of it, which was excavated and studied in situ. Part of a cross-section of the hull at frame M109 was also sawn and raised, to complete the study of the construction and resolve gaps impossible to clarify otherwise. Thanks to an experimental reproduction of this portion of hull, a better understanding of the construction of the Giraglia 2 has been obtained. Although the wreck lies at an accessible depth (35 m) and in water of excellent visibility, long dives are not possible. Its location on a Mediterranean island, in an area with a complex maritime climate, makes the wreck difficult of access, so every step in its recording and study required considerable effort. For these reasons, in the 2010 campaign, after all the surviving structural elements had been labelled and identified, measurements taken and some sections drawn, the recording of the hull was based on photogrammetry.2
Nautical space and interpretation. In the author’s opinion, if this ship took part in an organised system of maritime trade, in some ways particular or different from the other forms of trade of its period, the ship too had to be adapted to the needs of that trade, as the tool or machine that it is. When the ratio between the weight of the container and the volume of the contents is considered, transport in dolia is more favourable than in any known type of amphora. However, the relatively short length of ships such as Giraglia 1, Grand Ribaud D and Ladispoli (Marlier 2008: 158) shows that the merchants were not interested in owning very large tankers for this trade, as the Madrague de Giens was in its day, to draw a loose comparison. The author proposes as a working hypothesis that the merchants’ aim was to have an agile, versatile ship able to sail in various nautical spaces, so as to reach the wine production sites or storage ports themselves, taking advantage of the favourable ratio of the dolia. The very forward position in the Giraglia 2 of a mast step, attested only by a copper bolt in the keel, led the author to suggest that this ship might be one of those shown in iconography with a spritsail (stele in the Thasos Museum or funerary relief from Thessaloniki, Basch 1987: 478), with a forward mast that could be used as a towing pole for taking the ship through deltaic, lagoon or river waters. The sum of these arguments leads to an interpretation of the construction of the Giraglia 2 as specific to the type of transport and model of trade in which it took part, reflecting in part the practices and traditions of shipbuilding in southern Latium, where the Pirani produced their dolia. Significantly, the particular features of this wreck can at present only be compared with the Ladispoli wreck, which leads the author to propose that the architectural system of these ships reflects an architectural signature.3
Figure 1. View of the hull of the Ouest Giraglia 2 at the end of the study of its construction in the 2010 campaign (photo: C. de Juan Fuertes).
Figure 2. Diagram of how the weight of the cargo is distributed over the hull with amphorae or dolia. A cargo of amphorae, here Dressel 7-11, of similar volume weighs about 1,200 kg, spread through 8 toes over 2.25 m², unlike the approximate weight of a dolium, 2,000 kg, which rests on a square only 25 cm a side (0.0625 m²).
Description of the site
The surviving fragment of the ship’s hull corresponds to the central part of the vessel, measuring in situ 7.10 m long and about 3 m wide. It preserves the keel, 5 strakes of planking to the east of it and 9 to the west, with lengths joined by simple bevelled scarfs, and 28 framing timbers in varying states of preservation, in which floor timbers, half-frames and four futtocks follow one another without rhythm (Figure 3).
Figure 3. Plan of the Ouest Giraglia 2 at the end of the 2011 campaign, with the associated longitudinal section (drawing: C. de Juan Fuertes).
Cargo
Wine shipped in bulk in dolia, together with Dressel 2-4 amphorae of the 1st century. The data on the cargo are still preliminary.
Ballast
Not reported.
Anchors
Not reported.
Guns
Not applicable.
Iron concretions
Numerous iron concretions were found next to the copper bolt at frame M118, which could not be interpreted. Lines of iron concretions on the upper faces of frames M112 to M115 relate to the nails of a stringer that has not survived.
Hull Remains
The preserved hull corresponds to the central part of the ship (Figure 3). The orientation of bow, stern, starboard and port follows from the interpretation of the copper bolt at M118 as the forward end of the mast step (see below), and corresponds practically to the cardinal points.
Keel, keelson and mast step
The keel is the longest surviving piece4 of the wreck, at 7.10 m. For the joint with the garboard strakes (by mortises and tenons) it has a simple bevelled rabbet on the upper edge of its side faces, which in the central part of the hull is so gentle (almost 180° to the horizontal) that it is imperceptible to the eye. The rabbet is about 6.5–7 cm wide. Towards the end, it changes from an angle of 88°/102° to the horizontal in the central part of the hull to practically 60°/120° at its southern end, without its width changing (Figure 11). The hull lines thus pass from very flat, with a barely prominent keel in the area that may correspond to the maximum beam, to somewhat more angular ones towards the beginning of the sternpost.5 Seen from above, the mortises were cut in a staggered pattern between the two side faces along the whole keel. Their order and spacing change along the rabbet between the central part of the hull and the end, since their placing is alternating in the centre (Figure 4) and becomes aligned at the end (Figure 5). In the central part, the average dimensions of the mortises in the keel are 8.5 cm wide, 9 cm deep and between 1 and 0.9 cm thick, with a spacing of 5 cm in the fragment observed. The tenons have bevelled edges on their narrow sides; in some cases they fit the mortise perfectly, and in others they have a little lateral play.
In the central section of the keel the locking pegs were driven obliquely, apparently from the inside of the hull outwards, although only their diameters on the upper face of the keel, about 10 mm, could be established. Likewise, in the section of hull raised to the surface the pegs in the keel were seen to be somewhat conical, but this has not been confirmed with complete certainty. At the end (Figure 5), on each side face of the keel, the mortises are cut rhythmically in a staggered pattern, spaced 6.3–6.5 cm apart. The dimensions of the mortises, practically the same as those of the tenons, are 7.5 cm average width, 6.8 to 7.8 cm depth and 0.9 to 1.1 cm thickness.6 In the end fragment of keel raised in 2009 (Figure 5), two mortises were mechanically emptied, showing in one case a trapezoidal shape with well-defined, not rounded, deep corners (Figure 7, 1). Only one half-tenon could be measured, slightly shorter than the depth of the mortise, at 6.8 cm.
The locking pegs are of truncated-cone shape, with diameters between 0.9 and 1.5 cm, spaced 13.9 to 14.5 cm apart for each joint between keel and garboard strake, and are located, seen from above, about 1.5 cm from the edge of the rabbet. One of them, which could be well observed after mechanical emptying of part of the wood, showed no knife marks and seemed to have been turned on a lathe, unlike another faceted truncated-cone peg, possibly made with a carpenter’s plane, knife or small adze, which was driven from the outside of the hull and either improved the joint or was redundant (Figures 7, 2 and 8). Apart from this exception, the pegs do not pass through the keel from side to side, so in this end section they were driven from the inside of the hull during construction. In one case, one of them closed a mistaken hole that did not reach the tenon (Figure 7, 3).
During the study of the backbone in situ in 2010, it was observed that the keel lay in a lower plane than the garboard strakes along the whole hull, as shown in one of the cross-sections drawn (M109); but the erosion of the upper faces of these pieces was so severe that their shapes were difficult to see correctly (Figure 12), and the small distance between the frames in this sector did not allow proper observation, as was later confirmed (Figure 9). This difference in height could also be seen at the end (Figure 11). In the section of hull raised in the 2011 campaign it became clear that under frame M109 the erosion of the upper faces of the keel and garboard strakes was less severe, and that the garboard strakes were about 1.5 cm above the plane of the keel, creating a kind of bilge channel, as documented in the Ladispoli (Carre 1993: 14, fig. 5), with no other known parallel at present except the unusual Nemi ships. The author relates its interpretation, an evolution of the backbone with “turned” garboard strakes (retour de galbord), more to the fitting and reinforcement of the framing than to facilitating the passage of water. After the removal of the transverse portion of hull, this position of the keel could be seen in the space between frames M109 (already removed) and M110 (Figure 10).
As a final observation on the keel, in the cut made at the southern end in 2009 a small nail appeared, driven right into the rabbet7 before the garboard strake was fitted (Figure 11), with no apparent structural purpose. Since the inner edge of the nail head coincides with that of the garboard strake, the working hypothesis is that it could be related to the marking-out and construction of the ship’s bottom and/or the fixing of guide lines, plumb lines, etc.
The species chosen for the keel is holm oak,8 a member of the Fagaceae with one of the highest known densities (robust and less flexible) and high mechanical properties, but unusual in ancient backbone construction, with only one known parallel9 in the Dramont C wreck (Guibal and Pomey 2002: 95). Moreover, Quercus ilex, with an average height of 12 to 15 m (Rival 1991: 84), does not have a trunk entirely usable for making a keel, which limits a priori the maximum length of the Giraglia 2. Without numerical data, it can still be said that it should be in line with the values given for other dolia wrecks such as Giraglia 1, Grand Ribaud D and Ladispoli (Marlier 2008: 158). Even without knowing the scarfs joining the keel to the stem and sternpost, or their size and orientation, the length should not have exceeded 20 m.
The morphology of the keel is also unusual in ancient construction, since both in its central part (20.5 cm wide and 15 cm high) and at its ends (18 cm wide and 14 cm high) it is wider than it is high, although at the end of the surviving section to the south of the site it has a somewhat polygonal, flattened shape that recalls the end of the Ladispoli keel (Carre 1993: 14, fig. 5). In its central section the very rectangular keel of the Giraglia 2 resembles in shape those of, for example, the Cap del Vol (Nieto and Foerster 1980: 171) or Perduto 1 (Foerster 1988: 179); as noted, this morphology is less frequent in ancient shipbuilding (Gianfrotta and Pomey 1981: 237), but not exceptional.
Figure 4. View of the east side face of the keel (central section) raised to the surface, showing the rabbet. The mortises were deliberately not cut along the same grain of the wood, so they are alternating, practically separated by an imaginary line touching the lower side of one and the upper side of the next (photo: C. de Juan Fuertes).
Figure 5. Section of the keel, wider than it is high, and of garboard strake B1O. Note the staggered arrangement of the mortises on the upper face (drawing: Poveda, CCJ-CNRS, Aix-en-Provence).
Figure 6. View of the southern cut at the joint between the keel and garboard strake B1E. There is an apparent absence of rabbet, but the angle may be so gentle as to be difficult to observe. The tenon in the keel mortise has an angular end, showing its symmetrical bevelled finish. Both upper faces are eroded by the sea. Only under the frame can a difference in height between these two pieces be seen (photo: C. de Juan Fuertes).
Figure 7. Top view of one of the two fragments of the keel (southern end, stern) (west side face at top). 1: trapezoidal shape of the mortises, with well-defined internal corners; 2: corrective or redundant peg; 3: mortise that a second peg fails to reach (compare with Figure 5, pegs at upper left of the keel) (photo: Seguin).
Figure 8. View of the west side face where, after mechanical emptying, the making of the two pegs can be seen. The one inserted from the outside of the hull (arrow) is faceted, made with a carpenter’s plane. In the one inserted from the inside of the hull, which does not pass right through the keel, no such marks could be seen (photo: Seguin).
Figure 9. View of garboard strake B1O, the keel and frame M109, reassembled by hand on the surface. Note the heavy erosion of the upper and side faces of B1O and the keel, and of the underside of M109 (photo: C. de Juan Fuertes).
Figure 10. View of the northern cut after removal of the portion of hull. Despite the damage, the shape of the keel, the garboard strakes and frame M110 can be seen. The frame has a central heel in which a limber hole has been cut, and the garboard strakes, with their upper edges next to the keel slightly bevelled, lie in a higher plane, creating a kind of channel (photo: Seguin).
Figure 11. View of the cut/section at the southern end of the keel, showing the nail driven into the keel rabbet (1) and how the garboard strake lies in a higher plane than the keel, as can be seen from the bevel at its upper left corner (2). The deformation of the right-hand tenon shows that the garboard strake now lies at a lower angle than originally (photo: Seguin).
Figure 12. Part of the cross-section at frame M109 (south face) drawn in 2010, showing some of the errors mentioned in the text. On the one hand, the shape of the inner heel of the frame (marked by an ellipse) was not shown correctly, and the transfer of the shapes of the keel, rabbets (red arrow) and garboard strakes (1) from the end to the central part was completely wrong. On the other hand (2), the difference in level between the keel and the adjacent garboard strakes, and the rise of these planks (5°), were shown correctly.
No notch for the mast step exists in the framing of the surviving portion of hull, but at frame M118, over the keel, there is a copper bolt 1.5 cm in diameter and 27 cm high, finished at the top by a square rove 4.3 cm a side, and therefore placed much higher than the frames (Figure 13). Its most probable interpretation is that it fastened a keelson/mast step (not preserved) to floor timber M118 and the keel, improving the longitudinal strength of the assembly. The author believes it was a bolt placed at the forward end of this piece, so that the foot of the mast would have been very far forward, which suits a cargo in dolia requiring space for filling and emptying the wine. The severe erosion of the frames, caused by an east-west underwater current,10 and the poor state or absence of frames in the northern area allow no further information to be obtained. The very forward position of the mast foot opens the possibility that the Giraglia 2 sailed with a spritsail, and a mast very far forward could also serve as a towing pole in river waters.
Figure 13. Cross-sections at frames M109 and M118. The first shows a curious difference in height between the asymmetrical floor timber (M109) and its futtock (AL109). The second shows the difference in height between the rove of the bolt and the frame, suggesting the presence of a mast step or keelson (drawing: C. de Juan Fuertes).
Frames
The framing consisted of 28 pieces of basically rectangular-trapezoidal section11 (Figure 3, longitudinal section), many of them poorly preserved because of Teredo navalis (Figures 14 and 15), while of others only some treenails and impressions on the hull remain. Although three half-frames were identified, the framing does not show the classic alternation, but rather a widespread succession of flat floor timbers, with 18 identified.12 Of the half-frames recorded, M105 has its two arms in direct contact with each other over the keel, unlike half-frame M111, where there is a slight gap of about 3 cm. The four futtocks recorded are on the west side of the hull, and their placing, like that of their floor timbers (off-centre or asymmetrical), does not seem to follow any clear pattern, although the position of AL107 and AL121, which begin at strake B5O, suggests that there were probably other futtocks (impossible to record on the wreck on the east side) near strake B5E. Surely the availability of timber of the maximum lengths these single-piece timbers could reach determined the final arrangement of the framing. Another interpretation is that the building process of this ship, with active frames as explained below, may be related to the position of the futtocks. Between floor timbers and futtocks there is a slight gap of 2 to 4 cm, with no fastening between these two pieces, which have straight-cut ends.
The longest surviving piece is 267 cm long (M108), and the frames are on average about 10 cm sided, with a moulded dimension13 over the keel between 11 and 22 cm (average 16.8 cm). At strake B4O they average 17.75 cm moulded. These measurements must be treated with caution, since all the framing is heavily eroded by an underwater current that has removed, in most of them, the sharp edges of the sawn timber, even reducing their height by several centimetres, as could be seen from iron concretions on the upper faces of M112, 113 and 114. M113 increases from 12 cm moulded nearest the keel (eroded) to 20 cm at the point where both the dolium (Figure 1) and an iron concretion had kept the wood very hard (Figure 14). It is most likely that the moulded dimension of the framing was originally very uniform, with its maximum values over the keel. The shapes of the flat floor timbers in the south-west area of the wreck suggest that they lose some of their height from the keel towards the side, perhaps to compensate slightly for the shape of the hull and so create a more level hold floor.
As regards the morphology of the frames, the area over the keel is notable. As noted, the keel lies in a lower plane than the garboard strakes, and despite the erosion (Figure 15) almost all the frames were seen to have a heel on their underside (2 cm deep on average) to rest directly on the “sunken” keel. There are few known parallels, partly the Roche Fouras wreck (Joncheray and Rochier 1976: 178) and Titan (Benoit 1958: 13), more clearly Ladispoli (Carre 1993: 22–23) and also the unusual palace ships of Nemi (Ucelli 1950).
The frames have a main central limber hole, rectangular with a slightly rounded top, right over the keel (Figure 15) (7.5–8 cm high), and small secondary limber holes of triangular section (2–3 cm high) towards the sides of the ship,14 such as that of futtock AL109, between strakes B6O and B7O, 1.20 m from the keel15 (Figure 16). On the keel, under frame M109, as at other points of the backbone (Figure 17), a resinous fixing putty, possibly pitch, was found; the working hypothesis relates its presence to the building process, since there is no mechanical role for it in the hull structure other than levelling framing timbers and correcting sawing imperfections before their final fastening to the hull (Figure 18).
The room-and-space (malla) is 14 cm on average (between 11 and 22 cm), a value comparable to other wrecks of the period (e.g. Albufereta I, AD 60; Juan Fuertes 2008: table 6), but the moulded dimension of the framing, above all in the south-west part of the hull16 where there is less damage from Teredo navalis and erosion, gives this framing a very tight and robust appearance.
After the systematic review of the framing carried out in the 2011 campaign, and of the photographic record of 2010, it can be concluded that the treenails fastening the frames on the east side of the keel (starboard) are systematically in the south side face, and on the west side (port) in the north side face of the frames. Their density is high, with 2 to 4 treenails per planking strake. The order and rhythm of the treenails driven at an angle through the side faces lead the author to suggest that the holes were drilled from the inside of the hull,17 neatly arranged in the side faces and with precise auger angles so that they emerged centred on the underside of the frame (Figure 19). This does not mean that the very slightly truncated-cone treenails could not have been driven from the outside of the hull, but there are no data to say whether this was done from outside or inside.
Interpreting this particular placing of the treenails in the side faces is not simple. Several pinched-bottom wrecks of Hellenistic-Republican date have floor timbers with treenails inserted through their side faces (Marsala, Frost 1976: fig. 156; Cap Gros, Joncheray 1989: 67; Roche Fouras, Joncheray and Rochier 1976: 178; Madrague de Giens, Tchernia et al. 1978: 77), normally in the tallest area near the keel, justified by economy of effort, since mechanically it is not necessary for the treenails to pass vertically through the great height of the floor timbers. In the case of the Giraglia 2, the first point to note is that the wood species recorded for the framing are very homogeneous, with only oak (Quercus sp.) and elm (Ulmus sp.) present, in practically equal proportions, two species of high mechanical properties and hardness. This rather contradicts a direct comparison with most ancient shell-first ships, in which the framing, with a secondary role, is very heterogeneous (Guibal and Pomey 2002: 197). Oak and elm are extremely hard to drill, which, together with the moulded dimension of the framing of nearly 20 cm, makes it logical to drill the holes at an angle to save effort. But why are the holes drilled systematically in one side face or the other according to the side of the ship? Significantly, the only other wreck in which the author has documented this practice, though never as profusely as in the Giraglia 2, is again the Ladispoli (Carre 1993: 22), wrecks belonging to the same system of maritime trade, about 50 years apart.18
Building process. The portion of hull raised to the surface, its dismantling and the making of a model for experimentation allowed some conclusions on the building process of the Giraglia 2. The ship has the construction of its time: an architectural system characterised by a longitudinal, shell-first conception, in which flush-laid strakes are joined to the backbone by mortises and tenons locked by pegs, the framing being a reinforcement of the hull structure. However, some details of its building sequence deserve closer analysis. The ship was begun with the keel, to which the stem and sternpost were presumably joined by trait de Jupiter scarfs with the locking key inserted through the side face, a practice with many known examples. Then the garboard strakes were added, creating a robust backbone of holm oak and oak. From the dismantling of the portion of hull, the second strake on the starboard side would then have been added, and it is here that the active frame19 M109 comes into play, the earliest known case. From the maximum and minimum diameters of the truncated-cone pegs locking the tenons, it can be deduced whether they were driven from the inside or the outside of the hull. If a frame (floor timber, half-frame, futtock, etc.) was fitted before the planking strakes to which it was to be joined, any tenon joining strakes that coincides with the space occupied by the frame must necessarily be drilled and pegged from the outside. This procedure was masterfully analysed in the La Bourse wreck (Gassend 1985: 119), although the author does not share, at least for now, the interpretation that the participation of these active frames in the building process alters the longitudinal, shell-first conception of the period. In the case of M109, its role as an active frame shows an even more visible detail than the presence of pegs with their minimum diameters on the inner face of the hull: when the holes were drilled, they reached the frame itself, as did the pegs, which were left embedded in it.20 Since both pegs were driven from the outside of the hull, the position of the tenons coinciding with the active frames must logically have been scribed, though the erosion of the planks does not allow this to be seen. In the frame these peg holes were detected at the seam between B1O and B2O on the port side and at the pegs joining B3E and B4E, while M109 is a floor timber that could be described as off-centre or asymmetrical. This suggests that there must have been other active frames in the building of the Giraglia 2. After the second strake on the starboard side, M109 would have been fitted. Its role is to help construction, an advance in the building process, marking the lines of the ship before it is assembled, but the structure still lies in the hull, the design principle remaining unchanged, so there is no evolution towards a transverse conception of the ship.
To fasten it to the hull, a resin was applied as a fixing or levelling putty at the base of the two central heels separated by the limber hole, and the working hypothesis is that the only nails recorded in the two garboard strakes helped to stabilise the piece before the hull was assembled and these timbers were physically joined. Why there are active frames in a ship that could be considered medium-small is not easy to explain. The working hypothesis is that, except in the navy and in very exceptional ships,21 shipyards mainly reproduced, or cloned, well-known and optimised models (Pomey 1998: 55; Bonino 1985: 37; Beltrame and Bondioli 2006: 89) that the fabri navales had always known how to build. It is in this reproduction of a type of ship answering a specific need, here the transport of dolia for the bulk wine trade, that active frames would have their place, helping to reproduce or clone fairly quickly a model demanded by the navicularius.
Figure 14. General state of the framing between M111 and M116. Note the greater height, by several centimetres, of frames M112, 113 and 114, coinciding (highlighted area) with the position of the dolium (see Figure 1). The wood was reinforced by iron concretions (white arrows) related to the nails of a lost stringer (photo: C. de Juan Fuertes).
Figure 15. View of frames M103 and M104 from their south face (2010 campaign). Note the central rectangular limber hole for the circulation of water in the bilge, coinciding with the keel. Despite the erosion, two heels can be seen on the underside of the frame, one on each side of the limber hole, with a total width matching that of the keel. Garboard strake B1E (lower right) can be seen to have lost its original rise over time (photo: Seguin).
Figure 16. Triangular limber hole made by two saw cuts in the underside of futtock AL109, coinciding with strake B6O (photo: Seguin).
Figure 17. Views of the keel (Q) and the west garboard strake (B1O) from the northern area of the wreck, showing remains of resin together with the treenails and nail of the fastening of a lost frame (photo: Seguin).
Figure 18. Underside of frame M109, showing one of the two heels with resin adhering, which suggests the working hypothesis that it was used to level the piece during the building of the ship (photo: Seguin).
Figure 19. Cut made for the wood identification of frame M109, coinciding with a treenail, which is driven at an angle and has a slightly truncated-cone profile (photo: Seguin).
Figure 20. Example of the upper face of a frame with the hole of a square-section iron nail, 2 mm a side, together with the characteristic stains of this type of oxidation (photo: Seguin).
Planking
The garboard strakes have a rather particular morphology, which changes along the hull from the central part towards the end, judging by the samples raised, although there are no data to analyse this change. At frame M109 they are completely rectangular, about 20 cm wide and 8.5 cm thick, with well-defined angles, except for the upper edge of each plank next to the keel, which seems slightly rounded or bevelled (Figure 10). At the end of the keel (Figure 11), on the other hand, they have a somewhat polygonal section, 18.8 cm wide and up to 7 cm thick. In both areas there is a difference in height between the keel and the garboard strakes. At the end, in section, the garboard strakes can be seen to have a small bevelled shoulder on their upper edges next to the keel, of gentler angle than in the central part.
As for the hull lines, the garboard strakes appear to form an angle of 0°/180° with the keel, perhaps partly because of the deformation of the hull, which has tended to flatten (Figure 11). The section drawn in 2010, despite some errors later corrected (Figure 12), showed very well a slight angle of 5° to the horizontal. At the end, the measurements taken in 2009 on the sample raised gave an angle of 18° between the upper face of the keel (the horizontal) and the garboard strake.
The mortises in the garboard strakes for the joint with the second strakes (B2E and B2O) are aligned, 2 cm from the upper edge, with average dimensions of 7.5 cm wide, 7 cm deep and 0.9 cm thick, spaced 5 cm apart in the portion of hull raised, which indicates a high density. The truncated-cone pegs have maximum diameters of 1.3 cm and minimum diameters of 0.8/0.9 cm, and were driven both from inside and outside the hull, which is related to the building process explained above. The treenails fastening the frames also reach the garboard strakes, apparently passing through them from the upper to the lower face, as seen in the hull at M109. From the diameters measured on the upper and lower faces of the garboard strakes, these treenails seem to have a slightly conical profile and to have been driven from the outside of the hull. In the section raised, blind iron nails of square section, 8 mm a side, had been driven from the outside of the hull into the garboard strakes to fasten frame M109. They had rounded heads 2 cm in diameter, perhaps truncated-pyramidal in shape, although the iron concretion did not allow a proper view (one in B1O and two in B1E).
The species chosen for the garboard strakes (Quercus) is unusual, without documented parallels (Guibal and Pomey 2002: 95–96); only the Giraglia 1, Fiumicino 1 and County Hall wrecks (Sciallano and Marlier 2008: 144) show that there are other cases of ships with oak planking, the wreck sunk in the Thames, the County Hall, being the only example of a ship built entirely of Quercus.22 In the author’s view, the species chosen for the keel (Quercus ilex) and the garboard strakes (Quercus) of the Giraglia 2 create in themselves a rather particular bottom that cannot at present be related to any other wreck.
The second strakes at M109 have a slightly polygonal section, the side face of the edge in contact with the garboard strake being 4.9 cm thick, decreasing to 3.9 cm at the side face in contact with the third strake; their role is clearly transitional. The only complete mortise that could be measured in this plank is 7.5 cm wide, 0.9 cm thick and 6 cm deep. The spacing of the mortises has increased to 9.5 cm on the face in contact with B3E, compared with the 5 cm measured on the face joining B1E. This spacing of 9.5 cm cannot be extrapolated to the rest of the hull, since measurements of mortise spacing on the east face of this plank elsewhere in the hull give 7.5 cm (with mortises 8.5 cm wide and 0.8 cm thick). The truncated-cone locking pegs in B3E (in the portion raised) were driven from the outside of the hull, with maximum diameters of 1.4 and 1.2 cm and minimum diameters of 0.8 and 0.9 cm. For these planks the Fagaceae were abandoned in favour of a very common resinous species, Pinus halepensis/brutia, both for the second strake and for the remaining strakes on the east (starboard) side of the Giraglia 2.
As for the planking strakes as a whole, 5 strakes survive on the east side and 9 on the west. The 8th strake consists of planks B8O, B9O and B10O, joined by simple bevelled scarfs. The strakes are laid edge to edge and are completely rectangular in section. The section raised shows no significant difference in the thickness of these planks or in the shape of their outer face.23 Their widths vary between 15.5 cm (B4E) and 37 cm (B8O), with an average thickness of 3.8 cm in the central part of the hull, although other measurements in situ gave 3.3 cm. The spacing of the mortises and tenons, basically staggered on each side of the strake (seen from above), is fairly similar, as deduced from the position of the pegs, about 15 cm apart in line, though this value is neither precise nor constant. A mortise measured in strake B9O at M109 is 8.7 cm wide and 0.7 cm thick. Those of strakes B3E and B4E at the same point are 8.5 cm wide and 0.8 cm thick, spaced 7.5 cm apart, locked by pegs 1.2 cm in diameter on the inner face of the hull and 1.4 cm on the outer face. In too many cases at least one of the pegs securing a tenon is missing. Where no pegs at all can be seen, this is very probably because they lie under the frames (Figure 3). The author interprets this lack of pegs as errors by the shipwrights, since there are also cases of two redundant pegs. Could this be related to very rapid construction of the ship?
Ceiling, thick stuff, and wales
None of the planks (ceiling strakes and/or stringers) forming the floor of the hold was found, although the systematic observation of the upper faces of frames M109 to M121 revealed concretions and holes related to the nails that held these planks. Longitudinal lines of iron concretions, or even the square holes of the nails that held a lost stringer, can be seen on the upper faces of frames M112 to M115, always in the western part of the wreck (Figures 14 and 20). The severe erosion of the framing does not allow further conclusions (fixed planks, removable ones, etc.).
Caulking
A resinous putty, possibly pitch, was found under frame M109 and at other points of the backbone, interpreted as a levelling material used during construction (see Frames).
Size and scantlings
The preserved portion measures 7.10 × c. 3 m. The keel is wider than it is high (20.5 × 15 cm in the centre), the garboard strakes very thick (8.5 cm) and the framing tight and robust (about 10 × 16.8 cm on average over the keel, room-and-space 14 cm). Given the limited height of holm oak trunks, the ship’s length should not have exceeded 20 m.
Table 1. Scantling of the timber remains of the Ouest Giraglia 2 Shipwreck
| Timber | Sided [cm] | Molded [cm] |
| Keel At the end | 20.5 18 | 15 14 |
| Keelson | – | – |
| Maststep (copper bolt Ø) | 1.5 | 27 (height) |
| Buttresses | – | – |
| Frames | c. 10 | 11–22 (avg. 16.8) |
| Futtocks | – | – |
| Room-and-space At central frames | 14 (avg.) | |
| Planking Garboard (at M109) Garboard (at end) Second strake | 15.5–37 20 18.8 – | 3.3–3.8 8.5 7 (max.) 4.9 to 3.9 |
| Ceiling | – | – |
| Main mast Ø Chock | – – | – – |
The longest surviving frame is 267 cm (M108). Planking: sided = width, moulded = thickness.
Wood
The wood identifications were carried out by Greck. Keel: holm oak (Quercus ilex). Garboard strakes: oak (Quercus). Second strakes and other planking: Aleppo/Calabrian pine (Pinus halepensis/brutia). Frames: oak (Quercus sp.) and elm (Ulmus sp.), in practically equal proportions (Figure 21).
Figure 21. Distribution of wood species in the hull of the Giraglia 2. Note the homogeneity of the backbone and framing, with very hard species (oak, holm oak and elm) (drawing: Greck).
Bibliography
Basch, L. 1972. Ancient wrecks and the archaeology of ships. The International Journal of Nautical Archaeology 1: 1–58.
Basch, L. 1987. Le musée imaginaire de la marine antique. Athens: Institut hellénique pour la préservation de la tradition nautique.
Beltrame, C. and Bondioli, M. 2006. A hypothesis on the development of Mediterranean ship construction from Antiquity to the Late Middle Ages. In L. Blue, F. Hocker and A. Englert (eds), Connected by the Sea. Proceedings of the Tenth International Symposium on Boat and Ship Archaeology, Roskilde 2003, 89–94. Oxford: Oxbow.
Benoit, F. 1958. Nouvelles épaves de Provence. Gallia 16.1: 5–39.
Bonino, M. 1985. Notes on the architecture of some Roman ships: Nemi and Fiumicino. In H. Tzalas (ed.), Tropis I. 1st International Symposium on Ship Construction in Antiquity, Piraeus 1985, 37–53. Athens.
Carre, M.-B. 1993. L’épave à dolia de Ladispoli (Étrurie méridionale). Archaeonautica 11: 9–29.
Casson, L. 1971. Ships and Seamanship in the Ancient World. Princeton: Princeton University Press.
Foerster, F. 1988. A third ancient ship of the 0°-class, with cargo from Catalonia. The International Journal of Nautical Archaeology 17.2: 178–180.
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.
Gassend, J.-M. 1985. La construction navale antique de type alterné: exemple d’un mode de construction. In H. Tzalas (ed.), Tropis I. 1st International Symposium on Ship Construction in Antiquity, Piraeus 1985, 115–127. Athens.
Gianfrotta, P. A. and Pomey, P. 1981. Archeologia subacquea. Storia, tecniche, scoperte e relitti. Milan: Mondadori.
Guibal, F. and Pomey, P. 2002. Essences et qualité des bois utilisés dans la construction navale antique. L’apport de l’étude anatomique et dendrochronologique. Forêt méditerranéenne 23.2.
Joncheray, J.-P. 1989. L’épave du Cap Gros. Matériel du Ier s. av. J.-C., pompe de cale en bois. Cahiers d’Archéologie Subaquatique 8: 57–84.
Joncheray, A. and Joncheray, J.-P. 2009. L’épave romaine de la Rabiou, Saint-Tropez (Var). Cahiers d’Archéologie Subaquatique 17: 63–102.
Joncheray, J.-P. and Rochier, R. 1976. L’épave de la Roche Fouras. Coque du navire, estampilles sur amphores. Cahiers d’Archéologie Subaquatique 5: 167–180.
Juan Fuertes, C. de 2008. La construcción naval del pecio de la Albufereta (Alicante). Un velero de comercio del s. I procedente de la Bética. Saguntum 40: 161–176.
Marlier, S. 2008. Architecture et espace de navigation des navires à dolia. Archaeonautica 15: 153–173.
Nieto, X. and Foerster, F. 1980. El pecio romano del Cap del Vol: campañas de 1978 y 1979. Cypsela 3: 163–177.
Pomey, P. 1998. Conception et réalisation des navires dans l’Antiquité méditerranéenne. In E. Rieth (ed.), Concevoir et construire les navires. De la trière au picoteux. Technologies, Idéologies, Pratiques 13.1: 49–72. Ramonville Saint-Agne: Érès.
Pomey, P. 2009. Principes et méthodes de construction en architecture navale antique méditerranéenne. De la conception à la réalisation. In X. Nieto and M. A. Cau (eds), Arqueologia nàutica mediterrània, 337–342. Monografies del CASC 8. Girona.
Pomey, P. and Guibal, F. 1997. Dendrochronologie et dendromorphologie des épaves antiques de Méditerranée. Bilan scientifique du DRASSM 1996, 102–103. Paris.
Pomey, P. and Rieth, E. 2005. L’archéologie navale. Paris: Errance.
Rival, M. 1991. La charpenterie navale romaine. Matériaux, méthodes, moyens. Travaux du Centre Camille Jullian 6. Paris: CNRS.
Sciallano, M. and Marlier, S. 2008. L’épave à dolia de l’île de la Giraglia (Haute-Corse). Archaeonautica 15: 113–151.
Tchernia, A., Pomey, P., Hesnard, A. et al. 1978. L’épave romaine de la Madrague de Giens (Var) (campagnes 1972–1975). Gallia suppl. 34. Paris: CNRS.
Ucelli, G. 1950. Le navi di Nemi. Rome: Libreria dello Stato.
Notes
1 Direct trade from the wine production centres to the storage and redistribution ports.
2 A methodology developed by the team, considered valid despite a number of handicaps. With the metric information taken in situ on the site (cumulative levels) and cross- and longitudinal sections, it was possible to complete a detailed plan.
3 The adaptation of a structure to a specific type of transport is made through the customary knowledge of the shipwrights who build the ship, and is therefore visible in particular details of construction.
4 Not counting a 30 cm fragment raised to the surface (Figure 5) during a verification by Sciallano and Minvielle (unpublished 2009, DRASSM archive), which corresponds to the after end of the wreck.
5 The orientation of the hull is explained below.
6 Marlier established a ratio of mortise density based on their spacing (6.3/6.5 cm) and average width (7.5 cm), giving a coefficient of 0.85 for the joint between the keel and the garboard strake at the end of the hull.
7 The Rabiou wreck, dated between the end of the 1st century BC and the 1st century AD, has a similar nail in the side face of the keel, for which no explanation has been given (Joncheray 2009: 71).
8 The identification of the wood species of the wreck was carried out by Greck.
9 The Dendrochronologie et dendromorphologie des épaves antiques de Méditerranée project of Pomey and Guibal has recorded the use of other Fagaceae and hardwoods such as Quercus sp. and Ulmus campestris (Madrague de Giens, Pointe de Pomègues, Laurons 1 and 2), Castanea sativa (Tradelière), Juglans sp. (Dramont A), Carpinus sp. (Chrétienne C) and Fagus silvatica (Chrétienne A). Quercus ilex would thus appear in the Dramont C and the Tour Fondue (3rd century BC), where part of the keel was of holm oak.
10 As was confirmed during the work of the 2011 campaign.
11 With the narrower side at the upper face.
12 Floor timbers: M101, M102, M103, M104, M106, M107, M108, M109, M110, M112, M114, M115, M116, M117, M118, M119, M120, M121; half-frames: M105, M111, M113; futtocks: AL107, AL109, AL117, AL121; unidentified: 2 cases.
13 Very low values from badly damaged pieces are excluded from these measurements.
14 Recorded only on the port side, but presumably with symmetrical counterparts on the starboard side, which is less well preserved on the wreck.
15 On the east side the planking does not survive up to the 6th or 7th strake, so the symmetrical limber holes cannot be observed.
16 The moulded dimensions of the frames in this sector are very high, more than 20 cm on average.
17 It is even possible that the holes were drilled before the frames were placed in the hull, the hull then being drilled last.
18 Not a very long interval in the evolution of ancient ship construction.
19 Active frame is a term defined by L. Basch (1972) for those pieces that help in fitting the planking strakes. They reveal an advance in the building process without altering the design principle (Pomey 2009: 339), as against the argument of Gassend (1985: 120, fig. 4a), who proposes a variation in the building process that consequently affects the design principle.
20 This detail was observed in situ by the team at the moment of dismantling.
21 For example the Syracusia, the Isis or the ship that carried Caligula’s obelisk; cf. Casson 1971: 184, Part 2: Oversize freighters.
22 Could the choice of this species be related to a way of building ships in yards of river tradition? Theophrastus, in Historia Plantarum V, 4, 3, notes that oak is a good wood for building ships for rivers and lagoons because of its resistance to rot.
23 This observation is made in the light of what was documented in the Chrétienne A (cf. Pomey and Guibal 1997: 103).