Portal de Periódicos da CAPES

The Mazarrón 1 Shipwreck: an iron-age boat with unique features from the Iberian Peninsula

2018; Taylor & Francis; Volume: 47; Issue: 2 Linguagem: Inglês

10.1111/1095-9270.12318

1095-9270

Carlos Cabrera Tejedor,

Archaeology and Historical Studies

Two iron-age shipwrecks, associated with Phoenician ceramics, were discovered at the Playa de la Isla in Mazarrón, Spain. This preliminary report describes hitherto unknown boatbuilding features of the Mazarrón 1 hull remains. The vessel presents hybrid boatbuilding techniques using both pegged mortise-and-tenon plank-edge fasteners and sewn seams employing longitudinal continuous stitching, and a unique keel scarf. It is an important source of information for the development of shipbuilding in the western Mediterranean during the Iron Age. El naufragio de Mazarrón 1: resultados del estudio preliminar de una embarcación de la Edad de Hierro de la Península Ibérica con características constructivas únicas Dos naufragios de la Edad de Hierro, asociados a cerámicas fenicias, fueron descubiertos en la Playa de Mazarrón, España. Este artículo discute las características constructivas, hasta ahora desconocidas, que se documentaron en los restos del casco del naufragio de Mazarrón 1. La embarcación presenta técnicas de construcción hibridas que emplean elementos de sujeción en los cantos de las tracas de mortaja y lengüeta, y costura de uniones. Es una fuente de información importante para el desarrollo de la construcción naval en el Mediterráneo occidental durante la Edad de Hierro. Palabras clave: Naufragio de Mazarrón 1, construcción naval, arquitectura naval, escarpe en forma de T, costura longitudinal de puntada continua, fenicios, Edad de Hierro, Península Ibérica. 马萨龙1号沉船: 对这艘带有独特造船特征的伊比利亚半岛铁器时代船舶的初步研究 在西班牙马萨龙的Playa de la Isla岛上发现了两艘与腓尼基陶瓷有关的铁器时代沉船。本文将讨论马萨龙1号 (Mazarrón 1) 船体遗存上保留下来的迄今为止未知的造船特征。该船混合了榫卯边接和缝制拼合的造船技术。这是铁器时代西地中海地区造船业发展的重要原始资料。 关键词: 马萨龙1号沉船 (Mazarrón 1) , 造船, 造船学, T型嵌接, 纵向连续拼接, 腓尼基, 铁器时代, 伊比利亚半岛 馬薩龍1號沉船: 對這艘帶有獨特造船特征的伊比利亞半島鐵器時代船舶的初步研究 在西班牙馬薩龍的Playa de la Isla島上發現了兩艘與腓尼基陶瓷有關的鐵器時代沉船。本文將討論馬薩龍1號 (Mazarrón 1) 船體遺存上保留下來的迄今爲止未知的造船特征。該船混合了榫卯邊接和縫制拼合的造船技術。這是鐵器時代西地中海地區造船業發展的重要原始資料。 關鍵詞: 馬薩龍1號沉船 (Mazarrón 1) , 造船, 造船學, T型嵌接, 縱向連續拼接, 腓尼基, 鐵器時代, 伊比利亞半島 Of the 117 wreck-sites found along the shores of the Mediterranean dated earlier than c.300 BC catalogued by Parker (1992: 10–12), only 11 had preserved hull remains (McGrail, 2001: 145). Thus, ancient shipwrecks found in the western Mediterranean that provide us with information about shipbuilding and its development before the Roman era are relatively scarce. Since Parker's publication less than a dozen shipwrecks of this period with preserved hull remains have been reported. Among them, two iron-age shipwrecks associated with abundant ceramics of Phoenician origin were discovered at the Playa de la Isla in Mazarrón, Spain (Fig. 1). This article discusses the presence of mixed shipbuilding techniques and hitherto unknown boatbuilding features documented on the Mazarrón 1 hull remains. Through comparative study of analogous wrecks, it is argued that the Mazarrón 1 shipwreck represents an important source of information for our understanding of ancient shipbuilding and its development during the Iron Age. The hull remains of Mazarrón 1, currently on display at the National Museum of Underwater Archaeology (ARQVA), Cartagena, Spain, reveal a number of shipbuilding features that, despite their being unique, have barely been mentioned in previous publications of the wreck, which partly inspired this detailed, though forcibly partial study and reconstruction of the vessel to be undertaken. The Mazarrón 1 hull was not archaeologically recorded or documented following the excavation of the timber remains in 1995 as they immediately underwent a lengthy conservation treatment that lasted until 2007. This project therefore is the first and only officially approved post excavation study of the Mazarrón 1 timber remains, other than the original excavators’ site reports.1 This study was completed in different phases from 2006 to 2009, which included a one-day direct inspection of the hull remains conducted in 2008 (Cabrera Tejedor, 2017: 190–193). This article aims to supplement previous publications that described but sketchily a number of unique shipbuilding features of Mazarrón 1; it provides a concise but complete summary of all previously published data about the Mazarrón 1 shipwreck and, indirectly, some of the Mazarrón 2 features used here as comparanda. The results of a preliminary reconstruction of the Mazarrón 1 hull are also briefly presented (see also Cabrera Tejedor, forthcoming; 2017). The results of this preliminary study of the Mazarrón 1 remains are used to re-examine previous interpretations that suggested the vessel was of Phoenician origin, and to propose new hypotheses regarding the nature, function, and origin of the construction techniques documented in the hull. The underwater site of Playa de la Isla was discovered in 1988 during a series of coastal surveys by a team from the Museo Nacional de Arqueología Marítima y Centro Nacional de Investigaciones Arqueológicas Submarinas (MNAM-CNIAS). The ongoing project has been overseen by a series of museum directors: Víctor Antona del Val, Paloma Cabrera, and Ivan Negueruela, but from 1988 until 1995 all fieldwork at Playa de la Isla was directed by archaeologist and permit co-holder Juan Pinedo Reyes. The remains of a wooden boat (Fig. 2), designated Mazarrón 1, was found and protected in situ in July 1991 (Cabrera et al., 1992: 38; 1997: 151; Barba et al., 1999: 196; Negueruela et al., 2000: 1671). From October 1993 to June 1995 a systematic survey of the area and documentation of the Mazarrón 1 wreck was initiated as part of the ‘Nave Fenicia’ project (Arellano et al., 1999; Barba et al., 1999: 197; Negueruela et al., 2000: 1671). In 1994 the remains of a second shipwreck were discovered and designated Mazarrón 2 (Arellano et al., 1999: 221; Negueruela et al., 2000: 1673–1674). The hull remains of Mazarrón 1 were raised in 1995 and transferred to the MNAM-CNIAS to begin conservation, while Mazarrón 2 remains in situ (Gómez-Gil and Sierra, 1996). The hull of the Mazarrón 1 shipwreck was found largely incomplete without any cargo, but the surviving timbers were relatively well preserved (Fig. 2). The hull timbers found consist of a complete keel, nine incomplete strakes of planking including one fragmented garboard, and four incomplete and fragmented frames (Fig. 3) (Negueruela, 2000a: 183; 2002: 167; 2004: 230; 2006: 24; 2014: 243; Negueruela et al., 2000: 1673–1674). When inspected in 2008, the timber remains were in two groups of articulated timbers: the first comprising keel, garboard, and second strake; the second comprising the preserved side of the hull from the third to the eighth strakes (Fig. 4); a third group of the disarticulated ninth strake and the frames was not examined. Both groups of assembled timbers were lying over supports made ad hoc, resting on the outboard surface, which allowed the inspection of the internal surfaces only. In previous publications of Mazarrón 1 and 2 (Cabrera et al., 1992; 1997; Roldán et al. 1994; Arellano et al., 1999; Barba et al., 1999; Negueruela, 2000a; 2000b; 2002; 2004; 2006; 2014; Negueruela et al., 1995; 1998; 2000; 2004; Miñano et al., 2012; Miñano, 2014), Mazarrón 1 hull is reported to have ‘the same construction method and fairly similar overall dimensions’ as that of Mazarrón 2 (Negueruela, 2004: 230; 2006: 24; 2014: 243). The claim that both hulls were made using the same construction method can be disputed as, although both hulls have pegged mortise-and-tenons fasteners, other construction elements present in the hull of Mazarrón 1 have not been reported in the hull of Mazarrón 2, such as longitudinal continuous stitching. The following presents a brief summary of the main shipbuilding elements of the Mazarrón 1 boat, and this information is compared with some features of the Mazarrón 2 boat as published. The keel of Mazarrón 1 has been preserved almost completely (Fig. 3). The aft end has some damage produced by biological attack (Cabrera et al., 1997: 151–52; Negueruela et al., 2000: 1672). The keel dimensions are 3.98m long, 170mm sided and 100mm moulded (Cabrera et al., 1997: 151–52; Negueruela, 2002: 165 fig. 3 and 167; Negueruela et al., 2000: 1675, fig. 1), although in 2008 the maximum keel width recorded was c.155mm. Elsewhere, it has been reported as 4.50m long (Negueruela, 2004: 230; 2006: 24; 2014: 243), yet this is not compatible with the evidence (Fig. 3). It has a rectangular cross-section amidships, presents no rabbets, and its bottom longitudinal edges are chamfered. Wood species identification of samples taken in 1997, determined that the keel is made of Cupressus sempervirens L. (Negueruela, 2004: 236–237; 2006: 25), commonly known as cypress, yet it has been mistakenly described as cedar (Negueruela, 2004: 237; Miñano, 2014: 6). The fore end of the Mazarrón 1 keel presents a unique type of scarf, that was mentioned by Negueruela et al. (1995: 196; 2000: 1673; Negueruela 2002; 167), but only recently fully published (Cabrera Tejedor, 2017: 210–213) (Fig. 5a, b and e). The scarf between the keel and the stem has two perpendicular tenons, one positioned above, though not touching, the other. The top tenon is horizontal and the lower one vertical, creating a T-shaped scarf. The vertical tenon is slightly damaged at its upper end. Additionally, the uppermost edge of the scarf was carved to create an inverted oblique angle of approximately 70°; this feature would have helped to secure the joint between keel and stem under vertical stress. The stem of Mazarrón 1 was not preserved, so matching mortises can only be hypothesized. The aft end of the keel appears to have had the same type of scarf (Fig. 5c, d), although it is not so clearly recognizable due to damage. This hypothesis cannot be confirmed without further analysis of the timber. Nine incomplete strakes of planking, including one fragmented garboard plank, were found. In 2008, it was observed that the width of the planks ranges from c.130mm to c.140mm, except for the eighth strake that is wider, at c.210–220mm. It was observed that plank thickness is c.36mm. The third, fourth, sixth, and eighth strakes have diagonal scarfs (Fig. 3), which do not present pegged mortise-and-tenon joints on their edges, except that in the eighth strake. The eighth strake is wider and reported to be thicker than the rest of the planking; it also has mortises that do not match the adjacent strakes and, consequently, it has been interpreted as a reused plank from a different hull (Negueruela et al., 2000: 1673), perhaps used in the Mazarrón 1 hull as a repair (Negueruela, 2002: 167). One fastening system used to assemble the hull of Mazarrón 1 was pegged mortise-and-tenon joints (Basch, 1972: 15). This type of fastener was used for shipbuilding in the Mediterranean at least since the Late Bronze Age as it is archaeologically documented in the Uluburun shipwreck, dated c.1320 BC (Pulak, 1998; 2005; 2008), where only fragments of the keel and planking survived. Half of a tenon survived from the hull of the Cape Gelidonya shipwreck dated to c.1200 BC (Bass, 1967; Pulak, 1998); however, the directors believe that its construction method was similar to the Uluburun vessel (Pulak, 1998: 210). The two boats found in Mazarrón are the next earliest-known archaeological examples using pegged mortise-and-tenon joints (Negueruela, 2002: 167; 2004: 246; 2006: 27; 2014: 243; Negueruela et al., 1995: 195; 2000: 1673; Negueruela et al., 2004: 480). Cato the Elder, writing in c.160 BC, described the use of pegged mortise-and-tenon joinery (De Agri, XVIII, 9) called them ‘Punic joints’ (punicanis coagmentis) (Sleeswyk, 1980: 243). Pegged mortise-and-tenon joinery was widely used for shipbuilding in the Roman period (Steffy, 1994: 43, 46, 77–78, 83–84; Casson, 1995: 203). In 2008, it was observed that mortises in the Mazarrón 1 hull are on average 30–36mm wide, 8–10mm thick, and 60–80mm deep. Tenons fit their mortises tightly both in thickness and width. Pegs are cylindrical in section with diameters of 7–10mm, although the majority are 8mm (Fig. 6). The spacing between pegs was rather difficult to document in some strakes since a layer of pine tar covers large areas of the hull planking (Gómez-Gil and Sierra, 1996: 219) (Fig. 2). However, when pegs were covered by the tar, their approximate position was documented by locating them in the open seams of the strakes. The spacing of those pegs documented varies substantially depending on the strake (Fig. 7), and two distinct zones with different arrangements were identified. First, on the keel, garboard, and second strake spacing between pegs varies from 110–240mm (195mm on average) and pegs are more closely spaced towards the ends of each timber. In contrast, a second zone was observed from the seam of the third and fourth strakes up to the ninth, where pegs are more widely spaced, varying from 280–520mm (400mm on average). It was also observed that tenons seem to be slightly larger in this second zone than in the lower part of the hull. Wood species identification of samples taken in 1997 determined that the planking is made of pine (Pinus sp.), whereas tenons and pegs are made of olive (Olea europea L.) (Negueruela, 2004: 236–237; 2006: 25). It was reported that at the seams the planks have chamfered edges and that thin ropes were placed in the resulting grooves to serve as waterproofing material, held in place with simple running stitching (Negueruela, 2000a: 196; 2002: 167; Negueruela et al., 2000: 1673). The initial hypothesis was that the stitching served to strengthen the union between planks and that the waterproofing ropes made the plank seams watertight (Negueruela et al., 2000: 1673); although in later publications only the waterproofing function was maintained (Gómez-Gil and Sierra, 1996: 219; Negueruela, 2002: 167). The rope and stitching are visible in the only in situ detailed photograph published to date (Fig. 8b). The position of the stitching is shown schematically in one drawing (NAVIS, nd), and one published site drawing (Negueruela et al., 1995: fig. 11) marked as diagonal parallel lines along the seams (Fig. 8a). Unfortunately, only scant remains of the waterproofing ropes and the original stitching have survived the extraction/conservation processes, since they were not readily apparent after conservation in 2008 (Fig. 8c). Consequently, the longitudinal continuous stitching is now only attested by the chamfered plank edges, the pre-drilled sewing holes and imprints of the stitches left on the protective coating of pine tar that was applied to the hull. Chamfered edges, waterproofing rope, and stitching were reported for the seams between the second and third strakes up to the seams of the ninth strakes (Negueruela et al., 2000: 1673). In 2008, however, no chamfered edges or sewing holes were seen along the keel-garboards seams or garboard/second strake seam. From the outer edge of the second strake upwards these features were seen in the remaining strake seams and scarfs up to the seam between the seventh and eighth strakes (Fig. 9) but not between the eighth and ninth strakes (cf. Negueruela et al., 2000: 1673), neither are chamfered plank edges and sewing holes found only in two small areas within the surviving part of the hull as reported elsewhere (De Juan, 2017a: fig. 8). The only preserved end of any of the surviving strakes of the Mazarrón 1 hull is the fore end of the second strake (Fig. 3). This end was originally fastened to the stem with pegged mortise-and-tenon joints alone as it presents neither chamfered edges nor sewing holes to accommodate stitching (Fig. 10). This limited evidence suggests that the stem and hood-ends were not sewn. Each plank edge is chamfered at an angle of c.45° starting about c.5mm from the edge of the plank, penetrating about c.5mm into the plank thickness (c.36mm); the chamfered edges of two adjacent strakes create a V-shaped groove in their seam. The sewing holes were also made at about c.5mm from the edge of the plank and they sit just within the groove along the chamfered edges (Fig. 11). It was noted that the sewing holes on adjacent planks were disposed diagonally to one another and not in opposing pairs (Fig. 12). They have an approximate diameter of c.2mm and are fairly regularly spaced c.20–25mm apart. The holes appear to be drilled through the plank at an approximately 90° angle—perpendicular to the plank's face and parallel to its edge—through the entire thickness of the plank, but this could not be confirmed since the external surfaces of the hull were not available for inspection. However, this hypothesis is supported by the following observations: first, given the proximity of the perforations to the edge of each strake (c.5mm), there is little space for the perforation to traverse the plank at a significantly oblique angle. Moreover, the edge surfaces of some planks were visible but no exit holes could be seen (Fig. 12b, c). Second, in the edges of some the planks cracks were documented that appear related to the perforations; these may have resulted from stresses sustained by the planking cracking the small wood interstice between the perforation and the edge of the plank. The few documented cracks are perpendicular to the face of the plank (Fig. 12c). Third, one plank fragment apparently recovered from Mazarrón 2 has perpendicular perforations traversing the thickness of the plank and also has a perpendicular crack (vide infra, Fig. 19). For all of the above, the sewing holes seem to traverse the strake thickness completely and are perpendicular to the face of the plank. The scant surviving evidence suggests that the sewing string fitted tightly in the sewing holes; accordingly, it would have had an approximate thickness of c.2mm. On careful observation of the published photograph, it could be tentatively estimated that the waterproofing ropes had an approximate diameter of c.6–8mm (Fig. 8). The balance of evidence suggests that the waterproofing ropes were placed in the inboard planking seams only; in the only schematic drawing of the stitching published by the excavators (Fig. 9a) it is represented inboard only. Nonetheless, this cannot be confirmed since the external surfaces of the hull were not available for inspection. It was reported that the waterproofing ropes were formed of a number of strands (Negueruela et al., 2000: 1673). Careful observation of the published photograph (Fig. 8b) suggests that both waterproofing ropes, and the string for the stitching, were formed of a number of strands; the waterproofing rope depicted seems to have been formed by two strands (Fig. 11). Fibre identification of a rope sample taken in 1997 revealed that it was made of esparto grass (Stipa tenacissima L.) (Negueruela, 2004: 236–237). Because the external surface of the hull planking was not available for inspection, the characteristics of the stitching on the outboard surface can only be hypothesized. However, as J.F. Coates rightly noted (1985: 15) ‘if wadding were semi-circular in section and fitted on both sides of the seam, the stitching would lose all capacity to pull the planks together and much of any resistance to sliding it would otherwise have had’. Therefore, the arrangement proposed in Figure 11 shows the simplest solution: the string, after exiting a sewing hole, enters the next one located diagonally on the opposite plank, creating a pattern of short, parallel, diagonal stitches. More complex solutions could have been used as documented in different sewn-plank boats around the world (McGrail and Kentley, 1985; McGrail, 2001; McCarthy, 2005: 10–21), but, until inspection of the external surface of the hull planking of Mazarrón 1 is permitted, it remains conjectural. Four fragmented frames were found (Figs 2 and 3). These were described as lightly worked branches (Negueruela, 2004: 237; 2014: 244). They were not available for inspection in 2008 but are reported to be cylindrical in section and 60–65mm (Gómez-Gil and Sierra, 1996: 219) or 70–80mm in diameter (Cabrera et al., 1997: 152), with room and space of c.0.45m (centre to centre). Wood species identification of samples taken in 1997 determined that the frames are made of fig wood (Ficus carica L.) (Negueruela, 2004: 236–237; 2006: 25, 29). The frames of Mazarrón 2 are reported to be cylindrical in section also, 40mm in diameter, with room and space ranging 0.40–0.50m (Negueruela, 2004: 249; 2006: 29; 2014: 244). They lay over the keel but were not fastened to it (Negueruela, 2004: 249–250; 2006: 29–30). According to wood species identification conducted in 2010, they are made of wood from the genus Juniperus (Miñano, 2014: 9). In the case of Mazarrón 1, the frames were lashed to the planking and some of the lashing cords were preserved in situ (Negueruela et al., 2000: 1673); when examined in 2008, only lashing points survived (Fig. 13). The holes were drilled in two pairs on either side of each frame so that the ligatures crossed over the top of the frame forming an X shape (Negueruela, 2004: 249; 2006: 29; 2014: 244). The sets of paired lashing points were positioned with two holes on each side of a strake seam (Negueruela, 2004: 245, 249–250; 2006: 29–30; 2014: 244; Miñano, 2014: 10). The lashing points of Mazarrón 1 are c.6mm in diameter and c.40 mm apart (Fig. 13). It was not possible to document the angle at which the lashing points were bored. After drilling, several individual lashings were tied to fasten each frame to the hull. De Juan (2014: 29; 2017b: fig. 7) offers a hypothetical reconstruction of how the Mazarrón 2 individual lashings were tied. This type of cylindrical frames may have an archaeological parallel in the tentatively dated 7th- or 6th-century-BC Golo wreck found off the island of Corsica, France (Pomey, 2012: 26, 28). Additionally, the type of lashing used to fasten the frames to the hull according to the excavators may have an archaeological parallel in the 4th-century-BC Benissafúller wreck found off the island of Menorca, Balearic Islands, Spain (De Juan et al., 2010: 65). The frames of this wreck, however, are trapezoidal in section, not cylindrical (De Juan et al., 2010: 66). In the Mazarrón 1 shipwreck no mast-step timber was found but its existence is suggested by the presence of six mortises on the upper side of the keel to which it could have been secured (Fig. 14). The mortises are all longitudinally orientated: four are located towards the bow in an approximately square arrangement; the remaining two are aligned aft of the other mortises. If the four fore mortises echo the position of the step in the mast-step timber, a similar position for the mast is seen, slightly towards the bow and at the fore end of its mast-step timber, as in the c.400 BC Ma‘agan Mikhael shipwreck (Linder, 1992; Linder and Rosloff, 1995; Linder et al., 2003). If correct, additional information regarding the Mazarrón 1 hull can then be deduced: the position of the missing mast allows us to identify the bow and the stern of the ship and thus the preserved strakes correspond to the starboard side of the original vessel (Fig. 14). Study of the Ma‘agan Mikhael shipwreck proved the need of a through-beam acting as mast-partner to secure the mast in conjunction with the mast-step (Steffy, 1994: 40–41; Linder et al., 2003: 105; Kahanov and Pomey, 2004: 7–8). For that reason, a preliminary reconstruction of the Mazarrón 1 shipwreck (Cabrera Tejedor, 2017) included through-beams. In Mazarrón 2 a 0.98m-long mast-step timber was reported (Negueruela, 2004: 241; 2006: 26), whereas Miñano (2014: 7), has stated that the mast-step timber is 1.04m long, 0.10m wide, and 0.06m deep. It is fastened to the keel with five mortise-and-tenon joints: four longitudinally arranged in a straight line, with the fifth, aft, orientated perpendicular to the others (Negueruela, 2004: 241; Miñano, 2014: 7). The mast-step timber is in direct contact with the keel and is notched over two of the cylindrical frames with transverse approximately rectangular slots made ad hoc on its base (Miñano, 2014: 7, fig. 5). This data has also been used to identify the bow and the stern of the boat (Fig. 15). In the case of Mazarrón 1 no through-beams were reported, and the uppermost parts of the hull were lost (Figs 2, 3). In contrast, in the Mazarrón 2 shipwreck seven through-beams were reported: four astern, one amidships (c.1m abaft the mast), and two more at the bow (Negueruela, 2004: 241–244; 2006: 24–26; 2014: 243). However, this description seems incomplete since an eighth through-beam is attested by notches in the top of the eighth strakes visible on published photos (Negueruela, 2004: figs. 12, 14, 25; 2006: figs. 13, 15, 20; Miñano et al., 2012: fig. 10). The starboard notch is close to a rectangular hole cut by the excavators to take a sample for radiocarbon analysis (Negueruela, 2004: 267 fig. 12)(Fig. 15). Judging by its position next to the mast-step, this through-beam may have served as mast-partner (Cabrera Tejedor, forthcoming), as seen in the c.400 BC Ma'agan Mikhael shipwreck (Steffy, 1994: 40–1; Linder et al., 2003: 105; Kahanov and Pomey, 2004: 7–8), although this mast-partner was placed aft of the mast. This through-beam and mast-partner is not mentioned in any published reports of the Mazarrón 2 shipwreck. In the only published reconstruction of Mazarrón 2 (Negueruela, 2006: figs. 6, 19, 30, 31), this through-beam fore of the mast-step is not included. Miñano (2014: 8), however, reports that ‘it is currently possible to detect the presence of more (through-beams)’ without additional detail. The through-beams in the Mazarrón 2 shipwreck were secured to the hull with dovetail ends fitted into ad hoc notches made in the upper edge of the eighth strakes (Negueruela, 2004: 241, 250–251; 2006: 26, 31–32; 2014: 244). The ends of the through-beams protruded outside the hull (Fig. 16) and were locked in place by the row of strakes inserted above them (Negueruela, 2004: 242, 272 figs. 21, 22; 2006: 40 fig. 26; 2014: 244). On the preserved parts of the eighth and ninth strakes of Mazarrón 1, there are two features that could be partial remains of dovetailed notches: on the upper aft edge of each strake, there is an angular feature in which a through-beam could have been fitted (Fig. 17). In Mazarrón 2, the through-beams were fitted on the eighth strakes (Negueruela, 2004: 241, 250–251; 2006: 26, 31–32; 2014: 244). The eighth strake of Mazarrón 1 is wider, reported to be thicker than the others, and a reused timber from another ship (Negueruela et al., 2000: 1674; Negueruela, 2002: 25). Marlier has suggested it is a wale (2005: 138). Together these observations support the suggestion that the eighth strake of Mazarrón 1 carried through-beams. Unfortunately, the aft broken fragment of the ninth plank of Mazarrón 1 was not present during the examination of the hull remains in 2008 (Fig. 4), and it has been impossible to inspect the Mazarrón 1 timber remains in detail since. Observation from outside the museum cabinet does not allow these hypotheses to be confirmed or dismissed. Nevertheless, there seems to be indirect and analogous evidence to suggest that through-beams were used in the construction of the Mazarrón 1 hull. Consequently, for the preliminary reconstruction of Mazarrón 1 (Cabrera Tejedor, 2017), a minimum solution construction was adopted resulting in six through-beams: two at the stern, two amidships, and two at the bow all fitted on the eighth strakes (see Fig. 24). The final step in the construction process consisted of waterproofing the hull with a protective coating of resinous material. Abundant remains of a protective coating can be seen on the hull remains of Mazarrón 1 (Negueruela, 2002: 167) (Fig. 4). A preliminary chemical analysis of samples taken in 1997 stated that the coating material was copal resin, which led to some preliminary erroneous interpretations of the boat (Negueruela, 2004: 255; 2006: 25). Further analyses conducted in 2002 concluded that the coating was pine tar, made from heating pine resin (Negueruela, 2004: 235–236; 2006: 25). A waterproof coating of pine tar was applied both internally and externally (Negueruela, 2002: 167), as also seen on Mazarrón 2 (Negueruela et al., 2000: 167; Negueruela, 2004: 251–252, Negueruela et al., 2000: 1674). The distinctive T-shape scarf from the keel of Mazarrón 1 is contemporary with the trait de Jupiter scarf in use at least from the 6th century BC as documented archaeologically in the Jules-Verne 7 and 9 shipwrecks discovered in the ancient port of Marseilles (Pomey, 1999; 2001; 2003). The Jules-Verne 9 vessel was c.9m long and c.1.60m in the beam (Pomey, 1999: 148) and thus similar in size to Mazarrón 1; its trait de Jupiter scarf is the closest archaeological parallel to the Mazarrón 1 T-shape scarf in form, geographically, and chronologically (Fig. 5). The Mazarrón 1 T-shape scarf differs from the ‘trait de Jupiter’ in that it does not have a self-locking mechanism by means of a pegged tenon (Fig. 5). The T-shape scarf was locked by the garboards and the second strakes being fastened to the keel and stem respectively with pegged mortise-and-tenon joinery (Figs 3 and 7). The design of the Mazarrón 1 T-shape scarf would withstand stress from pressures derived from hogging and sagging of the hull and the keel of the ship caused by the imbalance of weight and buoyancy along the length of the hull (McGrail, 2001: 147). Its design would prevent the separation of the keel and stem despite stresses in the vertical and transverse axes or through torsional forces in the longitudinal axis. In contrast to the trait de Jupiter scarf, however, it would not be able to contain tensile stress in the longitudinal axis because of the absence of a self-locking feature. This tensile stress would have been sustained instead by the pegged mortise-and-tenon joinery of the garboards and the second strakes. Consequen