A Question of Marks: Timber Marks from the Ses Llumetes Shipwreck, 1st Century AD, Mallorca (Spain)

ABSTRACT This article presents the timber marks found on one of the Ses Llumetes shipwreck’s frames (1st century AD; Mallorca, Spain). From this find, a review over time of this type of archaeological evidence is developed, comparing the use, meaning and interpretations in the different comparative contexts found in the Mediterranean. Timber marks on vessels are particularly rare to find in underwater archaeological context, especially for the Roman Period. This makes the discovery of the Ses Llumetes shipwreck an important contribution to the existing examples. The location of a mark in a specific location in a vessel can mean that its interpretation can vary, and also shows relevant details about the people behind this action. In particular, the analysis of the tria nomina represented by the Ses Llumetes mark has proved to be especially important for investigating where the ship might have been built.


Introduction to the Ses Llumetes Shipwreck: A Possible Italic Vessel with a Southern Iberian Cargo
Porto Cristo harbour (Mallorca, Spain) is a very suitable anchorage, especially for protection against the north-west winds.The harbour has an additional advantage in that a natural spring is in close proximity, which could have served to supply ships in Antiquity with fresh water.However, for inexperienced sailors, Porto Cristo was also a potential trap during severe storms, which may have contributed to the high number of shipwrecks (up to five shipwrecks ranging from the Roman Republican Period to the 19th century) found in a relatively small area (Munar et al., 2017).One of these, the so-called Ses Llumetes shipwreck, was discovered in the mid-1920s due to new harbour infrastructure works (Figure 1) and was dated to the 1st century AD.Consequently, in 2015, the Institut Balear d'Estudis en Arqueologia Maritima (IBEAM) launched an underwater archaeological project which permitted the excavation, research and protection of the archaeological remains from the shipwreck (Aragon et al., 2016;Catalá et al., 2019).
The Ses Llumetes shipwreck was found 30 m from the beach and in 2 m depth.The site is covered by a layer of sand of 30-40 cm, with some inclusions of stones and seaweed.The Ses Llumetes shipwreck was discovered as the new harbour infrastructure caused a change of sediment dynamics that exposed the hull remains together with its cargo.The loss of sediment and the shallow depth opened the site to a continuous process of degradation and looting over the years.However, it was also the subject of occasional studies since its discovery (Massanet, 1957;Pinya, 1954).
The vessel was carrying a fairly homogeneous cargo containing amphoras corresponding to productions from the south of Hispania, mainly Dressel 7-11 types, together with Haltern 70 and Dressel 20 types.The amphoras are complemented by Italian materials, including a representative collection of oil lamps that give the site its name.Finally, an interesting addition was the presence of sediment that covered most of the preserved site.This sediment contained volcanic sand, identified as puzzolana, mixed with various construction materials.The cargo potentially represents the route of the vessel's last voyage, connecting the province of Baetica, towards the Balearic Islands to Italy or vice-versa.The Iberian Peninsula is represented by the cargo of salted-fish products, wine and oil, and the Italian materials are represented by the oil lamps and other artefacts that complete the archaeological remains (Munar et al., 2017, p. 219).
The preserved hull remains show several elements that allow us to define some of its characteristics.A keel with a trapezoidal section, higher than it is wide, that is wide at the upper surface and narrows at its lower surface, and with no rabbet, was found north-west of the site.The garboard strake was formed by two planks made out of two different wood species (pine and oak), joined to the keel using mortise and tenons.The upper strake, as well as all the other hull strakes, are 6 cm thick.The lower strake, however, decreases in size, from 8.8 to 5.5 cm in thickness, ending in a small moulding 2 cm wide on its side face.The fastenings of the Ses Llumetes shipwreck are characterised by the presence of nails and treenails.Finally, height (18-20 cm) and width (13-15 cm) of some of the frames indicates a very robust structure (Munar et al., 2017, pp. 221-222).
The hull of the Ses Llumetes shipwreck reveals features that provide a better understanding of the changes in naval architecture that took place between the Late Republican Period and the beginning of the Imperial Era: such as changes in waterlines (tending to flat bottoms), the keel fitting arrangement, or the use of metal fastenings, among other details (Figure 2).This article, however, focuses on details from the vessel that highlight different aspects of the people involved in the ship's construction and the construction process itself.More specifically, this paper discusses a number of timber marks discovered on the hull.Marks related to ship construction and assembly appear to be uncommon or are frequently overlooked due to their subtle nature (Atkinson, 2007).This paper aims to present the significance of the timber marks from the Ses Llumetes shipwreck, gviing an overview of the use of these marks throughout history and adding new evidence to the discussion of their use and interpretation within the context of shipbuilding during the Roman Period.

Marks on the Ses Llumetes Shipwreck
As will be seen through different examples in this article, it can be argued that timber marks on ships are particularly rare, especially for the Roman Period.This makes their discovery on the Ses Llumetes shipwreck an important contribution to the existent compendium (Figure 3).One example of the rarity of timber marks is that of the 37 wrecks recently found in the Yenikapı excavations (Istanbul, Türkiye) only one (YK35) has a mark on its timbers (Kocabaş, 2015, p. 25).
There are four timber marks in the case of the Ses Llumetes shipwreck, which appear on the west side of frame C-111; however, only three are clearly visible.The marks are separated by a regular distance of 20 cm, and are a stamp that is 5 cm long and 1.2 cm high.The four stamps appear to have been burned into the wood, presumably using a signaculum.From the marks, it is possible to read: C•I ( VL ) • ( TELE ) S (Separation points.Ligation of V and L of IVL.Quadruple ligation of T, E, L and E from TELES).
This small set of references suggests that if the man in the mark from the Ses Llumetes shipwreck was named C. Iul(ius) Teles(inus), his most probable origin was Italy.The second hypothesis considers the development of Teles(…) in Telesphorus, or in one of its many variants (Telesfor, Telesforus, Telesporus, etc.).Unlike the previous one, this name was widely used throughout the Empire.Its Greek origin (Τελεσφόρος: this name may have had, at least originally, a relationship to religious initiation rites: Pape, 1875, p. 1504) also would explain why the men who carry this cognomen belong to a varied palette of legal statuses and social levels: slave, libertus or citizen, since it was frequently given to slaveswhether they were from a Greekspeaking region or the Italic Peninsula (Figure 5).
This cognomen was widely used throughout the Empire with more than 200 inscriptions known (based on Epigraphik-Datenbank Clauss / Slaby-EDCS online): attestations can be found in Moesia, Britannia, Numidia, Pannonia and even Arabia.But in these provinces, Telesphorus appears once or twice, which does not compare to the presence of this name in Italy: more than 120 texts come from Rome and almost 40 from Lazio and Campania, including 14 in Ostia.Even the epigraphic material provides evidence of four individuals named C. Iulius Telesphorus.None of the three men from Rome can be identified with the person mentioned on the Ses Llumetes wreck, due to the chronology and because it is a frequent personal name (CIL, VI, 19837, 20295 (Telesporus, 2nd century), 31234 (on a list of vigils, dating from the time of Marcus Aurelius).In Ostia: CIL, XIV, 1170: C(aius) Iulius Telesphorus, domo Ravennans, sibi et Calpurniae Gemellinae uxori fecit et libertis [l]ibertabusq(ue) posterisq(ue) [e]orum, [in fro]nt(e) p(edes) XX in agr(o) p(edes) XXV).On the other hand, an inscription from Ostia, dated from the 1st century AD, must be considered: C. Iulius Telesphoruslibertus or ingenuusduring his lifetime, had a marble epitaph erected for himself, his wife, his freedmen and their children (CIL, XIV, 1170).The man was certainly from Ravenna, where he lived as he expressly states (L.2: domo Ravenna {n}s).He had enough ties to Ostia to decide to establish his final resting place there with his wife and freedmen.Of course, the utmost care must be taken in comparing the marks from the wreck and the inscription in Ostia.Nevertheless, the hypothesis that he could have been a navicularius maris Hadriatici has already been put forward, even if it could not be based on solid arguments (2006, p. 298, n°95; the navicularii of the Adriatic Sea were probably wine merchants from the Adriatic coast).Another hypothesis proposes that he was a sailor of the Ravenna fleet (Cébeillac-Gervasoni et al., 2006, p. 298, n°95;De Salvo, 1992;Pellegrino, 1987).
If this is the same Telesphorus from the Ses Llumetes wreck as in Ostia, a third hypothesis can be put forward: Telesphorus owned shipyard(s) in Ostia, and perhaps in Ravennaif at least owning a shipyard under the Empire made senseor ran a business (of naval carpenters?) involved in the construction of merchant ships.
In summary, the study of this name suggests that the development of Teles(…) as Telesphorus rather than Telesinus is more likely and that Telesphorus was certainly from Italy, and more precisely from Lazio-Campania, with a preference for Ostia: these two remarks are based on a simple reason of frequency and chronology of examples as the Ses Llumetes shipwreck dates to around the 1st century AD.Furthermore, the Greek origin of the cognomen is well known among freedmen, making it probable that the person represented in this mark corresponds to someone of former servile status, even if it is possible to exclude that he was a citizen.The hypothesis that Telesphorus from the Ses Llumetes wreck marks is to be identified with the character of the epitaph of Ostia must be considered with the utmost caution.

Timber Marks from Early to Modern Times
The evidence from the Ses Llumetes shipwreck requires a review of the different examples of timber marks that present similar characteristics.This brief overview extends not only to Roman but also to the medieval and Modern Periods, giving a wide framework to visualise the function of these marks in relation to the construction ships and the different actors involved.Evidence of timber marks from wrecks are rarely recorded in the archaeological context.One of the earliest examples of timber markings is found on the Cheops ships (Lipke & Moustafa, 1984) discovered near the Great Pyramid in Egypt (Jenkins, 1980, pp. 124-130;Throckmorton, 1987, p. 93).The marks appear on a significant number of timbers and were painted in black.Along with these marks, there were other symbols interpreted as directly connected to the location of the timbers on the structure, with the objective to serve as a guide to the construction of the boats (Atkinson, 2007, p. 26).
Incised construction marks were also found on two of the Place Jules-Verne shipwrecks (ca.late 6th century BC), excavated in Marseille, France.In the Jules-Verne 9 wreck, a fishing boat built in the sewn tradition, a pattern of three lines were regularly incised to indicate the locations for sewing holes, to ensure proper spacing.In the Jules-Verne 7 wreck, likely a merchant ship, arrows were incised in the hull planks to indicate the locations of internal framing (Creasman, 2010;Pomey, 1998, p. 152).Such marks, often termed 'scriber marks' (see Steffy, 1994, p. 43) can also be found on the Kyrenia ship (ca.290 BC), excavated off the north coast of Cyprus (Steffy, 1985, p. 85).
Another early example of the use of timber marks are the ones found on the Marsala shipwreck, dated to the 3rd century BC (Frost, 1972(Frost, , 1976)), excavated off the west coast of Sicily.In this case, the marks, representing Punic letters, are painted in black on the hull planks (Johnstone, 1981(Johnstone, , pp. 198-219, 1983)).The different spellings reveal the interpretation of at least five characters.They highlight a system of work organization based on the pre-assembly of certain strakesa form of prefabrication probably intended for mass production (Frost, 1993, p. 55).Other marks, only incised, are also present in large numbers and some indicate the location of the frames.A large number of these marks and symbols were obscured by either adjoining hull components or other processes such as the covering of the outer lower hull with lead sheathing (Frost, 1972, p. 44).
Incised lines on hull planks, similar to those found on the Marsala shipwreck, were also present on the Ma'agan Mikhael wreck, dated to the 5th century BC (Mor, 2004, pp. 174-175), excavated off the coast of Israel, probably with the same function to indicate frames location during the assemblage process.The use of 'X' marks was found on the primary structural frames, especially the keel, stem, and hull planks (Mor, 2004, p. 167).They were used to indicate the exact location of the timbers in relation to one another.
The presence of 'X' marks and other incised lines on the inner face of the strakes appears to continue throughout the Roman Period, indicating the location of the frames.An example of 'X'-shaped graffiti is found on the inner side of repair strakes in the wreck of Grado (off Aquileia, Italy) dated 2nd century AD.These particular marks are always located in the gap between the frames (Beltrame & Gaddi, 2007, p. 144).There is also evidence of marks on the socalled wreck of Mandirac (Narbonne, France) dated to the early 5th century AD (Jézégou & Chaussade, 2020, p. 79).On the other hand, in this same wreck there is less evidence of numbers indicating the order in which the main parts of the hull were assembled.On the stern foot, an engraved mark takes the form of two vertical lines which suggest it correspond to a Roman numeral 'IIø (Jézégou et al., 2015, p. 35).This number is probably an indication that the stern foot is the second structural element erected just after the keel.The use of numerals connected to shipbuilding during the Roman Period is exhibited in the case of the Plavac wreck near the Île de Zlarin, Croatia, a site that also shows the use of numerals as a guidance system to its structure (1st century AD) (Radić Rossi et al., 2019).
In regard to the use of timber marks, it is necessary to mention the one found in the Caesarea shipwreck in Israel and dated to 1st century BC (Derenne et al., 2019).This wreck reveals a number of marks interpreted as 'inscriptions' found on the hull framing (Raban, 1985, p. 175).Two groups of letters were carved into the upper face of the inboard extremity of futtock 21.More specifically, the letters were located in the connection between the floor timber and the futtock.The two groups of letters repeat a single word that may be read 'ΧΙΛΙαΙΨΙΝ', interpreted as a number instead of a name (Fitzgerald, 1994, p. 166;1995).Despite the original significance and purpose being unknown, it was suggested they might have been engraved by a stevedore involved with loading the ship and perhaps related to the tonnage (Raban & Oleson, 1989, p. 188, Pl. III: 190).Unfortunately, these two marks are no longer visible, and reinterpretation cannot be discussed (E.Nantet, 2020, pers. comm.).
The inscription found by V. Scrinari on the Fiumicino 1 wreck, near Portus, dated to 4th-5th centuries AD, is closely related to the evidence discussed so far.There is not much information about the mark's appearance, but the author describes it is as a mark made by fire on a frame.It is interpreted as a Latin epithet trituta meaning 'saved three times', which leads the author to interpret the mark as the name of the ship placed on a piece that would be reusedthree timesto avoid bad luck (Scrinari, 1979, p. 37, ill. XXVIIII).G. Boetto (2008, p. 45) re-studied this restored vessel, noticing the mark was no longer visible.She suggests a more plausible interpretation as a shipyard mark, similar to the one found on the Saint-Gervais 3 wreck (Liou et al., 1990).Indeed, the Fiumicino 1 wreck, showing the inscription on a frame, which was not visible, could hardly correspond to the name of the ship.Moreover, in iconographic (Basch, 1987, pp. 467-468) and ancient textual sources (Casson, 1971, p. 356ff.)or epigraphy for the case of military fleets (Buonopane, 2017), ships often bore their name using a tutelary deity.They are frequently found on anchor stocks, reflecting the name of the vessels to which they belonged (Brody, 2008, p. 3;Gianfrotta, 1980, p. 109;Fenet, 2016).In other cases, anchors stocks show tria nomina, and it has been argued there is a possibility they would refer to the name of the owner of the ship to which the anchor belonged (Gianfrotta, 1980, pp. 110-112).These names are sometimes found in the same form on all or part of the cargo, indicating that the owner of the vessel is also the merchant responsible for the maritime transport (Hesnard & Gianfrotta, 1989, p. 397;Hesnard et al., 1999;Liou & Domergue, 1990, p. 92).Returning to the timber marks, the most relevant of this kind can be found during the Roman Period in France, where three wrecks show marks among their timbers.On the wreck of Saint-Gervais 3 dated in the middle of the 2nd century AD (Liou et al., 1990, p. 245) the mark 'C ( ANT ) IOR ( VM )' is repeated six times, appearing at the bow, on the keelson corresponding to a plural genitive name from Celtic origin Cantius.
The mark appears together with the letters 'R R F G' repeated four times on the same frame and positioned slightly before the previous mark; in this case, B. Liou did not propose any interpretation.The mast-step also displays two other marks that were not described by B. Liou but some authors mention these marks as 'CANTI[?Et ]ASAI' (Marty et al., 2016, n. 10, p. 275).One last mark was also registered: Q • M • F • (probably the initials of tria nomina), in this case in a frame.It was imprinted four times on the base of the timber floor number 150, located at the forward face, inside the limber hole, at a place where it was invisible once the floor timber was placed on the keel (Liou et al., 1990, p. 234).Due to the narrowness of the limber hole, these marks were not visible throughout any stage of the assemblage.As in the case previously, the mark 'C ( ANT ) IOR ( VM ) ' is located in a spot that was incised when the ship was being assembled, making it inaccessible once the boat was in use.Furthermore, these kinds of marks seem to be necessarily associated with shipwrights (Liou et al., 1990, p. 245).
Another tria nomina was recorded in the La Bourse shipwreck (late 2nd century-early 3rd century AD), excavated in Marseille (Gassend et al., 1982).In this case the stamp 'M•VIB •AVXE' is repeated up to four times and it can be read as: M(arcus) Vib(ius) Auxe(tus), interpreted as corresponding to the shipwright (Tran, 2014, p. 169).
However, when referring to timber marks, one of the most representative studies are those from the shipwreck Arles-Rhône 3, a river barge specially fitted with a central box to hold a cargo of stones (Marlier & Djaoui, 2014, p. 271).Here seven marks were recorded together with seven graffiti.Tran (2014) interpreted one of the marks 'C-L-POS(TV)' as C(ai) (et) L(uci) Postu(miorum), repeated twice, one on the inner face of the upper starboard side, thus partially visible.The other mark was in the centre of the upper part of a timber floor, thus covered by the ceiling.Tran interprets the marks as representing a society of shipwrights from the same family.This interpretation seems to coincide with the fact that it is the only mark that appears on the primary structures of the hull.The other six marks, 'T•C•L•CE (TA)'; 'FID•T++'; 'NOBIL•M[-?]'; 'ɔOC'; '[-?] SS•SOL'; and 'EM++' (see Tran, 2014 for proposed reading of each mark) were placed on the sides and on the floor of a removable central box containing the stone cargo.Some of these planks were replacements (Marlier, 2014, p. 160); they were dismantled from another ship on which they may not have had the same function.The presence of reused pieces complicates the analysis since these pieces may have been marked as part of their original use but does not fundamentally alter the interpretation proposed by N. Tran, where Caius and Lucius Postumius would have formed a society of marine carpenters.Their hallmark could have been used to distinguish their products from those of other carpenters working in the same yard (Tran, 2014, p. 169).The seven graffiti represent a series of 'X' or 'IX' marks that are superficially incised on parts of the central box that are not assembled but only laid, with the exception of a strake used as a repairon the starboard side.Without this observation, one might have been tempted to consider these crosses as possible numerical indications for carpenters to use as reference points (Tran, 2014, p. 170).At least in this case they do not seem to fulfil this function, although we cannot deny that in their original position they could have been used as assembly markers.
More direct parallels beyond the above presented shipwreck contexts are found in ancient port structures.The first parallel is located in Rezé (Ratiatum), the Gallo-Roman port of Nantes (Loire Atlantique).This site has structures built in stone and wood using the technique of juxtaposed boxes, dating from the end of the 1st century AD to the middle of the 2nd century AD (Mouchard et al., 2016, p. 260, fig. 17).The timbers that are part of the construction appear to be consecutively marked and laid out horizontally on the ground, following a pre-established pattern.It can be stated that at an early point of the production, each element of the planned structure was given a numerical mark so that they could be positioned at their specific location when the entire structure was raised and permanently installed (Mouchard et al., 2016, p. 251, fig. 5).Following this, each part had a precise position because of their morphology.The marks are inscribed on the right side, identically on each module, using Roman numerals in ascendant order (Mouchard et al., 2016, p. 259, fig. 16).Once the marking was completed, the structure could be transported to its final location and installed.During the excavation, after dismantling one of the described structures, it was possible to see that a stamp appeared to be repeated several times, consisting of the three letters: 'P•A•S', that might correspond to a tria nomina (Mouchard et al., 2016, p. 261, fig. 18).The stamps are located at the bottom part of the wooden piles (repeated eight times), which were to be placed on the ground and inside of a carved box, where the stamp appears again three times.Once assembled in the final position, they would have been invisible.
As observed in the Saint-Gervais 3 and Arles-Rhône 3 wrecks, these stamps are therefore not intended to be seen once the structure is in use.They are only partially visible during the pre-assembly phase.The same pattern is found at Fos-sur-Mer (Bouches-du-Rhône), where a series of quadrangular wooden structures were put in place in AD 65/85, consisting of horizontal beams supported by piles (Marty et al., 2016, p. 266).Two of them have inscriptions on the base.On one is the inscription 'M•D•SE[', read as M(arci) D() Se(), that does not correspond to any possible gens or cognomen (Marty et al., 2016, p. 274).On the second, there are two inscriptions, one repeated twice: 'T•C•SAT•', read as T(iti) C() Sat().The second, partially covered by the previous one, therefore earlier than the first: ']VL Q•LAVR' or ']IVL Q•LAVR', interpreted as two names without cognomen: () Iul(i) and Q (uinti) Laur() (Marty et al., 2016, p. 274).
Following this line of evidence, in Antium (Anzio, Italy), a breakwater of formwork was built during the reign of Nero.The wooden piles have only one face roughly squared, and the end is blunt.On some of them were found several stamps branded on the wood using a so-called signaculum, a metal tag with personal details used as seal or stamp (see Loreti, 1994).This instrument was frequently used on barrels or wooden plugs (Desbat, 1991, p. 168) to authenticate objects or their contents.One of the piles bears five different marks, repeated several times (17 marks in total) and the second bears four of these marks, also repeated several times (nine marks in total), with an identical organizational scheme from the top of the pile to the end (Felici, 2002, p. 108, fig. 14, fig. 16).These five marks seem to refer to two tria nomina: 'L.L() M(), C. S() M()', which have not been interpreted to date; two public properties published as (scil.operum) pub(licorum) vel pub(licum), O(pus) P-(ortus) A(ntiatum) vel A(ugusti), and a possible collective of builders: Calpurnii Pisones, documented during Nero's time.On the two stakes, an inscription in the name of a citizen, 'L.L() M()', covers the public inscription, O(pus) P(ortus) A(ntiatum) vel A(ugusti), which was therefore made before it.Observation of the spelling of 'L.L() M()' shows two different matrices depending on whether the inscription is isolated or overlaps the public inscription.Felici (2002) noticed that the public inscription was never physically related to any other of the documented marks.Finally, this author argued that the inscription O (pus) P(ortus) A(ntiatum) vel A(ugusti), printed several times, is carefully placed at regular intervals and in the same orientation.It seems to have received scrupulous attention, interpreted by Felici as a procedure to prevent indiscriminate harvesting of all or part of the timber, or even as a marking, dictated by fiscal reasons, given the importance of its final destination (Felici, 2002, p. 114).
In Laurons (Martigues, Bouches-du-Rhône), 22 piles, numbered from 'II' to 'X', were observed on a quay installation on formwork.The date for this structure is estimated between the 1st century AD and ca. 4 th -5 th centuries AD.The matching numbers are identical on the piles and on the stones that receive them.The numbers sometimes appear to be superficially incised and sometimes printed (Ximénès & Moerman, 1988, p. 232, figs. 11, 26-27, 1989, p. 184, figs. 4-10).As for the case in Ratiatum (Nantes), the numbers 4 and 9 were engraved respectively as 'IIII' and 'VIIII', while 6 and 11 have the expected forms of 'VI' and 'XI', to obviously avoid any form of confusion in reading.The numbers based on 'V' appear upside down on the mounted structure, which means that they were engraved in such a way as to be read correctly in the pre-assembly stage on the ground.Finally, it is important to mention the foundation piles of the circus of Arles built in AD 149 (Sintès & Arcelin, 1996, p. 78).They bear, on the main body of the pile, a series of letters corresponding to tria or duo nomina and, at the top, was placed a number indicating the order of installation.
In shipbuilding, marks to indicate the frame positions during assembly become more frequent in the archaeological record moving forward in time.Representative for the early medieval period is the Tantura B wreck dated to the 9th century (Wachsmann et al., 1997, p. 6, fig.3), which for the first time the letters 'HX' are known to have been branded in wood used for shipbuilding, located on a timber in the north-western quadrant of the hull, although its interpretation is uncertain (Atkinson, 2007, p. 27).It was not until the 14th century that the use of wood markings became evident in the process of assembling a boat.This fact seems to be closely linked to the development of skeleton-first construction in the Mediterranean (Atkinson, 2007, p. 28;Hutchinson, 1994, p. 42).The Cala Culip VI wreck (Rieth, 1998, pp. 154-157) is undoubtedly the most representative of Mediterranean shipbuilding in the late medieval period.Therefore, it provides the first clear archaeological evidence of the use of the so-called 'master mould' and 'rising square' method.From the 15th century on, it seems that the use of timber marks on boats increased, always linked to the construction process.
The use of a variety of marks expanded over time.In particular, 'X'-shaped marks became more frequently used to indicate the specific location of frames, and examples can be found on wrecks from the Middle Ages to the present (Figure 6).Thus, on the keel of the wreck the Bataiguier in Cannes (Alpes Maritimes), dated to the 10th century, an 'X' inscribed in a rectangle is engraved on the keel, at the presumed location of the midship frame (Jézégou & Joncheray, 2015, p. 148).The Listel 1 wreck at Sète (Hérault, France), dated to the 16th century (Jézégou, 2009, p. 22), has different markings, inscribed on the outside of the inner planking.Finally, on the Agde J wreck, dated to the 18th century (Letuppe, 2016, p. 33) and discovered at Cap d'Agde (Hérault, France), an 'X' is deeply engraved inside the mast-step.There are multiple examples such as Grace Dieu, Southampton, UK (Friel, 1993), the Copper Wreck, in Gdansk, Poland (Litwin, 1980, p. 217), or the Ria de Aveiro A wreck in Portugal (Alves et al., 2001, p. 21).
Stamps on wood became a popular practice in shipbuilding during the modern era where the high volume of boat production forced the need for a more effective system of management (Atkinson, 2007, p. 234).Great examples in this regard are the wrecks such as La Belle that 'was built as a prefabricated vessel, philosophically similar to inexpensive assemble-at-home furniture kits available today' (Creasman, 2010, p. 47).Finally, one of the best examples of wood markings is HMS Victory, which has recently been the subject of an exhaustive study (Atkinson et al., 2019).This vessel shows in detail the typological variety, distribution, use and meaning of the markings on timber and their relationship with the architecture of the vessel itself.It only remains to mention a stealthy inscription engraved at the end of a futtock on the Jeanne Elisabeth shipwreck, dated to the mid-18th century.The letters 'RP' are interpreted either as the initials of a carpenter or as an indication of payment for wood delivered to the shipyard (Jaouen et al., 2017, p. 49).
It is not the objective of this article to cover in detail the evolution of the timber marks process throughout history.However, it is important to note the need for further studies to identify the different historical and functional implications that these marks entail regarding the understanding of the evolution of shipbuilding.Besides, from the brief overview presented in this paper, it can be argued that, at least during Antiquity, two groups of marks can be distinguished based on their function.A first groupnumbers, letters or graffitithat seem destined to mark a reference in the shipbuilding process and a second group of identifying marks of individuals involved in different ways and at different stages in the construction process.Due to the nature of the archaeological evidence from the Ses Llumetes shipwreck, the discussion will focus on this second group.

Discussion
The use of timber marks is frequently interpreted as representing a stage during the construction process that must have more or less developed in certain social, political and economic contexts (Atkinson, 2007, p. 19).Thus, some authors associate the use of certain marks on wood with changes in shipbuilding practices (Atkinson, 2007, p. 28).The evolution of these practices through technological and industrial advances leads to a larger workforce and production.The increase in production or larger size vessels forces the creation of more structured and standardized processes in shipbuilding (Atkinson, 2007, p. 24).It has been argued that this process illustrates technological transitions such as the passage from shell-first to skeleton-first construction in the Mediterranean and later in northwest Europe (Harpster, 1997;Arenson, 1990;Steffy, 1994).It is in this context that marks on wood seem to proliferate and play a specific role in the assembly process.
As we have seen in different examples, the letters on woods from hulls or docks are responses to a different nature and function.In the majority of cases, these painted or incised letters, and in rare cases also branded, correspond to simple markings intended for the assembly (or reassembly) of these structures.However, it is not possible to include all of them in this function, so many of them are still difficult to interpret (e.g.carved into the futtock of the Caesarea wreck or the case of the Fiumicino 1 wreck).Despite this, it can be argued that the longest inscriptions, composed of several letters, always appear branded using hot metal or a signaculum, which makes them more complex to define.The use of this instrument suggests that they belonged to a more professional structure.The use of a signaculum is frequently suggested for timber marks (Felici, 2002, p. 111;Tran, 2014, p. 195).These tools were most often made of iron and sometimes bronze, and are associated with wood and leather craftsmen (Desbat, 1997, p. 115).
In general, the difficulty of fully understand these marks is due to the scarcity of evidence: the number of finds is very small in relation to the number of wrecks excavated.These marks, which are hardly discernible in the water due to their location and small dimensions, could only be read during the partial or complete dismantling of certain frames and planks, an operation that is still rare in underwater archaeology.This is why the corpus of inscriptions, for the first two centuries of the Roman Empire, is limited to four wrecks (Ses Llumetes, Saint-Gervais 3, Arles-Rhône 3 and La Bourse) and to the piles discovered in three port facilities (Antium, Fos-sur-Mer, Rezé) and in a circus (Arles).These inscriptions have four points in common: 1) they are always more or less developed abbreviations; 2) they are repeated several times on the same piece or on two different pieces, as is the case for 'C.L. POSTV' on the Arles-Rhône 3 wreck; 3) they are engraved/printed on in shaped wood, which makes it possible to exclude them from the outset a function of these marks in the context of the work of the raw wood or a fiscal function in relation to their transport; 1 and 4) in the majority of cases, they are not intended to be visible once the structure was assembled.
These inscriptions are most often limited to simple initials (Q•M•F, etc.) and the most developed cases make use of ligatured letters.This demonstrates a conciseness in a standardised process, which goes hand in hand with using small-sized letters, showing that the intention was not to waste time or money in making them.The most highly developed inscriptions make it possible to identify names in the form of a duo or tria nomina, representing individuals (Ses Llumetes, Saint-Gervais 3) or groups (Arles-Rhône 3).The two public corporations and a collective of corporations, the Calpurnii Pisones, found at Antium, are the best examples.These 'more complex' marks suggest that the shortest inscriptions were also names, reduced to their simple initials, 'Q•M•F', would thus correspond to Q(uinti) M() F().Nevertheless, some abbreviations, such as 'R R F G' (Saint-Gervais 3), are still hard to interpret.Although the inscriptions on piles and ships' hulls have some common features, the reviewed corpus nevertheless allows us to distinguish specific features of the wood used in docks.Rezé and the circus of Arles provide the only examples of piles where both abbreviated names and assembly marks corresponding to the numbering of the pieces for their assembly are found.Another characteristic is that in Antium, as in Fos-sur-Mer, certain inscriptions cover others, which implies at least two stages between the squaring of the log and the assembly on-site of the various pieces of wood.The piles from Antium provide additional information with the presence of different matrices of the same mark, depending on whether the inscription in the individual's name overlaps the public inscription or not (Felici, 2002, p. 112).Finally, although the marks on piles are similar in form to the marks found on ships, the difference is that those on docks often refer to public facilities.Only the small port of Les Laurons, near Fos-sur-Mer, France (Vella et al., 2000) could be considered a maritime villa and therefore a private property.Its piles revealed only an ordinal numbering intended to indicate an assembly order.
In considering marks on ancient ships, ordinal marks are very rare (Jézégou et al., 2015, p. 35).This is most probably due to the 'shell-first' construction method of these ships, which, unlike 'skeleton-first' construction method, did not require an anticipated order for fitting the frames.Each element was shaped ad hoc at the last moment.It should also be noted that the few ancient recorded ordinal marks never appear associated with names.As shown above, individual and corporation names are all abbreviated.Some questions then arise from this practice: was the location on the hull privileged?The marks are inscribed on various parts of the ship: the keelson on Saint-Gervais 3, a frame on Ses Llumetes, Arles-Rhône 3 and Saint-Gervais 3, and for the case of Arles-Rhône 3, the inside part of a wooden side plank on the starboard side at the top and the planks of a central cargo box (some of which were re-used).Were these pieces visible?Apart from the side plank on the upper part of the Arles-Rhône 3 wreck-where the mark was visible given its size, neither the keelson nor the hull frames were intended to be seen.On the Arles-Rhône 3 wreck, the six inscriptions on the walls and floor of the removable central box were only visible when it was installed, and when the ship was sailing without cargo.Were they key pieces?Structurally speaking, the keelson of Saint-Gervais 3 wreck can appear as a major part of the boat because it carries the small foremast and the wooden side plank from Arles-Rhone 3 wreck, too.The identical repetition accentuated the importance of these inscriptions, which were to be visible at a stage of construction after the wood had been transported, dried, squared and shaped, but before the ships were launched.As N. Tran points out, these markings therefore do not refer to the owners of the boats (Tran, 2014, p. 169).They were therefore important only to the person who marked them.
Two stages can then be envisaged: the storage of shaped but not yet finished carpentry parts, and their assembly.Taking into account the absence of ancient sources that could provide information on the functioning of shipyards in the Roman Period, in particular on the different trades that were practised there, comparisons must be made with later periods.A good reference are the Greek Coptic and Arab papyri of the 7th-10th centuries, even if they reveal nothing about the internal organization of Egyptian yards, they do mention the diversity of trades (carpenters, sawyers, blacksmiths, caulkers) and thus the specialization of the workers who had to coordinate their work on each ship (Bramoullé, 2019, p. 424).In the Venetian Arsenal, at the end of the Middle Ages and the beginning of the Modern Era, the organization of work is better known.The different trades were organized into corporations.
In Venice, a foreman-builder was chosen among the many foremen of the Arsenal.The shipyard was owned by a third party who was neither the foreman nor the shipowner (Lane, 1934, p. 116).The foreman-builder was in charge of choosing the wood that was marked in the forest and then transported to Venice before being sawn.A quality control of the wood was carried out before shaping (Lane, 1934, p. 150); measurements to check the shapes, especially of the frame, before assembly, were carried out afterwards (Bondioli, 2009, p. 254) without it being known whether they had been marked and, if so, of what type, symbol or person identified by name.The storage of timber was regulated from 1564 onwards.After sawing, the wood was stored in dry conditions in the various holds where several vessels were under construction simultaneously and, from the end of the 16th century, the approved and individualized equipment was carefully stored under the responsibility of a foreman according to the galleys for which it was intended.The entire outfit of a galley was grouped together in the stores and bore the same number, which was also affixed to the ship (Concina, 1993, p. 49).
The location of these marks on the vessels, on the one hand, and their low visibility, on the other hand, is probably the key to understanding their function.Specific timbers, like the mast-step, are almost permanent structural parts.The frames were intended to be covered by the ceiling planks and consequently the inscriptions are invisible.On the wreck Saint-Gervais 3, one mark is located, as we have seen, inside the limber hole which allowed water to circulate in the bilge or to the pump.A long rope, used to clean the limber hole in order to maintain perfect water circulation, was still preserved in situ (Liou et al., 1990, p. 233).This location is small, about 3-5 cm wide and high, and the inscription is stamped on it four times.This location is vital because of the care given to its maintenance.Although it is not covered by the mast-step, the inscription stamped several times on the keelson was not intended to be visible.
The location of the 'M•VIB •AVXE' brand on the wreck of La Bourse cited by N. Tran (Tran, 2014, p. 169;Gassend et al., 1982) could not be re-located by the authors.Without confirmation, the inscriptions on the frames of Ses Llumetes and Arles-Rhône 3 do not contradict this hypothesis.Although repeated several times, these marks remain small, 11 cm for the largest on Arles-Rhône 3 and on the latter, only partially visible.It seems that the tria nomina of Saint-Gervais 3, marked several times inside a limber hole, could perhaps have a propitiatory intention; the mark on the Ses Llumetes wreck, repeated four times, also seems to relate to the signature of a marine carpenter at the very moment of the installation of the piece, as B. Liou has already noted about the Saint-Gervais 3 wreck (Liou et al., 1990).This practice is found in more recent periods, with the letters 'RP' engraved at the end of a frame on the wreck of Jeanne Elisabeth dated to the mid-18th century.The marine carpenter who participated in the excavation of the wreck reported that this was a common practice in traditional wooden construction and that he himself had discreetly left his mark on several of his works (S.Bertoliatti, 2021, pers. comm.).The interpretation proposed here does not close the discussion.The very numerous inscriptions from the Arles-Rhône 3 wreck, also repeated several times but only partially visible, probably did not all respond to a single objective.Those of Anzio and Fos-sur-Mer, intended to remain visible at two different stages of their implementation, are probably, as the authors have indicated, the result of commercial and especially fiscal constraints.

Conclusion
The discovery of the marks on the wood from the wreck of Ses Llumetes has allowed us to take a closer look at archaeological evidence that is not at all common in underwater archaeological contexts.The different examples shown for the Archaic, Classical, Roman, medieval and Modern Periods, as well as the close parallels in port structures, provide us with a relevant compendium with significant potential.The division into two groups seems clear: some marks were intended to mark a process in shipbuilding, and the second group were dedicated to identifying individuals related to the structure.The classification of the Ses Llumetes markings in the second group has allowed us to expand this research, developing theories about what the tria nomina represent and analysing the marks' function based on the position they occupy within the architecture of the vessel itself.The identification of Telesphorus as an individual probably originating from central-southern Italy and the hypothesis that he participated as a carpenter in the construction of the ship allows us to consider that the ship could have been built in the Italic Peninsula.Of course, dendrological analyses will likely reveal the geographical origin of the timbers, or a more detailed architectural study will allow us to confirm this hypothesis.Note 1.For these reasons, we share the opinion of Tran (2014, p. 169), who rejects the hypothesis once put forward by Manacorda (2005, p. 27), that the marks at the circus of Arles or the RRFG of the Saint-Gervais 3 wreck are the names of redemptores, contractors linked to the timber trade.

Figure 1 .
Figure 1.Map of Balearic Islands and Ses Llumetes shipwreck location (marked with a star) at Porto Cristo (Mallorca) (Authors).

Figure 6 .
Figure 6.Distribution of Telesphorus: the map shows the areas where five or more examples of this cognomen have been located.The Ses Llumetes shipwreck is marked by a star (Authors).