• 1 We are grateful to Jan Theo Bakker, Leiden, for useful suggestions and comments, and Richard Holto (...)

1Jørgen Christian Meyer and Eivind Heldaas Seland

• 2 Seyrig, Amy and Will 1968/1975, p. 17-19, pl. 2, 7 and 21. We are grateful to Peter G. Steyn, Glob (...)
• 3 http://syrianheritagerevival.org/temple-of-bel/ (accessed 2 April 2019).

2In 2004, Meyer photographed a graffito on a column in the eastern peristyle of the temple of Bel, unfortunately without making a proper registration of the exact column and the position of the graffito on the shaft, and the lens of the camera had some dirt making a small section of the photo a bit blurry. By combining other photos of the columns in the eastern peristyle with the drawings and plans in the publication of the temple of Bel, it has been possible to identify the exact column with its characteristic cracks, and the exact position and measurement of the graffito. 2 The column is the second broken one from the southeastern corner in the eastern peristyle (fig. 1 and 2). It is one of the few columns still standing after the destruction of the temple by ISIS in August 2015, though the surface with the graffito was blasted away. 3 The graffito was located on the lower part of the shaft of the Corinthian column with shallow flutes, facing the interior of the peristyle, about 62 cm above the base of the shaft and about 1,5 m above the stylobate. As documented in the pictures from before the destruction of the temple, the graffito covers two flutes and measures about 35 cm horizontally and 17 cm vertically (fig. 3 and 4).

• 4 Dickson 1949, p. 522-523; de Voogt 2010.
• 5 De Voogt 2010.

5The terminus ante quem of the grafitto is c. ad 1100. After 1100, the courtyard of the temple was occupied by a dense pattern of mud-brick houses. To the west, the habitation reached the walls of the temple, confining the columns of the peristyle, but to the east French photos from 1919-1921 show a corridor in the eastern peristyle (fig. 5). A game board, the popular Arab game of al-Hawailah or mancala, 4 cut into the stylobate abreast of the space between the fifth and the sixth columns, may belong to the Ottoman period (fig. 1 and 6). Our gameboard represents an addition to those already known from Palmyra. 5

• 6 Langner 2001, pl. 116-146; Basch 1987, p. 373-380, 480, 483.

7The combination of the horizontal lines with three sets of vertical lines with a loop, a square and a triangle indicates that the graffito depicts some kind of ship or vessel with masts. There is no indication of any prow or stern, brailing ropes, rudders, oars, oar ports or a deeper hull below the horizontal lines. This may, however, be due to several factors. The graffito is crudely executed, and the hard material imposes limits on the rendering of at least smaller details, in contrast to painted graffiti or graffiti incised into walls covered with stucco. It may also be due to a lack of engraving skills or empirical knowledge about ships. A graffito is a very personal way of expressing oneself. Perhaps the engraver did not want to go into details or had no intention of correctly rendering such details as, for example, the exact position and number of masts, the size and shape of sails, or the proportions of the different elements. Many ship graffiti from the ancient period are poorly executed and some of them are insufficient for an exact classification of the vessel, irrespective of the material used as the base for the graffito. 6

• 7 Langner 2001, no. 2019 (Pompeii), 2044 (Ostia), 2045 (Pompeii), 2202 (Ostia); Bagnall et al. 2016, (...)
• 8 Bagnall et al. 2016, no. 92, 184, 342, 391 (Smyrna); Langner 2001, no. 1861, 1910, 1953, 1990 (Nym (...)
• 9 Bagnall et al. 2016, no. 176, 177, 178, 184 (Smyrna); Langner 2001, no. 1959, 1990, 2033, 2060 (Ny (...)

8With all these limitations in mind, what kind of vessel did the engraver want to convey either to himself or to a spectator? At first sight the lengthy parallel lines suggest a long ship with several masts, and a yard on the central mast. The triangles might be an indication of a bulwark. The loop on top of the right mast might be a crow’s nest, though it ought to have been placed on the top of the yard, where it would normally stand. 7 Another possibility is that it might be a merchant ship. The parallel lines would represent the tiers of planks above the waterline, as is normal in many graffiti. 8 The absence of a hull, however, argues against this interpretation. In almost all graffiti, cargo vessels are depicted with rounded hulls below the water line, indicating the capacity of the ship. The absence of a prow and a stern, and the position of the masts, especially the right mast, which is incised with a double line indicating an important function, do not fit either with a long ship or a merchant ship. Normally a double line is used at the central mast with the main yard. 9

• 10 Place 1867, pl. 43; Ritter 1919; Naval Intelligence Division 1944, p. 558; Rollinger 2013; Rolling (...)
• 11 Ritter 1919, p. 141-143.
• 12 Ritter 1919, p. 141-143.

9We may attribute all these peculiarities to the engraver’s strong subjectivity. There is one type of vessel, however, where some of them make more sense. Our suggestion is that the engraver did not want to convey a seagoing vessel, but a vessel used for river transport, a so-called kelek (fig. 7, 8 and 9). Keleks were square rafts, hence without prow and stern, made of inflatable goat-skins mounted in a wooden frame and covered with brushwood and reeds. The design, unchanged in its basics, was used for river transport in Mesopotamia at least from the Neo-Assyrian period until the first half of the 20th century, and is also well attested in the Hellenistic and Roman periods. 10 The rafts would be used for downriver transport. Upon reaching their destinations they were dismantled. The floating-skins were emptied, dried, and preserved, and brought back upriver by animal transport. 11 A kelek could be of considerable size (fig. 8), and rafts with up to 400 skins were attested in the early 20th century. 12 The craft depicted in our graffito has the shape and proportions of a kelek, and the horizontal lines resemble the frames of a raft like this.

• 13 Seland 2017, p. 111.

10Two triangles, pointing up, are depicted just left of the middle mast. We suggest that these represent the cargo of the craft. This is based on two parallels: (1) a ship-graffito from Dura-Europos, probably depicting a river-craft and with a terminus ante quem of ad 256, showing similar triangles along the rail (fig. 10), and (2) a second graffito from Dura showing a camel caravan, where the cargo has been indicated by triangles (fig. 11). The latter example is in need of explanation. Superficially the graffito seems to show two-humped camels without cargo, but tethered to each other in a caravan. 13 A closer look at the third animal in the line, where the lines of the foremost triangle cover part of the neck of the animal as well as parts of the rear triangle, reveals that the triangular forms are not humps, but bales of cargo mounted on each side of a single hump (hidden from sight). It thus appears that the triangles next to the mast of our suggested kelek show a standard way of depicting cargo in the region in the mid third century.

• 14 Naval intelligence division 1944, p. 168-171.
• 15 Ritter 1919, fig. 42.
• 16 Ritter 1919, p. 141-143.

11An argument against our interpretation is that according to ethnographic accounts and depictions masts do not seem to play an important role for navigation by kelek. It is significant, however, that our craft has three masts, if that is indeed what they are, and that the artist took pains to depict them differently. The one to the left is crowned by a small triangle pointing right. This seems to be a flag or banner, probably helping visibility between crafts and/or signalling ownership. Similar banners were mounted on 19th century keleks on the Tigris (fig. 8). The middle mast has what we interpret as a small square sail, which might have helped propulsion after the Euphrates straightens out and slows down below Hit, taking advantage of the northwesterly winds that prevail in Mesopotamia through most of the year. 14 The use of small sails, although apparently not widespread, is ethnographically attested. 15 The right vertical double line with a loop at the top is the most puzzling. To have a mast with a crow’s nest does not make sense in relation to river transport. A suggestion is that it represents one of the steering oars (large crafts in the modern period carried six), 16 with their large, round, paddles (fig. 8 and 9), though the vertical position and the double line are peculiar, or perhaps it could depict a person. The interpretation remains an open question.

• 17 Schmidt-Colinet 2016, also “Addendum,” below; Schmidt-Colinet 2005, p. 91, fig. 150; Seland 2017, (...)
• 18 Seyrig 1963, p. 161-166.
• 19 Gawlikowski 1983; Gawlikowski 1988; Rollinger and Ruffing 2013, p. 408-410; Seland 2016, p. 47-51.

12Ancient Palmyrenes were cosmopolitans. Many of them would know ships of the kinds that plied the Mediterranean, the Persian Gulf, the Red Sea, the Nile, and the Indian Ocean. As the graffito of the lighthouse of Pharos recently published by Schmidt-Colinet and the ship on the famous relief from the Marona tomb / Qasr el-Hayye (Palmyra Museum A24/1226) demonstrate, 17 we should not be surprised to find maritime motives in the iconographic record of the desert city. Nevertheless, it seems fitting that a depiction of the kind of craft most familiar to the Palmyrenes should be incised on a column in the city’s main temple. The importance of the kelek to Palmyrene trade was first recognized by Seyrig on the basis of an inscription from the colonnaded street dedicated by the association of floating-skin makers (askonautopoioi), whose products may have served no other conceivable purpose than riverine navigation in the otherwise arid environment of the Syrian Desert (IGLS 17.1.59). 18 The argument has later been supported and elaborated by others. 19 In all likelihood, Palmyrene caravans travelled to the Euphrates, where goods were transferred to rafts, only to be unloaded at ports in Middle and Southern Mesopotamia. Besides the inscription mentioned, the reconstruction builds on the ethnographically attested use of kelek in modern times and on the functional argument that riverine transport was less strenuous and expensive than a c. 30-day/600-kilometer-long caravan haul between the middle Euphrates and Southern Mesopotamia. If our interpretation of the graffito from the temple of Bel is correct, it adds an iconographic element to the epigraphic, ethnographic and functional evidence of Palmyrene trade already cited.

• 20 For the excavation of Tomb no. 173, see al-Asʿad 2013; for the graffiti, see Schmidt-Colinet 2016.
• 21 Schmidt-Colinet 2016, p. 158-161, fig. 7-11.
• 22 See Seland 2016.

14Graffiti detected in 2010 on the third of the front steps (from below) of Temple Tomb no. 173 at Palmyra (fig. 12) are a welcome supplement to the graffito discovered in the Temple of Bel. 20 Two of these graffiti from Tomb no. 173 can be interpreted as representations of a lighthouse and a ship at its side (fig. 13). For several reasons, the lighthouse can be identified as the famous Pharos of Alexandria. 21 The ship at its right side shows a rectangular sail and the top of the mast just over the sail. The rectangular form of the sail as well as the vicinity of the Pharos make clear that this representation depicts a ship that can cross the sea, not a boat used on a river (like the kelek described above). Thus, this graffito is another proof that the Palmyrenes did know “Ships of the Desert and Ships of the Sea.” 22 Together, the two graffiti are rare visual proofs documenting the otherwise well-known far reaching economic and cultural connections of ancient Palmyra.