• Carmel coast;
  • iron anchors;
  • shipwreck;
  • Ottoman;
  • Mamluk;
  • bombard


  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References

In 1977, a shipwreck assemblage was discovered off Atlit, Israel. Finds included three iron anchors, a large bombard, four swivel guns, stone and lead shot, and bronze helmets. The bronze bombard (2210 kg, 3.247 m long) contained a wooden wad and remnants of what may have been gunpowder. The swivel guns (each 185 kg) were bronze, with swivels and the tillers of iron. Twenty (or 21) bronze helmets were recovered. The findings provide rare evidence for the mounting of heavy ordnance on the bow of a galley or ship in the 15th century.

In December 1977, Mr S. Shapira, a member of the Israeli Association for Underwater Archaeology, reported spotting metal helmets, an iron anchor and stone shot while diving in the vicinity of the Atlit North Bay (Fig. 1). One of the helmets, presently displayed in the National Maritime Museum, Haifa, was retrieved, studied and published as an Assyrian helmet (Radan, 2002). On 3 September 1985, underwater survey of that coast carried out by the University of Haifa, headed by the first author with volunteer divers, discovered that the sand covering the site had shifted. The anchor previously reported was exposed, as well as a concentration of finds c.3 m to its south, including a large bombard, several swivel guns, stone shot, and metal helmets (Fig. 2). A routine underwater survey carried out by the same group a year later, in an area some 150 m south-west of the site, discovered two large grapnel anchors which are thought to be related to the assemblage.


Figure 1. The location of the site. (EG)

Download figure to PowerPoint


Figure 2. Distribution of the finds on the site. (EG)

Download figure to PowerPoint

An underwater excavation lasting seven days was carried out at the site by a team from the University of Haifa and volunteer divers during September 1985. The loose sand covering the site was removed using a dredging system operated by a portable water pump placed on a boat. The finds were located with the help of a metal detector, measured, photographed and recorded in situ. Heavy artefacts were lifted by airbags and dragged ashore; smaller finds were dug, lifted and brought ashore manually, and transported to the University of Haifa facilities at the Israel National Maritime Museum at Haifa.

The site

  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References

Underwater excavations and surveys in the Atlit North Bay, carried out over the past 45 years, have revealed numerous remains of shipwrecks, cargoes, anchors, harbour installations, isolated finds and submerged prehistoric settlements (Galili et al., 2002). Archaeological remains attest to extensive maritime and coastal activities in the area during the past eight millennia. The site discussed here (Fig. 1) is located 10 km south of Haifa (34:56′ E, 32:42.5′ N), at a water depth of 2.5–3 m, some 35–50 m off shore. The artefacts defining the site were scattered on top of a dark clay matrix, filling the trough between a submerged sandstone ridge and the shore. The clay and the archaeological material have been covered, most of the time, by a thick (1–2 m) layer of quartz sand. Waves and sea currents have occasionally removed the sand layer and exposed the artefacts. The main concentration of artefacts was in an area 15 x 20 m (Fig. 2), suggesting that they came from a distinct shipwreck, as discussed below.

The iron anchor was located on the north side of the concentration and marked the site's northern border. It was lying on an E-W axis with the cable ring pointing east. A few metres south of the anchor there were four swivel guns, generally lying on a SW-NE axis (Fig. 3). The muzzles of three of them were pointing north, while one was pointing south. South of the swivel guns lay the massive bombard on a N-S axis with the muzzle pointing south (Fig. 4). A few metres west of the bombard were several groups of bronze helmets, some stacked one inside another. Several stone shot were located around the bombard. The two iron grapnel anchors found some 150 m south-west, seaward of the main artefact concentration, are thought to be associated with this assemblage as discussed below. No other objects and no hull remains were found on the site.


Figure 3. The swivel guns as found on the sea-floor. (EG)

Download figure to PowerPoint


Figure 4. The bombard and a stone shot as found on the sea-floor. (EG)

Download figure to PowerPoint

The finds

  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References


Three iron anchors were recovered. A single two-armed anchor was located with the main concentration of artefacts (Fig. 2). The total length of the anchor is 2.9 m and it weighs c.320 kg. Its two arms span 1.855 m and end in triangular palms (Fig. 5). The head of the shank is pierced, holding an iron ring, 0.55 m in diameter. The shank cross-section is square while the ring was manufactured from a rod with a round cross-section. On the shank, about 150 mm below the ring, there are four nuts for the attachment of the wooden stock, which has not survived.


Figure 5. The two armed iron anchor. (Ben-Yehuda S. and EG)

Download figure to PowerPoint

Two iron grapnel anchors (Fig. 6) were found some 150 m west and south-west respectively from the main artefact concentration and were left in situ. The four-armed grapnels are c.2.2 m long and lay c.60 m from each other, at a water depth of 5.5 m. One of the arms of the southern anchor is broken (Fig. 6b marked with a black ‘a’). They are covered by sand for most of the year and, when found, the small sections of the arms and palms which protruded from the bottom were covered by torn fishing nets.


Figure 6. a) Grapnel anchor in situ; b) Grapnel anchor with a broken arm marked by an ‘a’. (Galili J.)

Download figure to PowerPoint

Bronze bombard (perrier) (IAA 1988–743)

The terminology used for the different parts of the naval ordnance is mainly based on Roth (1996) and Manucy (1992). The bronze bombard weighs 2210 kg and is 3.247 m long (Figs 7, 8). The reinforce is slightly conical, while the chase has an almost imperceptible taper. Starting from the breech ring and working toward the muzzle, it is roughly divided into a single reinforce and a chase. Its outer diameter at the breech is 0.46 m and at the muzzle 0.43 m. The breech base is flat, without a cascabel. It is circumscribed by a base ring fillet in the shape of three stairs. Two opposing square-based protrusions, plano-convex in section, are placed about 0.17 m from the base. The touch hole, protected by a semi-circular vent pan, is situated on the top of the reinforce, just behind and between the two protrusions. The reinforce measures 1.35 m, the chase 1.897 m and the muzzle 0.16 m long. Between the reinforce and the chase there is a reinforce moulding including an astragal fillet. About 0.23 m from the end of the reinforce there is a chase astragal fillet. A third astragal fillet is situated at the end of the chase before the muzzle hoop. About 2.66 m from the breech base there are two opposing lugs into which lifting rings are inserted. The muzzle is decorated by a muzzle hoop. The interior of the bombard is divided into a chamber, 1.045 m long with a diameter of 127 mm, and a bore, 2.045 m long with a diameter of c.281 mm.


Figure 7. The bronze bombard. (Roth R. and EG)

Download figure to PowerPoint


Figure 8. The bombard and the stone shot, scale is 0.5 m. (EG)

Download figure to PowerPoint

The composition of the alloy used to cast the bombard was subjected to Inductively Coupled Plasma (ICP) analysis (I. Segal pers. comm., unpublished report 1995), the results of which are given in Table 1. It proved to be cast in tin bronze. The alloy is made of copper with some tin and insignificant amounts of lead and arsenic.

Table 1. Metallic composition of the bombard (I. Segal, unpublished IAA report 24.7.1994)
IAA NoTest NoTin (SN) %Arsenic (As) %Lead (Pb) %Zinc (Zn) %Copper (Cu) %

Wooden wad

When the bombard was examined after lifting, the mouth of the chamber was tightly sealed by a wooden wad, c.155 mm long, with a maximum diameter of 244 mm (Fig. 9). The disposable wad, intended to be wasted on firing, was carefully extracted. Judging by the 16 concentric visible tree rings, it was made from the centre of a branch or a young tree c.16–20 years old. The natural centre of the tree is located in the centre of the wad. It was crudely formed, most probably using an adze, into a polygonal, almost round-sectioned, conical plug. The wood was identified as Pinus sylvestris L. a non-local species of pine (Liphschitz N. pers. comm.) which is common in Europe and Anatolia (Atalay and Efe, 2012; Liphschitz and Biger, 1992). A sample taken from the outer surface of the wad was radiocarbon-dated to AD 1312–1353 with 40% probability and AD 1384–1424 with 51% probability, or AD 1312–1424 with probability of 100%; or an age of 553±53 years BP calibrated (sample RT 1593 Weitzman Radio-dating Laboratory; Carmi I. and Segal D. report from 23.3.1992; Segal and Carmi 2007: 85). The dendro-chronological calibration was carried out according to Stuiver and Pearson (1986), and Niklaus (1991) allowing the radiocarbon-date to be recalibrated to AD 1298–1440 at 2 σ with 95% probability (Carmi I. and Segal D., Unpublished report 20.03.2013; calibration according to Bronk Ramsey, 1995; 2001).


Figure 9. a) The wooden wad (Ben-Yehuda S.) b) The wooden wad. (EG)

Download figure to PowerPoint

Gunpowder residue?

The material sealed by the wad was extracted, analysed visually by the authors and subjected to ICP analysis (I. Segal unpublished report 1995, job number 53). The material contained about 40% silicon dioxide, visually identified as local sea-sand. The considerable amount of sulphur in the examined samples could indicate the presence of traces of gunpowder, originally composed of charcoal, sulphur and potassium nitrate.

The presence of sea-sand indicates that the chamber was not well sealed, and that minute particles and solutions could leak in and out of the chamber. The presence of sulphur alone would not normally be sufficient to prove the presence of a charge in the chamber: sulphur residues could have remained from the previous firing and sulphur exists in coastal sediments in various forms and could have entered the chamber with the sand. However, charcoal would not be expected to have remained in either of these scenarios: if the chamber was loaded, the charcoal, in the form of a fine powder, would have leaked out and floated away, while charcoal remains from previous firings would most likely have been completely consumed. The third component, potassium nitrate, is water soluble, and would have dissolved and washed away. In this instance, we consider that the traces of sulphur and the in situ wad are sufficient evidence that the gun was loaded when the ship sank.

Stone shot

Thirteen round shot were retrieved from the site during the excavation and examined. It is known that at least 18 round shot were present at the shipwreck site, of which four were stolen by treasure hunters and one more was brought to the National Maritime Museum by an amateur diver. Of the 13 examined, two were limestone and 11 were basalt (Table 2, Fig. 8). Their diameter ranged from 271 mm to 275 mm. The diameter of ten out of the 13 is 273.8 mm, which probably represents the desired diameter of the shot. Their weight ranged from 27.1 kg to 28.6 kg. The diameter of all the stone shot was less than the diameter of the bore (281 mm) and thus they could have been loaded into the bombard. Given the size of the ten standard shot, the desired difference between bore and shot diameter (windage) was about 7 mm.

Table 2. Properties of the stone shot
No.Weight (kg)Type of stoneDiameter (mm)

Swivel guns

Four swivel guns (Figs 2, 10-12), each weighing 185 kg, were retrieved from the site. The best preserved gun is described and illustrated here, on the assumption that all four were similar and as the other guns are no longer available for further research. The barrel, 2.19 m long, is made of bronze while the swivels (yokes) and the tillers (tails) are made of iron. The bore is 50 mm in diameter and 2.09 m long. The chase is slightly tapered and conical. An iron tiller is inserted into a square hole formed in a lug extending from the breech ring. The tiller is flat, with an elongated slit in the middle. It has a rectangular cross-section at the end, where it is inserted, and is evenly tapered toward the far end. A reinforce circumscribing the barrel extends 150 mm from the breech base. On the base of the breech there is a circular hole, 70 mm in diameter, filled with bronze. Apparently, by sealing the breech of the bore, a chamber was created. Two rectangular holes are seen on opposing sides of the reinforce which are filled with bronze cast in place. After casting, the bronze was hammered like a rivet. The cast bronze sealed the breech, thus creating a muzzle-loading cannon. A touch hole is drilled above the beginning of the bore. Two cylindrical trunnions are found some 150 mm back from the middle of the gun, excluding the tiller. The trunnions are set in an iron yoke with a slightly conical peg. The end of the chase, before joining the muzzle, is decorated by a circumscribing fillet. The expanded muzzle hoop is decorated by three circumscribing fillets.


Figure 10. Swivel gun. (Roth R. and EG)

Download figure to PowerPoint


Figure 11. Swivel gun. (EG)

Download figure to PowerPoint


Figure 12. Isometric reconstruction of a swivel gun from the Atlit wreck. (Ben-Yehuda S.)

Download figure to PowerPoint

The body of the swivel gun studied was cast in bronze (Table 3). The alloy is made of copper with some tin, insignificant amounts of lead and arsenic and no zinc. The filled holes are similar to key holes found on breech-loading swivels with separate chambers. Perhaps the four swivel guns were converted breech-loaders.

Table 3. Metallic composition of the swivel guns (I. Segal, unpublished IAA report 18.5.1995)
IAA NoTin (SN) %Arsenic (As) %Lead (Pb) %Zinc (Zn) %Copper (Cu) %Remarks
1908-7478.100.450.4589.80whole iron handle
1908-7488.420.630.8990.00broken iron handle

Lead shot

Two lead shot, 50 mm in diameter, each weighing c.1.2 kg, were recovered from the site. They could have been used as ammunition for the swivel guns.

Bronze helmets

Altogether 20 bronze helmets were recovered from the wreck-site. Some were found in three stacks, each holding 2–3 helmets, and several others as individual helmets (Fig. 13). An additional, heavily corroded helmet was handed over to the IAA, apparently originating from this site. These 21 helmets are described and discussed here (Tables 4, 5).


Figure 13. left The bronze helmets (EG); right One of the bronze helmets. (EG)

Download figure to PowerPoint

Table 4. Nominal metallic composition of the helmets
Helmet No.Test No.Sampling surfaceCopper (Cu) + trace elements %Zinc (Zn) %Arsenic (As) %Lead (Pb) %Tin (Sn) %Sample location
Table 5. Properties of the helmets
Helmet No.Diving report No./IAA No.Weight (g)/% lost by corrosionHeight (mm)Circumference at base (mm)Tip buttonNo. of holes: Ø (mm)Thickness (mm)Notes
 1HY 257



IAA 404/6–91


200 (210 before deformation)690+1.4–2.5Tip button deformed

IAA 404/8–91



240668+2: 6 mm1.5–2.2 

IAA 404/2–91


220680+2: 4 mm1.5–2.5 

HY 257

IAA 404/2–91



220695+2: 4 mm1.4–2 

HY 257

IAA 404/1–91



210 (possibly more before corrosion)643?1: 6.5 mm1.5–2.5 

IAA 404/10–91



213 (possibly more before erosion)670?1.5–2.5 

IAA 404/11–91



210 (23 before deformation)695+2: 6.5 mm0.6–1Very corroded

IAA 404/2–91



1856302: 7 mm/4.5 mm1–1.5 
10IAA 1637/653660?690??1–2Three nested helmets
11IAA 404/9–915%?680??  
12 2107002: 5.5 mm  
13IAA 404/9–913280???1: 4.5 mm1–2.3Three nested helmets
15 6802101: 6 mm  
16HY 2572000???1–1.5Two nested helmets
18HY 257950220620+2: 5 mm/7 mm0.5–1 
19940218685+2: 7 mm1–2.2 
IAA 404/4–912%      
201050230665+2: 4.5 mm1.8–2.9 
IAA 404/7–91       
21 265     Private collection
22       Haifa Maritime Museum
23       Private collection

The site has been exposed by storms several times in the past, and information received from local fishermen indicates that at least two additional helmets have been retrieved. One of these was illegally taken by treasure hunters and another was donated by a diver to the National Maritime Museum, Haifa.

All the helmets are bullet-shaped with a circular cross-section. Although they look similar, their dimensions and proportions show that it is a heterogeneous assemblage (Table 5). Their height ranges from 170 to 220 mm and their diameter from 195 to 225 mm. The intact, less-corroded ones each weigh about 1.020–1.070 kg. Some helmets were heavily corroded or have concretions attached, thus their weight is not representative. Nine of them are adorned with a flattened, mushroom-like top-piece, while others lack such decoration. Nine have two opposing holes on the bottom rim, c.5 mm in diameter. The holes were punched from the outside and the edges carefully hammered flat on the inside. These holes must have had some function such as securing inner cushions or chin straps and could indicate that some of the helmets were in use.

The elemental composition of five of the helmets was analysed by XRF. Some of the tests were taken from polished surfaces while others were taken from corroded surfaces (Table 4). The helmets were made of an alloyed copper with a low percentage of tin (up to 4.5%). A sample taken from the unpolished point of helmet no 1 (test no. 180) shows it to contain an exceptionally high percentage of tin, 11.7%. The possibility that the point of some of the helmets was made from a separate piece of metal that was then added to the main body was checked by X-ray (S. Shalev pers. comm. 2011, unpublished report), but no signs of connection or welding between the tip and the body were detected.

The tin percentage in the helmets is low relative to bronze alloys in general, which contain up to 20% tin. The analysis indicates that the composition of the bronze helmets varies, though they were all made of an alloy relatively poor in tin. Arsenic, zinc and lead may have been introduced unintentionally in scrap metal used in the cast mix for the metal sheets.

Such a copper alloy, with a low percentage of tin, is relatively soft (HV 65) and is suited to processing by hammering. Signs of shaping by hammering can be seen in the interior of many of the helmets. The helmets were formed by heating and hammering sheet metal on a mould, either into a cavity or over a male form. The helmets are symmetric in section; therefore missing parts could be estimated from the complete sections.

Lead ingots

Two lead ingots, weighing 3 and 4 kg, were found lying c.2 m south of the two-armed anchor, and were recovered. They are crudely formed in a plano-convex shape. They were covered with corrosion products to which local materials have concreted. XRF tests suggest that the ingots are made of lead with some traces of copper (S. Irena, pers. comm., unpublished report 20.2.1994).


  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References

Site formation and wrecking

The site lies on the Carmel coast just north of Atlit North Bay, an important natural ship shelter providing anchorage for sea-going vessels from the Middle Bronze Age on. Archaeological evidence attests to an almost continuous maritime activity there. The Carmel coast further north lacks natural shelters for ships. Storms have regularly trapped ships there over the past millennia, where they have been crushed in the breaker zone (Galili et al., 2009). Analysis of the composition of about 200 ancient shipwreck sites along the coast has demonstrated that the remains of ships wrecked in the surf zone were generally separated into three main classes by the action of the sea: 1) the hull disintegrated and wooden parts drifted ashore; 2) people and livestock also drifted ashore, along with light cargoes, wooden parts and objects firmly attached to them; 3) heavy metallic or stone objects sank into the sediment during the storm or soon thereafter, accumulating on the substratum (clay or stone) under the sand, at depths of 2.5–4 m. In most cases amphoras were broken by the surf. Therefore only the heavy objects remained at wreck sites.

In stormy sea conditions in this region, any act of jettisoning should have occurred in deeper water further off shore and away from the coastal breakers zone. Once in the breaker zone, jettisoning was useless and the people aboard would, by that point, have been aiming to save their own lives as it would have been clear that the ship was lost (Galili et al., 1986; Galili et al., 2011). Thus the distribution of heavy artefacts leads us to interpret the assemblage under discussion as a shipwreck rather than a group of jettisoned artefacts.

The ship was likely wrecked while sailing along the Carmel coast or anchoring north of Atlit North Bay. It is likely that some of the cargo was composed of materials that floated away from the site. A less probable scenario, but still a possibility, is that the ship was the victim of hostile action.

The location of the grapnels, south-west of the main artefact concentration, suggests that they were the working anchors, cast prior to the wrecking. This is also indicated by the orientation of the grapnels: the shanks and cable rings are pointing in the general direction of the main site. The position of the grapnels suggests that the ship was using them when hit and driven ashore by a SW-W gale, which is the dominant direction for gales and storms in the region. The broken arm of grapnel B suggests that it may have suffered a traumatic shock, possibly caused by the ship being driven ashore by the breakers. Typologically the grapnels could not be associated with any other known wrecks in the vicinity. The position of the two-armed anchor, within the main concentration of artefacts, indicates that it was on board when the wrecking occurred.

Post deposition processes in the past century, including sand quarrying and the construction of marine structures along the Carmel coast, have created a shortage of sand and changed patterns of coastal sedimentation. Wide areas of sea-floor have been uncovered and, consequently, hundreds of sites, including the one presented here, have been exposed and discovered.

Dating the wreck

Shipwrecks may be dated by radiocarbon analysis of organic materials securely associated with the ship, coins, inscriptions (if available) and the typology of the artefacts recovered.

The wad

The pine wad is the only wooden object that can be securely associated with the wrecking event through its position within the bombard, where it was protected by copper salts which have prevented biological contamination. As a disposable item, it is unlikely to have been of a great age when the ship was wrecked. Radiocarbon dating found that the wood grew AD 1384–1424, recalibrated to AD 1298–1440 within 2 σ with 95% probability. Judging by the nature of the wad it could have been produced from a young tree trunk or a branch with a suitable diameter which would have required minimal work, and thus little waste. It is less probable that the centre of an older and thicker tree was used for its manufacture. Such a branch or narrow trunk is unlikely to have been left long in a shipyard (probably no more than 30 years) or on board ship before being used. Thus the wrecking event is likely to have occurred before the middle of the 15th century. The date range of AD 1298–1440 provides a terminus ad quem in the mid 15th century, with a high probability that the ship was wrecked within that range.


The earliest-known archaeological example of a grapnel is the small anchor recovered from the Yassi Ada Byzantine shipwreck, dated c.AD 625; it probably belonged to the ship's boat (Bass and van Doorninck, 1982: 132). Michael of Rhodes, the Venetian 15th-century seaman, described and drew five grapnel anchors with four arms (dated 1434–1436, with later additions in 1437 and 1444), which were carried by a Venetian galley of his era (Fig. 14) (Long et al., 2009, vol. 1: 324, 325; Long et al., 2009, vol. 3: 176). When exactly large ship's grapnels came into use is not clear, but the earliest conclusive evidence is from the first third of the 15th century, although they may go as far back as the mid 14th century (C. Pulak pers. comm.). This type of anchor was often carried by Mediterranean ships, especially galleys in the Middle Ages.


Figure 14. A Venetian galley under sail carrying a grapnel anchor. (After Long et al., 2009, vol. 1, fig. 154b)

Download figure to PowerPoint

Two-armed anchor

The orientation of the two pairs of nuts used to attach the wooden stock is unusual, being parallel to the arms rather than perpendicular to them. This suggests that the anchor is no later than the 16th century. Based on its size, shape, and its manufacture technique, it is unlikely to be earlier than the 15th century (C. Pulak pers. comm. 2013).


The nature and date of the bombard were appraised by L. Tarassuk of the Metropolitan Museum of Art in New York and R. G. Ridella of the Istituto di Storia dell'Europa Mediterranea. Tarassuk identified it as a ‘stone-throwing cannon’ or stone cannon, cast in the 15th–16th centuries. According to him, this specific cannon could have been in use during the years 1425–1450.

A similar cannon, dated to 1420–1430, exists in the Germanisches National Museum (Essenwein, 1877: plates XXIV–XXV) (inventory number W571). It was donated to the Germanisches Museum in 1869 by the Ottoman Sultan Abdülaziz and was part of several similar big artillery pieces from the island of Rhodes (Thomas Eser pers. comm., Germanisches National Museum).

No precise equivalent to the Atlit bombard has yet been located. However several bombards with similar characteristics have been identified. The Atlit bombard is characterized by a pair of rectangular rear lugs, placed near the breech, and a pair of protrusions fitted with rings positioned before the muzzle. Such a combination was already known during the reign of Sultan Mehmet the Conqueror (1451–1481) (Turkish Military Museum, Askeri Müzesi, Istanbul inv. N. 657) and continued in the reign of Sultan Suleiman the Magnificent (1520–1566) (A. M. inv. N. 6 and 7). The typological characteristics of the bombard may be dated to around the first half or the mid 15th century. The radiocarbon-date of the wad could suggest an even earlier date, which would make this a significant piece.

Swivel guns

According to Tarassuk the four small bronze swivel guns were widely used from the 15th-century onwards on board ships. The swivel guns look like they have been converted from breech-loaders to muzzle-loaders.

Given all the above it seems that the wrecking event can be dated to the first half of the 15th century. However a late 14th century date is also possible.

Ordnance aboard

The artefacts discovered, and their distribution on the sea-floor, are consistent with the identification of the site as the remains of a ship carrying a bow bombard with four flanking swivel guns. The location of the anchor in relation to the guns contributes to this interpretation. This is a typical armament for a Mediterranean galley in the 15th and 16th centuries, but it could also be the armament for an early armed sailing ship.

In general, guns were carried aboard ships and galleys from the 14th century onwards. It has been suggested at that early period heavy guns were intended for use on shore rather than afloat (Guilmartin, 1974: 295, 296) as wrought-iron ordnance was very expensive, scarce and more useful in siege warfare than aboard ships. It was considered not to have been placed aboard ordinary galleys for use at sea until late in the 15th-century (Guilmartin, 1974: 295). However, according to written documents, the Royal galley of Alfonso the 5th of Aragon had two bombards mounted on board in 1418 (Guilmartin, 2007: 662).

The artillery found in the Atlit wreck was most probably active weapons rather than cargo. The presence of swivel guns, which were not typical of shore-based weaponry, is further evidence that the ship was armed, as does the wad in the bombard, indicating that the bombard was probably loaded. The presence of stone and lead shot also supports this contention. The spatial distribution of the bombard and the swivel guns on the sea-floor indicates that they were mounted on the open prow of a vessel, probably a galley. Evidence from the second half of the 15th and the first half of the 16th centuries suggests that such an arrangement, a main centreline bow gun flanked by swivel guns, existed on Mediterranean galleys (Guilmartin, 1974: 295). Evidence for the mounting of effective, heavy ordnance on the open bows of east Mediterranean galleys in the first half of the 15th century is rare. Constantine (1984: 206) noted a very early document recording some bombards and 60 stone shot on a Catalan ship going from Barcelona to Alexandria in 1394. A German woodcut of 1486 depicts a panoramic view of the port of Venice with two war galleys, each carrying a large bombard mounted as a centreline bow gun; these galleys appear to be laid up for repairs, thus small swivel guns could have been dismounted and taken ashore (Guilmartin, 1974: 295–9). The set of four swivel guns and a centreline bombard recovered from the Atlit shipwreck is similar to later sets of guns mounted on the prows of 16th-century Genoese, Sicilian, Venetian, Neapolitan and Spanish galleys (Guilmartin, 1974: 295). It is reasonable to conclude that the five guns from Atlit were mounted similarly on the prow of a ship or a galley (Fig. 15).


Figure 15. Possible reconstruction of the Atlit guns mounted on the prow of the ship. (After Guilmartin, 1974: 298)

Download figure to PowerPoint

Later, in the second half of the 15th century, and almost until modern times, ship's guns were commonly operated from a gun deck and mounted to port or starboard. It has been suggested that in the 15th–16th centuries, the operation of heavy cannon, the first artillery used at sea, from an open bow of a galley was much easier than from a confined and crowded gun deck (Guilmartin, 1974: 295). Stone-throwing cannon had disappeared from Spanish galleys almost completely by the beginning of the 16th century (Guilmartin, 1974: 295). Whatever the nationality of the Atlit wreck may be, it appears to provide early archaeological evidence for a stone-shooting bombard and swivel guns mounted on a ship or galley.


Helmets similar in shape to those recovered from the wreck-site were commonly used in the region during the 14th–16th centuries, thus they do not provide a precise date for the wreck. Several similar artefacts, though decorated and complete, are known from European, American and Near Eastern collections; for example an Ottoman Turkish çiçak can be found in the Topkapı Museum in Istanbul (Robinson, 1989: 62–63), and an another in the Metropolitan Museum of Art in New York (Heilbrunn Timeline of Art, 2006). Judging by the parallels mentioned, the recovered helmets appear to be of a Muslim origin, possibly Mamluk. Some of the helmets were stacked together, and none have attachments facilitating immediate use. This may indicate that some of them were not personal items belonging to active warriors, but rather components that had to be finished prior to their use in battle. Finishing could have been executed in any port by a local tinsmith, or even by one on board. Possibly nasal and cheek pieces, as well as a neck guard, cushioning and chin straps, would have been joined with rivets. Without indicative decorations, accessories and added parts, pinpointing an origin and date beyond the eastern Mediterranean in the 14th–16th centuries would be overly speculative.

Historical background and nationality

It is hard to identify the ship's nationality. At the end of 14th century and the first half of the 15th century, the estimated date of the wreck, the Levant mainland was under the rule of the Mamluks (Fig. 16). It is traditionally accepted that they had little interest in nautical affairs (Ayalon, 1967). Their armies were essentially trained to fight as armed horsemen. Their primary goal was to prevent the return of the Crusaders by warding off capture of coastal fortifications which could be used as a foothold on the east coast of the Mediterranean. To achieve this goal, the Mamluks resorted to a ‘scorched earth’ policy as a defensive measure, systematically destroying coastal facilities such as ports, anchorages and fortifications (Ayalon, 1967; Fuess, 2001; Terdiman, 2007). The coast was considered not as a base but rather as a border, a ‘no man's land’; although at times watchtowers were built on the shore so as to give warning of attack.


Figure 16. The Eastern Mediterranean during the mid 15th century. (After McEvedy, 1992; Kinder and Hilgemann, 1964–1966).

Download figure to PowerPoint

Mamluk ships were built for a specific job: to transport armies (Ayalon, 1967). After use, these ships were usually neglected, they rotted, their crews were dispersed, and no naval tradition was established nor know-how accumulated. The Mamluks tried to substitute political treaties with the European powers for the sea power they lacked. This never worked satisfactorily. Christian pirates continued to attack and plunder Muslim merchant ships, even when anchored in ports ruled over by the Mamluks, such as Beirut and Sidon in 1403 (Fuess, 2001). Cyprus remained a Christian island even after 1426 when the Egyptian Mamluks of the Burji dynasty attacked it successfully. The Mamluk victory there, under the command of Sultan Al-Ashraf Barsbay (1422–1438), shows some ability to conduct war involving action at sea, but that ability was not developed. The Christian rulers of the island paid yearly taxes and promised to prevent raids by Christian corsairs, but these promises were not fulfilled. Some captains of the well-equipped and expertly manned navies of the Italian coastal cities, while trading with Cyprus and the coastal Muslims, also engaged in piracy. Toward the middle of the 15th century, coastal raids by the Knights Hospitallers, from a base in Rhodes, added to the increasing attacks by Christian pirates from the western Mediterranean. As a result the Mamluk Sultan Jaqmaq (1438–1453) attempted to fight the Knights of Rhodes, but to no avail (Fuess, 2001).

So the wrecked ship could have been a European ship, perhaps from one of the Italian or Iberian port-cities, or a pirate ship on a raid in the eastern Mediterranean. Less probable, but still possible, is that the assemblage represents a Mamluk naval vessel; in which case it could inform us about the Mamluk navy, of which we know so little.


  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References

We wish to thank S. Shalev for the analysis of the helmets' metal; L. Tarassuk, T. Eser and R. G. Ridella for seminal comments on the ordnance; C. Pulak for commenting on the ground tackle; N. Liphschitz for identifying the wood; I. Segal for the metal analysis of the guns; I. Carmi and D. Segal for the radiocarbon analysis; S. Arenson and D. Syon for critical remarks and English editing; R. Roth for measuring and drawing the cannon; the late A. Raban for his support; the Israel Antiquities Authority; the National Maritime Museum; the University of Haifa. We also thank the anonymous reviewers for their useful comments and corrections.


  1. Top of page
  2. Abstract
  3. The site
  4. The finds
  5. Discussion
  6. Acknowledgments
  7. References
  • Atalay, I. and Efe, R., 2012, Ecological attributes and distribution of Anatolian black pine [Pinus nigra Arnold. subsp. pallasiana Lamb. Holmboe] in Turkey. Journal of Environmental Biology 33, 509519.
  • Ayalon, D., 1967, The Mamluks and Naval Power: A Phase of the Struggle between Islam and Christian Europe, Proceedings of the Israel Academy of Sciences and Humanities 1, no. 8. Jerusalem.
  • Bass, G. F. and van Doorninck Jr. F. H., 1982, Yassi Ada I: A Seventh-Century Byzantine Shipwreck. College Station, TX.
  • Bronk Ramsey, C., 1995, Radiocarbon Calibration and Analysis of Stratigraphy: the OxCal Program. Radiocarbon 37.2, 425430.
  • Bronk Ramsey, C., 2001, Development of the Radiocarbon Program OxCal. Radiocarbon 43.2a, 355363.
  • Constantine, P., 1984, War in the Middle Ages. Oxford.
  • Essenwein, A., 1877, Quellen zur Geschichte der Feuerwaffen. Leipzig.
  • Fuess, A., 2001, Rotting Ships and Razed Harbors: The naval Policy of the Mamlukes. Mamluk Studies Review 5, 4571.
  • Galili, E., Gale, N., and Rosen, B., 2011, Bronze Age metal Cargoes off the Israeli coast. Skyllis 11, 6473.
  • Galili, E., Raban, A., and Sharvit, J., 2002, Forty Years of Marine Archaeology in Israel, in H. Tzalas (ed.), Tropis VII, Proceedings of 7th International Symposium on Ship Construction in Antiquity 1999, 927961. Pylos (Greece).
  • Galili, E., Shmueli, N., and Artzy, M., 1986, Bronze Age ship's cargo of copper and tin. IJNA 15, 2537.
  • Galili, E., Zviely, D., and Rosen, B., 2009, Ancient sounding weights and navigation along the Mediterranean coast of Israel. IJNA 38, 343368.
  • Guilmartin, J. F., 1974, Gunpowder and Galleys. Cambridge.
  • Guilmartin, J. F., 2007, Shipboard Gunpowder Ordnance: An Analysis of Its Technical Parameters and Tactical Capabilities. The Journal of Military History 71, 649669.
  • Heilbrunn Timeline of Art, 2006, Helmet (Turkish) (04.3.456a), The Metropolitan Museum of Art, New York, (October 2006) (Accessed 13/8/2013).
  • Kinder, H. and Hilgemann, W., 1964–1966, DTV-Atlas zur Weltgeschichte. Munchen.
  • Liphschitz, N. and Biger, G., 1992, Building in Israel throughout the Ages. Geojournal 27, 345352.
  • Long, O. P., McGee, D., and Stahl, M. A., 2009, The book of Michael of Rhodes, a Fifteenth-Century Maritime Manuscript. Cambridge, MA.
  • Manucy, A., 1985, Artillery Through the Ages: A Short Illustrated History of Cannon, Emphasizing Types Used in America. Washington DC.
  • McEvedy, C., 1992, The new Penguin Atlas of Medieval History. London.
  • Niklaus, T. R., 1991, CalibETH 1.5b, Program for Calibration of Radiocarbon Dates. Institute for intermediate Energy Physics, ETH, Zurich, Switzerland.
  • Radan, G., 1981, Assyrian Helmet found off the coast of Atlit. Sefunim 6, 4957, pl. VIII:3.
  • Robinson, H. R., 2002, Oriental Armour. New York.
  • Roth, R., 1989, A Proposed Standard in the Reporting of Historic Artillery. IJNA 18, 191202.
  • Segal, D. and Carmi, I., 1996, Rehovot Radiocarbon Date List V. `Atiqot 29, 79106.
  • Stuiver, M., Pearson, G. W., and Braziunas, T., 1986, Radiocarbon age calibration of marine samples back to 9000 cal yr BP. Radiocarbon 28, 9801021.
  • Terdiman, M., 2007, The Mamluk Maritime Policy in the Eastern Mediterranean and the Red Sea. Ph. D Thesis, University of Haifa, Haifa.