Post-classical history


The types of ships involved in the crusades at various times were referred to in contemporary sources by a wide variety of different names, and in most cases the types of vessels to which the terms corresponded are known reasonably well; however, there are exceptions that are sometimes difficult to categorize with certainty. This is particularly true of the Muslim world. Only a handful of scholars have addressed the issues, and none have examined the nature of the ships involved, their historical evolution, and their performance capabilities, issues that influenced, indeed governed profoundly the actual participation of naval forces in the crusades.


The first Western fleet to sail to the East in the period of the crusades was probably that commanded by Guynemer of Boulogne, who anchored off Tarsos (mod. Tarsus, Turkey) in September 1097. His vessels are characterized merely as naves by the chronicler Albert of Aachen, who was undoubtedly using the term as a generic for ships. The Genoese fleet that sailed in summer 1097 reportedly consisted of 12 galee (galleys of a new Western design) and 1 sandanum. The term sandanum was a Latinization of the Greek chelandion, for a transport galley. The Pisan fleet that left in summer 1099 reportedly numbered 120 naves, while that of the Venetians that sailed in summer 1099 had at least 30 naves. The fleet of Sigurd Jorsalfar, which left Norway in 1107, was said by Thôrarin Stuttfeld to have consisted of 60 ships (ON skip). In 1123 the Venetian fleet that sailed for Outremer reportedly numbered 120 naves plus some small boats (Lat. carabii).

During the Third Crusade (1189-1192), a Northern fleet of 50 ships referred to as cogas (cogs) reached Acre (mod. ‘Akko, Israel). The fleet of Richard I the Lionheart, king of England, consisted of ships variously called esneccas or enekes, galees, naves or nefs, dromonz, and bucee. For the Fourth Crusade (1202-1204) the Venetians supplied a battle fleet of 50 galeae (galleys) and a transport fleet consisting of naves for the men and uissiers for the horses. In 1217 Count William I of Holland led a fleet of coccones to Damietta in Egypt. For his Crusade of 1227-1229 the Holy Roman Emperor Frederick II prepared various fleets of ships described as naves, galee, usseria, chelandre, and taride. King Louis IX of France contracted with Marseilles and Genoa for squadrons of naves, taride, and galee for his Crusade of 1248-1254.

Byzantine sources of the crusade period mostly use classical Greek terms such as trieres (pl. triereis) or generic terms such as naus (pl. nees) or ploion (pl. ploia) for ships, although Anna Komnene and Niketas Choniates occasionally use dromon (pl. dromones), the term par excellence for Byzantine war galleys since the sixth century. From the late twelfth century a new term, katergon (pl. katerga), began to be used, displacing dromon and becoming the generic for war galleys during the late Byzantine Empire.

Arabic sources of the period use terms such as qârib (pl. qawârib) for small boats and also for galleys, shini (pl. shawâm) and ghurāb (pl. ‘aghriba/ghirbān/‘aghrub) for galleys; markab (pl.marākib) and qit‘a (pl. qitaf for vessels in general, but often with reference to war galleys; safina (pl. sufun, safā‘in) for vessels in general, but often for transport ships; musattah (pl.musattahāt) for transport ships; batsha (pl. butash) for large sailing ships; ‘ushari (pl. ‘ushāriyyāt) for transport galleys, tarrida (pl. taraid) for horse transports, and harrāqa (pl. harrāqāt) for fire ships, as well as loan words such as shalandi (pl. shalandiyyat) from the Byzantine chelandion, frequently used for Byzantine war ships but also for galleys built in Egypt.

Such words were rarely used in a technical or technological sense. Most authors were unfamiliar with the sea, were not writing for maritime audiences, and used approved or literary terms for ships even if they did know what particular types ought to be called. Generic terms such as naves, nees, or ploia, and sufun/safain or marākib were used for all types of ships in fleets. The Latin word naves was used for all ships of the Pisan and Venetian fleets in the First Crusade (1096-1099), even though they included galleys. But as well as its generic meaning, by the twelfth century the term navis/naves had acquired a specific reference in a Mediterranean context to lateen-rigged, round-hulled sailing ships.

Evolution of Ship Types

The historical evolution of ship types was continuous. A type could change in its fundamental technology and yet retain the same name, as in the case of the cog. Alternatively a name could become applied to a completely different type of ship, as occurred with dromon in its Latin and Arabic variants. By the twelfth century it was used for sailing ships that had nothing in common with Byzantine galleys.

In the centuries before the First Crusade, Mediterranean maritime commerce was carried on a cloud of small sailing ships. Judging from excavations of shipwrecks (the Yassi Ada ship of the seventh century, the Bozburun ship whose timbers were felled in 874, and the early eleventh-century Serçe Limani ship, all from within modern Turkey, and the thirteenth-century Contarina ship found near Venice), such ships averaged around 20 meters (c. 651/2 ft.) in overall length, around 5 meters (c. 161/2 ft.) maximum beam, and around 2-2.5 meters (c. 6V2-8V4 ft.) depth in hold. They probably had single masts and lateen sails, were steered by two stern-quarter rudders, and carried multiple anchors because of their inefficient design and light weight. Smaller versions had only half-decks at bow and stern. Larger ones were fully decked with a small, low cabin at the stern, such as was the case with the Yassi Ada ship.

There certainly were large ships, such as the fine one sailing up the Bosporus that reportedly incensed the Byzantine emperor Theophilos (d. 842) when he learned that its owner was his wife Theodora, or the three-masted “pirate” naus reported by Anna Komnene during the First Crusade. Three- masted sailing ships certainly existed in the eleventh century, as is proved by a depiction on glazed Pisan bacini, which were glazed pottery bowls from the Muslim world placed on the facades of churches at Pisa and elsewhere to reflect sunlight and give the churches a glittering aspect. However, such large ships have left no trace in the documentary record before the thirteenth century.

In the Mediterranean, ships had always been built with strakes edge-joined: that is, with the planks of the hull joined edge to edge and not overlapping as in clinker construction. In classical antiquity, hulls had been shell-constructed from the keel out by holding the strakes together with closely spaced mortise-and-tenon joints pegged with treenails (wooden pegs). Frames were inserted only when hulls had been built to a point where they could be usefully positioned. This produced light and strong hulls but was extremely expensive in terms of the labor and carpentry skills required. By the fourth century (as evidenced in another wreck found at Yassi Ada), tenons had become less tightly fitting, wider but shorter, and spaced more widely apart. The evolutionary process was yet more clearly apparent in the seventh- century Yassi Ada wreck. The Bozburun wreck showed no signs of mortise-and-tenon edge-joining, and by the eleventh century, in the SerçeLimani wreck, skeleton construction over a framework of ribs and stringers had replaced shell construction.

Coating ships with a layer of pitch over the whole hull was replaced by driving caulking into the seams between the planks. The first usages of the Greek word for a caulker (kalaphates) occurred in sixth-century Egyptian papyri and appeared in Byzantium itself in the tenth century, in inventories for expeditions to Crete in the De cerimoniis attributed to Emperor Constantine VII Porphyrogennetos, which included linen for, and the cost of kalaphateseos: caulking. The first picture of caulkers working on a ship is in a manuscript of the De materia medica of Dioskorides Pedanios, probably made for Emperor Constantine VII (MS New York, Pierpont Morgan Library, 652 fol. 240r).

In northern Europe, ships were also shell-constructed, but there strakes were not edge-joined but rather clinker- overlapped. Up to the eleventh century, the sailing ship par excellence was the Norse knorr (ON, pl. knerrir). Two such ships, dated to around 1000, were among 5 ships excavated off Skuldelev in Roskilde Fjord (Denmark). The smaller was around 13.72 meters (45 ft.) long and around 3.20 meters (101/2 ft.) in beam, the larger around 14.94 meters (49 ft.) long and 4.57 meters (15 ft.) in the beam. Both had half decks forward and aft and an open hold amidships. The mast of the smaller was socketed into the kelson, with a stringer laid over the floor timbers and keel to provide fore-and-aft rigidity and to lock the floor timbers to the keel, while that of the larger was set in a complex mast-step. Both would have been steered by a starboard stern-quarter rudder. The larger may well have ventured out into the Atlantic Ocean, since it was heavily built with sturdy frames. Larger knerrir were primarily sailing ships, using what oars they carried only for entering and leaving harbor, for close maneuvering, and in emergencies.

The Skuldelev ships also included two long ships (ON langskip) like the famous ships excavated at Oseberg and Gokstad. The Oseberg ship (from around 800) and the Tune ship (from around 875) had dimensions of 21.4 meters (701/4 ft.) in length, 5.1 meters (163/4 ft.) in the beam, and 1.4 meters (41/2 ft.) depth in hold and around 20.0, 4.6, and an unknown depth in hold respectively (651/2 and 15 ft.). The Gokstad ship (of around 900) was 23.4 meters in length, 5.2 meters in the beam, and 1.9 meters depth in hold (763/4, 17, and 61/4 ft.). Those of the two Skuldelev ships were approximately 18, 2.6, and 1.1 meters (59, 81/2, and 31/2 ft.) and 30, 4.2 meters (981/2 and 133/4 ft.), and unknown depth in hold, respectively. The Oseberg ship carried 15 pairs of oars, the Tune and smaller Skuldelev ships 12 pairs. The Gokstad ship carried 16 pairs of oars and the larger Skuldelev ship between 20 and 30 pairs. In Norse literature the terms langskip and “snake” (ONsnekkja, pl. snekkjur) were interchangeable, although snekkjur may have been larger. As the length of northern longships began to increase, the largest became known as drekar, “dragon ships” (sing. dreki). Those such as the Gokstad ship and the larger Skuldelev ship would certainly have been capable of voyages to Outremer; however, whether Norse langskip, snekkjur, or drekar actually made such voyages is debatable. Crews would have been much more comfortable in knerrir, and there would have been no reason to use longships unless naval combat was expected.

In England knerrir became known as céolas (OE sing. céol), that is keels, perhaps reflecting originally the pronounced keels of Norse ships as opposed to keel-less Frisian hulks and cogs. By the eleventh century the Old English term céol was used commonly for sailing ships in the lower North Sea and English Channel. Another name that appeared increasingly from the eleventh century was bussa/busse/bus/buza (Lat. bucia/bucius); however, whether the Mediterranean Latin bucia/bucius was adopted in the north or vice-versa is unclear. No descriptions of these terms permit attribution of particular characteristics to ships. Nor do illustrations name types until much later. By the thirteenth century ships like knerrir but with light and possibly demountable castles added at the stern, and sometimes the bow, appeared on town seals. Beam ends through the hull of some were probably deck beams. Such ships were probably the buciae in the English fleet of the Third Crusade, and according to Richard of Devizes, the naves of the English fleet, which were only half the size of the buciae, carried forty foot soldiers and forty horses as well as their supplies. The Norse snekkja was also imitated in England under the names esnecca or esneke. King Henry II had a sixty-oared esnecca that made the voyage to Outremer during the Third Crusade. Other flotillas of snekkjur and smaller oared ships from France, Flanders, and Denmark also participated.

From the sixth to late eleventh centuries, the warship par excellence in the Mediterranean was the dromon and its Muslim and Latin imitations, although Byzantines also developed the chelandion, originally an oared horse transport, which was also imitated. By the tenth century dromons had become biremes with two banks of 50 oars, one below and one above deck. A standard ship’s complement (Gr. ousia) consisted of 108 men, excluding officers, marines, and specialists such as helmsmen and a carpenter. Around 31.5 meters (103 ft.) long, they carried two masts with lateen sails and had quarter rudders on both stern quarters. They also had fighting castles on each side just aft of the foremast and a foredeck at the prow, below which was housed their siphon for hurling Greek fire: this was a weapon that used a force pump to eject a stream of petroleum naphtha fuel that was set alight, creating a tongue of flame that could destroy enemy ships.

Monoreme dromons with only 50 oars were known as galeai, and it seems certain that the Latin galea developed from them. The earliest uses of the Latin term were by a group of eleventh-century Italo-Norman chroniclers, which suggests an adaptation of this ship type in southern Italy from Byzantine originals encountered there. Even though references to galee proliferated rapidly, however, they were never described in detail and all that is known about early galee is that they had fine lines and were fast. The earliest documents with construction details are from the Angevin kingdom of Sicily between 1269 and 1284.

Almost certainly, the oarage system was reconfigured so that two oarsmen could row from one bench position above deck. This made immeasurable difference to the oar mechanics, increasing oarsmen’s combined power delivery markedly. No longer did one bank row below deck in stygian darkness and foul air. It also freed holds for supplies, water, spare gear, and armaments. No wonder that the galea spread rapidly in the West and came to be emulated in both the Byzantine and Muslim worlds.

The earliest datable illustrations of galee are three miniatures in the Madrid manuscript of the Synopsis historiarum of John Skylitzes, produced in Palermo around 1160 (MS Madrid, Biblioteca Nacional, vitr. 26-2). These clearly show bireme galleys with a different oarage system. One file of oars was rowed through oarports, but the other was worked above the gunwale. The same system is depicted even more clearly in an early thirteenth-century manuscript of the Carmen ad honorem Augusti of Peter of Eboli (MS Bern, Burg- erbibliothek, 120).

This bireme oarage system became known as the alla sen- sile system. Two oarsmen each rowed single oars from the same bench above deck. Using a stand-and-sit stroke as opposed to the fully seated stroke of classical and Byzantine galleys, they threw their whole weight and the power of their legs into the stroke by falling back onto the bench. The inboard oar was pulled through an oarport in the outrigger, while the outboard oar was pulled against a thole, a pin set in the gunwale (Gr. apostis, It. posticcio, apposticio), to which an oar was held by an oar thong or grommet.

Such bireme galleys remained the norm throughout the main age of the crusades. There is no clear evidence for galleys using any other type of oarage system, and occasional literary references to triremes or to more than one oarsmen pulling each oar are either classical allusions or mistakes. Only in the early fourteenth century did Marino Sanudo Torsello report that: “in [...] 1290, two oarsmen used to row on a bench on almost all galleys which sailed the sea. Later more perceptive men realized that three oarsmen could row on each of the aforesaid benches. Almost everyone uses this nowadays” [Marino Sanudo Torsello, “Liber Secretorum Fidelium Crucis,” in Gesta Dei per Francos, ed. Jacques Bongars, 2 vols. (Hannover: Typis Wechelianis, 1611), p. 57].

Standard bireme galee came to measure around 39.5 meters (129 1/2 ft.) in overall length; they were longer than dromons because they mounted 108 rather than 100 oars and because the stand-and-sit stroke needed a distance between any two tholes (Lat. interscalmium) of around 1.20 meters (4 ft.) rather than the 1 meter (3 1/4 ft.) for fully-seated oarsmen. Their beam was around 4.6 meters (15 ft.) at the deck amidships, and their depth in hold around 2 meters (6 1/2 ft.). They still carried only two masts with lateen sails, the foremasts being almost 16 meters (521/2 ft.) long with yards nearly 27 meters (88 1/2 ft.) and the midships masts being 11 meters (36 ft.) long with yards of 20.5 meters (67 1/4 ft.). The stern-quarter rudders were 6 meters (19 3/4 ft.) long. By the later thirteenth century, standard crews on Angevin galleys consisted of 108 oarsmen, 2 masters, 4 helmsmen, 36 marines, and 2 ship’s boys. Standard armaments included around 30 crossbows, 8 cases of quarrels (crossbow bolts), 40 shields, 200 lances, 10 halberds, 47 axes, 400 darts, 108 helmets and padded jackets for the oarsmen, 40 glass bottles for Greek fire and 100 pots of powdered quick lime, 2 iron grapnels, rigging cutters, and possibly iron rockets for shooting Greek fire.

Thirteenth-century warships. Illustration from the Cantigas de Santa Maria, by King Alfonso X of Castile. (Bridgeman- Giraudon/Art Resource)

Thirteenth-century warships. Illustration from the Cantigas de Santa Maria, by King Alfonso X of Castile. (Bridgeman- Giraudon/Art Resource)

Although no substantiatable evidence suggests that any oarage system other than the alla sensile system was used until the late thirteenth century, a variety of names for galleys other than galee appeared sporadically in the sources: sagene, sagittae, gatti, and garabi in particular.

The term sagena appeared first for vessels of Muslim and Croatian corsairs in Byzantine sources. The type was developed among the Slavs on the east coast of the Adriatic Sea.

Table I: Primary specifications of Genoese 20-horse

taride of 1246 and Angevin 30-horse taride of 1274-1281




Overall length



Depth in hold amidships



Beam on floor


3.56 or 3.69m.

Beam at wale


3.96 or 4.09m.

Beam of deck beams


4.88 or 5.01m.

Number of stern quarter ports


1 or 2

Length of foremast



Length of foremast yard



Length of midships mast



Length of quarter rudders



Number and length of oars


108 or 112m., half 7.38m., half 7.65m.

Length of landing bridges



Number of stalls



Length of stalls



Sagitta (lit. “arrow”) was also applied to corsair galleys, but to those of the Latin West. They were smaller than galee and presumably very fast, to judge by their name. Thirteenth-century Genoese documents refer to sagittae with 48, 58, 64, and 80 oars. Gattus was derived from the Arabic qit‘a and appeared mainly in late eleventh- and twelfth-century sources, referring to galleys larger than the norm.

The ships were sometimes described as triremes, although that may have been a classicizing literary affectation. Garabus was again derived from Arabic: ghurāb, or possibly also qārib, since it is unclear whether these terms were not simply variants of the same name. ‘Aghriba were sometimes said to have carried 140 oars, and garabus may also have been applied to galleys larger than the norm. If they really did row 140 oars, then they must have been triremes.

Horse Transports

Byzantines had oared horse transports equipped with landing ramps as early as the ninth century and retained that capability into the twelfth. The chronicler William of Tyre reports Byzantine horse transports supplied for a combined Byzantine-Frankish attack on Egypt in 1169 as “also having accessible ports at the poops for embarking and disembarking them [horses], also with bridges by which ease of entrance and exit both of men and of horses might be attended to as usual” [William of Tyre, Chronicon, ed. Robert B. C. Huygens (Turnhout: Brepols, 1986), p. 927].

Muslims too were transporting horses on specialized ‘ushāriyyāt and taraid by the tenth century. Twelfth-century taraid could hold forty horses. The Venetians were apparently the first Latins to transport horses to Outremer in 1123, but it is unknown what types of ship they used. By this time the Normans of Sicily definitely could transport horses by galley, because during the Mahdia campaign (1087) 500 cavalry disembarked from beached ships to attack Muslim troops; only galleys could be beached.

Horses were transported in both sailing ships and galleys from the West to Outremer from 1129 up to the time of the Third Crusade. Venice used oared horse transports with stern-quarter ports and landing bridges for the Fourth Crusade (1202-1204), and both appeared among the miscellaneous fleets that reached Damietta during the Fifth Crusade (1217-1221). In 1224 Emperor Frederick II prepared a great fleet of 50 oared horse transports, each carrying forty horses, and in 1246 agents of King Louis IX of France contracted with Genoa for 12 taride to carry twenty horses each for his Crusade to the East.

The Byzantine term for an oared horse transport was che- landion while those of the Muslims had been ‘ushari and tamda. In the West uscerius and variants became the generic term for an oared horse transport, while chelandion became adopted as chelandre and variants and tarnda as taride/tarida. Whether there was ever any real difference between Western chelandre and taride is debatable, but as the thirteenth century wore on, tarida became more common and chelandre disappeared.

Louis IX’s contract of 1246 has the earliest specifications, and these may be compared to those of thirty-horse taride constructed for Charles I of Anjou between 1274 and 1281.

The Later Middle Ages

The twelfth and thirteenth centuries saw no innovation in the technology or construction of Mediterranean sailing naves; however, the length and beam, number of masts, and number of decks and depth in hold of some increased dramatically. By the mid-twelfth century, iconography depicted as a matter of course two-masted naves with multiple-tiered sterncastles and substantial forecastles, such as the Genoese ship conveying Conrad of Montferrat to the Holy Land in the Paris manuscript of the Annals of Genoa [MS Paris, Bibliothèque nationale de France,]. By the thirteenth century, the mosaics of San Marco in Venice also showed three-masted ships. No documentary evidence for three-masted ships survives; however, it is available for two- masted ships provided by Genoa and Venice in 1246-1269 for the two crusades of Louis IX of France. The data vary considerably but may be averaged out or calculated as laid out in the accompanying table.

Table II: Primary dimensions of 2-decked and 3-decked

Genoese and Venetian crusader naves, 1246-1269


2-decked naves

3-decked naves

Length of keel

18.7 m.

22.6 m.

Overall length

28.9 m.

35.2 m.

Depth in hold

3.45 m.

4.11 m.

Height of first deck

2.06 m.

2.91 m.

Height of second deck


2.03 m.

Height of corridoria part decks



(calculated for 3-decked naves)

1.57 m.

1.76 m.

Height of bulwark

1.13 m.

1.18 m.

Height of stem- and sternposts



above keel (calculated)


12.77 m.

Internal beam on floor (calculated)

1.7 m.

3.14 m.

Internal beam at first deck (calculated)


8.92 m.

Maximum internal beam (calculated)

7.75 m.

9.5 m.




Talamus (number of)



Pontes (number of)



Suprapons (number of)






Paradisi (number of)



Castellum (number of)



Vanna (number of)



Supravannum (number of)



Length of quarter rudders (calculated)

13.98 m.

17.0 m.

Length of foremast



(calculated for 3-decked naves)

29.3 m.

36.48 m.

Length of midships mast



(calculated for 3-decked naves)

26.8 m.

34.25 m.

Length of foremast yard (calculated)

36.2 m.

46.8 m.

Length of midships yard (calculated)

33.3 m.

41.8 m.

Ship’s boats (number of)



Barca de canterio (52 oars)



Barche de parischalmo (32 or 34 oars)






Deadweight tonnage (calculated)

323 tonnes

805 tonnes

With the amount of deck space legislated by Marseilles for each pilgrim or crusader, ships of the size of the average three-decked ship could carry around 500-550 passengers. One thirteenth- century Genoese ship, the Oliva, is known to have had a capacity of 1,100 passengers. Such ships were intercepted and captured by Saladin’s squadrons in the 1170s and 1180s and were referred to in Arabic sources as butash. According to Arabic sources, one batsha wrecked off Damietta in 1181/1182 was carrying 2,500 passengers, of whom 1,690 were taken alive. The figures are probably exaggerated, but perhaps not by a great deal if the ship was indeed very large. They could also carry up to 100 horses, normally on the lowest deck, as revealed by a Marseillese contract of 1268 with Louis IX, which specified a fare of 25 shillings for passengers if horses were not stabled there. As the French chronicler Joinville remarked: “On the day that we entered into our ships, the port of the ship was opened and all the horses we wanted to take to Outremer were put inside, and then the port was closed again and plugged well, as when a cask is caulked, because, when the ship is on the high sea, the whole port is under water” [John of Joinville, Vie de Saint Louis, ed. J. Monfrin (Paris: Classsiques Garnier, 1998), p. 62]. Another Northern ship, the cog (MLG kogge), appeared in crusader fleets as early as the Second Crusade (1147-1149). Originally cogs had been flat-bottomed estu- arine and river craft in Frisia, but in the early twelfth century cogs appeared that had flat floors and high sides, a radical new rudder hung off a straight sternpost, and straight stemposts, and that could hold the high seas. The wreck from Kolding Fjord (Denmark), dated to the late twelfth or thirteenth centuries, was around 18.3 meters (60 ft.) long and 6.1 (1 3/4 ft.) meters wide, with a mast step around 1.98 meters (6 1/2 ft.) forward of midships. It had a flat bottom with edge-joined strakes and sides with clinker strakes. Traces of rust on the sternpost revealed a sternpost rudder rather than the Norse quarter rudder. One of the earliest depictions of such a cog is on the first seal of the city of Elbing (mod. Elbląg, Poland) in Prussia (1242). Whether they evolved from earlier Frisian cogs or whether this name was taken over by a completely new ocean-going craft is debatable.

Cogs of the new type appeared in northern fleets for the Third and Fifth Crusades. In 1217 Count William I of Holland left with a fleet of cogs to join the Fifth Crusade at Damietta. Ports were cut into their sterns to embark the horses, and when the horses were on board, the ports were covered and sealed with pitch and tar. The Mediterranean technology for embarking and disembarking horses by ramps through ports in the hull became adopted in northern Europe.

Maritime history was an evolutionary process. The bireme galee, one-decked and single-masted Mediterranean naves, and Norse knerrir and snekkjur of the age of the First Crusade were replaced 200 years later by trireme galee, multiple- decked and masted naves, and northern cogs.

If you find an error please notify us in the comments. Thank you!