Advanced search
CrossRef citations to date

Ship graffiti at the Zanzibar Gereza (Old Fort), Stone Town, Unguja, Tanzania

Pages 90-120
Received 23 Feb 2021
Accepted 10 Jun 2021
Published online: 30 Mar 2022


A number of graffiti of ships are to be found engraved into the plaster of the Gereza (Old Fort) of Stone Town on Unguja, the main island of the Zanzibar Archipelago, Tanzania. Most of those reported here appear on the ramparts of the southwestern tower, while some are on the western face of the main partition wall separating the western and eastern wards. Although sometimes sketchy, the images suggest a number of vessel types, including a frigate or frigate-built vessel and a number of settee-rigged ocean-going vessels referred to exonymically as ‘dhows’. Some appear to have transom sterns, hinting at particular vessel types, such as the baghla, ghanja, sanbūq or kotia. Two graffiti might also depict the stem heads of the East African mtepe. The graffiti are documented and interpreted in the context of the fort, Oman’s East African empire, and the Indian Ocean dhow trade. The construction history of the building and the vessel types depicted date the graffiti to the mid-late nineteenth century.


Un nombre de graffitis de navires se trouvent gravés dans le plâtre du Gereza (Le Vieux Fort) de Stone Town sur Unguja, l’île principale de l’archipel de Zanzibar en Tanzanie. La plupart des graffiti dont il est question ici apparaissent sur les remparts de la tour sud-ouest, alors que d’autres sont sur la face ouest du mur de séparation principal divisant les quartiers ouest et est. Bien que parfois sommaires, les images laissent deviner plusieurs types de navires, y compris une frégate ou un navire de type frégate, et un certain nombre de navires destinés à la navigation maritime, équipés de voiles arabes, nommés de façon exonyme ‘boutres’. Certains semblent avoir des poupes élevées, suggérant des types de navires particuliers, tels que le baghla, le ghanja, le sanbūq ou le kotia. Deux graffitis pourraient également représenter l’étrave d’un mtepe d’Afrique de l'Est. Ces graffitis sont documentés et interprétés dans le contexte du fort, de l'empire est-africain d’Oman et du commerce de l’océan Indien conduit par le biais de boutres. L’histoire de la construction du bâtiment et les types de navires représentés datent les graffitis du milieu ou de la fin du dix-neuvième siècle.


A number of graffiti of ships are visible on the plaster walls of the Gereza (Old Fort) in Stone Town, the capital of Zanzibar, United Republic of Tanzania (Figure 1); they depict types of vessel that once anchored in the roadstead just north of the fort. These images are an important, albeit nebulous, record of the ships that served Stone Town, embodying the connection between the inhabitants of this Omani-built fort and the Indian Ocean, which, at the time of their execution, came within metres of its main door (Rhodes 2014: 104).

Figure 1. Map showing the location of the Gereza (Old Fort), Zanzibar (Image: Alessandro Ghidoni).

The depictions of the ships are incised into the stucco work of the interior faces of the fort (Figure 2). They are graffiti in the technical sense, in that each is ‘a drawing or writing scratched on a wall or other surface’ (‘graffito, n.’ OED Online), although we do not imply by the term that they are inherently ‘illegal, involving the unauthorised marking of public space by an individual or group.’ (Curry and Decker 2020). We investigated one group of graffiti of ships at ground level on the western face of the major wall partitioning the western inner ward of the fort from its eastern counterpart (Figure 3(a)); this wall has its origins as the fort’s eastern curtain wall before the building’s eastward extension (see below). Most, however, appeared on the interior faces of the parapet wall and merlons of the fort’s southwestern tower (Figures 3(b) and 4). We noted others — alas, too late to record — on the parapet of the western curtain wall. In total, we identified 16 of the 32 recorded graffiti (as we grouped them) as depicting at least one watercraft, most of which contained detail beyond simply signifying a ship or boat. The graffiti that we did not interpret as watercraft were all unidentifiable abstractions, whether because of their incompleteness or otherwise: we identified no depictions other than those of watecraft. Shatter facies seen on a number of the graffiti created by the drawing of the inscribing implement across the surface indicate that these ones, at least, were created some time after the plaster into which they were etched had dried.

Figure 2. Outline plan of the Gereza showing the locations of the graffiti (Image: Alessandro Ghidoni).

Figure 3. (a) View down onto the western inner ward of the Gereza, Stone Town, Zanzibar, from the southwestern tower. The graffiti within the ward are behind the row of shops; (b) the southwestern tower from within the western inner ward. (Images: Alessandro Ghidoni).

Figure 4. Outline plan (and flattened elevation view) of the southwest tower parapet wall showing the relative locations of the graffiti.

The graffiti are, for the most part, rather sketchy outlines of sailing vessels seen in profile; the most detailed show the outline shape of the hull and rigging elements such as stays,1 masts, yards and, occasionally, a flag or the outline shape of the sail. These allow us to hazard broad identifications of type. At least one of the graffiti can be said to depict a scene, in that the main vessel is shown with a tender moored to it. Despite their sketchy nature, they nevertheless remain useful to researchers in a part of the world where the physical remnants of past vessels are scarce (Pollard and Bita 2017: 174–175) and iconographic evidence is limited.

Ship graffiti in the western Indian Ocean

The Zanzibar Gereza images reported here complement a small but growing body of scholarly work on graffiti in and on plaster, as well as rock engravings, depiciting ships and boats in the western Indian Ocean. The earliest work focusing on ship graffiti in East Africa is that by Peter and Margaret Garlake (1964), who examined images from mosque and house walls at sites across Tanzania and Kenya dating from the fifteenth to the nineteenth/twentieth centuries AD that had previously been excavated by James Kirkman (1957; 1959) and Neville Chittick (1974: 264–268). More recently, an article in this journal by Edward Pollard and Cesar Bita (2017) has reported images of ships engraved into the walls of a fifteenth-century building on the island of Kilepwa in Mida Creek, Kenya. Across the Indian Ocean in Oman, meanwhile, works by Sarah Kenderdine and Tom Vosmer (1994) and later by Vosmer alone (20102011) have reported impressive ship graffiti at a number of fort sites dating to the 1700s onwards: their Omani and military settings makes them particularly salient to the present paper. Dionisius Agius (2002: 20, 2005: 14) notes ship graffiti on other buildings in Oman, as well as rock engravings, but these have not been extensively surveyed. Meanwhile excavations in Qatar at the site of the eighteenth/nineteenth-century pearling town of al-Zubara (الزُبَارَة ) have also yielded images of ships engraved into plaster walls (Cooper et al. 2015). The connection, through the slave trade, between the pearling towns of the Gulf and ‘exporting’ locales in East Africa again makes these graffiti pertinent to our present discussion, although we do not presume a common authorship between these disparate images, ethnic or otherwise. In dating, the Gereza graffiti are some decades more recent than those of al-Zubara, which date to the late eighteenth or early nineteenth centuries (see below). Also in Qatar are the striking rock art representations of watercraft at Jabal Jassasiyah (جَبَل الْجَسَاسِيّة; Glob 1957; Kapel 1983) and Jabal Fuwairit (جَبَل فُوَيْرِط; Facey 1987: 201–202) in the northeast of the peninsula: the former have been dated to the modern era (Hassiba et al. 2012: 12–14). Across the Gulf in Iran, meanwhile, a graffito of what appears to be a square-rigged vessel with three masts was recorded during excavations by the British Institute of Persian Studies at Siraf (سِيرَاف ) between 1966 and 1973 (Whitehouse et al. 2009: 108). Finally, in the Red Sea, and particularly in Egypt, a huge number of graffiti and rock engravings of past watercraft have been recorded. Most, however, are from so early a period as to have little direct relevance to the world of the Zanzibar fort images discussed here (cf. Lankaster 2013): Julian Whitewright (2011: 203–205) presents one (or perhaps two) exceptions from more recent times from Wadi Quseir al-Qadim, Egypt.

To the limited extent that published work on these ship and boat graffiti has taken a hermeneutic turn, it has been to suggest that their locations suggest some sort of talismanic or votive function: in the East African cases reported to date, the graffiti tend to appear on the outsides of mosques or within houses, near their entrances; in al-Zubara, many are near the corners of rooms or in recesses (Garlake and Garlake 1964: 199–200, 205; Cooper et al. 2015: 46; Pollard and Bita 2017: 179, 185, 186–187). The very different location of most of the Zanzibar Fort graffiti we discuss here — on the ramparts of a corner tower — does not seem to accord obviously with such an interpretation, although the group on the wall partitioning the inner wards of the fort may well have been executed within a building that is no longer standing. Rather, the presence in addition of ship graffiti in several other Omani forts suggests a common medium of expression across Omani military contexts.

Previous work has, rather inevitably, drawn the prosaic conclusion that the images are emphatic of the connection between the communities that created them and the sea (Garlake and Garlake 1964: 205; Cooper et al. 2015: 46; Pollard and Bita, 2017: 185). This can hardly be doubted. Pollard and Bita, meanwhile, consider a didactic or communicative function for them among shipbuilding communities: again, the location of the Gereza images does not lend itself to this interpretation.


We inspected and recorded the graffiti in August 2018 while visiting Zanzibar to conduct an ethnographic survey of contemporary boatbuilding practices on the island of Unguja that was funded by the British Academy and the Honor Frost Foundation (SL-08385: principal investigators Dr John P. Cooper and Dr Lucy Blue). The subsequent process of documentation was therefore opportunistic and necessarily rapid. It was certainly not part of a systematic survey of the entire fort, let alone the historic town surrounding it. Our attention was kindly drawn to the graffiti of the main wall separating the western and eastern wards and those of the southwest tower by Mr Abdallah Khamis Ali of the Department of Museums and Antiquities, Zanzibar, and by Professor Mark Horton of the Royal Agricultural University, United Kingdom, who was at the time excavating inside the fort’s western (inner) ward.

We focused the limited time available to us on recording the graffiti in these locations using a relatively simple and quick on-site methodology. This comprised straightforward colour digital photography of each graffito using a scale, followed by a three-dimensional photogrammetry survey covering the entire area of the rampart wall. In the end, the photogrammetry provided us with no greater diagnostic detail for each graffito than the individual photographs. However, it did mitigate lens distortion and, particularly in the case of the southwest tower graffiti, create a record of the precise location of each graffito relative to the tower and to the others (Figure 4). The photogrammetry was processed using Metashape Pro photogrammetry software; the line drawings of each engraving produced here are hand-tracings of the basic photographs, done in Affinity Design and Adobe Illustrator softwares.

The graffiti in the western inner ward

The graffiti of the western inner ward appear on the major wall that partitions that ward from its eastern counterpart, some 19 m south of the door between the two. At the time of the survey, they were to be found behind a line of temporary shop structures (Figure 3(a)), but there is reason to think that they might have been executed within a building, abutting the wall, that no longer exists, something that we discuss further below.

According to Clark and Horton (1985: 13, 70) the partition wall is the former curtain wall of the original Omani fort, built in the early eighteenth century. It is made of stone (probably coral limestone) and cement that have been covered with white plaster. However the graffiti are incised onto a thickened section of the wall that Clark and Horton attribute to a late nineteenth-century date (see below). The lowest part of the graffiti are some 1.4 m above the modern ground level, and cover an area of approximately 1.3 m2. A large portion of the wall surface had disintegrated by the time of our survey, effacing parts of the graffiti (Figure 5); even when it is largely intact, the plaster often has cracks and holes that can make it difficult to recognise the individual lines of the graffiti and thus make an interpretation.

Figure 5. The ship graffiti on the east wall of the western inner ward: (a) graffito WWEW1, interpreted as the port profile of a two-masted settee-rigged vessel with quarterdeck, interpreted as a ghanja, baghla, or kūtiyya type; (b) the damaged graffito WWEW2, interpreted as a hull with raking prow, washstrake and wale forward of the damage. The grid at the right might be a flag, palm awning, or talismanic grid; (c) the damaged graffito WWEW3, interpreted as the stern of a badan or battīl (Image: authors)

We interpret the graffiti in this ‘group’ as representing three ships or boats. The first, WWEW1, is relatively easy to identify as a ship shown in port profile; although they are fragmentary, we also interpret WWEW2 and WWEW3 as the truncated images of two other boats.

Graffito WWEW1

The highest graffito in the group, WWEW1, is the easiest to recognise as a ship and shows a vessel in profile viewed from the port side (Figure 5(a)). About three-quarters of the length of the hull is visible, from the bow aft; the stern has been lost due to the collapse of the wall’s surface. The bow is raked and slightly curved, with a considerable overhang. The hull appears particularly shallow at its forward section, perhaps depicting its freeboard portion rather than its full depth. Amidships, the sheer slopes suddenly upwards before levelling off again towards the stern, indicating the presence of a quarterdeck. The relatively large size of the quarterdeck area suggests a ship with a transom stern and hence, perhaps, one of a number of large ocean-going cargo vessels used in the western Indian Ocean into the mid-twentieth century, such as the ghanja (غنجة), baghla (بغلة), sanbūq (سنبوق ) and kotia (rendered kūtiyya (كُوتيّة) in Arabic), all of which had a raking stem, stepped sheer line and quarterdeck (Hawkins 1977; Agius 2002: 49–58; Burningham 2007; Weissman and Dziamski 2010; Cooper et al. 2015). The ghanja, baghla and sanbūq were Arabian vessels mainly involved in the monsoon trade with India and East Africa (Figure 6), while the kotia was generally built in the Gulf of Kutch, Gujarat, India.

Figure 6. A Some of the vessel types mentioned in the paper: (a) Guillian’s (1856: plate 51) drawing of a ghanja; (b) the ghanja ‘Fatah al-Khair’ on display in Sur, Oman (Image: Tom Vosmer); (c) Guillian’s (1856: plate 51) drawing of a baghla in Zanzibar; (d) a scale model of a baghla on display at the Museum of the Frankincense Land, Salalah, Oman (Image: Robert Jackson); (e) a sanbūq undergoing maintenance in Zanzibar (mid-twentieth century) (Image: Bait al-Zubair).

Like these vessel types, the ship in graffito WWEW1 also appears to have two masts, although the interpretation of the second of these is open to question. A vertical line rising from the sheer immediately forward of its slope up to the quarterdeck is interpreted as the main mast of the ship: this is corroborated by the presence of a diagonal line set at its top that runs down towards the stemhead. This is most likely the main yard, suggesting that the vessel was rigged with a lateen or settee sail, as indeed these ocean-going vessels were. It could, alternatively, be a fore stay of the main sail, although this seems less likely. The slightly curved line aft of the mast could be the artist’s depiction of the leech of the sail, as seen in other nautical depictions from East Africa such as some of the engravings from the (broadly contemporary) Captain’s House in Fort Jesus, Mombasa (Garlake and Garlake 1964: 199, 203, 204, figures 2(2), 4(3), 5(4, 6 and 10)). A series of lines aft of the main mast and yard, above the quarterdeck, might point to the presence of a mizzen mast, although the pattern of lines is unclear, and certainly not definitive. One vertical line is well positioned to represent a mast, although it does not connect to the sheer line. Meanwhile, two horizontal lines seem to suggest the yards of square sails, rather than the lateen or settee rig that might be expected.

No other ship-related elements such as a rudder, oars, anchors or rigging cables are discernible from the graffito. A short line extending upwards diagonally from what is probably the top of the stern might be interpreted as a jack staff, suggesting the presence of a flag, which in turn is represented by a series of wavy lines running vertically downwards. A flag, often very large, is a commonly observed feature in various illustrations of western Indian Ocean ships of a similar period (Nicolle 1989: 194, figures 66a–b; Guillain 1856: plate 51; Weismann 2012: 427–429; cf. Kenderdine and Vosmer 1994). This practice persisted until recently, as noted by Hawkins (1977: 127), who remarks that the ‘dhows of the Gulf countries appear to compete with each other’ on the colour, number and size of their banners.

One curious feature of graffito WWEW1 is a series of vertical lines connecting the sheer to the lowest part of the depicted hull, which reflect nothing visible on the outside of the hull of a real craft. They might be intended to hint at internal framing, even though they are so few, but this is by no means certain. In any case, similar lines are etched on several of the broadly contemporary ship graffiti (labelled D1, D2, D10 and D14) at al-Zubara (Cooper et al. 2015), as well as some of the sixteenth- to eighteenth-century boat carvings from Jabal Jassasiyah (Facey 1987: 210–211, figures 1–4).

Graffito WWEW2

A few centimetres below graffito WWEW1 is another engraving, only partially preserved, which we also interpret as a ship (Fig. 5(b)). The image is very stylised, with a large portion of it missing due to the collapse of the wall surface; enough of it survives, however, to discern the bow of a vessel with a straight stem, raking at some 55°.

Some structural elements of the hull can be made out. A line at the base of the hull suggests a keel. Half way up, a further pair of horizontal parallel lines suggest a wale. Meanwhile, at the forward portion of the sheer, two diverging lines suggest the presence of a wash strake or other extension plank. There are no additional marks suggesting other commonly depicted elements, such as a rudder, mast or sail.

The graffito seems to reappear on other side of the wall’s intervening damaged surface, with an oblique line perhaps suggesting the raking lower face of a counter. Set immediately above the stern is a square divided into a three-by-four grid into which oblique lines and exes have been engraved. Given the oblique line, interpreted as a counter, leading up to it, this ‘grid’ might be interpreted as a cabin, sterncastle or poop deck, overhanging the stern. Some scholars have interpreted criss-cross and zig-zag patterns on the hulls of nautical engravings from the Indian Ocean as suggestive of a sewn fastening method (Kenderdine and Vosmer 1994: 40; Pollard and Bitta 2017: 181). However, the fact that the maker of graffito WWEW2 has not extended this pattern to the main portion of the hull suggests that this is not the case here. Cooper et al. (2015) interpret roughly comparable cross hatching on the upper stern quarter of three of the ship graffiti at al-Zubarah (graffiti D1, D4 and D12) as stern-quarter decoration typical of several types of Arabian ocean-going sailing vessels, among them the baghla, ghanja and sanbūq. The artist might also have used this particular motif to render the pattern of wooden latticework or woven palm matting. The latter is commonly used in western Indian Ocean boatbuilding to make sails and wash strakes (Lyddeker 1919: 88; Hornell 1941: 57; Prins 1982: 89). Alternatively, the grid might be a further depiction of a flag or of a cabin thatched with coconut palm leaves (makuti) or woven matting of the type seen on the East African mtepe and dau la mtepe (Figure 7(a)) (Lydekker 1919; Hornell 1941, 1942; Prins 1959, 1982, 1986: 64–92; Hatchell 1961; Gilbert 1998) and common on other Indian Ocean watercraft since the fifteenth/sixteenth centuries (Weismann 2002). Thatched cabins are also depicted on vessel depictions from the eleventh/twelfth centuries, as suggested by the illustrations of the ‘Argo’ in various manuscripts2 of the Ṣuwar al kawākib (Nicolle 1989: 171–175, figures 9 and 14a–b). Another possibility, however, is that the grid is a kind of talisman in the form of a geomantic grid, connected with the vessel in order to confer some degree of protective power to someone embarking on, or expected to arrive from, an ocean journey. We know of no parallels to this, however.

Figure 7. (a) a scale model of the East African mtepe with the distinctive thatched cabin at the stern (Image: Alessandro Ghidoni). The model, previously part of the Exeter Maritime Museum collection and the Eyemouth International Sailing Association (EISCA), is now held by at the Institute of Arab and Islamic Studies, University of Exeter; (b) Admiral Pâris’s (1843: plate 9) watercolour painting showing ‘bedens’ (badans) seine fishing off the coast of Muscat; (c): Guillian’s (1856: plate 52) drawing of a “béden” (badan) in Zanzibar.

Graffito WWEW3

Little remains of this graffito except what is perhaps the end of a hull that, in form, vaguely resembles the stern of a badan (بَدَن ) or battīl (بَتِيل ), two vessels known in particular from what is now Oman and the United Arab Emirates, but also East Africa (Figure 5(c)). The high vertical point is reminiscent of the fashīn (فَشِين ; the stern fin) of both types. The fashīn of the badan — at least in the twentieth century — was higher and sharper, while that of the battīl was more elaborate. However, the maker of the graffito may simply have glossed over such detail in this rendition. That the graffito might credibly represent a badan or battīl is supported by the appearance of these vessels in Charles Guillain’s (1856: Volume 2, plate 52) depictions of East African watercraft of broadly the same period and also by photographic evidence presented below. In fact, the putative stern profile appears closer to Guillain’s ‘beden’ or indeed the Omani ‘beden seyad’ recorded by François-Edmond Pâris (1843: plate 9) in 1839 (Figure 7(b–c)). One further surviving detail of the graffito is a line running midway along the side of the hull, broadly parallel to the sheer and bottom of the hull; this might indicate the waterline or else be a stylised rendering of the planking.

The graffiti on the rampart of the southwest tower

We identified 30 individual graffiti on the fort’s southwestern tower, 13 of which we interpreted as depicting watercraft. The remainder we could not interpret as figurative. All were incised into the stucco work of the circular parapet wall of the tower and could have been executed by any adult standing on the tower’s flat roof. The tower itself is roughly circular, with sloping walls, and has a single room at rampart level: the ground-floor level below is infilled with rubble. Access to the roof is today by ladder through a hatch in the ceiling of the tower. The roof area is a little under 8.5 m in diameter. The parapet wall is pierced with gun slits and topped with merlons, the whole standing some 2 m above roof level.

The graffiti occurred around some 280° of the parapet; the remaining 80°or so of the parapet had been replastered relatively recently. Indeed, photographs of the fort found on the internet suggest that this portion of the parapet wall, which overlooks the inner ward, has been entirely rebuilt since the onset of the colour-photographic age.

The tower graffiti were, for the most part, incomplete, due either to the loss of plaster onto which they had been engraved or to the growth of a dark lichen-like organism over the plaster surface. Evidence of shatter around the edges of some of the incisions suggests that these were created some time after the plaster had dried, not while it was wet.

Of the visible watercraft graffiti occuring in the tower, two — SWT2 and SWT11 — are particularly striking.

Graffito SWT2

This graffito (Figure 8(a)) appears to be a depiction of a square-rigged vessel with a three masts, viewed from the port side. We interpret it as a frigate or frigate-built vessel such as a corvette. The image is incomplete, not least because a band of cement — probably chasing for an electrical cable — runs through the grafitto amidships. This excises the presumed main mast, which is nevertheless implied by a remnant sloping line interpreted as a forestay. Detail of the hull is limited: the sheer is flat, while a line below and parallel to it implies a wale. The bow rakes at around 60° from vertical. The stern is largely lost. Most information lies instead in the depiction of the rigging. At the bow, a prominent and raking bowsprit is visible; its length implies extension by means of a jib-boom. Two lines rising aft from this line are interpreted as foremast stays. Also at the bow is a topgallant forecastle, indicated by a roughly oblong polygon. Moving aft past the area of new plaster, a mizzen mast is visible: a horizontal line near the top is interpreted as a mizzen topsail yard; lower down, a diagonal line running aft is interpreted as either the main gaff of a spanker or, less likely, a jack staff; the horizontal line at the same level is interpreted as either a crossjack yard (if the mast was gaff-rigged) or mizzen course yard (if the vessel was ship-rigged). A horizontal line high in the rigging is interpreted as indicating stays running between the foremast and the (missing) main mast and from there to the mizzen mast: again, this is interrupted by the electrical chasing in the plasterwork.

Figure 8. Three of the most detailed ship graffiti on the parapet wall of the southwest tower: (a) graffito SWT2, interpreted as a three-masted frigate-built vessel; (b) graffito SWT3, interpreted as a transom-sterned, settee-rigged vessel such as a baghla, ghanja, kūtiyya or sanbūq; (c) graffito SWT11, interpreted as a scene comprising a settee-rigged vessel with raking stem and (perhaps) two others behind, and a ship’s tender moored to it. The red lines on the photographs on the right-hand side are post-field tracings by the authors to highlight the engraved lines for the sake of clarity. (Image: authors).

Graffito SWT11

Graffito SWT11 (Figure 8(c)), meanwhile, is the only image discussed here that appears to depict not just a vessel, but a wider scene in which it occurs. Its inclusion of a ship’s tender moored to the main vessels recalls a similar scene in a graffito at Fort Jesus, Mombasa (Garlake and Garlake 1964: 204). The image is of a two-masted, lateen-rigged sailing vessel with promiment quarterdeck. Its convex bow curves slightly, and rakes at an average of about 40° from the vertical. The sheer line of the hull curves upwards as it moves forward from amidships and steps up to the quarterdeck aft. A triangular rudder blade is visible at the stern; it tapers to a point just below the counter of the quarterdeck, just where the rudder stock would transition inboard. At the same point, a painter runs out to the bow of a ship’s tender, the aft portion of which is missing.

The main mast of the vessel is clearly indicated by a vertical line running from the waterline (or perhaps the keel) of the hull to the highest point of the graffito; the fact that the engraver has indicated the mast below the sheer line, even though this would not be visible from outside the vessel, reflects similar choices evident in modern-era graffiti at the Palace of Mwenyi Mkuu at Dunga, also in Zanzibar (Garlake and Garlake 1964: 205), and at al-Zubara (Cooper et al. 2015: 38). Meanwhile, a diagonal line rising up to the mast from the bow indicates a forestay, while a corresponding diagonal line aft of the mast and joining the latter higher up suggests a halyard. A diagonal line crossing the upper part of the mast at 50° to the horizontal is interpreted as the yard of a lateen or settee sail. The long horizontal line at the top of the image could also be interpreted as a yard in another positon. However, it appears disproportionately long, and is not the position of a yard at rest: our first candidate for the yard is a better one in both these respects. The mizzen mast of the vessel is depicted on the quarterdeck as a vertical line, crossed by a line at approximately 45°, which probably indicates its yard. Finally, the artist has depicted a large flag hanging downward from a jack staff set at the aft point of the quarterdeck; this is the second appearance of a flag among the Gereza ship graffiti.

Graffito SWT3

A third graffito in the tower has enough detail to be suggestive of broad type. Graffito SWT3 has at its centre the outline of a vessel viewed in starboard profile (Figure 8(b)). The ship has a sharply raking bow (approximately 60° from vertical), a flat sheer and a poop deck with counter. The broadly horizontal line above the sheer can be interpreted as a lowered yard, while two oblique lines rising from the bow area can be interpeted as a forestay and/or moveable stays. The image is comparable to others at the eighteenth-century Captain’s House at Fort Jesus, Mombasa (Garlake and Garlake 1964: figure 5 (3 and 6)) and also at al-Zubarah (Cooper et al. 2015: 37–39, 42–43), where they are interpreted as indicating the Arab baghla or similar, smaller ghanja.

Graffito SWT17

While not indicative of type per se, graffito SWT17 (Figure 9(a)) is highly evocative of the lateen- or settee-sailed ocean-going vessels of a type that would have regularly been seen at anchor from the ramparts of the fort (see below). SWT17 is perhaps again a depiction of a scene, albeit a highly truncated and impressionistic one. In the foreground, the rough outlines of two hulls are seen from the starboard side: their separate, raking prows suggest that this is not a single vessel. Above, a series of diagonal lines rising from the sheer of the upper of the two hulls converge towards a point. The arrangement is highly evocative of the forward-raking mast of a lateen- or settee-rigged vessel with its attendant moveable stays and, sloping furthest aft, a halyard. Slightly higher up the wall, two further clusters of diagonal lines echo the same rakes and convergence patterns and are again evocative of the rigging, masts and yards of lateen- or settee-rigged vessels, even without the attendant evidence of hulls.

Figure 9. Two of the most detailed ship graffiti on the parapet wall of the southwest tower of the Old Fort: (a) graffito SWT17, showing one or two settee-rigged vessels with raking stems below, and possible disembodied settee-rigging above; (b) graffito SWT3, showing a single-masted sailing vessel. (Image: authors).

Graffito SWT19

One further graffito on the tower clearly depicts an individual craft with component features, but its identity is more nebulous. Viewed in starboard profile, or a little to stern, Graffito SWT19 is clearly that of a single-masted sailing vessel (Figure 9(b)). However, there is very little within it to suggest a local lateen-, settee- or square-rigged vessel. Indeed, its mast is very much vertical and the diagonal lines descending from it — particularly those forward, which reach beyond the bow — are more redolent of a fore-and-aft rigged yacht than the rigging of any Indian Ocean vessel. The two sweeping diagonal lines descending from top left to bottom right across the graffito are not interpreted as part of the rigging as they appear too large.

Other probable ship graffiti

Eight of the tower graffiti that we indentified as depicting watercraft do no more than give an impression of the broad shape of ships. Graffito SWT4 (Figure 10(a)) gives the impression of a vessel in port profile view with mast and yard(s?), raked bow and poop deck; to its right the bow, horizontal bow sprit, sheer line, mast and stays of a vessel can be discerned. Whether these two nearby graffiti are part of the same graffito in their creation is, in fact, moot. Meanwhile, graffito SWT5 (Figure 10(b)) can be interpreted as showing a large, amorphous hull in port profile without rigging. There may be another, smaller vessel depicted above, again in port profile. An s-shaped curving line might be interpreted as the famous ‘swan’s neck’ prow of a mtepe, with the two vertical lines to the right the leeches of its square sail, although this is, admittedly, tenuous. Graffito SWT6 (Figure 10(c)) appears to show the raking bow or stern of a hull, complete with stempost, onto which is set a mast and square(?) sail. Meanwhile, graffito SWT13 (Figure 10(d)) shows little that can be interpreted as a hull, but does seem to show a prow that might be interpreted as that of a mtepe. Graffito SWT14 (Figure 11(a)) is a complex set of lines within which it is possible to interpret the outlines of hulls, but there is little tangible. However, graffito SWT23 (Figure 11(b)) is the simple outline of a hull in starboard profile showing a raking prow at 60° from the vertical, a rising sheer line at the bow and, perhaps, a counter at the stern that might place it alongside graffito SWT3 in terms of the type it represents. We interpret the nearby graffito SWT26 (Figure 11(c)) as being a sketchy representation of a mtepe, largely because of its characteristic stem head, heavily raking prow and flat sheer line, sloping here because of the angle of depiction. Finally, the nearby graffito SWT27 (Figure 11(d)) appears to preserve the vague outline of a hull with quarterdeck.

Figure 10. Fragmentary graffiti, probable depictions of vessels, on the parapet wall of the southwest tower of the Old Fort: (a) graffito SWT4; (b) graffito SWT5; (c) graffito SWT6; (d) graffito SWT13, perhaps showing the head of a mtepe.

Figure 11. Fragmentary graffiti, probably depictions of vessels, on the parapet wall of the southwest tower of the Old Fort: (a) graffito SWT14; (b) graffito SWT23; (c) graffito SWT26; (d) graffito SWT27.

Dating the graffiti

Barring further investigation using scientific dating methods, the age of the graffiti in the Gereza can be determined only on the basis of the phasing of the walls on which they are engraved and by the type of vessels depicted. It is clear from several of the images that these at least were inscribed into the plaster after it had dried: there is evidence on these of shatter facies along the chased lines of the images. Hence, the graffiti do not necessarily date to the time of the construction of the walls — or even the latest plaster rendering of the walls — on which they occur, but to some time after.

Our understanding of the development of the Gereza as a building complex is largely based on work by Catherine Clark and Mark Horton (1985: 13, 70) during their 1984–1985 Zanzibar Archaeological Survey and, to a lesser extent, by that of Stéphane Pradines and Pierre Blanchard (2019) when investigating Omani forts across East Africa. It has to be noted, however, that neither of these publications is particularly transparent in presenting the historical or archaeological evidence underpinning their dating of the various phases and that there is considerable scope for clarification of this evidence in future work that might impact the dating of the graffiti discussed here. However, since such an elucidation is beyond the scope of this article we proceed on the basis of this existing scholarship.

The origins of the fort can be traced to the early period of Oman’s rule over the Zanzibar archipelago, which began under the Yaʿāruba dynasty (1624–1741) following its 1698 expulsion of the Portuguese from all of East Africa north of Cape Delgado (Gray 1962: 52–54; Lyne 1987: 9; Breen et al. 2016: 192). John Grey (1933: 226–227) cites an account of an English piratical raid on Zanzibar that mentions an ‘Arab fort’ in, or soon after, 1700. In its first iteration, the building appears to have been little more than a fortification of the structure of the former Portuguese church. The latter had originally been built in 1569 and was rebuilt after an Omani raid in 1652 (Gray 1962: 51, 83; Clarke and Horton 1985: 13, 70). Remnants of the north transept of the church and part of the nave can still be seen in the modern fort, incorporated into the northern curtain wall of the western ward (Clark and Horton 1985: 13; Pradines and Blanchard 2019: 53). There is no evidence of a Portuguese fortification having existed at the site, although James Lancaster’s observation of ‘a small Factorie’ there during his visit of 1591–1592 suggests that the church was not the only significant Portuguese building present (Hakluyt 1904: 392).

According to Clark and Horton (1985: 13, 70), the first Omani fortification of the site comprised simply a square tower butted to the western elevation of the church — together, presumably, with the blockage of windows and doors of the church — with gun emplacements set on the north transept. Their reason for ascribing a 1710 date to this phase is not clear, although the existence of some sort of fortification by that date is supported by Gray’s evidence mentioned above.

As late as 1774, a chart of the waters around Zanzibar compiled by Alexander Dalrymple (1774) ascribes to the nascent Gereza a similar character, describing it as ‘the fort or factory, where there is some small guns, which appears like a ruined church’. This could, however, be based on out-of-date sources. Indeed, according to the phasing for the building arrived at by Clark and Horton (1985: 13), the Omanis had by this time — indeed, by 1750 — developed the site into a quadrilateral fort with a curtain wall that enclosed the area of today’s western ward. Relevant to our dating of the southwest tower graffiti is the fact that the fort in this phase had just two round towers, in its northwest and southeast corners, but none in its southwest. It thus followed what Pradines and Blanchard (2019) call a ‘classic’ Omani fort design, one also found at Siyu and Lamu in Kenya and Chake Chake and Kilwa in Tanzania. According to Clark and Horton’s (1985: 13, 70) account, the southwest tower, on which most of the graffiti discussed in this paper appear, was not built until around 1838. It was at this time that the Omani Busaidi ruler Sayyid Saʿīd subtantially extended the fort, creating the present eastern ward and adding as he did so the southwest tower as well as a further tower at the new southeastern corner (Freeman-Grenville 1966: 139; Clark and Horton 1985: 13, 70; Horton and Clark 1985: 170–171; Pradines and Blanchard 2019: 54). This expansion followed Sayyid Saʿīd’s relocation of his political base from Muscat to Zanzibar after 1832 and was intended to accommodate a garrison of Baluchi and other soldiers (Middleton and Campbell 1965: 4; al-Maamiry 1979: 64). It is this fort that Guillain (1856: plates 4 and 9) recorded during his visit of 1846, showing clearly the southwest tower (Figures 12(a-b)).

Figure 12. The Gereza fort in 1846, according to Guillain: (a) as seen from ‘the anchorage’ (Guillain 1856: plate 4), showing the enlarged post-1838 fort, including the southwest tower; (b) plan of the fort, taken from his Plan de Zanzibar map (Guillain 1856: plate 9). Note the fine lines indicating internal divisions.

According to Clark and Horton once more, the partition wall between the two wards of today’s fort has its origins in the smaller 1750 fort, at which time it constituted part of the eastern curtain wall. However, these authors date the wall’s thickened part — on which appear the graffiti of the western inner ward — to adaptations made around 1890, when the fort was repurposed as a prison (Zanzibar 1952: 44; Clark and Horton 1985: 13, 70). A photograph of the fort from around 1875 (Figure 13) shows a rooftop panorama across the fort towards the southwest, probably from the royal palace — the Bayt al-Sāḥil — where the House of Wonders now stands. Both the closest (eastern) ward in the image and the further (western) one are packed with buildings — the western one less densely — and the fort appears to have not one but two partition walls running north to south. The closest one to the camera appears to correspond to the contemporary partition wall. It may be these two walls that Guillain represents with fine lines in his plan the fort on his 1856 map (Figure 12(b)): if so, he places them somewhat incorrectly.

Figure 13. A photograph from around 1875 showing a view across the rooftops of the Gereza fort from the northeast, perhaps taken (like the image in Figure 14) by British consul John Kirk. The contemporary partition wall is the closest to the camera of the two dividing the inner space of the fort. The southwest tower is the furthest away: note its short missing section of parapet wall. (Image courtesy of Heilke van der Wijke).

The Gereza was abandoned soon after its conversion to a prison, following damage caused by a British naval bombardment in 1893 (Lyne 1987: 196–205; Pradines and Blanchard 2019: 54). At the beginning of the twentieth century, it was used for a time as a workshop for the Bububu railway, which terminated outside its door, but which in 1946 was restored and used for a time by the Zanzibar Ladies Club (Ommanney 1957: 37; Clark and Horton 1985: 13).

According to Clark and Horton’s phasing of the building, we can therefore establish a terminus post quem for the southwest tower graffiti of about 1838 — the year around which the tower was built — and another (perhaps more tentative) of c. 1890 for the graffiti on the wall partitioning the wards. Establishing a terminus ante quem is less easy. However, it does seem less likely that the graffiti were created while the building was used as a railway workshop or a women’s club that allowed women in purdah to exercise unseen by men. Neither of these uses relates to the sea as strongly as the earlier billeting of Baluchi soldiers originating from across the Indian Ocean in the service of a thallasocratic state, although we cannot rule them out.

The graffiti themselves do little to narrow the dating window further. The three-masted frigate-built vessel that we identify in graffito SWT2 is a type with origins that predate the fort, and certainly the southwest tower, but that was progressively replaced with the onset of steam power from the mid-nineteenth century. We therefore presume a broad date for that graffito of between 1838 and the mid-to-late 1800s.

Other graffiti (SWT11, SWT13 and SWT17) hint at lateen- or settee-rigged, ocean-going vessels that might have been seen off Stone Town at any time during the fort’s history up to the late 1960s. The same is true of transom-sterned vessels such as those shown on graffiti WWEW1, SWT3 and perhaps SWT23, which might represent sanbūqs, ghanjas, baghlas or kotias. Finally, if graffiti SWT13 and SWT26 really do show the stem heads of the mtepe, then these must date to before the disappearance of these vessels from the scene early in the twentieth century (Prins 1986: 66).


Although highly variable in level of detail and completeness, the Zanzibar Gereza ship graffiti open a window onto the vessels of the western Indian Ocean at the time of their execution, probably some time in the mid- to late nineteenth century. The Zanzibar Archipelago at the time constituted the southern terminus of a trans-oceanic trade network that harnessed the Indian Ocean monsoon, a relatively stable seasonal wind system that has been utilised by ocean-going mariners in the region for at least two millennia (Gilbert 2004: 23; Sheriff 2010: 10–12; Breen et al. 2016: 192–196). Under this régime, winds blow across the ocean from the southwest to the northeast as the Asian continent heats up during the northern hemisphere summer and then reverse direction as winter cools the Asian land mass. These stable oceanic winds facilited an annual cycle of trade and contact between East Africa on the one hand and Arabia, the Red Sea, the Arabian-Persian Gulf, Iran and India on the other (Sheriff 2010: 20–23). In particular, merchant vessels referred to exonymically as ‘dhows’ carried people, together with cargoes such as dates, dried fish and textiles, from Arabia and the Gulf to East Africa, loading up with mangrove poles, gum copal, ivory, rhinoceros horn and enslaved persons for the return journey (Gilbert 2004: 31–33). According to Gilbert (2004: 21), ‘Zanzibar [i.e. Stone Town] in the mid-nineteenth century was a boom town.’ Visiting it in 1868, the British naval captain G.I. Sulivan (2003 (1873): 75) describes as ‘literally crowded with dhows of every kind, and among them several full of slaves.’

Stone Town emerged as a key locale in the monsoon trade system in part because of its position within the Zanzibar Channel, where it was sheltered from the open ocean to the east by the island of Unguja itself and from the west by the Africa mainland, some 35 km away. More locally, the headland of Shangani, on which Stone Town sits, provided further shelter, with vessels anchoring on its northern side during the southwest monsoon, and on its southern side during the northeast; the locale also offered ample fresh water to navigators (Gilbert 2004: 29). A series of small islands close offshore added an extra level of shelter to vessels at anchor there (Breen et al. 2016: 191). It was on Shangani’s northern side that the Portuguese established their first settlement, which in turn became the kernel around which the subsequent Omani fort grew. At the time the graffiti were executed, vessels anchored, beached and unloaded their cargoes along the length of the waterfront in front of the fort. Such a scene is amply captured in a photograph taken by John Kirk, British consul in Zanzibar from 1866 to 1886 (Hazell 2012: 75, 321) (Figures 14(a–b)). The southwest tower of the fort can been seen clearly in the lower righthand corner of Kirk’s image: clearly visible, too, is part of its parapet wall, on which the graffiti were, perhaps, by then already inscribed (Figure 14(b)). The first development of ‘hard’ waterfront infrastructure only came about with the mid-nineteenth-century boom in economic activity in Zanzibar following the Omani Āl Bū Saʿīd dynasty’s development of Stone Town as their political capital, that is, after 1832. A sea wall was built to protect the fort and the new royal palace, Bayt al-Sāḥil, that was constructed to its east, to which a wharf had been attached by the 1850s (Breen et al. 2016: 195). Today’s ‘Dhow Harbour’, meanwhile, is the product of a port development completed in 1929 during the period of the British protectorate, which began in 1890 (Gilbert 2004: 29–30, 99–100; Breen et al. 2016: 195–196): set to the east along the shore, largely out of site of the fort, it almost certainly long postdates the inscription of the graffiti.

Figure 14. (a) A panoramic view of Stone Town waterfront in 1875 by John Kirk, showing the crowded anchorage immediately outside the fort. (Image: British Library: Visual Arts, Photo 355/1(120), retrieved from the Qatar Digital Library×000002); (b) a close-up of the anchorage area taken from the same photograph, showing the variety of boats that could be seen. The image includes a frigate-rigged vessel, several baghlas and ghanjas (or possible kotias) and a badan, all types identified among the graffiti. (High-resolution image courtesy of Torrence Royer).

The emergence of the Gereza fort itself, meanwhile, reflects Oman’s growing political involvement in the Zanzibar archipelago and adjacent mainland coast, echoing and amplifying the economic interests and social connections that the monsoon system engendered. Oman has a long and deep connection with East Africa that reaches back into the medieval period. The Omani historian Reda Bhakar (2004: 6) infers, from the flight to East Africa of the Julanda rulers of Oman’s interior at the turn of the eighth century AD, that links with the region were already in existence at that time. Meanwhile, the Arab traveller and geographer al-Masʿūdī reports Omani-crewed trading ships travelling to East Africa and penetrating as far south as Sofala in Mozambique in the tenth century (Barbiers de Meynard and Pavet de Courteilles 1861: Volume I 231–233). Oman’s strategic and commercial interests in the region were dealt a blow by the violent arrival in 1498 of the Portuguese, who asserted their naval dominance along the East African coast, building forts at Ilha de Moçambique and Sofala in Mozambique and also Fort Jesus in Mombasa, Kenya. They also took Shangani in 1509 (Breen et al. 2016: 192). Oman was able ultimately to reassert itself, however, and from the mid-seventeenth century it was again appointing governors (Arabic: wālī (singular)) in the region (Bhacker 2004: 3). A process of naval reassertion culminated in the Yaʿāruba dynasty’s ejection of their European rivals from all of the East African coast north of Cape Delgado in 1698, following a siege of Fort Jesus. It was from this time that Oman’s rulers began to develop the Zanzibar Gereza as one of their main fortifications in the region. From it they oversaw and controlled the trade in raw materials and enslaved people from the African interior that passed through Zanzibar. However, it was the subsequent development of the spice trade, and in particular the prospect of esablishing clove plantations on the Zanzibar Archipelago (Rhodes 2014: 25; Breen et al. 2016: 193), that ultimately shifted the political centre of gravity of Oman’s thalassocratic state from Arabia to Zanzibar, giving it also the wherewithall to support a substantial navy. When Sultan Saʿīd bin Sulṭān Āl Bū Saʿīd relocated his court from Muscat to Zanzibar in 1832, he established his new Bayt al-Sāḥil palace alongside the Gereza fort. The surrounding Shangani district increasingly became the focus of Arab, Persian and Indian merchant houses, markets and foreign consulates (Nicholls 1971: 79, 246; Sheriff 1995: 146; Breen et al. 2016: 193–195). It was during this period of Omani-Arab political and economic primacy that the fort was used to billet Baluchi and enslaved local soldiers, these troops being our best candidates for the creators of the ship graffiti. Nicholls (1971: 255) has the number of Baluchi soldiers present varying from 50–60 in 1819 (i.e. before the fort was expanded) to 30 in 1843 and as many as 80 in both 1848 and 1856.

When Sultan Saʿīd died in 1856, the Āl Bū Saʿīd realm was divided between two of his sons, and Zanzibar established as a sultanate independent of Muscat: this independence came to an end in 1890, when the then Sultan, ʿAlī bin Saʿīd, accepted British protectorate status. An attempt by his successor three years later to break free of the British led not only to failure, but also to the bombardment and abandonment of the fort as a military and prison structure.

Perhaps the most intriguing image among the graffiti — graffito SWT2 — is a rendition of a three-masted frigate or frigate-built vessel such as a corvette. It is certainly the case that frigate-built and other square-rigged ships from Western powers visited Zanzibar during this time and it might be that the graffito is of just such a vessel. However, it is also the case that the Āl Bū Saʿīd owned a substantial fleet that included several such vessels. An American merchant, Edmund Roberts, describes Sultan Sayyid Saʿīd’s arrival in Muscat some time around 1828: ‘bringing with him a force consisting of a 64 gun ship [the Liverpool], three frigates of 36 guns, and two brigs of 14 guns’ (Gray 1962: 124). In the same year, Sayyid Saʿīd threatened his Mazrūʿī rivals in Mombasa with a force including the Liverpool, the 50-gun Shah Alam and two corvettes (Gray 1962: 261). Gray (1962: 261) presents a list of vessels in the Sultan’s fleet that includes not only the Liverpool and Shah Alam, but also the 36-gun Piedmontese, which Said-Ruete (1929: 91) describes as a frigate. In 1833, Sayyid Saʿīd also famously commissioned in Bombay a three-masted corvette, the Sultanah, which sailed on a diplomatic mission to Europe and the United States (Facey 1979: 181–188). It may well be, therefore, that the square-rigged vessel depicted in this graffito was, in fact, Omani, rather than belonging to a Western power.

It is noteworthy, meanwhile, that square-rigged vessels are among the graffiti reported from Siraf in Iran (Whitehouse et al. 2009: 108), al-Zubara in Qatar (Cooper et al. 2014: 39) and the forts of Fulaij, Jabrin and Bayt al-Na’aman in Oman (Kenderdine and Vosmer 1994: 34, 37–38, Vosmer 20102011: 102, 105–106). This is certainly an impressive naval vessel type, and one that evidently inspired the efforts of makers of graffiti.

The settee- or lateen-rigged vessels depicted in the graffiti may also represent Omani ocean-going merchant vessels participating in the monsoon-based trade, although the sultan also made sure that such ships were adaptable for conflict and thus to carry guns. Alternatively, they might represent non-Omani trading craft ariving from Yemen, the Arabian-Persian Gulf or India. It is tempting to identify the transom-sterned vessels as representing ghanjas, since these are particularly associated with Oman and in particulary the town of Sur where so many of them were built (Hawkins 1977: 24; Agius 2002: 53–54). However there is nothing in these very rough images to allow us to differentiate between the ghanja and other transomed sterned ships such as the sanbūqs (particularly associated with Aden and Salalah), baghlas (associated with the northen Gulf) or kotias (associated with Gujarat). That said, the prominent counters on graffiti SWT3 and SWT11 are more redolent of a ghanja, baghla or kūtiyya than, say, of a sanbūq (Jewell 1969: 42–57, 62–65, 72–75; Burningham 2007).

The graffiti are a further example of the amenability of the fine-plaster coating found on rubble-stone walls in many parts of the western Indian Ocean to engraving with a sharp metal implement and of the desires of engravers to make ships the subject of their action (Garlake and Garlake 1964; Cooper et al. 2015; Pollard and Bita 2017). What sets the Gereza graffiti apart from so many of those already reported from East Africa, however, is that they are not set on the outside of mosques or (certainly those in the tower) at significant locations within domestic spaces as others reported elsewhere were, for example near doors or recesses. Hence, those in the tower at least do not lend themselves to the same interpretations as talismanic within a domestic setting or as votive in a public-religious one. We cannot rule out a spiritual or religious function for these images, but their location does not suggest such an interpretation. The situation is less clear for the graffiti near ground level on the partition wall between the western and eastern inner wards. The photographs reproduced in Figures 13 and 14(b) show a complex huddle of buildings within the western ward, leaving open the possibility that these were indeed created in an indoor, perhaps domestic, space and that the enclosing building was later removed. However, we cannot be sure of this from the evidence of the photographs alone.

In contrast, the graffiti do sit within a pattern of images of vessels appearing on the walls of forts in Oman itself, as the reports of Kenderdine and Vosmer (1994) and Vosmer (2010-2011) show. Graffiti — some inscribed, but most painted — are reported from the forts of Fulaij (فُلَيْج ), Hazm (حَزم ), Jabrin (جَبْرِين ), Nizwa (نِزْوَة )، Bayt al-Na‘aman (بَيْتُ اَلْنُعْمَان ) and Liwa (لِيوَا ), suggesting a reletively widespread practice of inscribing ship graffiti within Omani military buildings. Set within the confines of the fort, the Gereza graffiti were not for public consumption in the way that they might have been had they been set on the fort’s outer faces, where people flocking to the busy Soko Uku market under its walls might have seen them, as would the families of Arab and Indian merchants and notables who built their houses around the fort (Sheriff 1987: 139). For the most part, the graffiti must have been made for and by members of the internal community of the fort itself. Moreover, those in the southwest tower and the western ramparts of the Gereza must have been made by people with access to these more reserved upper reaches of the fort, hence our suggestion of the Baluchi or slave soldiers garrisoned in the fort by Omani or Zanzibari sultans for much of the nineteenth century. They were probably made by people with time on their hands, soldiers on guard duty or spending their leisure time in the breezier upper reaches of the building. The Baluchi soldiers would themselves have arrived, and ultimately departed, by such ocean-going craft. The images are less likely to have been created by prisoners during periods when the fort was functioning as a gaol, given the location of so many on the ramparts.

All of the graffiti depict ships that would have been easily visible from the ramparts of the fort itself or by stepping a few paces outside its door. Indeed, those on the ramparts could have been executed by making almost constant reference to the ships themselves, visible at anchor along the waterfront of Stone Town. There is a striking similarity between some of the graffiti — in terms of the orientation of the vessels shown and the salient features selected — and historic photographs of ships taken from the ramparts of the fort, such as that of Kirk (Figure 14), which shows the scene along the waterfront just north of the Gereza around the time that the graffiti were being executed. Indeed, the southwest tower itself is visible in the right of the photograph, which was evidently taken from an adjacent building. Not only does the photograph provide a general impression of the crowded roadstead, but it also captures several of the vessel types we have identified from the graffiti: a three-masted, frigate-built vessel, several transom-sterned baghlas, ghanjas and possible kotias and also a badan are all visible in the photograph. Interestingly, all of the vessels in the photographs have their port sides facing the camera, presumably because the photograph was taken during the southern (kusi) monsoon: all of the ship graffiti are also shown in profile view, although not all on their port side.

A plausible interpretation is therefore that the vessels were engraved by soldiers on guard duty with little else to do. Even if that is the case, the choice of subject matter remains salient: above all, the engravers opted to represent watercraft among all the things that could be seen from the vantage point of the fort. And among watercraft, it seems to have been ocean-going vessels that commanded their attention. Those graffiti containing enough detail to be identified are all ships: a frigate, baghlas or ghanjas, possible mtepes. They never depict a small fishing canoe, for example. This again suggests a connection between the engravers in the fort and ocean-going, monsoon-based traffic. Furthermore, although the graffiti are not highly detailed, the best do contain enough information to show that their engravers either understood how ships were broadly constructed and operated or at the very least were good draughtspeople when it came to capturing the essence of what they saw.


We are grateful to the British Academy/Leverhulme Small Grant scheme and the Honor Frost Foundation for the funding that enabled this research (grant SRG/171201), as well as to Mr Abdulla Khamis Ali of the Zanzibar Department of Museums and Antiquities and Professor Mark Horton of the Royal Agricultural University, United Kingdom, for bringing the existence of the ship graffiti to our attention. Thanks are also due to our colleague Dr Lucy Blue for assisting us in the recording of the graffiti in the field.


The following nautical glossary is intended as a simplified reference for non-specialist readers of this article. It is derived largely (and loosely) from Smyth’s The Sailor’s Word-Book (2005 (1991)) and Kemp’s The Oxford Companion to the Sea (1976). Terms used in the definitions that themselves appear in the Glossary are underlined.

Aft towards the back of the hull.

Amidships towards the central section of the hull.

Bow the forward part of the hull.

Bowsprit a large spar that is run out over the bow of a vessel, extending forward the point to which the foot of a settee sail can be anchored.

Counter­ an overhang at the stern of a vessel, above the waterline.

Crossjack yard the lower yard on the mizzen mast of a square-rigged ship, to which the corners of the mizzen topsail are attached.

Fashīn a tall, fin-like set of timbers at the stern of a vessel, from which the rudder is hung.

Foremast the foremost mast on the vessel.

Forestay the stay used to secure the foremast.

Foretopgallant mast the third timber of a mast, counting from the deck, extending the overall mast upwards.

Freeboard—the vertical distance between the waterline and the deck amidships.

Gaff rigged rigged with a quadrilateral fore-and-aft sail that is laced to a gaff, the latter being a spar that is hoisted on the after side of a mast, against which it pivots.

Halyard the rope and associated tackle used to raise the sail, attached to the yard.

Hull the main body of the ship, excluding rigging, superstructure and internal fittings.

Jack staff a short pole erected at the stern of a vessel on which the national flag is hoisted.

Jib-boom a separate spar run forward of the bowsprit, extending the foot of the jib sail and the stay of the fore topgallant mast.

Jib a triangular sail set on the stays of the foremast.

Keel the lowest and main timber of the hull, running longitudinally and supporting the framing timbers.

Lateen sail a triangular sail set on a long yard, its forward end set low, so that its aft end forms a high peak.

Leech the aft edge of a settee or lateen sail; either of the vertical edges of a square sail.

Luff the leading edge of a settee or fore-and-aft sail.

Mizzen course yard the yard from which is set the lowest square sail on the mizzen mast of a square-rigged ship.

Main gaff—The spar to which the head of a quadrilateral fore-and-aft sail is attached.

Mizzen mast—the aftermost mast of a sailing ship.

Poop deck the aftermost and highest deck of a vessel.

Prow the foremost end of a vessel.

Quarterdeck that part of the deck that is aft of the main mast.

Rake (raking/raked) the angle of projection of a timber from the vertical; having a pronounced angle of projection.

Settee sail a quadrilateral sail with a very short luff, giving the impression from afar of a lateen sail.

Sheer/sheer line the longitudinal curve of a hull’s upper edge.

Ship-rigged having three or more masts, all of them set with square sails.

Square-rigged having mainly or exclusivley square sails.

Spanker an exta sail raised on the mizzen mast to benefit from a following wind.

Spar a general term for any length of timber, often roundwood, used as a mast, yard, boom, etc.

Starboard the right-hand side of the vessel when looking forward.

Stay a part of the standing (i.e. fixed) rigging of a vessel, set fore and aft, that supports the mast.

Stem the substantial timber at the bow of a vessel to which the forward end of the planking is attached. It is also jointed to the keel.

Stem head the upper part of the stem.

Stern the rear end of the hull.

Stern post the timber at the stern of the hull corresponding to the stem at the bow; the hull planking is jointed to it aft and it, in turn, is jointed to the keel.

Sterncastle a structure built behind the mizzen mast and above the transom that houses the principal cabins of a ship.

Strake a line of hull planking running along the hull of a vessel.

Tender a small boat serving the shore communications of a larger vessel.

Topgallant forecastle a short forward deck above the main deck.

Topsail the second sail set on a mast above the deck.

Transom side-to-side timbers fixed to the stern post of some vessels, creating a ‘flat’ stern.

Wale an extra thickness of wood running along the sides of a hull to provide additional strength and protection.

Wash strake a strake set at the upper part of the bow to keep spray out of the hull.

Yard a large spar crossing the mast to which the head (top) of a sail is attached.

Additional information

Notes on contributors

John P. Cooper

John P. Cooper is Associate Professor in Maritime Archaeology and Arabic at the Centre for Islamic Archaeology of the University of Exeter’s Institute of Arab & Islamic Studies. He is a maritime archaeologist and Arabicist, with research interests lying in the maritime past of the Islamicate world, approached through archaeology, ethnography and text, and has worked in the southern Red Sea and Gulf of Aden, the Arabian-Persian Gulf and the Indian Ocean, as well as along the Nile. He is a Fellow of the Nautical Archaeology Society.

Alessandro Ghidoni

Alessandro Ghidoni recently completed his doctoral studies at the Centre for Islamic Archaeology of the University of Exeter’s Institute of Arab and Islamic Studies. His research revolves around shipbuilding and seafaring in the western Indian Ocean, with a particular focus on medieval Islamic watercraft. He has been involved in maritime-focused ethnographic, archaeology and experimental archaeology projects in Oman, the United Arab Emirates, Qatar, Bahrain Zanzibar and India ranging from the Early Bronze Age to the modern era.


1 A glossary of nautical terms is provided for the lay person at the end of the article.

2 Bodleian Library., Ms Marsh 144, folios 365–366 and 367–368, Oxford): AD 1009/10; Topkapi Library, Ms., Faith 3422, ff. 198z and 198 V, Istanbul: from Mardia AD 1134/35.


  • Agius, D.A. 2002. In the Wake of the Dhow: the Arabian Gulf and Oman. Reading: Ithaca. [Google Scholar]
  • Agius, D.A. 2005. Seafaring in the Arabian Gulf and Oman: the People of the Dhow. London: Routledge. [Google Scholar]
  • Barbiers de Meynard, C. and Pavet de Courteille, A. (eds). 1861. Maçoudi. Les Prairies d'Or. Paris: Imprimerie Impériale. [Google Scholar]
  • Bhacker, R. 2004. Trade and Empire in Muscat and Zanzibar: the Roots of British Domination. London: Routledge. [Google Scholar]
  • Breen, C., Forsythe, W. and Rhodes, D. 2016A maritime archaeological survey of Stonetown, ZanzibarInternational Journal of Nautical Archaeology 45: 191198 [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Burningham, N. 2007. “Baghla, ghanja and kotia: distinguishing the baghla from the Suri ghanja and Indian kotia.” International Journal of Nautical Archaeology 36: 91111. [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Chittick, H.N. 1974. Kilwa: An Islamic Trading City on the East African Coast. Nairobi: British Institute in Eastern Africa. [Google Scholar]
  • Clark, C. and Horton, M. 1985. Zanzibar Archaeological Survey 1984/5. Zanzibar: Ministry of Information, Culture and Sport. [Google Scholar]
  • Cooper, J.P., Agius, D.A., Collie, T. and al-Naimi, F.A. 2015. “Boat and ship engravings at Al Zubarah, Qatar: the dāw exposed?Proceedings of the Seminar for Arabian Studies 45: 3548. [Google Scholar]
  • Curry, G.D. and Decker, S.H. 2020. “Graffiti.” Encyclopedia Britannica, Site accessed 8 February 2021. [Google Scholar]
  • Dalrymple, A. 1774. Plan of the Road or Harbour of Zanzibarra or Zanzibar, From an English M.S. London: By Act of Parliament. [Google Scholar]
  • Dziamski, P. and Weismann, N. 2010. Fatah al-Khair: Oman’s Last Ghanjah. Muscat: Al Roya Press and Publishing House. [Google Scholar]
  • Facey, W. 1979. Oman: A Seafaring Nation. Muscat: Ministry of Information and Culture. [Google Scholar]
  • Facey, W. 1987. “The boat carvings at Jabal al-Jussiyah, northeast Qatar.” Proceedings of the Seminar for Arabian Studies 17: 199222. [Google Scholar]
  • Freeman-Grenville, G. 1966. The East Coast of Africa: Select Documents. Oxford: Clarendon Press. [Google Scholar]
  • Garlake, P. and Garlake, M. 1964. “Early ship engravings of the East African coast.” Tanganyika Notes and Records 61: 197206. [Google Scholar]
  • Gilbert, E. 1998. “The mtepe: regional trade and the late survival of sewn ships in East African waters.” International Journal of Nautical Archaeology 27: 4350. [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Gilbert, E. 2004. Dhows & the Colonial Economy of Zanzibar 1860–1970. Oxford: James Currey. [Google Scholar]
  • Glob P.V. 1957. “Prehistoric discoveries in Qatar.” Kuml 7:  167178. [Crossref][Google Scholar]
  • Gray, J. 1962. History of Zanzibar from the Middle Ages to 1856. London: Oxford University Press. [Google Scholar]
  • Grey, J. 1933. Pirates of the Eastern Seas (1618–1723): A Lurid Page in History. London: Sampson Low, Marston & Co. [Google Scholar]
  • Guillain, C. 1856. Voyage à la Côte Orientale d'Afrique: Exécuté pendant les Années 1846, 1847 et 1848 par le Brick le Ducouëdic, sous le Commandement de M. Guillain: Album. Paris: A. Bertrand. [Google Scholar]
  • Hakluyt, R. 1904. The Principal Navigations, Voyages, Traffiques and Discoveries of the English Nation, Volume 6. Edinburgh: J. MacLehose & Sons. [Google Scholar]
  • Hassiba, R. Cieslinski, G.B., Chance, B., Al-Naimi F.A., Pilant, M. and Rowe, M.W. 2012. “Determining the age of Qatari Jabal Jassasiyah petroglyphs.” QScience Connect. 2012: 116. [Crossref][Google Scholar]
  • Hatchell, G.W. 1961. “The ngalawa and the mtepe.” Tanganyika Notes and Records 56: 210215. [Google Scholar]
  • Hawkins, C.W. 1977. The Dhow: An Illustrated History of the Dhow and its World. Lymington: Nautical Publishing. [Google Scholar]
  • Hazell, A. 2012. Dr John Kirk and the Struggle to End the East African Slave Trade. London: Constable. [Google Scholar]
  • Hornell, J. 1941. “The sea-going mtepe and dáu of the Lamu Archipelago.” The Mariner’s Mirror 27: 5468. [Taylor & Francis Online][Google Scholar]
  • Hornell, J. 1942. “The sea-going mtepe and dáu of the Lamu Archipelago.” Tanganyika Notes & Records 14: 2737. [Google Scholar]
  • Horton, M.C. and Clark, C.M. 1985. “Archaeological survey of Zanzibar.” Azania 20: 167171. [Taylor & Francis Online][Google Scholar]
  • Jewell, J. 1969. Dhows at Mombasa. Nairobi: East African Publishing House. [Google Scholar]
  • Kapel, H. 1983. “Rock carvings at Jabal Jusasiyah, Qatar.” Arrayan 8: 1126 and Appendix 1–53. [Google Scholar]
  • Kemp, P.K. 1976. The Oxford Companion to Ships and the Sea. London: Oxford University Press. [Google Scholar]
  • Kenderdine, S. and Vosmer, S. 1994. “Maritime graffiti in Oman.” Bulletin of the Australian Institute for Maritime Archaeology 18: 3344. [Google Scholar]
  • Kirkman, J.S. 1957. “Takwa — The mosque of the pillar.” Ars Orientalis 2: 175182. [Google Scholar]
  • Kirkman, J.S. 1959. “Excavations at Ras Mkumbuu on the island of Pemba.” Tanganyika Notes and Records 53: 161178. [Google Scholar]
  • Lankaster, F. 2013. Desert Boats: Predynastic and Pharaonic Era Rock Art in Egypt’s Central Eastern Desert: Distribution, Dating and Interpretation. Oxford: Archaeopress. [Crossref][Google Scholar]
  • Lydekker, C.J.W. 1919. “The ‘mtepe’ dhau of the Bajun Islands.” Man 46: 8892. [Crossref][Google Scholar]
  • Lyne, R.N. 1987. Zanzibar in Contemporay Times: A Short History of the Southen East in the Nineteenth Century. London: Darf Publishers. [Google Scholar]
  • al-Maamiry, A.H. 1979. Oman and East Africa. New Delhi: Lancers Books. [Google Scholar]
  • Middleton, J. and Campbell, J. 1965. Zanzibar: Its Society and Politics. London: Oxford University Press. [Google Scholar]
  • Nicholls, C.S. 1971. The Swahili Coast: Politics, Diplomacy & Trade on the East African Littoral, 1798–1856. London: George Allen & Unwin. [Google Scholar]
  • Nicolle, D. 1989. “Shipping in Islamic art: seventh through sixteenth century AD.” The American Neptune 49: 168197. [Google Scholar]
  • Ommanney, F.D. 1957. Isle of Cloves: A View of Zanzibar. London: The Travel Book Club. [Google Scholar]
  • Pâris, F.E. 1843, Essai sur la Construction Navale des Peuples Extra-Européens, ou Collection des Navires et Pirogues Construits par les Habitants de l’Asie, de la Malaisie, du Grand Océan et de l’Amérique Dessinés et Mesurés pendant les Voyages autour du Monde de ‘l’Astrolabe’, ‘la Favorite’ et ‘l’Artémise’. Paris: Arthus Bertrand. [Google Scholar]
  • Pollard, E. and Bita, C. 2017. “Ship engravings of Kilepwa, Mida Creek, Kenya.” Azania: Archaeological Research in Africa 52: 173191. [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Pradines, S. and Blanchard, P. 2019. “From Zanzibar to Kilwa: eighteenth and nineteenth century Omani forts in East Africa.” Fort 47: 3969. [Google Scholar]
  • Prins, A.H.J. 1959. “Uncertainties in coastal cultural history: the ‘Ngalawa’ and the ‘Mtepe’.” Tanganyika Notes and Records 53: 205213. [Google Scholar]
  • Prins, A.H.J. 1982. “The mtepe of Lamu, Mombasa and the Zanzibar Sea.” Paideuma: Mitteilungen zur Kulturkunde 28: 85100. [Google Scholar]
  • Prins, A.H.J. 1986. A Handbook of Sewn Boats: The Ethnography and Archaeology of Archaic Plank-built Craft. Greenwich: National Maritime Museum. [Google Scholar]
  • Rhodes, D.T. 2014. Building Colonialism; Archaeology and Urban Space in East Africa. London: Duckworth. [Google Scholar]
  • Said-Ruete, R. 1929. Said bin Sultan, 1791-1856: Ruler of Oman and Zanzibar, His Place in the History of Arabia and East Africa. London: William Ouseley. [Google Scholar]
  • Sheriff, A. 1987. Slave, Spices and Ivory. London: James Currey. [Google Scholar]
  • Sheriff, A. 1995. The History and Conservation of Zanzibar Stone Town. London: James Currey. [Google Scholar]
  • Sheriff, A. 2010. Dhow Cultures of the Indian Ocean: Cosmopolitanism, Commerce and Islam. London: Hurst & Company. [Google Scholar]
  • Smyth, W.H. 2005 (1991). The Sailor’s Word-Book. London: Conway Maritime Press. [Google Scholar]
  • Sulivan, G.I. 2003 (1873). Dhow Chasing in Zanzibar Waters. Zanzibar: The Gallery Publications. [Google Scholar]
  • Weismann, N. 2002. “A type of ship on the Indian Ocean in the fifteenth and sixteenth centuries.” The Mariner's Mirror 88: 132143. [Taylor & Francis Online], [Web of Science ®][Google Scholar]
  • Vosmer, T. 2010–2011. “Maritime graffiti and imagery.” In Pride 2010–2011: Celebrating Oman's Maritime Heritage, edited by H. Kirkbride, 102109. Muscat: al-Roya Press and Publishing House. [Google Scholar]
  • Whitehouse, D., Whitcomb, D.S. and Wilkinson, T.J. 2009. Siraf: History, Topography and Environment. Oxford and Oakville: Oxbow Books. [Crossref][Google Scholar]
  • Whitewright, J. 2011. “Maritime Rock Art from Wadi Quseir al-Qadim.” In Myos Hormos — Quseir al-Qadim: Roman and Islamic Ports on the Red Sea. Volume 2: Finds from the Excavations 1999–2003, edited by D. Peacock and L. Blue, 199205. Oxford: Archaeopress. [Google Scholar]
  • Zanzibar. 1952. A Guide to Zanzibar, A Detailed Account of Zanzibar Town and Island, Including General Information about the Protectorate, and a Description of Itineraries for the Use of Visitors. Zanzibar: Government Printer [Google Scholar]

Alternative formats


Related research

People also read lists articles that other readers of this article have read.

Recommended articles lists articles that we recommend and is powered by our AI driven recommendation engine.

Cited by lists all citing articles based on Crossref citations.
Articles with the Crossref icon will open in a new tab.