Sunday, July 12, 2020

“Woodskin” Canoes of Guyana

Logboats are probably the best-known Amerindian watercraft in Guyana, but another boat type in common use – at least into the first half of the 20th century – is the bark canoe. Although terminology differs among various writers, the term “woodskin” is commonly applied to all Guyanese bark canoes.

Bark canoe on the Mazaruni River, Guyana
Akawai open-ended woodskin on the Mazaruni River (Roth, W., 1924:plate 177) Click any image to enlarge.

The most common type of woodskin appears to have been the one with open ends, which was used by many of Guyana’s Amerindian peoples, both near the coast and in the inland “hinterlands”. This was usually made from the bark of the purpleheart tree (Copaifera pubiflora). Several aspects of its construction are unusual, if not unique.

The tree is felled with the bark still attached. Cuts the desired length of the canoe are then made along both sides of the trunk, then these lineal cuts are connected by circumferential cuts at both ends around the top and sides of the trunk and the bark is pried off with wood wedges. There is thus no need to roll the trunk to get at the surface that rests on the ground. Once it is off the trunk, the bark is propped open with sticks between the opposite edges to keep it from closing up again.

This is quite different from the method of harvesting birch bark for North American canoes. There, the tree is left standing and is not killed by being barked. A single slit the length of the canoe (or as long a piece as the tree allows) is made along the height of the trunk, and cuts are made around the entire circumference of the trunk at the top and bottom of the slit, so that the bark is removed in a single piece that completely surrounds the trunk, thus maximizing its width. This is possible because birch bark is relatively thin and quite flexible, while the bark of purpleheart is so thick and stiff that a full circumference could not be opened up around a single split to remove it from the trunk without cracking.

Detail of bark preparation for bark canoe, Guyana
Outer bark removed (right); inner bark folded (left) (Roth, W., 1924:615)

The purpleheart bark is of two layers – a thick, stiff outer one, and a more flexible inner one. The two are removed together from the trunk, then wedges of the outer layer are cut and removed from both edges 2-3 feet (70-100cm) in from both ends, leaving the inner layer intact. With one man standing amidships, another raises one of the ends so that the flexible inner bark folks in upon itself. Holes are punctured through the four layers of bark and the overlapping sections are stitched together with “bush rope” – presumably thin roots, withies, or possibly natural fibers taken from palm leaves or similar. The other end is treated the same way.

Fully-outfitted bark canoe, Guayana
Fully-outfitted woodskin with inwales, thwarts, spreaders and tightening ropes (Roth, W., 1924:616).

Details seem to differ from one boat to the next, or possibly according to the practices of different communities or Amerindian peoples, but one common modification is the addition of inwales, which are stitched along the upper edge of the bark amidships, and extend into the raised ends below the top edges, where they help keep the open ends elevated above the waterline. Sometimes sitting thwarts are added, suspended by hangers attached to the inwales. Beams are tied in place to keep the sides apart amidships. Conversely, ropes are used to keep the sides from spreading out too far toward the ends.

Dimensions are typically about 15-16 feet (450-500cm) LOA (although lengths of 25-30 feet/8-9m are reported), 4 feet (125cm) beam, and depth 6-8 inches (15-20cm), with freeboard a mere 3-5 inches (8-12cm).

Closed-end woodskin. Top: bark cuts.Closed-end bark canoe, Guyana: construction details.
Closed-end woodskin. Top: bark cuts. Middle: ends folded up. Bottom: outfitted. (Farabee, W. C., 1918:75)

An alternative form, used by the inland Wapisiana Arawak people, has pointed, closed ends. The bark is harvested in the same manner, but after it is removed from the trunk, the top corners at both ends are removed, so that the bark is pointed at both ends. The bark is placed open-side down over a low fire to soften it, then it is expanded and sticks are placed between the opposite sides to keep them spread apart, but apparently not as wide as in the open-ended type. The ends are then folded and raised as above, except that the wedge-shaped sections from which the stiff outer bark is removed are longer, almost touching each other from opposite sides on the bottom of the hull. This seems to produce a hull with a rounder bottom and greater freeboard than the open-ended type.

Bark canoe on Rupununi River, Guyana
Woodskin on the Rupununi River (Roth, W., 1924:plate 179)

Woodskins generally carried one to three people and were used for fishing and general transportation. They drew little water (about 3 inches/8cm), so were useful on shallow and rocky streams, and could be more easily portaged around rapids and falls than heavier logboats. Their low freeboard, however, was a disadvantage because they could afford to take on very little water, the purpleheart bark being so dense that the boat would sink if swamped. Propulsion was with single-bladed paddles, an example of which can be seen clearly in the first photo.

I have found no recent references to woodskin use, but hope to determine whether they are still in use during a planned visit. If you have “on the ground” knowledge, please contact me.



Arnold, B. (2017) Bark-canoes of South America: from Amazonia to Tierra del Fuego (English text without illustrations; French original: Les canoës en écorce d’ Amérique du Sud: de l ’Amazonie à la Terre de Feu). Le Locle: Editions G d’Encre (Le tour du monde en 80 pirogues, fascicule 3).

Brindley, M. D. (1924) ‘THE CANOES of BRITISH GUIANA’, The Mariner’s Mirror. Routledge, 10(2), pp. 124–132. doi: 10.1080/00253359.1924.10655267.

Farabee, W. C. (1918) The Central Arawaks. Philadelphia: University of Pennsylvania Anthropological Publications. Available at:

Roth, W. E. (1924) An introductory study of the arts, crafts, and customs of the Guiana Indians. Washington DC: US Government Printing Office. Available at:

Worcester, G. R. G. (1956) ‘Notes on the canoes of British Guiana’, Mariner’s Mirror, 42(3), pp. 249–251.

Wednesday, April 15, 2020

Boat Iconography at the British Museum #2: Pre-Classic

This is the second in a series of posts on ancient boat iconography at the British Museum. The first post looked at ancient Egyptian boat models. (Click any image to enlarge.)

Naqada jar with sailing ship image
Dating from the second phase of the Naqada culture (3500–3200 BC) in what is now Egypt, this is one of the earliest undoubted images of a boat with a sail from anywhere in the world. The medium-aspect squaresail, hung from a mast stepped far toward the bow, appears to have a boom along the bottom edge. As the mast crosses the sail somewhat off-center, it could conceivably have been a lugsail, although there is no good evidence for the use of lugsails in ancient Egypt, and with the mast so far forward, the boat could only have sailed before the wind in any case, so it was likely used only while traveling upstream on the Nile. No rigging is shown, but this is surely a function of the illustration’s overall paucity of detail, not an indication that none was used. The prow rises vertically very high and the stern is also raised. There is a great deal of rocker and sheer. Just behind the raised stem and beneath the leading edge of the sail is what appears to be a tiny platform: perhaps this was a pilot’s station or a base for a votive image. Aft, vertical posts support a forward-sloping platform, roof, or awning. Rectangular blocks of “waves” all around the boat represent the sea. Here is the British Museum’s record and another photo.

Naqada jar with sailing ship image
The redware pottery jar stands 58.5cm tall. 
Meopotamian bitumen boat model
This big (75cm long) model, from a grave in Ur, in southern Mesopotamia, dates to the Akkadian period (2300-2150 BC). It’s made of bitumen mixed with earth and is very similar in form to plank-built boats called taradas used in the Iraqi swamps well into the 20th century. Taradas and boats made of reeds were both coated with bitumen, which occurs naturally in the area. It’s unclear if the model represents a boat of reeds or planks, but to me it feels more like the former. Grave boats in Ur were originally loaded with containers thought to have held provisions for the afterlife, or possibly as bait to distract evil spirits. British Museum record and photos.

Bronze Age Cyprian jar with ship image
We’ve leapt forward well over a milennium, to 750 BC-600 BC. The vessel depicted on this Bronze Age Cypriot jar has its mast stepped amidships and would have been more capable than the earlier Naqada boat of sailing across or into the wind on the open Mediterranean. The furled sail, of low or medium aspect, has no boom along its bottom edge. Rigging is clearly shown but is hard to interpret. (Guess: the lower, upside-down V represents shrouds; the upper, rightside-up V represents braces.) Both bow and stern turn up abruptly and rise to great heights, with decorative flourishes at their upper ends. There is a large structure in the bow (right side), and a helmsman stands at the stern managing double steering oars or side rudders. Large amphorae, probably containing wine, oil, or fish sauce, constitute the cargo. Just out of the frame to the left side, a crewman defacates over the stern, making this probably the world’s earliest depiction of shipboard sanitary arrangements. The Nautical Archaeology Society uses this image (minus the biological function) as its logo. More about this item in this article from the International Journal of Nautical Archaeology. British Museum record and photos.

Cyprian Bronze Age boat models
Three terracotta ship models from tombs on Cyprus. Top and middle: 600BC-500BC. Bottom: 750 BC-475 BC. With no suggestion of a rig, they appear to be rowing galleys, probably warships, judging by the rams on two of them. Although they’re similar, they all differ slightly in the forms of the ram or prow, the upper extensions of the stem and sternpost, and decoration. All three have oculi and are 16-17cm long. British Museum records and photos: top, middle, bottom.

Cyprian Bronze Age boat models
The starboard sides of the Cypriot galley models.
Cyprian Bronze Age ship model
This terracotta merchant ship from Cyprus (600-500 BC) has a mast step amidship in the bottom, and the vessel was surely rigged with a square sail. The ends of the posts have fishtail-like shapes. The broken parts of the hull aft (right) may have been the location of steering oars. Where the sides bend inward at the top probably represents a bulwark, not a tumblehome hull shape. Forward is a cross-beam that probably served as catheads for anchor handling. British Museum record and photos.

Cyprian Bronze Age ship model
This more elaborate merchant ship model is also from Cyprus (750-500BC) and also has a mast step amidships. There are several cross-beams and an elaborate sterncastle and poop deck, with structures to secure steering oars or side-rudders. As this was a sailing merchant ships, the rows of holes on both sides do not represent oar ports, and they are too low and too numerous to be fastening points for shrouds. I believe they are scuppers that would have been located at deck level, at the bottom of the bulwarks. British Museum record and photos.
Cyprian copper "ox hide" ingot
Not iconography, but an example of an important type of cargo carried by Cypriot ships. This is a 37 kg copper ingot, dated to about 1200 BC. It’s thought that the distinctive "ox hide" shape made them easier to carry. Copper was the primary metal required for the establishment of the Bronze Age. British Museum record and photo.

Wednesday, April 1, 2020

Boat Iconography at the British Museum #1: Egypt

When it comes to studying the earliest watercraft, direct archaeological evidence, in the form of artefactual boats and ships, is extremely rare and fragmentary. In contrast, the iconography of ancient boats -- in the form of models, relief carvings, images on pottery, etc. -- is relatively abundant, and often well-preserved. If you read enough nautical history or archaeology, you'll come across a number of oft-used images that provide some of our best clues about the design and construction of early watercraft. Although subject to differing interpretations, these mostly well-known examples of boat iconography are crucial to current understandings of such fundamentals as when sails were first used, how Egyptians built reed, and then wooden, boats, and what Greek and Roman galleys looked like. Interpretation of fragmentary shipwrecks can be greatly hampered by a lack of relevant iconography.

On a visit to London in January, I realized for the first time how so many of these "iconic examples of iconography" are held in the British Museum. It was like walking into a well-illustrated textbook, and a thrill to see these classic examples up close and in 3D. There are too many to include in a single post, so I'll begin with the Egyptian boat models, all of which are from funerary contexts. Later posts will examine examples from other areas and cultures. Photos were taken through display case glass, so image quality is poor, for which I apologize.

Ancient Egyptian boat model at British Museum
From tomb 56 at the necropolis of Asyut, the burial of Hetepnebi, a local official, 1st Intermediate Period, about 2090 BC. I'm unsure if this represents a papyrus raft or a plank-built boat. Two masts are present (possibly used as towing posts?). Aboard are the owner, a pilot, six oarsmen, who I believe are kneeling, and five soldiers, who stand. Shields and staves are stacked amidships. Paddles and a steering oar have been lost. (Click any image to enlarge.)
Ancient Egyptian boat model at British Museum
From the same tomb: crews of two papyrus rafts. The raft on the left is a replica of the original, which disintegrated. Unlike the oared craft above, these are propelled by forward-facing paddlers, who work from a crouching posture. A pilot originally stood at the bow, and a figure of the owner at the stern.
Ancient Egyptian boat model at British Museum
The museum's online catalog search isn't working and the display signage only indicates that this and the next model are from a tomb from "the end of the Old Kingdom to the mid Twelfth Dynasty", which is roughly 2181 BC to 2000 BC. This model appears to represent a wooden boat, propelled with 8 oarsman, with a pilot in the bow and a helmsman astern. The oarsmen wear a skirt-like garment that covers their legs, making it difficult to say if they crouch or kneel. The model once included a mast, sail, and rigging as well.
Ancient Egyptian boat model at British Museum
Apparently found with the previous model, this one is set up strictly for sail, although two of the crew were apparently using poles (lost) as well. Three other crew manage the rigging, and again there's a pilot forward and a helmsman aft. The owner sits with boxes of cargo beneath the decorated canopy. The rudder arrangement is interesting. The upper end of the stock rests on a post forward of and high above the helmsman; the lower end, just above the blade, rests on top of the transom. A tiller (lost) extended down from the stock between these two pivot points. The significance of the grid-like lines painted on the deck of this and all of the models below is unclear. Did they represent removable deck panels? Perhaps thwarts and a longitudinal strength member?
Ancient Egyptian boat model at British Museum
This funerary boat (12th Dynasty, about 1850, from Thebes) carries the deceased owner's mummy. attended by a mourner, a priest, and provisions for the afterlife. The boat represented was probably wooden, but the upturned ends are reminiscent of papyrus rafts, a design holdover from the older, more "traditional" technology. Twin quarter-rudders are supported at the upper end by an A-frame that is topped by a carved falcon head, and managed by a single helmsman who has two tillers to manage. 

Ancient Egyptian boat model at British Museum
This top view of the previous model shows a painted grid pattern on the deck, similar to the previous two models.

Ancient Egyptian boat model at British Museum
12th Dynasty, about 1985-1795 BC, provenance unknown. With the crew sitting on boxes and posed for rowing, and the boat is rigged for sail, the deceased's soul will be able to travel both upstream and down. 
Ancient Egyptian boat model at British Museum
The same model as in the previous photo.  The square-profile squaresail rig has a yard with multiple lifts and a boom with multiple halyards.The rudder has lifting tackle to raise it in shallow water or for beaching. 

Wednesday, February 12, 2020

Corfu's Reed Raft, the Papyrella

Corfiot papyrella, a reed raft
The papyrella built for the Exeter Maritime Museum, now held by the National Maritime Museum of Poland, photographed in Suffolk, England by Bob Holtzman (click any image to enlarge).
Rafts made of reeds are among the oldest types of watercraft, and remained in use in many areas through the end of the previous millennium “wherever there is a good supply of reeds” (McGrail, 2001:21, 104). Papyrus, among the most common reeds used for raft building, may have been among the earliest as well. Papyrus rafts are known from Egyptian tomb carvings, paintings, and funerary models as early as the Fifth Dynasty (2492-2345 BCE) (Hornell, 1970:47-48; McGrail, 2001:22), and these appear quite finely modeled and highly developed, so that much earlier use can be safely assumed (Hornell, 1970:48-49).
Reed raft under construction in an Egyptian 5th Dynasty tomb relief
Papyrus reed raft under construction in an Egyptian 5th Dynasty tomb relief (McGrail 2001:21).
Modern examples of papyrus rafts are known from Lake Tsana (i.e., Tana) in Ethiopia; among the Buduma and Kuri people on Lake Chad; in Palestine; and elsewhere (Hornell, 1970:53-56). Among the last European reed rafts was the papyrella of Corfu, whose use apparently petered out only in the 1970s. Although the Corfiot name papyrella is clearly etymologically related to papyrus, the reeds from which they were built is disputed. They have been identified as giant fennel, Ferula Communs L. (Tzamtis, 1990:329), although a later observer states that they were Scirpus lacustris L. ssp lacustris, of the same Cyperacea family as Egyptian Cyperus papyrus (Tzalas, 1995:456).

The basic papyrus raft can be built with no tool other than a stone blade for cutting reeds, and with no specialized techniques other than rope-making (McGrail, 2001:21-22), and this goes a long way to explain the papyrus raft’s early appearance in history. The papyrella, however, is a unique hybrid design incorporating some wooden components and requiring a few additional tools and techniques.

Construction began with six to eight light cypress saplings about 3m long. These were tied together at their narrow ends – the end that would be the bow – and splayed to a width of 1.1-1.3m at the stern. Three broad planks were laid beneath this framework, one at the stern, one about a meter back from the bow end, and one halfway between. Papyrus reeds were gathered into bundles with their butt ends all aligned and tied with cordage, so that the bundles were roughly 20cm in diameter at the butt end, narrower at the other end. These bundles were laid over the frame, perhaps six or seven bundles across, in two layers. Three more planks were then placed on top of the bundles, directly above the lower planks. The two layers of planks were then lashed with heavy rope passing through two holes in both ends of each plank, sandwiching the reed bundles and holding them in place. More reeds were bundled and tied into a horseshoe shape on top of the raft to serve as a coaming or gunwale. Finally, the saplings at the bow were bent upward and back and tied in place to form a prow (Tzamtis, 1990:330).

Papyrus is delicate and easily crushed and abraded, and when damaged, it absorbs water readily and loses its buoyancy. To protect the reeds on the bottom of the bundles when the boat was grounded, “common reeds” of a more durable type than papyrus were sometimes placed between the cypress bottom poles. A second set of cypress poles, tied like the first set, was sometimes placed on top of the reed bundles and beneath the top set of planks (Tzamtis, 1990:330). This might have been done to stiffen the structure, although it is unclear if rafts without this feature were subject to excessive flexing. Since the poles run mostly parallel with the bundles, it does not appear to be useful for containing them. A third modification sometimes present was the substitution of short cypress poles for the transverse boards (Tzamtis, 1990:330).

Papyrella raft stern view
Stern of the papyrella (photo: Bob Holtzman)
Descriptions of a few aspects of construction are ambiguous. Tzamtis states that the bottom cross-planks were placed beneath the cypress longitudinals (1990:330), which would hold the saplings in place sandwich-wise. In the example shown, however, the planks are on top of the longitudinals, so the saplings had to have been tied to the boards to remain in place. As one of the final steps, “two cypresses were placed on the stern, and bound there from the bottom to around the top, thus completing the caging in of the papyri bundles” (Tzamtis, 1990:330). This feature did not appear on the example I examined, and the disposition of these two poles on the stern is unclear. Also, the cypress base poles which extended beyond the reed bundles at the bow, were bent up and back and tied in place to bring the tips out of the water and create somewhat of a prow shape. Tzamtis (1990:330) indicates that this was done as the very last step in construction, but Tzalas (1995:444) states they were bent into this shape when green and held under pressure until the wood dried, after which they held the shape on their own. It is unclear if the square stern was the natural result of aligning the bottoms of the bundles, or if the aft ends of the bundles were all sheared even and square after being bound together. There is an account of papyrellas with a rounded stern (Tzamtis, 1990:330), which surely would have been cut to shape after the bundles were bound together.

The entire building process took two to three days. Finished rafts were 2.5-3m LOA, with their greatest beam of 1.1-1.3m at the stern and maximum thickness of 45-50cm, also at the stern (Tzamtis, 1990:330; Tzalas, 1995:443-445).

Papyrella were used for fishing in sheltered waters, “confined to lakes and bogs, rarely faring out to sea and far from the coast” (Tzamtis, 1990:330). Reports refer to their use for lobster fishing, but it would be surprising if other forms of fishing did not occur. There are reports that double-length papyrellas of 5-6m LOA were built by joining two single boats stern-to-stern, and that these larger boats would venture offshore for lobster fishing, but no such boats have been properly documented. Long poles were used to hold the two boats together, but it is unclear if the poles were forced through the bundles of both boats or lashed to their exterior surfaces or frameworks (Tzamtis, 1990:330-331).

Papyrella raft in Corfu with standing paddler using a double-bladed paddle
Papyrella under way in Corfu by a standing paddler using a double-bladed paddle (Tzalas 1995:466)
The standard “single” papyrella was a one-man craft, propelled from a standing position with a double-bladed paddle 2.3-2.5m long, including the two 50cm blades (Tzamtis, 1990:330; Tzalas, 1995:449). With the paddler standing, the gunwales would have done nothing to protect him from waves, so their purpose must have been to protect his gear and his catch.

Fisherman standing a papyrella on end to dry
Fisherman standing a papyrella on end to dry (Tzalas 1995:465)
The more tightly the bundles are tied, the better a reed raft will resist waterlogging and decay (McGrail, 2001:22, 104). The bundles in the example examined were still tightly bound some 50 years after the raft was built. Another key to longevity was to dry the bundles after every use by pulling the raft out of the water and standing it on its stern end. Treated this way, the bundles might last two to three years, while the framework could be used over and over (Tzalas, 1995:443).

Regular use of the papyrella continued into the late 1970s or early 1980s, at which time a single user remained in Palaiokastritsa, in northwest Corfu (Tzalas, 1995:443). During the 1970s, three had been built for museums – one in Corfu, one in Piraeus, and one in Exeter, England (Tzalas, 1995:443), the original home of the papyrella I examined. The Exeter collection was subsequently transferred to the World of Boats collection in Eyemouth, England, and when that museum closed abruptly in 2017, the papyrella and a few other craft were purchased at auction by Valerie Fenwick, a renowned British maritime archaeologist who kept them secure in a barn in Suffolk until a proper caretaker organization could be found. This is where I photographed it and documented its basic features in concert with the fellow student who appears in the photos. The papyrella, along with the rest of the small collection, was recently transferred to the National Maritime Museum in Gdansk, Poland, where, one hopes, it will be stabilized and displayed (Fenwick and Pink, 2020).

For a 50-some-odd-year-old reed raft, it remains in quite good condition. The reeds are beginning to disintegrate, however, and lightweight monofilament fishnet has been fastened over them in an attempt to hold them together. It is unclear if the bundles at the bow were originally larger than in the photos or if they extended further forward. Some of the ropework is sloppy and haphazard, the result, I suspect, of attempted repairs by an individual unskilled in knots.

Harry Tzalas's raft Papyrella
Harry Tzalas's double-ended papyrella, named Papyrella (Sampson, 2018:23)
Certainly the most famous individual papyrella was one named Papyrella and used in an experimental voyage from the Greek island of Melos to the mainland in 1988 (Tzalas, 1995). Seafaring by Mesolithic people has been indirectly but firmly established by the presence on the Peloponnesian mainland of obsidian from Melos, and Harry Tzalas (following Tzamtis’s lead) hypothesized that this was mostly likely accomplished by means of a reed raft. Tzalas had a double-length, double-ended papyrella built which he and five crew paddled from the mainland to Melos over the course of 16 days (Tzalas, 1995). Tzalas claimed that this voyage supported his hypothesis, but I and some others (Cherry and Leppard, 2015:745) find the experiment to be fraught with errors of theory, logic, and methodology and view its results as dubious.

map: route of reed raft Papyrella from Melos to Greek mainland
The route of Papyrella from Melos to the Greek mainland (Cherry and Leppard, 2015:746)


Cherry, J. F. and Leppard, T. P. (2015) ‘Experimental archaeology and the earliest seagoing: the limitations of inference’, World Archaeology. Routledge, 47(5), pp. 740–755. doi: 10.1080/00438243.2015.1078739.

Fenwick, V. and Pink, J. (2020) ‘The Fate of the ISCA Collection: The World’s Largest Collection of Traditional and Vernacular Boats’, Nautical Archaeology Society. Available at: (Accessed: 12 February 2020).

Hornell, J. (1970) Water transport: origins & early evolution. Newton Abbot: David & Charles.

McGrail, S. (2001) Boats of the World: from the Stone Age to Medieval times. Oxford: Oxford University Press.

Sampson, A. (2018) ‘The Mesolithic Hunter-Gatherers in the Southeastern Mediterranean and Their Contribution in the Neolithisation of the Aegean’, Archaeology and Culture, 1(1), pp. 11–36. doi: 10.22158/ac.v1n1p11.

Tzalas, H. E. (1995) ‘On the Obsidian Trail with a papyrus raft in the Cyclades’, in Tzalas, H. E. (ed.) Tropis III: 3rd International Symposium on Ship Construction in Antiquity. Athens: Hellenic Institute for the Preservation of Nautical Tradition, pp. 441–470.

Tzamtis, A. I. (1990) ‘Papyrella: remote descendant of a middle stone age craft?’, in Tzalas, H. E. (ed.) Tropis II: 2nd International Symposium on Ship Construction in Antiquity Proceedings (1987). Delphi: Hellenic Institute for the Preservation of Nautical Tradition, pp. 329–332.