Home Maritime Resources
Ship Models Machine Tools Other Stuff Site history

updated 11/12/08

Canoe from the Gilbert Islands (Kiribati), Micronesia

click on pictures for larger images

(1) Prototype in the
Ethnographical Museum
Berlin-Dahlem ©

(2) Map of the
Gilbert Islands
(3) GoogleEarth view of
(4) GA sketch (KOCH, 1965) (5) The model in its glass case
(Scale 1:87)


Historical Notes

The picture Europe has (had) of the Southern Seas is still very much the product of the great voyages of the discoverers of the 18th Century: Wallis, Cook, Bougainville, to name just a few of more the important commanders. Experiences made and impressions gained by various travellers found their way into the World literature, think of Melville, Stevenson or Chamisso. Fine art, particularly the paintings by Gaugin, also shaped our mental picture. The liberal social customs of the (female) islanders fired the imagination of the male European population - who doesn't know the story of the mutiny on HMS Bounty !

The interest in and the scientific instrument to investigate foreign cultures slowly developed over the past 200 years. So much of the material and spiritual culture of Oceania disappeared or was replaced (often deliberately) by European goods and concepts before it could be documented. A process that accelerated since the early decades of the 19th century due to increasingly frequent contacts with European and US whalers, traders and missionars. This affected also the perhaps most important cultural and technological achievement of Oceania, namely the sea-going boat and the navigational traditions without which the settlement of the Pacific Islands would not have been possible.

Based on the geographical situation and the ethnographic results of a complex history of settlement (e.g. IRWIN, 1994) today we distinguish the three mega-regions of Melanesia (i.e. the islands of the dark-skinned people), Polynesia (i.e. the region of the numerous islands), and Micronesia (i.e. the region of the small islands). The cultures of Poly- and Micronesia are particularly characterised by the absence of metals and pottery, which is due to lack of suitable raw materials. In pre-European times all artefacts were made without the use of metals. Considering the type and limited availability of raw materials the boat building techniques were at an amazingly high level of development. In many regions sophisticated techniques for assembling big boats, or rather small ships, from small planks had been developed. However, such boat largely had ceased to be built before the first Europeans had the opportunity to view them. Nevertheless the first travellers were suitably impressed by their size and numbers. Today, traditional boat building is only practised in a few remote regions that do not have the resources to acquire the products of Western technology.

These important cultural achievements, as consequence, have almost completely disappeared from the daily life and relegated to museums (e.g. the Bernice P. Bishop Museum on Hawaii). The Ethnographical Museum in Berlin is lucky enough to have collection of several of the most important boats from Oceania. Most have been collected during the early years of the 19th century, when parts of Micronesia and Melanesia were colonies of the German Empire. Others have been built to order for the museum after WWII or collected during the expeditions of its then director Gerhard Koch. Next to original specimens and models in other museums (in Europe e.g. those in Hamburg, Bremen, Paris, London) provide the field research and compilations by HADDON and HORNELL (1936-38) and NEYRET (1976) the most important source for the prospective modeller. In addition there is a great number of monographs and journal articles concerned with the material culture of the various archipelagos, including their boat building traditions. An overview over this literature provides e.g. the bibliography by GOETZFRIDT (1992).

The Prototype

The sail-powered outrigger boat (Fig. 1) from the Gilbert Islands (Fig. 2) attracted my attention already during my very first visit to the museum in Berlin in 1973. With permission by the museum I have been able to photograph the boat with all its details and to make sketches and measurements in preparation for the model (Fig. 5) in 1995. The most important scientific study on these boats is that of GRIMBLE (1924), whose observations have been largely confirmed by DREWS (1945). According to the field research by KOCH (1961, 1984) in the 1950s the techniques did not fundamentally change, but the construction of the museum boat, which was built in 1951 and transferred to the museum in 1964, was somewhat simplified in comparison to the sources quoted.

The culture and, hence, the boat building tradition was largely determined by the scarce resources of the low coral islands of the Gilbert archipelago (Fig. 3). Nevertheless, the indigenous pandanus, coco nut and mangrove provided suitable raw materials for boats, though really suitable timber is extremely scarce. To become boat owner required the acquisition of a suitable supply of timber first. In spite of those limitations the Gilbert islanders managed to develop the fastest sail-powered boats of World that are not based on modern science and engineering technology. Given the right wind and sea conditions on the inner lagoons of the atolls, these boat reach a speed of 20 knots

The boats from the Gilbert Is. (Fig. 4) differ from most of the other boats of similar size in Oceania in that they are not real dugouts or 'extended' dugouts, but are constructed from planks on a keel (Fig. 6); though the keel could be possibly interpreted as a rudimentary dugout (c.f. GREENHILL, 1976). This rather sophisticated technique probably was developed by necessity, as no real trees were available on these poor islands. In the old days the planks were shaped from the timber using axes and adzes. The latter traditionally having blades made from the Tridacna spec. mussel. In historic times tools made from imported steel or even imported tools were used. Given the rather knotty and crooked trees the rule was: one tree trunk = one plank. Since ready cut planks became available from European and particularly North American sources these were preferred and the sides of the boats built up from continuous strakes. It was interesting to observe that the old techniques were revived when timber imports stalled during WWII.

(6) Planking diagram

(7) Keel piece set up
(8) Cross section
(9) Rabetted stem piece
(10) Spacer sticks
(GRIMBLE, 1924)

The following paraphrases the prototype building description by GRIMBLE (1924). After various preparations of logistic and spiritual nature the building proper began by propping up the three-sided keel (Fig. 7) that receives a groove along its central line to act as a bilge (Fig. 8). Ideally the keel is made up from one piece. Trestles and pegs coerce the timber into the desired shape, that is a slight hog. In addition, it is made slightly concave, opening towards to the outrigger, the objective of which is to counteract the drag caused by the outrigger. The stems (the boat is symmetrical) also have a triangular cross-section an receive a shallow rabbet (Fig. 9). The actual shape of the stem pieces are determined by the available timber, but the builder tries to achieve a symmetrical appearance. Keel and stem pieces are butted together only. The fit is tested using a touching-up paste made from soot. The pieces are lashed together using twine made from coconut fibre. The passage of the cord is recessed to limit chafing when the boat is hauled ashore etc.

The hull is built up by the shell-first method. The main cross-section is slightly asymmetric (Fig. 8), as with most outrigger boats, in order to counterbalance the turning moment of the outrigger. The side facing away from the outrigger is almost flat and vertical. This asymmetry is obtained by a simple former that spreads out the keel-planks (Fig. 10). The strakes run parallel to the keel (Fig. 6). Their individual pieces are carefully fitted to each other using the touching-up paste. The planks are lashed together along their edges using twine made from coconut fibre and caulked with pandanus leaves that had been chewed(!) through and oiled. The hull is coerced into its final shape using thin branches as spacers. Once the master boatbuilder is satisfied with the shape, half-frames are fashioned from suitable pieces of wood and tied down with twine (Fig. 8). All cord running on the outside of the hull are recessed.

The frames reach somewhat above the topmost planks (Fig. 8) in order to provide a fixing point for two pairs of stringers that run in- and outboard along the sides of the hulls and fashioned from rather straight thin branches.

A further stiffening of the shell is provided by the various members of the outrigger that run across the wales (Fig. 11). The main beams are recessed into the 'wale'
(Fig. 8). Whenever members of the outrigger cross each other or suitable parts of the hull, they are tied together using coconut twine. The main beams also serve as seats when paddling and steering. The outrigger parts are constructed from well-seasoned pieces of the coconut palm tree (Fig. 13). As the native wood has a too high density to be useful as float, the islanders have to rely on flotsam. Suitable pieces are treasured to the day they are needed for this purpose. Each archipelago and each island has its own technique for joining the float with the outrigger. Here short, Y-shaped pieces are lashed on top of the float first (Fig. 14). The main beams are then lashed to these pieces (Fig. 15).

After completion, hull is sanded down using shark or ray skin and receives protective coatings of oil and a lime wash.

(11) Outrigger diagram

(12) Rig
(13) Outrigger
(14) Float
(15) Float
(16) Sail
(17) Yard-arm
(GRIMBLE, 1924) Ethnographical Museum Berlin-Dahlem

The triangular sail originally was woven from strips of pandanus leaves (Fig. 16). The process of bleaching, beating soft and weaving is very labour-intensive and sails cut from imported canvas had replaced the traditional ones for decades. Canvas, however, has the disadvantage of drying more slowly after the boat should have happend to be ditched than sails made from oiled pandanus leaves. Today, sails made from synthetic fibres are preferred. The sail is bent to a 'yard' and  a 'boom' using short pieces of twine (Fig. 16). A rather massive piece of wood with several depressions serves as mast spur. These depressions allow for varying trims of the rig. The 'standing' rig consists of two stays and a branch with a crotch at its end that supports the mast on its windward side (Fig. 12 and 13). There are also a couple of shrouds fastened to the outrigger. The yard is hoisted by a halliard running through a hole in the mast. The sheet is not directly bent to the boom, but to a primitive block running along a bridle, but not actually forming a purchase. The shrouds may be rigged in a similar fashion. Perhaps this is an imitation of European rigging practice. The yard is either made from a long branch ending in a crotch, or has a short piece with a crotch tied to it. This crotch rests on the forward-facing of the two short cross-pieces that are provided at each end of the boat (Fig. 17).

When going through the wind the sail is shifted around and the crotch rested in the cross piece on the other bow. The boat is always sailed with the outrigger wind-wards. It is steered with a paddle fashioned from a single piece of wood. The paddle is tied to a special rest on the respective stern of the boat. Conversely, the paddles used for propulsion are built up from two pieces of wood. The boat's equipment is completed by a bailer and a pierced big chunk of coral serving as anchor.

The Model

The model was constructed in 1:87 or HO-scale, mainly because the German company PREISSER produces figurines that can be used to create the 'islanders'. The hull of the model was shaped from Westindian boxwood (Costelo). Costelo is rather similar to the European or Asian boxwood in grain, texture and workability, but rather cheaper. On the PROXXON table saw a billet of 6 by 7 mm cross-section and of 95 mm length was cut from the stock. The further shaping was carried out very much like you would do on a solid hull model, but on a rather small scale. Side elevation and 'deck' plan were constructed on the computer from the available drawings. This has a number of advantages: the endless reproducibility also at different scales; mirror images of parts can be produced without loss in quality; etc. The billet was marked with an inscribed centre line for reference and printouts of elevations deck plan stuck on with just a bit of PVA glue. The shape of the hull was then roughened out on the table saw first. For the next step the stationary drilling machine was converted something like a drum sander by mounting one of those expanding mandrels for emery paper tubes. The hull billet was passed along the drum free-handedly to round off the plan view. Both ends were cut off according to the rakes of the stem pieces. A couple of paper templates serve as guides for sanding the billet to the desired unsymmetrical V-shape. The hull surface is smoothed down with emery paper of increasing fineness. The hull is now ready for hollowing out.


Scenic Display

The hollowing out is done with various round and flame-shaped milling cutters mounted in the mini-drill. The remaining thickness is easily controlled by holding the hull against a light source. Ideally it should shine through uniformly over the whole hull. Of course the area of the keel is left thicker. Prototype hulls are only 2 or 3 cm thick, which translates into about 0.3 mm at the chosen scale ! The inside of the hull is carefully smoothed using rotary diamond tools and rubber-bound grinding wheels (CRATEX®-wheels). On the outside the planks seems are lightly incised using a knife-shaped engraving tool. These lines are later picked out in dark ink. The recessed ties between the planks are also marked using the engraving tool (Fig. 7). The wood was lightly stained with a dark grey wood stain. Costelo does not take stain very well, which was actually of advantage here as a slightly weathered appearance was intended. The surface was sealed using wood primer (Schnellschleifgrundierung G1 from Clouth, Germany) and lightly rubbed down using no. 000 steel wool.

The float was fashioned from a short stick of beechwood, its rougher appearance than boxwood being more akin to the prototype wood. The tree 'branches' that make up the outrigger and strengthening pieces of the hull are made from copper and brass wire of different thickness. To this end pieces of appropriate length were rather abused by bending, straightening, hammering and covering coarsely in electronic soldering tin - until they looked somewhat like branches. Were needed also crotches were soldered on. These wire branches then received a base coat of HUMBROL British Army Desert Yellow that comes close in hue to the woods used.

All parts of the outrigger were tied together prototype fashion ! The material used for this is very fine two-ply twisted polyester thread as once used for stitching up ladies' tights. Since mending tights seems to have gone out of fashion, new supplies of this thread seem to be impossible to come by. If you are lucky it may turn up on flea markets (sellers always seem to raise an eyebrow when a man buys such stuff ...). The thread is dyed using mahagony wood stain, which gives it the colour of coconut twine. Several length of this material are prepared and the ends stiffened with a drop of zapon lacquer (also from Clouth in Germany) to facilitate threading. I prefer this lacquer over cyanoacrylate glue because it is easier to apply controlled (with a toothpick or an old paintbrush), does not degrade to this whitish powder - and can be re-dissolved in an emergency. The bindings are done as per prototype, for which sketches were available from the published literature and my own observations. Some compromises had to be made, as the thread is extremely fine though, but still some five times overscale. Therefore only two rounds were taken at each binding as opposed to several in the prototype. The knots were secured with a fine drop of lacquer.

The whole boat was given a thin coat with light-brown furniture varnish (CLOU® Möbellasurlack), which makes the appearance more uniform and gives it warm tint. The hull again was lightly rubbed down with steel wool to give it a silky appearance and to provide a better key for the final wash in dilute white water colour. This will appear patchy by intention.

The paddles were carved from thin Costelo shavings. The handles were reduced to the correct diameter (0.5 mm) using a home-made drawing plate. The bailer (2 mm long!) with its characteristic inside handle was carved from a scrap piece of Costelo.

The prototype pandanus leave matte sails have a rather coarse appearance compared to canvas, but still were of incredible fineness as is evidenced by museum samples. Starting material for the model sail was the very fine silk weave that is used for model airplanes. A scaled copy of the sail plan was stuck to a piece of stiff cardboard that then was covered with cling-film. Strips of the silk were cut of a width that is to scale with the width of the individual mattes from which the sail is composed. This strips were laid out on the sail plan with a narrow overlap to represent the seams and pinned down outside of the sail. The whole arrangement was then lightly brushed in the wood-primer mentioned. This stuff has many good properties including of drying within a few minutes. Once dried the sail is cut out allowing a margin for turning over as doubling. The doubling is bent over a thin copper wire laid into the seam and tacked down using cyanoacrylate glue - the only place were CA was used. The wire allows to sculpt the sail and prevents the silk from ripping when bent to the yard and boom. The next step was painting the sail in acrylic paints, whereby the base coat was applied by airbrush. The texture of the silk was somewhat highlighted by 'dry brushing'. As the matte sails have the appearance of polished wood due to their repeated treatment with coconut oil for water-proofing was model sails were given a light coat of semi-gloss acrylic varnish.

The model was finally rigged up using the thread mentioned above and all knots secured with drops of lacquer.

The Figurines

Human figures provide the necessary scale and illustrate the mode of use of such boats. A scenic display illustrates also the environment in which these boats were built and used.

The well-known German company PREISSER makes a range of nude figurines, called 'Adam & Eve', in various scales. Even those at the 1:87 scale are broken down into individual bodies, arms, legs and heads, allowing the composition of different postures. The tooling is incredibly detailed, allowing even to distinguish individual finger joints ! If one can trust Western photographers and their selection of models, many Polynesian and Micronesion were well-proportioned by 'classical' standards, often athletic, but sometimes with a slight inclination to obesity. My personal theory on this is that only people of a robust constitutions with certain reserves of body fat would have survived the arduous voyages during their migration over Oceania, but I cannot reference this scientifically. Anyway, the 'academic' proportions of the PREISSER set makes them very appropriate for those islanders.

Photographs in KOCH (1961) and KENNEDY (1931) show most people still wearing the traditional Pandanus leave skirt. It is not certain, of course, whether they did this to just please the ethnologist or whether their limited cash would not allow them to buy Western goods. The otherwise ubiquitous printed cotton had replaced the traditional tree leaves and bark tissues already on Gaugin's pictures. The scenic display tries to capture some epic time before Western goods swamped the markets, particularly in the wake of WWII.

From the supply of body parts I composed three figures. Two of them operate the boat, while the third one is of more decorative nature. In many regions fishing boats were 'tabu' for women, but they operated boat themselves to cross lagoons or even to travel between islands. The skirts or aprons are made from no. 000 steel wire and very thin paper. The body parts were assembled using liquid polystyrene cement, which was also used to attach the pieces of clothing. Gaps were filled using HUMBROL putty, which was painted over with the liquid cement to smooth and blend-in the joints. The figurines then were given a base coat in a dark-flesh artists' acrylic paint using the airbrush. This base coat was left to set for several days. Shadows and lights were added using appropriately toned down and diluted acrylic paints. A light touch from the airbrush in a reddish-brown translucent paint particularly in the shadows adds depth and unity to the paintwork. The the flesh parts of the figurines were finally varnished in a mixture of gloss and matte acrylic, while the fabrics were left matte (Fig. 10).

The Scenic Display

It is helpful to think of a 'story board' for the scene first. Thus, I have been thinking of a sunny day with good, but not too strong breeze blowing across the inner lagoon of one of the islands. Consequently, there ripples and a bit foam, but no high seas. The water is only a few metres deep, so it would appear crystal clear and in a more greenish than deep blue colour. The scenery forms an integral part of the glass case described in the following section.

While one has to be cautious using clear plastics for representing the sea, this is case where it is almost a must. By chance I came across a piece of daylight fluorescent acrylic glass and thought this would make a nice effect.  The fluorescence becomes only visible at the cut edges. However, the acrylic artist's gel used to model the ripples and wavelets has a refractive index close to that of the acrylic glass. Thus some of the fluorescence appears on the surface of the 'water'. The acrylic glass plate is cut to size to fit on the base of the glass case. Work continued with milling a suitable recess for the boat using various mills and diamond tools in the mini-drill. In the next step the water was modelled. The base of the case was painted in something like a turquoise colour and, after the paint had thoroughly dried, the acrylic glass plate was glued on. Then the boat was set into the recession and stuck into place using acrylic gel and the water modelled around it. Finally white highlights were set on the waves, around the boat and the steering paddle using artist's acrylic paint.

The Display Case

The idea of a glass case with a slanting front came to me when I saw some jewellery display and I thought it makes an attractive change to the usual brick shape. As one views the main pane almost perpendicularly this reduces also reflections. Also, this semi-pyramid shape is rather appropriate to a sailing boat.

Since having read McNarry's book on Miniature Shipbuilding I have constructed my glass cases all on the same principle. The brass edges have largely decorative function, the structural strengths comes from glueing the glass panes together. McNarry uses clear silicone sealant, which I also did until one Saturday afternoon I discovered that the cartridge had set. I then tried in my desperation the general purpose glue UHU® and literally 'stuck' to it. To date my glass cases have survived several moves around Europe.

Glass is not exactly cheap and oblique cuts and precise dimensions add to the cost. In fact, the display case is the most expensive part of the model, not counting the work on it. I have the glass cut in the shop and go over the edges with emery paper, making sure that no grains get onto the surface. The panes are carefully cleaned using first water with washing-up liquid and then white spirit. The case is then provisionally assembled using clear tape. making sure not to touch the edges that need to be glued and ensuring proper alignment. One pane at a time is taken off the assembly and stuck back on, the top being the last piece. The glue takes some time to dry, so there is enough time to correct the alignment.

I begin with the glass part of the case because it so much easier to cut wood to exact dimensions than glass. The base is built to fit the glass case. It usually consists of a piece of thick plywood or builders board around which a frame like an inverted picture frame is constructed. I love the combination of mahagony and brass and so dye and varnish the wood to resemble this wood.

With the glass case in place on the base, the 'frames' of L-shaped brass profiles can be fitted. I begin with the socket around the case, which is actually stuck to the wooden base rather than to the glass. The rough mitres can be cut with a pair of nippers. Fine fitting is done by grinding down the mitres on the PROXXON pillar drill fitted with an abrasive stone and a mitre guide. Carefully done, the joints become almost invisible. Before glueing on, the brass is de-oxidised using fine steel wool and degreased by wiping it down with acetone. The different pieces are provisionally attached using sellotape and the fit adjusted step by step. Once I am satisfied with the fit, the pieces are taken off one by one and glued down.


DREWS, R.A. (1945): A Gilbert Islands Canoe.- American Anthrop., New Ser., 47: 471-4.

FALCK, W.E. (2002): Boote der Gilbert-Inseln (Kiribati, Mikronesien).- LOGBUCH, 2/2002: 146-150.

GOETZFRIDT, N.J. (1992): Indigeneous Navigation and Voyaging in the Pacific — A Reference Guide.- Bibliographies and Indexes in Anthropology, No. 6, New York etc. (Greenwood Press).

GREENHILL, B. (1976): Archaeology of the Boat.- 319 pp., London (A. and C. Black).

GRIMBLE, A. (1924): Canoes in the Gilbert Islands.- J. Royal Anthrop. Inst. ...: 54: 101-39.

IRWIN, G. (1994): The Prehistoric Exploration and Colonisation of the Pacific.- 240 S., Cambridge (Cambridge University Press).

IRWIN, G. (1998): The Colonisation of the Pacific Plate: Chronological, Navigational and Social Issues.- J. Polynesian Soc., 107(2): 111-43.

HADDON, A.C., HORNELL, J. (1936-38): Canoes of Oceania.- 3 Vols., Bernice P. Bishop Mus. Spec. Publ., 27/28/29, Honolulu, Hawaii (Reprint 1976).

KOCH, G. (1965): Die materielle Kultur der Gilbert-Inseln.- Veröff. Mus. Völkerk. Berlin, Neue Folge 6, Abt. Südsee III: 216 pp., Berlin (SMPK).

KOCH, G. (1984): Boote aus Polynesien und Micronesien.- in: KOCH, G. [Hrsg.]: Boote aus aller Welt — Katalog der Ausstellung 30.01.-05.05.1985, Berlin-Dahlem: 11-31, Berlin (SMPK/Frölich & Kaufmann).

NEYRET, J. (1976): Pirogues Océaniennes, Tome II — II. Polynésie, III. Micronésie, IV. Indonésie, V. Inde, VI. Autres Continents.- 315 pp., Paris (Assoc. des Amis des Muséés de la Marine).

Contact: webmaster at wefalck dot eu

Home Maritime Resources
Ship Models Machine Tools Other Stuff Site history Top of Page