BOROBUDUR SHIP RECONSTRUCTION: DESIGN OUTLINE

The intention is to develop a reconstruction of the type of large outrigger vessels depicted at Borobudur in a form suitable for ocean voyaging DISTANCES AND DURATION OF VOYAGES and recreating the first millennium Indonesian voyaging to Madagascar and Africa. Distances: Sunda Strait to Southern Maldives: Approx. 1600 n.m. The vessel should be capable of transporting some Maldives to Northern Madagascar: Approx. 1300 n.m. 25-30 persons, all necessary provisions, stores and a cargo of a few cubic metres volume. Assuming that the voyaging route to Madagascar was via the Maldives, a reasonably swift vessel As far as possible the reconstruction will be built could expect to make each leg of the voyage in using construction techniques from 1st millennium approximately two weeks in the southern winter Southeast Asia: edge-doweled planking, lashings months when good southeasterly winds can be to lugs on the inboard face of planks (tambuku) to expected. However, a period of calm can be secure the frames, and multiple through-beams to experienced at any time of year and provisioning strengthen the hull structure. for three-four weeks would be prudent. The Maldives would provide limited opportunity There are five bas-relief depictions of large vessels for re-provisioning. It can be assumed that rice with outriggers in the galleries of Borobudur. They sufficient for protracted voyaging would be carried are not five depictions of the same vessel. While from Java. the five vessels are obviously similar and may be seen as illustrating a distinct type of vessel there are differences in the clearly observed details. The depictions are probably not all by the same artist. Following van der Heide (1929), I shall use the plate (Afbeelding) numbers from van Erp’s (1923) paper to identify the Borobudur ships. Volume Component Calculation basis Amount Weight

<2 tonnes Water for approx Thirty persons, twenty-five days, not 1500 including drinking and 33m less than 2 litres per person/day litres containers, cooking

Average 50kg per person Ships' Sleeping space 1.6m x 0.45m for 25 1.5 tonnes compliment persons 182m

Thirty persons, sixty days, 0.5kg per approx Rice day 900kg 1.53m

A fire would be kept smoldering at all Firewood for times when a stronger fire was not ~1 tonne 23m cooking required

Salt fish, plantains, tubers, tamarind, ~0,5 tonne 0.53m Other foodstuffs etc.

Approx 2 tonnes of spices and other Cargo ~2 tonnes 43m high-value commodities

Personal belongings, e.g. sleeping ~0.5 23m Other mats, tools; spare ship's gear; tonnes

The total stowage space indicated by these However, these are not sailing vessels — they are approximations is 133m. The weight calculated is motorised and use multiple stays to hold up the approx 8.5 tonnes. Depending on design, the vessel outriggers. Sailing vessels impose much more load might also need to carry a tonne or two of ballast. on their outriggers and outrigger booms. The Approximately 182m of sheltered space would be largest sailing outriggers of recent decades have required for most of the ship’s company to sleep. been the perahu sande and perahu pangkur of the It is assumed that some persons would be on watch Mandar people from the west coast of Sulawesi, at all times. and some very large jerangkat built on the western coast of the Gulf of Bone and used as FADs (Fish MAXIMUM SIZE OF OUTRIGGER SAILING Aggregating Devices) anchored in very deep water VESSELS to the south of the Gulf. These vessels have been There are vessels without outriggers depicted at up to about 10-11m length and the largest pangkur Borobudur but the five large vessels depicted in have been able to load more than 5 tonnes. detail all have outriggers. In the first half of the 20th century, outrigger perahu The Borobudur ships appear to be fairly large. paduwang from Madura were probably slightly Sizes up to 25m have been posited for the largest larger than the vessels mentioned above. However example: Erp 6. Heide (1929) offers a more sober they seem to have regularly employed human estimate of 12-15m based on the number of oar ballast on the weather outrigger to enhance ports. This allows about 1m for each oarsman. stability. In recent decades some very large outrigger vessels The largest pangkur, sande and jerangkat had very have been built in Indonesia and the Philippines. large and long bamboo outriggers. Pangkur and boom is increasingly susceptible to breaking off when pitching into a headsea.

A Selat Badung jukung with a 6m long hull has outriggers 9m long, mean diameter at least 125mm and buoyancy of about 100kg. (0.0625 x 0.0625 x π x 9m = 0.1103m) Using these proportions and assuming that the largest bamboos would have diameter about 200mm, the largest possible outrigger canoe is about 10m long. perahu pangkur Extrapolating from that calculation, outriggers of sande usually have outriggers more than 150% the about 300mm diameter with buoyancy of about length of the hull. They normally depended entirely 1.3 tonnes would be required by an outrigger vessel on outriggers for stability and sometimes made a little more than14m long. Such a vessel, designed fairly long, open-sea voyages. with a long narrow hull would scarcely have the The maximum size of pangkur, sande and jerankat capacity to carry the proposed passengers, is probably determined by the maximum size of provisions and cargo. Therefore it can be bamboos available. Since the sail area that imposes concluded that it is not possible that an outrigger heeling loads increases as the second power of the vessel, designed to derive all its stability from length of the vessel, the volume (buoyancy) of the outriggers, could be built of a size large enough to outriggers needs to increase at the same rate. In carry the proposed passengers, provisions and other words the diameter of the outriggers must cargo. increase in proportion to an increase in hull length. This leads to engineering problems as size The outriggers shown on the Borobudur ships are increases: if the length of the outriggers increases not like the long outriggers of the large outrigger in proportion to the hull length, the long projection vessels of the 20th century. They appear short and of the outriggers forward of the forward outrigger small relative to the size of the ships.

THE DESIGN AND PURPOSE OF THE BOROBUDUR OUTRIGGERS

The relatively small outriggers of the Borobudur ships, with their short projections forward and aft of the booms would be relatively robust but they would provide little buoyancy and stability relative to the size of the ships and their sail area.

Single or double outriggers? In four of the five Borobudur ship depictions only the windward side of the vessel is shown. Whether 19th century perahu paduwang the vessels are double outriggers with another dhoni means “pilgrimage boat”. Single outrigger canoes are used on some parts of the north coast of Central and West Java. They have a single outrigger boom and they shift the outrigger from one side to the other when changing tack. The arrangement is simple, loose, and temporary in appearance. It is not suitable for larger sea-going vessels. The Borobudur ships have three or four outrigger booms of complicated construction: it does not look like an arrangement that could be shifted when changing tack. Large single outrigger, non-shunting, canoes were outrigger on the leeward side or single outriggers built at Macassar, South Sulawesi. These vessels carrying a flying outrigger on the windward side were built exclusively for racing and were a single only cannot be determined with certainty. outrigger version of the double outrigger However, Erp 10, which has its sail partly furled, jerangkat. They had one very large outrigger and has that sail on the side of the mast towards the employed a lot of movable human ballast when viewer which suggests that we are looking at the racing (Collins 1936). leeward side. Four of the five depictions show the It seems unlikely that the Borobudur ships were port side of the vessel but Erp 9 shows the single-outrigger craft. starboard side so outriggers are seen on both sides Double outrigger canoes are not widely used on of the hull though not simultaneously. the coasts of Java but they are more common on Most single outrigger vessels are designed to tack neighbouring islands including Madura and Bali by “shunting” in order that the outrigger remains where sophisticated designs exist. on the windward side when they change tack. When a vessels tacks by shunting it reverses The Borobudur outriggers have a number of direction and reverses ends — the bow becomes characteristics that make them significantly the stern and vice versa — the steering gear must different from the outriggers of sailing canoes of be shifted from one end to the other and the rig more recent times. reconfigured to drive in the opposite end. Such vessels are necessarily longitudinally symmetrical Length of the outriggers — the ends are very similar — and they have a The outriggers of Indonesian vessels, including simple rig usually with the mast positioned those first depicted by Europeans in the late 16th midships. century, have generally been similar in length to The Borobudur ships do not exhibit longitudinal the hulls of the canoes they were fitted to. In many symmetry and do not have rigs that could be easily cases they are longer than the hulls. The Borobudur reversed. ships all have outriggers shorter than the waterline There was a type of a large single outrigger vessel length of their hulls. Erp 6 (which appears to be from Sri Lanka, the yatra dhoni, which did not the largest vessel) has an outrigger only 0.54 the shunt. Unfortunately the last of these fell into length of the hull. Erp 8 has the longest outrigger disuse early in the 20th century and it is not known at 0.79 the hull length, it has four outrigger booms how the yatra dhoni was operated. The hull form while the other vessels have three outrigger booms. of the yatra dhoni had enough beam and stability if ballasted to sail without an outrigger (Vosmer Erp No. Outrigger length as decimal 1993). It seems possible that the outrigger could fraction of hull waterline length be shifted from one side to the other (with some Erp 6 0.545 difficulty) and that the yatra dhoni was used Erp 7 0.585 mainly for long monsoonal voyages to and from Erp 8 0.79 India when wind direction might be consistent for Erp 9 0.64 the entire duration of the voyage. The name yatra Erp 10 0.750 On all but one of the Borobudur ships the there is another outrigger on the inboard-underside outriggers are doubled — there are two outrigger of the boom. components which I presume are bamboos, one The booms or connectives (it is not clear which on the inboard side of the outrigger boom/ they are) project below the outriggers. connective, one on the outboard side.

Some of the Indonesian vessels recorded by Captain Paris in the 19th century (Paris 1841, Reith 1992) had outriggers of relative lengths that fell within the range illustrated in the Borobudur ships. They were mostly from the Moluccas and Detail from Erp 6 shows outrigger connectives that project neighbouring regions of Eastern Indonesia where well below the outrigger outriggers have remained relatively short in more recent times. Light weight timber rather than Volume of the Outriggers bamboo is often used for outriggers in Eastern The Borobudur outriggers do not appear large in Indonesia (in some areas suitable bamboo is diameter. They are shown with diameter about the unavailable) and outrigger craft depend on same as that of the outrigger booms or less than movable human ballast for stability. that of the booms. This relative proportion By contrast, in areas closer to Java (e.g. Bali, probably reflects the large size of the ships and Madura) outriggers are usually significantly longer the upper limit to the size of bamboo available. than the canoe they are fixed to. Typically A Borobudur vessel of about 14m length would relatively large and fast sailing canoes have a crew have outriggers about 8m in length: if the of only one or two persons and depend on the outriggers were about 200mm diameter, and the buoyancy and hydrodynamic lift of the lee outriggers were doubled (inner and outer outrigger for stability. outrigger) the buoyancy of the outrigger would The shortest outriggers (0.525 relative to hull be about 0.5 tonne which is significant but not in length) in a survey of sailing vessels illustrated in proportion to the size and sail area of the ship. the literature were on a large kora kora from Dorey off the western end of New Guinea (Paris 1841). Fairings The kora kora had a fairly capacious planked hull Erp 6 shows a fairing on the forward end of the and a length beam ratio of less than 4:1 which outrigger. The other four Borobudur ships have could sail without outriggers if properly ballasted. no outrigger fairings. Fairings similar to those on Most outrigger craft have a quite different type of Erp 6 have been used in recent times so that the hull, built up from an unexpanded dugout canoe bamboo outriggers can cut smoothly through the and therefore of very narrow beam and not suitable water. They are a standard feature of most sailing for sailing without outriggers. outrigger canoes.

Attachment of Outriggers to Outrigger Booms All the features discussed above suggest that the There are several different ways of attaching outriggers of the Borobudur ships were not outriggers to the outrigger booms used in intended to provide stability to the extent that the Indonesia. In some cases there are curved timbers leeward outrigger of double outrigger canoe connecting the outriggers to the booms (these are normally does: they lack volume and therefore the termed “outrigger connectives” by Haddon and buoyancy to resist heeling. They are not faired to Hornell). Whether there are connectives or the cut through the water. The outrigger booms or outriggers are secured directly to the booms, the connectives project below the outriggers and outrigger usually lies under the boom or would cause significant drag to the detriment of connective. Alternatively the boom or connective speed and steering if depressed into the water. Erp penetrates the outrigger. On all the Borobudur 9 shows its lee outrigger apparently flying clear ships there is an outrigger attached on the outboard of the water since the outrigger boom (connective) side of the boom/connective. On all except Erp 8 ends can be seen projecting below the outrigger. If it is concluded that the outriggers are not Larger and wider galleries, running the full length primarily intended to provide stability another of the hull, were a feature of 18-19th century perahu explanation of their use is required. Like the lancang. outriggers of the 19th century bouanga from New Guinea, drawn by Capt M. Paris, the outriggers The galleries seen on the Borobudur ships are might be seats for paddlers to propel the vessel in large, enclosed, and exhibit complex structure. calms and in martial use. Again martial use is suggested with a need to If the outriggers are not designed to be the vessel’s protect the oarsmen from spears and arrows. primary source of stability then the hull form will not be a typical outrigger canoe hull form. Instead All the Borobudur ships except Erp 9 have a deck it will have a broader, more stable and more house with a pitched roof positioned aft of the main capacious hull form. mast. The deck houses appear to be small and are unlikely to represent all the accomodation space in the ship.

THE SUPERSTRUCTURES HULL PROFILES The Borobudur ships all have considerable The upper parts of the hulls of the Borobudur ships superstructures obscuring their hulls. There are are hidden by the galleries and the bow and stern outboard galleries for rowers along the full length screens. of the hull. In the bow and stern tall screens Four of the five Borobudur ships have a straight, surround the stem and sternpost. Heide has forward raked profile to the bow. Erp 9 has a proposed that the bow and stern screens are curved, forward raked bow profile. In Heide’s protective structures for use in warfare. This seems interpretation, Erp 7 and 8 both have an external very likely. Later warships used by Sulu sea pirates stem. Erp 6 does not obviously have a stem in the had a heavier timber screen in the bow to offer lower part of the bow, but there is a high prow some protection from cannon shot. timber projecting above the bow screen. Some types of Indonesian perahu are built without a stem (and also without sternpost). These include the Javanese perahu jegongan. None of the Borobudur ships show a sternpost. However, very little detail of hull structure is

19th C Sulu Sea vessel with bow screen

The tall screens seen on the Borobudur ships would create much windage which would be detrimental to sailing performance and might not be fitted to ships engaged on long voyages. The tall screens are fitted on top of large through beams and wing-like projections in the bow and stern. Both the heavy through beams (polangan) and the wing-like transoms (kopengan) have been features of some Javanese and Madurese vessels until the late 20th century. In some examples those wings support the ends of small galleries (ambeng) Stern of perahu kacik showing kopengan built out from the hull. supporting an aft gallery shown in any of the depictions. All show an inward HULL FORM raking, recurved profile to the stern. This kind of Little information about hull form can be derived profile has remained a stylistic feature of a number from the Borobudur iconography. The bows and of Javanese perahu types. sterns appear to be sharp rather than bluff. The Heide traces some evidence of hull shape below underwater body of the hull is not shown and the the waterline in Erp 6 and 7. The bows do not show bas-relief does not provide a fully three- a cutaway forefoot but in the stern there is a curved, dimensional representation that would allow cutaway heel profile. Again, this is typical of assessment of beam. Javanese perahu design. Indeed the perahu Hull form must be inferred from archaeological konteng of East Java when planked up to increase data and from critical use of more recent freeboard and cargo capacity has a profile much ethnographic data. like a Borobudur ship.

Midships cross-sectional shape A fairly large number of Medieval shipwrecks have been discovered and investigated in Southeast Asia. The vessels discovered at Butuan in the Philippines illustrate a relatively small and sharp type. A larger and more capacious type has been identified as a South China Sea hybrid type combining some Chinese construction techniques with Southeast Asian hull form and edge-doweling of planks. A large wreck discovered near Palembang, Sumatera probably represents a version of the type with entirely indigenous Profile of a planked up perahu konteng structure. All these wrecks exhibit similar cross- sectional shape. They all show hollow deadrise and fairly slack turn to the bilge. None of the Because of the full-length galleries it is not wrecks include the remains of the topsides (with possible to see the rail of the hull and therefore the possible exception of Butuan 5). Ethnographic the freeboard cannot be judged. However, it is data suggests that the topsides were flared rather unlikely that outrigger booms would cross the hull than vertical since vertical topsides are virtually above the height of the rail. Clearly the lower unknown in the region. However, if the rowing outrigger booms pass through the hull; possibly galleries are fitted outboard of the topsides, the the upper outrigger booms rest on the rail. The flared topsides would seem to interfere with the hulls do appear to have considerable freeboard and use of the galleries. the lower outrigger booms appear to be well above the waterline. It is not proposed that the freeboard/ Length-beam ratio, beam-depth ratio length ratio measured from the depictions could The iconography gives no clear indication of the be taken as representing the actual ratio. Distortion, ratio of proportions of the hull. As noted above, exaggerating freeboard, is more or less a standard outrigger vessels usually have very little beam feature of ship iconography. relative to their length and depth. Non-outrigger If the hulls do have considerable freeboard, then Indonesia vessels very often have length-beam they will either have fairly considerable beam to ratio of about 3:1 and beam-depth ratio of not less provide stability or they are narrow canoe like hulls than 2:1. (“Depth” here means moulded depth of deriving all their stability from outriggers. But for the midsection and has no connection to draft.) reasons given above that seems unlikely. These ratios produce a hull with very large righting moment (stability) at low and moderate angles of heel making outriggers totally unnecessary. Indeed, such a hull form, when rolling in a beam sea would tend to immerse and damage outriggers References if they were fitted. The 19th century kora kora illustrated by Paris has a length-beam ratio of 4:1 Burningham, N. 1989. The Structure of Javanese and a beam-depth ratio of 1.8:1. It is proposed that Perahu. The Beagle, Records of the these proportions be used as a guide in developing Northern Territory Museum of Arts and a design for the Borobudur ship reconstruction. Sciences, 6(1): 195–219.

Burningham, N. and Stenross, K., 1994. Mayang: The Traditional Fishing Vessels of Java. The Beagle: Records of the Museums and Art Galleries of the Northern Territory, 11:73–132.

Collins, G.E.C., 1936. East Monsoon, Scribner, New York.

Erp, T. van. 1923. Vorstelling van vaartuigen op de reliefs van den Boroboedoer. Nederlandsch-Indië Oud en Nieuw 8:227– 57.

Heide, G.J. van der, 1929. De samanstelling van Hindoe-vaartuigen: iutgewerkt naar beeldwerken van den Boroboedoer. Nederlandsch-Indië Oud en Nieuw 12: 343–57.

Paris, F.E. 1841. Essai sur la Construction Naval des peuples extra-Européens. Artus Bertrand, Paris.

Piollet, P. 1995. Equipages et Voiliers de Madura. Paul Piollet, Ternant.

Reith, E., 1992. le Voyage de la Favorite, Amiral E. Paris. Editions ANTHÈSE, Arcueil.

Vosmer, T., 1993, The yatra dhoni of Sri lanka. Bulletin of the Australian Institute for Maritime Archaeology, 17.2: 37-42.