Thursday, November 2, 2017

“Popo” Outrigger Canoes of the Central Caroline Islands

Flying proas of the Caroline Islands, from Admiral Paris
Flying proas of the Caroline Islands, from Adm. Paris. (click any image to enlarge)
In 1983 and 1984, Steve Thomas, who would later host the television series This Old House, lived intermittently on the tiny atoll of Satawal in the central Caroline Islands, an experience he documented in the book The Last Navigator: A Young Man, An Ancient Mariner, The Secrets of the Sea. While there, he lived with and studied under Mau Piailug, a master of the traditional Micronesian art and science of navigation, who had previously gained notoriety as the navigator on the early voyages of Hokule’a (which, as a Polynesian double canoe, was quite a different craft from the single-outrigger canoes of Micronesia to which he was accustomed).

proa from Satawal (From The Last Navigator, by Steve Thomas)
A proa from Satawal (From the cover of The Last Navigator, by Steve Thomas)
The Last Navigator is a good book, an engaging, sensitively-written memoir of a young man attempting to learn about and fit into a very different society – in this case, one whose traditional, preindustrial culture was under extreme pressure from the forces of modernization and Westernization. While it contained what I suspect might be some valuable ethnographic observations and a good general description of traditional Micronesian methods of navigation (for a detailed explanation, see We, The Navigators: The Ancient Art of Landfinding in the Pacific, by David Lewis), it is scanty on information about the boats of Satawal and their construction.

Micronesia was home to a great number of outrigger canoe types – so great that James Hornell, in The Canoes of Polynesia, Fiji, and Micronesia, devoted many pages documenting separately those of the Carolines, the Marshalls, the Gilberts, the Marianas, and certain other of the region’s smaller groups and individual islands. Even within the central Carolines, there was too much variation between islands for it to be practical to document every differentiating detail, necessitating some generalization of a “western and central Caroline” type. Since I came to the subject of the central Carolines canoe by way of The Last Navigator, this post focuses, to the extent possible, on boats most like those that appear on Satawal, and relies on Hornell’s generalizations for a more encompassing view.

Flying proa or popo of the Caroline Islands, after Paris
Flying proa or popo of the Caroline Islands. (From Hornell, after Paris) 
The ocean-voyaging canoes of Satawal are “flying proas” of a type which, according to Hornell, were called popos. Like almost all single-outrigger craft of the Pacific, they always sail with the outrigger to windward, using the outrigger float for stability in light or no air and as a counterbalance – supplemented by crew weight – against the force of higher winds. As such, popos are shunting craft, with identical ends that alternately serve as the bow and stern as the boat changes direction relative to the wind.

canoes of western and central Carolines and Marshall Islands (Kubary)
Notable differences in canoe design existed between the main island groups of Micronesia, as shown in this comparison of proas from the western and central Carolines (left) and the Marshall Islands (right). (Kubary)
Among the most distinctive characteristic of the popo are:
  • a narrow, deep hull with identical ends but lateral asymmetry, in which the windward side is much less curved than the leeward. In fact, on some boats, the windward side of the hull is nearly a flat plane.
  • hull construction of stitched planks on a dugout base
  • a platform extending from the hull on the lee side, opposite the outrigger
  • an oceanic lateen rig, consisting of a single triangular sprit sail hung from a mast stepped exactly amidships that pivots fore-and-aft with each shunt, allowing the sail’s direction and its center of effort to shift ends (not, however, without some complex evolutions on the part of the crew)
When Thomas lived on Satawal, Piailug was in the process of building a popo 33’ feet long and 8’ high from the keel to the “eyes” at the ends. This was the largest canoe built there in living memory, most others being around 26’ or 27’ long. According to Hornell, the canoes present in the central Carolines early in the twentieth century were smaller than those of the nineteenth, so lengths in the mid-thirty-foot range and, judging from Paris's drawings, even longer, may have been common in the past.

Hull Form and Construction

End, plan and construction views of the Caroline proa
End, plan and section views of the Caroline proa (Hornell, after Paris)
On Yap, in the western Carolines, and a few other islands were large trees were available,
...all but the upper part of the sides and the two curved heads are hewn from a single log. In atolls where no large timber is available the dugout portion shrinks to a wedge-shaped piece channeled longitudinally on the upper side so as to give two everted edges upon which the garboard strakes are sewn.... Usually the shapes and sizes of the strake planks are irregular, suitable wood being too precious in atolls to permit either of long running lengths or of adzing opposite edges parallel. So the hull in these islands is a mass of patchwork, all, however, fitted together with remarkable accuracy. (Hornell)
West/Central Carolines proa plan and construction views
Paris's plan and construction views on which Hornell's illustrations are based. (Source: Wikipedia. Thanks to Walter Stanish for uploading this version of the image and alerting me to it.)
According to Thomas, the hull timber was breadfruit or “a mahoganylike tree called rugger...” from which the keel was hewn and the planks were split. Thomas describes how a chainsaw was used to fell the tree for the keel of Piailug’s new boat. The chainsaw was a fairly unfamiliar tool on Satawal at the time, and few of the men knew how to use it. After the “blade-like legs” of the breadfruit were cut through and the tree was felled, limbed, and cut to its desired length, the log for the keel was 6’ in diameter and 30’ long. With the chainsaw, a couple of operators performed the process in one morning “what usually took six men more than a day” to perform with axes.

Thomas does not describe how planks were split or otherwise gotten out, but it is clear that adzes were used to shape all the timbers. The hull has two end pieces cut from solid timber, each of which attaches at an end of the keel and extends upward as a cutwater and stem “and ends throughout the western and central Carolines in a peculiar and most characteristic fork” (Hornell).

No plans or standard measuring devices are used in shaping the hull, the builders working instead by eye and by a series of proportions or ratios between various of the boat’s features. (For example: height of the mast equals length of the hull; length of the outrigger is one half the length of the hull. Thomas.) Regarding the pronounced asymmetry of the hull, Hornell explains that is counteracts the asymmetrical resistance imposed by the single outrigger, allowing the boat to travel straight with minimal steering input when the float is in the water.

Planks are added to the keel to build up the hull’s freeboard with, as Hornell says, “little or no need to fit strengthening frames and this is not the custom in the Carolines.... Adequate stiffness is obtained when necessary by the insertion of solid bulkheads or partitions beneath the transverse supports of the lee platform.” Stiffness is increased further with heavy thwarts and gunwales.

Thomas stated that Piailug, in building his new boat, lashed the planks together temporarily, but Thomas did not explain why. Perhaps it was not enough to ensure that each individual plank fitted properly against its neighbors but that, instead, the entire hull had to be test-assembled to ensure good fits. The wood of the cross-beams (i.e., thwarts) that Piailug wished to install was too tough to be worked with an adze, so the pieces were buried in moist sand for a period to soften them.

Planks are stitched to the keel and to each other with discontinuous stitches of coconut-fiber rope (i.e., coir), made, in Thomas’s account, by the island’s old men, hand-rolling the coconut fibers against their thighs. Planks were caulked with a compound of dehydrated breadfruit sap applied to strips of coconut husk. According to Thomas, this caulking gradually dries out and loses its efficacy, requiring the boats to be disassembled and rebuilt “every two years or so.”

During one of these rebuilding episodes on Satawal on a boat named Suntory (after a brand of Japanese whiskey), the builder in charge decided to lighten the boat to make it faster. After disassembly, all the planks were therefore adzed down to make them thinner. Thomas did not describe whether this process improved the boat’s performance or affected its strength or water-tightness.

Thomas agrees with Hornell in saying that the outer surface of the hull was finished very smooth, although the process of sanding or otherwise smoothing the planks it is not described by either. (Throughout much of the Pacific, shark skin was used as sandpaper for this purpose.) Hulls were traditionally painted in patterns of red, black, and white.

Outrigger and Lee Platform

According to Hornell, the two main outrigger booms, which are fairly straight, “pierce both washstrakes of the hull in large canoes and (extend) a few inches outboard on the lee side.” (In Hornell's version of Admiral Paris’s drawing, however, it is unclear if the booms are entirely surrounded by the washstrake or if, possibly, they are notched full-depth into its top surface). In the popo’s most characteristic form, the outrigger booms also serve as the main support for a large triangular platform which carries crew and cargo. Poles extend diagonally from the windward gunwale at points near both ends of the hull to the main booms near their outboard ends, and the resulting isosceles triangle is covered by planks or light poles.

A second rectangular platform extending from the hull’s leeward size further increases the boat’s cargo capacity. This platform is supported by another pair of heavy timbers that cross the hull. These timbers are angled sharply upward toward their outboard end, allowing the platform to remain dry with the boat at a significant angle of heel. Sometimes the surface of the platform is built directly on these sloping supports; on other cases, another, lighter framework is built over this main structure and decked over, creating a surface that is horizontal when the boat is on an even keel.

Enclosures of basketwork, which may be round, oval, or rectangular, often appear on one or both platforms. These are used primarily to protect cargo, but it appears they occasionally serve as shelters for passengers or off-duty crew.
Connection between the outrigger booms and the float
Connection between the outrigger booms and the float. In the image on the left, the inner two crutches have been omitted for clarity. (Hornell)
The outrigger float, canoe-shaped in plan, hexagonal in section, and about half the length of the hull, is hewn from a solid baulk. It is connected to each boom by a pair of short crutches, the forks of which straddle the booms from below. The lower ends of these crutches are pointed and driven several inches into the upper facets of the float. Between each pair of crutch forks is a yoke which rests across the top surfaces of both booms and extends several inches beyond them. Holes are bored horizontally through the angled top surfaces of the float, and ropes are routed through these holes, over the ends of the yoke, and in a complex path around the crutches to hold the float securely to the booms, with the crutches in compression between them. Additional, lighter-weight braces are lashed to the undersides of the booms, further connecting them to each other and to the crutches, and from one pair of crutches to the other. The entire outrigger structure is thus highly complex and highly engineered for strength and flexibility.

Sail Rig and Steering

Caroline proas
Caroline proas (Paris)
The defining characteristic of all true proas is the oceanic lateen or oceanic sprit sail hung from a pivoting mast stepped amidships. (I find the term “crab-claw” used often inappropriately, preferring to apply it only to sails whose leech is deeply concave and therefore somewhat reminiscent of the shape of a claw. This is not the case in the canoes of the central Carolines.) In combination with a double-ended hull, this permits the boat to be shunted, and sailed with either end forward. Sails were traditionally hand-woven from fibers stripped from the leaves of the pandanus or screw pine, but commercially manufactured cloth has apparently been in common use for some time.

Hornell takes pains to define the nomenclature of the sail’s spars, calling the one at the luff (the upper/forward edge) a yard, and the one at the foot a boom. He describes the rigging as follows:
A shroud runs from near the masthead to the yoke [i.e., the short, stout timber connecting the outrigger booms at their outer ends], to which it is secured after passing through a hole bored through its center. There is also a fore-and-aft running stay made fast respectively to the endmost thwart at each end of the hull.
(T)he sail is hoisted by a peak halyard attached far out on the yard and rove through a sheave hole in the masthead; the heel of the yard rests against a sail step set on a short thwart right in the bows, to which the tack is made fast.

There is no shroud or other line serving as such to leeward in this description, and in Thomas’s account of a voyage from Satawal to a nearby atoll, the rig was nearly lost once when the sail was backwinded, due to this lack of support on that side of the boat. No leeward shroud is shown in the diagram or cover photo of Thomas’s book, but in Paris’s diagram and one of his drawings (fifth and first images in this post), a shroud is visible leading from the masthead to the lee platform. Perhaps the practice varied on different islands even within the central Carolines.

Main features of a Satawal proa
Main features of a Satawal proa (from Thomas)
In the past, popos were generally steered by means of a quarter rudder with a tiller that extended laterally from its aft edge near the top. The rudder was not hung from the hull, but instead held by the helmsman’s foot against a wooden pin that projected from the hull. A rope from the top of the blade to the hull, visible in Paris's diagram (image #3), prevented loss of the rudder and perhaps provided some stability to it but apparently did little to hold it in its proper working position for, according to Hornell, “The steersman’s duty is the heaviest aboard and on a journey he has to be relieved frequently.” It rough seas when the rudder could not be controlled, it was lifted from the water, and several men with regular paddles would steer. In Thomas’s account, however, steering was always by means of a single steering paddle, and no rudder appears in his description or in the book’s diagram.

Canoe Houses

Caroline sailing canoes are built and kept in a canoe house. On Satawal, in Thomas’s account, each of the island’s eight clans owned a canoe and kept it in its own house. All the houses were on the same beach, near the main (only?) channel through the island’s surrounding reef, with the house belonging to the clan of the island’s chief directly opposite and closest to the channel. The canoe houses serve as social centers for the men, who typically gather there to talk and drink even when work is not under way on a canoe.

Experienced men of middle age do most of the work on the canoes, assisted and observed by younger ones -- or, in Thomas's account, those few younger ones who could be induced to take an interest any longer. Older men typically observe the work and offer suggestions but restrict their hands-on contribution to the making of rope.

Primary Sources:

Hornell, James, The Canoes of Polynesia, Fiji, and Micronesia, B.P. Bishop Museum, 1936, in Canoes of Oceania (with A.C. Haddon), B.P Bishop Museum, Honolulu, 1975

Kubary, J.S., Ethnographiphische Beiträge zur Kenntniss des Karolinen-Achipel, Leiden, 1889-95

Pâris, François-Edmond, 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")

Thomas, Steve, The Last Navigator: A Young Man, An Ancient Mariner, The Secrets of the Sea, International Marine, Camden, ME (no date). Originally published by Henry Holt, 1987

Additional sources, useful websites and pages: and


  1. Excellent summary. I liked Thomas's book very much but left it feeling sad for lack of the information you have discussed. I consoled myself by thinking he had two main purposes: document the navigational method, which he did very well, I thought, and narrate a typical "personal journey" story about forming one's identity (here, through father-figure Mau Pialuig). No disrespect intended, but Thomas having already been an accomplished blue-water sailor, I was bemused by his need for further social ID formation. I suppose interesting people are always striving the next step....

    Interesting story about the adzes. Steel having been introduced long ago, scholars have been curious about how native materials would affect canoe building (thick shells, sometimes basalt). basalt is not a great tool material, but better than nothing (I tried it once as an archaeology student); shell I have never tried. But I read somewhere that Mau and colleagues were asked to try shell adzes -- they gave them up very, very quickly. I would love to know more details about that brief experiment in re-use of native canoe building materials!

  2. I wouldn't want to have to hollow a dugout with an adze blade made of a shell either. I guess it works if you use the char-and-scrape method and have plenty of time on your hands. Obsidian is available in Micronesia, and that might be the best choice of natural materials -- delicate, but very sharp.
    Regarding Thomas and his emotional objectives -- the social identification you refer to was mostly about resolving internal issues regarding his relationship with his own father. That is the kind of thing that some people simply never get over. Although I agree with you that it can be unseemly for an accomplished adult to be going on in public about it, I thought he handled it well overall. Somewhat too "sensitive" at times, but some people like that kind of thing.

  3. On the island of Nomwin, a small canoe from the Gilbert Islands (Kiribati) is known. The canoes of Kiribati are much lighter; Kiribati straddles the doldrums and has much lighter winds and seas than those of the Carolines. Today in Kiribati, small canoes are made up of strakes sawn from a sheet of thin marine plywood, stitched together with monofilament fishing line. I do not know details of construction of the canoe known on Nomwin, called the "Tara Waa" (Waa = canoe); however, it is esteemed in racing, as it is light and speedy.

  4. Bob, great blog but I regret to say that I think I've found a .....mistake! When it comes to the asymmetry of the Vaka section, it is the lee side which is flattened and the windward side which is curved. If you don't believe me, take a look at Gary Dierking's T2 (p23 of his book).
    I remember reading somewhere, maybe in @East is a big bird' by Thomas Gladwin, that the sailors and builders of these Waa had no notion that the asymmetry had anything to do with preventing leeway. To them it was all about wanting the canoe to tend to head down wind rather than to luff. This was because what they feared most was luffing through the wind and being caught aback, which could lead to serious injuries as the rig came crashing down on crew and craft.
    I've been vaguely wondering if, among the shunting proas of the Pacific there is any correlation between having mast props and not needing asymmetric sections. (Probably not). With a mast prop, being caught aback is a lesser crisis. see Gary's post on mast props here: