Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Arts and Crafts of the Cook Islands

Canoes

Canoes

Introduction

According to traditional records, the ancestors of the Cook Islanders made long sea voyages to islands as distant as Tahiti and the Marquesas. Similarities in certain artifacts reveal that there must also have been some communication between the Cook and Austral Islands. Each of the Cook Islands, except Mangaia, has traditions concerning the coming of a named ancestor in a named canoe generally from a region termed 'Avaiki, which was probably the Society Islands; and in some traditions a district in Tahiti is specified. One of the page 177Rarotongan ancestors is said to have come from Samoa. Traditions also record the coming of voyagers after the first settlers. Though the Mangaians are unique in the belief that their island emerged from the depths with the first settlers upon it, their oral history records the influx of later settlers who reached Mangaia by canoes. Hence the voyaging canoe plays an important part in the history of the group.

The canoes seen at Atiu during Cook's third expedition (20, p. 196) were double canoes about 20 feet long and were evidently the type used for coastal service and fishing. Similar double canoes were seen by Cook (20, p. 209) at Hervey Island (Manuae), but the vesselhe saw at Mangaia (20, p. 173) was a small outrigger canoe about 10 feet long. John Williams, the official discoverer of Rarotonga, mentions that a large double canoe came off to his ship (81, p. 122). He also mentions that a chief visited his ship in a large double canoe at Atiu (81, p. 95). Portlock (46, pp. 228, 229), who visited Aitutaki in 1792, states that of a dozen canoes that visited the ship, four were double canoes.

With the advent of Europeans, particularly missionaries, interisland communication in native craft ceased, hence the building of double voyaging canoes ended. It is evident, however, that double canoes of a fair size continued to be used for coastal service and fishing, particularly in Atiu and Mauke where a number are still to be found. Their use in the other islands of the group appears to have been discontinued fairly early. The double canoes in Atiu and Mauke were used for torching at night to catch flying fish. When I was in Atiu in 1929, I was told that they were seldom used. As no more are being made, they will disappear when the present canoes rot away. The small out-rigger canoes will continue to survive for years to come, because the people depend upon fish as a great part of their diet. A fishing canoe is thus a necessary part of every able-bodied man's equipment. It is preferable to a boat, because it is lighter and of a shallower draft and can be maneuvered over the reef. The pre-European type of outrigger canoe is still made in Atiu, Mauke, and Mitiaro. Changes in hull construction have taken place in Rarotonga, Mangaia, and Aitutaki. It may be more than a coincidence that the islands where change has taken place are those where white missionary stations were established early. For loading and unloading cargo from trading ships and steamers, the Atiuans now use whale boats. However, as their draft is too deep to cross the edge of the reef, they are loaded and unloaded at the outer edge of the reef. The Mangaians, on the other hand, have preferred to use large outrigger canoes which have a shallower draft and can be hauled over the reef edge, and paddled from and to the shore. The Mangaian method reduces the distance that the goods have to be carried, but the canoes cannot carry such large loads as the boats carry.

page 178

Cook Islands canoes are described by Hornell (44, pp. 156-176), but they are dealt with here to complete this work and to include some field details that were not available to Hornell.

Outrigger Canoes

The outrigger canoe consists of a hull, or underbody, and the outrigger.

The Hull

Tamanu timber was preferred for the canoe hull because it had a tall straight trunk and was durable (pakari), lasting from eight to ten years. The puka was also used, but it had a short trunk that necessitated joining and lasted but one to two years. The introduced mango (vi) has been used in late years as it has a straight trunk and lasts two to four years.

In addition to the dugout underbody, there is a gunwale strake, or rail; a bow cover; a stern cover; and sometimes a raised stern piece. The gunwale strake, or rail, was made of breadfruit wood, the bow and stern covers of 'utu, and the stern piece of tamanu, miro, or tou.

The entire hull is sometimes called the takere, but the term is specifically applied to the under part or keel. The interior hold is termed riu. The hull is a dugout, hollowed out of a selected length of the tree trunk and carefully shaped at the bow and stern. It is possible to form the hull of a small canoe out of one length of timber, but in large canoes or canoes made of puka, it is necessary to join two and sometimes three pieces together to get the necessary length.

The canoe hulls now in use in Rarotonga and Mangaia have departed from the native type in that certain features of European boats have been copied, just as they have been in Tahiti. These features are described by Hornell (44, p. 159) as follows: "… the cutwater sharp and nearly vertical, rounding at the lower angle into the keel line with the lower side of the after end curving fairly quickly from the horizontal bottom to terminate in a bluntly pointed stern." The Aitutaki canoes have a similar bow and the stern follows the same shape.

The old type of hull is retained in Atiu, Mauke, and Mitiaro, a group of three islands close together which had frequent communication with each other. Hornell (44, p. 161) describes these hulls as follows, "The dugout underbody has no sheer except for a short distance at each end." The bottom is straight until it rounds upward to form the bow and stern, which are fairly sharp. The bow curve is longer and more gradual than that of the stern, a feature of Hawaiian canoes. Transversely, the hull is well rounded, and in most the sides are inclined inward, or, as Hornell (44, p. 161) puts it, "… so that the gunwale width is less than that at the level of the bilges." In Mauke, this form page 179of hull is termed tua'ura and is used in fishing canoes. Another form of hull in which the sides did not slope inward so that the greatest width was at the gunwales is termed tuapa. Informants said that the tuapa was faster in the water but that it held less than the tua'ura.

On the underside of the extreme fore end, a downward angular projection is formed which is termed the tara kokiri (spine of the kokiri fish). Hornell (44, p. 161) states that this projection is found on Atiu and Mitiaro canoes but not on those of Mauke. Though it is absent on some Mauke canoes, I did see it on some single and double canoes in Mauke in 1929. Hornell thinks it analagous to the heel at the stern of the Samoan paopao canoe, but it is more like the angular projection on the double canoe models of Manihiki and Rakahanga (75, p. 149, fig. 59). When the Samoan paopao hull was roughed out in the woods, it was notched at the stern to provide a holding place for a rope by which the log was dragged to the village. The notched part was trimmed into a rounded knob which had a fairly wide flat surface on the after side (73, p. 374), whereas the Atiu bow projection has a sharp mesial edge.

The hull join (patunga or potu) was a transverse butted join in which the two square ends of the hull segments were fitted together accurately (fig. 119). Paired holes were made through the full thickness of the two ends
Figure 119.—Butt join of hull: a, longitudinal section of hull, through and through lashing; b, oblique lashing on either side of mid line; c, underside of hull join showing clear space on either side of mid line where oblique lashings do not show through; d, hold side of hull join showing upper side of oblique lashings (1, 2) on either side of mid line.

Figure 119.—Butt join of hull: a, longitudinal section of hull, through and through lashing; b, oblique lashing on either side of mid line; c, underside of hull join showing clear space on either side of mid line where oblique lashings do not show through; d, hold side of hull join showing upper side of oblique lashings (1, 2) on either side of mid line.

parallel with the end surfaces. On either side of the middle or keel line, a pair of holes were made obliquely from the hold side to meet in the end surfaces without reaching the under surface. Thus one lashing on either side of the keel line does not appear on the under side of the hull and is prevented from being frayed when the canoe is dragged over shallows or the outer edge of the reef. Owing to the convex curve of the hull, the other through and through lashings do not contact the coral bottom.

In joining the ends, coconut husk fiber (puru) was used for caulking, and when so used was termed vai torea. Sennit (ka'a) was used for lashing. One end of the sennit was frayed out and thinned to a two-ply twist. A green coconut-leaflet midrib (tuaniu) was frayed at its butt end and twisted in with the coir twist which was termed moamoa ka'a. The midrib was used as a needle to guide the sennit through the paired holes. The Samoan continuous loop (73, p. 387) was not known. Five turns through the paired holes were page 180usual. Each turn was drawn taut with a forked stick termed keke in Aitutaki, the free end of the sennit being passed between the legs of the fork and turns made around the stem. Resting the two legs against the hull, leverage was applied by pulling outward with the handle (70, p. 261). In Atiu, however, a straight stick of taiki wood (a maera) was used in a similar way with one end against the hull. After the turns were drawn taut a sharp wooden stick termed a titi (peg) was plugged into the hole.

In dubbing out the hold, one or two raised flanges (vae) were sometimes left in larger hulls to act as ribs in strengthening the hull. These ran from the bottom to the gunwale edge of the hull.

The forepart of the hull is termed the aumi'i vaka; and the aft part is termed the miri vaka, a dialectical form of muri vaka (muri, aft). The Atiu name for the after part is noko. In a two-piece hull with one join, the two sections are designated by the above terms. In a three-piece hull with two joins, the fore and aft sections are named as above and the extra middle section is the moe. The general term for a section is potonga.

Hull Accessories

The accessory parts that go with the hull are the gunwale strake, the bow and stern covers, and an elevated stern piece.

The gunwale strakes (oa), which, because of their width are better termed the gunwale rails (44, p. 161), are lashed to the gunwale of the underbody, one on each side. They extend to the junction with the bow and stern covers and are made preferably of breadfruit wood. They are from 1.5 to 2 inches wide and about 2.75 inches deep. In cross section, they vary slightly in the different islands but all have a V-shaped longitudinal groove on the under surface to fit against the gunwale of the underbody. The Atiu, Mauke, and Mitiaro rails have an inner flange projecting horizontally a little distance below the top edge to form a ledge termed the paevai. The V-shaped groove on the under surface is termed the rua puru (recess for coconut husk) because it receives the husk packing. The hull gunwale is beveled from both sides to fit into the V-shaped groove of the rail in Atiu and Mangaia, but in Mauke the double bevel is not used (fig. 120). The gunwale rails of the Atiu and Mauke canoes were so much thicker than the gunwale of the underbody that it was possible to bore oblique lashing holes from the angle of the inner flange to the under surface of the rail (fig. 120, a, b). Horizontal holes were bored through the gunwale of the underbody to pair with the holes through the rail. Owing to the oblique direction of the holes through the rail, only a short part of the lashing showed on the outer side. The distances between lashings average five inches. In the Mangaian canoes (fig. 120, c), the holes through the rail were horizontal and thus more of the lashing showed on the outer side.

page 181

The "bow cover" combines the structure of what Hornell (44, vol. 3, p. 6) has defined as the fore end-piece and the head-board. It consists of a triangular piece of wood that fits over the fore part of the hull to form a short fore deck. At the same time, by means of downward projecting sides, it fills in the space between the fore end of the gunwale rail and the bow. The side pieces project backward on either side beyond the horizontal decking part. In Rarotonga and Aitutaki, the structure is called poki, a general term for cover. This is probably a recent term, for Aitutakians say the old name was papa'ura. In Mangaia, it is called the tuaru; in Atiu and Mauke, 'utu aumi'i. In Aitutaki and Rarotonga, the wood used for the cover was tamanu, 'au, puka, and in recent times mango. In Atiu and Mauke, however, the wood used was 'utu. The name 'utu aumi'i (bow Barringtonia) is therefore descriptive of the wood used.

Figure 120.—Lashing of gunwale rail to underbody: 1, gunwale rail; 2, underbody; 3, lashing; 4, inner ledge (paevai); 5, outer ledge. a, Atiu; b, Mauke; c, Mangaia. The outboard side is to the left.

Figure 120.—Lashing of gunwale rail to underbody: 1, gunwale rail; 2, underbody; 3, lashing; 4, inner ledge (paevai); 5, outer ledge. a, Atiu; b, Mauke; c, Mangaia. The outboard side is to the left.

The lower projecting flanges of the bow are lashed to the gunwale of the underbody at intervals. The short after projections, which are about 1.5 inches long are of the same depth and width as the gunwale rail, and their ends are lashed together. It is evident that the aft projections of the cover were made to render the lashing to gunwale rail easier. The fore end of the cover follows the upward curve of the underbody. The front edge of the pointed cover and the short edge of the hull below form a vertical mesial edge, which is termed the rae 'utu or 'e.

The stern cover of the Atiu, Mauke, and Mitiaro canoes is also made of 'utu and is termed 'utu noko. It is shaped on the same principle as the bow cover but is much shorter, as in these canoes the aft outrigger boom is closer to the stern than the fore boom is to the bow. The lower projecting side flanges are lashed to the gunwale and the aft end of the gunwale rail to the forward projections of the cover in a similar manner to the bow cover. In the shorter length of the stern cover, the aft end rises more sharply and the sharp end is continued upward from the underbody to form a slightly elevated sternpiece. The short elevation is used in the canoes of Atiu and Mauke but in Mitiaro an extra sternpiece is added for greater height.

page 182

In the modern canoes of Aitutaki, the sharp stern as well as the bow is provided with a cover. The bow cover prevents water from breaking over into the hold and, as the Aitutaki canoes are always sailed with the outrigger on the windward side, the stern cover is necessary when tacking with the stern in front. The Aitutaki covers are flat with no rise at the sharp ends.

The elevated sternpiece ('iku tira) occurs in the outrigger canoes of Mitiaro and in the double canoes of Atiu and Mauke. This was formerly a feature of the Atiu outrigger canoes, for Cook (20, p. 173) says, "They are long and narrow and supported by outriggers. The stern is elevated about three or four feet, something like a ship's stern post." There is every probability that the elevated stern post was also used in Mauke. Hence Mitiaro, the smallest of the three islands, is the only one to preserve an old form of technique in the outrigger canoes.

Figure 121.—Rarotongan canoe ornament (?) (British Mus., C.C.2086). a, front view: height, 680 mm.; top width, 345 mm.; middle width, 245 mm.; bottom width, 400 mm.; thickness at upper border, 17 mm.; thickness at lower border, 40-50 mm.; at upper border there are two serrated edges (1) followed by three rows of carved figures; lateral rows commence with a head (2) with typical flanged eyes and well-marked ears forming a connection with the main slab, and defined below by a flange (3) with a notched lower edge; below follows a set of four complete figures, each being attached to slab by the ears (4), backward arm projection (5), gluteus (6), and foot formed by a notched flange (7); in each of the three upper figures, notched foot flange (7) rests on head of figure below; in lowest figure, lower leg (8) rests on an outward projection at base and is without a notched foot. Middle row consists of five figures, four being similar in detail to the lateral figures but the lowest (9) much larger, having three-fingered hands, and resting on a curved base (10) which is 77 mm. wide and projects forward for 100 mm.; curved base with two notched edges with points upward; lower square holes (11). b, back view: upper edge (1) cut at backward downward slope, forming a lower curve owing to greatest thickness in middle line; lateral figures show same as on front; back surface plain without any median figures, c, side view of lower end: shows full face view of two lowest lateral figures (1, 2) and three lowest middle figures with forward projection of lowest (3) and curved base (4); fourth median figure (5) has no notched foot flange.

Figure 121.—Rarotongan canoe ornament (?) (British Mus., C.C.2086). a, front view: height, 680 mm.; top width, 345 mm.; middle width, 245 mm.; bottom width, 400 mm.; thickness at upper border, 17 mm.; thickness at lower border, 40-50 mm.; at upper border there are two serrated edges (1) followed by three rows of carved figures; lateral rows commence with a head (2) with typical flanged eyes and well-marked ears forming a connection with the main slab, and defined below by a flange (3) with a notched lower edge; below follows a set of four complete figures, each being attached to slab by the ears (4), backward arm projection (5), gluteus (6), and foot formed by a notched flange (7); in each of the three upper figures, notched foot flange (7) rests on head of figure below; in lowest figure, lower leg (8) rests on an outward projection at base and is without a notched foot. Middle row consists of five figures, four being similar in detail to the lateral figures but the lowest (9) much larger, having three-fingered hands, and resting on a curved base (10) which is 77 mm. wide and projects forward for 100 mm.; curved base with two notched edges with points upward; lower square holes (11). b, back view: upper edge (1) cut at backward downward slope, forming a lower curve owing to greatest thickness in middle line; lateral figures show same as on front; back surface plain without any median figures, c, side view of lower end: shows full face view of two lowest lateral figures (1, 2) and three lowest middle figures with forward projection of lowest (3) and curved base (4); fourth median figure (5) has no notched foot flange.

page 183

Cook (20, p. 211) noted a similar construction in Hervey Island and in Mangaia (20, p. 178). Hornell figures it for Mangaia in a model in the Australian Museum, Sydney (44, p. 160) and in a small canoe in the Dominion Museum, New Zealand (44, p. 159).

Belcher (5, p. 229) figures a Rarotongan canoe with a tall carved stern post, but the details are not clear enough to merit reproduction here. However, a wooden slab in the British Museum, carved with typical Rarotongan human figures, may have been used as a canoe ornament. The slab is 26.75 inches (680 mm.) high, with straight upper and lower edges and concave sides. A row of human figures is carved on each side edge, looking out laterally, and a median row of similar figures is carved in relief on the front surface. The bottom edge is wider than the top, and at each lower corner there is a square-cut hole for lashing the slab to some other object. The slab is figured here for record, but its identification as a canoe stern-piece ornament is purely provisional and open to future correction. For details see figure 121.

Portlock (46, p. 229) mentions that some of the canoes in Aitutaki had "a kind of gallows erected at the stern, almost six feet high and decorated with man-of-war bird's feathers." These references prove that the raised stern piece was once present in all the islands of the Cook group.

Figure 122.—Stern piece of Mitiaro canoe. a, junction of stern piece (1) to hull (2) and stern cover (3); stern piece lashed to stern cover at each upper corner. b, on either side of posterior median edge (4), three oblique holes are bored near edge of stern piece (1) and hull (2) so as to meet. Commencing on right, a length of sennit (5) passes through upper hole (6) to emerge through lower paired hole (7) from inside; sennit is then carried over on outside to enter upper hole (6) and passed down on inside to again emerge through lower hole (7) from whence it is carried obliquely to next upper hole (8); sennit now passes down on inside to emerge through lower paired hole (9) whence it is carried upward on outside to again enter upper hole (8); it passes down on inside to again emerge from lower hole (9) whence it is carried obliquely over to next upper hole (10). Sequence of turns is continued with remaining paired holes to the left and results in two turns being made on inside between paired holes but only one on outside including oblique turns connecting paired holes. Turns are all made loosely with stern piece raised to allow threading, and when all turns are completed they are drawn taut in sequence from right to left. c, ends of sennit at first and last pairs of holes are passed around nearest oblique turn, tied in an overhand knot, and ends cut off.

Figure 122.—Stern piece of Mitiaro canoe. a, junction of stern piece (1) to hull (2) and stern cover (3); stern piece lashed to stern cover at each upper corner. b, on either side of posterior median edge (4), three oblique holes are bored near edge of stern piece (1) and hull (2) so as to meet. Commencing on right, a length of sennit (5) passes through upper hole (6) to emerge through lower paired hole (7) from inside; sennit is then carried over on outside to enter upper hole (6) and passed down on inside to again emerge through lower hole (7) from whence it is carried obliquely to next upper hole (8); sennit now passes down on inside to emerge through lower paired hole (9) whence it is carried upward on outside to again enter upper hole (8); it passes down on inside to again emerge from lower hole (9) whence it is carried obliquely over to next upper hole (10). Sequence of turns is continued with remaining paired holes to the left and results in two turns being made on inside between paired holes but only one on outside including oblique turns connecting paired holes. Turns are all made loosely with stern piece raised to allow threading, and when all turns are completed they are drawn taut in sequence from right to left. c, ends of sennit at first and last pairs of holes are passed around nearest oblique turn, tied in an overhand knot, and ends cut off.

page 184
Figure 123.—Outrigger canoes: 1, hull or underbody; 2, gunwale rail; 3, bow cover; 4, stern cover; 5, float; 6, fore boom; 7, after boom; 8, thwart; 9, carrying stringer; 10, bow projection (tara kokiri); 11, stern post. a, Atiu: length, 14 feet 2 inches; note slope of bow and tara kokiri projection (10), blunter stern with short upward projection of stern cover (4); carrying stringer (9) passes above after boom (7) and thwart (8) and below fore boom (6). b, Mitiaro: length, 13 feet 10 inches; inside width between gunwale rails at fore boom, 9.5 inches; at after boom, 10.25 inches; fore cover (3) with marked upward curve; length of gunwale rail (2), 10 feet 4.5 inches; stern cover (4) width in front, 9.75 inches and after width at junction with stern post, 3.75 inches; fore limbs of stern cover, 1.25 inches long, 2 inches wide, and 1.75 inches deep; stern post (11), 15.25 inches high, triangular in section with front surface 3.75 inches wide at base and 3.5 inches on each side; at top each stern post surface 1.5 inches wide; float (5), round in section, fore end sharp, after end blunt, and total length, 12 feet 3 inches; fore boom (6) curved, 3 feet 9 inches long with outboard projection on starboard side of 5 inches, crosses gunwale rails 1.5 inches behind their fore ends; after boom (7) also curved, 4 feet long; both booms with short branches and two stays to float; thwart (8); carrying stringer (9), 11 feet long, 1.5 inches thick, passes above both booms and thwart projection, c, Mangaia: length, 15 feet 10 inches; inside width between gunwales at middle, 9.5 inches and at square stern, 6 inches; canoe depth, 16 inches in middle, 15.5 inches at bow, and 11.5 inches at stern; gunwale rail (2), 5 inches deep in middle and 2.25 inches at stern; fore cover (3) depth ranges from 3.25 inches at after end to 2.5 inches at bow; no after cover; float (5), 13 feet 10 inches long; fore boom (6), 3 feet 5 inches long, projects outboard on starboard side for 2 inches; 1.4 inches thick, boom branch 1.1 inches thick, height of boom above float 13.75 inches, and distance between hull and float 29 inches; after boom (7), 3 feet 6 inches long, projects outboard on starboard side for 0.75 inch, same thickness as foreboom, height above float 12.5 inches, distance between hull and float 30 inches; carrying stringer (9), 1.5 inches thick, passes above thwart and below both booms, and 1.75 inches from hull.

Figure 123.—Outrigger canoes: 1, hull or underbody; 2, gunwale rail; 3, bow cover; 4, stern cover; 5, float; 6, fore boom; 7, after boom; 8, thwart; 9, carrying stringer; 10, bow projection (tara kokiri); 11, stern post. a, Atiu: length, 14 feet 2 inches; note slope of bow and tara kokiri projection (10), blunter stern with short upward projection of stern cover (4); carrying stringer (9) passes above after boom (7) and thwart (8) and below fore boom (6). b, Mitiaro: length, 13 feet 10 inches; inside width between gunwale rails at fore boom, 9.5 inches; at after boom, 10.25 inches; fore cover (3) with marked upward curve; length of gunwale rail (2), 10 feet 4.5 inches; stern cover (4) width in front, 9.75 inches and after width at junction with stern post, 3.75 inches; fore limbs of stern cover, 1.25 inches long, 2 inches wide, and 1.75 inches deep; stern post (11), 15.25 inches high, triangular in section with front surface 3.75 inches wide at base and 3.5 inches on each side; at top each stern post surface 1.5 inches wide; float (5), round in section, fore end sharp, after end blunt, and total length, 12 feet 3 inches; fore boom (6) curved, 3 feet 9 inches long with outboard projection on starboard side of 5 inches, crosses gunwale rails 1.5 inches behind their fore ends; after boom (7) also curved, 4 feet long; both booms with short branches and two stays to float; thwart (8); carrying stringer (9), 11 feet long, 1.5 inches thick, passes above both booms and thwart projection, c, Mangaia: length, 15 feet 10 inches; inside width between gunwales at middle, 9.5 inches and at square stern, 6 inches; canoe depth, 16 inches in middle, 15.5 inches at bow, and 11.5 inches at stern; gunwale rail (2), 5 inches deep in middle and 2.25 inches at stern; fore cover (3) depth ranges from 3.25 inches at after end to 2.5 inches at bow; no after cover; float (5), 13 feet 10 inches long; fore boom (6), 3 feet 5 inches long, projects outboard on starboard side for 2 inches; 1.4 inches thick, boom branch 1.1 inches thick, height of boom above float 13.75 inches, and distance between hull and float 29 inches; after boom (7), 3 feet 6 inches long, projects outboard on starboard side for 0.75 inch, same thickness as foreboom, height above float 12.5 inches, distance between hull and float 30 inches; carrying stringer (9), 1.5 inches thick, passes above thwart and below both booms, and 1.75 inches from hull.

page 185

The stern posts seen in Mitiaro were triangular in section with the apex coinciding with the median edge of the hull stern and with the base to the front to continue the narrow end surface of the stern cover (fig. 122).

To show the relationship between the hull and the outrigger parts, drawings of canoes from Atiu, Mitiaro, and Mangaia are shown in figure 123.

The Outrigger

The outrigger consists of a float and two booms with attachments of the booms to the hull and the float.

The float

The float (ama) is a straight length of 'au which is pointed at the fore end and cut off square at the after end (fig. 123, a-c, 5). Most floats are trimmed flat on the upper surface, but in some of the Mitiaro canoes, the spar retained its natural form. The trimming of the upper surface formed side edges that made it easier to pierce holes for the supporting stays of sennit. The fore end was trimmed from the sides and from below to form the point which gave the effect of an upward curve. The float was always shorter than the hull, because, according to informants, the canoe would be more difficult to steer if they were of the same length. The after end extended for a short distance behind the after boom, but the fore end extended for a greater distance beyond the fore boom. The fore end was also closer to the hull than the after end. As pointed out by Hornell (44, p. 163), the floats of the Aitutaki sailing canoes were sharp at both ends. The float is almost always on the port side of the canoe, the reason given being that right-handed men can make long sweeps with the paddle on the starboard side and can steer from that side without interference from the aft boom. A left-handed man attaches his outrigger on the starboard side.

Booms

There are always two booms (kiato), a fore boom (kiato mua) and an after boom (kiato muri). Poles a few feet long and about 1.5 inches in diameter at the middle were used. In all the islands, except Aitutaki, the selected poles had a stout branch at one end for insertion into the upper surface of the float. In Rarotonga and Mangaia, the poles were perfectly straight and were of ironwood. In Rarotonga, saplings of the matira tree that grows in the uplands were also used. The roots spread out and one of them served as a branch. In Atiu, Mauke, and Mitiaro, the poles were curved; and my informants said that they were always of tamanu wood, though I have seen ironwood used (fig. 123, a-c, 6, 7). The straight ironwood poles require a long branch to reach the float, whereas the curved tamanu poles require a page 186shorter branch. It may be that the slightly different technique in lashing to the float has been brought about by the kind of wood used for the booms. The booms are laid over the top of both gunwale rails, close to their ends, and lashed to them through the holes pierced in the rails. The butt ends project slightly on the starboard side. The booms are spaced widely apart with the after one close to the stern, whereas the fore boom is a greater distance from the head. As the two covers fill in the spaces between the booms and the ends of the hull, the after cover is much shorter than the fore cover.

The Aitutaki booms, which are straight poles of ironwood without a branch, are connected with the float by means of a Y-shaped connective (fig. 124, g).

Outrigger Attachments

Boom to gunwale rail. The booms are lashed to the gunwale rails before they are attached to the float. In small canoes, one hole is pierced through the gunwale rail on the level just above its inner flange and immediately below the middle of the boom. Sennit is tied around the boom close to the inner side of the rail with a running noose, and the free end is passed through the hole in the rail from the inside. The sennit is then crossed diagonally over the boom to enter the hole from the inside. The next turn is crossed obliquely over the first turn, and so the turns are made obliquely on either side of the first two turns to develop a neat lozenge pattern. When sufficient turns were made, transverse turns were made around them, the sennit passing horizontally between the boom and the rail to tighten up the lashing. A final half-hitch in the transverse turns completes the lashing, and the slack is cut off.

Boom to float. The attachment of the branched booms to the float is usually made in the water, where the set of the canoe in relation to the float may be seen. A hole for each branch is made in the middle of the upper surface of the float, and the branch is driven in for the required depth. The rigid branch keeps the float at the right distance from the boom, but, when the canoe is lifted and carried, the weight of the float is apt to loosen it. The float is therefore lashed to the boom, and the lashing technique differs slightly with the straight booms and the curved booms.

In the straight booms of Rarotonga and Mangaia (fig. 124, a-c), the boom is continued outward for a short distance beyond the junction of the branch. This outer projection is grooved transversely for the attachment of the suspensory lashing. Oblique holes are pierced through the float, and the sennit passes through them and around the boom projection and the upper part is seized. The number of stays so formed ranges from two to four.

In the curved boom of Atiu, Mauke, and Mitiaro, there is no outward prolongation of the boom beyond the junction of the branch, for it is cut off page 187
Figure 124.—Boom attachments to float. 1, float; 2, boom. a-c, Mangaian and Rarotongan straight boom with long branch and projecting end (3) for lashing with: a, two stays; b, three stays; and c, four stays. d-f, Atiu, Mauke, and Mitiaro curved boom with short branch, no end projection, and stays attached to boom on inner side of branch: d, Mauke with one inner stay; e, Mitiaro after boom with one outer stay; f, Mitiaro fore boom with two outer stays; g, Aitutaki with Y-shaped connective (4).

Figure 124.—Boom attachments to float. 1, float; 2, boom. a-c, Mangaian and Rarotongan straight boom with long branch and projecting end (3) for lashing with: a, two stays; b, three stays; and c, four stays. d-f, Atiu, Mauke, and Mitiaro curved boom with short branch, no end projection, and stays attached to boom on inner side of branch: d, Mauke with one inner stay; e, Mitiaro after boom with one outer stay; f, Mitiaro fore boom with two outer stays; g, Aitutaki with Y-shaped connective (4).

close and only a slight knob may be visible which cannot be used for lashing (fig. 124, d-f). Oblique holes are bored through the float, and the sennit that passes through them must be lashed to the boom on the inner side of the branch junction. These turns are also seized. Some booms have only one stay, others have two.

In the branched booms, the branch is termed tito and the lashing stays titotito. In modern canoes, the stays are attached to the float by staples or bent nails, though some canoe builders still prefer the holes.

In Aitutaki, an indirect attachment by means of a Y-shaped stanchion (patiatia) is used (70, p. 266). The stem is driven into a hole in the middle line of the upper surface of the float with the arms at a slightly oblique angle, so that they clasp the boom on either side. The stem may pass through the entire thickness of the float. The boom does not rest in the crotch of the fork but is slightly above it and is lashed separately to each limb (fig. 124, g). Suspensory lashings were not used. In an early article (72, p. 214) quoted by Hornell (44, p. 165), I made the statement that the Y-shaped connective was used also in Manihiki. My information was obtained from a native in Rarotonga, but on a subsequent visit to Manihiki I did not see the Y-shaped connective.

I was informed in Aitutaki that four stick stanchions that crossed below the boom were formerly used (70, p. 266, fig. 231, b). This statement is confirmed somewhat by a drawing of an Aitutaki canoe made by Lieutenant Tobin (46, p. 134) in 1792, but it is impossible to tell from the drawing whether the sticks are undercrossed or overcrossed. This raises the question as to whether or not the Y-shaped stanchion is a late introduction in Aitutaki. See p. 452.

page 188
Thwarts

The thwart, or seat (no'oanga), which is single in the small outrigger canoes, is placed nearer the stern than the bow. It rests on the gunwale rails to which it is lashed on either side.

The thwarts of Atiu, Mauke, and Mitiaro canoes are somewhat elaborate in that they have a downward projection for some inches along the fore edge. This projection is notched to fit against the inner ledge of the gunwale rail, and its lower edge is pierced with a number of holes to give attachment to the fore end of a net which covers the hold from the seat to the after cover.

The net over the after hold prevents live fish from jumping or from spilling out when the canoe heels over. A plaited coconut mat is attached over the net to protect the fish from the sun. The upper surface of the thwart is slightly concave. A projection juts out from the middle of the port side to give support to the carrying stringer (tango). Below the projection two holes are pierced through the gunwale rail at the level of its inner ledge and the sennit lashing passes through them and around the projection close to the seat edge. In some seats two holes are pierced vertically through the projection to take the lashing for the carrying stringer, but in others the stringer is lashed directly.

On the starboard side there is a shorter projection, which may be a vestige of a former lashing projection; but the lashing is now made through two holes pierced through the seat and another pair through the gunwale rail (fig. 125, a, b).

The Mangaian thwart does not have the downward projection at the front edge, and no net is used over the after hold. The seat is lashed to the gunwale rails through a series of holes extending across the full width of the seat. On the port side, the thwart projects outboard for its full width to give support to the carrying stringer to which it is also lashed (fig. 125, c-d).

In Aitutaki canoes, the thwart does not project beyond the gunwale rails, hence it does not give support to the longitudinal stringer. A thwart is provided at each end of the hull and pierced with a median hole through which the mast passes into a rounded socket on the floor of the hold.

The Carrying Stringer

The carrying stringer (tango) is a constant feature in canoes of all the islands except Aitutaki. It is a pole about 1.5 inches thick, and its length is determined by the distance between the booms to which it is lashed at its two ends, about two inches from the gunwale rail, and also to the seat projection. There is a little variation in the manner of lashing to the seat projection and its position above or below the booms.

In Mangaian canoes, the stringer passes under both booms and above the seat projection (fig. 123, c, 9); in Mitiaroan canoes, it passes above all three page 189(fig. 123, b, 9); and in Atiuan and Maukean canoes, it passes above the after boom and seat projection and below the fore boom (fig. 123, a, 9). The fore end is notched to fit under the fore boom and so render the lashing more secure for carrying the canoe. The lashings to the booms are oblique crossings which form a lozenge pattern. Then, some transverse turns are made between the two wooden elements to tauten the previous turns. When the seat projection is perforated, the first two turns are passed around the stringer and through the holes; but subsequent turns are passed around the projection with finishing transverse turns as in the boom lashings (fig. 125, b). The Mangaian lashing is through a set of five holes (fig. 125, c, d). In Atiu, Mauke and Mitiaro canoes, the stringer projects for a few inches beyond the after boom.

Figure 125.—Canoe thwarts, viewed from bow with outrigger side to right: a, b, Mauke; c, d, Mangaia. a, upper surface of seat (1): length without projections, 13.6 inches; width in middle line, 7.75 inches, at two ends, 7.25 inches; port projection (2) perforated with two holes, 4 inches long, 2 inches wide, 1.5 inches thick; short starboard projection (3); two holes on starboard side for lashing seat to gunwale rail. b, front view of seat (1) with downward projection (6) 6.5 inches deep and 10.25 inches wide at lower edge, showing perforations for net; sides cut to fit inner ledge (paevai) of gunwale rail (4) resting on hull (5). Port projection (2) supporting carrying stringer (7); short starboard projection (3). c, rectangular seat (1): 14.5 inches long, 9 inches wide, and 1 inch thick; perforated with holes for lashing to gunwale rails (3) and projecting well over port side (2) to support carrying stringer (4) which is lashed to seat projection by four courses of sennit passing through five holes. d, front view: seat (1) with port projection (2); gunwale rails (3); carrying stringer (4); hull (5).

Figure 125.—Canoe thwarts, viewed from bow with outrigger side to right: a, b, Mauke; c, d, Mangaia. a, upper surface of seat (1): length without projections, 13.6 inches; width in middle line, 7.75 inches, at two ends, 7.25 inches; port projection (2) perforated with two holes, 4 inches long, 2 inches wide, 1.5 inches thick; short starboard projection (3); two holes on starboard side for lashing seat to gunwale rail. b, front view of seat (1) with downward projection (6) 6.5 inches deep and 10.25 inches wide at lower edge, showing perforations for net; sides cut to fit inner ledge (paevai) of gunwale rail (4) resting on hull (5). Port projection (2) supporting carrying stringer (7); short starboard projection (3). c, rectangular seat (1): 14.5 inches long, 9 inches wide, and 1 inch thick; perforated with holes for lashing to gunwale rails (3) and projecting well over port side (2) to support carrying stringer (4) which is lashed to seat projection by four courses of sennit passing through five holes. d, front view: seat (1) with port projection (2); gunwale rails (3); carrying stringer (4); hull (5).

To carry the canoe, a person stands beside the middle of the hull on the outrigger side. Seizing the stringer with both hands, he lifts the canoe so that the stringer rests on his right shoulder. He holds the stringer with his right hand, and places his left hand on the float. (See pl. 10, A.)

page 190

In Aitutaki, the canoes are practically all sailing canoes with long booms. The stringer pole, which is lashed to both booms at some distance from the hull, cannot be used for carrying purposes. When the canoe is sailing, the float is always on the windward side, so that when it heels over, the float rises out of the water. To prevent the float rising too high and capsizing the canoe, the occupant leans out on the stringer to force down the float. Hence the stringer is termed rakau taomi ama (wood that keeps down the float) and differs in function from the tango. The term tango is now used as an alternative to oa for the gunwale rail. Portlock (46, p. 229) states that the canoes were "strengthened by a piece of wood one and a half inches thick, that is securely lashed within and just below the gunwale." It is probable that this was the original Aitutaki tango; and, as the thickness corresponds to that of the tango of the other islands, one cannot help thinking that in Portlock's description "within" should have been written "without." The outward shift of the stringer in the sailing canoes, with a change of function leading to a change of name, evidently left the term tango free; and, evidently, because the canoe was lifted by grasping the gunwale rail (oa), the gunwale rail performed the lifting function of the original tango and so received the additional term.

Paddles

The old form of paddle ('oe) is still used in Atiu, Mauke, and Mitiaro, the islands that kept the old form of canoe hull. Naturally enough, Aitutaki, Mangaia, and Rarotonga which changed the form of hull after European
Figure 126.—Atiu paddle (Bishop Mus., C2914). a, front of paddle: length of shaft, 1,538 mm., of blade, 585 mm., total, 2,123 mm.; diameter at end of shaft, 35 mm., of smaller projection, 22 mm., at middle, 44 mm., at shoulder with blade, 38 mm. transversely and 32 mm. front to back. b, enlargement of front of blade: width at shoulders (1) 92 mm., at widest part (2), 222 mm.; median ridge (3), 212 mm. long and 15 mm. wide, notched on sides and incised at lower end with owner's initials, M.K.; two curved lines (4) across blade of no particular significance.

Figure 126.—Atiu paddle (Bishop Mus., C2914). a, front of paddle: length of shaft, 1,538 mm., of blade, 585 mm., total, 2,123 mm.; diameter at end of shaft, 35 mm., of smaller projection, 22 mm., at middle, 44 mm., at shoulder with blade, 38 mm. transversely and 32 mm. front to back. b, enlargement of front of blade: width at shoulders (1) 92 mm., at widest part (2), 222 mm.; median ridge (3), 212 mm. long and 15 mm. wide, notched on sides and incised at lower end with owner's initials, M.K.; two curved lines (4) across blade of no particular significance.

page 191 contact, also changed the form of their paddles to resemble the general shape of the English oar.
A type specimen collected in Atiu in 1929 has a long shaft (kakau) and a blade (rapa) square shouldered at the shaft junction and with the greatest
Figure 127.—Aitutaki paddle (Cranmore Mus., Kent). a, front showing carved upper end (1) of shaft, narrow shoulders (2), carving on blade below shoulder (3) and long mesial edge (4). Total length, 2,425 mm.; shaft length, 1,073 mm.; blade length, 1,352 mm.; shaft diameter in middle, 52 mm., at shaft junction, 54 mm. by 45 mm.; blade width at shoulder, 82 mm., at widest part, 342 mm. b, details of front carving (a, 3) 183 mm. deep; these may be analyzed into transverse panels with serrated lower edges and rows of motifs which are V-shaped or triangular. with their points directed downward toward alternate points in the lower edge and also elliptical figures, panels are as follows: 1, lower serrated edge, row of V-shaped figures, row of triangles, upper edge with spaced serrated projections directed upward; 2, serrated lower edge, alternating triangles, and serrated upward projection from upper edge; 3, serrated lower edge, alternating triangles, middle panel projecting upward with two elliptical figures, and narrow serrated panels projecting from upper edge laterally to reach side edges of blade; 4, lower serrated edge, alternatiing triangles, and straight upper edge; 5, narrow transverse panel with both edges serrated; 6, lower serrated edge, alternating triangles, row of elliptical figures, upper edge scalloped in series of convex curves; 7, median vertical panel extending down 100 mm.; upper end in trident shape; three median lozenges with short lateral serrated projections on either side from middle and junction of lozenges; lower end continues into raised median ridge which extends to blade tip. c, carving on back of blade, similar to front (b) without serrated projections from upper border of panels: 1, serrated lower edge, row of alternating V-figures, row of triangles, straight upper edge; 2, panel with both edges serrated; 3, serrated lower edge, alternating triangles, straight upper edge; 4, both edges serrated; 5, serrated lower edge, alternating triangles, scalloped upper edge with double curves (elaboration of upper edge of b, 6). d, carved upper end of shaft, length, 82 mm.; width of each of four surfaces, 73 mm.: each surface has two human figures (1) at upper end, faces without details, hands clasped on abdomen, legs flexed, and bounded below by serrated edge; 2, three raised crescents in two vertical rows, bounded below by serrated edge; 3, row of raised geometrical figures formed of upper and lower lozenges with split tails directed to middle and connected by larger lozenge-shaped body (resembles figures on Mitiaro gods, fig. 228); 4, horizontal panel with both edges serrated, e, various painted motifs: 1, side edge of blade; 2, motif based on the carved motif (b, 7).

Figure 127.—Aitutaki paddle (Cranmore Mus., Kent). a, front showing carved upper end (1) of shaft, narrow shoulders (2), carving on blade below shoulder (3) and long mesial edge (4). Total length, 2,425 mm.; shaft length, 1,073 mm.; blade length, 1,352 mm.; shaft diameter in middle, 52 mm., at shaft junction, 54 mm. by 45 mm.; blade width at shoulder, 82 mm., at widest part, 342 mm. b, details of front carving (a, 3) 183 mm. deep; these may be analyzed into transverse panels with serrated lower edges and rows of motifs which are V-shaped or triangular. with their points directed downward toward alternate points in the lower edge and also elliptical figures, panels are as follows: 1, lower serrated edge, row of V-shaped figures, row of triangles, upper edge with spaced serrated projections directed upward; 2, serrated lower edge, alternating triangles, and serrated upward projection from upper edge; 3, serrated lower edge, alternating triangles, middle panel projecting upward with two elliptical figures, and narrow serrated panels projecting from upper edge laterally to reach side edges of blade; 4, lower serrated edge, alternatiing triangles, and straight upper edge; 5, narrow transverse panel with both edges serrated; 6, lower serrated edge, alternating triangles, row of elliptical figures, upper edge scalloped in series of convex curves; 7, median vertical panel extending down 100 mm.; upper end in trident shape; three median lozenges with short lateral serrated projections on either side from middle and junction of lozenges; lower end continues into raised median ridge which extends to blade tip. c, carving on back of blade, similar to front (b) without serrated projections from upper border of panels: 1, serrated lower edge, row of alternating V-figures, row of triangles, straight upper edge; 2, panel with both edges serrated; 3, serrated lower edge, alternating triangles, straight upper edge; 4, both edges serrated; 5, serrated lower edge, alternating triangles, scalloped upper edge with double curves (elaboration of upper edge of b, 6). d, carved upper end of shaft, length, 82 mm.; width of each of four surfaces, 73 mm.: each surface has two human figures (1) at upper end, faces without details, hands clasped on abdomen, legs flexed, and bounded below by serrated edge; 2, three raised crescents in two vertical rows, bounded below by serrated edge; 3, row of raised geometrical figures formed of upper and lower lozenges with split tails directed to middle and connected by larger lozenge-shaped body (resembles figures on Mitiaro gods, fig. 228); 4, horizontal panel with both edges serrated, e, various painted motifs: 1, side edge of blade; 2, motif based on the carved motif (b, 7).

page 192 width hearer the tip (fig. 126). A median ridge (arero) extends upward from the pointed tip for a couple of inches on one surface (aro, front) of the blade; the other plain surface is called the tua (back). A shaft end projection (parepare) is less in diameter than the shaft. The long shaft enabled the paddler to stand if he chose. The paddles of Mauke and Mitiaro are of similar form.

In Aitutaki, the old form of paddle no longer exists, but I was told that it resembled the Atiu paddle (70, p. 269). My informant made a drawing which was subsequently found to be correct as regards shape, but the mesial ridge was omitted. Portlock (quoted by Hornell, 44, p. 168) gives the length as 5 feet 7 inches and the greatest width of the blade as 9 inches. Portlock adds (46, p. 229): "One side of the blade, nearly at the grasp, is cut out like Cornish work and on the other side, just above the point of the blade, is a ridge about half an inch above the surface of the blade. Besides ornamenting they take pains in staining them with a black dye in variety as their fancy directs."

Figure 128.—Paddle, probably from Aitutaki (W. O. Oldman coll.): 1, narrow shoulder; 2, incised curved lines below shoulders, two double curves with convexity downward; 3, second set of three triple curves with convexity upward; 4, median notched ridge at tip end.

Figure 128.—Paddle, probably from Aitutaki (W. O. Oldman coll.): 1, narrow shoulder; 2, incised curved lines below shoulders, two double curves with convexity downward; 3, second set of three triple curves with convexity upward; 4, median notched ridge at tip end.

Portlock (46, p. 227) procured four paddles in 1792, and these should be somewhere in England. Fortunately, a fine specimen in the Cranmore Museum, Kent, labeled "Cook Group" is without doubt from Aitutaki (fig. 127). It is very large, being 8 feet long with a blade 13.5 inches across the widest part. It has square shoulders and a pointed tip, and the greatest width of the blade is toward the tip end. It thus conforms to the main features of the Atiu type. From the shoulders downward for some inches, the blade is carved with motifs that are identical with those on the authenticated chief's seat from Aitutaki in the Auckland Museum (70, p. 369). This is evidently the "Cornish work" referred to by Portlock. The carving is on both sides of the blade, but on the back it extends downward in the middle line as a narrow band for an extra four inches and from its lower end a mesial ridge extends to the tip of the blade. This paddle thus departs from the Atiu form in that the ridge is not confined to a few inches at the tip. Blade and shaft are further ornamented with various motifs stained with a black dye, which further conforms to Portlock's description. The upper end of the shaft is enlarged and carved.

page 193

A paddle in Oldman's collection, attributed to the Hervey Islands, is peculiar in that the blade is narrow at the shoulder and has a distinct concave curve on the sides before reaching the greatest width which is nearer the tip than the shoulder. A raised median ridge, with notched edges, on the front extends upward for some inches from the tip. The blade is carved with curved lines near the shoulder and lower down. As this curved motif is on the Cranmore paddle (fig. 127, d, 2), the Oldman paddle (fig. 128) may also be from Aitutaki. The shaft is plain without any enlargement or carving at the top end.

Figure 129.—Aitutaki paddle without shoulders (W. O. Oldman coll., 368): total length, 78 inches; blade length, 35 inches; blade width, 6 inches, a, upper part of back of blade showing slope to shaft and carving (1). b, lower part of front of blade with long mesial ridge and, above it, four short transverse raised panels with serrated edges. c, enlargement of carving (a, 1): 1, slightly raised horizontal panel with serrated upper and lower edges, row of V-shaped figures with points opposite every alternate point in lower serrated edge, and row of sunken triangles above each V-shaped figure; 2, narrow horizontal panel with serrated upper and lower edges; 3, repetition of lowest panel (1) but with two rows of V-shaped figures instead of one row; 4, lozenges, triangles, and V-shaped figures in horizontal rows. d, enlargement of lower end of blade (b). e, head of shaft, enlarged four sided with human face incised on each surface. (From collection of late Thomas Brynton, F.S.A., York Museum.)

Figure 129.—Aitutaki paddle without shoulders (W. O. Oldman coll., 368): total length, 78 inches; blade length, 35 inches; blade width, 6 inches, a, upper part of back of blade showing slope to shaft and carving (1). b, lower part of front of blade with long mesial ridge and, above it, four short transverse raised panels with serrated edges. c, enlargement of carving (a, 1): 1, slightly raised horizontal panel with serrated upper and lower edges, row of V-shaped figures with points opposite every alternate point in lower serrated edge, and row of sunken triangles above each V-shaped figure; 2, narrow horizontal panel with serrated upper and lower edges; 3, repetition of lowest panel (1) but with two rows of V-shaped figures instead of one row; 4, lozenges, triangles, and V-shaped figures in horizontal rows. d, enlargement of lower end of blade (b). e, head of shaft, enlarged four sided with human face incised on each surface. (From collection of late Thomas Brynton, F.S.A., York Museum.)

page 194

The square shoulders in the last two examples are much narrower in proportion to the blade than the Atiu type paddle. A further exaggeration of the narrowing tendency in Aitutaki is shown by a paddle in the Oldman collection which has no shoulders (fig. 129). The blade is long and narrow without the characteristic widening toward the tip. It is carved on the upper front part of the blade with typical Aitutaki motifs. About half way down the back are four raised transverse zigzag ridges, also typical of Aitutaki; and immediately below them a long mesial ridge extends downward to the tip of the blade. The upper end of the shaft is enlarged and squared, with a human head carved on each surface. Except for a doubtful object attributed to Mangaia (fig. 234), the use of a human motif seems to have been confined to Rarotonga and Aitutaki. The Cook Islands as the location of the paddle would be doubtful because of its form, were it not for the carving motifs which make its identification with Aitutaki absolutely certain.

An even greater anomaly is presented by a paddle in the Peabody Museum, Cambridge, Massachusetts, which is said to have been collected by Agassiz in Aitutaki (fig. 130). It has no shoulders, the blade is evenly elliptical with the
Figure 130.—Aberrant paddle, Aitutaki (Peabody Mus., Cambridge, Massachusetts, 55419): length, 5 feet 9.25 inches, elliptical blade with greatest width, 9.75 inches. a, full length. b, blade enlarged: shows mesial ridge (1) extending full length of blade. c, side view of blade: lateral edge (2) of blade, showing projections (1, 1) of mesial ridge on each surface.

Figure 130.—Aberrant paddle, Aitutaki (Peabody Mus., Cambridge, Massachusetts, 55419): length, 5 feet 9.25 inches, elliptical blade with greatest width, 9.75 inches. a, full length. b, blade enlarged: shows mesial ridge (1) extending full length of blade. c, side view of blade: lateral edge (2) of blade, showing projections (1, 1) of mesial ridge on each surface.

greatest width at the middle, and there is a mesial ridge on both surfaces that extends for the full length of the blade. The top end of the shaft is carved with linear incisions which do not aid identification. It is such a departure from Cook Islands form that its intrusion from elsewhere is probable.

A Mangaian paddle (fig. 131) in the British Museum was figured by Edge-Partington (24, I-5-5) and reproduced by Hornell (44, p. 168). The locality page 195is confirmed by the typical K-motif in carving on the mesial ridge and on the upper end of the shaft. While the blade conforms to the Atiu paddle in form, it has two features that mark a slight local variation that may be useful for identification. The blade shoulder is very wide with a concave upper edge; and the sides of the blade, instead of being evenly curved at the widest part, form an obtuse angle. The widest part is also nearer the tip than it is in the Atiu type.

Figure 131.—Carved paddle, Mangaia (British Mus., 3000). a, full blade with shortened handle: total length, 1,740 mm.; blade length, 745 mm.; blade length from shoulder to widest part, 502 mm.; length widest part to tip, 243 mm. b, upper part of blade: concave shoulder, 150 mm. wide and 31 mm. thick in middle line; four rows of small triangles with bases forming continuous line. c, lower end of blade: greatest width, 220 mm. where sides form an obtuse angle; mesial ridge, 312 mm. long and 15 mm. wide, carved throughout. d, enlargement of section of mesial ridge, showing transverse K-motif in carving. e, enlargement of top of shaft: squared and divided into three horizontal panels on each surface and carved with K-motif; middle part of middle panel perforated with rectangular hole from front back and from side to side.

Figure 131.—Carved paddle, Mangaia (British Mus., 3000). a, full blade with shortened handle: total length, 1,740 mm.; blade length, 745 mm.; blade length from shoulder to widest part, 502 mm.; length widest part to tip, 243 mm. b, upper part of blade: concave shoulder, 150 mm. wide and 31 mm. thick in middle line; four rows of small triangles with bases forming continuous line. c, lower end of blade: greatest width, 220 mm. where sides form an obtuse angle; mesial ridge, 312 mm. long and 15 mm. wide, carved throughout. d, enlargement of section of mesial ridge, showing transverse K-motif in carving. e, enlargement of top of shaft: squared and divided into three horizontal panels on each surface and carved with K-motif; middle part of middle panel perforated with rectangular hole from front back and from side to side.

A painted paddle in the American Museum of Natural History was attributed to "Rarotonga Island, Hervey Group", but its shape and the addition of characteristic carving prove conclusively that it is from Mangaia (fig. 132). It has wide shoulders with concave upper edge and an obtuse angle at the widest part, in every way similar to the British Museum specimen. A narrow strip below the shoulder, both front and back, is carved with two rows of the K-motif which fuse together. A median panel, extending the whole length page 196of the blade on the front, is carved with the K-motif. The upper end of the shaft is enlarged and also carved with the K-motif. Both sides of the blade and the whole length of the shaft are stained black with a relief pattern formed by the unstained parts of the wood.

Figure 132.—Carved painted paddle, Mangaia (American Mus. Nat. Hist., S.4801). a, shaft (1), length 815 mm., thickness 37 by 30 mm., with carved upper end (2); blade (3), length 970 mm., width at shoulder (4) 180 mm., width at widest part (5) 222 mm.; median carved panel (6) on front, 16 mm. wide, extending from tip to just above shoulder junction; carved transverse panel (7) extending from edge to edge just below shoulder edge. b, carved end of shaft, rectangular, 43 mm. wide, 40 mm. thick; each surface divided into three panels with characteristic K-pattern. c, front of shoulder showing concave upper edge with transverse carved panel (7) 180 mm. deep, median carved panel (6), and examples of painted motifs (8) in outline. d, tip of blade, showing obtuse angles at widest part (5) of blade, and continuation of median panel (6). e, back of shoulder with transverse carved panel (9) similar to front, no median panel, and painted motifs (8) on blade and shaft. f, painted motifs that cover blade and shaft with background painted black; stalks joining lozenges to vertical main stalk are usually alternating (1) but a few are opposite (2); note small pointed projections between side stalks which may alternate (1) or be opposite (2).

Figure 132.—Carved painted paddle, Mangaia (American Mus. Nat. Hist., S.4801). a, shaft (1), length 815 mm., thickness 37 by 30 mm., with carved upper end (2); blade (3), length 970 mm., width at shoulder (4) 180 mm., width at widest part (5) 222 mm.; median carved panel (6) on front, 16 mm. wide, extending from tip to just above shoulder junction; carved transverse panel (7) extending from edge to edge just below shoulder edge. b, carved end of shaft, rectangular, 43 mm. wide, 40 mm. thick; each surface divided into three panels with characteristic K-pattern. c, front of shoulder showing concave upper edge with transverse carved panel (7) 180 mm. deep, median carved panel (6), and examples of painted motifs (8) in outline. d, tip of blade, showing obtuse angles at widest part (5) of blade, and continuation of median panel (6). e, back of shoulder with transverse carved panel (9) similar to front, no median panel, and painted motifs (8) on blade and shaft. f, painted motifs that cover blade and shaft with background painted black; stalks joining lozenges to vertical main stalk are usually alternating (1) but a few are opposite (2); note small pointed projections between side stalks which may alternate (1) or be opposite (2).

Hornell (44, p. 168), who described a paddle in the Pitt-Rivers Museum, Oxford, thought that it probably represented the ordinary paddle of Mangaia. A similar paddle in the British Museum (fig. 133) conforms to the Atiu type in having a square shoulder, the blade wider toward the tip, and a short mesial ridge at the tip; but it varies slightly in that the mesial ridge is on the back as well as the front. The shaft is plain with no enlargement at the upper end. Like the Oxford paddle, it has transverse rows of chevrons below the shoulder, but there are 19 rows as compared with 12 rows on the Oxford paddle. The straight shoulder is of medium width, and the sides of the blade are evenly curved at the greatest width which differentiates the two paddles from the Mangaian form. The zigzag lines seem, at first sight, to have some affinity page 197with the carving below the shoulder of the Mangaian paddle (fig. 131). However the Mangaian carving consists of four rows of small triangles with their bases touching to form a continuous line; and this, with the differences in the form of the blade, outweigh the apparent similarity. The greatest width of the shoulder distinguishes them from Aitutaki paddles, and it is probable that the two paddles belong to Atiu, Mauke, or Mitiaro or to Rarotonga where the zigzag-line motif occurs.

Figure 133.—Cook Islands paddle (British Mus., 1912-123). a, with handle shortened: total length, 2,020 mm.; shaft length, 1,070 mm.; shaft diameter at upper end, 43 mm.; at shaft junction, 52 mm. by 41 mm.; blade length, 950 mm.; width at shoulder, 128 mm.; greatest width, 250 mm. b, enlargement of upper part of blade: shows zigzag incised lines for depth of 102 mm. on both surfaces and on side edges which are 23 mm. thick. c, lower end of blade: shows median ridge, 9 mm. wide, 4 mm. thick and 104 mm. long on one surface and 96 mm. on the other.

Figure 133.—Cook Islands paddle (British Mus., 1912-123). a, with handle shortened: total length, 2,020 mm.; shaft length, 1,070 mm.; shaft diameter at upper end, 43 mm.; at shaft junction, 52 mm. by 41 mm.; blade length, 950 mm.; width at shoulder, 128 mm.; greatest width, 250 mm. b, enlargement of upper part of blade: shows zigzag incised lines for depth of 102 mm. on both surfaces and on side edges which are 23 mm. thick. c, lower end of blade: shows median ridge, 9 mm. wide, 4 mm. thick and 104 mm. long on one surface and 96 mm. on the other.

In spite of the Aitutaki variations, the paddles now in use in Atiu, Mauke, and Mitiaro and the museum specimens from Aitutaki and Mangaia establish the old form for the Cook Islands as being shouldered blades with the greatest width nearer the tip than the shoulder and with a raised median ridge extending for a varying distance from the tip on the front of the blade.

The correct locality of a number of painted paddles has caused some perplexity. Two in the American Museum of Natural History, New York (S.4601, S.4602), are attributed to Jarvis Island, which is an uninhabited coral atoll near the equator. Neither has shoulders, but the sides of the blades form obtuse angles at the widest part, which is nearer the tip. The blades are painted with circles, crescents, and tree forms in black with smaller motifs on the surfaces and edges of the blade and on the shaft. One of these motifs resembles a Mangaian tattooing motif (fig. 71, b), and the lower part of one paddle is carved with a continuous lozenge pattern also used on some of the pedestals of the Mangaian ceremonial adzes. Another paddle, in the Peabody Museum, Salem (E.6440), is similar in that it has no shoulder and it has identical painted circles, but the obtuse angle at the widest part of the blade is not so marked. The wood is very light, and it is evident that the paddles were made for some festival rather than for use in propelling canoes. The key paddle was discovered in the Peabody Museum, Cambridge (fig. 134). Though it has no page 198shoulder, it has the obtuse angles at the widest part toward the tip and the circular and other motifs on the painted blade. The upper end of the blade is carved in horizontal panels divided by incised lines, and the shaft is cut into horizontal panels carved with the typical Mangaian K-motif. The carving thus proves that the painted paddles are Mangaian. In paddles for some social function, the craftsmen did away with the blade shoulders and omitted the mesial strengthening ridges of the tip. One of the painted motifs (fig. 134, f) on this paddle provides the key for the identification of certain round bowls as Mangaian (fig. 5, h). In two of the paddles, the owners' name is printed in black.

Figure 134.—Mangaian painted paddle (Peabody Mus., Salem, Massachusetts, 53508). a, front: total length, 1,645 mm.; greatest blade width, 265 mm.; shaft (1) length, 978 mm.; rectangular in section, 58 mm. wide, 26 mm. thick, carved with K-motif in horizontal panels; upper part of blade (2) carved on raised part for 155 mm.; lower blade (3) painted with circles and smaller motifs in black. b, back: shaft (1) carved like front; upper blade (2) painted with tree-like motif, lower blade (3) with circles, crescents, and smaller motifs; transverse panel (4) with peculiar motif (see f) which is also used in a median panel extending to blade tip, and along margins of sides. c, side view: shows transverse panels on shaft (1) continuous with front and back panels; raised upper part (2) on front, carved; side edge of blade, 13 mm. thick, painted with motif shown in f. d, junction of shaft with blade, front: 1, lowest panel on shaft, with typical K-motif, similar panels extend throughout length of shaft both front and back; 2, commencement of blade surface (a, 2) with horizontal lines; 3, two upper carved panels separated by narrow groove, with similar panels for rest of raised part. e, side view of shaft showing three carved panels. f, painted panel (1) on back of blade (b, 4) with notched lozenge motif, seen also on side margin (2). This motif led to identification of round bowls. (See figure 5, c, h.)

Figure 134.—Mangaian painted paddle (Peabody Mus., Salem, Massachusetts, 53508). a, front: total length, 1,645 mm.; greatest blade width, 265 mm.; shaft (1) length, 978 mm.; rectangular in section, 58 mm. wide, 26 mm. thick, carved with K-motif in horizontal panels; upper part of blade (2) carved on raised part for 155 mm.; lower blade (3) painted with circles and smaller motifs in black. b, back: shaft (1) carved like front; upper blade (2) painted with tree-like motif, lower blade (3) with circles, crescents, and smaller motifs; transverse panel (4) with peculiar motif (see f) which is also used in a median panel extending to blade tip, and along margins of sides. c, side view: shows transverse panels on shaft (1) continuous with front and back panels; raised upper part (2) on front, carved; side edge of blade, 13 mm. thick, painted with motif shown in f. d, junction of shaft with blade, front: 1, lowest panel on shaft, with typical K-motif, similar panels extend throughout length of shaft both front and back; 2, commencement of blade surface (a, 2) with horizontal lines; 3, two upper carved panels separated by narrow groove, with similar panels for rest of raised part. e, side view of shaft showing three carved panels. f, painted panel (1) on back of blade (b, 4) with notched lozenge motif, seen also on side margin (2). This motif led to identification of round bowls. (See figure 5, c, h.)

page 199

Though the shape of paddle blade may vary in different groups of islands in Polynesia, there is one constant feature—the lower end of the blade is pointed. A wide convex lower edge, such as that in the modern paddles of Rarotonga, Aitutaki, and Mangaia, is such a marked departure that the influence of the European oar is obvious. The preservation in museums of old paddles from Mangaia and Aitutaki proves conclusively that the present departure from form is post-European. A similar departure from an old form occurs in the Society Islands, where the change may be attributed to post-European copying of the oar as regards shape but not manner of use. In Manihiki, the paddles are narrow and consequently have a very sharp point. A specimen in Bishop Museum (75, pl. 9, A, 5) was catalogued as a paddle from Manihiki, but its wide convex lower edge made me regard it with suspicion. Later comparative study confirmed this suspicion, for the specimen proved to be a modern copy of a Marquesan paddle-shaped war club.

Poles

A pole (toko) of hibiscus sapling was carried by every canoe in Aitutaki for poling the canoe in the shallow parts of the lagoon, where poling was more effective than paddling. Paddles could be used for poling in much shallower water, and it is probable that the median ridge extending from the tip of the blade was for the purpose of strengthening the tip of the paddle that came in contact with the coral bottom when poling. In the other islands, the narrow shore lagoons were not so favorable for poling, but poles supplemented paddles on occasion.

Bailers

A bailer (tata or a'u) forms part of the equipment of every canoe. For small fishing canoes, a half coconut shell was sufficient; but, for large canoes, large wooden bailers were made. The technique was to cut a suitable section of tree trunk, usually hibiscus or breadfruit; trim it on the round; and dub out a flat plane to form the upper surface of the bailer. Two rectangular slots were cut on the flat surface on either side of a median bar which was to form the handle. Through the slots and the front end of the section, the wood was hollowed out to form a cylinder, the front end open and the back end closed. In the Atiu and Mauke bailers (fig. 135), the slots for the fingers to pass through are rectangular; but I saw a bailer in Aitutaki (70, p. 269, fig. 235) in which the outer sides of the slots were slightly curved. In Rarotonga and Mangaia, only half coconut shells and small cans were seen in use, but the Cambridge University Museum has a small wooden bowl of the beaker type (fig. 4, a, b) from Mangaia that is said to have been used as a bailer (tata te riu, bail out the bilge). Though it may have functioned as a bailer, I do not page 200think that it represents the old Mangaian type. It may be assumed that the form of bailer found in Aitutaki, Atiu, and Mauke represents the Cook Islands type. It differs from the open scoop in general use in Polynesia, and its significance is discussed on page 451.

Figure 135.—Canoe bailers: a-c, Atiu (Bishop Mus., C2836); d-f, Mauke (Bishop Mus., C2837). a, side view with sloping back: greatest length, 241 mm.; front depth, 117 mm. b, upper surface: square front, curved back; front width, 152 mm.; width at back of slots, 155 mm.; slots on either side of median bar handle, 119 mm. by 39 mm. and 116 mm. by 38 mm.; handle, 22 mm. wide and 21 mm. thick, c, front opening: upper surface, 152 mm. wide; widest part, 155 mm. d, side view with square back: depth front, 157 mm., middle, 167 mm.; and back, 151 mm.; length, 286 mm. e, upper surface: front width, 172 mm.; width at back of slots, 184 mm.; slots 210 mm. long by 50 mm. in front and 47 mm. back and 205 mm. long by 54 mm. in front and 52 mm. back; bar handle, 37 mm. wide by 32 mm. thick; part of front crosspiece broken away. f, front opening: top width, 172 mm.; greatest width, 204 mm.

Figure 135.—Canoe bailers: a-c, Atiu (Bishop Mus., C2836); d-f, Mauke (Bishop Mus., C2837). a, side view with sloping back: greatest length, 241 mm.; front depth, 117 mm. b, upper surface: square front, curved back; front width, 152 mm.; width at back of slots, 155 mm.; slots on either side of median bar handle, 119 mm. by 39 mm. and 116 mm. by 38 mm.; handle, 22 mm. wide and 21 mm. thick, c, front opening: upper surface, 152 mm. wide; widest part, 155 mm. d, side view with square back: depth front, 157 mm., middle, 167 mm.; and back, 151 mm.; length, 286 mm. e, upper surface: front width, 172 mm.; width at back of slots, 184 mm.; slots 210 mm. long by 50 mm. in front and 47 mm. back and 205 mm. long by 54 mm. in front and 52 mm. back; bar handle, 37 mm. wide by 32 mm. thick; part of front crosspiece broken away. f, front opening: top width, 172 mm.; greatest width, 204 mm.

Mast and Sail

The mast was termed tira, and the general term for sail was kie, but in Aitutaki the k was replaced by the glottal ('ie). As the Aitutakians have the k sound in their dialect, it is probable that the change was due to the early influence of Tahitian missionaries. [A similar change took place in the name 'inaki (fish trap) which in Aitutaki is 'ina'i, the initial glottal representing the general Cook Islands dropping of h and the second glottal the Tahitian dropping of k.] In Aitutaki, 'a'angi is an archaic term for sail. Except in Aitutaki, a canoe with a sail was referred to as a vaka takie.

page 201

The use of sails has long been abandoned, except in Aitutaki, where practically every canoe has a sail made of canvas, with a European rig. The masts are poles of 'au which pass down through a median hole in the thwart and are stepped in a wooden ring on the bottom of the hull (70, p. 270). There is a perforated thwart fore and aft, for the canoes always sail with the outrigger on the windward side because they have no balance board on the other side. When sailing, the outrigger float comes up out of the water and thus reduces friction. If the float rises so high as to threaten capsizal, the operator leans out on the after boom or the longitudinal stringer to weigh down the float. At the end of the board, the mast is quickly unshipped and transferred to the other thwart; and the canoe makes the next board stern first. The continued use of sails in Aitutaki is undoubtedly due to its large lagoon with deep water, a feature lacking in the other islands of the group.

The old sails must have been triangular in one form of the oceanic spritsail (44, vol. 3, p. 11), but I could get no information as to the old form used on the larger canoes. I was told in Aitutaki, however, that four-sided sails, sometimes formed from a sleeping mat, were used. I was also told that one side was fastened to the mast and that a diagonal sprit (toko) was used to expand the sail (70, p. 270). Hornell (44, p. 167) justly remarks that such a sail is foreign to Polynesia but was adopted in various islands, notably Tahiti, soon after European contact. He gives this as an explanation for its possible occurrence in Aitutaki. My informant's contention that the four-sided sail was old is borne out by a drawing by Tobin (46, p. 134), who visited Aitutaki in 1792. The drawing depicts an outrigger canoe with three men. A man in the bow is shown holding an upright spear in each hand with a square sail, probably a mat, attached by the corners to the two spears. However, the use of two hand-held spears as masts shows that this rig was a makeshift. K. P. Emory (ms.) states that in the Tuamotu a rectangular sleeping mat is often used as a sail when visits are made to the smaller islets on the reef. Two corners are tied to the mast and a diagonal strut is tied to the lower corner at the mast and the free upper corner. A rope is then tied to the lower free corner as a sheet to adjust the angle of the sail to the wind. It is thus possible that rectangular sails were used in pre-European days as a temporary expedient and that they owe their shape to the form of the sleeping mats that were conveniently at hand during short trips across the lagoon to the islets on the reef. Lashings were effected by bulging out the corner of the mat with a pebble and tying the cord or rope around the constricted neck so formed. This method applied to both the lashings to the mast and the diagonal sprit. The journey ended, the lashings were unfastened and the temporary sail resumed its normal function as a sleeping mat.

page 202

Anchors

Anchors (tutau, 'akamou) were stones of suitable weight attached to a rope. I did not see any that were shaped or perforated.

Skids

Skids ('akapapa) of medium-sized logs or branches were laid down in a series ('akapapa) between the water and the canoe shed to facilitate the hauling of large craft. The term rango applied to skids in other islands was applied in Aitutaki to the pieces of wood in the canoe shed upon which the canoe was rested to keep it off the ground (70, p. 271).

Decoration

From the brief accounts of early voyagers, decoration purely for art included carving, painting, and the use of feather streamers. Such decorations were used only on the canoes of important persons.

In Mangaia, Cook (20, p. 173) saw only one small canoe with the upper end of the stern post forked. The upper part of the hull, probably the gunwale rails and end covers, was painted black, perhaps for utility as well as for color
Figure 136.—Painted hull of Mangaian model canoe (Dominion Mus., N.Z.): shows triple motif similar to perforated motif used in poncho garment; also vertical lines with curved upper end forming single spiral.

Figure 136.—Painted hull of Mangaian model canoe (Dominion Mus., N.Z.): shows triple motif similar to perforated motif used in poncho garment; also vertical lines with curved upper end forming single spiral.

page 203 contrast. A model canoe figured by Hornell (44, p. 160) has carving every place but on the hull. A small canoe figured by Best (7, pp. 206, 207) is painted on the hull with various motifs in black (fig. 136). Models are usually overdecorated for sales appeal, but it is probable that some canoes were both carved and painted.

Mr. Anderson, with Cook's third voyage (20, pp. 196, 197), states that two of the ten double canoes he saw in Atiu "were most curiously stained or painted, all over with black, in numberless small figures; as squares, triangles, etc. and excelled, by far, anything of that kind I had ever seen at any other island in this ocean."

Belcher (5, p. 19) says that the Rarotongan canoes were beautifully carved, and he gives a drawing of an outrigger canoe with the high stern piece elaborately carved.

Portlock (46, p. 229) mentions that the canoes of Aitutaki were ornamented all round with a kind of little red pea with a black eye stuck on the outside of the canoe and that some of the large canoes had a high stern piece decorated with many man-of-war bird feathers. The feather decoration of the stern piece was also a feature of the New Zealand war canoes.

With the passing of large canoes, the art of canoe decoration has died out in all the islands. Modern small fishing canoes are evidently not deemed worth the trouble involved in carving and painting.

Amulets

Hornell (44, p. 170) quotes Williams (81, p. 99) in stating that a wooden image was carried by Rarotongan fishermen in the bows of their canoes to give them luck in fishing. Several such images have been preserved in museums, and some are shown in this work (figs. 191-193). The memory of them has completely faded and the showing of photographs to various informants failed to elicit any information.

In Mangaia, the tip end of a coconut leaf with a piece of sennit tied in a knot around the midrib was used as a charm in the bow of each canoe (Gill, 30, p. 104, quoted by Hornell, 44, p. 170). These were distributed to the fishermen by the priest known as the Ruler of Food on the beach before the fishing fleet shoved off. (See figure 231, b.) When the Ruler of Food accompanied the fleet, he had four amulets of bark cloth, shaped like cones and termed poani (plug), placed under the port and starboard lashings of the fore boom to the gunwales and under the bow and stern covers for the purpose of plugging the holes of the four cardinal winds (76, p. 146). It is probable that the other islands had a similar custom but information is lacking.

page 204

Double Canoes

Double canoes (vaka katea or vaka purua) are now found only in Atiu and Mauke. They have practically ceased to be used in their original function for torching for flying fish. The individual canoes are made on the same lines as the single outrigger canoes with slender bows ending in the hull projection (tara kokiri) and with raised stern pieces. (See plate 10, B.) The double canoes are larger and thus have one or two hull joins. The starboard hull is always longer than the port hull, as the measurements of five double canoes verify (table 2).

Table 2.—Canoe Measurements
Starboard Hull Port Hull Difference
a 25 ft. 7 in. 21 ft. 11 in. 3 ft. 8 in.
b 27 ft. 5 in. 21 ft. 9 in. 5 ft. 8 in.
c 29 ft. 11 in. 22 ft. 0 in. 7 ft. 11 in.
d 25 ft. 9 in. 22 ft. 3 in. 3 ft. 6 in.
e 25 ft. 7 in. 20 ft. 2 in. 5 ft. 5. in.
average 26 ft. 10 in. 21 ft. 7.5 in. 5 ft. 4 in.

Hornell (44, p. 173) says, "The two hulls are approximately equal in dimensions, the port hull slightly the shorter. …" However, table 2 shows that the average difference between the two hulls is 20 percent.

Informants told me that canoes of equal length would not steer well. The longer starboard canoe is regarded as the tuakana (elder brother) and the port canoe, the teina (younger brother). As regards position, the starboard canoe is katea and the port canoe, which takes the place of the float of an outrigger canoe, is termed ama (float).

The terms applied to the various parts of the canoes are the same as those for the outrigger canoe, but the terms katea or ama were added to distinguish between the similar parts of the two canoes. Thus, the stern cover of the starboard canoe is the 'utu noko katea and that of the port canoe, the 'utu noko ama.

The holes for the hull join are two inches from the butt edges, and the paired holes are three inches apart. Commencing with the first pair of holes below the gunwale edge, three turns were made with the lashing braid, the next pair four, the next five, and the rest six turns. When the other side of the hull was reached, they diminished in like manner toward the gunwale. There are usually strengthening ribs (vae), two in the larger canoe and one in the smaller. The ribs which extend from gunwale to gunwale across the hull are rounded off and are about 1.5 inches high by 5 inches wide.

The two canoes are placed side by side with bows usually level, but some of the Mauke canoes had the bow of the port canoe a little (6 inches) behind the other. The sterns were slightly farther apart than the bows. In one canoe examined, the hulls were 22 inches apart at the fore boom, 23.5 inches at the page 205mid boom, and 26.5 inches at the after boom. This corresponds with having the fore end of the float closer to the hull than the after end in outrigger canoes. The canoes are connected by three booms termed kiato mua (fore boom), kiato roto (mid boom), and kiato muri (after boom). In the study canoe, the booms are 9 feet, 10 inches; 10 feet, 7 inches; and 10 feet, 8 inches respectively in length, and they average about 3.5 inches in diameter. They project outboard on either side for about 30 inches. The parts over and between the hulls are squared and the outboard parts left on the round, except that the outer ends are leveled off on the upper side to form a bed for a longitudinal spar which may be termed the thole pin stringer. The booms are always tamanu wood.

Figure 137.—Plan of double canoe, Atiu. a, side view, b, view from above: 1, section of platform between canoes; 2, thole pin stringer, c, section of hull (1) and gunwale rail junction (2). d, boom lashing to gunwale rail, e, bow cover, f, stern post (1); 2, canoe stern; 3, stern cover, g, carving on fore end of stern cover. (Drawn by T. W. Downes.)

Figure 137.—Plan of double canoe, Atiu. a, side view, b, view from above: 1, section of platform between canoes; 2, thole pin stringer, c, section of hull (1) and gunwale rail junction (2). d, boom lashing to gunwale rail, e, bow cover, f, stern post (1); 2, canoe stern; 3, stern cover, g, carving on fore end of stern cover. (Drawn by T. W. Downes.)

The booms are securely lashed to both gunwale rails of each hull through two holes pierced through the rail below the boom and set closer together than the width of the boom. Sometimes a third hole is made below and between them. The lashing turns pass through the holes and obliquely over the boom to form the simple lozenge pattern termed puna.

The thole pin stringer ('au) was a length of tamanu wood about 3.5 inches in diameter, slightly greater at the fore end and less at the after end. This was placed upon the ends of the booms and lashed to them. In the double canoe shown in figure 137, the stringer reached all three booms; but in the canoes page 206I examined, the stringer was placed on the after and mid booms and did not reach the fore booms. The shorter stringers were pierced on the upper side with two pairs of holes for thole pins, one pair placed in front of the after boom and the other in front of the mid boom. The thole pins (pine) were an inch thick and projected 4.5 inches above the stringer.

A platform ('ata'ata) was formed of 7 or 8 poles 2 to 2.5 inches thick. These were laid side by side over three cross bars, and lashed firmly to them. The cross bars were so placed that when the platform was over the booms, the front and middle ones fitted down immediately in front of the fore and mid booms and the after bar fitted behind the after boom. The width of the platform corresponded to the open space between the two canoe hulls. The platform was then lashed to the booms in such a manner that it could be easily disengaged. In some of the Mauke canoes, a plank was used instead of the pole platform. Informants stated that if the canoe filled (tomo) with water the platform was untied and thrust end on under the canoe for use as a lever. Paddles were quickly tied together and used in a similar manner on the other side. By the use of such levers, the canoe was lifted up in the water and bailed.

Washboards ('ura) were sometimes set up on edge to the outer side of the gunwale rails to keep out the water. Slots were cut on one edge to fit over the boom.

The double canoes were carved on the bow and stern covers and on the stern posts with geometrical motifs such as chevrons, triangles, curves, and crossed lines. The plan of a double canoe in Atiu, which was drawn for me by the late T. W. Downes of Wanganui, New Zealand, is shown in figure 137.

It is evident that though the hulls of the double canoes adhere to the old form, the superstructure has been changed, owing to the use of oars instead of paddles.

Double canoes are carried by men who stand behind each of the three booms and lift the booms with their arms. In carrying the canoes up a slope, the leading two men lifted the booms on their arms and the rear man lifted the boom onto his shoulders.

Miscellaneous

The following saying is used when an important canoe is being built by master builders.

Ka mate te vaka i roto i te 'are, If misfortune occurs to the canoe within the building shed,
Ka mate katoa te vaka i te moana. Misfortune will also follow it to sea.

The saying is really a hint to the owner to feed his builders well, for, if they allow slovenly work because of lack of attention, the canoe will be unlucky at sea.

page 207

The space between the two hulls or hull and float is termed aro'a. If the owner finds some defect in his newly built canoe, the craftsman asks, "I mate ki'ea" (Where is the defect)? The owner replies, "I mate ki katea" (It is defective to starboard) or "I mate ki aro'a" (It is defective to port). In this sense, the Mauke people never use the term ama for port. In steering, however, the word ama distinguishes the port side. To steer is ruri and the steersman turns to katea or ama as the look-out directs. The term for keeping straight ahead with the current flowing between the canoes is aro'a or ta'e ngatai.

Navigation

As Williams (81, p. 82) has pointed out, voyages were made from definite starting points where there were certain landmarks by which they steered until the stars became visible. In Atiu, in 1823, he asked the high chief Rongo-ma-tane the sailing directions from Atiu to Rarotonga which at the time had not been officially reported by any European. The native's starting point was at Cook's Landing on the west coast. Rongo-ma-tane, after the ship was maneuvered into line with his landmarks, turned about and pointed to sea. Williams took a compass bearing of the direction which was southwest by west "and proved to be as correct as if he had been an accomplished navigator."

At Cook's Landing in 1929, I asked Mataio, the great-grandson of Rongo-ma-tane, the same question. Mataio immediately replied, "The course my great-grandfather Rongo-ma-tane gave to the missionary John Williams was obtained by getting Oravaru and Toka-'o'onu in line and carrying that line out to sea." He pointed out two huge rocks; Oravaru stood up sheer from the shore and Toka-'o'onu towered up from the rising ground some distance away.

On the north coast of the island I was shown the landmark from Atiu to Tahiti. It was a long straight ridge named Te Mimi-o-Te-Arau. The double canoe was maneuvered until the line of the ridge corresponded with the line between the two hulls and then the canoe sailed or paddled on that line until the hill line faded as the guiding star came out.

It is evident that landmarks served but a temporary function. It seems feasible that early voyagers experimented in the favorable seasons by selecting a star and sailing on it to avoid going round in circles. It is also probable that they selected a star astern to guide them home when circumstances permitted. When land was discovered, they fixed their guiding star for that island. In subsequent voyages, the place for leaving was naturally selected as that part of the coast that was nearest to the course of the star. What would be more natural than to look out to sea in the line of the star and then to turn the gaze shoreward to see what natural objects lay in the same line ? Having established page 208shore landmarks, the mariner could set out in daylight instead of waiting until his star rose.