Source Documents
edwards : rafts and canoes, pacific sth america, 1965 

Clinton R. Edwards : Rafts and Canoes, Pacific South America, 1965.

Extracts and illustrations from
Edwards, Clinton R.
: Aboriginal Watercraft of the Pacific Coast of South America
University of California Press, Berkley and Los Angles, 1965.

A intensely detailed work.

Page 1

Many a traveler to western South America has noted strange little watercraft fashioned from bundles of reeds by fishermen of the coast and lake dwellers of the highlands.
Although reed floats have disappeared from much of their former range, in some places they persist virtually unaltered from the ancient form.

Abundant archeological evidence attests to their use in pre-Columbian times.
At the pre-ceramic coastal site of Huaca Prieta, Peru, nets and net floats suggest the use of some sort of watercraft for ocean fishing.
Many representations of reed floats have been found in pre-Inca sites, mostly in the form of pottery vessels dating from Mochica and Chimu times.
There are indications that reed craft far larger than those known in historic timess were built, capable of carrying both cargo and a number of passengers. (1)

Historical records begin with an account of Hernando Pizarro's penetration the Titicaca Basin in the mid-1530's.
The anonymous chronicler mentions the reed floats (balsas de ened) of the Rio Desaguadero, as well as a pontoon bridge constructed of reed bundles.(2)

According to Garcilaso de la Vega (1609), reed floats were used for carrying passengers and cargo across streams and for fishing in the sea.
Two reed bundles, "of the thickness of an ox," wider in the after part tapered forward to a raised point "like the bow of a ship so it will it strike and cut the water," formed simple but serviceable craft.
Coastal men, kneeling atop their floats and paddling with split canes, ventured "four, five, or six leagues out to sea, or farther if necessary."
Garcilaso describes in minute detail the manner of paddling: the hollow side of the cane served as a blade, and the ends were dipped alternately as with a double paddle.
"When such a little craft travels at full speed, the rest-horse will not overtake it."
He gives no details of distribution, attributing their use to "the Indians of the entire coast of Peru."(3)

Jose de Acosta (1590) was intrigued by the reed vessels, and wrote the following vivid account of their use in the waters off Callao:
 "... [the Indians] make floats of sedges or reeds, well tied, which they call balsas.
They carry them on their shoulders to the shore, toss them energetically the water, climb aboard, and thus like horsemen they enter the sea . . . They carry on these bundles their nets and lines ... It was certainly a great di-

Page 2

version for me to see them going out to fish in the Callao de Lima, for .. . each one sits on his little balsa like a . .. jockey, cutting through the waves of the sea, which is very rough where they fish.
They seem like Tritons or Neptunes, painted on the surface of the water.
On reaching the shore, they set their boats on their shoulders, and later take them apart and put the bundles on the beach so they will drain and dry.(4

The earlier descriptions of form, construction, and use are substantiated and expanded by Padre Bernabe Cobo (1653):

The most common [balsas] ... are made of dry reeds or other kinds of sedges, and are formed in this manner:
They tie with cords two bundles of reeds, of a size appropriate for the balsa.
They are well tightened and rounded, with the point of the bow thin, so that they are thicker in the middle and progressively thinner towards the ends.
They are not equally tapered, however, since the end which serves as stern remains thicker, unless both ends are to have the form of a bow, as many balsas have.
In this case the ends of the bundles are tapered equally, the pairs joined together lengthwise, with the points and heads to­gether, tied very strongly . . . The smallest of these balsas are a little more than four codos [about 6 feet] long.
Their circumference in the thickest part amounts to as much as a man can encircle with his arms.
The larger ones are as long as fifteen to twenty feet and ten or twelve feet wide.
The first will not carry more than one or two persons; of the second there are some that carry a dozen.
Two of the large ones joined and tied together comprise a single one capable of carrying horses and cattle.

The various kinds are used in the sea, rivers, and lakes, although it is true that the large ones usually serve only for crossing rivers and lagoons, and the small ones for fishing in the sea.
None of them carry sails; due to their exceed­ing lightness they would capsize with very little wind.
Thus they are propelled with paddles and poles.(5

It is evident that the early Spaniards observed and were somewhat amazed at the natives' strong seagoing tradition and familiarity with the sea.
Experienced seamen of later times, often confronted with the problem of landing under difficult and dangerous conditions in open roadsteads of the Peruvian coast, also remarked on the ease with which the native mariners negotiated the surf.
At Huanchaco, in 1822, George Coggeshall, a Yankee sea captain and trader, entered in his journal the following description of reed float construction and use:

The ships that touch here cannot with any safety use their own boats, and always employ the boats or canoes of the Indians, the surf being too high to venture off and on without the aid of these men, who are almost amphibious.
They are trained to swimming from their infancy, and commence with a small 'Balsa,' in the surf within the reefs, and by degrees, as they grow older and larger venture through the surf, and out upon the broad ocean.
These 'Balsas' are made of reed bound firmly together, with a hole near the after end, for one

Page 3

person; the forward end is tapered, and turned up like a skate or a Turkish shoe.
Those for children are perhaps from five to eight feet long, and those used by the men generally about ten or twelve, and about as large in circumference as a small-sized barrel ... I have seen the men go off through [heavy surf] . . . and have with great anxiety observed them when a high rolling sea threatened to overwhelm them, watch the approaching roller, duck their heads down close to the reed boat, let the billow pass over them like a seal or a wild duck and force their way with perfect confidence through the surf, where no white man would for a moment dare to venture ..."

W. B. Stevenson, in 1829, described reed floats used at coastal villages somewhere north of Lima; the locations are not given specifically.
His description differs from those of his contemporaries in that he speaks of double-ended craft rather than ones with upcurved bows and truncated sterns.
He may have observed a survival of the double-enders described by Gobo, but by the nineteenth century the upraised bow and truncated stern construction seems to have been standardized.
Stevenson goes on to mention the mobility of the coastal Indian fishermen:

When dry, the balsa only weighs a few pounds, so that on one mule the fisherman can carry his boat, his net, and even sufficient materials to build his hut: in this manner they range up and down the coast in search of fish, which they often salt and take either to Lima or some other market.(7)

The extent to which pre-Conquest fishermen ranged along the coast is an interesting and as yet unanswered question.
Did this coastal roaming start after introduction of Old World beasts of burden, or was it an older tradition?
There is no indication that llamas were ever used thus.
In addition to their continued use for fishing, the reed craft were commonly employed during the nineteenth century as courier vessels at open roadsteads where the landing of ships' boats was hazardous.
Such use was recorded by Ruschenberger at Huanchaco in 1833, by which time the cabattito (little horse) was being applied consistently. He saw reed also at Pacasmayo.(8)
In the early 1850's, Lieutenant Gilliss of the United States Naval Astromomical Expedition picked up his ship's mail at Huanchaco anchorage, the mailbag being delivered from shore "by a
courier mounted on a little balsa called a caballito."(9)

The most complete description of caballitos from the nineteenth century literature is that of the botanist Antonio Raimondi, who observed them near Moche in 1859:

The caballitos are formed of four bundles of reeds cut at one end and tapering to a point at the other.
Two of these bundles are underneath and comprise

Page 4

the full length of the caballito, and the other two above are placed longitudinally over the first.
The upper ones are shorter, thus leaving a cavity in which the fishermen place their catch . . . The position which the mounted man takes can be either seated with his legs extended forward, or on his knees.
In difficult passages [through the surf] or when the sea is rough, he lowers his legs and sits astride; thus the name caballito . . . The fisherman manages his boat by means of a double paddle which he holds in his hands in the center and maneuvers paddling to right and left alternately.
This peculiar kind of boat lasts only a month, because litde by little the reed absorbs water and gradually grows heavier.
At times they take the bundles apart, and reassemble them when they have dried out a litde. When they have been rendered useless for this purpose, they use the material for the construction of their houses . . .(10)

Evidently design and construction had not changed since early post-Conquest times.
A number of descriptions by twentieth-century investi­gators of Peruvian fisheries indicate that no subsequent changes have taken place.(11
) Page 8

Page 21


In western South America
, the only displacement-type, open-hulled vessels occurring south of northern Ecuador at the time of first European contact were canoes made of bark slabs sewn together to form double-ended craft or of wood planks used similarly.
The bark canoes were built only by the Yahgan and Alacaluf, nonfarming, nomadic sea-mammal hunters and gatherers who ranged the waterways and islands of the Chilean Archipelago between Taitao Peninsula and Cape Horn.
The plank canoes, for which the aboriginal term
dalca has survived, were used by the nonfarming Chonos between Chiloe Island and Taitao Peninsula, and also by farming Indians of Chiloe and the mainland immediately to the north.
These craft have been replaced completely by modern small boats, but an adequate historical record of their appearance and frequent mention of their occurrence provide information on their design and dis­tribution.

Page 25

Historical Records—Dalcas

For the first record of dalcas we again turn to Goicueta's narrative; in the Gulf of Coronados, "canoes are made of three planks, like batiquines [small skiffs] of Flanders.
They are very light on the water, and we saw a great quantity of them..." In describing the country between the Gulf of Corcovado and Cabo Tres Montes, Goicueta wrote: "In this land live some maritime Indians who use canoes of three planks, in the same man­ner as those of the Coronados... their habitation is in the canoes."(12)
Since Goicueta used the word "tabla" (plank) in connection with the dalcas of the north, and "corteza de arbor (bark) in referring to the southern canoes, there is no doubt of the aboriginal difference in primary construc­tion material.

The first Spanish expedition (1558) to cross the Canal de Chacao and land on Chiloe did so in the dalcas of the native inhabitants. One of the conquistadores, Alonso de Gongora Marmolejo, who participated in the occupation of the island in the 1560's, described the vessels thus:

. . . piraguas made of three planks, one for the bottom and one for each side, sewn with thin cords. In the seam formed by the planks they put a split cane lengthways, and under it, above the seam, the bark of a tree which they call maque, which crushes easily as the stitches are taken. This bark forms a joint which keeps the water from entering very well. They are some 30 to 40 feet long, with a beam of one yard, and narrow at bow and stern like a weaver's shuttle. They [the Spaniards] collected fifty of these piraguas, as they are called by the Christians; they are called dalca by the Indians.(13)

Page 26

In the early 1570's, when Diego Maso de Alderete sailed into the Gulf of Ancud in a small brigantine, he was attacked by Indians in "a great number of piraguas made of planks sewn with tree bark and caulked with crushed herbs in place of caulking stuff and tar ..."
In 1578, the Spaniards induced friendly Indians in the vicinity of Valdivia to build fifty
piraguas for a raiding voyage to the Gulf of Ancud and Seno Reloncavi.
After razing an Indian village they were intercepted on their return voyage by a large group of natives in
dalcas, and a pitched naval battle ensued.(15)

From these early accounts it is evident that the
dalca was in widespread aboriginal use from the mainland country north of Chiloe Island as far south as Taitao Peninsula.

Jesuit missionary activity in Chiloe and the Chonos Archipelago began in the second decade of the seventeenth century, and missionary as well as military expeditions made use of the native boats for travel among the islands.
By the 1670's many natives of the archipelago north of Taitao Peninsula had been gathered into the
reducciones of Chiloe, and the Spaniards began to penetrate the northern fringes of Alacaluf country.
Bartolome Diez Gallardo in 1674-1675 and Antonio de Vea in 1675-1676
19 led expeditions to capture Indians in the Guaianeco Islands south of Taitao Peninsula, crossing the Isthmus of Ofqui.
Having run his ship on a rock in the treacherous Canal de Chacao, de Vea requisitioned nine
dalcas in Chiloe with which to continue his voyage.
He had their gunwales strengthened and wide thwarts added so that they could be rowed with oars in the European manner; perhaps this was the first type of alteration from the aboriginal form.
By December of 1675 de Vea had penetrated south to the Isthmus of Ofqui, where he dismantled several of the
dalcas, transported the pieces across the isthmus, and assembled them to continue the voyage.
In this he probably followed the Indian practice; in his com­pany were a number of "reformed" natives from Chiloe, some of whom had formerly lived in the Chonos Archipelago and were acquainted with this route.
At Ofqui he left a guard to intercept any Indians crossing by the portage, a further indication that the canoe people probably made common use of it.
The expedition reached the English Narrows before turning back, thus penetrating well into the aboriginal bark canoe coun­try.
In this region de Vea encountered several groups of local inhabitants, but although he mentions their
"piraguillas," there is no clue as to their

Page 27

construction material.
Such expeditions undoubtedly served to acquaint the bark canoe builders with sturdier planked craft.

In 1765, one of John Byron's officers reported having encountered some Indians in the Straits of Magellan just west of Cape Monday "who had with them a canoe of a construction very different from any that they had seen in the Straight before
[i.e., bark canoes in the eastern part]; this vessel consisted of planks sewed together: "A few days later more natives were seen at Cape Upright: "Their canoe was not of bark, but of planks sewed together.""

The Spanish surveys of the Straits of Magellan were extended in 1788— 1789 to the western part, again under the leadership of Antonio de Cordoba. Observations of Churruca and Cevallos, who saw
dalcas at Cape Upright, are quoted directly from their journals by Vargas y Ponce:

"The construction of their canoes shows some superiority of workmanship over those of the rest of the Strait.
They are not made of the weak and badly united pieces of bark, like the others, but of planks joined together by a thick cord a half inch in diameter.
The joints are treated with a compound the composition of which seems to be herbs and a certain clay, so sticky and adhesive that it prevents leaking.
Each side is composed of two very strong planks which are skillfully given the necessary curve to diminish the beam regularly towards the bow and stern; the maximum beam is amidships.
The bottom is a thick, long, narrow plank, united to the side planks in the same manner as they are to each other. The thwarts and floortimbers are the same as in the canoes of the other Indians, but stronger.
Although their design does not seem so well adapted for speed, they at least have the advantage of greater stability and strength, and are not so likely to sink from leaking as the others.
They row in the European manner, and their oars are of proper proportions, from which it can be seen that they knew the advantage of this type over the paddles used by the other Indians; that is, the necessity of a certain length to derive the greatest efficiency from their efforts. They used only a paddle as a rudder, as the oar serves the Europeans, particularly in rough seas.
It seems that to inhabit such regions where the seas and winds are usually rougher than in the rest of the Straits, it is necessary for these men to improve the construction of their canoes.
Thus the progress of human ingenuity has always followed the law of necessity.

Perhaps the young surveyors would better have credited their own nation with the introduction of certain features of this craft, for the use of oars in the European manner and the five-plank construction are cer­tainly manifestations of European influence. It will be recalled that de Vea had installed rowing thwarts and oars in his
dalcas over a century earlier. The first record of change from the aboriginal three-plank to five-

Page 28

plank construction occurs in the 1736 chronicle of Miguel de Olivares, as noted by John Cooper.(19)

By the end of the eighteenth century the
dalca had taken on so many European characteristics that aboriginal techniques of construction seem to have been preserved only in the sewing of the seams and caulking materials.
Padre Gonzalez de Agiieros (1791) described the sewn plank craft of Chiloe,
as built during the 1780's:

piraguas] are built of five or seven planks, each two to four brazas [about 12 to 24 feet] long and one-half to three-quarters of a yard wide, and two or three inches thick.
They shape them so that the ends are narrow to form the stern and bow, and apply fire to them for shaping.
To put the piragua together, uniting the planks, they drill some small holes two inches apart, and sew them with some cords which they make from solid canes called Colegues.
Thus they form a true seam as if they were sewing together two pieces of cloth.
So that this joint between the planks will not leak, they put, both inside and outside, along the seam, some crushed leaves.
Over these they pass the stitches, and with the same leaves they caulk the holes.
Thus built they resemble a normal boat, but without keel nor deck.
For strength they place inside some curved pieces for beams, secured by wooden wedges in place of nails."

In Gonzalez de Agiieros' time many natives, especially in the south of Chiloe, were still making their living by fishing, gathering shellfish, and seal hunting, for which they built
dalcas of three planks.(20)

Although more than one Spaniard complained about the fragility and intractability of these vessels, they were well adapted for exploration. They were easily disassembled for portaging, and could be repaired with a minimum of equipment.
Both in their explorations southward among the islands and in their navigation of lakes leading inland to the Andean
Cordillera, the Spaniards built piraguas as they needed them.
In accounting for materials used in building four vessels with which to cross Laguna Todos los Santos and Nahuel-huapi in the late eighteenth century, Fran­cisco Menendez listed only nails, drills, adzes, and chisels, the rest of the materials coming from the surrounding forest.

The sewn-plank vessels continued in use through colonial times and afterwards, serving as small cargo boats for inter-island trade and for transportation from point to point on Chiloe. Captain King, of the British surveying voyages, wrote of their role in coastal transport at the time of his visit in the late 1820's:

"As the only mode of supplying the town of San Carlos [Ancud] with pro­visions is by water-carriage, it is frequently ill supplied during winter, when

Page 29

N. W. winds prevent the arrival of the piraguas.
A southerly wind for two days, at that season, brings from fifty to a hundred piraguas from Dalcahue and Castro, laden with hams, potatoes, pigs, grain, fowls, calves, dried fish, and charcoal.. .

King continues with a description of the
piragua, in which it can be seen that it was becoming even more European in design.
The bow and stern resembled those of a whaleboat, and the hull was planked with three or four courses on each side.
Sewing and caulking methods, however, re­mained the same.

During their surveying cruises in the southern part of the Chilean Archipelago, King and his officers frequently noted
At Fortescue Bay,

"The canoes of these natives were very different in their construction from any we had seen to the eastward [with reference to many sightings of bark canoes in the eastern part of the Strait].
Instead of being paddled, they were pulled with oars . . . The canoes were large; at the bottom was a plank, twenty inches wide, to which were sewn the sides, in the manner of the piraguas [referring to those of Chiloe], and they were caulked with bark in a similar way.

Plank canoes were also seen at Cape Gloucester, the southwest point of Isla Carlos, where they were "rather like the Chilote piraguas, made of boards sewed together,"(
24) and among the fjords and channels between the Straits of Magellan and the Gulf of Penas.
Bynoe, cruising with Skyring in the schooner
Adelaide, recorded large dalcas in the Gulf of Trinidad, which they took at first to be whaleboats:

". . . they proved to be large plank canoes, pulled with oars . . . the size . . . was quite beyond anything hitherto noticed; they were near thirty feet in length and seven feet broad, with proportionate depth, being made of planks sewn to­gether with strips of twisted bark and rushes: the bow and stern were flat, and nearly upright.
Six round pieces of wood formed the thwarts, which were fastened to the gunwale by ropes of twisted rushes: and there were six short oars on each side.

In King's abstract of Skyring's journal of the
Adelaide cruise, we find that plank canoes were used also in the inner channels just north of the Straits.
Somewhere near the entrance to the dead-end labyrinth of channels leading in to the Ultima Esperanza region, the surveyors met some natives who "in appearance and manner... were exactly similar to the Fuegians; and by their canoes only, which were built of planks, could they be distinguished as belonging to another tribe."

were still in use in Chilotan waters during the 1850's, as recorded

Page 30

by Lt. Gilliss of the United States Navy Astronomical Expedition,(27) but during the next fifty years they were replaced by small sailing vessels and skiffs of European and North American design.
By the turn of the century, they had disappeared entirely from Chiloe, surviving for a few years more only in the remote channels between Taitao Peninsula and the Straits of Magellan.
The last record of one in use here is from 1915.

Changes in Distribution

From the historical records cited above, it is evident that significant changes in the distribution of the bark canoe and
dalca occurred since first Euro­pean contact.
Aboriginal distribution of the bark canoe conformed to the territory inhabited by the Yahgan and Alacaluf, extending from Cape Horn and the eastern Straits of Magellan to the Gulf of Penas and Taitao Peninsula.
Its range did not extend beyond the inland limits of the many fjords and inlets of the intricately embayed south Chilean coastline.
No records of bark canoes are found from the Chonos Archipelago; from Taitao Peninsula northward at least to the Gulf of Coronados, and possibly to the present site of Valdivia, the
dalca is recorded in the earliest chron­icles.

Beginning in the last quarter of the seventeenth century, such expedi­tions as those of Gallardo and de Vea introduced the
dalca in bark canoe country.
This dispersal, which continued throughout the eighteenth and nineteenth centuries, was attended by changes in form, construction, and mode of propulsion resulting from Spanish influence.

By the time of Byron's passage through the Straits of Magellan in 1765, the
dalca had penetrated southward at least as far as Cape Monday and Cape Upright in the western part.
Its occurrence at Cape Upright about a quarter of a century later, as recorded by Churruca and Cevallos, sug­gests that the plank craft had come to be preferred over bark canoes among many of the Alacaluf.

The limits of this southern and eastern dispersal seem to have been Fortescue Bay, on the Straits, and the island region between Canal Ballen-ero and Cape Gloucester, where
dalcas were reported by members of the British surveying voyages of the 1820's.
These limits approximate the zone of contact between
the Alacaluf of the western Straits and theYahgan of the Fuegian channels and islands southward to Cape Horn.
Evidence to be presented later indicates that the Yahgan continued to use bark

Page 31

canoes for almost a century after their Alacaluf neighbors had adopted the dalca.

Although by the 1820's the
dalca had penetrated well into the Straits, bark canoes were still in use here, for they were seen by King and his officers at Warrington Cove, just across from Fortescue Bay.
Further to the northwest, and well north of the
dalca's southern limit, bark canoes were being built at the native "boatyard" described by Bynoe.
The northernmost occurrence of bark canoes in what by then seems to have been predominantly
dalca country was in Messier Channel. Thus the dalca, incorporating features of European influence, extended its range considerably but did not completely replace the bark canoe among the Alacaluf, and made no inroads among the Yahgan.

It was a very different kind of vessel, the dugout canoe, that eventually replaced the bark canoe in Yahgan territory.
Introduced by missionaries based at Ushuaia, on the northern shores of Beagle Channel, in the last quarter of the nineteenth century, this new craft was apparently so rapidly accepted that by the early 1900's the bark canoe was rare enough to warrant a place in the Salesian Brothers' museum at Punta Arenas.(

The dugout canoe also spread rapidly among the Alacaluf of the western Straits, and northward into the sparsely inhabited archipelago south of Taitao Peninsula, replacing the
dalca completely by the second decade of the twentieth century.
In turn, the dugout has been replaced by planked, framed skiffs of various more or less modern designs over virtually all of the Chilean Archipelago.

The dugout had little to do with the vanishing of the
dalca from Chilotan waters.
The transition seems to have been made directly from
dalca to conventionally planked and framed skiffs and sailboats, with metal fastenings, common to the region today.
By the turn of the twentieth century, sewn boats had disappeared entirely from Chiloe, surviving for a few years more only in the remote channels between Taitao Peninsula and the Straits of Magellan.
Perhaps their last record in Chiloe is from the journal of Roberto Maldonado, who led a surveying expedition along the west coast in 1895. At Isla Metalqui he saw two double-ended
chalu-pones (small sailboats) of " suigeneris" construction, navigated by seal hunters and guano collectors from Quehui.
It seems doubtful that Mal­donado would have referred to the craft thus if they were modern skiffs or launches.

page 32

Processes of Dispersal

The close similarity between the bark canoe and the dalca is evident.
The general shape, caulking techniques and material, and the manner of joining the bark slabs or planks, are substantially the same.
The principal dif­ference, the substitution of wooden planks for bark slabs, did not constitute a radical change for the bark canoe makers, many of whom shifted easily to building the sturdier
This change, it must be remembered, re­sulted in great measure from European influence.
An important question is whether or not it followed the introduction of more effective wood­working tools.

The meager archeological record indicates that axes were mostly un­known in Alacaluf country.
The Chonos, on the other hand, had hafted axes, presumably acquired from the more technologically advanced peoples of Chiloe and the adjacent mainland.
Planks could have been worked laboriously with flints, shells, and fire, of course, and Cooper, citing Byron, Garcia and Rosales, says: "[the] Chonos and more southern Canoe Indians made their
dalcas without axes or adzes, by the use of fire, flints, and shells."(31)
It is not always clear, however, that the
dalca and not the bark canoe was the craft manufactured with these rude tools. Byron's ship, the Wager of Anson's fleet, was wrecked in 1741 on one of the Guaianeco Islands, in what was then perhaps still predominantly bark canoe country.
Garcia's mention of the tools used comes from an interview with an old man who accompanied him on an expedition from Cailin to the Guaianecos in 1766-1767.
The old Indian told Garcia that he had been born in the Chanaquelya region, on the mainland near the Gulf of Penas, and de­scribed the construction of canoes there:

"They constructed their boats using fire and shells.
They were two brazadas [about
12 feet] in length.
After the loss, a long time ago, of a ship in these parts [the
Wager?], they found some spikes which, when tapered at the end, served as tools for hewing the planks of their boats. (32)

Was it this sort of acquisition, plus trade hatchets, knives, and other iron tools, that spurred the southward dispersal of the dalca ?
In view of its his­torical spread largely as the result of European penetration, this seems likely.

Another interesting question is the meaning of the aboriginal southern limit of the dalca.
Was an aboriginal southward dispersal from a hypo-

Page 33

thetical origin in Chilote country merely given additional impetus by the Spaniards, or was the Taitao Peninsula a dividing line of long standing between dalca and bark canoe?
This involves the problem of derivation of these craft, one from the other; the question is, which came first?


The historical southward dispersal of the
dalca seems to have suggested to Graebner that a similar process involving the bark canoe took place in pre-Spanish times.
He postulated the derivation of the bark canoe from the
dalca, the influence stemming from the farming Araucanians of main­land Chile north of the archipelago.
He saw the bark canoe as a crude imitation of the
dalca, perhaps because the southern folk did not possess the woodworking skills necessary for plank construction.(33)

Friederici maintained the opposite view, deriving the
dalca from the bark canoe on the basis that the thick pieces of bark and the planks were used in the same fashion.
Assuming that sewn craft of bark were present in southern Araucanian territory before these technically advanced wood­working people arrived, the Araucanians would easily have adapted wooden planks to the design and thus created a sturdier craft.

A third alternative, independent invention and development of the two craft, is noted along with those above by Cooper, but there are so many similarities in design and construction that this notion does not warrant consideration.

Graebner's thesis has two possible implications: that the southern sea nomads had no boats previous to the introduction of the three-piece sewn boat, or that they previously possessed some other type of craft.
The first may be discounted in view of the antiquity assigned to shell middens on the islands; they almost certainly predate the Araucanian occupation of Chile.
 Archeological investigation of pre-Araucanian inhabitants of the region north of Chiloe has not progressed sufficiently for conclusions.
If the evidence that they, as well as the later Araucanians, came to temperate Chile from the east, across the Andes, should stand correct, it seems doubt­ful that displacement-type boats would have been part of their equip­ment.
The only primitive craft known for the region immediately east of the Andes here are crude hide
pelotas ("bull boats"), rafts, and reed floats. As for the second alternative, if the canoe people ever did use other

Page 3

types of vessels or floats, no record of them has as yet been found.
This, of course, does not discount the possibility.

Cooper enlarges on Friederici's thesis that the plank boat originated among the southernmost Araucanians, with derivation from the bark canoe.
He mentions in passing the "Chumash" sewn plank boat of south­ern California, but makes no attempt to relate it to the
He rightly discounts European introduction, and expresses doubts concerning Oceanic influence.
Introduction from Peru is ruled out on the grounds that the Peruvians did not have this type of vessel and Peruvian influence did not extend to southern Araucanian country.
Basing his judgment on the "powerful stimulus" of geographic conditions, the "pressure of local needs," and the intelligence and inventiveness of the Araucanians, Cooper favored the derivation of the
dalca from the bark canoe, in agreement with Friederici. He cautions, however, that this is "not strictly demonstrated."(35)

Cooper does not treat of the larger problem posed by acceptance of Friederici's view: that of the origin of the bark canoe. The final solution of the problem of derivation between the
dalca and the bark canoe must await that of a much more difficult one, the origin of the sewn displace­ment hull.

Page 35


Lack of data in early sources on design and construction of aboriginal canoes has been noted by Robert West for Colombia.(1)
This is generally true for the rest of dugout country in western South America.
There is more information on their distribution, and although many art not very precise as to place, there is enough evidence to construct a reasonably reliable picture of their occurrence at or shortly after first Spanish contact.


Nuuez de Balboa mentions briefly the Indian canoes of Darien; and Andagoya, the first Spaniard to record the forested coastal lands south of Darien, did much of his voyaging in native canoes.(2)
On Francisco Pizarro's expeditions, large canoes presumably built for him by Indians of Panama were used as shore boats, as related by Xerez: "When they thought they saw signs of habitations, they went on shore in three canoes they had with them,rowed by sixty men . . ."(3)
This refers to the Colombian coast, during Pizarro's second voyage toward Peru.

Several accounts relate Pizarro's encounter with Indians who came out from the shores of Bahia de San Mateo in northern Ecuador.
Diego de Trujillo merely says: "Many Indians came downriver in canoes to take a look at us."
Xerez, another eyewitness, makes the encounter at the Indian town of Tacamez, just south of the Rio Esmeraldas, and gives the number of canoes as fourteen.(5)
Oviedo, who obtained his information from members of the expedition, agrees with Trujillo on the location north of the river, and notes eighteen canoes.
Despite these discrepancies, the accounts agree on the large size and ornamentation of the vessels, and Oviedo notes
the use of a sail:

... there came eighteen large canoes, the biggest of them larger than the Christians had ever seen in those parts.
The bows and sterns were very large with certain structures of wood the height of a man on them.
They came under sail and paddle, filled with people with fittings of gold and silver seeks and arms and heads.
On the structure which they carried on the sterns of he canoes there were many pieces of gold."(6)

This description of sailing canoes stands with the widely quoted account of a sailing raft encountered by one of Pizarro's pilots (see p. 67 ff.) as evidence for the aboriginal mariners' knowledge of sail.

Page 36

The presence of canoes along the coast and on highland rivers of western Colombia is demonstrated by West, who consulted colonial documents that show the Spaniards' early dependence on Indian canoe builders for the means of river transport.(7)
But in none of the first contact accounts are dugout canoes mentioned south of the Esmeraldas region in northern Ecuador.
The Manabi coast was balsa-wood raft country.
When the Indians of the coast just north of the Peninsula of Santa Elena fled the Spanish advance, they took to the sea on rafts, as related by Diego de Trujillo (quoted on p. 70).
A later description of the Manabi region by Girolamc Benzoni mentions only sailing rafts, and small log rafts on which the fishermen sat astride.(8)
Garcilaso de la Vega states that the dugout canoe was unknown in Peru: "... they did not know how to or could not make piraguas or canoes like those of Florida or the Windward Islands or Tierra Firme, which are like wooden troughs, because in Peru there was no thick wood suitable for them."(9)

Canoes also seem not to have been used in the Guayas Basin, the ex­tensive lowland drained by the several large rivers and tributaries which join the Rio Guayas just north of Guayaquil.
During the first penetration of the northern part of this region by the Spaniards, in 1569, they used rafts rather than canoes in their expedition up the Rio Daule to collect Indians and transport them downriver.(10)
According to a 1572 (or 1573) relacion, the Rio Canaribamba (upper Jubones) could be navigated "for fifteen leagues upstream with balsas.
The navigation would be easier with canoes, which up to now have not been used."
The same recommendation is made for navigating the Rio Catamayo (a tributary of the modern Chira), on which only balsas had been used up to that time.(11)

There is no mention of canoes in the balsa-wood raft region until after responses to the 1577 questionnaire (later compiled as the Relationes Geograficas) were in.
The relacion for Piura records both canoes and rafts on the rivers of the Sechura Desert- presumably the Chira and the Piura.
Only rafts are mentioned for the Rio Tumbez.(12)

Thus, dugout canoes seem not to have occurred aboriginally on the coast or on rivers of Pacific drainage for a great distance south of northern Ecuador.


For form and construction materials of aboriginal dugout canoes, there is no detailed information in early accounts from the country between

Page 37

Darien and Esmeraldas.
Acosta (1590) and Cobo (1653) mention the common use of the ceiba tree (Ceiba sp.) for canoe building, without specifying particular places.(13)
Wafer observed in the 1680's that "the Indians burn the Trees hollow"; this in reference to canoe building in Panama, using the "cotton tree," without doubt a ceiba.(14)
Dampier differentiates the "red Cotton-tree" from the larger one, with wood "somewhat harder," which was also used for canoes.
Canoes of "cotton-tree" wood "will not last long, especially if not drawn ashoar often and tarred; otherwise the Worm and the Water soon rot them."(15)
Perhaps Dampier observed the application of a beeswax preservative, a practice that persists among certain Indians today (see below).

I observed no ceiba canoes during my travels; modern canoe builders consider it too soft and spongy.
Possibly ceiba was no longer used after the native builders acquired more effective tools for dealing with hardwoods, although Stevenson, writing in the 1820's, listed "ceibo" along with hard­woods used for canoe building in the vicinity of Guayaquil.(18)
An un­answered question is the capability of the pre-Conquest Indians in build­ing canoes from the harder cross-grained woods used by their modern descendants.
There is no technical reason that a well-formed and skillfully made canoe could not have been produced from hardwood by use of shell, stone, or soft metal (bronze?) tools, with the aid of fire and a measure of patience.

Whether or not the acquisition of better tools affected the form of canoes is another question, for which we have virtually no evidence from colonial sources.
In modern craft, a distinction can be made between types obviously patterned after European small boat designs and others that do not show such derivation.
The former are most common in Panama, but occur sporadically throughout the range of dugouts in northwestern South America.
The latter show little difference in general form as built by Indians and Negroes along Pacific drainage from Panama to the Gulf of Guayaquil.
The feature which serves to distinguish some is the shape of the ends, especially the design of the platforms at bow and stern.
All foreign influence on design has tended to eliminate these plat­forms, indicating that they represent an aboriginal feature.
Similar plat­forms occur on native canoes of wide distribution in the Amazon country," in contexts that do not indicate foreign influence or the use of metal tools.
Various Indian groups of Pacific coastal lands, for example the Choco,

Page 38

Waunama, and Cayapa, have a continuous history as canoe builders since Conquest times.
It seems very doubtful that the acquisition of new tools, however much sharper and more durable, would have the effect of chang­ing radically the traditional designs.
The hand tools used today, in fact, do not perform any basic cutting, scraping, smoothing or boring functions that were unknown to the aboriginal builders- they just make the job faster and easier.

Page 61

There is little doubt that simple rafts of logs lashed or pinned together have been used in western South America since very ancient times. Their aboriginal occurrence is well documented in many first-contact and early chronicles.
However, rafts were among the few types of aboriginal water-craft in the New World which were familiar to the Europeans, and consequently they seldom remarked on details of construction, except for those that bore sails.
The term used almost universally by Spaniards for rafts of wooden logs was balsa, usually without qualification as to material.
This term was also applied generally to any other type of floating device, except displacement hulled boats, but commonly with some qualification such as balsa de totora, balsa de calabazas, etc.
Thus, with careful attention to context we can often assume that the unqualified term applied to flat rafts of parallel wooden logs.

Page 64

The earliest evidence for aboriginal use of nonsailing rafts along the coast is in the form of model or toy rafts found in pre-Columbian burials in northern Chile.
They are of three logs of light wood, with the center log projecting a third of its length forward from the others.
Double-bladed paddles are found with them.(8)
Full-size counterparts of these rafts were pictured together with a sailing raft by Benzoni (pi. i6,b) in 1563, as ob­served on the Ecuadorian coast.(9)
The use of such simple craft for coastal fishing persists today, and although the tradition of building them with a center log projecting beyond the ones on the sides is lost, most other aspects of construction and use recall what must have been aboriginal practices.

On the coast of northern Peru, fishermen who cannot afford to build a sailing raft or boat, or migratory farm workers who fish part time, use small balsa-wood rafts (balsitas) without sails for fishing.
Balsitas are also much used as tenders and lifeboats by the crews of such larger vessels as sailing botes and powered fishboats. The rafts are rather crude affairs, often built partly of driftwood pieces of various lengths and diameters, but usually containing at least two good logs purchased for the construction.
The logs are joined by two or three crosspieces of hard algorrobo wood, chosen so that their curves will depress the smaller logs to form a relatively flat undersurface for floating stability.
The forward ends of the logs are commonly tapered from the underside by chopping with axe or knife to form a crude bow, while the after ends are left blunt and more or less square across.
On a central log aft, a large block of balsa wood is often lashed to form a bogadero (sculling fulcrum

Page 65

Propulsion is by a wide-bladed, tapered plank, identical to that which serves sailing raftsmen as centerboard, rudder, and paddle (pi. 14,12).
The retention of this piece of equipment for use on the sail-less rafts is curious; as a paddle it is very unwieldy and makes for slow progress through the water.

Along the north coast of Peru the best inshore fishing grounds are often near headlands or stretches of cliff-lined shore strewn with rocks. Although these rugged shores seem unapproachable to European or North American seamen, the native fishermen make use of many small coves that offer landing places.
One such region is the rocky Paita Peninsula, between the bays of Paita and Sechura.
Here the face of the tablazo drops abruptly to the edge of the sea, which in most places surges directly against the base of the sheer cliff.
Here and there, however, the cliff face has been worn back into short, precipitous ravines with steep and narrow beaches of coarse sand. Around the base of the peninsula over a dozen tiny fishing hamlets are occupied seasonally by men who fish the inshore waters.
These settlements are reached by trails from the heads of the ravines, nar­row footpaths that wind down the steep sides and often traverse loose slopes of precarious footing.

Buildings at the hamlets are simple flat-roofed shacks built of driftwood propped against pole frames stuck in the sand.
Offering some shelter from wind and sun, they serve as storage places for the meager belongings of the fishermen; and here the fish are cleaned and salted, for this isolated stretch of coast is too far from markets to allow delivery of fresh fish.

The only fishing craft observed here were the small balistas (pl. 14,b).
Well suited to the fishermen's needs, they are unsinkable, float lightly upon the water, and do not capsize easily in the surf.
The surge in the rocky coves, even on calm days, is quite treacherous, and in most places the landing of an ordinary small boat would be extremely hazardous.
The fishermen, however, launch their rafts and paddle out through the swirling waters around the rocks with a nonchalance almost incomprehensible to the uninitiated.

After negotiating the surf, the fisherman paddles out to anchor just off a rocky point, or continues his voyage out perhaps a mile or two offshore to fish in deeper water.
For the rock fishing inshore he tosses overboard a large round stone to which a short stick is lashed- a crude anchor that holds the raft in position while he casts his handline towards likely look-

Page 66

ing crevices among the rocks.
Further offshore, the craft simply drift and bob while a few handlines or short trawls are set.

If we substitute in our mind's eye a shell or bone hook for the store-bought metal one, and visualize the fishermen in slightly different garb, we may with little difficulty close out the modern world for a time and obtain a vista into the ancient past.
The fleet of small rafts, bobbing in the seas off this remote desert edge of the great continent, each with its solitary fisherman peering into the depths and feeling for a bite on his line, repre­sents a picture little changed from aboriginal times.
At intervals of a few days, however, the aura of the past is shattered abruptly by the noisy arrival of a truck from Sullana or Piura to pick up the salted fish.
One is reminded forcibly that for all its primitive appearance this fishery is di­rected toward modern commercial marketing of the catch.

Page 85
Reed Floats

The notion of tying pieces of buoyant plant material together to form a float is simple enough to have been conceived more than once; but it is also simple enough to be very ancient.
It is only necessary to gather ma­terial from any of a wide variety of sources—reeds, leaves, sticks, bark, stems—and tie it together with a vine, thong, strip of bark, fiber, or some manufactured lashing.

In comparing South American bundle floats with those of other regions, it is probably unimportant that hull forms are not identical.
Hull forms in Peru itself are as different from each other as are forms from widely separated parts of the world.
Most unprofitable would be an assumption that the Titicaca
balsas, the three-part caballitos of south-central Peru, and the two-part ones of the north, all quite different in hull design, represent independently conceived ideas of using reeds for flotation.

A construction feature common to all reed floats of western South America is the tying of separate bundles that are in turn lashed together to form the hull.
This is by no means the only possible way to use reed or similar material to build an adequate float.
Another feature in common is the turning up of one or both ends, usually to a point.
To my knowledge, upturned ends are lacking only on the modern Huanacache Lagoon vessels.
These characteristics suggest the possibility of common heritage; and such a common heritage would explain much in attempting to ac­count for their presence in South America.

Reed bundle floats and similarly shaped craft fashioned from other materials are found today or known from the past among primitive, and some not so primitive, folk of every inhabited continent.
Thompson notes an early Spanish report of Mexican reed floats at Lake Chapala, Lake Tlaxcala, and in Nayarit.
He also mentions modern reports of their occurrence in Morelos, Guerrero, and the State of Mexico.
 A task for the future is investigation of these craft for comparison with the South Ameri­can floats.
The only suggestive note in Thompson's brief remarks on these floats is that in the State of Mexico they "apparently consisted of bundles of reed tied together."
Between central Mexico and Peru there is a lengthy stretch of coastline for which we have no information on reed floats.
I would surmise that investigation of the dispersal of the dugout

Page 86

canoe here might allow some inferences as to its replacement of other craft, perhaps including reed floats.

From west-central Mexico the distribution appears to have been almost continuous, with a few interruptions where other craft were reported, through northwest Mexico, Baja California, along the Pacific coast of California, and ending inland in the Klamath country.
Reed floats have also been reported from scattered locations in the Great Basin of western United States, and they were the predominant watercraft of interior Californian lakes and rivers.(
Detailed mapping and analysis of all the re­ported occurrences is yet to be done, but a distinct west coast orientation is clear.
A sampling of the reports also indicates that the comparative criteria of composite bundle construction and tapered, upturned ends are met widely throughout this range:

"... Canoe ... about fourteen feet long and consisted of three or four bunches of bulrushes fastened together with thongs and tapering at both extremities ... [San Francisco Bay, California]"(

"... canes tied in three bundles, each part tied separately, and then all tied together, the middle section being larger than the laterals. [Bahia de San Luis Gonzaga, east coast of Baja California]."

Reed floats built by the Porno of Clear Lake, in northern California, were very similar in appearance to Lake Titicaca
A float from Pyramid Lake, Nevada, now in the U. S. National Museum, has an up­turned, tapered bow, truncated stern, and composite bundle construction.
In a description of Tiburon Island in the Gulf of California, Manuel Correa recorded Indian reed vessels as they appeared in 1750:

"[The Indians passed between the island and the mainland] in some balsas made of thin reeds disposed in three bundles, thick in the middle and thin at the ends ... tied with lashings... from five to six yards long, and about a yard and a quarter wide; from the center the width diminished proportionately towards the ends... Their oars or paddles are some wooden rods two yards long, at the ends of which are fixed two blades. The rower holds it in the middle, stroking from one side to the other."

This conforms very well to descriptions of Seri floats by modern ethnog­raphers.

From this cursory sampling, which a study in progress will augment considerably, it is apparent that similarities of construction technique and form among North and South American floats are suggestive enough so

Page 87

that an alternative hypothesis to that of independent development is worthy of consideration.

In further following this trail, and using the same comparative criteria, we may consider some similarities in the far reaches of the Pacific Ocean, as well as some scattered throughout an immense territory in the Old World.
Pioneer work on the world distribution of bundle floats of reed and other materials was done by Shinji Nishimura and Hans Suder. Nishimura notes representations of reed craft in the archeological record of ancient Egypt and Assyria, records of bundle floats from Australia and Tasmania, and modern descriptions from Korea, Seistan, and Lake Tchad, as well as from North and South America.
Brindley added to Suder's list; Hornell increased our knowledge of distributions; and Elsdon Best, and A. C. Haddon together with Hornell, included some descriptions of bundle floats in Oceania in their works on Pacific Island canoes.
A recent study of Oceanian rafts by Schori incorporates the work of the above authors (except Nishimura).'

The Old World distribution of bundle craft, as derived from the many records from ancient and modern times cited by these authors, is almost universal throughout the tropical and temperate lands.
Floats range dis-continuously from eastern Asia to southern Africa, and from Europe to Australia and Tasmania.

Descriptions extracted by Best from the early literature of New Zealand exploration show that all bundle floats for which he found records were built of separate bundles lashed together, and where the shape is indicated the float is tapered at the ends.

Three bundles of
raupo [Typha angustijolia, "cat-tail"] about eighteen feet long and two feet in diameter at the centre, but tapering towards the ex­tremities, were first constructed separately, each being tightly bound and se­cured with flax [Phormium tenax, "New Zealand flax"]; and were then fastened together so as to form a flat raft.
Another bundle similarly made was next laid along the middle of this, and secured in that position, forming a sort of keel; the hollow intervals left between the keel and sides were filled up with
raupo packed carefully and tighdy in layers, and secured with bands of flax.
The bottom of the
mo\ihi [raft or float in general; also applied to rafts of wooden logs] being thus finished, it was turned over, and two smaller bundles were laid along its outer rim, from stem to stern, for topsides, and all the vacancies within were filled up with layers of raupo tied down with flax.(8)

In Seistan, the border region between Iran and Afghanistan where the extensive Hamun-i-Hilmend spreads its shallow waters over a virtually

Page 8

flat plain, reed floats were reported by Sven Hedin in the early 1900's.
They were used for fishing and for carrying cargo and passengers; Hedin crossed the lake in one, here called

"The canoes are constructed of bundles of rushes tied together and combined into a large raft which has considerable buoyancy . . . My tutin was . . . nearly 20 feet long, and amidships, where it was broadest, it measured nearly 4 feet... all the material consisted of long dry handfuls of soft yellow rushes {tut), which were bound together in very stout bundles ... Two bunches of rushes tied along the edges ran from the bow backwards to a distance of a third of the boat's length from the stern, and served as a gunwale."

Hedin's photographs show bow and stern tapered and turned up slightly; some of the vessels have small vertical prows, apparently formed by turn­ing the tapered ends sharply upward. The tutins were built by lakeside dwellers who subsisted by fishing and herding cattle. The extensive growths of tut were used for houses, mats and other household goods, and cattle fed on the rhizomes.9 Hedin did not identify the reed, but Hornell, quoting Annandale and Hora, supplied the binomial Typha angustata.1"

In the Huleh Swamps of northern Israel and in the floodplain and delta lands of the Tigris-Euphrates system, reed bundle floats were built by people who lived by fishing and hunting wildfowl.
Bundle floats, pre­sumably of reed, are represented in ancient bas-reliefs at Nineveh, in­dicating that little change in form has taken place from ancient to modern times.
These representations, like similar ones from ancient Egypt, seem to precede any of wooden vessels.
11 In 1939, reed floats were still in regular use in Kuwait Bay, as recorded by Alan Villiers.(12)

Reed floats are also recorded from Africa, ranging from the Upper Nile and the Ethiopian Lake Tana in the east, through the interior streams and swamps of the Tchad Basin to the Ras-ed-Doura lagoons of coastal Morocco and the River Loukkos.
They also occur along the string of rift valley lakes in the east.
The southernmost record is from the Okovango Swamp and Lake Ngami district in Bechuanaland.

On the River Loukkos, the middle course of which flows through an extensive swampland, Angel Cabrera, in 1913, recorded the reed floats built by fishermen.
They were composed of several bundles of
carrizo (Phragmites), tied tightly with lashings of grass cord.
The bow was turned up and tapered, the stern square across, and sideboards, extending well aft but short of the stern, were lashed above the hull bundles. The craft were

Page 89

"el madi"—with little doubt, as Cabrera states, the origin of the Spanish "almadia."
They are among the very few reed floats on which sails were carried; if the wind was favorable, a small lateen sail was hoisted on a simple mast.

This vessel, as well as floats built by the lake dwellers of the Tchad, with high, pointed prows, side bundles lashed to the hull, and truncated sterns, demonstrates a tradition of form widespread in Africa.(14)
Brindley described briefly some floats of the White Nile in which the "forward part turns up vertically as a 'beak-head'
... it has 'sides' formed by small additional bundles superimposed on the edges . . . The after-end of the papyrus raft is cut square, but there seems to be usually some narrowing in width towards this.
Similar rafts are used in the Lower Congo and other Central African regions."
Floats made of both reed and
ambatch (Herminiera elaphroxylon) in the Okovango Swamp and Lake Ngami region, and of papyrus on Lake Tana, show a most striking similarity in form to those of Lake Titicaca.(18)

The foregoing examples will suffice to demonstrate the magnitude of the problem of explaining this distribution.
Although the criteria are admittedly tentative, pending more detailed information, we cannot ignore the data suggesting affinities by the gratuitous assumption that in each case the similarities of construction and design are chance parallels.
In this distribution, reed craft are built and used by people, mostly nonagricultural, who make their livings in similar ways.
The similarity of their watercraft may be only a first clue to significant affinities in ways of life among waterside-dwelling people, regardless of the particular regions or ethnic groups. The widespread distribution of bundle floats suggests great antiq­uity for this mode of water transport.
They represent but one aspect of a way of life that probably extends back in time far beyond any record of civilization.
The Egyptian and Assyrian archeological records are late in the story of need and knowledge of watercraft.
One can conjecture that these simple bundle floats were used by Old World waterside dwellers before man had extended his shoreline living to the New World.
In this view, the New World reed floats simply represent variations on a very ancient theme.

Page 90

Sewn Bark Canoe and Dalca

There are past and modern records of sewn-plank vessels from southern California, Oceania, southern Asia, and Africa; and the technique of sewing planks to dugout hulls or keels has an equally wide distribution, including the Pacific Northwest of North America.
The stripping of bark for canoe construction is known sporadically from eastern tropical South America to the North American subarctic region, and in the Old World from Australia and Africa.
Some authors have suggested affinities among Oceanian or Asian sewn-plank canoes and
dalcas, based on hull form, but convincing similarities have not as yet been demonstrated. Another

Page 91

difficulty is imposed by the great distances, both land and oceanic, that separate them.

There are also difficulties in a thesis of local South American invention, especially as attributed to the Alacaluf or Yahgan.
Despite the general crudeness of their construction, the ideas involved in manufacture of the bark canoes embody a curious mixture of rudimentary and advanced techniques.
From what we know of Alacaluf and Yahgan material, they possessed in aboriginal times only the simplest tools: pointed flints for cutting and punching (or drilling?), and stone and shell scrapers—crude equipment with which to fashion a boat.
Bark is certainly a rudimentary building material, available even to people incapable of felling large trees.
But other, more sophisticated elements of construction, such as the stitching of the sides to the bottom, caulking the seams, installing internal stiffeners analogous to frames and stringers, headng of bark and wood for ease in bending, and the whole idea of a displacement-type hull, seem inconsistent with the otherwise sparse talent for handicraft found among these people.
Other elements of their material culture support disbelief in their capacity to invent such a vessel after isolation in their island habitat.

Although I have objected to Graebner's thesis of derivation of the bark canoe from the dalca partly on the grounds that there is no trace of boat or float types other than the bark canoe among the aboriginal islanders, this concept cannot be ruled out entirely.
The rapid substitution of other craft for their canoes in historical dmes is well documented, indicating an inclination to accept new ideas, at least in watercraft.
Thus the pos­sibility that their ancestors arrived among their islands with some other kind of vessel is by no means remote.
But no known possessor of sewn bark craft anywhere near southern South America can be indicated as an outside source for the complicated construction techniques.

Perhaps a profitable approach is to focus attention on techniques of assembly and basic ideas employed in obtaining, treating, and applying the materials for construction among the Chilean craft and their putative counterparts in other parts of the world.
If investigation in other regions should reveal identifiable complexes of these techniques and ideas which can be related to the Chilean craft, perhaps we can avoid the subjectivity of hull form comparisons.
In these terms the
dalca and the sewn bark canoe go together; the question of which came first is submerged in the larger one of the origins and dispersals of sewn-hulled vessels.
This, to-

Page 92

gether with further archeological research in the many shellmounds on both coasts of the continent, will no doubt increase our knowledge of the roamings of coastal fishermen, sea-hunters, and gatherers, and a trail to the southern islands may some day emerge.

Page 94

Log Rafts

The design of raft hulls has received proper attention by J. G. Nelson (28) as suggesting pre-Columbian affinities among South American rafts and those of other parts of the world.
The tradition of tapering one or both ends, or using odd-numbered logs with the center log the longest, is wide­spread.
This feature can no longer be identified in modern rafts of western South America, but it was common from Conquest times to the late nineteenth century.
By Juan and Ulloa's time, the design incorporating a single longer center log seems to have been in vogue,(28) but a century later Francois Paris showed a raft tapered at both ends.
His other illustrations, however, show untapered rafts, as does an earlier picture of a Guayas raft by von Humboldt.(27)

Another widespread tradition is the holding together of softwood logs by sharpened hardwood pins instead of by lashings.
The only instance I saw in western South America was on the small banana rafts of the upper Rio Cojimies in Ecuador.
The balsa-wood logs are held together by slender crosspieces, but the lashings, instead of passing around the logs, are tied to hardwood pins driven diagonally into the upper surfaces of the logs.
The pins bear on the crosspieces, gripping them against the logs, and the lashings merely reinforce their hold.
That this may be a survival of an old practice and not a local inspiration of Cojimies raftsmen is suggested by Wafer's seventeenth-century description of Panamanian rafts:

". . . they take Logs of this Wood ['Light-wood,' probably balsa-wood] not very big, and bind them together collaterally with Maho-Cords, making of them a kind of Floor. Then they lay another Range of Logs across these, at some distance from each other, and peg them down to the former with long Pins of Macaw-wood; and the Wood of the Float is so soft, and tenacious withal, that it easily gives admittance to the Peg upon driving, and closes fast about it.(28)

Dampier, in his description of sailing rafts quoted in Chapter VI, also mentions the pinning of logs for raft construction.
Several trans-Andean occurrences have been noted by modern ethnologists:

"The Leco [of the upper Beni River and tributaries in northern Bolivia] . . . descend the river on rafts made of light, corky balsa, pinned together with palm spikes.
Three of these rafts bound together with stout cross logs tied with strips of bark or vine form a type of craft called callapo."(29

"The Mosetene [neighbors of the Leco] travel only on rafts . . . made of seven logs of palo de balsa, a very light wood, nailed together with chonta spikes [chontaduro palm, Bactris (syn. Guilielma) sp.] and provided with a platform

Page 95

to keep goods dry.
The long central logs consisted of two trunks laid end to end.(

The Campa [of the central Peruvian montana] make pointed balsa rafts held together with chonta nails and cross-beams.31

In 1720, during his stay at "Puerto Segura" in Baja California ("about two leagues to the North-eastward of Cape St. Lucas, which is the South­ernmost land of California"), Shelvocke observed small rafts used by the natives for fishing:

. . . they go out to sea on their bark-logs, which are only composed of five logs of a light wood, made fast to one another by wooden pegs; on these they venture out rowing with a double paddle .. ."(32)

Among the craft described by Hornell as incorporating the pinned fastening technique, are rafts from India, Australia, Melanesia, New Zealand, Brazil, and Africa.(33)
Many of these, like South American examples noted above, combine pinning with the tapered, odd-log shape.
These features in combination invite inquiry as examples of possibly very old practices.

Nelson's theory of Old World origin for rafts exhibiting tapered form and odd-log construction was applied to sailing and nonsailing rafts alike.
His concentration on "defining the hull or main body of the raft as rigidly as possible" for comparative purposes led him to hypothesize separate origins for "large freighter rafts" with non-tapered, square ends, and "finger or organ-shaped" (tapered) rafts.
His classification of Ecuadorian and Peruvian rafts into these two types was based on that of Emilio Estrada;
(31) but Estrada classified them according to their appearance and use in colonial, not aboriginal times.
His "large freight" raft type was based solely on Dampier's late seventeenth-century description of large sailing rafts used for voyages from Peru and Ecuador to Panama.
Nelson seems to have inferred the separation into two basic types from Estrada's rendering of Dampier's account, in which Estrada separated the description into two parts, one dealing with small rafts and the other with large.
When these descriptions are read in proper sequence, however, they state clearly that the large rafts incorporated the same tapered design as the smaller ones (see pp. 71-73 for the full quotation from Dampier's journal).
Thus the tradition of tapered bow, or of installing a center log longer than the others, applies to all sailing rafts of Ecuador for which we have adequate early descriptions.

Page 96

With little doubt this tradition preceded the introduction of sails and centerboards to western South America.
Theoretically, sails and center-boards were superimposed on the traditional raft design when the idea of their use in combination was introduced.

Sailing rafts with centerboards occur elsewhere in South America, and also overseas. Brazilian coastal fishermen, plying the waters between Bahia and Ceara, build small sailing rafts of light wood, called jangadas.
They are used for offshore fishing in much the same manner as the Sechura craft. Hull construction combines pinning and lashing, and the logs are often tapered at the bow.
A slender, curving, flexible mast is stepped on deck and provided with a tabernacle in the form of an open box, much like that described for Peruvian rafts by Amasa Delano.
The position of the mast can be varied by inserting the heel in any of several holes in the plank which forms the top of the box.

The jangada sail is triangular, with diagonally sewn panels, and the luff is lashed directly to the mast.
Although it is not a true lateen rig, it is referred to as
vela latina. It bears a strong resemblance to the sails pictured by Madox and Spilbergen.
The Brazilian sails are loose-footed, with the clew extended by a light boom which is attached to the mast by a yoke.

Jangadas carry a single centerboard, thrust between the logs just aft of the mast, and a wide steering oar.
Unfortunately, the record is not quite so clear on the method of sailing as it is for the Peruvian and Ecuadorian rafts.
Bill Burk, a missionary of the Amazon country who made a fishing trip on a
jangada out of Fortaleza, Ceara, provides the following descrip­tion of steering: As they got under way, the "Skipper was aft working into place and up and down the fixed-direction rudder ..." Later, he refers to the same motion: "... [he] worked his seven-foot rudder up and down in its slot . . ." As they prepared to anchor on the fishing grounds, the helmsman "paddled us into position with the rudder now free of its socket," and when the day's fishing was done they "set sail with the rudder and centerboard also in their opposite slots."(36)

The function of the "fixed direction rudder" was apparently to change the sailing balance and thus the course by raising and lowering.
Such interaction between the stern board or oar and centerboard would produce the same results as on the Peruvian rafts.

Hornell describes the jangada steering oar as having a long and broad

Page 97

blade, serving "both for steerage and to reduce leeway."(37)
Alves Camara provides a detailed description of
jangadas used on the coast from Bahia to Ceara in the late 1800's; the rudder was an oar with a long blade, which was kept in one of the slots at the after ends of the logs while sailing to windward, but was taken out when the wind was astern.
The centerboard was adjusted for the particular point of sailing: ". . . conforming to the force of the wind, and the course which they follow in relation to it, they lower and raise them, or pull them out when running before the wind; they also carry them vertically or inclined."
Other descriptions and illustrations show wide-bladed steering oars or paddles inserted between the logs at the stern,
(39) but only Burk indicates the possibility that their function is the same as that of the stern board of ancient and modern west coast rafts.
If it has this function, rather than serving as a conventional steering oar which swivels from side to side, levering on the slot, it might indicate a common origin for the Brazilian and Peruvian-Ecuadorian centerboard rafts.
The possibility is strengthened by the similarity of
jangada sails to those of aboriginal western South America.

However, an opposing opinion must be considered. In a monograph on the life and customs of the Brazilian jangadeiros, Luis da Camara Cascudo states that neither the sail nor the centerboard were aboriginal in Brazil, but were introduced for use on the previously sailless rafts after European contact, probably after the end of the sixteenth century.
He cites various historical sources, including early Portuguese narratives and geographical descriptions which, although mentioning rafts and other aboriginal craft, make no mention of sails or centerboards.
From this negative evidence he concludes that sailing techniques were learned from the Europeans.
40 Until further research clarifies this story, we can only point out that the Portuguese in Brazil, like the Spaniards in the west, were ignorant of the function of centerboards, and did not possess the triangular sail with mast stepped on deck, nor the technique of moving the mast and sail thwartships.
Thus, the possibility that the native raft sailors learned their art from them seems remote; but for the present the matter must be left open.

A number of writers have discussed relationships among rafts of eastern Asia and South America.
The combination of sail, of whatever design, and centerboards, is thought to represent a navigation technique too com­plicated to have been invented more than once.
In considering affinities

Page 98

between Asian and American craft, much attention has been focussed on rafts that have persisted to modern times in Formosa. According to Ling, however, these Formosan craft "are built by fishermen, whose ancestors came from Southern Fukien about 300 years ago."(41)
On the China coast replacement of rafts by displacement vessels of wide variety has blanked out the older picture; perhaps a few rafts still exist, but information is lacking.
Sailing rafts do occur, however, in areas peripheral to the coast of modern China, in Korea and Annam (now Vietnam).
They also persist on the east coast of India and on Ceylon.

These Asian rafts carry sail types characteristic of other sailing craft in their respective regions.
The Formosan and Korean types carry typical matting or canvas lugsails, as do sampans and junks of the China Sea.
The Annam vessels have lugsails rigged with a steeply dipping yard and canted boom.
On the Coromandel coast of India the lateen sail with truncated tack, common to canoes and other displacement craft of these waters, is carried.
However, the Formosan, Annamese and Indian rafts have in common the use of centerboards (information lacking for Korea).

Among the several types of sailing rafts he observed on the Coromandel coast, Hornell mentions specifically the use of centerboards for only one, the "kola maram" or "flying-fish catamaran" of the Tanjore District north of Point Calimere:

"But though of such shallow draft and without a keel, she is able to beat against a wind fairly well, for when this is necessary two powerful leeboards are brought into action, one abreast the forward mast (two are carried), the other at the stern nearly abreast of the steering paddle, which of itself functions as an efficient leeboard and is of the same shape.
When close-hauled we may justly say that the craft is employing three leeboards, a curious and significant fact seeing that the great sailing rafts of Formosa employ the same number.

The two masts are short, raked forward, and stepped laterally on the outside log on whichever happens to be the leeward side."(42)

In the Thanh Hoa region of northern Annam, Charles Robequain noted sailing rafts along the coast between the Song Koi River and Cape Boung-Quioua.
One of his illustrations (plate xxv, a) of a two-masted bamboo raft shows clearly a thin centerboard inserted between the logs just aft of the after mast.

Detailed descriptions were later included in extensive studies of An­namese maritime activities by French ethnographers.
The following observations combine those of J. Y. Claeys, Pierre Paris, and J. B. Pietri.

Page 99

The rafts are equipped with two, rarely three, high-peaked lugsails with yards and booms.
By means of crosspieces with several holes, the masts may be shifted from one side to the other, in much the same manner as the masts of Brazilian
Forward, midships, and stern centerboards are carried.
A wide stern board may be used as a "sliding rudder"
(gou-vernail coulissant), but steering is usually with a stern sweep on the port side, which is also used for sculling.

With the three centerboards shoved all the way down, a two-masted raft carries too much weather helm, so the forward centerboard is shoved only halfway down.
To luff, the stern board is raised.
When only the midship (or forward, in the two-mast rig) sail is in use, the raft is balanced with the forward and midship centerboards.
When only the after sail is used, the midship and stern boards are carried."

It is apparent that the Annamese raft sailors know the principle of balance by means of interaction of centerboards and sails. There is the suggestion that the stern board, or "sliding rudder," is used occasionally as the sole steering device.
Does this represent an ancient practice which has been almost completely replaced by steering with a stern sweep?

Sailing balance and course changes also seem to be accomplished by movement and adjustment of centerboards on the Formosan rafts, although in published descriptions the relationship between the centerboards and stern sweep oars in maneuvering is not entirely clear.
According to Ling, six centerboard positions are provided by slots in the raft's hull, but ap­parently only three boards are employed for any given point of sailing.
There are slots on each side of the bow and stern, and two in the center-line amidships.
Normally, the after centerline board and either port or starboard bow and stern boards are used.
On the starboard tack, the port or leeward bow and stern boards are submerged; on the port tack they are raised and the starboard boards are brought into play, a practice that appears analogous to the shifting of the centerboards to the leeward side in tacking or jibing a Sechura raft, or to the shifting of rudder and center-board on Burk's
In tacking a Formosan raft, the stern board is raised completely, and the centerline boards are raised part way.
The full length of the forward board is used when sailing close to the wind.(
The rafts are usually equipped with several sweep oars, including one or two at the stern for steering and sculling.
As with the Annamese craft, there

Page 100

is the suggestion that the sweep oar may be superimposed on a more ancient system of steering by centerboard.

Thus the feature common to all of these Asian and American rafts, besides the use of sails and the generalized raft form of the hull, is the use of centerboards to prevent leeway while sailing close to the wind and reaching.
Although the rafts differ in construction material and details of form, the centerboards and their use indicate the possibility of common heritage.
This hypothesis, however, is in need of further substantiation by detailed and knowledgable observation of the manner in which the centerboards of Asian and Brazilian rafts are employed.
If further in­vestigation should disclose real similarity of centerboard navigation tech­niques, the case for trans-Pacific dispersal would be very strong.
The com­plex of raft hull, sail, centerboards, and the sophisticated techniques of maintaining sailing balance, preventing leeway, and maneuvering and steering by changing the relationship of lateral plane to center of effort, represents a set of ideas much more complicated than many that are generally accepted as evidence of dispersal rather than fortuitous parallel developments.

If these affinities were to be established convincingly, there would still remain the problem of determining a common time base for the prehis­toric existence of rafts and the centerboard navigation technique in Asia and in South America.
Nishimura and Ling provide some evidence for early voyaging in Asian waters and agree that raft navigation probably preceded the development of displacement vessels in the Orient.
An ex­tensive search in Chinese literature led Ling to the conclusion that sailing rafts were present in China in the fifth century
b.c, and he presents leg­endary material suggesting their existence as far back as the thirty-third century b.c.(46)
Further investigation is needed to determine the nature and extent of early raft-borne commerce, and especially to determine a time level for the occurrence of centerboards and the centerboard navigation technique.
As yet we have no data for India or Annam, nor for Brazil.

This does not mean that we can dismiss out of hand hypotheses of trans­oceanic raft voyaging.
It must be admitted that the capacity of balsa and other light woods to float long enough for extensive voyaging has been sufficiently argued and amply demonstrated by Heyerdahl in his
Kon-Tiki voyage.
Accounts by experienced seafaring men quoted herein record favorable impressions of the sailing qualities, maneuverability, and sea-

Page 101

worthiness of sailing rafts.
South American craft regularly made long voyages carrying large cargoes, both with and against the set of the current and the prevailing wind.
The ability of ancient canoe sailors to traverse the great waters must also be conceded to raftsmen.
The argument that lengthy raft voyages were impossible is no longer acceptable; the ques­tions now are: who were their navigators; what were the routes; and what role did the rafts play in the dispersal of ideas?

1. For discussions of float representations in archeological material see Philip A. Means, "Pre-Spanish Navigation off the Andean Coast," The American Neptune, II (1942), 115-119; Max Schmidt, Kunst und Kultur von Peru (Berlin, 1929), pp. 88-90; Thor Heyerdahl, American Indians in the Pacific (Rand McNally, 1953), pp. 586-591.
For Huaca Prieta materials see T. W. Whitaker and Junius B. Bird, "Identification and Significance of the Cucurbit Materials from Huaca Prieta, Peru," American Museum Novitates, No. 1426, 1949.

2. "Relacion del Sitio del Cuzco y Principio de las Guerras Civiles del Peru . . . 1535 a t539." in Coleccion de Libros Espanoles Raros 6 Curiosos, XIII (Madrid, 1879), p. 179.

3. "Primera Parte de los Comentarios Reales de los Incas," in Biblioteca de Autores Espanoles, CXXXIII (Madrid, i960), pp. 107-108.

4. Historia Natural y Moral de las Indias . . . (Madrid, 1894), I, 235-236.

5.  "Historia del Nuevo Mundo," in Biblioteca de Autores Espanoles, XCII (Madrid, 1956), pp. 265-266.

6. Thirty-six Voyages to Various Parts of the World . . . between . .. 7799 aT>d I^4I (3d. ed., New York, 1858), pp. 333-334-

7. Historical and Descriptive Narrative of Twenty Years' Residence in South America (Lon­don, 1829), II, 18.

8. W. S. W. Ruschenberger (who signed his book "By an Officer of the United States Navy"), Three Years in the Pacific, including notices of Brazil, Chile, Bolivia and Peru . . . (Philadel­phia, 1834), pp. 379, 384.

9. J. M. Gilliss, The U. S. Naval Astronomical Expedition to the Southern Hemisphere . . . Vol. I, Chile (33d Congress, 1st Session, Executive Document No. 121) (Washington, 1855), p. 426.

10. Notas de Viajes para su obra "El Peru" (Lima, 1942), I, 186-187.

11. R. E. Coker, "The Fisheries and Guano Industry of Peru," U. S. Bureau of Fisheries, Document No. 663 (Washington, 1910); Erwin Schweigger, Pesqueria y Oceanografia del Peru, Compania Administradora del Guano (Lima, 1943).


1. The Pacific Lowlands of Colombia (Louisiana State University Studies. Series, No. 8, Baton Rouge, 1957), p. 252, fn. 114.

2. Vasco Nunez de Balboa, "Letter from Balboa to the King, dated Darien," in Narrative of the Proceedings of Pedrarias Ddvila . . . by the de Andagoya, transl. and ed. by Clements R. Markham (Publications or' tie XXXIV, London, 1865), p. x; Pascual de Andagoya, Narrative . . . , passim.

3 Francisco de Xerez, "Verdadera Relation de la Conquista del Peru y Provincia del Cuzco . ..," in Biblioteca de Autores Espanoles, XXVI (Madrid, 1947), p. 321.

4. "Relation del Descubrimiento del Reyno del Peru," in Publicaciones de la Escuela de Estudios Hispano-Americanos de Sevilla, Serie 7a., Num. 4 (1948), p. 45. (Written in 1571.)

5. "La Relation Samano-Xerez," in Raul Porras Barrenechea, Las Relaciones Primitivas de la Conquista del Peru (Paris, 1937), pp. 66-67. For discussion of the authorship of this relation see pp. 20-21.

6. Gonzalo Fernandez de Oviedo y Valdes, Historia General y Natural de las Indias, lslas y Tierra-firme del Mar Oceano . . . , ed. by Jose Amador de los Rios (Madrid, 1851-1855), IV, 122.

7. The Pacific Lowlands of Colombia, pp. 191, 251, fn. 109.

8. Historia del Mondo Nuovo (2d. Italian ed., Venice, 1572), pp. 164-165. (First publ. 1563.)

9. "Primera Parte de los Comentarios Reales de los Incas," in Biblioteca de Autores Espanoles, CXXXIII (Madrid, 1960), p. 107.

10. Martin de Carranza, "Relation de las Provincias de las Esmeraldas . . . ," in Relaciones Geogrdficas de Indias, Peru, III (Madrid, 1897), p. cxxxvi.

11   "Relation y Description de la Ciudad de Loxa," in Relaciones Geogrdficas de Indias, Peru, III, p. 201.

12. "Relation de la Ciudad de Sant Miguel de Piura," in Relaciones Geogrdficas de Indias, Peru, II (Madrid, 1885), pp. 229, 230.

13. Jose de Acosta, Historia Natural y Moral de las Indias . . . (Madrid, 1894), I, 407; Bernabe Cobo, "Historia del Nuevo Mundo," in Biblioteca de Autores Espanoles, XCII (Madrid, 1956), p. 264.

14. Lionel Wafer, A New Voyage and Description of the Isthmus of America (Publications of the Hakluyt Society, Series 2, LXXIII, Oxford, 1933), p. 52.

15. William Dampier, A New Voyage Round the World (5th ed., London, 1703), I, 214.


8. Junius B. Bird, "Excavations in Northern Chile," Anthropological Papers (American Museum of Natural History), XXXVIII (1943), 224, 227 et passim; Thor Heyerdahl, Amer­ican Indians in the Pacific (Rand McNally, 1953), pp. 553-555-

9. Girolamo Benzoni, Historia del Mondo Nuovo (2d. Italian ed., Venice, 1572).


1. J. Eric S. Thompson, "Canoes and Navigation of the Maya and Their Neighbours," Journal of the Royal Anthropological Institute of Great Britain and Ireland, LXXXIX (1951), 74.
His early Spanish source is A. de Ciudad Real,
Relacion Breve y Verdadera de Algunas Cosas de las muchas que sucedieron al Padre Fray Alonso Ponce en las provincias de la Nueva Espana . . . (Madrid, 1873).
Ponce traveled in western Mexico during the 1580's. Comments on modern occurrences derive from Thompson's correspondence with Roberto Weitlaner.

2. Georg Friederici, Die Schiffahrt der Indianer (Studien und Forschungen zur Menschen-und Vdlkerkunde, I, Stuttgart, 1907), pp. 17-21;
 A. L. Kroeber,
Handbook of the Indians of California (Smithsonian Institution, Bureau of American Ethnology, Bulletin 78, Wash­ington, 1925). PP- 243, 277. 3io. 329-330, 359. 4i6» 468, 531. 630, 652, 723, 739, 813;
James Hornell,
Water Transport (Cambridge 1946), pp. 44-46;
Robert F. Heizer and William C. Massey, "Aboriginal Navigation off the Coasts of Upper and Baja California,"
Anthro­pological Papers, No. 39 (Smithsonian Institution, Bureau of American Ethnology, Bulletin 151, 1952), pp. 291-296.

3. Heizer and Massey, p. 292, quoting Archibald Menzies, "Menzies' California Journal," introd. and notes by Alice Eastwood, California Historical Society, Quarterly, II (1924).

4. From the account of Francisco de Ulloa, a sixteenth century explorer of the Gulf of California, as quoted by Heizer and Massey, p. 296, from Henry R. Wagner's translation in Spanish Voyages to the Northwest Coast of America (San Francisco, 1929).

5. Kroeber, p. 243. The similarity is noted by Bill Durham, Canoes and Kaya\s of Western America (Seattle, i960), p. 98; Durham, however, regards this similarity as "interesting, if meaningless."

6. "Descripcion de la Isla de Tiburon . . . ano de 1750," Archivo General de la Nation (Mexico), Provincias Internas, 176, f. 115. (Photocopy, Department of Geography, University of California, Berkeley.)

7. Shinji Nishimura, A Study of Ancient Ships of Japan, Part I: Floats (Tokyo, 1936); Hans Suder, Vom Einbaum und Floss sum Sckiff: Die Primitiven Wasserfahrzeuge (Veroffent-lichungen des Instituts fur Meereskunde, Universitat Berlin, Neue Folge, B, Heft 7, Berlin, 1930);
H. H. Brindley, "The Sailing Balsa of Lake Titicaca and other Reed-Bundle Craft," The Mariner's Mirror, XVII (1931);
James Hornell, Water Transport (Cambridge, 1946);
Elsdon Best, "The Maori Canoe," Dominion Museum Bulletin (Wellington), No. 7 (1925);
A. C. Haddon and James Hornell, Canoes of Oceania (Bernice P. Bishop Museum Special Publications, Nos. 27^ 28, and 29, Honolulu, 1936-1938);
Dieter Schori, Das Floss in Ozeanien (Volkerkundliche Beitrage zur Ozeanistik, I, Gottingen, 1959).

8. Best, p. 139, quoting E. Shortland, The Southern Districts of New Zealand (London, 1851).

9. Overland to India (London, 1910), II, 263—272.

10.  Water Transport, pp. 58-59;
N. Annandale and S. L. Hora, "The Fishes of Seistan,"
Records of the Indian Museum, XVIII (1920), 193.

11. Hornell, Water Transport, pp. 56-57;
his source for the Huleh Swamp occurrence is
J. Macgregor, Rob Roy on the Jordan (London, 1869);
for Nineveh, Sir A. H. Layard,
Discoveries in the Ruins of Nineveh and Babylon (London, 1853); for the Tigris-Euphrates region,
P. A. Buxton and V. H. W. Dowson, "The Marsh Arabs of Lower Mesopotamia,"
Indian Antiquary, I (1922).

12. Monsoon Seas (McGraw-Hill, 1952), p. 98.

13. "Balsa de Juncos en el Bajo Lucus," Revista del Instituto de Antropologia de la Univer-sidad Nacional de Tucuman, I (1938), 39-41.

14. E. van Konijnenberg, Shipbuilding from its Beginnings, Brussels, n.d. (ca. 1905?), II, 3, photo of Lake Tchad papyrus float.

15. H. H. Brindley, "The Sailing Balsa of Lake Titicaca and other Reed-Bundle Craft," The Mariner's Mirror, XVII (1931), 14.

16. Hornell, Water Transport, pp. 53, 55.

17. Shinji Nishimura, A Study of Ancient Ships of Japan, Part I: Floats (Tokyo, 1936);
James Hornell, Water Transport, pp. 20-34.

For an east Asian example, see W. Robert Moore, "Raft Life on the Hwang Ho," The National Geographic Magazine, LXI (1932), 743—752.

18. "Canoes and Navigation of the Maya .. . ," p. 73.

20.  Francisco de Aguero, "Discripcion de Zapotitlan, Tuscacuesco y Cusalapa (1579);" in Noticias Varias de Nueva Galicia, Intendencia de Guadalajara (Guadalajara, 1878), pp. 295-296;
Isabel Kelly, The Archaeology of the Autlan-Tuxcacuesco Area of Jalisco. II: The
Tuxcacuesco-Zapotitldn Zone (University of California Publications: Ibero-Americana, 27, Berkeley, 1949), pp. 27-28.

21. Roberto Weitlaner, "Chilacachapa y Tetelcingo," El Mexico Antiguo, V (1941), 284.

22. Thompson, p. 73, from personal communication with Roberto Weitlaner.

Pedro R. Hendrichs Perez, Por Tierras Ignotas: Viajes y Observaciones en la Region del Rio de las Balsas (Mexico, 1945), I, 40-64.

24. Nishimura, pp. 12-69; map, facing p. 156.

25"The Geography of the Balsa," The American Neptune, XXI (1961), 157-195.

26. Juan and Ulloa's original drawing (pi. 17) shows the center log projecting at both bow and stern. In the several erroneous copies made for various foreign editions, the projecting center log is not shown.

27. Paris, Essai sur la Construction Navale des Peuples extra-Europeens . . . (Paris, 1841-1843), plate 129;
Alexander von Humboldt,
Vues des Cordilleres, et Monuments des Peuples Indigenes de I'Amerique (Paris, 1810);
both are reproduced in Thor Heyerdahl,
American Indians in the Pacific (Rand McNally, 1953), plates Ixviii (Paris) and xxxiii (Humboldt).

28. Lionel Wafer, A New Voyage and Description of the Isthmus of America (Publications of the Hakluyt Society, Series 2, LXXIII; Oxford, 1933), p. 59.

29. Alfred Metraux, "Tribes of the Eastern Slopes of the Bolivian Andes," Handbook of South American Indians (Smithsonian Institution, Bureau of American Ethnology, Bulletin 143, III, Washington, 1948), p. 505.

30. Ibid,, p. 494.

31. Julian Steward and Alfred Metraux, "Tribes of the Peruvian and Ecuadorian Montana," Handbook of South American Indians, III, 544.

32.  George Shelvocke, A Voyage Around the World (London, 1928), p. 226, drawing on p. 224.

33.  Water Transport; India (Ganjam, Orissa, Vizagapatam, Bengal, Assam, Tanjore), pp.67-68; Australia (Northwest coast, from north end of Ninety-Mile Beach to Bathhurst Island),p. 72; Melanesia (Buka, Bougainville, and the Nissans in the northern Solomon Islands; Tevai, near Vanikoro in the Santa Cruz Islands), pp. 74-75; New Zealand (east coast),
p. 79; Brazil, p. 82; Africa (Lobitos Bay, Angola), p. 85.

34.  "Balsa and Dugout Navigation in Ecuador," The American Neptune, XV (1955), 142-147.

35. Hornell, Water Transport, pp. 82-85;
R. H. Lane-Poole, "Primitive Craft and Medieval
Rigs in South America," The Mariner's Mirror, XXVI (1940), 333-338;
Luis da Camara
Cascudo, Jangada—Uma Pesquisa Etnogrdfica (Rio de Janeiro, 1957).

36.  "Voyage on a Jangada," Andean Air Mail and Peruvian Times, XX, No. 1046 (December 30, i960), 12-13.

37.  Water Transport, p. 85.

88. Antonio Alves Camara, Ensaio sobre as Construccoes Navaes Indigenas do Brasil (Biblio-theca Pedagogica Brasileira, Serie 5a., XCII), Companhia Editora Nacional (1937), pp. 22-44. (First publ. 1888.)

39Nineteenth century descriptions and illustrations in Henry Koster, Travels in Brazil (2d.ed., London, 1817), I, 4;
James Henderson,
A History of the Brazil . . . (London, 1821), pp.357-358, 366;
Daniel P. Kidder,
Sketches of Residence and Travels in Brazil (Philadelphia, 1845), pp. 176-180; and
Franz Keller,
The Amazon and Madeira Rivers (Philadelphia, 1875), pp. xi, 36.

40Jangada—Uma Pesquisa Etnogrdfica (Rio de Janeiro, 1957).

41.  Ling Shun-Sheng, "Formosan Sea-going Raft and its Origin in Ancient China," Bulletin
of the Institute of Ethnology (Academia Sinica, Taiwan), I (1946), 26.

42.  Water Transport, p. 66.

43. Le Thanh Hoa. Etude Geographique d'une Province Annamite (Paris and Brussels, 1929), II, 398-399; map, p. 389.

44. Pierre Paris, "Recherche de Parentis a Quatre Embarcations d'Indochine," Bulletins et Travaux, Institut Indochinois pour l'Etude de l'Homme, II (1939), 209-220;
Esquisse d'une Ethnographie Navale des Peuples Annamites
(Publicaties van het Museum voor Land- en Volkenkunde en het Maritiem Museum "Prins Hendrik," Rotterdam, 1955), pp. 59-64;
Y-Claeys "Les Radeaux de Peche de Luong-nhiem (Thanh Hoa) en Bambous Flottants," Bulletins et Travaux, Institut Indochinois..., V (1942, fasc. 1);
"L'Annamite et la Mer,"
ibid., pp. 17-28;
J. B. Pietri,
Voiliers d'Indochine (Nouvelle Edition, Saigon, 1949), pp. 11, 89-91. (1st ed. 1943-)

45. Ling Shun-Sheng, p. 26.

46. Ibid.,
p. 51.