Volume 54 1945 > Volume 54, No. 4 > Was there pre-Columbian contact between the peoples of Oceania and South America?, by James Hornell, p 167-191
                                                                                             Previous | Next   

- 167

THIS is a question that has given rise to much controversy. Unfortunately, some of the protagonists favouring the affirmative have shown notable partisanship, with the result that their arguments are often vitiated by overstatement and the too easy acceptance of unconfirmable legends of doubtful reliability, particularly in respect to place-names and routes. In the following examination of the different facets of the problem, I shall endeavour to maintain an objective attitude.

The subject will be considered under the following heads and sub-heads:—

1. The sailing craft available to the Polynesians for oversea voyaging;

2. The character of the prevailing drifts and currents in the Pacific ocean;

3. The evidence afforded by the geographical distribution of plant-products of economic value—mostly food plants;

Among these are:—

  • (1) The Yams (Dioscorea spp.);
  • (2) Taro (Colcasia antiquorum, Schott; the Arum esculentum of Linnaeus);
  • (3) The Coconut-palm (Cocos nucifera, L.);
  • (4) Hibiscus tiliaceus, L., and, most important of all—
  • (5) The Sweet-potato (Ipomoea batatas, L.).

4. Consideration of the Rival Hypotheses explaining the introduction of the Sweet-potato into Oceania:—

  • (1) By the agency of Polynesian voyagers;
  • (2) The alleged Spanish introduction;
  • (3) By drift voyages of Peruvians in sailing balsas.

5. The distribution of the Sweet-potato as ascertained between 1722 and 1825 in Polynesia;

6. Summary and Conclusions.


The first matter of importance to be settled is whether the Polynesians at any time possessed vessels large enough and weatherly enough to make them suitable for undertaking such a lengthy voyage as that separating the American mainland from the islands upon the eastern fringe of Polynesia.

- 168

So far as we are able to judge there is no difficulty in coming to a definite and entirely favourable conclusion. That long exploring-voyages were frequent during an early period in the colonizing phase of their history is made certain by the presence of people belonging unmistakably to the Polynesian race in all the islands scattered throughout the whole vast extent of ocean stretching eastward from Tonga and Samoa to the lonely outlier of Easter island (Rapanui), and southward from Hawaii to New Zealand—an area measuring over 3,600 miles in length and breadth. In every island within this region the people speak some form or dialect of the Polynesian language, and have numerous characteristic ideas, beliefs, and customs in common.

Apart from the plain fact that the ancestors of the present-day Polynesian inhabitants of all these islands must have used, of necessity, some form of water-transport to reach their new homes, abundant evidence exists in native oral records, fully confirmed by cross references, that adventurous voyages, covering distances up to 2,400 miles, were not uncommon during the period of maximum maritime activity which characterized Polynesian history from about the 10th to the 14th century. The most frequent and the best authenticated of these long voyages were those between Hawaii and Tahiti. Other voyages of about the same length were made between Tahiti and New Zealand, about 2,400 miles apart; even lonely Rapanui, looking across a waste of waters toward Chile, was reached, probably more than once, by intrepid adventurers, brave to the point of recklessness, who had originally set sail from the Society islands.

Few of these long voyages were made direct; whenever possible, halts to revictual the vessels and to refresh the crews were made at intermediate islands. Going toward Rapanui, Mangareva served as a half-way house, with a further halt at Rapa if desired. Sailing toward New Zealand, all canoes called at Raratonga, while a call at the Marquesas islands or at Fanning island served as a welcome break during the long voyage to or from Hawaii.

Protracted voyages of another kind were those desperate ventures which resulted from the expulsion from the homeland of one section of an island's population when vanquished in battle by a rival tribe or clan. Many times, storms or starvation brough death to the wanderers, but instances are recorded of eventual settlement on some far-distant island, either uninhabited, or peopled by kind-hearted folk who gave sanctuary to the hapless refugees, or by others who welcomed the newcomers as a fortunate reinforcement in their own struggle with aggressive rivals; others, less fortunate, met with a hostile reception and were destroyed or enslaved.

Some of the more important discoveries were probably made by crews caught by violent adverse winds; blown off their course, with no option but to run before the gale, these canoes, sailing in a direction never contemplated, occasionally were wrecked or brought up safely at some unknown and distant island. Within the Melanesian orbit we find here and there an isolated community of people speaking some variety of the Polynesian tongue; their ancestors could not conceivably have arrived there otherwise than by means of a drift voyage of this nature. As Elsdon Best (1923) has given typical instances of such occurrences, it is needless to recapitulate them here.

When Europeans cruised through Polynesia in the 18th and 19th centuries, the islanders of most of the important groups were still in possession of fine, sea-going vessels, capable of keeping the sea even in - 169 stormy weather. Two types were in use—the sailing outrigger-canoe and the double-canoe. The former is, in some respects, the handier to manoeuvre, and by far the more speedy; but even that exceptionally large vessel, the tongiaki, seen by Tasman (1898) at Namuku in the Tongan group, in 1643, lacked cargo space sufficient for prolonged voyaging. In this particular the double-canoe is markedly superior; besides the additional hull-space afforded by the duplication of the hull, a large deck-space is available whereon to build a cabin to serve both as a shelter for the crew and to house gear and perhaps to afford accommodation for some of the domestic animals necessary for the stocking of the new home if the vessel were to bring up at some uninhabited island. Double-canoes of large carrying-capacity were possessed by the people of every principal island group. Each group had its own distinctive local type before European contacts caused them to be replaced by small schooners which, in turn, are forsaking sail for the oil-driven motor.

The double-canoes of the Tuamotus were perhaps the finest ever built in Polynesia, for the dimensions of the twin hulls were not limited by the length and diameter of the tree trunks available as were those of other groups, because the Tuamotuan hulls were built up of strakes of planking, sewn together like those of Persian and Arab craft in the years before Portuguese, Dutch, and English ship-building example caused the disappearance of this primitive method of uniting planks (Hornell, 1936, 79 et seq.).

Next in carrying-capacity were the grand double-canoes of the Hawaiians, built of hulls hewn from mighty trunks of Oregon pine washed ashore by the north-west trade-wind or brought by the combined drifts of the Californian coastal current and the North Equatorial current, with assistance from the north-east trade-wind. This flotsam from North America was eagerly sought for, and any exceptionally fine tree-trunk was hoarded by the local chief for incorporation in a specially large double-canoe. Vancouver (1798, 2, 219) records that the chief Taio kept one gigantic trunk unworked for a long time in the hope of finding another to match it. Disappointed in this, eventually, and with great reluctance, he utilized it for the hull of one of the largest outrigger-canoes ever made in the islands.

In the account of Cook's third voyage (1784, 3, 148), the size of the largest double-canoe of Owhyhee (Hawaii) as actually measured, is given as 70 feet long, 12 feet broad, and 3½ feet deep; the breadth is the overall width of the double structure. There is, however, a record of the wreck of a double-canoe on the coast of Hawaii which had the extraordinary length of 108 feet, and a certain double-canoe of Oahu of the 18th century is reputed to have carried from 120 to 140 men (Best, 1925, 235).

Other fine double-canoes were possessed by the Tahitians when Cook visited these islands. From the records of old-time voyages which radiated from this group in every direction—to Hawaii, New Zealand, and Easter island—it is clear that Tahiti at one period produced numerous intrepid navigators and discoverers. During Cook's stay there, the high priest Tupaea was able to enumerate and give sailing directions for nearly all the islands in the Austral and Cook archipelagoes, for many in the Tuamotus, and for others farther afield, including Rotuma and some which appear to belong to Fiji; among the number enumerated were a number not at that time known to Europeans; several were subsequently found and identified with islands in Tupaea's list.

- 170

Whether schools of navigation were maintained in the Society islands as was the case in the Gilberts and the Carolines is not capable of proof, but the fact that the high priest of Tahiti possessed sufficient nautical knowledge to be able to fix the relative geographical positions of many islands, suggests that the Tahitian priesthood, as in Egypt, were the repositories of all the scientific knowledge of their time, and would be, in consequence, the educational authority in matters nautical. All this, taken in conjunction with our knowledge that the Tahitians were the most enterprizing and successful of Oceanic navigators—the discoverers of New Zealand and Easter island (Rapanui)—during the heroic age of Polynesian maritime enterprize, makes it seem extremely probable that some form of nautical instruction was available comparable with what we know existed in the Gilberts and Marshall islands (Grimble, 1924 and 1931), and the Carolines (Winkler, 1901, 495), down to a fairly recent period. In the schools in the islands named, the course of instruction was varied and extensive, covering every aspect of navigation within the limits of the scientific knowledge of the instructors. The pupils were taught to navigate by the stars, to shape definite courses for particular islands, to use as directive factors the run of the waves, the indications given by clouds of certain shapes, the local vagaries of the wind, and the recognition-signs indicating the proximity of land when the stars were obscured by cloud or storm. The seasonal flights of birds were also to be made use of, particularly on voyages between Tahiti and Hawaii, together with all available information concerning possible halting-places on the various routes. Polynesian and Micronesian navigation was no chance groping for an objective; it had a considerable body of scientific knowledge in its make-up, although the only charts they possessed were at best but a nework of bamboo splints indicating roughly and approximately the relative positions of a limited number of islands, with indications of the oceanic major currents and the local eddy-currents usually present at any given season.

A wise precaution taken on long voyages was for several vessels to sail in company, strung out in line abrest over a considerable distance. In setting a course for any particular island, especially if it were of small size, standing solitary and out of sight of any other island, the chance of missing a landfall was in this way minimized so far as lay within the power of the pilots.

Another safeguard provided for the future was the shipment of some domestic animals and of supplies of roots and seeds aboard any vessel proceeding on a long voyage, whether to a definite destination or on one of exploration. Even one or two women were sometimes carried in order to ensure the building up of a permanent population if an uninhabited island were found, suitable for colonization, or, alternatively, if the vessel had the misfortune to be wrecked thereon.


Before entering upon consideration of the alternative ways whereby certain plants of economic value may have been introduced into Polynesia and Melanesia, it becomes necessary to have a clear understanding of the bearing upon this problem of the normal interaction of the major drifts and currents present in the Pacific ocean. Knowledge of the use to which the ocean currents may be put by navigators is also needed in order to throw light upon the feasibility or otherwise of Polynesian voyagers being able, not only to reach the - 171 South American coast, but also to be able to set a correct course for home and to accomplish the return passage successfully.

It is unnecessary here to detail the causes which have as their effect the circulation of the surface water within this area. The first point we have to notice is that this surface-circulation is reversed on the opposite sides of the equator. In the north the circulation is clockwise; after sweeping down the west coast of the United States, the cold California current veers to the westward, south of the twentieth parallel, to form the North Equatorial drift.

South of the equator, the circulation runs in a counter-clockwise manner; running northwards along the west coast of South America, a cold current, the Peru (or Humboldt) current turns north-westward off the coast of Peru and Ecuador to form the South Equatorial drift. This continues due west between the equator and the parallel of 15° south; as it approaches the eastern Polynesian island groups, it begins to branch irregularly. Off the Marquesas islands the southernmost branch begins to diverge from the main stream, tending increasingly to follow a south-westerly course. This takes it through the northern Tuamotus and thence through the Society islands and the Cook group, where the obstruction caused by the many island-groups hereabouts, causes its divergence toward the south-east. Eventually it is joined by the Australian current; then, under the influence of the westerly winds prevalent south of the parallel of 40° south, it turns eastward to run straight across toward Patagonia. Arrived off that coast, one branch rounds the Horn into the Atlantic, while the other or northern branch, after being reinforced by an upwelling of cold water, returns up the west coast of South America as the cold Peru current, with which we began this tale of the elliptical course of the surface-waters of the South Pacific.


Several writers (notably Friederici, Rivet, and O. F. Cook) have claimed that the presence of certain plants of primary economic importance to the people of Polynesia and Melanesia, have had their distributional centre in Peru and Ecuador, whence they have been carried by human agency to the Oceanic islands far to the westward. These plants comprize the yam (Dioscorea alata, L.), taro (Colocasia antiquorum, Schott), the sweet-potato (Ipomoea batatas, L.), the coconut-palm (Cocos nucifera, L.), together with the irregularly branching Hibiscus tiliaceus, L., a tree of varied utilities in Polynesia, known there as the fau or hau.

Leaving for later consideration the problem of the origin of the sweet-potato, there is now adequate reason to believe that all the other plants originated either in south-east Asia, or in Indonesia, or else are of pantropic distribution.


Summarizing the position regarding the species Dioscorea alata, cultivated in Polynesia, there can be no reasonable doubt that it has spread eastward from its original home in Indonesia. The yams cultivated in America, China, and Japan are all of other and different species, with one exception; in Guiana, D. alata is found in cultivation, together with other species; this is due, according to Sagot (1871, 305), to a modern introduction from Indonesia and [or] Polynesia.

- 172


This plant grows wild in Indonesia and India according to authoritative botanists, as quoted by De Candolle (1886, 73), Miquel (De Candolle, l.c. 80) gives tallus, tallas, tales and taloes as Malay terms in use for it in Sumatra, words philologically related to the taro of Polynesia and the dalo of Fiji. Rivet (1928, 587-9) while ignoring the implication of affinity thus indicated, emphasizes the close resemblance of certain Melanesian names for the yam—ubi in New Caledonia, Sumatra, and Madagascar, as well as a long list of other related terms in use in Melanesia and Indonesia—with the ape of the Marquesas and Tahiti and with the kape of Tonga, names given in these islands to a species of taro, the Arum macrorhizum of Linnaeus and the Alocasia macrorhiza of Schott, which De Candolle says is sometimes cultivated in place of the true taro. According to Thwaites (as quoted by De Candolle, 1886, p. 75), this alternative species is found wild in Ceylon; as it has several unrelated names applied to it in Indonesia, in addition to those already given, we may assume that it was cultivated there at a very early period, and that it was a food-plant likely to be included among the sea-stores shipped aboard vessels which had to be provisioned for a long voyage. It may well be that both the tallus species and the ubi species were carried thus when the migrants sailed eastward into the Pacific; both appear to have been long established in Indonesia before the migration began, though one or other may have originated in Ceylon or in Indo-China.


The chief protagonist of the theory that the coconut, so characteristic of the strand-flora of Oceania and the Indian ocean, is of American origin is O. F. Cook (1901 and 1910). Various writers, especially Beccari (1917), and Merrill (1920), both possessed of intimate acquaintance with this palm in its home lands, have marshalled overwhelming reasons in disproof; there can now be little or no doubt that the coconut originated in Indonesia.

Cook attempts to show that the coconut does not occur wild in nature, and that it is seldom found growing within reach of the spray. Both these statements are recognized as mistaken by anyone who has travelled among the Indo-Polynesian islands. They know that on uninhabited islands there, and throughout Polynesia, the coconut-palm is to be found fringing the shores of the most isolated islands, islets, and reefs, provided that they be high enough to be free from frequent submersion and have some soil and humus in which to root; they have often seen these graceful palms thriving on the spray-drenched strand, their bulbous butts washed by the waves when the monsoon winds force the breakers over the protecting reef-flat. Under natural conditions this palm does not thrive in moist uplands; it must be in sight and hearing of the sea.

Cook also argues that wave-action is inimical to the casting ashore of floating nuts. This, to those who have lived on any tropic coast, is utterly opposed to fact. His argument is that any nuts falling from trees growing on the strand which fall into the sea, will be immediately thrown back upon their own shore. “High waves or tides, instead of floating shore debris away”, he says, “merely carry it further inland, as everybody knows”. If this were always the case, it would be impossible to account for the regeneration, without human intervention, of the coconut-element in the littoral flora of volcanic islands, such as happened on Krakatoa after the tremendous eruption which destroyed - 173 the entire vegetation of the island in 1833 (see also Guppy, 1906, and Morris, 1918).

Argument favouring America as this palm's original home rests largely upon the statement of botanists that a number of other species of the genus Cocos are found wild on that continent, the majority being endemic in Brazil. Unfortunately for this hypothesis, Cocos nucifera is not one of the number, which is given as eleven (De Candolle, l.c., 432). Further evidence that the coconut-palm is not indigenous in the New World is furnished by the fact that it was not till several years after the first voyage of Columbus that its cultivation was introduced into the Caribbean islands; Acosta (1598, 178) notes the coconut growing in Porto Rico in the sixteenth century. These introductions were due to Spaniards who had seen it growing on the Pacific coast; Dampier also records seeing it growing on Cocos island in the early years of the eighteenth century, uninhabited then as it remains to-day, apart from the sporadic visits of treasure-hunters during the present century. To what extent it was cultivated on the coast of the mainland is uncertain, although Oviedo (1851-1855) writing in 1526, does state that it was plentiful on the Pacific coast of Mexico in the lands under the jurisdiction of the Indian cacique Chiman, and, somewhat later, Pizarro and the other conquistadores met with it on the coast as they were going towards Peru.

Historical evidence in Asia is also antagonistic to Cook's hypothesis of an American origin. In Ceylon we have the authority of the Mahawamsa, the ancient Sinhalese record which goes back well nigh two thousand years, for the statement that King Aggrabodhi established a coconut-plantation, thirty-six miles in extent, in the south of Ceylon in the sixth century of our era. This was probably near Welligama, where a vihara contains a memorial to this king. In the twelfth century mention is also made of a similar benefaction by King Parakrama Bahu.

There is, however, a much earlier reference to the tree in the fabulous story of the contest between Dutthagamini and Elala (161 B.C.), when the warriors of the opposing kings armed themselves respectively with trunks of the coconut and the Palmyra palm, obviously an allegorical allusion, for these palms are the characteristic species that dominate severally the lives of the Sinhalese and the Tamils. The great antiquity of the cultivation of the coconut-palm is further attested by the occurrence of several different terms for it in Sanscrit.

The ordinary names for the coconut-palm and its nut in South India point decisively to Ceylon as the locality whence came the first coconut-cultivation to India; the nut is termed ten kai in Tamil and the tree itself tenna-maram 2 and as ten in Tamil is the term for “south”, when used as an adjective the two terms signify respectively the “southern tree' and the “southern-nut” or the “nut from the south”. The use of “south” as pointing to relation to Ceylon is due to the fact that Ceylon is reckoned in Tamil literature as lying to the south of India; this implication is further confirmed by the knowledge that the caste concerned with the coconut-industry in Malabar is termed Tiyyan, believed to be a corruption of Tivan, “the people from the island (tivu)”, i.e. “islanders”, an appropriate name for immigrants who came from Ceylon, bringing with them the knowledge of coconut-cultivation.

- 174

Ceylon is most unlikely to have been the original home of the coconut-palm. Trade-relations between Ceylon and Sumatra date back to the beginiing of the Christian era; unless nuts reached Ceylon by means of current action, it is reasonable to infer that the coconut was imported into the island at about the same time as the people we now know as Polynesians were moving outward from the Indonesian archipelago to seek new homes where they could live free from the inroads of warlike people out of Further India. The majority, we know, went eastward, but a considerable number voyaged to the westward, some to settle in Ceylon and perhaps in southern India, but the larger contingent to voyage onward to Madagascar—whether coast-wise or direct we cannot tell. Among the stores with which these migrants provisioned their sea-going vessels, fitted with double out-riggers, we know positively that coconuts were considered essential, for they provided both food and drink in an extremity when all other food was consumed and water was exhaused. We may be sure that these migrants would follow the same custom as that of their Polynesian descendants in later centuries—that of stocking their vessels with many plants and tubers in order to provide for the raising of new crops should they be compelled to make a prolonged stay in islands bare of useful cultivatable plants. So it was that besides coconuts, the voyaging canoes of Polynesians were stocked with young breadfruit-plants, tubers of the sweet-potato, taro, and yam, seeds of gourds, and even seedlings of Broussonetia papyrifera, the plant from which tapa was made. Pigs, dogs, and fowls were sometimes taken in addition, if available.

To anyone conversant with the ecology of the coconut-palm, the conditions of weather and currents generally prevalent in the Pacific ocean, and with the migrations and voyaging-history of the Polynesians so far as they are known, there can be little hesitation in coming to the denfiite decision that the coconut-palm is a species originating in or native to Indonesia. Without this amounting to a certainty, it is scarcely less inevitable that we must conclude that, as a direct consequence of the adventurous wanderings of the early Polynesians, particularly between the tenth and fourteenth centuries—the period of greatest maritime activity in the history of this people—the taro and the yam were distributed throughout Polynesia, from Hawaii to Tahiti, and from the Melanesian islands to Easter island.

HIBISCUS TILIACEUS, a strand-living tree characterized by the tangled growth of its low, spreading branches, has also been claimed by enthusiastic Americanists as having been spread in prehistoric times by human agency from America to the islands and shores of the Pacific and Indian oceans (O. F. Cook, 1918).

This is a tree almost as typical of littoral floras as is the coconut-palm. Like it, it bears seeds specially adapted for wide distribution by drift and current action; their protective outer coat is smooth and impervious, and it is reputed to float and retain its vitality during months of immersion. This quality dispenses at once with any idea of “human agency” having had any major share as the distributive factor which Cook claims for it. Merrill (1920) severely criticizes this view, pointing out that the argument “is largely based on the similarity between its local names in tropical America and in Polynesia, namely, maho, mahagua, etc., in tropical America, and mao, mau, vau, etc., in Polynesia”.

- 175

Unfortunately for Cook's reasoning, the many names used in Indonesia have been ignored. Many of these are given by Merrill, who reaches the conclusion that bago or baru found in the Indonesian series are the early terms from which the Polynesian names have been derived and that they belong to the mao series and not to that of the American maho and mahagua. The apparent resemblance between the mao and maho series is considered by Merrill's philological advisers in the Philippines to be superficial and therefore merely a coincidence or accidental parallelism.

This view is corroborated by a consideration of the distribution of the plant. This has a range throughout the entire coast-line of tropical Asia, Africa, and Australia, as well as being a typical strand-tree in Malaya and many islands in the Pacific.

Merrill concludes: “There can scarcely be any arguments as to other than its natural pantropic distribution, and claims to the contrary would appear not to be in conformity with the known facts regarding its occurrence and distribution in nature” (l.c. p. 380).

THE SWEET-POTATO, Ipomoea batatas, L.

The fact which appears on first consideration to be dominant when we attempt to find a solution to the problems of whether America gave this plant to Polynesia, and whether distribution was intentional or fortuitous, consists in the close approximation of many of its dialectal names in Polynesia with those in use by the northern Quichua-speaking people of Peru and Ecuador; in this region the local dialect knows the sweet-potato under the names kumar, komal or kumal, which have as counterparts in Polynesia, the terms kumara in New Zealand (Maori), Rarotonga, the Tuamotus, Mangareva and Easter island; umara and umaa in Tahiti; kumala in Tonga, 'umala in Samoa, uala and uwala in Hawaii, with kumaa in the Marquesas and ku'a'ra in Mangaia.

This practical identity of name has aroused much controversy, one group of writers arguing that Polynesian voyagers after reaching the Peruvian coast, presumably after a halt at Easter island, were able to return home safely, carrying with them the tubers and their Quichua designation.

Another group, of which the most important and cogent exponent is Dr. Georg Friederici, has sought to offer as a solution of this problem, a plausible and less difficult alternative explanation. This puts forward the view that Mendana and De Quiros, the earliest explorers of Polynesia and Melanesia, carried aboard their ships when leaving America, a supply of sweet-potatoes, other tubers and some seeds, with the specific intention of planting them in any islands considered suitable for occupation and settlement. The groups where planting is presumed to be specifically indicated are the Marquesas islands, the Solomons, the Santa Cruz group and the New Hebrides.

A third theory suggests that the transport of the sweet-potato from South America to Oceania was effected by means of the great sailing-balsas which were still in use when Pizarro and his followers passed down the coast from Panama to Peru in the early sixteenth century. This hypothesis relies to a great extent upon the story current at that time in Peru that the Inca, Tupac Yapanqi, had sailed far to the west-ward with a large fleet of sailing-balsas, and that he had returned after a period variously given as nine months or a year with “black” prisoners, precious metals, and other spoils.

- 176
A.—Importation from Peru by Polynesian Voyagers.

Two of the strongest objections to acceptance of this hypothesis consist in (a) the great distance to be traversed on a voyage from Polynesia to Peru—aboue 2,200 nautical miles—by the shortest route, which is from Easter island, the last Polynesian outpost on the east, and (b) the difficulties that would confront any voyagers who might succeed in making the crossing when the time came to select a course that would enable them to attempt the return trip with any fair chance of achieving success.

With the knowledge we possess of some of the long voyages successfully undertaken by Polynesians between the tenth and fourteenth centuries, as for example that from Tahiti to Hawaii, a distance of about 2,400 miles, and that from Central Polynesia to Easter island, the voyage to the South American coast would be well within the power of adventurers manning a well-found and well-provisioned double-canoe of adequate size. If the voyagers went viâ Easter island, they would experience no unsurmountable difficulties in reaching the Peruvian coast if their supply of food and water did not run short. Striking east or north-east from Easter island, they would be bound to come within the influence of the Peru current sweeping north, parallel with the Chilian and Peruvian coasts, and they would find it a most helpful factor in making a landfall on the coast of northern Peru or of Ecuador.

Great and well-nigh insuperable difficulties would surround the successful accomplishment of any attempt to return home. Several vitally important initial difficulties would have to be overcome; the active goodwill of the Peruvian authorities and people would have to be secured—an uncertain quantity. Even more difficult would be the determination of what course to set on leaving the Peruvian coast. To return by the same route as that by which they had come would be to invite disaster. As Dixon says (1934, 174) “it is one thing to sail eastward and strike a continent, and quite a different one to sail thence westward and find again a single island”. In this particular case, if the return were to be made viâ Easter island, the voyagers would experience added difficulty by having to battle with the Peru current which would now hinder them instead of favouring as it did on the outward voyage; the strong westerly winds generally prevalent in the South Pacific would also impede return by this route.

It is obvious that a different return-route would be essential for success if the journey were to be made in a vessel of clumsy build with its rig ill adapted to battle with head winds and contrary currents. And how were Polynesian sailors to become possessed of such knowledge? Out of their own experience they could not obtain it, but it may have been received from some of the Peruvian balsa-sailors who may be credited with having had sufficient acquaintance with the run of the local oceanic current to be able to indicate to the strangers that the northward-running Peru current veers toward the north-west in the latitude of northern Peru. Also, it may well be that the balsa-sailors had some notion, actual or legendary, of the existence of the Galapagos islands; if so, any reference to them as lying far out to sea in a north-west direction would be liable to give the Polynesian visitors the mistaken impression that the reference was to the existence of some of their own Polynesian islands, marooned far out in the ocean as is Easter island. As a natural consequence this information would - 177 tend to sway their decision in favour of a return route on a course identical with the run of the current indicated by their Peruvian friends. Their own specific knowledge of the far-stretching are of islands thickly strewn along the eastern fringe of Polynesia would further re-inforce the decision to set a north-westerly course; their most experienced navigator may have considered that the chance of sighting one of these islands was sufficiently promising to justify the hazard to be taken.

If this assumption was acted upon, and if the run of the current indicated—the South Equatorial current—was followed, we know from our present knowledge of the direction of the currents of the Pacific ocean, that such a course would take them through the Marquesas islands, a probability that will be demonstrated on a later page from the experience of the fully authenticated drift-voyages of derelicts and shipwrecked crews in this region, within the present century.

If the returning voyagers had any sweet-potatoes left uneaten on arrival at some Marquesan island, they are likely to have been given to the Marquesan people as a return for their hospitality—though a friendly reception is, again, a doubtful quantity.

If the wanderers had the misfortune to miss one of the Marquesan islands, they would have a good chance of sighting one of the islands in the northern Tuamotus or one or other of the mountainous islands, visible from afar, of the Society islands, if they had determination enough to continue to follow the run of the current which prevails in this region.

Such a sequence of happenings as I have outlined may well be admitted as a possibility for men of the adventurous spirit of the ancient Polynesians, wise in their knowledge of the many successful endings to other hazarous enterprises; but we can hardly go so far as to contend that it would have any considerable probability of achievement. All the same if, as we know for certain, the Polynesians succeeded in going backwards and forwards between Hawaii and Tahiti, and in finding New Zealand and may other isolated islands without charts or previous information concerning their presence or position, the difficulties besetting the success of a return voyage from Peru to Polynesia tend to dimish greatly. Even so, to make an island landfall in the Pacific without chart, nautical instruments, or sailing direction, is a gambler's throw with the dice loaded against the voyager; examples of this are seen in the paucity of island-landfalls in the early voyages across the Pacific, notably that of Mendaña before reaching the Solomons and that of Magellan on his way to the Philippines.

B.—The Alleged Spanish Introduction of the Sweet-potato.

Friederici (1929 and 1936) is the most important of those who have sought to consider the allusions to the planting of various vegetable-products by early Spanish explorers of the Pacific in the Marquesas islands, and in certain of the Melanesian islands, as offering a reasonable solution to the problem of whence and how the sweet-potato reached Oceania.

Unfortunately for this argument, there is not a single reference to the sweet-potato in any of these accounts; as a consequence the supporters of this theory fall back upon “probabilities”.

- 178

The facts so far as we know them are these:—

1. Marquesas islands.

In 1595, Mendaña discovered the Marquesas islands; he stayed about a week at the island of Tahuata. There he is recorded to have sown maize (De Quiros, 1, 23). No mention is made of any other product having been sown or planted, and it is rash for Friederici to assume that the sweet-potato was among the problematical “others” which he infers were introduced, seeing that he has no better reasons to offer than (1) the important food-value of this tuber in the economy of the Peruvian Indians, (2) the customary practice of the Spanish to introduce useful plants into whatever new lands they occupied or intended to colonize, together with (3) the absence of any mention of its prior occurrence in the islands when discovered, coupled with the fact that (4) its presence in the Marquesas is recorded as early as 1790 by Marchand (1798-1800, 1, 71).

The only plants mentioned by Mendaña's chronicler as noticed on the Marquesas islands are the coconut-palm, giant breadfruit trees, the graceful plantain or banana, plots of feathery sugar-cane, and a few others. All without exception are of Indonesian or of Asiatic origin and conspicuous either because of their large size or from some peculiarity of habit.

Strangly enough, although we have Marchand's definite statement of the presence of the sweet-potato in the Marquesas in 1790, neither Cook, nor Forster, nor Porter nor any one else who visited these islands toward the end of the eighteenth century and during the opening decades of the nineteenth, makes reference to its occurrence there; this suggests that it was not of sufficient importance to arrest the attention of the casual observer.

2. The Solomon islands.

These islands were discovered in 1568 by Mendanña in the course of his first voyage westward. Again there lacks any mention of the sweet-potato having been seen anywhere in cultivation. Neither is there any record of seeds having been sown nor of the planting of new products, although it is probable that something of this kind was done in view of the prolonged stay of several months made by the expedition whilst its members were busily occupied in the usual intensive quest for gold, characteristic of these early Spanish adventurers.

3. The Santa Cruz group.

The initial exploring success of Mendaña's first voyage led to a second, during which this group was reached in 1595. Here, if anywhere, we ought to hear of the planting of the sweet-potato, for the Spaniards made an unsuccessful attempt to establish a settlement on one of the islands. Some planting we may be sure was done, but, once more, no record of this is to be found in the account given by his pilot, De Quiros, neither is there mention of this plant having been noticed growing ashore. De Quiros does, indeed, state that two or three roots rather like the ordinary potato (Solanum) were in use by the natives; it is significant that he avoids employing the common Peruvian term for the sweet-potato for any of those roots which he noticed under cultivation. This could not have arisen from a want of knowledge, for the Spaniards were as familiar with the growing appearance of the sweet-potato as we are of the ordinary Solanum potato.

- 179

4. The New Hebrides.

The records of De Quiros' discovery of these islands in 1606, furnish the only definite historical evidence favouring the pre-Columbian presence of the sweet-potato in Oceania. De Leza (De Quiros, 1904, 2, 376) and Torquemada (1723, 5, cap. 69) in their respective accounts of the expedition, record the occurrence there of this plant, the former by the Mexican name camote, and the latter by the Antillean name of batata, both terms in use in America to distinguish the sweet-potato from the true potato, Solanum tuberosum. This definite identification gives fairly conclusive evidence that the sweet-potato was under cultivation in Melanesia before the first known contact with voyagers from South America. We are, therefore, justified in concluding that the sweet-potato must have been in cultivation in Melanesia in pre-Columbian times.

C.—The Hypothesis that the Sweet-potato was Introduced into Oceania by People from South America in pre-Columbian Times.

This hypothesis was suggested, but not elaborated, by Dixon (1932, 40) as the only possible alternative if we are to eliminate the Spaniards as the introducers of the sweet-potato into Polynesia and Melanesia, and if we refuse to believe in the possibility of Polynesian voyagers sailing to the Peruvian coast in pre-Columbian times and returning thence with tubers which were planted in their island home, with subsequent diffusion throughout Oceania.

Little attention has been bestowed upon this idea, due (1) partly to the well-known lack of maritime enterprise characterizing the Peruvian coastal tribes during the Inca period, and (2) partly to the belief that balsa-logs (of which ancient Peruvian vessels were built) remain buoyant for a very limited time when immersed in water. Too much importance has, in my belief, been given to these objections.

Regarding the supposed lack of maritime initiative and skill of Peruvian sailors in pre-Columbian times, we should assume a cautious attitude. We know that the Spanish conquistadors when sailing south from Panama, were greatly surprized to meet great rafts, rigged with large square-sails, sailing along the coast. Juan de Samanos (1842) writing in 1526, described one captured by Bartolomeo Ruiz, Pizarro's pilot, as big enough to have a burden of about thirty tons; it was built up in two tiers, the lower of thick and buoyant balsa-logs, the upper, which formed a kind of platform, formed of bundles of slender canes, upon which the passengers and cargo were placed to keep them dry from the water that swirled up through the lower tier of logs. “It carried masts and yards of very fine timber and cotton sails of proportionate size after the manner of our own ships and very good rigging of hennequen, which I say is like hemp, and some mooring stones [killicks] for anchors”. The mast was a bipod structure such as is used on Burmese rice-boats, and closely related to the masts of Indonesian vessels of the present day.

Juan and Ulloa who visited Ecuador in 1736 state (1760, 1, 189-92) that these sailing balsas made voyages from Guayaquil, carrying cargo to Puna, Tumbez, and even to Payta in northern Peru. These authors go on to say that “the greatest peculiarity of this floating vehicle is that it sails, tacks, and works as well in contrary winds as ships with a keel, and makes very little leeway”. This ability was due to the use of a series of boards let down vertically through slots between the - 180 logs; these functioned as leeboards and centreboards; it is said that the Indian sailors were so adept at manipulating these boards that a correct course could be kept under all but the most exceptional circumstances; these “leeboards” were fitted both toward the head and the stern; five or six might be in use at the same time.

In vessels of this kind, provided the logs did not become sodden and waterlogged, quite lengthy voyages would have been feasible, even to the islands of Polynesia. Ample supplies of food and water could easily be carried, sufficient for a prolonged period; the set of the Peruvian current was certainly known, and Peruvian sailors probably had some knowledge of the alteration in its course toward the north-west as it approaches nearer to the equator. With this factor in their favour, it amounts almost to certainty that their vessel, if dismasted in a storm, would be carried eventually to the immediate neighbourhood of the Marquesas islands.

The only instance, however, on record of a long westward voyage having been attempted is that made by the Inca Tupac Yupanqui in the reign of his father Pachacutec (1421-1478).

According to the account written by Sarmiento de Gamboa (1907), shortly after the Inca had conquered the coast and islands of the gulf of Guayaquil, some traders were brought to him, who stated that they had come by sea from the west in sailing vessels. Their marvellous stories of the riches of their island home fired the imagination of the Inca; he resolved to visit these wonderful islands and gain renown by their conquest. The construction of a mighty fleet of sailing-balsas was ordered, and when they were ready he embarked thereon a host of soldiers and set sail for the west. After the lapse of nine months, or, as some say, ‘about a year’, he returned victorious with the spoils of the islands—a number of prisoners, dark-skinned people, much gold, a throne of latón (literally ‘brass’, but probably copper), and the skin and jawbones of a horse (?). Some authorities, including Sir Clements Markham, consider that the islands visited were the Galapagos, discovered or rediscovered by Sarmiento de Gamboa in 1567; as it is Gamboa who tells the story of the Inca's expedition, it is natural that he should wish to identify his islands with those visited by the Inca; if these islands were indeed reached, the skin and bones of the reputed ‘horse’ may have belonged to a sea-lion such as the huge male which I once found basking on a sandy beach on Charles island in 1924.

As for the possibility that any Polynesian islands were found, this is immediately disproved by the mention of gold and latón; neither metal is found either in the Galapagos or in Polynesia. But both gold and copper are met with in Panama, and all the people of Central America were familiar with both metals.

To have found “black people” (Melanesians) living in an island where gold-deposits occur, the Inca would have had to reach the Fiji islands; this would have entailed a voyage so immensely long and difficult as to be quite beyond the capacity of the Peruvian sailors of that period, who seem to have had neither love of exploration nor any of the sea-lore and knowledge of navigation so marked a characteristic of the Polynesians and Micronesians.

That the Galapagos were not the only place visited by the Inca, seems to be ruled out by the fact that they were uninhabited when found later by Gamboa, and because neither gold nor any remains indicating former human occupation have ever been found there. A location possessing all the attributes necessary to fit in with the story of the spoils brought back by the Inca's expedition, cannot be indicated - 181 with any certainty; the mainland of the Gulf of Panama seems to be the locality most likely to have been raided by the Inca's armada but I am strongly inclined to think that the Galapagos islands were visited on the way there.

Lothrop's view that the expedition could not have reached the group because of the difficulty of finding its way thither without charts and navigational instruments may be discounted; we may be sure that the Inca's captains would distribute the fleet in such manner that the balsas would sail in line abreast, separated from each other by a considerable distance; this formation was the one best suited to achieve success when searching for the group of islands lying to the westward concerning which vague rumours were then current in the Peruvian ports.

These rumours must have originated from stories of an archipelago of considerable size seen far to the westward by the crew of some balsa blown seaward, far off its coastwise course.

If the Inca's expedition was inspired by these rumours, it is obvious that the voyage would begin by following a westward course. When and if the Galapagos were reached, their lack of inhabitants must have proved a grievous disappointment—no loot of any intrinsic value was to be obtained. Under these circumstances the Inca would seek counsel with his sea-captains. Those of them who had coasted northward to Panama would tell of the wealth of gold ornaments possessed by the chiefs of that country, and this would suggest that a raid on some of the important towns would yield booty rich enough to compensate for the failure to find the Galapagos worth attention.

If this proposal was adopted, a northerly course would take the fleet to the mainland of Panama. Gold was very plentiful in that land at the period in question, and golden ornaments would certainly have formed part of the booty brought back when raiding parties went ashore. If some of the prisoners captured were of darker skin than was then usual among Peruvians, this would be sufficient to give rise to the story that dark-skinned prisoners formed part of the plunder obtained.

Having visited the Galapagos and much of the coastline of Panama and Colombia, I am satisfied that this suggested sequence of events furnishes the most feasible and probable explanation of the course taken by the Inca's fleet on this memorable expedition.

The call made at the Galapagos would give a welcome opportunity to refit and reprovision the balsas. We may be sure that some of the giant tortoises perculiar to these islands were taken aboard to form a reserve store of livestock for use should the voyage be unduly prolonged, in the same way as was the common custom of whalers in the early decades of the nineteenth century. The fish so plentiful in the waters around the island would add a further addition to the larder, and if some of the Inca's soldiers killed a sea-lion on the beach, they probably carried off its skin and jawbone as trophies.

During the years between the reputed period of the expedition and the coming of the Spaniards, the tale seems to have become almost mythical and embroidered well nigh out of recognition. For example, the mention of the skin and jawbone of a ‘horse’, suggests the insertion of a topical allusion to the wonderful creature ridden by the Spanish invaders; it is easy to understand how the remains of a sea-lion might be ‘converted’ into those of a horse when the story was retailed to Spaniards.

- 182

Lothrop (1932, 238) besides denying the ability of the Inca's expedition to reach the Galapagos, is also unable to believe that the voyage was continued to Polynesia. A fatal objection to any such prolonged voyage is, he says, or rather meant to say, the inherent inability of a log-balsa to keep the sea “long enough to complete such a voyage”, 3 owing to the absorbent quality of balsa-wood, which soon loses its buoyancy unless frequently taken ashore and dried. Certainly no ordinary, untreated balsa-raft could make a prolonged oversea voyage unless the Inca's seamen knew of an effective method of treating its absorbent logs with some kind of waterproofing composition. This is likely to have been a preparation of gum, resin, or wax in some solvent, rubbed over the surface of the logs in the same ways as the Cayapa Indians of the interior of Ecuador coat the outsides of their dugout-canoes with beeswax in order to protect them from the danger of losing buoyancy (Lothrop, 1932, 230). If, as we have been told, the Ecuadorian sailing-balsas once plied south to Payta in Peru, and northward to Colombian ports, it seems reasonable to conclude that some waterproofing application was employed to enable these fairly lengthy voyages to be performed and the return trip to be made, without the necessity and labour of beaching and dismantling the whole structure for the purpose of drying out the component logs. The South American Indian of Inca times was certainly not less ingenious than the ancient Welsh, who carried a supply of butter with them on long coracle voyages for the purpose of smearing it over the hide covering of their frail craft to prevent decay, and of daubing it over the seams to keep them watertight; their descendants to-day coat the canvas which is their substitute for hide, with tar and pitch. In Ancient Peru and Ecuador forest-products alone were available, hence beeswax and vegetable gums would be the local alternatives.

Omitting as improbable, the possibility that Inca sailors ever voyaged westwards with the specific object of finding land except on the one historic occasion already mentioned, there remains the question of whether unpremeditated and involuntary voyages may have been made toward the Polynesian islands by some of their great sailing-balsas when dismasted and rendered unmanageable by an easterly or south-easterly gale. If carried toward the west or north-west by the storm, the forlorn craft would come under the influence of the South Equatorial current; thenceforward it would drift gradually toward the Marquesas islands; if it missed going ashore there, the craft would drift on through the Tuamotus to the Society islands, provided that the Tuamotus were passed without final disaster.

That this presumed sequence of happenings is to be accepted as fact and not unsupported fancy, is proved by the story of the drift-voyages of two modern derelicts, authenticated fully by the owners of the first, Messrs. Gracie, Beazley & Co., of Liverpool, and by the Secretaries of Lloyds' Corporation and Lloyds' Register of Shipping, in regard to the second. To the owners named, I am indebted for their great courtesy in looking up old records containing essential information; I also tender thanks to the officials of the two Lloyds' Institutions for the ready help accorded, nor must I omit to thank Monsieur Le Brunnec, the Breton store-keeper of Atuona in the island of Hiva-oa, for drawing my attention to these involuntary voyages, one of them an epic of the sea so fantastic, so full of human interest, - 183 of despair, hope, and courage, that it is with reluctance that I condense it and omit the details of the high lights and the tragic shadows.

The more notable of these two instances is the dramatic account of the capsizing and eventual stranding, after a lengthy voyage as a derelict, of the full-rigged ship Dagonar of Liverpool, a steel vessel of 2,665 tons gross, built in 1892, at Southampton.

Sailing out of Callao bay on 23 September, 1913, bound in ballast for Taltal, a port in Chile, the Dagonar ran into a violent gale from the south-west on 7 October; two days later a terrific squall struck the ship, throwing the ballast into the port wing; the next lurch threw her on to her beam ends. A luckless attempt was made to launch a boat—it was smashed against the ship's side and three men were lost. The captain was killed shortly afterward.

The position when the ship capsized was subsequently reckoned by the first officer to be approximately: Latitude 31° 45′ S., Longitude 85° 27′ W.

With great difficulty the masts were cut away but little relief was obtained. In this perilous condition hope was revived in the early hours of 10 October, by the sight of a green light approaching. The last rockets were instantly sent up and an answering signal received. At daybreak the vessel which had so providentially arrived on the scene proved to be the French four-masted barque Loire, of Dunkirk.

The miracle had happened; but for three dreadful days of misery and anxiety, during which the weather was too stormy to attempt a rescue, the Loire tacked to and fro, sometimes rounding the stern of the Dagonar to cheer her crew with the assurance that the Frenchman had no intention of abandoning them until resuce became possible. This was accomplished on 13 October, when a gallant attempt successfully took off the survivors during a lull in the weather.

From the time when the Dagonar capsized (9 October) until the rescue on 13 October, the helpless vessel was drifted by the Peru current for about 180 sea miles in a direction generally N. 30° W. This brought the derelict within the influence of the south-eastern section of the South Equatorial current. Thereafter, her course veered more and more toward the westward until it became due west. On 28 October, a passing vessel sighted the drifting hulk in latitude 27° S., longitude 95° W. (letter from the Secretary, Lloyds' Register, dated 5 October, 1944). The course taken thereafter, according to the data available, brought the Dagonar into the vicinity of the Marquesas islands; hereabouts the southern portion of the current stream becomes deflected toward the south-west. On this course the derelict would be carried through the northern Tuamotus, the most dangerous area of the Pacific ocean on account of the many closely scattered clusters of low-lying reefs and atolls that have earned for these islands the name of the ‘Dangerous Archipelago’.

Helpless as she was, the ship appears to have passed through scathless, eventually to run ashore on the coastal reef of Mopeha island, near Ra'iatea, in the Society islands. The date of stranding is given vaguely as “early in 1914”. As the derelict seems to have been afloat early in March, it is probable that it went ashore toward the end of that month. If we fix the date arbitrarily as 30 March, this would indicate a drift voyage, on the curving course followed, of between 5,000 and 5,500 miles performed in about 170 days, without allowing much for zig-zagging. The time afloat and the distance

- 184
Drift-voyages of the Dalgonar and Loire.
- 185

covered give the daily drift as from thirty to thirty-two miles, equal to a rate between one and a quarter and one and a third miles per hour. 4

The second vessel, the wooden Chilian ship Hindostan, of 1,629 tons gross, owned by Sñr. M. B. Peede, of Valparaiso, was burned out and lost on 8 March, 1900, in longitude 123° W., latitude 8° S., while on a voyage from port Blakeley to Caldera in Chile, loaded with a cargo of lumber. The crew were forced to take to their boats and “after a voyage lasting eleven days, the current carried them to the Marquesas islands” (letter from the Secretary, Lloyds' Committee, 12 October, 1944). This gives a rate of drift very much greater than the Dagonar experienced; the reason for this is probably that progress was accelerated by the use of sail in the boats, as the officers would be anxious to reach the islands with all speed.

The story of the third voyage ending in a drift which cast the vessel ashore eventually not far from where the Dagonar stranded, is less-well documented. I heard of it first when in the Marquesas in 1926; it concerns an American barquentine said to have been named the Red River, laden with a cargo of lumber, on her way to Callao. She is reputed to have sprung a leak off the South American coast; the captain decided to go before the wind, then blowing strongly from the south-east. On reaching the vicinity of the Marquesas islands the crew abandoned the vessel, which drifted to the south-west, and was eventually found derelict off Ra'iatea in or about 1919.

Unfortunately I have, as yet, been unable to obtain confirmation of these particulars. In view of the accuracy of the information given by the same informant regarding the course of events in the case of the two other instances here detailed, I feel inclined to believe that the facts are substantially correct, though it may be that there has been some mistake in regard to the name of the third ship.

Omitting consideration of the third instance of drift-voyages because of the need for fuller information, we are able now to see how the other two demonstrate how possible it was during the centuries of Inca domination in the north-west of South America for the Peruvian sea-going balsas, if coated with some form of waterproofing, to be drifted when dismasted in a storm to one or other of the eastern Polynesian islands if they happened to have aboard a good supply of fresh water and a sufficiency of food. The ordinary provisions carried would comprize usually a quantity of edible roots and tubers, such as those of the sweet-potato; there would often be a supply of various fruits containing seeds, from which new plants could be raised. Sometimes part or even the whole of the cargo might have been made up of such vegetable products.

So long as food and water lasted, a prolonged voyage aboard a balsa, even if crippled by a gale, would be easily bearable by the stoic Indians of Peru; if they eventually struck an island in the Marquesas - 186 or in the Tuamotus, and if they escaped death at the hands of the natives, we may be sure that they would indicate the food-value of the tubers and roots and fruits that remained uneaten. If the sweet-potato were among these, this would receive special commendation and its Quichua name would pass into the islanders' vocabulary.

There can be no proof that this ever did occur, but the hypothesis is easier to imagine than that of a voyage from Polynesia to the Peruvian coast, with a friendly reception there and a safe return home. As already explained, it is scarcely conceivable that return by the same route could be effected safely. By any other route, return would be equally or even more difficult and dangerous unless they obtained much more information about sea-routes from the Peruvians than their knowledge of navigation was likely to be able to furnish.


Before summarizing the body of evidence presented in the foregoing pages, we must know what was the distribution of the sweet-potato in the various groups of islands which make up Polynesia when this area was visited by explorers in the eighteenth century and in the early part of the nineteenth century, the period when Europeans and North Americans were most actively engaged in exploring and charting the island of Oceania. We have to know in particular what were the chief centres of the cultivation of this tuber, and what was its distribution in the outlying or peripheral islands. These comprized the trio, Easter island in the extreme east, Hawaii in the extreme north, and New Zealand in the distant south-west.

The available evidence for the answering of these questions is the following:—


Roggeveen, who discovered the island in 1722, and all subsequent visitors down to Beechey in 1825, speak of the sweet-potato as abundant there; it was grown in large plantations and was recognized as one of the most important of the islanders' food supplies.


Cook, who discovered the group in 1778, and those who followed in his track during the next quarter of a century, testify to the abundance and excellency of the sweet-potatoes grown there.


Cook and Banks in 1769 and many others agree in their testimony that the sweet-potato occupied an important place in the dietary of the Maori of North island, and it was later found in common use on the northern section of the east coast of South island.

Conversely, the evidence of eighteenth century voyagers through Central Polynesia tends to belittle the importance attached to the sweet-potato in the dietary of the people of this region. Dixon (1932, 45-49) goes into details which may be summarized as follows:—

In the Tuamotus, an archipelago consisting of low coral islands, conditions are unfavourable for the cultivation of the sweet-potato, and no references to it are on record. In Mangareva (Gambier islands) Beechey records the sweet-potato as of comparatively little importance but later records point to a revival of its cultivation after the arrival of the early French missionaries.

- 187

In the Society islands the sweet-potato occupied only a minor position in the islands' food-production. In the Cook islands, Williams (1837, 578) specifically attributes its introduction to missionary enterprise. In Tonga, it also had little importance as indicated by the fact that Mariner does not mention it save in his vocabulary; in Samoa it is again a late missionary introduction and was not in general use according to Turner (1861, 192), who is explicit that its introduction was due to European intervention.

The outstanding conclusion to be drawn from these facts is that the cultivation of the sweet-potato was of minor importance in Central Polynesia prior to the arrival of European missionaries in the early part of the eithteenth century, whereas it was flourishing in the three islands or island-groups situated at the points of the Polynesian triangle—Hawaii, Easter island, and New Zealand—when these were visited by Europeans in the eighteenth century and in the first quarter of the nineteenth.

To be in profuse cultivation in peripheral islands so distant from one another, presupposes a distribution at an earlier period throughout the whole area; the small importance which it had in the central area at the time of eighteenth century exploration, suggests that either some factor or factors were present there which adversely affected its cultivation, such as a high, humid, temperature, or else that, in the favoured central area, there were other food supplies available in abundance, better adapted to suit the conditions prevailing there; this question deserves attention, for at present we have not sufficient information on which to base any reasoned conclusion.


1. Of the principal food-products obtained from plants which were in general use in Polynesia when its islands were first visited by Europeans, the only one which seems, without reasonable doubt, to be a native of America, is the sweet-potato. All others of importance, such as taro, yams, and the coconut are derived, on the contrary, from species native to India, South-east Asia, or Indonesia; these appear to have accompanied those migrating bands who pushed eastward into the Pacific at some remote period, to become specialized eventually into the Polynesian race.

2. The problem before us resolves itself mainly into consideration of how and when this one plant-exception, an American-born species, was introduced into the island world of Polynesia.

3. Failing the possibility of diffusion along a land-bridge, we are obliged to conclude that it must have been transported by human agency by sea from its home in South America to one or other of the Polynesian islands, and that, by the same agency, it was gradually distributed to all the other islands. This diffusion must have taken place at a very early period, seeing that the cultivation of the sweet-potato had time to spread from the focal point of first introduction to every group throughout Polynesia—from Hawaii to New Zealand and from Tonga as far as Easter island—by the time when Europeans were making their major discoveries in the eighteenth century. This diffusion must have taken centuries to complete, seeing how widely scattered the islands lie and that intercommunication was greatly restricted after the close of the heroic age of Polynesian voyaging which ended sometime about the fourteenth century. After that we have no records of voyages from Tahiti to New Zealand, or to Easter island, and none to or from Hawaii and Tahiti, whereas it is known - 188 that the sweet-potato was held in high esteem as a valuable foodstuff by the people of Hawaii, New Zealand and Easter island when these islands were first visited by Europeans.

4. Consideration of the foregoing proves that the cultivation of the sweet-potato in Polynesia dates back to such an ancient period that the question of pre-Columbian contact between America and Polynesia must be answered in the affirmative.

5. How this pre-Columbian contact and the resulting introduction of this tuber into Polynesia were effected we cannot decide on the data available. Two alternatives have been envisaged; either Polynesian vessels sailed to Peru and returned with a quantity of sweet-potato tubers suitable for planting, or Quichua-speaking people in sailing-vessels of the balsa type, were drifted to Polynesia and there were enabled to plant the tubers in the islands which they reached. I incline to consider the latter explanation to be the more probable, as the Polynesian names for the sweet-potato are variations of the Quichua term kumar or kumal.

6. Did Polynesians voyage to South America and return? That a Polynesian voyage of discovery may have reached South America between the tenth and the fourteenth century, is extremely probable when we consider the extremely long voyages successfully accomplished by Polynesians during this period of intense maritime activity. It may, however, be urged that had such a voyage been made, its memory would have lived in tradition as have the voyages between Central Polynesia and New Zealand, Hawaii and Easter island.

The only route to South America that could conceivably have resulted in success is that from Tahiti or Samoa, viâ Mangareva, Rapa, and Easter island. After resting at Easter island, the voyagers would soon have fallen in with the northbound Peru current; this would prove of great help and would lead to a successful landfall in Peru if an easterly course were continued, gradually edging across the current.

A voyage from the Marquesas islands would have been much more difficult and uncertain, and this route may be eliminated at once because the drift of the South Equatorial and Peru currents would be contrary over the entire distance to be traversed; the length of the voyage would also be considerably greater than that by Easter island. The winds would also normally be adverse.

7. If the voyage were made viâ Easter island, the factors of winds and currents, so extremely favourable on the outward journey, would operate equally adversely for a return in the reverse direction; any chance of succeeding would be so reduced that it becomes negligible and has to be ruled out as hopeless.

Only by steering a course north-west from Peru would it be possible for the return to be made; this would lead viâ the Marquesas islands to the Society islands. This fact could not be known to the Polynesian voyagers; although unlikely, it is possible that information pointing to the favourable run of the currents on the first stage of this route was available from some of the Peruvian balsa-sailors; if so, a return voyage by this route may have been effected. Every fact considered, it seems extremely doubtful if any Polynesians ever succeeded in making the return voyage.

8. Peruvian drift voyages to Polynesia. The prominence given in Peruvian history to the expedition of the Inca Tupac Yupanqui, is evidence that a long voyage westward was a most exceptional event. - 189 Nowhere have we any suggestion, apart from this solitary exception, that the ancient Peruvians had either inclination or necessity to go voyaging in search of new lands.

But there remains the possibility and even the probability that transmission of the sweet-potato may have resulted from an involuntary drift voyage from Peru, consequent upon the dismasting and crippling of a balsa-raft when on a coastwise voyage. When such an occurrence happened, the northbound Peru current would take charge until a position was reached where the northward current merges into the South Equatorial current. This in turn would take charge and would carry the helpless craft westward to the Marquesas islands, where, granted a friendly reception, any tubers uneaten, would be taken ashore and planted, the Quichua name going with them.

This hypothesis is rejected by Lothrop owing to the known inability of balsa-logs to remain buoyant for more than a limited period; this, he contends, is a fatal objection to the idea of the ability of balsas to make long voyages.

But this defect may be remedied. In Ancient Egypt and in Assyria reed-bundle craft were made watertight, when this was considered necessary, by a coating of bitumen, just as the ‘bulrush ark’ of baby Moses was waterproofed. In ancient Peru a similar result may have been obtained by coating the logs with some preparation of beeswax in the way that certain Indians or Ecuador waterproof their dugouts; gum and resin are alternatives. Balsa-wood so treated would be well adapted to remain afloat long enough to enable a drifting balsa-raft to reach the neighbourhood of the Marquesas islands.

9. Proof of the value of the Peru and South Equatorial currents for the fortuitous transport of objects of material culture from South America to Polynesia, is afforded by the drift-voyage of the derelict ship Dagonar, which, after capsizing off the coast of Peru, drifted into the South Equatorial current and was carried by it past the Marquesas islands and through the Tuamotus, to strand at length upon one of the Society islands.

10. As an alternative to the pre-Columbian introduction into Oceania of the sweet-potato whether by Polynesians or South American Indians, Friederici has championed the view that this plant was introduced into the Marquesas islands and into certain Melanesian groups, by the Spaniards under Mendaña and Quiros in 1568, 1595, or 1606, the years when they had opportunity to plant the tubers on the islands which they had discovered.

This plausible explanation which, if well founded, would solve the problem in an easy and satisfactory manner, cannot be accepted for the following reasons:—(1) there is no evidence that the sweet-potato was not already under cultivation in Oceania at the period indicated; (2), there is no proof that the Spaniards did introduce the cultivation of this plant at any of the islands visited even so early as 1568; the reasoning that this did occur is based entirely upon “probabilities”, particularly upon the common custom of Spanish explorers to plant seeds and roots of useful fruits and vegetables at any place which they had hopes of colonizing or occupying, and (3) had the sweet-potato been introduced even so early as 1568, this would not have allowed sufficient time for its diffusion to Easter island, New Zealand and Hawaii by 1722, 1769 and 1778, the years when these islands were discovered respectively by European explorers, because long inter- - 190 island voyages such as those from Hawaii to Tahiti, and those from Tahiti to Easter island, had been discontinued after the close of the fourteenth century.

What appears reasonably certain is:—

  • A. That there was some degree of contact between Polynesia and South America in the pre-Columbian period;
  • B. That the sweet-potato's introduction into Polynesia was due to this contact, possibly arising from an isolated incident;
  • C. Whether this introduction was due to Polynesian voyagers who sailed to South America, and took back the sweet-potato on their return, or whether its diffusion originated in an involuntary drift-voyage from Peru, remains uncertain; I incline to favour the latter explanation as being the more likely.
  • Acosta, Joseph de, 1894—Historia natural y moral de las Indias, 2 vols., Madrid. A translation has been issued by the Hakluyt Society.
  • Beccari, O., 1917—“The Origin and Dispersal of Cocos nucifera”, Philippine Journ. of Sci., vol. 12, bot. 27-48, Manila.
  • Beechey, F. W., 1831—Narrative of a Voyage to the Pacific in H.M.S. Blossom in the years 1825-1828, London.
  • Best, Elsdon, 1923—“Polynesian Voyagers”, Dominion Mus. Monograph, no. 5, Wellington, N.Z.
  • Best, Elsdon, 1925—“The Maori Canoe”, Dominion Mus. Bull., no. 7, Wellington, N.Z.
  • Cook, Capt. James and King, James, 1784—A Voyage to the Pacific Ocean (1776-80), London.
  • Cook, O. F., 1901—“The Origin and Distribution of the cocoa palm”, Contrib. U.S. Nat. Herb., vol. 7, 257-293.
  • Cook, O. F., 1910—“History of the coconut palm in America”, Contrib. U.S. Nat. Herb., vol. 14, 271-342.
  • Candolle, A. de, 1886—Origin of Cultivated Plants, 2nd edn, London.
  • Dixon, R. B., 1932—“The Problem of the Sweet Potato in Polynesia”, American Anthropologist, N.S., vol. 34, 40-66.
  • Dixon, R. B., 1933—“The American Aborogines”, Fifth Pacific Science Congress.
  • Dixon, R. B., 1934—“The long Voyages of the Polynesians”, Proc. of the Amer. Philosophical Soc., vol. 74, no. 3, 167-175.
  • Friederici, G., 1920—“Zu den vorkolumb. Verbindungen der Südsee mit Amerika”, Anthropos, vol. 24, 441-487.
  • Friederici, G., 1936—“Die Süsskartoffel in der Südsee”, Mitteil. der Gesellschaft für Völkerkunde, no. 7, Leipzig.
  • Gamboa—See Sarmiento de Gamboa.
  • Guppy, H. B., 1903-6—“Observations of a Naturalist in the Pacific between 1896 and 1899”, vol. 2, Plant Dispersal, London, 1906.
  • Hornell, J., 1936—“The Canoes of Polynesia, Fiji and Mirconesia”, being vol. 1 of The Canoes of Oceania, Haddon & Hornell, 3 vols., Honolulu, 1936-38.
  • Juan, George, and Ulloa, Antonio de, 1760—A Voyage to South America, 2 vols., London.
  • Lothrop, S. K., 1932—“Aboriginal Navigation off the West Coast of South America”, Journ. Roy. Anthrop. Institute, 62, 229-256.
  • Marchand, E., 1798-1800—Voyage autour du monde pendant les années 1790-92, Paris.
- 191
  • Markham, Sir Clements, 1910—The Incas of Peru, London.
  • Mendaña, Alvaro de, 1901—“The Discovery of the Solomon Islands. . . in 1568”, Hakluyt Society, 2nd ser., vols. 7-8, London.
  • Merrill, E. D., 1920—“Comments on Cook's theory as to the American Origin and prehistoric Polynesian Distribution of certain economic plants . . .”, Philipp. Journ. of Science, vol. 17, 377-384, Manila.
  • Morris, Sir Daniel, 1918—“A Chapter in the Geographical Distribution of Plants—the Dispersal of Fruits and Seeds by Ocean Currents and Tides”, The South-eastern Naturalist . . . for 1918, 1-31, London, 1918.
  • Oviedo y Valdez, G. F. de, 1851-55—Hist. general y natural de los Indios, islas y tierra-firme del Mar Océano, 4 vols., Madrid.
  • Quiros, P. F. de, 1904—“The Voyages of Pedro Fernandez de Quiros, 1595 to 1606”, Hakluyt Society, ser. 2, vols. 14-15, London.
  • Rivet, P., 1928,—“Relaciones commerciales precolombianas entre Oceania y America”, Anales de la Facultad de Ciencias de la Educación, Tomo 3, 165-193, Parana, Rep. Argentina.
  • Sagot.,—., 1871—In the Bull. de la Soc. Bot. de France, 1871, 305-6 (quoted by de Candolle).
  • Samanos, Juan de, 1842—“Relación de los Primeros Discubrimientos de F. Pizarro . . .” in Colección de Documentos Inéditos para la Historia de España, vol. 5, Madrid.
  • Sarmiento de Gamboa, 1907—“History of the Incas”, Hakluyt Society, ser. 2, vol. 22, London.
  • Tasman, A. J., 1898—Abel Janszoon Tasman's Journal, English trans., Amsterdam.
  • Torquemada, J. de, 1723—Los veinte y un libros rituales y monarchia Indiana, Madrid.
  • Turner, G., 1861—Nineteen years in Polynesia, London.
  • Vancouver, G., 1798—Voyage to the N. Pacific Ocean .. 1790-5, London.
  • Williams, J., 1837—Narrative of Missionary Enterprises in the South Sea Islands, London.
  • Winkler, Capt. 1901—“On sea charts formerly used in the Marshall Ids”, Smithsonian Inst. Annual Report for 1899, 487-501, Washington.

NOTE.—For bibliographies of wider range, see Dixon (1932), Lothrop (1932) and Hornell (1936).

NOTE.—It would seem that the kumara (sweet-potato) was not in New Zealand before the year 1350. Toi-kai-rakau, who was driven from his home by a storm and made New Zealand without the kumara, later sent for it to Hawaiki, but the ones sent did not return. It was brought in several of the Fleet Maori vessels, so it must have been growing in the region from whence they came—presumably Ra'iatea, and consequently in the neighbouring Tahiti: see among other papers in Journ. Pol. Soc., Judge W. E. Gudgeon, “Maori Tradition as to the Kumara”, J.P.S., 2 (1893) 99-102; E. Best, “Maori Agriculture”, J.P.S., 39 (1930) 358 on; S. Locke, “The Visit of Pou to Hawaiki to procure the kumara”, J.P.S., 30 (1922) 40-47.

As regards possible contact with Peru, see notes on the musical instrument nguru, and its close similarity to an instrument from Peru, Johannes C. Andersen, Maori Music (1934) 264-266.

—J.C.A., Ed.
1   The present note does not deal with the problem of certain artifacts of apparently Polynesian origin which have found their way to South America in some unknown manner. This has been done because the late Professor R. B. Dixon has already dealt with the question in an able and dispassionate manner, leaving nothing more to be added, as I am in accord with his conclusions. See his article “Contacts with America across the Southern Pacific”, in The American Aborigines (Fifth Pacific Science Congress, pp. 315-353).
2   In Tamil maram signifies “tree” and kai, “fruit”.
3   Owing to a printer's error, this author has been made to state the opposite, that the balsa (jangada) has “inherent ability to remain at sea long enough to compete such a voyage”.
4   Great difficulty has been experienced in verifying, reconciling, and understanding several of the statements relating to the position of the Dagonar on specific dates: one error has arisen through the confounding of one date with another, and some are due to errors in transcription. To recapitulate would be tiresome. Apart from the positions given in the present account, which are derived from the first mate's log and from information kindly supplied by the owners and by the Secretaries of Lloyds' Corporation and Lloyds' Register of Shipping, confusion and inconsistencies are so general that we have to fall back upon the three basic observations which are fully substantiated, viz. (1) that the ship was abandoned in latitude 28° 58′ S.; longitude 87° 4′ W. (2) that the derelict was seen on 28 October in latitude 27° S.; longitude 95° W. and (3) that she eventually stranded on Mopeha island.