Volume 70 1961 > Volume 70, No. 2 > Polynesian navigation to distant islands, by Andrew Sharp, p 219-226
                                                                                             Previous | Next   

- 219

In this article Mr. Sharp, well-known as the author of “Ancient Voyagers in the Pacific,” further scrutinizes the arguments of those who believe that Polynesians performed feats of deliberate long voyaging and colonization. He pays special attention to the views of the late Harold Gatty.

IN A BOOK ENTITLED Ancient Voyagers in the Pacific, which was published by the Polynesian Society in 1956 and reprinted by Penguin Books in 1957, 1 I argued that deliberate navigation to and from remote ocean islands was impossible in the days before the plotting of courses with precision instruments, and that in prehistoric times neither the Polynesians nor any other people in the Pacific performed the feats of deliberate long voyaging and colonization with which they have been credited. I cited world-wide evidence in support of Captain Cook's theory that the detached ocean islands of the world in general, and the Pacific in particular, were settled by accident, 2 and applied this theory to a re-examination of the prehistoric Pacific migrations.

Since then a number of writers have continued, as I see it, to confuse myth with history on these themes. Prominent among them, because of his expertise in modern navigation, was the late Harold Gatty, who in his book Nature is Your Guide 3 repeated with modifications the theories of Polynesian long navigation given in his previous publication The Raft Book. 4 Roger Duff, a New Zealand ethnologist, in a recent article, 5 stated that he was impressed by Gatty's theory that the Polynesians sailed to distant islands by bringing stars overhead which were known from previous contacts to be overhead at those islands at a given time of the night and a given time of the year. While Duff did not come out unreservedly in this article for the Polynesian deliberate settlement of New Zealand, he thought the New Zealand Maori traditions of a Tahitian colonizing fleet contained evidence of fairly accurate navigation, because most of the Fleet canoes were described as arriving on the same part of New Zealand's coast. Robert C. Suggs, an American archaeologist, in his book The Island Civilizations of Polynesia, 6 similarly crossed swords with me over what I regard as peripheral and inconclusive issues, but failed to get to grips with the fundamental navigational issues which are the crux of the whole matter.

- 220

I cannot in an article repeat the world-wide evidence of the character and range of early navigation without precision instruments, which I cited in my book. What I can do is prove to any perceptive mind that the statements of Polynesian long navigation by Gatty and other deliberate voyage theorists are fallacious, and in particular that the deliberate settlement of Hawaii and New Zealand by the Tahitians or other Polynesians was impossible. Everybody is interested in the prehistoric Polynesian migrations to Hawaii and New Zealand because of the thousands of miles which separate them from Tahiti and the other islands of Eastern Polynesia, and because of the importance of Hawaii and New Zealand as centres of Eastern Polynesian culture.

Before proceeding to demonstrate the specific fallacies of the theories on Polynesian long navigation of Gatty and others who share such views, let me mention five facts which have an important bearing on the matters at issue, and which one does not ordinarily hear of from the deliberate voyage theorists.

The first of these facts is that the idea of deliberate settlement implies at least three voyages, two of them navigated ones—an initial voyage of discovery, a navigated voyage back to the home islands, and a navigated voyage of settlement back again to the discovery. The crucial voyage would therefore be the return one from Hawaii or New Zealand to report the discovery, following on which a further navigated voyage on the outward leg would be required to complete the picture implied by the deliberate settlement theory.

The second fact is the variable and unknowable character of the ocean currents on these supposed long journeys. In the case of Hawaii the voyagers would have had to pass through three separate belts of currents operating across their path, each characterized by marked local variations in width, direction and speed. In the case of New Zealand the voyagers would have had to negotiate the zone of the South Equatorial Current extending for hundreds of miles across their path on both legs of the supposed journeys, as well as a zone of variable currents extending for hundreds of miles round New Zealand. Such currents are imperceptible out of sight of land without the pin-pointing of courses by precision instruments, because the whole body of water surrounding a vessel which is being influenced by them moves with the vessel. The speed with which the Pacific currents move rafts, bottles and other floating objects is well known and remarkable. The action of currents on vessels is known as set. The phenomenon of set, therefore, is a major issue in a consideration of Polynesian navigation.

The third fact is the variable character of winds on the supposed long journeys. In the case of Hawaii, the voyagers would, on the leg from Hawaii to Tahiti, have had to buck the south-east trade wind which blows for 1,200 miles and more north of Tahiti and the other central islands of Polynesia. They would also have had to pass through the equatorial doldrums on both legs, not to mention the north-east trade wind round Hawaii itself. In the case of New Zealand, the voyagers would have had to pass through the belt of the south-east trade wind as well as an area of variable winds extending for hundreds of miles round New Zealand on both legs of the supposed journeys. In - 221 such circumstances the accurate calculation of drift or leeway with the wind was and is impracticable. Unknown drift of considerable dimensions was inevitable on the supposed journeys.

The fourth fact is that set and drift are independent variables, set continuing whether the wind is blowing or not.

The fifth fact is the relatively small target offered by remote islands as compared with islands near at hand. New Zealand, the largest land mass in the Pacific Islands, with a generous margin of 75 miles (Gatty's figure) for visibility or other natural homing aids on either side, occupies only one-sixth of the south-to-west quarter of the compass in relation to Tahiti. Hawaii is less again, as are the central groups of Polynesia on the return leg. A comparatively small error in the course to these objectives, therefore, would result in their being missed, as a glance at a map will show. For this reason the local Tahitian voyages of several hundred miles in the trade wind zone where set and drift were relatively predictable bear no relationship to the supposed two-way contacts over 4,000 miles and more of variable currents and winds. Putting on a golf green is a deliberate art because the hole offers a relatively large target. Holing out from a golf tee is achieved only occasionally and with luck, and is a one-way affair.

We are now in a position to consider the specific fallacies of the theories of Gatty and others, and the lack of any valid theory of Polynesian long navigation.

Gatty's main statement of Polynesian long navigation is that the voyagers came within 50 to 75 miles of a distant objective by bringing stars overhead which were supposedly known from previous probes to be overhead at the objective at a given time of the night and a given time of the year. It was this theory of Gatty's which impressed Duff. The time of night, however, is a relative and local phenomenon, because of the apparent constant motion of the sun. The local times of night at which a star is overhead at various points as it passes across the celestial sphere at a given time of the year are all within a few minutes of one another. It is not possible, therefore, to differentiate between one point and another where a star is overhead by judging the time of night at a given time of the year. The problem of longitude cannot be solved so easily. The range of positions on the surface of the Pacific Ocean where a star appears to be overhead at a given time of the night and a given time of the year is, therefore, limited only by the western and eastern sides of the Pacific Ocean itself. The margin of error in Gatty's theory, therefore, is not 75 miles, but 8,000.

The reason why the time of night when a star is overhead is not specific to locality is because the stars and sun follow one another round the celestial sphere during a rotation of the earth, and the time of night and day is determined in relation to the sun. If, therefore, it is four hours after sunset or eleven hours after high noon at one point on the earth's surface where a star is overhead during a rotation of the earth, it will be four hours after sunset or eleven hours after high noon at evey other point where the star is overhead as it completes its apparent circuit, apart from the negligible difference arising from the variation between sidereal and mean solar time.

- 222

In view of these facts, Gatty's argument that the voyagers, having come within 50 to 75 miles of their objective, used natural homing aids to make their final landfall, is beside the point. Gatty, in any case, like many others, exaggerates the applicability of these aids. Cloud effects over islands are not always present, and are frequently simulated over tracts of ocean where no land exists. The smell of land is discernible only when one is close to land, and only when the wind is blowing off shore. Releasing tropic-birds is no use if one has no way of knowing whether or not one is closer to Australia or Norfolk Island than to New Zealand, or to the Marshalls or Gilberts than Hawaii. Concentrations of sea-birds are not always present near land, and are detectable only in the vicinity when they are. The mountains of Savaii in Samoa are visible at 70-odd miles. Most of the Pacific Islands are visible at fifty to fifteen miles.

The further theory of Gatty and others that the Polynesians followed migratory birds is subject to the fatal defect that to keep in touch with the birds in the face of set and drift it would be necessary to see some birds every hour or two. In the hours of darkness at sea, however, birds are invisible, and are heard only intermittently, if at all, particularly in winds. It would therefore have been something of a miracle if the birds were still in sight after the first night, let alone the many nights required for the supposed long journeys. Furthermore, whereas migratory bird flights last two or three weeks, Captain Cook, when he made the first historical voyage from New Zealand to the Cooks, took five weeks to get there, and was still 600 miles short of Tahiti. 7

The reserve theory of Gatty that the Polynesians, having made a southing or northing to the latitude of their objective, then moved west or east along a line of latitude to their objective, is subject to the inherent defect that the voyagers had no way of knowing whether they were thousands of miles west of their objective while still thinking themselves east, or thousands of miles east while still thinking themselves west. (The reasons for this will be strengthened when the following two paragraphs have been read.) The chances of the voyagers' turning in the right direction on both legs of each supposed visit were, therefore, four to one against.

The time-honoured theory of Polynesian navigation, put forward by Teuira Henry, 8 Percy Smith 9 and many others, including Gatty himself in his earlier publication The Raft Book, 10 is that the Polynesians sailed toward guiding stars which were supposedly known from previous explorations to be in line with the desired objective. There are, however, no stable guiding stars on voyages to or from Hawaii, New Zealand or other remote Polynesian islands, and the stars do not shine during the day. Even if a theoretically ideal horizon star were to shine by night and by day beyond these distant objectives, the voyagers, as they sailed toward this beacon, would have no way of knowing if, because of set and drift, they had missed their objective by hundreds of miles on either side. This method of navigation is practicable only - 223 over short distances in areas where set and drift are predictable and can be allowed for. Even then it is only as good as the voyagers' dead reckoning of set and drift, as the observations of Thomas Gladwin, cited later in this article, show. Over courses of thousands of miles to and from New Zealand and Hawaii the method was useless, because of the changing bearings of the stars and the lateral displacements caused by set and drift.

The same defect applies to courses by bearings on the cardinal points, such as south-west, or north-north-west, or ‘a little to the left of the setting sun’. Again the voyagers would have no way of knowing whether they were off course as the result of set and drift, because their vessels, as they passed their objectives on one side or the other, would still be aligned to the south-west, or north-north-west, or ‘a little to the left of the setting sun’, as the case might be. This is the well-known defect of compass bearings also.

At bottom the techniques of sailing toward a point on the horizon marked by guiding stars and of sailing by a bearing on the cardinal points as determined by the stars including the sun are closely akin. In order to achieve anything like accuracy on voyages to and from Hawaii and New Zealand by such methods, it would have been necessary to have calculated and allowed for lateral displacement as the result of set and drift, which the Polynesians had no means of doing.

Horizon stars give no clue to lateral displacement because the objective is relatively near at hand but the stars are at a very great distance, so that the true course and all the possible false courses arising from lateral displacement all the way round the world are in effect parallel. Thus the widest separation between two possible courses is equivalent to the diameter of the earth, but the stars are astronomical distances away.

A reductio ad absurdum of theories of finding remote ocean islands by sailing toward guiding stars, aiming to bring stars overhead, and/or setting courses by bearings in relation to the cardinal points is given by applying these notions to courses toward the North Pole by sailing toward the Pole Star. The Pole Star, being almost in line with the axis on which the earth rotates, is relatively stable in its bearings in relation to all points in the northern hemisphere. It is unique in the heavens, there being no visible pole star in the southern hemisphere. If a number of amphibious vessels spaced across the Pacific at the equator were to sail toward the Pole Star, they would all at one and the same time be using the Pole Star as a guiding star, aiming to bring the Pole Star overhead, and sailing by a bearing on the north. Eventually in theory they would all come together at the North Pole with the Pole Star overhead. They would, however, have no idea of what their intervening courses had been, within the limits of 8,000 miles set by the west and east sides of the Pacific, nor what islands they might happen on, nor - 224 how far west or east of the direct course they might have wandered as the result of the inevitable vagaries of set and drift. The reason for this is that, wherever they were, they would at all times be sailing toward the Pole Star, aiming to bring the Pole Star overhead, and sailing by a bearing on the north. The North Pole is the only distant objective on the globe which there is any assurance of finding by sailing toward horizon stars, aiming to bring stars overhead, and/or sailing by bearings on the cardinal points. The Polynesians, however, did not live at the North Pole.

It has been pointed out earlier that the concept of the deliberate settlement of Hawaii or New Zealand or of other remote islands implies of necessity that the discoverers got back by deliberate navigation to the home islands. Let us suppose that after Hawaii or New Zealand was discovered by its original settlers, or re-discovered by later ones, some navigators set out on an attempted journey back to the central islands of Eastern Polynesia. That such attempts were in fact made is well within the bounds of possibility. Let us also assume that the navigators had carefully noted their courses on the sun, stars and cardinal points on the initial voyage of discovery, and that they set out for home at the beginning of the migratory season of the plovers or long-tailed cuckoos, so as to use them as an auxiliary navigation aid. The voyagers set sail, and after about fifty miles the coast behind them sinks beneath the waves. For the remaining hours of daylight the voyagers follow the remembered bearing on the sun's east-west path, as well as the birds. Then darkness falls. The birds, being independent of the set and drift which push the voyagers hither and yon, keep on course, invisible to the voyagers. The remembered courses on the stars, for the reasons given earlier, afford the pilots no clue whatever as to the influence of currents and winds in putting the vessels off course. Already, therefore, before the voyagers are more than a hundred miles or so on their journey of twenty times that distance, a concept of deliberate navigation ceases to have any reality. If by chance the navigators were still on course after the early part of their journey was completed, they would have no way of knowing where the westward set of the South Equatorial Current, extending for two thousand miles or more both north and south of the home islands, became dominant, nor the local changes of direction and speed within it.

What do these fundamental navigational issues add up to? Simply this, that the potential margin of error in navigation between Hawaii or New Zealand and the Tahiti area or anywhere else in Eastern Polynesia in any form of sailing or paddling vessel in the days before the plotting of courses with precision instruments was a great distance of hundreds or thousands of miles on either side of the objective on each and every leg of the supposed long journeys, with no way of knowing what the actual error was at any time. Deliberate long voyaging and colonization in such circumstances were impossible. Hawaii, New Zealand and the other detached islands in the Pacific were each settled by one or more accidental one-way voyages of waifs of the storm or exiles hoping to find new land. The evidence of design and planned - 225 transfer which Duff 11 and Suggs 12 think they see in the diffusion of artefacts, food plants, livestock and other cultural items to the peripheral islands can be explained by the setting forth of exiles with these items and their accidental arrival on these islands, or by the piecemeal diffusion of the items in several vessels over a period of time, or by both these types of transfer. Not only is it needless to think that the Polynesians first discovered distant islands, then returned to their home islands by deliberate navigation, and then set out once more on navigated voyages of settlement to explain the cultural facts, but also the facts of physical geography are such that there were no valid techniques of long navigation whereby this needless concept of deliberate long voyaging and colonization can be sustained. Accidental settlement by exiles who had made preparations and set out in the hope of finding new land, and by waifs of storms, could have achieved the peopling of the peripheral islands of Eastern Polynesia within a few centuries of the peopling of the central islands, since the winds and currents are relatively favourable to such voyages on the outward legs, although not on the return ones. 13

At the time when I was writing my own book, Thomas Gladwin, an American ethnologist, was independently writing up his own observations of voyaging in the Truk Islands, 14 where the Trukese voyagers still sail their outrigger sailing canoes without precision instruments. Gladwin's conclusions were the same as my own, namely that the inherent defect in such navigation was that it was fundamentally a dead reckoning system dependent on calculating and allowing for set and drift, and that when errors or failures in dead reckoning occurred, the voyagers were powerless to detect or correct them, in which case the landfalls were missed. Gladwin's observations were in due course published in the American Anthropologist. 15

Another formidable authority to the same effect is Brett Hilder, a captain of Burns Philp vessels with over thirty years' experience of Pacific navigation. In an article in a navigation journal, 16 he endorsed the view that deliberate navigation to and from remote islands in the Pacific in native vessels did not occur because of the mobility of currents and the difficulty of determining drift with winds.

Why do so many people hanker so determinedly after the notion that the Polynesians sailed to and from remote islands by deliberate navigation? After innumerable discussions I am driven to the conclusion that most people who cherish this belief do so for unconscious emotional rather than scientific reasons. These reasons are that the belief appeals to their romanticism, that it was taught to them at school, that it is flattering to the modern Polynesians, that many people have propounded it in the past, and that the modern believers have said it or written it.

- 226
  • COOK, J., 1784. A Voyage to the Pacific Ocean, Vol. 1. London, Nicoll & Cadell.
  • DUFF, R., 1959. “Neolithic Adzes of Eastern Polynesia.” Anthropology in the South Seas. New Plymouth, Avery.
  • — — 1960. “The Coming of the Maoris.” New Zealand Junior Encyclopaedia. Wellington, New Zealand Educational Foundation.
  • GATTY, H., 1943. The Raft Book. New York, Grady Press.
  • — — 1958. Nature is Your Guide. London, Collins.
  • GLADWIN, T., 1958. “Canoe Travel in the Truk Area: Technology and its Psychological Correlates.” American Anthropologist, 60:893-899.
  • GOODENOUGH, W. H., 1957. “Controls in the Study of Cultural and Human Evolution.” Journal of the Polynesian Society, 66:146-155.
  • HENRY, T., 1928. Ancient Tahiti. Honolulu, Bishop Museum. Bulletin No. 48.
  • HILDER, B., 1959. “Polynesian Navigational Stones.” Journal of the Institute of Navigation, 12:90-97.
  • SHARP, A., 1956. Ancient Voyagers in the Pacific. Wellington, Polynesian Society. Memoir No. 32.
  • — — 1957. Ancient Voyagers in the Pacific. Harmondsworth, Penguin Books. Reprint.
  • SMITH, S. P., 1921. Hawaiki. Wellington, Whitcombe & Tombs.
  • SUGGS, R. C., 1960. The Island Civilizations of Polynesia. Mentor, New York.
1   Sharp 1956 and 1957.
2   Cook 1784:202.
3   Gatty 1958:32-41, 76-86.
4   Gatty 1943.
5   Duff 1960:55-56.
6   Suggs 1960:83-85.
7   Cook 1784:167-170.
8   Henry 1928:401-402.
9   Smith 1910:185-186.
10   Gatty 1943:85-86.
11   Duff 1959:127-144.
12   Suggs 1960:83-85.
13   Sharp 1956:50-54.
14   Goodenough 1957:153, 155.
15   Gladwin 1958.
16   Hilder 1959.