- 284 - 285

THE OBJECT OF the following paper is to attempt the systematic identification of the Central Pacific Islands sighted by the Spanish explorers in their voyages across this ocean, and thus to close one of the two main gaps left in our record of their discoveries. 1

As will be seen, the Spanish made five voyages in all through this central region, in every instance from east to west, commencing with the initial entry of Magellan in 1521 and ending with the last expedition of Quiros in 1606. 2

Translations of the various accounts relating to all these expeditions have now been made from the original Spanish and Portuguese records, notably by the Hakluyt Society, and these have proved adequate for the purpose of enabling identification. Indeed, satisfactory identifications of the terminal islands—the objectives—of each expedition (whether the Philippines, New Guinea, the Solomons, Santa Cruz Group or the New Hebrides) have been already published in a number of books and specialist articles, as have those relating to the Marquesas and Tuamotus, which Mendana and Quiros sighted as their first landfalls in the Pacific. 3

Between the initial and final points of call there lay, however, the islands of the Central Pacific: a horseshoe of fifty-eight coral atolls and reef islands divided for practical purposes into five groups—the Gilberts (16), Ellice (9), Tokelau (4), Phoenix (8) and Line (11)—with 4 unattached outliers.

The Central Pacific Islands straddled right across the path of the Spanish and it is not surprising, therefore, that they discovered no less than nine of them on their expeditions. It is with the recognition of these nine islands that we are here concerned for, with two exceptions, 4 no serious attempt has been made to establish their identity since the days of Burney; 5 and painstaking and thorough though he was, Burney was inevitably handicapped by living at a time when the geography of the Pacific was still relatively unknown.

- 286

The reason for this omission is not hard to find, for the islands lie remote from shipping and air routes, 6 many of them being uninhabited and seldom visited by Europeans. Yet, where at first sight one low coral atoll so closely resembles another, detailed personal acquaintance with every island that might conceivably have been visited by a Spanish ship is an essential prerequisite to accurate identification.

This then is the author's main qualification for essaying a task from which the authorities on exploration have apparently recoiled. For during a period extending over twenty years I have visited all the groups and all but ten of the individual islands in the Central Pacific (including every one of the Spanish ‘possibilities’), and on most of them I have stayed for varying lengths of time. In several instances, e.g. Niulakita, my visit was made primarily for the purpose of checking an identification already made provisionally and I was therefore able to approach the island from the same direction as a Mendana or a Quiros and compare the descriptions of the old chroniclers with the visual image before me.

One has to remember that throughout the period of the Spanish discoveries the calculation of longitude was hardly possible, except by the often highly inaccurate method of estimating from dead reckoning. 7 It is fortunate, under the circumstances, that most of the courses across the Central Pacific after Magellan's involved only a comparatively simple “latitude sailing”, and that under the easy wind and weather conditions prevailing in the region the latitudes obtained can be relied on to within at any rate two or three degrees. 8

In making the identifications, therefore, I have depended in the first instance on latitude, and with islands as far spaced as they mostly are in the Central Pacific this usually narrows the field to a few possibles only. For further elimination and final identification, however, one must rely mainly on description, and it is here that a personal knowledge of the islands, their geographical characteristics and their inhabitants (if any), is important.

In a few instances the identification of one island is at least partly dependent on that of another, and for this reason it proved necessary to establish the identity of Magellan's San Pablo, which lies outside our particular area, before proceeding to fix that of Tiburones, which is within it. But here it must be confessed that the process of resolving the puzzle as to which Pacific Island was the first seen by European eyes had its own satisfaction.

Although references have been cited in the text to any previous identifications which merit consideration, no attempt has been made to - 287 list the academic conjectures of every arm-chair writer on the Spanish voyages. To do so would add considerably to the length of this work without serving any useful purpose; for in the main I have found them reckless in their disregard of known geographical facts, confusing large islands with small, reef islands with atolls, inhabited islands with uninhabited, and wandering far over the map in seeming indifference to probable course and estimated latitude.

Unfortunately, the resultant errors have with the passage of time tended to become accepted facts, to be found repeated in every reference work on the area. To give an example, the latest edition of the official Annual Report of the Gilbert and Ellice Islands Colony claims authority for supposing the discovery of “certain Gilbert Islands” by Grijalva and Alvarado in 1537, of others in the same Group by Luis Lopez in 1543 and Miguel Lopez in 1565, and of Nukufetau by Mendana in 1567; all understandable assumptions from the existing secondary source material, but as I hope to show none the less mistaken. 9 If the present paper results in removing a few of these time-hallowed misconceptions, it will have served its purpose.

In the case of Mendana's Isle of Jesus and Quiros' Buen Viaje, it was necessary to enlist the special skills of a practical navigator to confirm the identification, and here my sincere thanks are due to Captain G. H. Heyen, the acknowledged doyen of Central Pacific master mariners, whose appendix (the product of many hours of night work after a busy day as General Manager of the Commonwealth Shipping Lines) speaks for itself. Without his expert assistance this paper would lose much of any value it may possess.

Supreme feat of enterprise and endurance though it was, the crossing of the Pacific by Magellan which ushered in the era of European discovery almost succeeded, by some extraordinary mischance, in avoiding every one of the multitude of South Sea islands among which he threaded his route. For three months and twenty days his three ships, of which the largest was only 110 tons, traversed the for once calm ocean from the coast of South America to the Marianas, and during the whole of this time they sighted only two small and uninhabited islands, on which they found nothing but “birds and trees”. Pigafetta, the main chronicler of the voyage, states 10 that:

- 288

A second account, compiled by Maximilian of Transylvania from enquiries made after the return to Europe of the survivors of the expedition, runs as follows: 11

Although other sources, when they mention the islands at all, add nothing new to their description other than to state that the southern was named San Pablo and the northern Tiburones, they all vary in the estimates given of their latitudes, as will be seen from the following table:

	Latitude of		Bearing and Distance of Tiburones
Authority	San Pablo	Tiburones
1. Francisco Albo	16° S	10 2/3° S
2. Pigafetta	15° S	9° S	200 leagues from San Pablo
3. Genoese Pilot	18°-19° S	13°-14° S
4. Anon. Portuguese	18° S	14° S	800 miles from San Pablo
5. Barros	18° S	13° S	200 leagues NW from San Pablo
6. Santa Cruz	17° S	11° S	180 leagues NW from San Pablo

Herrera mentions that the islands were sighted thirty days after leaving the South American coast in 32° 20' S, Brito that San Pablo was found 1,600 leagues WNW of 32° S and Barros that it lay at a distance of 1,500 leagues from the Straits of Magellan. 13

By far the most detailed account of Magellan's route is contained in the log-book of the pilot Francisco Albo, as will be seen from the entries for the period covering the discovery of the Unfortunate Islands reproduced below:

Date—1521	Course	Position
January 24	W¼NW	16¼° S

- 289

January 25	NW¼W	15¾° S
January 26	NW¼W	15 1/3° S
January 27	NW¼W	15° S
January 28	WNW	14½° S
January 29	WNW	13¾° S
January 30	W¼NW	13½° S
January 31	W¼NW	13 1/3° S
February 1	NW	13° S
February 2	NW	12½° S
February 3	NW	11¾° S
February 4	NW	11¾° S

February 5

NW

10° S

In view of the primitive navigational equipment of Magellan's day, Albo's latitudes can scarcely be regarded as more than approximations, despite their ostensible accuracy within fractions of a degree; while longitudes, even of the most approximate character, are too much to expect at this early date. “At the present time”, says Pigafetta, in a much-quoted passage from his Treatise of Navigation, “the pilots content themselves with knowing the latitude; and are so proud they will not hear speak of longitude.” 15

Nevertheless, by making a rough calculation of distances run from the latitudes given in Albo's log, except when the ships were sailing due east and west, and by supplementing his information from other accounts whenever possible one can plot Magellan's course with a reasonable degree of probability, particularly for the earliest stages of the voyage.

This is confirmed by the fact that the various attempts which have been made to trace Magellan's route from the log show a remarkable similarity in their general outline, and all that I have seen take him to the north of the Tuamotus and thence in a north-west direction between the widely spaced islands of the Line Group. Indeed, any deviation would have either brought him within sight of the mountainous islands of the Marquesas, visible for a considerable distance at sea, or else lost him in the intricate maze of the Tuamotu Archipelago. To take Magellan to the north-east of the Marquesas or around to the south and west of the Tuamotus can hardly be done on the basis of Albo's courses.

We have, therefore, to look for a small uninhabited island, complete with trees and birds but with no good anchorage, somewhere to the east or north-east of the main labyrinth of the Tuamotus and between the parallels of say 14° and 20° S; and about 200 leagues away, in a - 290 roughly NW by W direction, another island possessing similar features, lying in shark-infested waters between say latitudes 8° and 15° S.

We can get closer than this with reasonable certainty, for owing to the greater accuracy of the positions given by Albo over the often hearsay estimates of the other informants our islands are likely to be situated not much more than a degree on either side of his latitude. Unfortunately, Albo gives the latitude of Tiburones but not of San Pablo, but taking this latter as the mean between his sights on January 24 and 25, or 16° S, we can expect San Pablo to be somewhere between 15° and 17° S and Tiburones between 9° 40′ and 11° 40′ S.

Now there is no lack of small islands in the Tuamotu Group which, as far as we can tell, were uninhabited (or only occasionally inhabited) at the time of Magellan's voyage; and several of these lie between 15° and 17° S. It so happens, however, that all except one are situated in the main complex of the archipelago and it is difficult to see how Magellan could have sighted one of them without, on the courses given, having also sighted several others either before or after: if at night he would likely have hit one. Many, probably most, of them were also at least partly planted with coconut palms (as they found later at the Marianas and Philippines), which would have been manna indeed to Magellan's crews, then near starvation; and those frequently visited would have presumably shown some signs of the fact for noting by Pigafetta.

One island, and one only, satisfies our criteria to a nicety: Pukapuka, in latitude 14° 50′ S and longitude 138° 50′ W, is the isolated north-eastermost outlier of the Group. It is small, and in 1930 was uninhabited and covered in trees, but with no coconut palms; 16 the bird life is prolific, as one would expect; and it has no navigable entrance into the lagoon. Neither in reaching nor leaving it would Magellan have passed within 60 miles of any other island on the courses set down by Albo, at which distance a coral island is, of course, well below the horizon even from the mast-head.

But if Pukapuka is San Pablo, we should be able to find the second of the Unfortunate Islands by sailing for approximately a further 200 leagues on the same courses that Albo took: and such a voyage will in fact bring us to the Southern Line Islands of Caroline, Vostok and Flint.

The problem of identifying the island of Tiburones is actually not very difficult, for the only other possibly uninhabited island between the parallels of 9° 40′ and 11° 40′ S, or for that matter between those of 5° 30′ and 14° S, is Manihiki, in the Northern Cook Group (that is to say within conceivable longitudes). Tongareva, Rakahanga and Pukapuka in the Cook Islands were all inhabited and even Manihiki was probably planted with coconuts and periodically occupied by 1521; 17 while to reach it Magellan would have to sail right round to the south of the Tuamotus (or the main part of them) and to the south and west of Tahiti, touching at uninhabited Mopihaa, Fenua Ura or Motu One— - 291 all north-west of Tahiti—on the way. But to make such a deviation we must scrap Albo and the evidence corroborating him and proceed on pure conjecture.

To place Tiburones away to the east in the Marquesas is to indulge in even more improbable guess-work; for then where is San Pablo? There is no land, inhabited or uninhabited, to the south-east of the Group. And in any case, none of the four uninhabited islands in the southern Marquesas could very well be reached without sighting others (they are all high and visible for miles at sea), while the northern Marquesas consist of two islands separated by a channel 2 miles wide, and a sandbank without trees.

There seems no option, therefore, but to identify Tiburones with one of the Southern Line Islands; and this whether San Pablo is Pukapuka or some other island in the north-eastern Tuamotus. With the latter we have no further concern, as it lies outside the area of this particular study; but the Line Islands are in the Central Pacific.

Which then of the three islands is Tiburones? From their position it could almost equally well be Caroline, Vostok or Flint; so, as in the case of many other identifications, we must turn to the internal evidence offered by the islands themselves.

Caroline island, in 10° S, can, I submit, be ruled out, for it was apparently already planted with coconuts when visited by Quiros in 1606 and sighted by Broughton in 1798, 18 while the marae ruins and basalt adzes found on the island, the former showing a close relationship to those in the Tuamotus, confirm the fact of former occupation. 19 Graves and stone platforms are still to be seen on several of the islets which extend along both shores of its 5 mile lagoon. 20

It is doubtful if Caroline was ever regarded as a permanent home, and it may well have been unoccupied at the time of Magellan's voyage. Whether this was so or not is, however, immaterial, the important point for our purposes being that the period of voluntary and involuntary colonization in the Eastern Pacific long antedated the 16th century and it is consequently reasonable to assume that Caroline had been visited and planted before 1521 and that the coconut palms would have been as conspicuous a feature then as they were 75 years later, when we learn of the island's “many cocoanuts” and “plenty of palms”.

Caroline Island, furthermore, consists of a cluster of some two dozen islets set on a reef 13 miles in circumference, and here the very fact that there is no mention of a lagoon in the accounts of either island seems to be significant; for the peculiar structure of the coral atolls almost invariably attracted the special attention of later expeditions, the land being described as “inundated”, “in pieces”, or by some similar phrase.

When one thinks of all the minor phenomena of the voyage so scrupulously reported by the ever curious Pigafetta, it seems likely

- 292 - 293

that he would have made some remark had he visited a typical atoll such as Caroline—quite spectacular when seen for the first time—or at least described Tiburones as a group rather than a single island. Since he did not, a reasonable inference is that both San Pablo and Tiburones were in fact reef islands, with at the most enclosed lagoons, and not a string of more or less separated islets on an exposed reef. Tiburones, therefore, was not Caroline.

Vostok Island, in 10° 05′ S, must also be dismissed, but for a different reason. It is the smallest islet standing on its own in the Central Pacific and so minute (about 700 yards at its greatest length 21) that with its round shape it provides no lee for even an island schooner, let alone for Magellan's three vessels. Even on calm days, as I can personally testify, it is safer and a good deal more pleasant to keep out at sea rather than approach this islet—a mere speck in the vast ocean—where the waves appear to break with almost equal force all round and the swift current rushing past creates eddies which add to the general turbulence. No-one who has ever been to Vostok would regard it for an instant as the island where Magellan's starving and exhausted crews stopped “for their health's sake, and general recruiting of their bodies”.

So we are left with Flint, in 11° 25′ S, and here we have an island that answers exactly to the descriptions of Tiburones given by the chroniclers of Magellan's expedition. Small (about 2½ miles long), it is the nesting place of innumerable sea-birds, who would have been even more in evidence before its occupation in 1872, and its pre-plantation vegetation included the trees recorded by Pigafetta. 22 Judging by the absence of archaeological sites it appears never to have been visited in pre-European days and, as a consequence, there were no coconut palms, for at least in the dry latitudes these have invariably been planted by man. 23

With its limited rainfall and sparse plant coverage, Flint would certainly have been described as barren by any Spanish voyager, but despite the lack of any proper anchorage there is good shark fishing to be obtained on the west, or lee, side of the island. This last feature, however, is admittedly not peculiar to Flint, for the lee of every uninhabited island in the Central Pacific teems with sharks.

It should perhaps be emphasized that the identification of Tiburones with Flint does not necessarily depend on San Pablo being Pukapuka. As far as latitude goes San Pablo could be any one of the uninhabited islands in the north-eastern Tuamotus, and there are others at approximately the right distance and direction from Flint. For Pukapuka as San Pablo there are, then, admittedly alternative possibilities, if Magellan could have reached them without hitting, or at least sighting, some other island in the process, which seems doubtful; but for Flint as Tiburones there is really no alternative wherever one looks.

- 294

Before concluding, however, it is necessary to mention briefly Nunn's contention that Albo's log, and presumably any corroborative evidence in the other accounts, was concocted to prove that the Spice Islands fell within the Spanish half of the world and bears no resemblance to the route taken by Magellan, who actually crossed the Pacific in the Northern Hemisphere. 24

To deal with all Nunn's points, which are ably argued even when they do not convince, would require a separate paper. But his main proposition that anyone following Albo's route would inevitably have been drawn to some island by observing the flight of birds would scarcely persuade a Pacific Islander, to whom birds are only reliable as an indication of land when it is close by. 25 As we shall see, fifty years later Mendana sailed between the Marquesas and Tuamotus, and indeed right across Polynesia, without sighting a single island until he reached the Ellice Group in longitude 177° E; a more remarkable feat than that of Magellan, who did sight two. And one must remember that Magellan required his pilots to follow “the course he had given, punctually” 26 and was not the man to be deviated easily by birds or anything else.

The vindication of Albo lies in the fact that the route as plotted from his positions passes just where it must if it is to avoid all land between South America and the Marianas with the exception of two small islands. When one considers the number of archipelagos lying in between it certainly would have been a most extraordinary feat if he had, in total ignorance of their existence, succeeded in constructing an imaginary course across the Pacific which touched land precisely where he said they touched—and nowhere else.

Magellan's voyage is a striking example of how wrong the navigators of his day could get by their reliance on dead reckoning alone in calculating distances run: from the Straits to the Philippines Albo's error was of the order of 53° ! 27

For this reason, until a practicable method of ascertaining longitude was discovered in the 18th century, latitude sailing was the standard practice on most ocean voyages. In other words, navigators were accustomed to make for the parallel on which lay their ultimate, or an intermediate, destination and run along it—or “make westing”—until it was reached.

As we shall see, all the remaining Spanish voyagers to visit the Central Pacific followed a latitude through the area with more or less fidelity; and as latitudes could be calculated with reasonable accuracy our task of identifying the places visited is greatly simplified. With a - 295 steady south-east trade wind blowing and clear skies overhead, latitude sailing in the Central Pacific should have presented fewer practical difficulties than almost anywhere else.

The next expedition to touch the area actually followed the Equator itself for its latitude; for this much seems established, though the accounts of the route followed are admittedly scanty, the best being by Antonio Galvano, Portuguese Governor of the Moluccas, and Couto. 28

It appears that Cortes, the conqueror of Mexico, sent Hernando de Grijalva and an officer named Alvarado to take troops to the aid of his colleague Pizarro, then having difficulties in Peru. Alvardo thereupon returned to New Spain from the Peruvian port of Paita with despatches, while Grijalva proceeded to explore the Eastern Pacific for rich islands rumoured to lie to the west of the Americas. 29 Abandoning his search on reaching 29° S he tried to make for California but was driven back by E and NE winds.

When again near the Equator, Grijalva's crew lost patience and demanded that he should make before the wind for the Moluccas. On his refusal they mutined, killed him, elected the Master of the ship Captain, and sailed due west.

Four months later, after a disastrous voyage across the Pacific with the fickle winds and frequent calms that one expects in the equatorial doldrums, the seven survivors sighted the northern coast of New Guinea. Here, after their rotting ship had fallen to pieces, they were captured by the natives and sold as slaves, and the following year some of them, who had been taken by their masters to the Moluccas, were ransomed by Galvano. 30

While Galvano treated the Spanish kindly, he was sure that they must have been ordered by Cortes to take the peculiar route they did in order “to discover that way a long under the equinoctial line, because the islands of Cloves stand under that paralele”. 31

Furthermore, although Couto writes that New Guinea “was the first land they reached”, Galvano was able to find out from them that they had, in fact, sighted a number of islands when crossing the Pacific, and of these, two were almost certainly in the Central Pacific. 32

To quote from Galvano's narrative, 33 the expedition sailed—

- 296

Later still they discovered several more islands, Haime, Apia, Seri, Coroa, Meosu and Bufu, inhabited by people described as “blacke, and have their hair frisled, whom the people of Maluco do call Papuas”. 34 But these are clearly well out of our area and the evidence suggests that they lay to the north of New Guinea.

Despite the absence of any description of Acea there is little difficulty in identifying it as Christmas Island, in latitude 2° N and longitude 157° 30′ W. 35 Christmas is approximately 1,000 leagues from the American mainland and, in any case, there are no islands between with the exception of the Galapagos Group, which is far too close to Peru. The other atolls of the Northern Line Group are well to the north of Acea's recorded latitude, Fanning Island, the nearest, being twice its distance from the Equator; while to the west there is no land for a further 1,100 miles.

On the rediscovery of Christmas by Cook in 1777, Buache, in his memoir Sur des Descouvertes a Faire dans le Grand Ocean, at once identified it with Acea and Burney, in expressing his agreement, says that, “By the article O being prefixed, though the word Acea is not - 297 found in the Spanish dictionaries, it may be presumed that the name was imposed by the discoverers; which is in favour of M. Buache's conjecture, as the island seen by Captain Cook was not inhabited”. 36

The identification of the Isla de los Pescadores is more conjectural. Its name indicates that it was inhabited, which eliminates Howland and Baker and makes it almost certainly one of the Gilbert Islands, for travelling along the Equator Grijalva's crew must have passed right through the centre of the group.

The Gilbert Islands nearest to the Equator are Aranuka (0° 12′ N), Kuria (0° 14′ N) 6 miles to the west, Abemama (0° 23′ N) 13 miles to the north-east, and Nonouti (0° 40′ S), the others lying at least a degree on either side of the line.

But if the mutineers, on a due west course, had passed near enough to Aranuka to have sighted the canoes fishing off the reef, it is difficult to see how they could have missed sighting Kuria, and possibly Abemama, as well. Had they sighted Abemama or Kuria in the first instance they would inevitably, while sailing along the Equator, have been even nearer to Aranuka. But the record speaks of them sighting “another” island: not two or more.

It would seem most probable, therefore, that the Isla de los Pescadores discovered in 1537 was Nonouti, and that the natives seen were engaged in fishing from canoes off the northern reef; the Nonouti people being renowned as deep sea fishermen to this day while the sea to the north of the island was one of their most frequented grounds.

It is necessary to consider the routes of these two expeditions owing to the already mentioned statement in the official Gilbert and Ellice Islands Colony Report that “Catholic histories record that Spanish vessels commanded by Luis Lopez in 1543 and Miguel Lopez in 1565 visited the Gilbert Islands en route to the Philippines”. 37

This is almost certainly a reference to Hartzer, the only Catholic historian to touch on the question in any detail. But a study of Hartzer's work will show that he made no claim that either Ruy Lopez de Villalobos or Miguel Lopez de Legaspi sighted any of the Gilberts: his observations relate entirely to discoveries made by them in the Marshalls. 38

It seems clear, furthermore, from the actual accounts of the voyages that the routes taken by both explorers did in fact take them through the Marshalls: in about 9° N. Villalobos is recorded to have touched this latitude on several occasions, but nowhere to the south of it; while Legaspi, in compliance with his orders, kept to the same parallel “a quarter of a degree more or less, according to the difference of the pilots in observing the sun”. 39 And 9° N is a long way from the northernmost of the Gilbert Islands in latitude 3° 16′ N.

- 298 - 299

Although Hartzer makes the general observation that “the Gilbert Islands . . ., on the way to the Moluccas, were certainly sighted several times during the sixteenth century”, 40 he offers no proof in support of this contention and I regard it as highly unlikely that any ships on the Navidad or Acapulco-East Indies run ever came so far to the south. All evidence points to their keeping to a defined track across the Pacific: 41

On the return journey, following Urdeneta's example, they sailed even further to the north, between latitudes 30° and 40°, so as to take advantage of the westerly winds and currents.

If any captains strayed hundreds of miles to the south of the regular route, and thus entered the Central Pacific, all one can say is that no record of the fact has come down to us; and in any event neither Villalobos nor Legaspi can be included among their number.

Apart from Loyasa, who appears to have roughly followed Magellan, and Grijalva's mutinous crew, who sailed along the Equator, all Spanish expeditions whose routes are known had hitherto taken a course across the Pacific from Mexico which passed well to the north of the Central Pacific Islands.

This procedure was ended, however, with the consolidation of the Spanish position in South America. Soon after this it was resolved to send out the first expedition to traverse the South Pacific, the ostensible reason being “to convert all infidels to Christianity”. The command was given to Alvaro de Mendana, nephew of the Governor of Peru, who left Callao with his two ships on November 19, 1567. As in the case of Magellan, the course taken looks on a chart almost as if it was designed to avoid any contact with the Polynesian islands; passing neatly between the Marquesas and Tuamotus, the Central and Southern Line Islands and the Phoenix and Tokelaus.

On January 16, 1568, they sighted their first island. There are no less than six accounts of this event, that of Mendana himself being as follows: 42

- 300

As was so often the case during the voyage, the narrative of Catoira, the Chief Purser, is almost identical. It is worth repeating, however, if only because it indicates that, despite Mendana's assertion, those on the Capitana were able to get a view of the island from the eastward as well as from the west. The fact that Mendana was below having a meal while his ship was passing to the south-west of the island, as is evident from his own record, may perhaps account for the discrepancy.

According to Catoira then: 43

Gallego, the Chief Pilot, confirms this already rather clear picture of the island, without adding anything significant to its description: 44

He estimated it as being “only 5 or 6 leagues in size at the most”, and added that when they finally left it he steered WNW from a position when “we were about 12 leagues off the island, and we had the island to the north-east of us, altitude 7½ degrees”. Later still, Gallego describes the island as being in “7 degrees south”. 45 The remaining two narratives give no description of the island itself, other than to state that it was small.

Mendana called his discovery the Island of Jesus, as it was discovered on the day following the feast of the Holy Name. It was inhabited and five canoes came nearly within bow-shot when their occupants raised their paddles and turned back with shouts. Mendana thereupon ordered signals to be made to them with a white cloth to try and get them to return, instead of which they landed and in turn stuck up signals along the shore “though we could not determine whether they were of palm-tree matting, or of cotton, because they were more or less white”. 46 At night, when one of the ships showed a light, it was copied by a fire lit on shore, and when it was put out the fire was extinguished

- 301 - 302 - 303

also. Gallego says that the natives were “naked and mulattoes” 47 and Sarmiento that the island “had a large fishery”.

As it was late, Mendana decided to defer landing until the morning and kept the ships tacking all night. With the dawn, however, a strong westerly storm blew up, January being in the middle of the “westerly” season in the Central Pacific, and although they tried all day to regain the island they were at length compelled to give up. It should be mentioned that while they were tacking Gallego thought that he saw another island to windward, but was unable to reach it as “the rain and darkness were very great”: Mendana considered it nothing but cloud. 48

Continuing west, Mendana's next discovery was the Candelaria Reef in 6° 15′ S, which has been identified with Ontong Java in 5° 00′ S, thus effectively establishing the fact that the Island of Jesus is in the Ellice Group, since there is nothing at all between the two. The only question is: which of the Ellice is it?

Hampered as we are by the lack of precision in Gallego's latitudes and the generous choice of positions offered us for the island, ranging from “barely 6 degrees” to 7½ degrees south, 49 we can nevertheless immediately strike off the three reef-bound islands of Niutao, Nanumanga and Niulakita, which have no “bays of the sea” in the middle of them. We can also delete Nanumea and Vaitapu, whose shape puts them out of question; Nukulaelae, which is too far to the south, even allowing for Gallego's errors in latitude; and Funafuti, which is too large to be described as 6 leagues in circumference and, furthermore, bears no remote resemblance to the Spanish descriptions. 50

We are left, therefore, with Nukufetau in 8° 00′ S and Nui in 7° 16′ S. C. M. Woodford, who was familiar with a large part of the Pacific and visited Nukufetau (but not, I think, Nui) on a voyage to the Gilberts in 1884, made a study of the identity of Mendana's discovery and came to the conclusion that it was Nukufetau. Lord Amherst of Hackney and Basil Thomson, in a note on Gallego's account of the voyage, concur with Woodford's view, on the grounds that “with the defective instruments of the time, a description of natural features is infinitely more trustworthy than nautical observation”. 51

- 304

One can readily agree with the editors of Mendana's voyage on the value of description as opposed to position in identifying the discoveries made by the Spanish explorers; but where in fact does it lead us in the case of the Island of Jesus? I have visited all the Ellice Islands on several occasions, and twice with the Spanish accounts particularly in mind. As we closed with the land at Nukufetau, and later at Nui, it was at once apparent that while the former could not be described as “a little low-lying island about 6 leagues in circumference, full of groves of trees, which were palms”, this was Nui exactly portrayed; while Nukufetau certainly looks nothing like “two galleys, with a copse in the middle which appears like a fleet of ships”, Nui does; and while the north and south points of Nukufetau are not connected in the west by a “strand lying lengthways, with reefs in different parts” or by “a wide beach beset with reefs”, this is a feature of Nui. Above all, from a distance Nui looks like an island, Nukufetau like a collection of islets, often widely separated, on a partly submerged reef.

Since obviously the appearance of the two islands from various positions at sea must be the deciding factor, it is fortunate that the Ellice Islands are all inhabited and consequently visited by administrative and trading vessels; for the masters of these have visual impressions of their various landfalls more vivid than mine, being based on more visits.

With this fact in mind I sent the accounts of Mendana, Catoira and Gallego already quoted to three of the most experienced captains in the islands, two of whom had been visiting the Ellice Group for a decade or more. No indication was given as to which island the descriptions might refer to, other than the fact that it was probably one of the Ellice.

From Captain G. H. Heyen's detailed analysis, the gist of which is given in the Appendix, it is evident that the description overwhelmingly favours Nui, as does the latitude, and that although size at first sight might appear to favour Nukufetau, this is not necessarily so; while in any case a visual estimate of the circumference of an atoll made from a passing ship could only be most approximate.

Independent confirmation of Captain Heyen's summing up is provided by Captain E. W. Harness, former master of the Gilbert and Ellice Islands Colony administrative ship Nimanoa and later Harbour Master at Suva, Fiji, who writes that: “I have come to the conclusion that Jesus Island appears to be Nui Island (Ellice)”; and by Captain Brett Hilder, of the Burns, Philip island fleet, who states his opinion that: “In regard to the Island of Jesus, only Nui fits the description of virtually three islets, two main ones with a group of trees between.” Even when I later urged the claims of Nukufetau, these local experts were not to be moved.

It is indeed difficult to argue why, when Nui almost exactly answers the description of the Island of Jesus and is almost exactly on the latitude where we are told the island lay, one should identify Mendana's discovery with an atoll 80 miles away in the wrong direction and in appearance unlike anything that he, or his colleagues, said that they had seen.

- 305

It should perhaps be mentioned that although Mendana discovered only one island in the Central Pacific during his first voyage, he only just missed sighting several more, for on his return from the Solomons he sailed northwards almost parallel with the line of the Gilbert Group. On September 2, 1568, he was in latitude 3° S and only a few miles to the west of the Gilberts when, according to Gallego: 52

Catoira also mentions that before reaching the Equator they saw signs of inhabited land, “such as palm-mats, leaves and burnt sticks”. 53 Believing it to be New Guinea, Mendana continued north (probably between Ocean Island and the Gilbert Group) until on September 17 he reached one of the Marshall Islands, where he saw a chisel made from an iron nail.

How this nail reached the Marshalls has been the subject of some conjecture, and Lord Amherst has argued that it must have been a relic of the Grijalva voyage in 1537, since “Asea was, doubtless, one of the Gilbert Islands”, and he assumed native trade connections between the Gilbert and Marshall Groups. 54 Apart from the mistake in his location of Acea, it is far more probable that the nail came from the expedition under Loyasa and del Cano (and on their death under Salazar), which passed through the Marshalls in 1526; or that under Saavedra, who lost two ships in a storm when in the northern part of the Group during the following year; or from those led by Villalobos in 1543 or Legaspi in 1565. It seems probable, furthermore, that other ships also passed through, or close by, the Marshalls between 1521 and 1568.

But why, in any case, is it necessary to suppose the visit of some ship to account for a piece of iron found in native hands, for in many parts of the Pacific nails are known to have floated ashore embedded in driftwood? And that the Marshalls were no exception is evident from Kotzebue who, although he disbelieved the islanders when they told him such a story during his visit there in 1817, later found some wood himself on the beach, “with nails remaining in it”. 55

Despite official indifference, and at times hostility, and a multitude of discouragements which would have daunted the enthusiasm of most men, Mendana never rested until he had gained permission to lead an expedition to follow up the discoveries made during his first voyage and to colonize the new lands.

- 306

It took a quarter of a century, by which time Mendana was nearly 50, before he finally succeeded in obtaining the necessary ships and men, the fleet leaving Callao in April, 1595, with Mendana himself in charge and Quiros the Chief Pilot.

There are two main accounts of the voyage: a long narrative probably dictated by Quiros to his Secretary, Belmonte Bermudez, and a shorter report written by Quiros to the Governor of the Philippines. 56 From these we learn that the expedition, after discovering several islands in the Marquesas, left Tahuata in that Group on August 5 and sailed west, keeping as far as possible between the parallels of 10° and 11° S. The wind was, as one would expect at that time of the year, E to SE.

After 14 days, at dawn on August 20, they were

This from the main narrative, the second account being as follows: 58

In the first account the island was said to be in 10° 20′ S and 400 leagues from the Marquesas, and in the second account in 10° 45′ S. It was named San Bernardo.

Leaving the identification of this island for the moment, let us proceed with Mendana for a further 9 days, steering a westerly course (between NW and SW) and still endeavouring to keep between 10° and 11° S, the wind being a fairly steady SE trade.

On the ninth day, August 29, Mendana sighted

A graphic description, which the shorter account adds little to, merely stating that on August 29 “we discovered a round islet, which might be a league round, all surrounded by reefs. We tried to land on it, and could not find where to do so . . .”. 60

Mendana called the island Solitaria and estimated its distance from Lima to be 1,535 leagues. In both narratives its latitude is given as 10° 40′ S.

Some years ago I called at Niulakita, the southernmost of the Ellice Group (10° 45′ S; 179° 30′ E), partly to see if it accorded with the above description, and I was satisfied that it did with remarkable exactitude. Furthermore, there is no other island within many days sail which even remotely meets the requirements.

The best account of Niulakita was given by Bennett, who visited the island in 1830, from which the following are brief extracts: 61

Niulakita has recently been estimated to be about 3½ miles (or almost exactly the league estimated by Mendana) in circumference; and is described in the Admiralty Pilot as “fringed by a reef, which extends about 2 cables from the western side, and a bank extends from it, with a depth of 14 fathoms at a distance of one mile, and 8 fathoms at half a mile”. 62 The only practicable boat passage through the coral reefs is on the south side. Several coral banks are reported around the island from the south to NNE, with depths as little as 14 fathoms.

Whether from the clear descriptions of the island by Quiros; its minute size, unique in the area; its estimated latitude (only 5 miles north of its true position); or from the time taken to reach it from Tahuata (23 days) and to proceed on to Ndeni (8 days); or the estimated distance of the island from the Marquesas (535 leagues) and from the Santa Cruz Group (315 leagues), I consider that there can be no real doubt that Solitaria is in fact Niulakita. Indeed, what other island could it be?

Having identified Solitaria, we are now in a better position to do the same for San Bernardo, which Quiros estimated to be 135 leagues to the east of it and 400 leagues from the Marquesas.

We know that Mendana was sailing west between 10° and 11° S when he saw the island: but in what longitude would we expect to look for it? Quiros estimates that from the Marquesas (Tahuata is in 139° W) to Santa Cruz (the east side of Ndeni is in 166° E) is 850 leagues. Now if 850 leagues = 55 degrees of longitude, then 400 leagues = approximately 26 degrees, and we would expect San Bernardo to be in the vicinity of 165° W. Or from the time taken from the Marquesas to San Bernardo (14 days) and from San Bernardo to Solitaria (9 days), remembering that from the narratives Mendana appears to have had a constant trade wind all the way, a rough calculation will show that we might expect the island to be not far from longitude 164° W.

As one can see from the chart there is only one atoll anywhere near a position of between 10° and 11° S and about 164° or 165° W, and that is Pukapuka (10° 53′ S; 165° 49′ W). 63 Nassau can be dismissed as not being an atoll, while Manihiki and Rakahanga are too far to the east: almost half-way between the Marquesas and Niulakita.

Quite apart from its position, however, the description of San Bernardo by Quiros is a description of Pukapuka, with its small, low islets and its conspicuous sand banks on the eastern side, particularly towards the north, rather than of Manihiki or Rakahanga, with their much larger and more numerous islets in a relatively closed ring

- 309

round the lagoon. Unfortunately, Mendana did not go close to the land, not even near enough to see if it was inhabited or not, so many details which we should like to know are missing, but nevertheless the picture he gives is, in my opinion, quite enough to enable an identification to be made.

I think that the only serious objection that can be raised to this identification is the fact that Quiros mentions four islets, whereas Pukapuka nowadays consists of only three (Pukapuka, Motu Ko and Motu Kotawa) and a sandy cay to the SW (Toka). In the first place, however, one must recognize that from a ship a good way off the atoll, Pukapuka can appear to consist of an indefinite number of islets, depending on one's distance and position.

More important, however, is the fact that Pukapuka is steadily diminishing in size and both hurricanes and tidal waves have been known to reduce the land area in recent times. Toka is a particularly unstable islet, having been entirely washed away in 1914 by a tidal wave, but is already re-formed into dry land. 64

- 310

According to traditional history, the most devastating of all tidal waves, probably accompanying a hurricane, occurred about 300 years ago, or after Mendana's visit: as recorded by Beaglehole, this wave reduced the population from between 1,000 and 2,000 to 15 men, 2 women and the remnants of their families. “All the island was broken and everything destroyed”, says the tradition: 65 one can well conjecture the effect on an outlying islet such as Toka, and indeed on other of the smaller islets that may have existed on the reef. And devastations such as this have occurred several times, though probably never as severe.

In brief, I suggest that Quiros probably did see four islets on the Pukapuka reef in 1595, but that one or more hurricanes and tidal waves have denuded the fourth of vegetation and reduced it to the size of Toka or else blotted it out entirely. Other islets were no doubt served likewise in the unstable history of this atoll.

Like Mendana, Quiros had a hard struggle to obtain permission to undertake a further journey to the South Seas, and it was through the intercession of Pope Clement VIII himself that the necessary instructions were at length issued to the Viceroy of Peru. Even so, there were delays in fitting out his two ships and launch, and it was not until Decvember 21, 1605, that he was able to set sail from Callao.

There are five published accounts of this expedition, the last of the great Spanish voyages of exploration: by Quiros himself, as dictated to his Secretary, Belmonte Bermudez; by the Chief Pilot, Gaspar Gonzales de Leza; by a Franciscan Friar, Torquemada; and lastly the Relacion of Diego de Prado y Tovar and a letter from Torres to the King of Spain. 66 Prado left Quiros and joined Torres at Taumaco and thus accompanied him on the journey through Torres Strait.

From Callao, Quiros sailed to the west and south until, on January 26, he sighted the lonely easternmost outlier of the South Pacific coral atolls, the little island of Ducie (La Encarnacion).

Three days later the ships were off Henderson Island (San Juan Bautista), where the course was changed to WNW which brought them, on February 3, to South Marutea (St. Elmo) and the following morning to the Actaeon Group (Las Cuarto Coronadas).

There can be no doubt as to the identification of these islands. The latitudes given (except Prado's, which appear to be merely approximations) are accurate enough to indicate fairly closely where the discoveries may be looked for, as will be seen from the following table:

	Actual Latitude	Leza	Quiros	Torquemada	Torres
Ducie	24° 40′	25°	25°	25°	24° 30′
Henderson	24° 20′	24° 45′	24° 45′	—	24°
South Marutea	21° 30′	—	20° 30′	—	21° 15′
Actaeon Group	21° 15′ to 21° 29′	20° to 21°	20°	—	—

- 311

The courses (W from Ducie to Henderson, and thence WNW) are what one would expect, and the distances are not unduly out, when one considers the method of calculation.

It is, however, the clear descriptions given of the islands, and particularly of Henderson and the Actaeon Group, which render identification certain. For, unlike all the other 13 islands discovered by Quiros between Callao and Taumaco, Henderson was not an ordinary coral island but “massive, moderately high, open, having groves and plains. It is steep, too, and its beaches are rocky” (Quiros); and lying “N. & S., all flat, with a hill to the S.” (Leza).

Where else, nearer than Makatea (over 1,000 miles away in lat. 16° S), can one find a raised coral island such as Henderson: flat-topped, 100 ft. high and steep too? Quiros noted the little bay on the SW coast and Leza considered that landing would be best on the beach to the NE, which has since been proved to be the case.

Where else in the Tuamotus, furthermore, other than in Actaeon Group, can one find “four islands in a triangle, 5 or 6 leagues each; low, uninhabited, and without soundings” (Torres)? But there is no need to argue at greater length, for the identifications agree with those originally made by Sir Clements Markham and, so far as I am aware, have been accepted generally.

Unfortunately, in his identifications of the remaining four islands in the Tuamotus discovered by Quiros—San Miguel, La Conversion de San Pablo, Decena, Sagittaria and Fugitiva—Markham would appear to have been led seriously astray by Sir William Wharton, who held that Sagittaria was probably the island of Anaa, which lies between latitude 17° 20′ and 17° 30′ to the south-east of the main Tuamotu chain. 67 He may also have been influenced by a statement made by Quiros that his course from Las Cuarto Coronadas to San Pablo was WNW.

Wharton, however, was mainly concerned with proving that Sagittaria (by which he presumably meant La Conversion de San Pablo) was not Tahiti, as had hitherto been generally supposed, but one of the low islands of the Tuamotu Group; and for courses and other navigational matters Leza, the Chief Pilot, gives by far the most detailed and consistent account, just as Albo did for Magellan′s voyage.

Leza says that the course from Las Cuarto Coronadas was NW and NW by W, while Torres gives it as NW throughout. These courses would not take Quiros near Anaa, especially as it seems probable that following the NW and WNW wind on February 7 the vessels experienced a fairly strong set of the current to the east until they finally left the Group. 68

- 312

In any case even if their courses had led them straight to Anaa, they still could not have reached it in the approximately 94 hours of sailing taken on the journey. 69 From the time taken in traversing known distances as well as from the estimated run per diem it will be seen that the ships sailed at a speed of between 3.6 and 4.3 knots: to reach Anaa in the time taken they would need to have sailed at over 6 knots. Furthermore, they saw an island bearing E by N when about 100 miles from San Pablo, and after leaving it on a NW course sighted an island to the N the same day, another to the N the following day and a final island to the E the day after that. A glance at a chart of the Tuamotu Islands will show that, all other considerations aside, San Pablo cannot be Anaa because in that case the necessary islands to fit into this clearly defined pattern simply do not exist.

Where then is San Pablo? If we take into account the course from the Actaeon Group, the distance, the latitude given for the island and the descriptions there can surely be only one answer: La Conversion de San Pablo is Hao, and the island of San Miguel which they passed on the way is Vairaatea.

Leza gives San Pablo as lying between 17° 45′ and 18° 10′ S. Hao is actually between 18° 3′ and 18° 26′ S, so his error was insignificant, as it was for San Miguel (19° instead of 19° 19′ S).

Quiros seems to have come up to Hao at its SE point (the easternmost extremity of the island), where there was an extensive grove of coconut trees with a village on the lagoon shore. 70 Standing off and on during the night, at dawn they approached the island farther to the west along the southern coast of the atoll, “a narrow and long reef almost covered by the water”.

At the SW corner of Hao lagoon is a second islet, also covered with coconut trees, and here instead of a village the landing party found a typical Polynesian marae, of which an excellent description is given by Leza and Torquemada. 71

From the various positions recorded for La Conversion de San Pablo, 72 it could conceivably be either Amanu or Marakau-Ravahere in place of Hao; but the former is ruled out if only because Quiros could not have traversed the southern coast without also sighting Hao, only 9 miles away, and the latter because of its shape, which bears no resemblance to the descriptions given.

These descriptions are indeed unusually clear: the large size of the atoll, which Leza estimated at 20 leagues long by 10 leagues wide; the narrow coast line, which “runs from the south end, E. to W., and from the north, N.W. to S.E.”, as it does; the great lagoon, which left “all - 313 the interior inundated, just as if it was a piece of the sea surrounded by land” ; and the two islets on the south coast with their coconut groves and connecting reef. 73 Would that the early explorers had always been so explicit in what they saw.

Rounding the SW point of Hao early on February 12, Quiros coasted along the western shore (which, as stated, runs from NW to SE) and continued on a NW course until, later the same day, the little island of Tauere was sighted to the north and christened Decena.

The following noon Rekareka was on the horizon, also to the north, and called Sagittaria. Both these islands were small and, being to windward, could not be reached. The next day saw the ships off the western shore of Raroia atoll, described as a large island, probably inhabited, lying in 15° S. 74

Quiros was now clear of the great Tuamotu archipelago, a temporary change of course on the 15th to NW by N taking him to the east of Takaroa, which was not sighted.

The courses, distances run and estimated latitudes from Raroia to the next landfall, on February 21, are set out in detail by Leza. 75 The expedition sailed partly NW and partly NW by N for 110 leagues, when in latitude 10° 45′ S (by Leza's noon sight), they changed their direction to W, with a view to continuing on that parallel until they should reach the island of San Bernardo, discovered during Quiros' previous voyage with Mendana, and which we have identified with Pukapuka.

According to Leza, they reached 10° 45′ S on the 18th and altered course that day: when the sun was taken on the 19th they were still in 10° 45′, and on the 20th and 21st in 10° 30′. That afternoon they discovered an atoll and, standing off and on all night, they found an anchorage on the west side, though it ultimately proved only suitable for the launch.

Quiros describes the island as “uninhabited, divided into four or five hummocks, and all the rest submerged”; Torquemada states that “it runs N. and S. . . . In its centre there is a great lagoon of salt water, like many of those already seen”; and Torres that “it was in pieces”: excellent descriptions of an atoll. Prado is even more precise, relating how at first the discovery “seemed to be three islands”, but that after they had landed they found that “it consisted of twenty-two islets, uninhabited and without water, trees or scrub for wood”. 76

While exploring ashore they found an old canoe, lying on its side, and great quantities of fish, which they caught in their hands, or with sticks and swords; lobsters and crayfish were also abundant and birds - 314 “so importunate that they seemed to want to attack the men”. Best of all, there were large quantities of coconuts lying at the foot of “plenty of palms”. But they could find no water and, though they dug wells, the water in them was salt. 77

The distance of the island from Lima was estimated at 1,400 leagues and its latitude, according to Quiros, was 10° 40′ S. Leza, however, gives 10° 30′ and Torquemada, who mentions that a sight was taken ashore, says that it was in “10° 30′ barely”; Prado, whose latitudes are approximate only, gives 10°. Estimates of its size also vary, from 12 leagues in circumference (Leza), 10 leagues (Quiros), to 8-10 leagues (Torquemada).

To confuse matters, the island was also called San Bernardo, since it was thought at first that they had arrived at Mendana's island: it was a natural enough mistake, for it was in the same latitude, the pilot of the Almiranta had predicted that San Bernardo would be sighted on that very day and, superficially at least, it had a similar appearance “because it was in pieces”. A few, with less reason, declared it to be La Solitaria, but eventually, as both Torres and Prado emphasize, it was seen to be a new discovery; and in Leza's journal (though not in that of Torres) it was renamed the Island of Fish. 78

The San Bernardo of Quiros was, in fact, far removed from the San Bernardo of Mendana; for if we calculate his position day by day from leaving Raroia, we can see that when it was sighted he should have been somewhere in the vicinity of latitude 11° 10′ S and longitude 150° 30′ W. Even using Leza's positions we cannot be more precise for we do not know the exact time at which the courses were changed; if they altered to W some hours after the noon sight on the 18th, for example, the estimated position would be a corresponding number of miles farther north and east.

Yet rough and approximate though our calculations may be, they will suffice for our purpose; for a glance at the chart shows that there are no islands within a radius of over 600 miles of this position other than Penhryn, 450 miles to the west in latitude 9° 00′ S and inhabited, 79 and

• (i) Flint (11° 25′ S, 151° 48′ W) : about 75 miles to the west;
• (ii) Caroline (10° 00′ S, 150° 14′ W) : about 75 miles to the north; and
• (iii) Vostok (10° 05′ S, 152° 23′ W) : about 125 miles to the northwest.

Fortunately for us, the new discovery was an atoll which rules out both Flint and Vostok and we are left with no alternative but to identify the San Bernardo of Quiros with Caroline Island.

- 315 - 316

Caroline, furthermore, fits our requirements to perfection. It runs north and south, as Torquemada said, and is usually held to consist of twenty-four islets (cf. Prado's twenty-two) irregularly spaced around a lagoon 6 miles in length. To make identification even more certain, it was the only uninhabited atoll planted with coconut trees within 200 miles of the parallel of 10° S other than Manihiki, over 600 miles farther west. 80

Accepting Markham's identification of San Pablo with Anaa (as I first did without checking, since it lay outside the area of this study), I had myself imagined that San Bernardo must indeed be Manihiki; and it was only the kind advice of Father Celsus Kelly, the authority on Quiros, that led me to re-trace the journey through the Tuamotu archipelago island by island, and thus to realize the impossibility of such a conjecture. 81

Leaving San Bernardo towards the evening of the 22nd, Quiros sailed west at a steady 25 leagues a day for 7 days, when at dawn on March 2, after covering an estimated 195 leagues (say roughly 670 miles) between the parallels of 10° and 10° 30′ S, they came to “a small island, 3 to 4 leagues in circuit, 82 covered with cocoanut palms, but no other trees”. The island was inhabited, with “a village under a grove of palm trees, near a lake which the island has in the middle”, and was estimated to lie in 10° 20′ S. 83

Quiros' account runs as follows: 84

As to the inhabitants of this island, of whom some 500 were seen assembled on the beach, they were “the most beautiful white and elegant people that were met with during the voyage”, “especially the women, who, if properly dressed, would have advantages over our Spanish women”. 85 Quiros was so struck with admiration that he called his discovery the island of Gente Hermosa, or handsome people; though he also called it Peregrino (or Peregrina), while Torres named it Matanga.

Again we are fortunate in that out of the two possible contenders for the honour of being Gente Hermosa one alone meets our criteria. For if we look along the parallel of 10° S it will be readily seen that Manihiki and Rakahanga are the only coral atolls within 1,300 miles to the west of Caroline.

- 317

Both these are at the right distance (approximately 650 miles, as against Leza's estimate of 670), and both in the right latitude (Manihiki in 10° 24′ S and Rakahanga in 10° 03′ S).

The descriptions, however, are of a fertile, well-planted island with an enclosed (or almost enclosed) lagoon and exceptionally bad landing conditions. “This island”, says Leza, “like the others, was inundated inland, but it looked much better, as it was all one palm-grove”. 86 In other words it was Rakahanga, and not the cut-up Manihiki.

The accounts, furthermore, speak clearly of a permanent settled island, with a well-built village consisting of houses “curiously and cleanly worked, each with a roof, open behind, and all the floors covered and lined with mats”, “very well made, of various colours”; and of the “great quantity of soft and very fine mats” which they used for - 318 clothing, which were also found in the houses, with tools and implements of a considerable variety, “knives, saws, chisels, punches, gouges, gimlets and fish-hooks” made of pearl shell, needles of bone and adzes. 87

Now Rakahanga possessed a permanent settled population in 1606, in fact since about A.D. 1350, when it was colonized by a single family from Rarotonga. Not so Manihiki, which was planted but only occasionally visited by the people of Rakahanga, who would naturally live there under camping conditions and not surrounded by all their material possessions. It was actually not until 1852 that the loss of life in these periodical visits led to their abandonment at the insistence of the missionaries, whereupon the population was for the first time divided and each island occupied permanently. 88

The various accounts given by members of the expedition add up to quite a useful epitomé of the main features of the pre-contact material culture of the Rakahanga people, as described later and in greater detail by Buck. Besides the inhabitants themselves and their village, 89 the Spanish were struck by the outrigger canoes, in which most of the natives eventually fled across the lagoon, the extensive puraka plantations, a noted feature of Rakahanga to this day, and the clothing worn by both sexes, including the “palm leaves made like tippets over their shoulders” by the women.

But enough has been said to confirm our identification 90 and we must leave Quiros to pass on: to the tragicomedy at Espiritu Santo and the anti-climax of the return voyage. A disappointed man, he was still confident that the importance of his discoveries would be recognized and rewarded by the command of another expedition to complete the work of exploration and colonization.

The decision to return would not, however, concern us here but for the fact that it resulted in the discovery of one more island in the Central Pacific. Leaving the latitude of Santa Cruz on June 21, Quiros sailed north, so far to the west of the Gilbert and Ellice Group that even the usual company of sea birds was missing. On the 26th course was changed to N by E and, after crossing the Equator on July 3, to NE. On July 8, says Leza, the Chief Pilot: 91

The account given by Quiros himself of the discovery is only slightly more discursive: 92

- 319

Torquemada says nothing of this island, and Torres, with Prado, was then, of course, on his way to New Guinea.

Despite the lack of descriptive detail it is possible to state with confidence that Buen Viaje is the island of Butaritari, in the Northern Gilberts, which, with Little Makin Island, stretches from 3° 00' N to 3° 19' N. The next island to the south is Marakei, in 2° 03' N, and there is no justification for holding that Quiros was as much as 1½ degrees wrong in his latitude. To the north there is nothing until one comes to Ebon, the most southerly of the Ralik chain in the Marshalls, in 4° 35' N, and Mille, the most southerly of the Ratak chain, in 6° 14' N.

Mille can be dismissed as being too far away and even the south of Ebon is over a degree away from Quiros' position as against only 11' for the north of Little Makin. 93 There is nothing in the descriptions to make one reject an island almost on the spot for one six times the distance away, and Ebon suffers from the additional drawback that if Quiros had discovered it and then proceeded on the courses described by Leza he could not very well have missed sighting islands in the south of the Ratak chain as well.

As various navigational points seemed involved the identification of Buen Viaje was referred to Captain G. H. Heyen, who first visited Butaritari when mate of the sailing barque Alexa and later lived there for several years, when he surveyed the whole of the lagoon and surrounding reefs. The substance of his reply is given in the Appendix, from which it will be seen that Quiros most probably first sighted Bikati Islet, on the NW corner of Butaritari lagoon: any other landfall would seem to necessitate a change of course to clear the atoll. While Quiros must have passed close to Little Makin Island it is unlikely that he actually saw it, since the difference between his noon sight and the position given for Butaritari rather indicates that on a NE course he would have passed it at night: if he did sight it, or lights on the shore, he evidently considered it to be part of Butaritari. 94

- 320 - 321

• AMHERST OF HACKNEY, Lord, & THOMSON, Basil, 1901. The Discovery of the Solomon Islands by Alvaro de Mendana in 1568, 2 vols. London, Hakluyt Society (Ser. I, nos. 7-8).
• ARGENSOLA, B. L. de, 1708. The Discovery and Conquest of the Molucco and Philippine Islands . . . London.

• BEAGLEHOLE, E. & P., 1938. Ethnology of Pukapuka. Honolulu, Bishop Museum, Bulletin 150.
• BECKE, Louis, 1897. Wild Life in Southern Seas. London, T. Fisher Unwin.
• BENNETT, George, 1831. “A Recent Visit to Several of the Polynesian Islands.” United Service Journal and Naval and Military Magazine, 31:194-202.
• BROUGHTON, William R., 1804. A Voyage of Discovery to the North Pacific Ocean . . . performed in His Majesty's Sloop Providence . . . in 1795, 1796, 1797, 1798. London, T. Cadell and W. Davies.
• BUCK, P. H., 1932. Ethnology of Manihiki and Rakahanga. Honolulu, Bishop Museum, Bulletin 99.
• — — 1938. Vikings of the Sunrise. New York, Frederick A. Stokes.
• BURNEY, James, 1803-17. A Chronological History of the Voyages and Discoveries in the South Sea, or Pacific Ocean, 5 vols. London, G. and W. Nicol.

• COUTO, D. do, 1612. Decada Quinta da Asia. Lisbon, Pedro Crasbeeck.

• DALRYMPLE, Alexander, 1770-71. An Historical Collection of the Several Voyages and Discoveries in the South Pacific Ocean, 2 vols. London, J. Nourse, J. Payne, P. Elmsly.
• DIXON, William, 1931. “Notes and Comments on ‘New Light on the Discovery of Australia’.” Royal Australian Historical Society, Journ. and Proc., XVII:289-330.

• EMORY, Kenneth P., 1934. Tuamotuan Stone Structures. Honolulu, Bishop Museum, Bulletin 118.
• — — 1947. Tuamotuan Religious Structures and Ceremonies. Honolulu, Bishop Museum, Bulletin 191.

• GALVANO, Antonio, 1862. The Discoveries of the World, from their First Original unto the Year of our Lord 1555. London, Hakluyt Society (Ser. I, no. 30).
• GATTY, Harold, 1958. Nature is Your Guide. London, Collins.
• GREAT BRITAIN, Colonial Office, 1957. Gilbert and Ellice Islands Colony and the Central and Southern Line Islands. Report for the years 1954 and 1955. London, H.M.S.O.
• GUPPY, H. B., 1887. The Solomon Islands and their Natives. London, Swan Sonnenschein.

• HARTZER, Fernand, 1900. Les Iles Blanches des Mers du Sud. Paris, Maison Vic et Amat.
• HENRY, Teuira, 1928. Ancient Tahiti. Honolulu, Bishop Museum, Bulletin 48.
• HERRERA TORDESILLAS, Antonio de, 1725-26. Descripcion de las Indias Ocidentales, 5 vols. Madrid, N. Rodriguez Franco.

• KOTZEBUE, Otto van, 1821. A Voyage of Discovery into the South Sea and Beering's Straits . . . 1815-18, in the ship “Rurick”, 3 vols. London, Longman, Hurst, Rees, Orme and Brown.

• MARKHAM, Sir Clements, 1904. The Voyages of Pedro Fernandez de Quiros, 1595-1606, 2 vols. London, Hakluyt Society (Ser. II, nos. 14-15).
• MORISON, Samuel Eliot, 1944. “Historical Notes on the Gilbert and Marshall Islands.” American Neptune, IV:87-118.

- 322

• NUNN, George E., 1934. “Magellan's Route in the Pacific.” The Geographical Review, XXIV:615-633.

• OGDEN, A., and SLUITER, E., (eds.), 1945. Greater America: Essays in Honor of Herbert Eugene Bolton. Berkeley, University of California Press.

• Pacific Islands Pilot, 1931. Vol. III, The Eastern Groups . . ., 6th ed. London, Admiralty Hydrographic Department.
• — — 1943. Vol. II, The Central Groups, 7th ed. London, Admiralty Hydrographic Department.
• — — 1946. Vol. III, The Eastern Groups . . ., 7th ed. London, Admiralty Hydrographic Department.

• RICKARD, T. A., 1934. “Drift Iron: a Fortuitous Factor in Primitive Culture.” Geographical Review, XXIV:525-543.
• ROYAL GEOGRAPHICAL SOCIETY, 1902. “Report on the Identification of the Bay of San Felipe and Santiago visited by Quiros in 1606.” Geographical Journal, XX:201-207.

• Sailing Directions for the Pacific Islands, 1952. Vol. III, Eastern Groups, 6th ed. Washington, D.C., Hydrographic Office.
• ST. JOHN, Harold, and FOSBERG, F. Raymond, 1937. Vegetation of Flint Island, Central Pacific. Honolulu, Bishop Museum, Occasional Papers Vol. XII, no. 24.
• SOLVER, C. V. and MARCUS, G. J., 1958. “Dead Reckoning and the Ocean Voyages of the Past.” Mariner's Mirror, 44:18-34.
• STANLEY OF ALDERLEY, Lord, 1874. The First Voyage Round the World, by Magellan. London, Hakluyt Society (Ser. I, no. 52).
• STEVENS, Henry N. and BARWICK, George F., 1930. New Light on the Discovery of Australia as Revealed by the Journal of Captain Don Diego de Prado y Tovar. London, Hakluyt Society (Ser. II, no. 64).

• WALLIS, H. M., 1954. The Exploration of the South Sea, 1519 to 1644. Oxford University, unpublished Ph.D. thesis.
• WHARTON, Sir W. J. L., 1902. “Note on the Identification of Le Sagittaria of Quiros.” The Geographical Journal, XX:207-209.
• WOODFORD, C. M., 1888. “Exploration of the Solomon Islands.” Proceedings of the Royal Geographical Society, X:351-376.

- 323

While the weather on the day of sighting is not mentioned, if the island was first seen “on the starboard to the south-west” and the ship sailed on a course SW¼S, and closed the island, it is obvious that the wind was favourable. It could have been from the north, through east to south-east. Probably winds were light and variable as is common in January, especially before westerlies. 95 It may reasonably be assumed, therefore, that she did sail past the island, and that it was observed from the eastward and westward. Furthermore, Almiranta must have sailed around the south side. This fact is important.

By having sailed partly round the island, Mendana and his shipmates would have been able to make a fairly accurate estimate of its circumference. The various narratives give the following estimates—

• (i) not more than 6 leagues . . . 20½ miles.
• (ii) 5 or 6 leagues at the most . . . 17 to 20½ miles.
• (iii) about 6 leagues . . . about 20½ miles.

These estimates must be treated with some reserve, since it is difficult to judge the length of low islands.

In considering identifying factors, latitude comes first. Across the whole Pacific, east of the Solomons, the only islands between the parallels of 6° and 8° south are Eiao, Hatutu and the coral islets in the Marquesas, and the central islands of the Ellice Group.

The former need not be considered, being in 140° west longitude. The latter comprise the islands of Niutao (6° 7′ S), Nanumanga (6° 18′ S), Nui (7° 16′ S), Vaitupu (° 28′ S) and Nukufetau (8° S).

Of these, Niutao and Nanumanga may be rejected; each is only 1½ miles long and the descriptions fit neither. Vaitupu could be considered—it is 3 miles long, but is compact and well-wooded, and hence has no copse in the middle. It is doubtful if it presented any very different appearance in 1568.

Nui and Nukufetau are the only islands to merit serious consideration, so their features must be compared with the narrative descriptions and with each other.

(a) Size

• Island of Jesus: estimated circumference—about 20 miles.
• Nukufetau: actual circumference—about 25 miles.
• Niu: actual circumference 96 —about 8 miles

- 324

• (i) Gallego I—6° 45′ S.
• (ii) Gallego II—7° 30′ S.
• (iii) Catoira—6° 00′ S (bare).
• (iv) Mendana—7° 00′ S (bare).
• (v) Anonymous—7° 00′ S (full).

In regard to (ii), the position is probably that of the ship. Assuming that the island bore NE (mag.) 12 miles, it would have been in 7° 23′ S.

• Observed average—6° 50′ S.
• Nukufetau—8° 00′ S.
• Nui—7° 16′ S.

It will be noted that Nui is only 26 miles south of the average recorded latitude, whereas Nukufetau is 70 miles to the southward. Considered on latitude alone, Nui is the obvious choice, but recorded latitude could have erred considerably.

The reference to a great reef on the north side and a lesser reef on the south side does not help greatly; in every Ellice and Gilbert island a sunken reef on the northern end extends further seaward than the reef at the southern or south-eastern extremities.

However, the accent on these reefs favours Nui, since the northern and southern sides of Nukufetau consist of reefs only. They do not “enter the sea” but are drying reefs, awash at high tide, and are 3 to 4 miles long. The sunken reef off the north end of Nui shows discoloured water for a quarter of a mile or more. The reef off the islet of Funuota on Nukufetau atoll is similar.

From the north-east and south-west, Nui appears as two larger islands separated by a cluster of small islets. This could well fit Mendana's description of “two galleys, with a copse in the middle which appears like a fleet of ships”. The eastern end of Nukufetau, 7½ miles long, gives the impression of almost continuous land. It does not have the appearance of two galleys and a fleet of ships.

On the western side, the atoll is also 7½ miles long, consisting mainly of reefs, with two small islands at the extremities and three small islets near the middle.

From the south-west, the northernmost island (Funuota) would barely be visible and would not look like a galley. An important factor in appearance is that Catoira describes the island as viewed from the eastward, and Mendana presumably from the westward, yet their descriptions are almost identical. The similarity of wording raises some doubt as to whether the descriptions apply to views from east and west, but Nui does have a similar appearance from both sides. The eastern and western sides of Nukufetau are entirely dissimilar.

There is no “strand” or “wide beach” on the west side of Nukufetau and from an offshore position the lagoon beach of the mainland (eastern side) would not be visible.

Nui has a series of sandy beaches along the entire western shore. A shallow lagoon, half a mile wide, separates the beach and an offshore - 325 barrier reef. The phrases “a wide beach beset with reefs” and “a strand lying lengthways” could well apply to the west side of Nui. They could not describe the west side of Nukufetau atoll.

Again the description fits Nui. The island is roughly pear-shaped, with land along the eastern and southern sides; a reef connects the northern and southern extremities. This gives the island a crescent shape with the appearance of a bay on the west side.

Nukufetau atoll is quadrilateral, with almost straight sides. It does not give the impression of having a bay in the middle, although such a description could loosely describe the lagoon.

Gallego “thought that he saw another island”, though to Mendana “it seemed to be nothing but the clouds”. When first raised, the islands of the Ellice Group appear as a serration, then a speckling on the horizon; later individual tree-tops appear, then the mass of palms appears like a dark line on the sea. They could not be mistaken for a low cloud except when less than a mile distant on a dark night. It is doubtful if the Almiranta got so close inshore, even in the daytime!

The reference to this sighting may thus be discounted in attempting to identify the Island of Jesus. In any case, the distance from Nui to Nanumanga (76 miles) and Nukufetau to Nui (82 miles) would mean a day's sailing if Almiranta had a fair wind; but she was actually beating against a head wind.

One factor of great importance is that all the narratives refer to the island: the singular is used in every instance, and no mention is made of an atoll or other islands or islets.

From a distance of several miles Nui appears to be an island, although it is, in fact, a small atoll. The islets between Fanutapu and Nui proper are set on a drying reef and give the impression of continuous low-lying, sparsely wooded land. Hence, the groups could easily be mistaken for a single island.

The mainland of Nukufetau is separated from the small western islands by 3½ miles of reef. The western islands are separated from each other by deep-water passages.

In sailing around the south side of Nukufetau, Mendana could not possibly consider the atoll as being a single island, whereas this impression could be gained when passing to the south of Nui.

One other factor that could be considered is the reference to a fishery, with the fact that the natives came off in seven canoes. More off-shore deep sea fishing is done at Nui than at Nukufetau. The latter has large fishing grounds in the lagoon, whereas Nui lagoon is shallow, small and enclosed by reefs: hence the necessity for pelagic fishing. Their canoes are also better in surf and rough water. Moreover, in 1568 the islets of Fale and Savave, on Nukufetau, would have had a very small population and would probably not muster many canoes—certainly not enough to give the impression of “a large fishery”.

It would appear that Mendana discovered Nui and named it the Island of Jesus.

- 326

The log states “Course N.E.”. This would probably be magnetic, therefore about 054° True, and would indicate the ship's head. The island was “sighted . . . 4 to 5 leagues ahead”. Since the tree tops on a coral island would be visible about 14 miles from a galleon's poop, the ship would have been approximately 12 miles offshore when the island was clearly visible, which coincides with Leza's estimate. A position about 12 miles 234° from Flink Point, on Butaritari, would be, say, 2° 55′ N, 172° 34′ E, in which case Leza's position erred 20 miles in latitude. This might be a reasonable error, but there could have been an alternative position. The island of Bikati, sighted ahead and distant 12 miles, would give an approximate position of 3° 02′ N, 172° 31′ E. This would reduce Leza's error to only 13 miles of latitude.

The alternative positions are shown at A and B on fig. 11. From both positions, with an are of vision of 2 leagues on the horizon, the impression would be created of an island about 6 leagues in circumference. The land at Aukiangang is bolder and more thickly wooded, but there is a big gap between Flink Point and Tukurere. From Tukurere to Bikati, the number of islands creates the impression of continuous land, partly wooded and partly flat, particularly so when the reefs are exposed at low tide.

All factors considered, it is most probable that Capitana was in position B, and the land sighted was thus Bikati Island. This is supported by the fact that the ship continued to sail to the north-east for some days. Had she been in position A, Bikati would have been sighted later; moreover, it is probable that an alteration of course would have been necessary in order to clear the island. It is likley that the mainland at Flink Point was just below the horizon.

A landfall at Little Makin may be ruled out—see position C on fig. 11. Also it would not have been possible to pass to the south and east of Butaritari. The ship could not have worked against wind and current and would have been on a lee shore—a risky procedure.

An interesting sidelight on this study is the apparent weatherliness of the Capitana. When making the passage Sydney to Butaritari in Alexa, it was customary to make easting to pass close westward of Fiji. This was to avoid being carried to the westward by the Equatorial Current, and so missing Butaritari. Alexa did so miss on one occasion; she became becalmed about 20 miles west of Tarawa on the 55th day out, drifted away, sailed through the Marshalls and then made easting again to the North Pacific. Eventually she reached Butaritari 155 days out from Sydney: this despite having been 400 miles to the eastward and windward of Espiritu. Some American schooners used to pass to windward of Fiji when bound from Sydney to Tarawa. Capitana must have been quite a handy ship with good windward qualities.

- i

- ii Page is blank