The Auroras and other doubtful islands

Another chapter from this fascinating forgotten work Oddities: A Book of Unexplained Facts (Allan, London 1928) by R.T. Gould. The Aurora Islands group of three phantom islands was first reported in 1762 by the Spanish merchant ship Aurora while sailing from Lima to Cádiz. They are referred to in an episode in Edgar Allan Poe's novel, The Narrative of Arthur Gordon Pym of Nantucket, where Pym and his crewmates search for but fail to find them. Gould was  an admirer of the 'divine Edgar' - whom he calls ' one of the greatest and most unfortunate of all writers... the only world-figure of American literature.' The islands were last reportedly sighted in 1856, but continued to appear on maps of the South Atlantic until the 1870s.

Aurora Islands on a map circa 1800 (far left)

THE AURORAS, AND OTHER
DOUBTFUL ISLANDS

Isla Grande (see Fig. 16).

  In 1675 Antonio de la Roché, who was either the discoverer or the rediscoverer of South Georgia, fell in, so he believed, with a previously undiscovered island in the South Atlantic. He described it as "a very large and pleasant island, with a good harbour towards the Eastward", and gave its latitude as 45° S., leaving its longitude unspecified–and its position, in consequence, quite uncertain.

  It should be remembered that while seamen in all ages have been able, when out of sight of land, to find their latitude more or less accurately by means of astronomical observations, the finding of longitude at sea remained an unsolved and apparently insoluble problem until the eighteenth century was more than half gone by. It was no academic problem; it overshadowed the life of every man at sea and the safety of every ship and cargo. Scientific men and practical navigators alike found themselves baffled by it, even with the stimulus provided by various large Government rewards, of which the most famous is the £20,000 offered by the British Government in 1714, and won, after a long struggle, by John Harrison, the Yorkshire carpenter.

  It is, perhaps, in the fitness of things that our own country should have led the way in the matter. Our legislators have generally shown themselves (except, originally, in the case of the Plimsoll line) sympathetic towards the needs of British seamen–even if in some of the resulting legislation those seamen have, not altogether unreasonably, been classed with children and lunatics. In view of our predominant shipping interests, it would have been a standing reproach to this country if we had not done more than any other towards solving the problem of finding longitude at sea. Happily, we have no such blot on our 'scutcheon. Governments, as we all know, are not usually over-clever at solving problems–but what the British Government could do in the matter it did. It offered a reward, for any practical method of finding longitude at sea, far larger than that offered by any other nation–and differing yet more wi;lely from the latter in that it was actually paid. And we got our money's worth–Harrison's marine time-keeper, incalculable benefit to shipping, and the bulk of the world’s chronometer trade, which we still retain.

  But in De la Roché’s time, and for long after, there were no chronmeters, and while an adept in the art and mystery of navigation could find his latitude at sea, his longitude was a matter of guess-work. Like the Bristol merchants of whom Thackeray sings, he could scarcely tell, on sighting land, whether it were Jerusalem or Madagascar, or haply North or South Amerikee. All that he could do was to keep a reckoning, called the "dead-reckoning", of the various courses and distances run by his ship, and make such allowances as he thought most suitable for errors of steering, errors in estimating the speed, leeway, the effect of tides and currents, the variation of the compass, and the innumerable other perplexities which combine to make the way of a ship in the sea, as Solomon has acutely remarked, no less mysterious than that of a snake on a rock, or of a man with a maid.

  In short, he guessed his way across the ocean, and he might well have taken for his slogan the refrain of the once-famous coster ditty "... 'E dunno where 'e are". So late as 1750 he might still easily be as much as 10° wrong in his longitude at the end of a six weeks' voyage–in other words, if running up-Channel in thick weather he might imagine himself a hundred miles west of Land's End when he was actually off Brighton. It follows that none of the early navigators can truly be said to have made "discoveries" in the modern sense; all that they really "told the world" was that they had fallen in with new land on a certain parallel of latitude and between certain wide limits of longitude. The only way to revisit such discoveries was to get into their latitude a long way to the eastward or the westward, and run along it; and, as will be seen in the case of Bouvet Island, even this plan was not infallible in its results.

  After De la Roché, many competent navigators did their best to rediscover his "Isla Grande" without success. Disbelief in its existence, however, was a plant of slow growth. It was natural to suppose that the difficulty in finding this "large and pleasant island with a good harbour" was due solely to the uncertainty attaching to its longitude. Not until the turn of the nineteenth century did doubt grow into scepticism, and scepticism swell into unbelief, as the result of long and exhaustive searches* along the parallel of latitude 45° S. from the South American coast to the middle of the South Atlantic. It seems quite clear that "Isla Grande" never existed at all. Cartographers in general appear to have reached this conclusion–and accordingly to have expunged the island from their charts–somewhere about 1820. As Mayda did, however, it may yet reappear–possibly, in view of the Bolshevik penchant for enunciating scientific "novelties" long discarded by less progressive nations, in the next Russian chart of the South Atlantic.

  * As late as 1795 Vancouver, on his way home after rounding C. Horn, thought it worth his while to make a detour in order to search for Isla Grande.

A careful search, planned on much sounder lines, made in October, 1801 by Capt. Matthew Flinders, R.N., on his way out to Australia, did nothing to dispel the doubt surrounding the island's existence. He ran from 31° 02' S., 26° W., to 30° 34' S., 20° 28' W., and thence E.S.E., passing very close to Lindeman's position. He saw some birds and a turtle–in themselves, indications of land–but nothing else.

  In 1804, however, confirmatory evidence of the most satisfactory nature was received from an American source. Captain Galloway, of the ship Fanny, reported that he had been in sight of the island for four hours, and that it exhibited a peaked hill in the centre. He agreed with Lindeman, also, as to its latitude, but made its longitude some two degrees farther eastward–a discrepancy of no moment.

  On the other hand Mr. Long, master of the sloop Columbus, who reported sighting the island in 1809, stated that it lay somewhat northward, and nearly 9° westward, of Lindeman's position. Here is an extract from his log:

"September 22nd, 1809, at 5 p.m., saw the island of Saxonburg, bearing ESE... found it to be in the latitude of 30° 18' S., longitude 28° 20' W., or thereabout.*

  "The island of Saxonburg is about four leagues in length, NW and SE., and about 2½ miles in breadth. The NW end is a high bluff of about 70 feet perpendicular form, and runs along to the SE about 8 miles. You will see trees at about a mile and a half distance, and a sandy beach."

  * Flinders, discussing this report in his Voyage to Terra Australis (Vol. I, pp. 34, 35) notes that on Sept. 28, 1801 his ship, the Investigator, saw many birds when about 80 miles from Long's position. He accepted this, as explaining why so many ships had previously missed the island. On the other hand Purdy, in his Oriental Navigator (1816) showed that it must be in error, since it fell almost exactly upon Capt. Cook's track in “1774" (a misprint for 1776. R.T.G.)—–and Cook certainly saw no land in the vicinity.

  Whatever its true position might be, no further confirmation of the island's existence would appear necessary–yet such was shortly forthcoming. In 1816 Captain Head, of the English ship True Briton, reported that he had spent six hours in sight of the island. He described it as having the same high peak in the centre which had previously been seen by Lindeman and Galloway,† while his position for it agreed exactly with that given by the latter.

  † But not, it will be noted, by the Columbus.

  Bearing in mind the isolated position of the island (six hundred miles from the nearest land) and its remoteness from the trade-routes, one would undoubtedly think that its existence, position, and appearance had been established beyond possibility of question; yet, from the repeated searches which have since been made for it, and the soundings of over two thousand fathoms taken in the vicinity of its reported position, one is driven to conclude that, in all probability, it never had any real existence.

  One of the searchers, the notorious American sealer Benjamin Morrell (popularly known in his day as "the biggest liar in the Pacific"), gives an amusing account of his endeavours.*

  * A Narrative of Four Voyages to the South Sea, by Captain Benjamin Morrell, Junr., pp., 276, 277 (New York, 1832).

  "... On Saturday, August 23rd (1828) we were roused by the cheering cry from the masthead of 'Land ho, land ho, about six points off the starboard bow.'

  "We now had the wind from west-by-south, which permitted us to haul up for it; but after running in that direction about four hours, at the rate of eight miles an hour, our tantalizing land took a sudden start, and rose about ten degrees above the horizon. Convinced that we could never come up to it in the ordinary course of navigation, we backed and stood to the northward."

  This extract provides us with a useful clue in a complicated maze. If we regard the accounts of Lindeman, Galloway, and Head as absolutely independent, their remarkable agreement both as to the position and the appearance of "Saxemberg Island" would be inexplicable except on the supposition that all three saw the same real island.† But we are not entitled to say that they were actually independent–i.e, that Galloway knew nothing of Lindeman's account, and Head nothing of either’s–and the strong probability is that they were not. If we assume that Lindeman was originally deceived by some such cloud-effect as that encountered by Morrell, and that, unlike the American, he did not close it sufficiently to discover his mistake–and if we further assume that Galloway and Head were acquainted with the reports made by their predecessors–the matter becomes fairly clear.

  † What the Columbus saw, or thought she saw, is a mystery. She was too near, one would imagine, to be misled by a mirage, a cloud-effect, or an iceberg–which last, also, could scarcely be met with in 30° S.

  The doctrine of "expectant attention" is familiar to psychologists. Broadly speaking, if you impress upon a person that he is to look for something, and that he will probably see it–or even if he makes a suggestion of this kind to himself–the chances are that he will ultimately come to imagine that he has seen what he is looking for. This result may come about as a pure effort of imagination–or, more probably, he will unconsciously graft on to some object which he really sees the qualities and appearance of the thing which he is expecting to see.

  For example, those naval officers who were serving with the Grand Fleet in the early days of the War are familiar with the remarkable but little-known Battle of Scapa Flow, which was fought in September 1914. One afternoon sounds of gunfire came from the light cruiser Falmouth, guarding the eastern entrance of that admirable, if scarcely exhilarating, anchorage. She signalled that she had shelled and sunk a German submarine (revealed by her periscope) in the act of entering the harbour.

  As it was quite possible that the submarine might not have been sunk, and that there might be more than one about, the harbour was soon black with destroyers dashing about at full speed, and hoping either to ram the U-boats or at least confuse their aim. One is reminded of the old lady whose custom it was, during air-raids, to perambulate the streets at a jog-trot, on the theory that a moving target is notoriously difficult to hit. The remarkable scene was further enlivened, every now and then, by the discharge of a four-inch gun from some battleship which imagined that she had sighted a periscope.

  In all probability, no U-boat had actually been within several hundreds of miles of Scapa Flow on that particular occasion.* "Expectant attention" provides the key both to the Falmouth's initial error and to the succeeding developments.

  * I am not losing sight of the fact that in 1939 the Royal Oak was torpedoed and sunk, in Scapa Flow, by a U-boat.

  Similarly, with regard to the case of Saxemberg Island, it is easy to see that if Galloway and his crew, knowing of Lindeman's report and being in the vicinity of his "discovery", had been scanning the horizon hoping to see a low island with a peak in the centre, they would quite easily (and, in actual fact, willingly) have been deceived by a cloud-effect of the kind that Lindeman probably saw in the first place. Given the necessary (but not absolutely indispensable) cloud, precisely the same thing was likely to happen in Head's case–more likely, in fact, for while it is by no means certain that Galloway knew of Lindeman's report, there can scarcely be much doubt, in view of the much shorter interval, that Head knew of Galloway's. It is not suggested in any of the three accounts that the "island" was approached more closely than, say, ten miles; and, in all probability, it was on the horizon most of the time. One can see a very long way from a ship's masthead in clear weather. Assuming that Morrell's masthead height was 70 feet, which is probably not far from the truth, the cloud which he at first took for the island must have been about forty miles distant when sighted, and about ten when he discovered its true nature.

  Summarizing, then, we may say that in all probability Lindeman, Galloway, and Head were all three deceived by cloud-effects; and that the remarkable similarity in their accounts, inexplicable otherwise, is due to "expectant attention", based upon knowledge of previous reports, and producing two successive cases of unconscious plagiarism.

Bouvet Island.
  Bouvet Island, about fifteen hundred miles south-westward from the Cape of Good Hope, is a real island which was long regarded as apocryphal. It came into prominence in 1928 owing to a dispute as to its ownership between this country and Norway–amicably settled in Norway's favour. Although of exceedingly slight value as a territorial acquisition, there is one respect in which it is unique among islands. It is the most isolated spot in the whole world–a fact which anyone who cares to spend an instructive five minutes with a pair of dividers and a good globe can easily verify. Around Bouvet Island, it is possible to draw a circle of one thousand miles radius (having an area of 3,146,000 square miles, or very nearly that of Europe) which contains no other land whatever. No other point of land on the earth's surface has this peculiarity.

  Bouvet Island was discovered on January 1, 1739, by J. B. C. Bouvet de Lozier, a Frenchman who has a clear title to be regarded as the first Antarctic explorer on record. At the time an employee of the"Compagnie des Indes", he afterwards became Governor of Mauritius, and seems to have been a man of fine character. Following the very useful custom of the early navigators, he christened his discovery, from its date, "Cap de la Circoncision".

  Like De la Roché, he failed to circumnavigate his new land (he was greatly hampered by fog), and after ten days spent off-shore in weather uniformly too bad to permit of landing he quitted it under the impression that he had at last discovered a promontory of the long-sought Southern Continent. In his time, and much later, geographers believed most firmly, on a priori grounds such as the necessity for balancing the known preponderance of land in the Northern Hemisphere, in the existence of a huge Southern Continent, extending northward into quite low latitudes. For example, Staten Island (east of the Horn), the Solomon Islands, Easter Island, and Kerguelen were all in turn taken, when first discovered, for promontories of such a Continent. It was not until Cook's second voyage round the world (1772–5) that clear proof was given that such a continent, if it existed (we know now, of course, that it does) must lie southward of about lat. 60° S. It is a significant comment on the geographical knowledge of Cook's time that, shortly before he sailed, his inveterate enemy Alexander Dalrymple (afterwards the first Hydrographer of the Navy) roundly asserted his belief in a temperate, fertile, and wealthy Southern Continent, inhabited by at least fifty millions of people.*

•  * He expressed this remarkable opinion in two letters which he addressed to Lord North in 1772.

  If Bouvet had only known it (perhaps, for his own peace of mind, it was as well that he was spared the knowledge), his "Cape Circumcision", instead of being the Cape North of the Southern Continent, was the northwest extremity of a tiny island about five miles in diameter. That he should have succeeded in falling in with this microscopic and isolated spot of land in the course of a brief excursion southward of his normal course is one of the most remarkable "lucky dips" in the whole history of exploration.

  Like the skilful navigator that he was, he did his best, with the means at disposal, to fix the position of his discovery. He made its latitude about 54.° S. (it is actually 54° 26' S.) and its longitude about 9° E. of Greenwich, with a "probable error" of certainly not less than 5°.

  The greatest of all navigators, Captain James Cook, R.N., F.R.S., now takes up the tale. In 1772, commanding H.M.S. Resolution, he searched unsuccessfully for Cape Circumcision in the position assigned to it by Bouvet, and by running some five hundred miles from east to west and back again in a considerably higher latitude he showed that, if it existed, it was in all probability situated on a comparatively small island.

  In 1775, on his way homeward to civilization from the second† circumnavigation of the Antarctic, Cook resumed his search, on the standard plan of getting into the latitude of Bouvet's discovery a long way to the westward, and then running it down along the parallel. He made a careful but unavailing search along lat. 54° S. from 6° E. to 22° E. and concluded, rather unjustly, that what Bouvet had really seen was an enormous iceberg.

  † The first, technically, was made more or less independently by the Resolution's nominal consort, the Adventure, under Furneaux, in 1772-1774. She parted company during a gale (Oct. 1773), and was not seen again throughout the voyage.

  Although Cook did not know it at the time, his colleague Furneaux, H.M.S. Adventure, had made a similarly unsuccessful, but not so thorough, search for the Cape in the previous year. He had explored the region between 19° W. and 11° E., but in a latitude varying from 54° S. to 53° S., so that he might easily have passed to the northward of Bouvet's discovery without sighting it.

  As the result of the view so definitely expressed by the one man in all the world best qualified to give an opinion, the existence of Bouvet's "Cape Circumcision" became, except in France, generally discredited. But not for very long.

  In 1808 a London firm, Enderby Bros., directed two of their sealing vessels to make a search for Bouvet's land along the parallel of 54° S. from 10° W. to 14° E., thus duplicating and completing the work already performed by Cook and Furneaux. The two ships (snow Swan, Captain James Lindsay, and brig Otter, Captain Thomas Hopper) left San Sebastian, Brazil, for this purpose on August 22, 1808.

  They were completely successful. On October 6th, Lindsay sighted an island about twenty-five miles ahead, and closed it the next day. It proved to be a small island some five miles in diameter, surrounded to a great distance by closely-packed ice. He was unable to land, and left again on the 13th, having in the interval been joined by his consort.

  By his observations, the island was in lat. 54° 15' S., long. 4° 15' E. (dead-reckoning). By chronometer, this longitude came out at 6° 15' E., but (with good reason, as afterwards appeared) he preferred to trust to his dead-reckoning. He had but the one chronometer, and had no doubt already found it untrustworthy. In those days chronometers were neither so well made nor so well understood as they are now, and the lot of a merchant skipper's chronometer in a small ship with no proper stowage for it must have been far from a happy one.

  It might be thought, then, that the question was settled. But actually, as will be seen, it had barely been opened.

  Lindsay's account seems to have been very sceptically received among the cognoscenti. They probably preferred to trust Cook and Furneaux, whom no one could suspect of having an axe to grind; and Lindsay's island, by his statements, lay so close to the tracks of those two explorers that it seemed hardly credible that they should have failed to see it–if it really existed. Ergo . . .

  I have never seen any contemporary chart which gave credence to Lindsay's perfectly truthful story. A short account of the voyage was printed by Burney in 1817 as an appendix to the account of Bouvet's voyage given in Vol. V of his Voyages; and so far as I am aware, this forms the sole authority for Lindsay's doings. His original log does not appear to have been preserved.

  In the course of the next twenty years, two men claimed to have seen–and not only seen, but landed upon–Bouvet's island. One, unfortunately, was the American sealer Morrell, already mentioned in connection with Saxemberg Island. There is no doubt that he actually made several voyages in which he combined sealing with a certain amount of exploration–for example, he was the first man to discover the rich guano deposits on Ichabo Island–but the only authority for his work is his own book A Narrative of Four Foyages . . ., published at New York in 1832. And what a book it is[ Written in the style of the Watertoast Gazette, it reveals Morrell as a boastful romancer, suffering from a constitutional inability to refrain from "orating" on all occasions (particularly emergencies calling for instant action) and quite unashamed to acknowledge that it is his constant practice to fill his pages with accounts of the doings of other men, related as his personal experiences. His style alone has a strange power of casting a most convincing air of unreality over his accounts of even quite credible events. As an Antarctic explorer once expressed it to me, "Reading Morrell, you have a feeling that if he came down and told you it was raining in buckets, you would be quite safe in leaving your oilskin below."

  By his own account, Morrell anchored off the island, which he seems to have had no difficulty in finding, for two days in 1822. He gives its position as lat. 54° 15' S., long. 6° 11’ E.–in which case Cook could not have failed to sight it.

  Much better authenticated is the visit paid to the island in I825 by two more of the Enderby vessels–the sealers Sprightly, Captain George Norris, and Lively. Norris fell in with the island on December 10, 1825, and on the 16th he landed there and took formal possession in the name of King George IV–thus antedating, by a little over a century, Captain Horntvedt of the Norwegian whaler Norvegia, who hoisted the Norwegian flag there on December 1, 1927, under the impression that the island had hitherto been a "no man's land".

Fig. 17.—“Cape Circumcision" (Bouvet island) as originally charted from the data recorded in Bouvet's journal. Copied from a plate in Alexander Dalrymple's Collection of Voyages, chiefly in the Southern Atlantick Ocean (London, 1775).   Inset: Bouvet I., from a running survey by Capt. Harold Horntvedt, S.S. Norvegia, Dec. 1927.   NOTE. The linear scale of the inset is 2½ times that of the chart. It will be noticed that while Bouvet charted the W. and NW. shores of his discovery with fair accuracy as regards outline and orientation, the foggy weather then prevailing led him to exaggerate their extent considerably.

Thompson Island.
  Norris's annexation of Bouvet Island, however, is chiefly interesting as providing grist for the diplomatic mills. He has a better claim to remembrance in that he has provided a puzzle, in connection with his explorations, which was long unsolved. In the interval between December 13th and 16th, he stood away to the north-eastward, and discovered a second island, which no one had ever seen before and only one man has reported again. It was a small, low-lying island, on which the sea was breaking with great violence. Three miles south-eastward of it was a little cluster of three rocky islets, with a solitary rock three miles farther southward still. He called the island "Thompson Island", and the three islets the "Chimneys".

  There seems to be no real doubt that he honestly believed he had made these discoveries. The question is, did he really see land–and, if so, where is it situated?

  Norris has done what he could to help us; but, unfortunately, when examining the data one finds oneself wishing that he had either done a little more or a little less. He drew a chart of all his discoveries, and entered their positions into his log. Chart and log, apparently, have long been lost, but in the Admiralty archives is a contemporary copy of both, communicated by Messrs. Enderby in November 1826, soon after the Sprightly's return. And it is when one examines this document that the difficulties begin.

  Norris called the larger island, of which he took possession, "Liverpool Island", and determined its position, by his observations, as lat. 54° 15' S., long. 5° E. Incidentally, he took the trouble to circumnavigate it, and so made certain that it was an island. From the sketches which he gives of it, and in the light of modem investigation, it is obvious that his "Liverpool Island" was Bouvet Island, whose position is now known within a mile or two, and which lies about 1½° westward of the position he gives–a discrepancy of little moment.

  It would seem fairly safe to assume, then, that Thompson Island and its satellites lie somewhere to the north-eastward of Bouvet Island. But as regards their exact position the information which Norris gives is both redundant and contradictory. Summarizing the statements contained in his log (and converting his bearings from points to degrees for convenience), he tells us that Thompson Island bears 22½°, forty-five miles from Bouvet Island; and, intending to make the matter still plainer, gives us the position of each island in latitude and longitude. Unfortunately, as his chart shows and calculation supports, these positions are only twenty-six miles apart, and make Thompson Island bear 42° from Bouvet Island. Furthermore, on this showing the variation of the compass (the divergence of the N. point of the compass-needle from the true north) at Bouvet Island in 1825 would be 19½° E., whereas Lindsay, in 1808, found it to be 17° W. in the same vicinity, and Ross, in 1845, 20° W. It is true that the variation alters, in course of time, at practically every point on our globe–but the process is slow, and such a change as a double swing of 38° in 37 years is a thing unheard of. As I say, one wishes that Norris had told us either more or less; but there is absolutely no reason to doubt his good faith. He was no scientist, it is true; he seems to have been a plain blunt seaman, doing his limited best to set down explicitly, for the benefit of his employers, what he had himself seen and done.

The re-discovery of Bouvet Island.
  Norris's work seems to have met with no more general credence than Lindsay's; and the Government appears to have received the news of a new addition to our far-flung Empire with the most awful calmness, unbroken by either announcement or acknowledgment. In view of the fact that particulars of Norris's exploits had been in the hands of the Admiralty since 1826, it seems incredible that Captain James Ross, R.N., when he searched for Bouvet Island in 1843, should never have heard of either Norris or Lindsay. Such, however, is the fact.

  Like his great predecessor Cook, whom he resembled in many ways, Ross was coming back from circumnavigating the Antarctic; having, in the course of three years, performed the greatest feat of Antarctic exploration which, so far, it has been given to one man to accomplish. As a useful piece of secondary work, he intended to determine the position of Bouvet Island, "which had so often been sought in vain,… with some degree of precision". But Fate, ably assisted by inefficient staff-work in the Hydrographic Department, ordained otherwise.

  It must be premised, and it will explain a good deal that would otherwise remain obscure," that the account which Ross gives (in his book) of his search for Bouvet Island is extraordinarily unreliable. To criticize the work of one of the greatest of all Polar explorers, who was also a countryman and a brother officer, is an invidious task; but I think that in this matter there has already been enough paltering with geographical truth, particularly since a Norwegian writer, Mr. Bjame Aagaard, has endeavoured to bolster up the Norwegian claim to Bouvet Island by vehement appeals to Ross's work as the final authority.

  I do not propose to go into the matter here in much detail. Broadly speaking, Ross, writing some years after the event, seems to have compiled the account of his search from his journal, and to have jotted down the courses and distances which he believed himself to have steered from day to day without verifying them from the only reliable source–the ship's log. The positions of his ships (they were, by the way, the famous Erebus and Terror in which Franklin and his men afterwards perished) at each noon, as given in his account, are correct; much of the remainder is a tissue of inaccuracies. I verified this for myself, many years ago, by consulting the original logs and Masters' journals of the two ships, from which their actual track can be computed with considerable accuracy. In what follows, it must be understood that I speak of what Ross actually did, and not of what he says he did.

  Running eastward across the South Atlantic, he crossed the meridian of Greenwich in lat. 54° 07' S., intending to run Bouvet Island down, if it existed, on the parallel of 54° 15' S. To avoid the chance of being shipwrecked on it in the night (for the Antarctic summer was over, and the nights rapidly darkening and lengthening), he hove-to every evening until daybreak.

  Had he managed to keep to his parallel for a couple of days, he must have accomplished his intention of finding the island. But by a strange chance a shift of wind took him north-eastward just before he could have sighted it, and when he regained his original latitude the island, as we now know, was just out of sight astern. He actually passed in sight of it, but eighteen miles off, a distance at which the best look-out might well be excused (unlike the seaman who received three dozen lashes because the earth was a globe*) for failing to distinguish it from one of the many icebergs then visible.

•   * The story was told in print long ago by Capt. Basil Hall, R.N. The seaman in question was the masthead-man of a frigate stationed on the landward beam of the flagship, and instructed to report by signal as soon as she made the land. Owing to the curvature of the earth's surface, the land was seen from the flagship's masthead (which was much more elevated) before it could possibly have been seen from that of the frigate. This not being realized at the time, the latter ship's look-out received three dozen lashes for having, it was considered, neglected his duty.

  As if to add point to Ross's failure, a similar fate befell Lieut. T. E. L. Moore, R.N., H.M. hired barque Pagoda, who made an unsuccessful search for the island in 1845. Moore had been dispatched by the Admiralty to complete, by running from the Cape to Australia in a high southern latitude, that portion of the magnetic survey of the Antarctic which Ross had been compelled to leave unfinished. He had been a mate in the Terror during the whole of Ross's voyage, and had, consequently, taken part in his chief's search for the island two years earlier. Further search for it formed no part of his programme: nor was this, as sometimes stated,† enjoined by his Instructions. None the less he went out of his way, although short of time, to approach the island's charted position (then 54° 16' S., 6° 14' E.) from the NE, and to quit it on a southerly course, thus connecting the tracks of two previous searchers–Cook and Ross–and reducing the unexplored area in which, if it existed, the island must lie.

  † By more than one Antarctic historian. I was under the same impression when this book first appeared (1928) and made some scathing and quite unjustified remarks about Moore in consequence. But in 2933 I managed to find his Instructions (P.R.O. Ad.2/1538. 1844) which had previously eluded me. See Geographical Journal, April 1934.

  The Admiralty, not unnaturally, seem to have decided, in view of the non-success of Cook, Furneaux, Ross, and Moore, that some peculiar spell cast upon Bouvet Island rendered it invisible to naval officers.‡ They sought for it no more–not even when an opportunity offered itself in the course of the Challenger's great oceanographic voyage of 1874-7. As a tardy measure of justice, however, "Lindsay Island", "Bouvet or Liverpool Island", "Thompson Island", and "The Chimneys" at last appeared (in 1853) upon the Admiralty charts; although, with singular fatuity, the three last-named were charted in the exact positions reported by Norris–and in which, as had been shown by the work of the unfortunate naval officers who had searched the vicinity, it was morally certain that they could not possibly be situated. That two members of this archipelago were identical, and three of them incorrectly charted, was learnt by the Hydrographic Department in x899 and rectified, with almost Spanish promptitude, in 1917.

  ‡ This spell, happily, is now broken. Bouvet Island was seen by Lieut.-Commander J. M. Chaplin, R.N., serving in the R.R.S. Discovery, Capt. J. R. Stenhouse, on November 17, 1926–also by the C. in C., Africa Station, Vice-Admiral (now Admiral Sir) E. R. G. R. Evans, R.N., flying his flag in H.M.S. Milford (Capt. H. C. Phillips, R.N.) on February 23, 1934.

  The action taken in 1853 is significant as indicating a reversal of the general opinion, held since Cook's time, that Bouvet Island was nonexistent. It must be admitted that for such a view there was considerable justification. Barring Bouvet, its discoverer, the only people who claimed to have seen it were three sealers, undoubtedly rule-of-thumb navigators, and not above suspicion of pitching cock-and-bull stories to enhance their own importance. On the other hand four naval officers, all men of disinterested character and two bearing European reputations, had been over the same ground and seen–nothing. Yet, although two of these searches had been made since the last report of sighting the island (Norris's in 1825), opinion had at Last swung definitely from scepticism to credence. So true it is that Truth will out, even in an affidavit.

  Fortified by this moral support, and possibly assisted by the reduced inaccuracy of the island's charted position, several American seamen (a class always notorious for their rigid adherence to literal truth) were emboldened to report having sighted Bouvet Island. Such were Captain Williams, of the Golden West (1878), Captain Church, of the Delia Church (1882), and Captain Fuller, of the Francis Allen (1893). Fuller also stated that he saw Thompson Island, to the north-eastward of Bouvet Island; a remarkable feat, in that he is the only man, except Norris, who has ever claimed to have done so.

  Doubt as to the existence of Bouvet Island having been removed, the question of its position was finally set at rest by C. Chun, in the German oceanographic vessel Valdivia, Captain Krech, in 1898. With truly Teutonic thoroughness her navigator, Sachse, took her straight over the reported positions of Norris's "Liverpool Island", Morrell's "Bouvette's Island", and "Lindsay Island" (as reported by Lindsay). No land was seen in any of these positions. Continuing to steer westward, however, a small island came in sight about 3 p.m. in the afternoon of November 25, 1898–a date worth recording, since it marks the definite solution of the problem which Bouvet set geographers on January 1, 1739.

  The island proved to be small and valueless–a volcanic cone rising some 3,000 feet, partially covered by an enormous glacier, and reaching the sea as a ring-fence of precipitous cliffs. It was pentagonal in plan, and about five miles in diameter.* Its position, definitely ascertained for the first time by modern methods, was found to be 54° 26' S., 3° 24' E. Le Monnier, of the French Academy, although discredited in his day, had been perfectly right when he maintained that Cook had thrown away his chance of finding the island by starting his search too far to the eastward; and that in all probability it was situated in about 3° 30’ E.**

  * As Fig. 17 shows, Bouvet's partial survey of his discovery agrees very fairly well with those made by the Valdivia and the Norvegia, which differ very little.

  ** He maintained this view in three memoirs read to the Académie des Sciences in 1776 and 1779. His conclusions were attacked by Wales (Cook's astronomer in his Antarctic voyage) in a paper read before the Royal Society and printed, with additions, in the Introduction to Cook's Third Voyage (London, 1784).

The searches for Thompson Island.
  One piece of verification remained for the Valdivia. It was obvious that Bouvet Island was "Liverpool Island"; in consequence it followed that Thompson Island should be situated to the north-eastward of it. Sachse made his way, in foggy weather, to the position given by Norris–45 miles north-north-east from Liverpool Island–but saw nothing, and obtained a sounding of 1,270 fathoms close to this position, rendering it highly unlikely that there could be an island within ten miles. Owing to the poor visibility he did not extend his search, but this was continued in 1926 by another German research-vessel) the Meteor. In clear weather, she steamed over the then charted position of Thompson Island (53° 56' S., 4° 13' E.). No land at all could be seen, although the visibility was estimated at eight miles. A sounding of 778 fathoms testified, in conjunction with the Valdivia's previous sounding, that the island, if it existed, probably lay further eastward.

  Its existence, however, still remained an open question. The position searched by the Meteor had been accepted for the Admiralty charts (1917) in consequence of an investigation which I had then recently completed. By carefully plotting the tracks of all recorded searches and the varying "areas of good visibility from mast-head height", it became clear that there was an area, some 300 square miles in extent and including this position, which had never been examined at all: and the Meteor's search, while reducing this area considerably, did not obliterate it altogether. There was still room for Thompson Island and its dependent islets, the "Chimneys", in about 54° S., 4° 35' E.

  I published this conjecture in August, 1928†–and four months later Consul Lars Christensen of Sandeford, Norway, despatched the Norvegia to test it. She made Bouvet Island on Dec. 20, and thereafter zig-zagged for eight days between 52° and 55° S., and from 4° 35' W. to 6° 17' E. No trace of Thompson Island could be seen, although she reported steaming over the positions indicated by Norris, Fuller and myself.* Furthermore, the R.R.S. Discovery II obtained, in 1930, a number of consistently-deep soundings on all sides of my position, more or less disposing of the suggestion that, since 1825, Thompson Island might have disappeared as the result of volcanic action. It was removed from the Admiralty charts soon afterwards.††

  † In the first edition of this book.

  * I have no details of her track. I should welcome them, since her week's search embraced an area of some 40,000 square miles.

  †† See Admiralty Notice to Mariners no. 406 of 1931.

  In all probability, Norris mistook an earth-encrusted iceberg for land–an error which, in the absence of deep-sea sounding equipment, the most experienced navigator might commit. He was aware of this, but could not always avoid it–when charting a detached rock (shown by the Discovery II to be non-existent) about 5½ miles 328° from Cape Circumcision, he remarks of it “…it is cased with ice and at first we imagined it to be an iceberg...". A similar appearance was probably at the bottom of Fuller's report, also.‡

‡ See Ad. N. to M. no. 407 of 1931.

“Vigias".
  Still, the curious history of the search for Bouvet Island may well make us cautious as to removing from the charts any island, however doubtful. For example, such action would, from a common-sense point of view, have been perfectly justifiable in the case of Bouvet Island in, say, 1775 (after Cook's search) or in 1846 (after the searches made by Ross and Moore). Yet it would, as we now know, have replaced truth by error–and it might have led to a Shipwreck. There is, indeed, a good deal to be said for the point of view indicated, not entirely seriously, in that rare and cynical classic, The Bogus Surveyor.

  "The most tedious process in boat sounding is that of searching for reported rocks or those known to exist in the old charts. In nearly all cases these are extremely difficult to discover, because–if they do exist–it seldom happens that the positions are exactly correct on the old charts, and also because very frequently these dangers only have their being in the brain of some foolish old merchant skipper who has reported their supposed existence. If, after an hour or two's search in the supposed neighbourhood, no traces of the rock or shoal can be discovered, it will be advisable to accept its rather doubtful existence as a fact.§... Place the rock on your field board and obtain the necessary angle to fix its position with the station-pointers.

  § "You will, of course, be guided by the preconceived opinion of your captain on this subject." (Footnote in original.)

  "Some few surveying captains have objected to this plan, but it is evidently the best under the circumstances, and Marine surveying is, after all, one continual struggle with difficulties, which must be overcome, or the work would not go on.

  "There can be no danger to shipping in adopting this course. On the contrary, navigators, seeing the rock marked in your chart, will naturally avoid the risk of approaching it too nearly; or, should they recklessly neglect precautions (as they will do at times), you have the satisfaction of knowing that there is really no rock there to strike upon."*

  * This extract is perfectly genuine, and from a published book. The latter's full title is: "The / Bogus Surveyor" / or / A Short History of a Peculiar People. / By Whitewash, / The Surveyor's Friend. / Price one shilling. Devonport, A. H. Swiss, "Bremner" Printing Works, 111 and 112 Fore Street (n.d.).

  This plan was followed, in the charts of a century ago, and even later, to such an extent that the chart of so frequented a highway as the North Atlantic became positively peppered with "vigias”–imaginary shoals originating in a misapprehension of such phenomena as discoloured water, floating kelp, a school of porpoises, a water-logged tree-trunk, or a dead whale. Nowadays we are more sceptical, and the list of the classical "doubtful islands" has been severely curtailed in consequence. The last of note to go, "unwept, unhonoured, and unsung", were the Nimrod group, long believed, on the strength of a single vague report, to exist in about 56° 20' S., 158° 30’ W.; but at least two others of equally dubious pedigree are overdue for removal from the charts.

  One is Pagoda Rock (60° 11' S., 4° 43' E.), reported by Lt. Moore, of the Pagoda, in 1845. This rock only made its début on the Admiralty charts in 1918 (at my suggestion). The idea was to direct attention to its vicinity, and so get the question of its existence settled. Three consequent searches, all fruitless (Quest 1922, Meteor 1926, Norvegia 1928) have left no reasonable doubt that what Moore saw was, as in the case of "Thompson Island", an earth-encrusted iceberg. It is true that Moore got, or thought that he got, a 250-fathom sounding close to his "discovery"–but the Pagoda was drifting rapidly before a strong breeze, and without the help of steam, deep-sounding, in the calmest weather, was a wearisome and unreliable business.

  The other is that hoary nuisance, Emerald Island. This was reported in 1821 by the ship Atlantic, C. J. Nockells master, which only saw it on the horizon, 25 miles away! It has never been seen again, and its charted position (57° 15' S., 162° 50’ E.) coruscates with notes recording the dates of numerous unsuccessful searches. No doubt there is some excellent reason for keeping it on the Admiralty charts, but I cannot imagine what this is–in my opinion, it should never have been placed on them at all. Yet, apparently, sedet aeternumque sedebit.

Dougherty Island
  There remains, however, one long-accepted island whose existence is still not absolutely disproved, and whose story is curiously similar to that of Bouvet Island. That is Dougherty Island, which, if it exists, is the farthest of all from any inhabited land.

  In 1800 an American whaler, Captain Swain, of Nantucket, sighted what he took to be an island south-westward of Cape Horn, "covered with snow, and abounding with sea-dogs and fowl". He named it "Swain's Island", and gave its position, roughly, as lat. 59° S., long. (dead or even corrupt reckoning) 90°–100° W.

  A few years later Captain Richard Macy, also of Nantucket, sighted an island "four or five miles in extent in south latitude 59° and west longitude 91°, his ship passing near enough to see the breakers. The island abounded with sea-dogs or seals, and the water was much coloured, and thick with rock-weed". There is also a vague report of a similar island having been seen about this time by a Captain Gardiner, of Sag Harbour, on his way home from New Zealand.

  A search, dictated by commercial motives, was made for "Swain's Island" by the American vessels Anawan, Captain N. B. Palmer, and Penguin, Captain A. S. Palmer, in February-March 1830—but such accounts of their cruise as have survived are vague and confused. It was certainly unsuccessful–they appear to have explored (but with what thoroughness does not appear) the region 54°–61° S., 63°–104° W. It is quite impossible that in so short a time they could have made a thorough search all over this area. Its widely extended limits are an eloquent tribute to the respect in which they held their compatriots' ability as navigators.

  The reports of Swain and Macy attracted little attention, and seem to have been entirely forgotten; for when, ten years after the Palmers (they were brothers) had abandoned their search, an island closely resembling Macy's was reported in a position considerably further westward, it was generally regarded as a completely new discovery. There can be little doubt, however, that if the island really exists, the honour of discovering and naming it should rightly belong to Swain.

  In this connection, it is interesting to recall the account given by Morrell, in the work already mentioned, of the fate of Captain Robert Johnson, of the schooner Henry. Apart from his possible connection with Dougherty Island, it may be noted that Johnson, in 1822, made an exhaustive but futile search for the Auroras–to which group, after a digression extending over two oceans, I am now slowly returning.

  Morrell remarks:
  "... In the year 1823, Captain Robert Johnson... left New Zealand on a cruise to the south and east, in search of new lands, between the sixtieth and sixty-fifth degrees of south latitude; and as he has never been heard of since leaving New Zealand, it is very probable that he made discovery of some new island near the parallel of 60°, on which the Henry was shipwrecked .... "

  Morrell's standard of probability does not seem to be very exacting; but the suggestion is interesting–if not for its likelihood, at least for the confidence which shines through it that a capable American skipper, as Johnson undoubtedly was, could not possibly lose even a small schooner through her foundering at sea.

  In 1841 the eponym of Dougherty Island attracted the world's attention for the first and, so far as I can gather, the only time. Here is a verbatim extract from the log of Captain Dougherty, of the British whaler James Stewart:

  "May 29, 1841, at 2 a.m., saw land ahead, luffed and cleared it. It appeared an island 5 or 6 miles in length, running N.E. and S.W., with a high round bluff on the N.E. end, with low land to S.W. : between N.E. and S.W. ends there appeared a valley covered with ice and snow; we passed it within a quarter of a mile, going ten knots: lat. 59° 20’ S., long. 120° 20’ W.: the position for lat. and long. may differ a few miles by reason of not having had proper observations for several preceding and following days."

  As in the case of Saxemberg Island, this report was followed, before many years had elapsed, by a second. In 1860 Captain Keates, of the Louise of Bristol, reported having sighted an island on September 4, 1859, which he placed, by good observations, in lat. 59° 21’ S., long. 119° 7' W. He described it as round and dark-coloured, about 80 feet high, with an iceberg aground on the north-west side of it: and he based this latter assertion on the fact that although the berg was tilted so that one end rode much higher out of the water than the other, yet it remained broadside-on to the wind, in opposition to several other neighbouring bergs, which all lay with their lower ends to windward.

  This report, in spite of the island's sadly-altered appearance, was generally accepted as confirming Dougherty's, and "Dougherty Island" became a familiar landmark (in Keates' position) upon the Admiralty and other charts. The fact that Furneaux, in 1774, had passed very close to both Dougherty's and Keates' positions without sighting any land escaped notice–and as the reports of "Swain's Island" had been forgotten, attention was not directed to them, or to the fact that Ross, in 1842, had sailed over two out of the three positions which they assigned to it.

  Yet again, in the course of a warm newspaper correspondence in the Otago Daily News of 1891 as to the island's existence, Captain William Stannard, of the Cingalese (its principal defender), stated that he had sighted the island in 1886, and that Captain Whitson of the Dunedin had done so a year earlier and had observed a large number of seal frequenting it. Stannard gave for its position 59° 20’ S., 120° 18' W., which agreed most singularly with Dougherty's, and in which, with the addition of a subsequent note, "Probably lies further eastward", it was shown on the Admiralty charts until 1935.*

  * In that year, it was removed from them "by square date" (i.e. without publication of a Notice to Mariners), as presumably non-existent.

  In support of his statements, Stannard produced a sketch of the island executed by himself. If it did not exactly form irrefutable proof that Dougherty Island existed, it at least demonstrated a minor point of some importance–namely, that he was no artist. It looks like a cross between a disreputable iceberg and a "dissipated saw".

  Two years later (Feb. 26, 1893) a New Zealand sealer, Capt. White, recorded sighting a considerable number of seal in the vicinity of Dougherty Island. Writing in 1909, he rernarked:†

  † In a letter to Mr. H. J. Bull, quoted in Mr. Lars Christensen's Such is the Antarctic (London, 1935). I am indebted to Messrs. Hodder & Stoughton, his publishers, for permission to make this extract.

"Any doubt as to its existence is all nonsense. We sighted the island about 6.30 a.m . . . . having passed round three sides, there cannot be much doubt about our having seen it. It is much like Macdonald Island‡ •. . . only a little longer. We spent 2-3 hours trying to lasso and club seals, which were plentiful and rubbing against our ship's side. Distance 5-6 miles off the island . . . . We had a very fine sight at noon . . ., so the latitude can be relied on. Longitude fairly good, but may be a few miles farther east. I have given you the position in which we placed it. As for some people saying it does not exist, I would bet all the tea in China it did exist on February 26, 1893."

  ‡ 53° 02' S., 72° 32’ E. It is about 1½ miles long.

His position was 59° 48' S., 118° 40' W.–slightly southward, and rather eastward, of Stannard's. Still, whatever slight uncertainty might attach to its situation, the case for the island's existence certainly looked, at this epoch, overwhelmingly strong.

  But, unfortunately, there is another side to the picture. Dougherty Island has been searched for, during the last sixty years, more often than Bouvet Island has ever been; and with a uniformly depressing lack of success. Take the following examples:

S.S. Ruapehu, 1889• Passed five miles north of charted position.
S.S. Atorangi, 1890. Passed very close to charted position.
S.S. Mamari, 1893. Passed over charted position•
S.S. Niwaru, 1907. Passed forty miles south of charted position.

  None of these vessels saw any signs of land.

  More complete searches have also been made by other navigators, who adopted the sound old plan of running along the parallel of 59°–59½° S. between wide limits of longitude. For example, the late Capt. H. E. Greenstreet, of the S.S. Rimutaka, in a most praiseworthy attempt to dispose of the problem, executed a whole series of such unsuccessful searches, as follows:

In	1894,	from	approximately	125°	W. to	115°	W.
"	1900,	"	"	125°	"	109°	"
"	1902,	"	"	124°	"	105°	"
"	1907,	"	"	120°	"	110°	"
"	1910,	"	"	123°	"	113°	"

  Again, Scott with the Discovery, in 1904, ran along this parallel from 125° W. to 104° W. without sighting the island, and obtained a sounding of 2,588 fathoms in the Dougherty-Stannard position: while the magnetic vessel Carnegie, in December 1915, passed within three miles of the same position, and stated that from the masthead the island could have been seen, had it existed, anywhere within a radius of thirty-five miles.

  Earlier than the Carnegie, too, the Nimrod, of Shackleton's first Antarctic expedition, had passed over the charted position of the island in June 1909; but as this search was made in mid-winter it does not carry the same weight as some of the others.

  Lastly, on Dec. 25-27, 1930 the Norvegia ran along 59° 48' S. from 121° 30’ W. to 115° 30’ W. without seeing anything of the island, although the average visibility was estimated at nineteen miles. She obtained soundings of 2,335 fathoms, and upwards, in and near Capt. White's position.

  It may be added that Ross, while not specifically in search of any island, contributed in 1842 a most useful piece of work to the data supporting the theory of its non-existence. He ran along the parallel of approximately 59° 10’ S. from 117° W. to 109° W., and from 102° W. to 89° W. and onwards. Similar aid can also be obtained from the tracks of Cook and Furneaux in the vicinity, which zigzag extensively among the more regular paths of the later explorers.

  The whole question is an extraordinary puzzle, far more perplexing than that of Bouvet Island's existence was in, say, i 85o; and it is still disputable. The positive evidence is strong–the negative, if anything, stronger still. Assuredly the island cannot exist in, or anywhere near, 118°-120° W. If it exists (and we can scarcely assert definitely that it does not) it must lie closer to the position originally reported by Swain in 1800–59½° S., 100° W.–although the discrepancy in longitude between this position and those of Dougherty and Stannard is an appalling one, even to those who know the haphazard navigation of the early sealers and whalers. As a class, these have never possessed, or been likely to possess, either the instruments or the skill necessary for accurate navigation; and the whalers, in particular, have always been credited, justly or unjustly, with holding the view that they didn't care two hoots for their position so long as they had plenty of whales in sight.

  As regards the positive evidence in favour of Dougherty Island's existence the remarkable discrepancy between the account of it given by Captain Keates and by everyone else who has reported sighting it prevents our giving the former much weight–so that one of (at first sight) the best pieces of corroborative evidence must be altogether discounted. The strongest real feature of the case for the island's existence is the agreement in its length and appearance, as described by Macy, Dougherty, and Stannard. Macy, it will be remembered, said that it was "four or five miles in extent": Dougherty stated that it was "five or six miles in length, with a high round bluff on the north-east end, with low land to southwest", and Stannard, in the course of some notes accompanying his sketch, remarks, "north-east end high bluff 300 feet. South-east end very rugged, and about six miles long".*

  * White (1893) described it as being a little longer than McDonald Is.–or some two miles only, at most.

  It is of course possible that Dougherty knew of Macy's account, but this certainly was not generally accessible at the time in any work of reference. The independence of Stannard's account is, unfortunately, more open to doubt. Broadly speaking, it is a question of his word v. several circumstances pointing against him. In a letter to the Otago Daily News, he stated that he had originally been doubtful about the island's existence, as he could find no published account of it. On the other hand, for that statement to be correct he must have been singularly destitute of sailing directions, since all available information relating to the island was given, at that date, both in the Admiralty Pilots and in Findlay's Pacific Directory, and similar works. And the remarkable agreement between his position for the island and Dougherty's own suggests, in a manner difficult to gainsay, that he may have obtained it from a chart (or a volume of sailing directions) and not from his own observations. The same cavil attaches to his description of the island.

  There is, however, one more or less plausible theory to be put forward–a theory which unfortunately cuts both ways. If accepted, it certainly clears up many difficulties, but at the same time it relegates Dougherty Island to the limbo of discarded and erroneous ideas.

  The parallel of lat. 59° S. forms, in the region within which all reports of Dougherty Island lie, a rough limit for the pack-ice and icebergs constantly emanating from the Antarctic Continent. This is not, of course, an inviolable limit; but in general terms one may say that a navigator who keeps two or three degrees to the north of it is unlikely to fall in with many bergs, while one who should persist in keeping a similar distance southward of it could scarcely avoid doing so.

  Now the Antarctic icebergs frequently attain a size which, at first sight, is apt to disconcert even experienced Arctic navigators. Flat-topped bergs five or six miles in length and rising several hundreds of feet out of the water are by no means uncommon. On the other hand, they are not as plentiful as blackberries, even inside the Antarctic Circle; and the occurrence of such bergs, allowing for the wastage which goes on perpetually from the time when a berg is "calved" from its parent glacier until it disintegrates altogether, is a comparative rarity in such a latitude as 59° S.: while by the inexperienced eye of a mariner not accustomed to Antarctic ice conditions, such a berg would be taken for an island far more often than not. And that a large berg should, in the absence of near-by land, be resorted to by numbers of seal is not in the least improbable.

  It seems not unlikely, then, that what Swain, Macy, Dougherty, and
the rest saw was a tabular iceberg; not of course the same berg in each case, but (so to speak) a standard pattern of berg, which they all met in much about the same latitude because that was a likely parallel in which to encounter such a berg, and because it was about as far south as they cared to go. Running along 59° S. they were bound, sooner or later, to fall in with a berg of the kind: they were equally bound to differ in the longitude where they met with their particular specimen: and, as explained, they were not unlikely to mistake such a berg for an island. If this reasoning be acceptable, cadit quœstio.

The Aurora Islands.
  And now, at long last, to return, after this well-nigh interminable digression "de omnibus et quibusdam aliis", to the question (now, perhaps more comprehensible) of the Aurora Islands.

  It may be as well, in the first place, to give the original authority for their appearance on the charts–an authority long regarded, with justice, as unquestionable. Here is an extract from the Transactions of the Royal Hydrographical Society of Madrid, 1809.*

  * My version is based on that given by Weddell in his Voyage towards the South Pole, pp. 61–9. I have made one or two slight amendments, and added an explanatory word or two in brackets.

"The Aurora Islands

  "We do not learn that they were ever seen before the year 1762, in which they were discovered by the ship Aurora, which gave them her name. In 1790 they were likewise seen again by the ship Princess, belonging to the Royal Philippine Company, Captain Manuel de Oyarvido, who showed us his journal in Lima, and gave us some information with regard to their situation. In 1794, the corvette Atrevida, having gone purposely to situate them, practised in their immediate vicinity from the 21st to 27th of January all the necessary observations, and measured by chronometers the difference of longitude between these islands and the port of Soledad in the Maluinas. The islands are three; they are very nearly in the same meridian; the centre one is rather low, and the other two may be seen at nine leagues distance."

  Then follows a very impressive set of calculations (which I will spare the reader) as to the position of the islands. The results are as follows:

Southernmost Island	53° 15' 22" S.	47° 57' 15" W.
Centre, or Low Island	53°   2' 40" S.	47° 55' 15" W.
Northernmost, or New Island	52° 37' 24" S.	47° 43' 15" W.

  The northernmost island was apparently called New Island by the Atrevida because it had not been sighted previously–for a most simple reason.

  In justice to Don Manuel Oyarvido, as will appear later, one further extract must be made:

  "The captain of the Princess says, that to E.S.E. of the southernmost island there is a bank or shoal, at the distance of eleven miles, but the corvette Atrevida, which made various efforts to find it, could not discover it. . . ."

  In an appendix to this account, there is a detailed account of the Atrevida's exploits, presumably from the pen of her commander, Captain J. de Bustamente. One or two passages in it explain a good deal.

  Apparently he was not enamoured, as every true surveyor ought to be, of exploration for its own sake. We read:

  "We took advantage of the winds, sometimes favourable, sometimes contrary to our course, keeping in the parallel of 53½° (S.) and with prudent moderation determined to lie-to at nights. . . . In these lyings-to, we suffered the double martyrdom of losing precious time, and encountering rollings and a cold that were insufferable even to those who had just experienced the intemperance of Cape Horn."

  Then, on the 20th of January, comes the sighting of the southernmost of the Auroras.

  "At 5½ p.m. we perceived to the northward, at a great distance a dark lump, which appeared to all of us like an iceberg. Notwithstanding, we bore away for it under a press of sail; and when we were near it, we saw distinctly a great mountain in the form of a tent, divided vertically into two parts; the eastern extremity white, and the western very dark; on which latter side was a belt of snow: and we noticed some breaks in the dark streak.

  "We all agreed that this was the island: but we saw no other, and none of the circumstances agreed with those reported of the Auroras.

  "We passed within one mile of the island, coasting it on the western side; and from that point, it presented us the view of a sharp rock, trending from north to south. The southern part, constantly exposed to the freezing winds from that quarter, was covered with snow; and, falling perpendicularly on the north-west side, with winds much more temperate and moist, the land was there perfectly discoverable."

  They lay-to during the night, hoping for finer weather next day.

  "At daylight, we saw another island at a great distance, also covered with snow, but not so high as the former one. At 6h. it might be distant ten miles, to the N. by E., and the first island was seen to the S.E., distant about eight miles. At 9h. we lost sight of it (the second island); and although the wind freshened from the N.W. we went round it without result, because, the clouds not having dissipated, we could not observe the latitude at noon. We nevertheless waited, and at one o'clock had an altitude, and another at three o'clock . . . .

  "The wind was now at S.W., and we hauled to the southward, seeking in higher latitudes more favourable winds to get to the westward and make the coast of Patagonia.

  "On the 24th, at midday, we were in 55° 28' latitude S.; and as we did not meet better winds, but rougher seas and more intense colds, it was resolved to lessen the latitude, in search of more favourable weather. We stood to the northward, on the port tack, with all sail; and on the 26th, at evening, discovered to the E. ¼ N.E. a white lump, which at first appeared to us an iceberg; but its immobility soon convinced us that it was an island. It is a large rock, making in sharp pinnacles, but formed like a saddle-hill. The N.E. was covered with snow, but the southern part, being perpendicular, would not retain it. At a mile from this last point, there extended several breaking reefs, terminating in small islands. We coasted along this great rock at a regular distance, and sounded frequently, without finding bottom. On the 27th, in the morning, we had good observations of latitude and longitude . . . . "

  Such, in epitome, in Bustamente's account; and, at first sight, it reads most convincingly. But there is a caveat to be entered, which may as well be done here.

  I am convinced that Bustamente, at the start of his cruise, had, like the Bogus Surveyor's captain, a preconceived idea as to what he would find; and I think I can indicate what that idea was.

  In the reported latitude of the Aurora Islands (53° 33' S.) and some six degrees further eastward (42° 02' W.) there is an undoubted group of small rocky islets, called the Shag Rocks. It is uncertain who discovered them; but they are known to have been shown on a chart which Bellingshausen, the Russian circumnavigator of the Antarctic, bought in London in August 1819. They are quite close together–not more than a mile or two apart–and form a line of three pinnacles some 150-200 feet high, running about north and south. A shoal awash lies ten miles S.E. by E. of the southernmost of the islets.

  Now there is very little doubt that some tidings of these Shag Rocks had reached Bustamente, probably from his informant Oyarvido. The coincidence between the actual shoal 10’ S.E. by E. of the Shag Rocks and the shoal stated by the latter to lie east-south-east of his Aurora Islands "at the distance of eleven miles", is proof positive, to my mind, that Oyarvido was, so far as we know, the discoverer of the Shag Rocks, and that he took them for the islands previously reported by the Aurora because he was unable to detect the considerable difference in the two longitudes. If so, the information he gave Bustamente actually referred, though neither knew this, to the Shag Rocks; and there are several indications to that effect in Bustamente's journal.

  For example, in a passage already quoted he remarks: ". . . We all agreed that this was the island; but we saw no other . . .," clearly indicating that he expected to find more than one island in sight at a time. Again, in an earlier passage (not previously quoted) he remarks:

  "At daybreak on the 16th, we saw two large icebergs distant 5 miles to the N.E. Their pyramidal shape would not have failed to flatter our hopes if their proximity had not destroyed the illusion .... "

  Assuredly the man who wrote that was expecting to fall in with "pyramidal shaped" islands.

  Cheered, no doubt, by the thought that he had performed a useful piece of exploration, and not suspecting that he had added to the woes of seamen fighting their way round the Horn as well as to the subsequent gaiety of nations, he went his way to the comparative comfort of Patagonia, where, in company with his chief, Captain Malaspina, of the Descubierta, he executed a number of excellent surveys which have better stood the test of time. His Aurora Islands remained, encumbering the charts--but not, happily, the ocean.

  The first seaman to relieve his professional brethren of this incubus was Captain James Weddell, a man of very remarkable character, who lacked nothing but opportunity to have won fame second to none as a Polar explorer. He will never be quite forgotten, for in 1823 he accomplished the remarkable feat of beating, by nearly two hundred miles, the "furthest south" record previously established by Cook, and reaching, with two tiny and feeble ships, the unprecedented latitude of 74° 15' S. The sea which he traversed, and to which he gave the cumbrous name of "Sea of George the Fourth", is now known, most justly, as the Weddell Sea.

  Weddell spent many years sealing and exploring in the South, but he never had a chance to repeat his famous exploit. He died at forty-six; esteemed and remembered as a fine seaman, a daring explorer, and–more important than either–a truly noble character; a man who achieved magnificent results with scanty means, and who never grudged either time or trouble in promoting the safety and the prosperity of his brother sailors.

  Of this trait, his search for the Auroras in 1820 provides an excellent example. He sailed from Staten Island, off Cape Horn, on January 27th, in the brig Jane, and ran eastward along the parallel of 53° 15’ S. On Feb. 1st, at noon, he observed his position to be 52° 47' S., 48° 47' W.; only 28 miles 254°, therefore, from the Atrevida's "New Island". I continue the story in his own words.*

  *A Voyage towards the South Pole . . . (London, 1825) pp. 60-74. The book is based, of course, on Weddell's journal: but I believe this to have been revised for publication by his friend William Jerdan, editor of The Literary Gazette from 1817 to 1850.

  "At seven in the evening we had passed over the (laid down) latitude and longitude of these islands, without observing the least appearance of land. We obtained and continued in the parallel of latitude, running through the place assigned to them till we arrived in the longitude of 46°. I consider this allowance for error in longitude to be pretty ample; particularly since the Atrevida sailed from port Soledad in the Falkland Islands; from which, to the place for our investigation, was but about three days' sail: hence her common reckoning could not have erred much, and she had chronometers which should have been nearly exact. These considerations produced in my mind a degree of surprise; and I could not, at that moment, reconcile my experience with the facts which had been asserted.† I was resolved, however, not to abandon the object of my pursuit, without being fully satisfied of the truth or falsity of this geographical problem.

  † This passage is not italicized in the original.

  "It was now remarkably clear; and, from the masthead, land of common height might have been seen at the distance of eight leagues; but still, nothing of the kind was observed. We next steered S.S.E. into the latitude of 53° 17', and then W. by S., in order to get sight of the southern island; but in vain--not the smallest indication of land appeared. . . .

  "The situation for the middle island bore now S. 33° E., distant eight miles. We had a clear view of 6 or 7 leagues, but nothing like land was to be seen. The only chance now left us for finding these Auroras, I conceived, was by making various courses between the latitudes of 53° 15' and 52° 37'; and this we did* . . . .

  * See Fig. 18. It may be noted that the track there shown is not in exact agreement with this description.

  "We had thus again passed over the site of these islands to no purpose. . . . Having thus diligently searched through the supposed situation of the Auroras, I concluded that the discoverers must have been misled by appearances; I therefore considered any further cruise to be an improvident waste of time; and to the gratification of my officers and crew, directed our course to the Falkland Islands."

  Perhaps the most striking feature of Weddell's narrative, apart from its evidence of his ability and thoroughness as an explorer, is its urbanity. Surely, if hard Fate should decree that one must publicly be called a liar, one could only wish that such a man as Weddell should have the doing of it. And how staunchly he and his editor must have resisted the temptation to put the word "discoverers" into inverted commas.

  After Weddell, several other explorers engaged in the excellent but rather purposeless pursuit of "flogging the dead horse". Johnson, and also Morrell, in 1822, and Biscoe in 1830 made similar protracted and careful searches for the Auroras without success.

  But tradition dies hard. As a class, cartographers, like lexicographers, are conservative and seamen credulous. The Auroras survived, on many charts, well into the second half of the nineteenth century–although in this particular the Admiralty charts, on which they never appeared, form an honourable exception. Even Ross, who knew of Weddell's work, accepted their existence, for they appear on the South Polar chart published in his book in 1847; although this may be merely a blunder, since in the same chart he omits the Shag Rocks. And many seamen made the existence of the Auroras an article of their simple faith, fathering upon them a wonderful legend that somewhere on their shores reposed the wreck of a Spanish galleon, offering untold wealth to its fortunate discoverer.

  For this "foc'sle yarn" there is a discoverable, if flimsy, foundation. In October 1819 a Spanish three-decker, the San Telmo, was lost with all hands on the South Shetland islands. Remains of her timbers, but no records or traces of survivors, were found on some of the islands by the sealers. William Smith, who had discovered the South Shetlands only a few months before her loss, possessed, and perhaps now inhabits, a coffin made from such timbers.
•

  The silent tragedy of the San Telmo, garbled and distorted by rumour, became, by an easy association of ideas, located at the Aurora Islands. And the treasure, of course, was an inevitable addition. For many people, such as those who persistently subscribe money for diving operations in Tobermory Bay, the ideas "Spanish galleon" and “treasure” are indissolubly associated. With reference to the "treasure-ship Florentia", sunk some fifty feet in the Tobermory sands, and devoutly believed in by the late Duke of Argyll and many humbler mortals, it may be pertinent to point out that she is quite certainly not the Florentia, which got safely back to Spain, but the San Juan Bautista; and that, in all human probability, there never was an ounce of treasure on board of her.

  In 1838 the mystery surrounding the Auroras attracted the attention of one of the greatest and most unfortunate of all writers, Edgar Allan Poe (pace Mr. H. L. Mencken* and other wielders of the critical broad-axe, who take a perverse pleasure in bespattering the only world-figure of American literature). He embodied it, not without some artistic shaping of the actual facts, in his extraordinary story Arthur Gordon Pym, the only long story–except the similarly-unfinished Journal of Julius Rodman–which he ever wrote. He seems to have mainly relied for his facts upon Morrell's book; but probably he also obtained a good deal of information from one J. N. Reynolds, a pertinacious advocate of American Antarctic exploration who became, as much as any man could become, Poe's intimate friend–one for whom he called vainly in his last hours.

  * Still, I can forgive Mr. Mencken much of his petulant mud-sligning for the sake of that slim book of essays which he has cheerfully entitled Damn: A Book of Calumny (New York, 1918).

  Another "sighting" of the Auroras took place in 1856. Here is an extract from the log of the Helen Baird.

  "1856, Dec. 6th.–Moderate breeze and foggy; at 4h. a.m. the chief mate reported icebergs to leeward (eastward); going on deck, pronounced them to be the Auroras covered with snow; two in sight (a large and small one) bearing east (true) distant eighteen miles; at 6h. a.m. more clear, saw two others to the northward–altogether five† islands; at 8h. a.m. the northern island was east 12 miles, with the N.W. and north parts clear of snow–the south covered, the top being flat, with snow; the south end higher than the north. By meridian altitude and chronometric observation the north island is in lat. 52° 40’ S., long. 48° 22' W.–and the whole may be about twenty to twenty-five miles in extent from north to south."

  † It is a curious coincidence that Purdy's Oriental Navigator, published in 1816–when the existence of the Auroras was unquestioned–remarks (p. 21):

  “16. Aurora Isles. Five small isles, distinguished by this name, are exhibited in the late Spanish charts . . . ."

  But by this date the tide had turned, and the fictitious existence of the Auroras was nearing its close. The captain of the Helen Baird (who seems to have possessed his full share of dogmatism and self-importance) might say what he pleased–he was preaching, in general, to deaf ears. Rosser and Imray, however, in their South Atlantic Directory, published in 1870, paid him the compliment of examining his report, in conjunction with all the former data, most seriously; and went out of their way to obtain particulars of the tracks of many vessels which had recently traversed the site of the islands. These routes they embodied in a chart in which the unfortunate Auroras are pierced and transfixed, by tracks of damning accuracy, in a manner which brings to mind the pictures showing the martyrdom of Saint Sebastian.

  Still, the tradition of the Aurora Islands long survived their removal from the charts: they were once more reported–finally, let us hope–in 1892. Here is an extract from the log of the barque Gladys, Capt. B. H. Hatfield, from Iquique to Hamburg, June 26, 1892.

  "Lat. 52° 55' S., long. 49° 10’ W., land was reported on the lee quarter about 12 or 15 miles distant; it appeared like a long island extending from N. 11° E. to S. 11° W., true, about 10 or 12 miles long, with two hummocks rising up from the top, dividing the island into thirds, which would appear like three islands . . . . It was free from any snow . . . . At 8 a.m. discovered another island . . . distance about ten miles. The part which I saw presented a bold bluff appearance of moderate height, and flat on top, slightly rising towards the south, with greyish appearance, free from snow . . . . It looked as if there might be a passage between this island and the long island . . . . Mu h reliance cannot be put on my reckoning, as I had no observations for several days."

  In a subsequent letter to the Journal of Commerce* (Oct. 10, 1892), Capt. Hatfield remarked:

  *Liverpool Journal of Commerce, Oct. 3, 1892.

  "I think those islands which I discovered would about tally with the position of the Auroras, which have been so many years unsuccessfully searched for. Captain James Weddell searched for them in 1820, and declared they did not exist. That man went in the middle of the long summer days, when they were covered with snow and ice, and declared them icebergs;† but in the dead of winter, when I saw them, such places are apt to be free from ice and snow."

  † See "Icebergs in the Southern Ocean", by Wm. Allingham (Nautical Magazine, vol. for 1893).

  Unfortunately, contemporary investigation showed conclusively that what the Gladys had actually seen, and mistaken for land, was a couple of very large tabular icebergs.

  What the Atrevida actually saw still remains, and will probably always remain, a puzzle. From his journal, her captain seems to have been perfectly capable of distinguishing between an iceberg and an island; he states emphatically that he saw naked land; and apparently he was by no means without opportunities of making accurate observations. Weddell suggests that what he actually saw was, in the case of the first and third islands, the Shag Rocks–their appearance altered by pack-ice adhering to them, over which their summits would show black–and that the second island was an iceberg encrusted with earth, similar to one which he had himself encountered farther to the south. But while Weddell's views are entitled to considerable weight, his solution is not altogether satisfactory.

  Unless we conclude that Bustamente and all his officers were incompetent to navigate a barge from Portsmouth to Spithead–in which case they were scarcely like to find employment as marine surveyors–we must assume that they were at least capable of determining their position (given such chances of observation as they had) within, say, five minutes of latitude and half a degree of longitude. In such a case, it is very difficult to believe that, if he saw the Shag Rocks, he could have been so grossly in error as six degrees of longitude; and it is simply inconceivable that his first and third islands, differing in latitude some 38', could both have been the Shag Rocks. Nor does his account suggest that the land which he saw resembled the Shag Rocks much in appearance; while in view of his emphatic assertions as to the nature of what was seen, which was certainly unlike any iceberg, Weddell's theory as to the origin of the second island seems also untenable.

  As suggested, it is possible that Bustamente had at the back of his mind a preconceived idea, inspired by what Oyarvido had told him of the Shag Rocks; and a psychologist might perhaps regard his Aurora Islands as a distorted Brocken-spectre of the Shag Rocks themselves, preserving their number and relative situation, but transferred bodily to the westward, and hugely enlarged both as to their size and as to their distance apart. That he could subconsciously put together, piecemeal, this dream of his and select, in the course of a voyage of considerable range and very varied direction, such appearances–icebergs or whatever else they may have been–as would bring it on to the chart, is, as the American papers used to say, "remarkable–if true": yet I put it forward, with diffidence, as the only attempt at an explanation which I can offer.

  "And now," in the manner of Herodotus, "let it suffice to have said this much about the Auroras and all other doubtful islands."