There are two names still in common use for American regions, which long antedate Columbus and most likely commemorate achievements of earlier explorers. They are Brazil and the Antilles. The former is earlier on the maps and records; but the case for Antillia, as an American pre-Columbian map item, is in some respects less complex and more obvious.
A good many decades before the New World became known as such, Antillia was recognized as a legitimate geographical feature. A comparatively late and generally familiar instance of such mention occurs in Toscanelli’s letter of 1474 to Columbus,237 recommending this island as a convenient resting point on the sea route to Cathay. Its authenticity has been questioned, notably by the venerable and learned Henry Vignaud,238 but at least some one wrote it and in it reflected the viewpoint of the time.
Nordenskiöld in his elaborate and invaluable “Periplus” declares: “As the mention of this large island, the name of which was afterwards given to the Antilles, in the portolanos of the fourteenth century, is probably owing to some vessel being storm-driven across the Atlantic (as, according to Behaim, happened to a Spanish vessel in 1414), those maps on which this island is marked must be reckoned as Americana.”239 The word “fourteenth” is probably an accidental substitute for “fifteenth.” The reference to Behaim undoubtedly means the often-quoted inscription on his globe of 1492, which avers that “1414 a ship from Spain got nighest it without being endangered.”240 This seems to record an approach rather than an actual landing. But at least it was evidently believed that Antillia had been nearly reached in that year by a vessel sailing from the Iberian Peninsula. Little distinction would then have been made between Spain and Portugal in such a reference by a non-Iberian.
Ruysch’s map of 1508 is a little more vague in its Antillia inscription as to the time of this adventure.241 He says it was discovered by the Spaniards long ago; but perhaps this means a rediscovery, for he also chronicles the refuge sought there by King Roderick in the eighth century.
Both of these representations show Antillia far in the ocean dissociated from any other land, but in the work of Peter Martyr d’Anghiera, contemporary and historian of Columbus, writing before 1511, we have an explicit identification as part of a well-known group or archipelago. He has been narrating the discovery of Cuba and Hispaniola and proceeds:
Turning, therefore, the stems of his ships toward the east, he assumed that he had found Ophir, whither Solomon’s ships sailed for gold, but, the descriptions of the cosmographers well considered, it seemeth that both these and the other islands adjoining are the islands of Antillia.242
Perhaps he meant delineations, like those we have yet to consider, and not descriptions in words; or writings concerning these islands may then have been extant which have since vanished as completely as the celebrated map of Toscanelli.
Among “the other islands adjoining” we may be sure he included that island of Beimini, or Bimini (no other than Florida), a part of which, thus marked, occurs in his accompanying map and has the distinction of owning the fabled fountain of youth and luring Ponce de Leon into romantic but futile adventure. Perhaps only one other map gives it the name Bimini; but its insular character is plain on divers maps (made before men learned better), with varying areas and under different names.
Peter Martyr was not alone in his identification of the “islands of Antillia.” Canerio’s map,243 attributed to 1502, names the large West India group “Antilhas del Rey de Castella,” though giving the name Isabella to the chief island; and another map of about the same date (anonymous)244 gives them the collective title of Antilie, though calling the Queen of the Antilles Cuba, as now. A later map,245 probably about 1518, varies the first form slightly to “Atilhas [i. e. Antilhas] de Castela” and shows also “Tera Bimini.” This is the second Bimini map above referred to.
It is true that the name Antillia, often slightly modified, was not restricted to this use but occasionally was applied in other quarters. Beside Behaim’s globe and Ruysch’s map already mentioned, a Catalan map of the fifteenth century (obviously earlier than the knowledge of the Portuguese rediscovery of Flores and Corvo)246 presents a duplicate delineation of most of the Azores, giving the supposed additional islands a quite correct slant northwestward and individual names selected impartially from divers sources. One of these is Attiaela, recalling the doubtful “Atilae” of the warning-figure inscription on the map of the Pizigani of 1367247 (Fig. 2), which may have suggested it, being applied in the same or a neighboring region. The islands remain mysterious, perhaps merely registering a free range of fancy at divers periods.
Again, at a much later time, when the exploration of the South American coast line had proceeded far enough to demonstrate the existence of a continent, some one speculated, it would seem, concerning an Antillia of the mainland. One of the maps248 in the portolan atlas in the British Museum known as Egerton MS. 2803 bears the word “Antiglia” running from north to south at a considerable distance west of the mouth of the Amazon, apparently about where would now be the southeastern part of Venezuela. Also, the world map249 in the same atlas (Fig. 8) bears “Antiglia” as a South American name, in this instance moved farther westward to the region of eastern Ecuador and neighboring territory.
But these aberrant applications of the name Antillia in its various forms were mostly late in time and probably all suggested by some novel geographical disclosures. The standard identification, as disclosed on the maps discussed below, at least from Beccario’s of 1435 to Benincasa’s of 1482, was with a great group of western islands; as was Peter Martyr’s, much later.
Naturally the origin of the word has been found a fascinating problem. Ever since Formaleoni,250 near the close of the eighteenth century, called attention to the delineation of Antillia in Bianco’s map of 1436, discussed below, as indicating some knowledge of America, there have been those to urge the claims of the suppositional lost Atlantis instead. The two island names certainly begin with “A” and utilize “t,” “l,” and “i” about equally; but “Atlantis” comes so easily out of “Atlas,” and the great mountain chain marches so conspicuously down to the sea in all early maps, that the derivation of the former may be called obvious; whereas you cannot readily or naturally turn “Atlas” into “Antillia,” and there is no evidence that any one ever did so. As to geographical items, both have been located in the great western sea; but that is true of many other lands, real or fanciful. Something has been made of the elongated quadrilateral form of Antillia; but Humboldt points out251 that in the description transmitted by Plato this outline is ascribed to a particular district in Atlantis, not to the great island as a whole, and that, even if it could be understood in the latter sense, there seems no reason why a fragment surviving the great cataclysm should repeat the configuration of Atlantis as a whole. There seems a total lack of any direct evidence, or any weighty inferential evidence, of the derivation of Antillia from Atlantis.
Humboldt, in rejecting this hypothesis, advanced another, which is picturesque and ingenious but hardly better supported.252 His choice is “Al-tin,” Arabic for “the dragon.” Undoubtedly Arabs navigated to some extent some parts of the great Sea of Darkness, and these monsters were among its generally credited terrors. The hardly decipherable inscriptions in the neighborhood of an island on the map of the Pizigani of 1367253 (Fig. 2), as we have seen (Ch. VI), seem to cite Arabic experience in proof of perils from fulvos (krakens) rising from the depths of the sea, coupling dragons with them in the same legend and illustrating it by a picture of a kraken dragging one seaman overboard from a ship in distress, while a dragon high overhead flies away with another. It is even true that Arabic tradition established a dragon on at least one island as a horrible oppression, long ago happily ended, and that another island (perhaps more than one) was known as the Island of the Dragon. But in all this there is nothing to connect dragons with Antillia, and that most hideous medieval fancy is out of all congruity with the fair and almost holy repute of this island as the place of refuge of the last Christian ante-Moorish monarch of Spain in the hour of his despair and as the new home of the seven Portuguese bishops with their following.
In passing, we may note that Antela, the version of the Laon globe hereinafter referred to, is identical with the name of that Lake Antela of northwestern Spain which is the source of the river Limia, fabled to be no other than Lethe, so that Roman soldiers drew back from it, fearing the waters of oblivion. But as yet no one has taken up the cause of Spanish Antela as the origin of the island’s name. Probably it is a mere matter of coincidence.
Humboldt admits that Antillia may be readily resolved into two Portuguese words, ante and illa (island). He even cites several parallel cases, of which Anti-bacchus will serve as an example. But he objects that such compound names have been used in comparison with other islands, not with a continent. In the present instance, however, the comparison would be with Portugal, not with all Europe, and the other member of it would be a map island which, he says, is as long as Portugal and seems curiously to borrow and copy Portugal’s general form and is arranged opposite to that kingdom far beyond the Azores across a great expanse of sea. It must be remembered that illa is the old form of ilha, found in many maps, that either would naturally be pronounced “illia,” and that you cannot say “anteillia” or “antiillia” at all rapidly without turning it almost exactly into Antillia. The “island out before,” or the “opposite island,” would be the natural interpretation. The latter seems preferable. Notwithstanding the great importance which must always be attached to any opinion of Humboldt’s, there really seems no need to let fancy range far afield when an obvious explanation faces us in the word itself and on the maps.
Nordenskiöld, practically applying his test of the presence of Antillia and arranging his materials in chronological order, heads his list of “The Oldest Maps of the New Hemisphere”254 with the anonymous map preserved in the Grand Ducal library in Weimar and credited to 1424.255 But it seems that this map does not deserve that position, for it is not entitled to the date; Humboldt, inspecting the original, made out certain fragments of words and the Roman characters for that year on a band running from south to north between the Azores and Antillia; also, in more modern ink, the date 1424 on the margin. Whatever the explanation, he was convinced of error by subsequent correspondence with the Weimar librarian and admitted that it was probably the work of Conde Freducci not earlier than 1481. Apart from all considerations of workmanship and map outlines, the use of “insule” instead of “insulle” and of “brandani” instead of “brandany” in the inscription concerning the Madeiras marks the map as almost certainly belonging to the last quarter, not the first quarter, of the fifteenth century.
The second map on Nordenskiöld’s New World list is “Becharius 1426,” a Latinization of the surname of Battista Beccario and at least not so weird a transformation as Humboldt’s “Beclario or Bedrazio.” Apparently the year of this map has not been doubted, but there is a lack of first-hand evidence that the original contains Antillia. No reproduction of this map had been published prior to the writer’s paper on St. Brendan’s Islands in the July, 1919, Geographical Review, nor, so far as is known, has its extreme western part been copied in any way. The section there reproduced, and herewith reprinted only slightly curtailed (Fig. 3), is one of several sent me in response to arrangements, made before the war, for a photograph of the map, but by some mistake the very portion that would have been conclusive was omitted, and all attempts to remedy the error have failed. But, if there were any inscription concerning recently discovered islands located as in his later map, some part of it at least would probably be seen on what I have; and for this and other reasons I do not believe that Antillia is delineated or named on the Beccario map of 1426.
The addition to fifteenth-century geography of a great group of large western islands roughly corresponding to a part of the West Indies and Florida rests mainly on the testimony of the following maps now to be discussed: Beccario 1435, Bianco 1436, Pareto 1455, Roselli 1468, Benincasa 1482, and the anonymous Weimar map probably by Freducci and dating somewhere after 1481. Of these the most complete as well as the earliest is Beccario’s256 (Fig. 20). He gives the islands the collective title of “Insulle a novo rep’te” (newly reported islands), which may refer to the discovery recorded by Behaim for 1414 or to some more recent experience. The interval would not be much greater than that between the first landing of Columbus and the narrative of Peter Martyr beginning with equivalent words. It is likely, however, that some lost map or maps preceded Beccario’s, for the artificially regular outlines of his islands, though in accord with the fashion of cartography in his time, seem rather out of keeping with a first appearance. The type had somehow fixed itself with curious minuteness and was repeated faithfully by his successors. In spite of these impossibly symmetrical details and some discrepancies as to individual direction of elongation and latitude, the fact remains that in the Atlantic there is no such great group except the Antilles and that the general correspondence is too surprising to be explained by mere accident or conjecture. Surely some mariner had visited Cuba and some of its neighbors before 1435.
This map of Beccario had been somewhat neglected, with misreading of the names, before it was taken in hand by the Italian Geographical Society and reproduced very carefully by photo-lithography. As regards the island names in particular, this eliminated some misunderstanding and confusion and made their meaning plain. Thus rendered, the map affords a convenient standard for the others, which, indeed, differ from it very little as to these “Islands of Antillia.”
This group, or more properly series—for three of them are strung out in a line—comprises the four islands Antillia, Reylla, Salvagio, and I in Mar. All these names have meaning, easy to render.
The largest and most southerly, Antillia, the “opposite island,” which I take to be no other than Cuba, is shown as an elongated, very much conventionalized parallelogram, extending from the latitude of Morocco a little south of the Strait of Gibraltar to that of northern Portugal. As Humboldt says, it is about a third as wide as it is long; and in this respect it is singularly even throughout its length. In its eastern front there are four bays, and three in its western. The intervals on each side are pretty nearly equal, and each bay is of a three-lobed form resembling an ill-divided clover leaf. In the lower end there is a broader and larger bay nearly triangular. The artificial exactness of these minute details is in keeping with the treatment on divers maps of the really well-known islands of the eastern Atlantic archipelagoes, except that the comparative smallness of a Teneriffe, a Terceira, or even a Madeira, offered less opportunity. The slant of the island is very slightly east of north, obviously quite different from the actual longitudinal direction of the even more elongated Queen of the Antilles.
Behind the lower part of Antillia, much as Jamaica is behind the eastern or lower part of Cuba, and about in similar proportions of relative area, Beccario shows a smaller but, nevertheless, considerable island, pentagonal in outline, mainly square in body, with a low westward-pointing broad-based triangular extension. He gives it the impressive name of Reylla, King Island, not ill suited to the royal beauty of that mountainous gem of the seas.
North of Antillia and nearly in line with it, but at a rather wide interval, he shows Saluagio or Salvagio (“u” and “v” being equivalent), which has the same name then long given to a wild and rocky cluster of islets between Madeira and the Canaries, that still bears it in the form Salvages. Wherever applied the name is bound to denote some form of savageness; perhaps “Savage Island” is an adequate rendering, the second word being understood. This Salvagio imitates the general form of Antillia on a reduced scale, being, nevertheless, much larger than any other island in the Atlantic south of the parallel of Ireland. Like Antillia, its eastern and western faces are provided with highly artificial bays, three in each. Its northern end is beveled upward and westward. I think this large island probably represents Florida, similarly situated to the northward of Cuba and divided from it by Florida Strait. Its area must have been nakedly conjectural, as much later maps show its line of supposed severance from the mainland to have been drawn by guesswork.
The inclined northern end of Salvagio is divided by a narrow sea belt from I in Mar, which has approximately a crescent form and a bulk not very different from that commonly ascribed at that time to Madeira. “I,” of course, stands for Insula or one of its derivatives, such as Illa, a word or initial applied or omitted at will. “Island in the Sea” is probably the true rendering, though formerly the initial and the two words were sometimes blended, as Tanmar or Danmar, to the confusion of geographers. A larger member of the Bahama group lying near the Florida coast would seem to fill the requirements, being naturally recognized as more at sea than Florida or Cuba. Great Abaco and Great Bahama are nearly contiguous and, considered together, would give nearly the required size and form; but it is not necessary to be individual in identification. Possibly Insula in Mar as drawn was meant to be symbolical and representative of the sea islands generally rather than to set forth any particular one of them.
The Roselli map of 1468,257 the property of the Hispanic Society of America, New York City, is nearly as complete as the Beccario map of 1435. It lacks only the western part of Reylla (a name here corrupted into “roella”), by the reason of the limitations of the material. These maps were generally drawn on parchment made of lambskin with the narrow neck of the skin presented toward the west, perhaps as the quarter in which unavoidable omissions were thought to do the least harm. Because of the island’s position on the very edge of the skin, its outline, although unmistakable, is faint and in a few decades of exposure of the original might have vanished altogether. This raises the question whether certain outlines, now missing but plainly called for, on other maps of the same period, have not met with the same fate. Probably this has happened. Antilia—spelled thus—is plain in name and outline; so is the island next above it, spelled Saluaega. The “I” is omitted from I in Mar, as was often done in like cases, and the words “in Mar” are uncertain, but seem as above. The island figure is correctly given by Beccario’s standard, and in general the representation of the island series is almost exactly the same. Perhaps the most discernible difference is a very slight northwestern trend given to Antillia, instead of the equally slight northeastern inclination in Beccario’s case.
The Bianco map of 1436258 (Fig. 25) was the first of the Antillia maps to attract attention in quite modern times but has suffered far worse than Roselli’s in the matter of limitation. The border of the material cuts off all but Antillia and the lower end of Salvagio, to which Bianco has given the strange name of La Man (or Mao) Satanaxio, generally translated “The Hand of Satan” but believed by Nordenskiöld to be rather a corruption of a saint’s name, perhaps that of St. Anastasio. It remains a mystery, though one hypothesis connects it with a grisly Far Eastern tale of a demon hand. The initial “S” is all that Satanaxio has in common with the names for this island on the other maps that show it; and, as nearly all of these present very slight changes from Salvagio, easily to be accounted for by carelessness or errors in copying, the latter name is fairly to be regarded as the legitimate one, while Satanaxio remains unique and grimly fanciful, perhaps to be explained another day. The most that can be said for its generally accepted meaning is that it corroborates Salvagio in so far as it intensifies savagery to diabolism. One is tempted to speculate as to whether any very cruel treatment from the natives had formed part of the experience of the visitors along that shore; but there is no known fact or assertion upon which to base such an idea. As to the delineation of the islands, it is quite evident that Bianco showed the same group as Beccario and Roselli so far as circumstances permitted; and there is no reason to believe that the islands for which he had no room would have differed from theirs in his showing, if admissible, any more than his Antillia differs; that is to say, hardly at all.
Humboldt was so impressed by this map of Bianco that he took the pains of measuring upon it the distance of Antillia from Portugal, making this about two hundred and forty leagues: an unreliable test, one would say, for the distances over the western waste of waters probably were not drawn to scale nor supposed to approach exactness. For that matter, the interval between Portugal and the Azores, as shown on maps for nearly a hundred years, was greatly underestimated, and the discrepancy becomes more glaring as the islands lie farther westward, Flores and Corvo being conspicuous examples. We should naturally expect to find the West Indies reported much nearer than they really are by anyone mapping a record of them. Perhaps the explanation lies in a disposition of cartographers to expect and allow for a great deal of nautical exaggeration in the mariners’ yarns that reached them. A careful man might come at last to believe in the existence of an island but doubt if it were really so very far away.
Pareto, 1455, has a very interesting and elaborate map259 (Fig. 21) showing Antillia, Reylla, and I in Mar (the latter without name) in the orthodox size, shape, and position, but with a great gap between Antillia and I in Mar where Salvagio should be. Very likely it was there once. Perhaps this is another case of fading away. One doubts whether the loss might not still be retrieved by more powerful magnifying glasses and close study of the significant interval. Pareto is unmistakably disclosing the same series of islands as the others. It may be that from him Roselli borrowed the inaccurate “roella” for Reylla, since Pareto is earlier in using a similar form (Roillo).
Benincasa’s map of 1482260 (Fig. 22) presents Salvagio as Saluaga, and I in Mar without name, but omits Reylla, both name and figure. The islands shown are in their accepted form and arrangement, except that Saluaga has but two bays on the western side, and his map adds a novelty in a series of names applied to the several bays, or the regions adjoining them, of the two larger islands. These names (Fig. 22) are twelve in number and seem like the fanciful work of some Portuguese who was haunted by a few Arabic sounds in addition to those of his native tongue. Several of them, like Antillia, begin with “An,” perhaps another illustration of the law of the line of least resistance. I cannot think that there is any significance in these bits of antiquated ingenuity, though, as we have seen in Chapter V, some have believed they found in them a relic of the Seven Cities legend.
The Weimar map,261 though long carefully housed, has suffered blurring and fading with some other damage in its earlier history. It is evidently a late representative of the tradition and begins to wander slightly from the accepted standard. It has been curtailed also from the beginning, like Bianco’s map of 1436, by the limitations of the border, which in this instance cuts off the lower part of Antillia, though the name is nearly intact; but enough remains to indicate a reduced relative size and a greater slant to the northeastward than on Beccario’s map. There is, of course, no room for Reylla, and there is none for I in Mar; but
Salvagio is given plainly and fully, with the letter S quite conspicuous. I cannot read more of the name on the photograph; but the Weimar librarian reads San on the original, being uncertain as to the rest. This map bears traces of local names arranged in places like those of Benincasa but fragmentary and illegible. Perhaps these names tend to show that the maps belong not only to the same period, but to the same general school of development. The other differences between this map and its predecessors are trivial. The general idea of the island series is the same so far as it is disclosed, and it is hardly to be doubted that all elements of the islands of Antillia would have been presented in the main on this map as they are by Roselli and Beccario, if there had been room to do so.
The Laon globe,262 1493, though mainly older, certainly had room enough, but it appears to have formed part of some mechanism and to have had only a secondary or incidental, and in part rather careless, application to geography. It shows two elongated islands, Antela and Salirosa, undoubtedly meant for Antillia and Salvagio. Perhaps the globe maker had at command only a somewhat defaced specimen of a map like Bianco’s or that of Weimar, showing perforce only two islands, and merely copied them, guessing at the dim names and outlines, without thinking or caring whether anything more were implied or making any farther search. This is apparently the last instance in which the larger two islands of the old group or series, marked by their traditional names or what are meant for such, appear together.
It may seem strange that certain other notable maps, for example Giraldi 1426,263 Valsequa 1439,264 and Fra Mauro 1459,265 show nothing of Antillia and its neighbors. Perhaps the makers were not interested in these far western parts of the ocean, or the narratives on which Beccario and the rest based their maps had not reached them; more likely they were skeptical and unwilling to commit themselves.
It is also true that the Antillia of Beccario and others is made to extend nearly north and south instead of east and west; that I in Mar is placed north of its greater neighbor instead of east; and that the whole chain of islands is moved into considerably more northern latitudes than the group which we suppose them to represent. Thus the eastern, or lower, end of Cuba is actually in the latitude of the lower part of the Sahara, and a point above the upper end of Florida would be in the latitude of the upper part of Morocco; whereas in the maps discussed the average location of the chain from the lower end of Antillia to the most northerly island, I in Mar, would run from the latitude of northern Morocco to that of southern France. There are slight individual differences in this matter of extension, but I believe Antillia always begins below Gibraltar and ends above northern Spain and a little below Bordeaux. But some dislocation, of course, is to be looked for in mapping exploration in an unscientific period. The changes of direction and extension are not greater than in the American coast line of Juan de la Cosa’s very important map of 1500,266 not to mention even more extravagant instances of later date; and the shifting of latitudes may partly be accounted for by ignorance of the southward dip of the isothermal lines in crossing the Atlantic westward. Thus a Portuguese sailor on reaching a far western island or shore having what seemed to him the climate and conditions of Gascony would be likely to suppose that it was really opposite Gascony, though in fact it might be more nearly opposite the Canaries; and the same cause of error would apply all down the line. Cuba is not really directly opposite Portugal but may easily have been believed so.
A more difficult question is raised by the absence of Haiti and Porto Rico from these maps, with all the more eastward Antilles. But it is possible that they may not have been visited or even seen. We can imagine an expedition that would touch Great Abaco, coast along Florida and Cuba, and visit Jamaica, returning out of sight, or with little notice, of the Haitian coast and barely passing an islet or two of the Bahamas, which, if not sufficiently commemorated in a general way by Insula in Mar, might well be disregarded. A report of such an expedition, adding that Antillia was directly opposite Portugal and of about equal size, would account fairly for the map which for half a century was faithfully repeated even in details by many different hands and evidently confidently believed in.
Unless we accept this explanation, we must assume an uncanny, almost an inspired, gift of conjecture in some one who, without basis, could imagine and depict the only array of great islands in the Atlantic. Certainly the outlines of Cuba, Jamaica, Florida, and one of the Bahamas will very well bear comparison with Scandinavia or the Hebrides and the Orkneys as given on maps of equal or even later date. Some glaring errors are to be expected in such work, as notoriously occurred in the sixteenth-century treatment of Newfoundland and Labrador. Applying the same tests and canons and making the same allowances as in these cases of distortion of undoubtedly actual lands, we may be reasonably confident that the Antillia of 1435 was really, as now, the Queen of the Antilles.