Short URL for this page:

[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

This webpage reproduces an article in
Geographical Review
Vol. 8 No. 1 (Jul. 1919), pp37‑46

The text is in the public domain.

This page has been carefully proofread
and I believe it to be free of errors.
If you find a mistake though,
please let me know!

 p37  St. Brendan's
Explorations and Islands​a

by W. H. Babcock

[With separate map, Pl. I, facing p40]

The Lismore Version of the Saint's Adventures

The fifteenth-century Book of Lismore, compiled from much older materials, tells us that St. Brenainn (evidently St. Brendan, the navigator) —

desired to leave his land and his country, his parents and his fatherland, and he urgently besought the Lord to give him a land secret, hidden, secure, delightful, separated from men. Now after he had slept on that night, he heard the voice of the angel from heaven, who said to him, "Arise, O Brenainn," saith he, "for God hath given thee what thou soughtest, even the Land of Promise" . . . and he goes alone to Sliab Daidche and he saw the mighty intolerable ocean on every side, and then he beheld the beautiful noble island, with trains of angels (rising) from it.​1

Thus far, in the rather redundant style of such literature, from the Life of Brenainn in the Lives of the Saints of this old manuscript. After a century and a half of disappearance this manuscript was accidentally discovered in 1814, in a walled-up recess, by workmen engaged on repairs.

Mr. Westropp holds​2 that this Lismore version it the "simplest and probably the earliest"; but its full-blown development of certain marvels (such as the spending of every Easter for at least five years on the back of a vast sea monster as a substitute for an island) may well awaken a question as to the validity of this conjecture.

However, the suggestion of the voyage by a dream seems likely enough, and his mood was in keeping with the anchorite enthusiasm of his time. Of course he promptly set forth to find his "promised land"; at first, in a hide-covered craft, with failure in spite of long endeavor; afterward, by advice of a holy woman, in a large wooden vessel, built in Connaught and manned by sixty religious men, with final success.

Another Version​b

Another version gives the credit of the first incitement to a purely human visitor, a friendly abbot, St. Brendan's aim being to reach an island "just under Mount Atlas." Here a holy predecessor, Mernoc by name, long vanished from among men, was believed to have hidden himself in "the first home of Adam and Eve." To all readers this was a fairly precise location for the earthly paradise. The great Atlas chain forms a conspicuous feature of medieval maps, running down to sea (as it does in  p38 reality) near Lanzarote and Fuerteventura, the innermost of the Canaries, which seem like detached, nearly submerged, summits of the range.

This narrative is longer and more detailed than that of the Book of Lismore and gives more plentiful indications of voyaging, especially toward the end, in southern seas. In its picture of volcanic fires it recalls occasional outbursts of Teneriffe and its neighbors. "They saw a hill all on fire, and the fire stood on each side of the hill like a wall, all burning." A visit is also recorded to a neighboring land, apparently continental, which the adventurers penetrated for forty days' travel to the banks of a magical river, whence they brought away "fruit and jewels." This may well be meant for Africa, obviously quite near these Fortunate Islands.

Attempts to Explain the Origin of the Brendan Narratives

It has been intimated that the narratives of "St. Brendan's Navigation" may have originated in misunderstood tales of his early sea wanderings around the coasts of Ireland seeking for a monastery site. He was successful in this at least, being best known (excepting as a discoverer) for the great religious establishment at Clonfert, not the first which he founded in the sixth century but the most widely known and the greatest.

Another explanation casts doubts upon his real existence and supposes the story of the discoveries to have arisen by confusion of language with the well-known pagan "Voyage of Bran," perhaps the earliest of the ancient Irish Imrama, or sea sagas.

It has also been said that the origin of the Brendan narratives may be found in "a ninth-century sermon elaborated up to its present form by the eleventh century."​3 A ninth-century manuscript is said to be in the Vatican library.

A Norman French Version

A Norman French translation was turned into Norman French verse by some trouvère of the court for the benefit of King Henry Beauclerc and his Queen Adelais early in the twelfth century and partly translated metrically into English for Blackwood's Magazine in 1836. It avers that the saint set sail for an

Isle beyond the sea

Where wild winds ne'er held revelry,

But fulfilled are the balmy skies

With spicy gales from Paradise;

These gales that waft the scent of flowers

That fade not, and the sunny hours

Speed on, nor night, nor shadow know.​4

They sail westward fifteen days from Ireland; then in a month's calm drift to a rock, where they find a palace with food and where Satan visits  p39 them but does no harm. They next voyage seven months, in a direction not stated, and find an island with immense sheep; but, when they are about to cook one, the island begins to sink and reveals itself as a "beast." They reach another island where the birds are repentant fallen angels. From this they journey six months to an island with a monastery founded by St. Alben. They sail thence till calm falls on them and the sea becomes like a marsh; but they reach an island where are fish made poisonous by feeding on metallic ores. A white bird warns them. They keep Pentecost on a great sea monster, remaining seven weeks. Then they journey to where the sea sleeps and cold runs through their veins. A sea serpent pursues them, breathing fire. Answering the saint's prayer, another monster fights and kills the first one. Similarly a dragon delivers them from a griffin. They see a great and bright jeweled crystal temple (probably an iceberg). They land on shores of smoke, flame, blast, and evil stench. A demon flourishes before them, flies overhead, and plunges into the sea. They find an island of flame and smoke, a mountain covered with clouds, and the entrance to hell. Beyond this they find Judas tormented. Next they find an island with a white-haired hermit, who directs them to the promised island, where another and altogether wonderful holy man awaits them, of whom more anon.

In this version, as in others, there are passages — such as the mention of extreme cold and the account of a great floating structure of crystal — which imply a northward course for their voyage in some one of its stages. So greatly was Humboldt impressed by this and by the insistence on the Isle of Sheep, which he identified with the Faroes, that he restricted in theory the saint's navigation to high latitudes.5

The Probable Basis of Fact

But it is noticeable that every version gives St. Brendan the task of finding a remote island, which was always warm and lovely, and chronicles the attainment of this delight, though he finds other delectable islands near it or by the way. The metrical description before quoted is surely explicit enough, but the Book of Lismore outdoes it in a very revel of adjectives. As though praises alone failed to satisfy the celebrant, he introduces the figure of a holy ungarmented usher — a living demonstration of the benignity of the climate. He was "without any human raiment, but all his body was full of bright white feathers like a dove or sea mew; and it was almost the speech of an angel that he had." "Vast is the light and fruitfulness of the island," he cried in welcome and launched forthwith on a prodigal expenditure of superextolling words outpoured on their new delightful home. It is all perfectly in keeping with the glow and luxuriance of sun-warmed shores and the unique airiness of his spontaneous  p40 raiment. Clearly "summer isles of Eden," and nothing that has to do with icebergs or wintry blasts, are called for in this case.

About six centuries lie between St. Brendan's experiences and the earliest writing purporting to relate them and generally accepted as to date. Doubtful manuscripts and miscellaneous allusions — also often doubtful — may lessen the gap; but at best we have several centuries bridged by tradition only, and that rather inferred than known. It seems likely that he really visited and enjoyed some remote lovely islands, removed but not very often reached from the mainland, such as could in any age have been discovered among the eastern Atlantic archipelagoes. In doing so he might well meet with surprising adventures, readily distorted and magnified; and the first tales of them would be basis enough for the florid fancy of Celtic and medieval romancers, growing in extravagance with passing generations.

The Cartographic Evidence

That he found some island or islands was certainly believed, for his name is on many maps in full confidence. But as to the particular islands thereby identified we find that conjecture had a wide range, varying in different periods and even with individual bias.

The Hereford Map of circa 1275

Probably its first appearance is on the Hereford map of 1275 or not much later,​6 the inscription being "Fortunate Insulae sex sunt Insulae Sct Brandani." It is about on the site of the Canary group, and the elliptical island Junonia is just below. The showing is uncertain and conventional; also the number six misses the mark by one; still there can be no doubt that the Canaries as a whole were intended. Concerning them Edrisi had observed,​c about 1154: "The Fortunate Islands are two in number and are in the Sea of Darkness." Perhaps he had Lanzarote and Fuerteventura, the most accessible pair, especially in mind. The surviving derivatives of the last eighth-century Beatus map also bear the inscription "Insulae Fortunate" where the Canary Islands should be, but they assert nothing of "St. Brandan." Doubtless, dimly known, they had been reputed Isles of the Blest from prehistoric times. If St. Brendan found them, he found them already the "Fortunate Isles."

A tradition long survived — perhaps survives still — in the Canary archipelago supporting this identification by the Hereford map. Thus Father Espinosa,​7 who long dwelt in Teneriffe and wrote his book there between 1580 and 1590, avers that St. Brendan and his companions spent several  p41 years in that archipelago and quotes a still earlier "calendar," date not given, as authority for their mighty works done there "in the time of the Emperor Justinian." Even as late as the eighteenth century an expedition sailed from among them for an island believed to be outside of those already known and to be the one discovered by St. Brendan.

The Dulcert Map of 1339

The second cartographical appearance of the saint's name seems to be in the portulan map​8 of Angelinus Dulcert, the Majorcan, dated 1339, where threeº islands corresponding to those now known as the Madeiras (Madeira, Porto Santo, and Las Desertas) and on the same site are labeled "Insulle Sa Brandani siue puelan." Since u was currently substituted for v, and m and n were interchangeable on these old maps, the last two words should probably be read "sive puellam." However the ending of the inscription be interpreted, there can be no doubt about St. Brendan and his title to the islands — according to Dulcert. And that this island group must be identified with Madeira and her consorts (though Madeira is named Capraria and Porto Santo is named Primaria) hardly admits of any question.

If the identification of them with the Fortunate Islands especially favored by St. Brendan were no more than a conjecture of Dulcert or some predecessor, it still had a certain plausibility from the facts of nature and the favorable report of antiquity. Strabo may have borne these islands in mind when he wrote: "the golden apples of the Hesperides, the Islands of the Blessed they speak of, which we know are still pointed out to us not far distant from the extremities of Maurusia, and opposite to Gades."​9 Apparently, too, Diodorus Siculus, writing half a century or so before the Christian era about what happened a thousand years earlier still, means Madeira by the "great island of very mild and healthful climate" and "in great part mountainous but much likewise champaign, which is the most sweet and pleasant part of all the rest";​10 whereto the Phoenicians were storm-driven after founding Cadiz and which the Etrurians coveted but the Carthaginians planned to hold for themselves. Even since those old days there has been a general recognition of Madeira's balminess and slumberous, flowery, enticing beauty.

The Map of the Pizigani of 1367

Divers maps of the fourteenth and fifteenth centuries do not contain the name of St. Brendan (it is perhaps never spelled Brendan in cartography)  p42 and hence do not count either way. But the identification of the notable map of 1367 of the brothers Pizigani​11 is the same as Dulcert's, the inscription being also given in the alternative. Like many oceanic features of this strange production it is by no means clear, but seems to read "Ysole dctur sommare sey ysole pone+le brandany." Perhaps it is to be understood as the "islands called of slumber or the islands of St. Brandan." There is at any rate no doubt about the last word or its meaning. But, as if to place the matter beyond all question, a monkish figure, generally accepted as that of the saint himself, is depicted bending over them in an attitude of benediction.

This map evidently does not copy from Dulcert, for the forms, proportions, and individual names of the islands all differ. It calls the chief island Canaria, instead of Capraria or the later Madeira, and appends a longer name, which seems like Caporizia, to what have long been known as Las Desertas, which appears greatly enlarged on it. Porto Santo is left unnamed on the map, perhaps because it lies so close to the general name of the group.

First Use of "Porto Santo" as Name of One of the Madeiras

A claim has been set up by the Portuguese that Porto Santo (Holy Port) was first applied to this island by their rediscoverers of the next century in honor of their safe arrival after peril, but this is abundantly confuted by its presence on divers fourteenth-century maps, notably the Atlante Mediceo​12 of 1351. Also the Book of the Spanish Friar,​13 dating from about the middle of that century, contains in his enumeration of islands the words "another Desierta, another Lecname, another Puerto Santo." It would seem to have been a familiar appellation about 1350 or earlier, and the suggestion naturally occurs that it may have originated in the tradition of the visit and blessing of the Irish saint. At any rate, the Portuguese can have had nothing to do with conferring it.

Animal and Bird Names of Islands

Concerning such names as Canaria, Capraria, etc., which, by reason of other associations, appear oddly out of place in this group, the more general question is raised of the tendency to apply animal and bird names to Eastern Atlantic islands. Goat, rabbit, dog, falcon, dove, wolf, and crow  p43 were applied to various islands long before the Portuguese visited the Madeiras and Azores, finding them untenanted; these names long held their ground on the maps, and some of them are in use even now. The reason for their adoption piques one's curiosity. If they could be taken as throwing any light on the fauna of these islands in 1350, they might also instruct us as to the probability of prior human occupancy or previous connection with the mainland. But, of course, in any significant instances some fancied resemblance of aspect may have suggested the name.


Madeira, meaning island of the woods or forest island, is a direct Portuguese translation from the Italian "I. de Legname" of the Atlante Mediceo and various later maps, and of the "Lecname" of the unnamed Spanish friar who tells us he was born in 1305. It is sufficiently explained by the former condition of the island, the northern part of which is said to preserve still its abundant woodland. Perhaps the modern name of Madeira (or Madera) first appears on the map of Geraldi of 1426, not very long after the rediscovery. But, with some cartographers, the Italian form of the name lingered on much later.

The Beccario Map of 1426

The alternative names, which had been given the Madeira group by Dulcert and the Pizigani, commemorating both the general fact of repose or blessedness and the delighted visit of St. Brendan, were closely blended (in what became the accepted formula) by the 1426 map of Battista Beccario, which unluckily has never been published in reproduction. Before the war, however, the writer obtained a good photograph of a part of it from Munich and herewith presents a section recording the words "Insulle fortunate santi brandany" (Pl. I). The first a of the final name may possibly be an e, having been obscured by one of the direction lines; but I think not. Beccario repeats the same inscription in his very important and now well-known map​14 of 1435, substituting "sancti" for "santi" by way of correction.

[image ALT: A blurry photograph of a portion of a medieval portolan map, crisscrossed by direction lines and showing a large swath of the mid-Atlantic Ocean, bordered on the right by the southwestern part of Europe and the northwestern part of Africa. In the middle of the photograph are three islands, above which is written 'sancti brandany'.]

Photograph of a section of a hitherto unpublished map for Battista Beccario, 1426, from the Munich Library, showing St. Brendan's Islands.

With no serious variations, this name, "The Fortunate Islands of St. Brandan" (or Brendan), is applied to Madeira and her consorts by Pareto (1455), Benincasa (1482), the anonymous Weimar map formerly attributed to 1424 but probably of about 1480 or 1490, and divers others. In several instances (the Beccario maps, for example) the words are almost  p44 as near to the most southerly pair of the Azores, next above them, as to the Madeiras below, and it is possible that the condition of special beatitude was understood as extending to the former also.

The Bianco Map of 1448

At any rate, the verdict of the fifteenth century for Madeira was by no means unanimous. The 1448 map of Bianco,​15 which is very unlike his earlier one of 1436 so far as concerns the Atlantic, was prepared after all the Azores had been found again by the Portuguese except Flores and Corvo. It shows the old familiar inaccurately north-and‑south string of the three groups of the Azores as they had come to him conventionally and traditionally, for evidently he did not dare or could not bring himself to discard them. But it also shows a slanting array of islands farther out, arranged in two groups respectively of two islands and five islands each and much more accurately presented as to location and direction than the old Italian standbys. These are quite clearly the Portuguese version, brought down to that date, of the newly rediscovered Azorean archipelago. But Bianco was obviously put to it to conjecture what islands these might be. He drew names from miscellaneous sources: in particular the largest island of the main group, corresponding to Terceira, bears the title "ya fortunat de sa. beati blandan." Nevertheless, he shows and names Madeira, Porto Santo, and Deserta in their usual places. Evidently he had given up, if he ever held, all thought of annexing St. Brendan's special blessing to them. He seems very confident of the St. Brendan's island of his slanting series, for it is drawn heavily in black and contrasts with the rather ghastly aspect of some neighbors. It has nearly the form of a Maltese cross, with long arms, but there is no reason to suppose that this has any significance.

Behaim's Globe of 1492

About the same period a Catalan map​16 of unknown author­ship, without copying details, adopted the same expedient of duplicating the Azores by adding the new slanting series. It is quite independent in details, however, omitting mention of "St. Brandan" in particular, though Ateallo (Antillia?) is given in the second group but not in the corresponding place. This may possibly indicate some confusion of Antillia with St. Brandan's Island, such as is more evident in the transfer of the traditional outline of the former to the latter, little changed, by Behaim on his globe of 1492.

As it stands, this globe undoubtedly gives an original and unique representation of St. Brandan's Island far west of the Cape Verde group and emphasizes it by showing Antillia independently in a more northern latitude  p45 and less western longitude and also of quite insignificant size and form. But Ravenstein, who made a very thorough study of the matter, tells us​17 that this globe has been twice retouched or renovated and that the only way to ascertain exactly what was originally delineated is to treat it as a palimpsest and remove the accretions. In particular, he relates the story of an expert geographer who found the draftsmen about to transpose St. Brandan's Island and Antillia; but they yielded to his protest. Of course, it is impossible to be quite certain that these map figures are such and in such place as Behaim intended or that they bear the names he gave. The presumption favors the present showing, generally accepted as authentic. It gives the saint only one island, but this a very large one, set in mid-ocean between Africa and South America.

Possibly this location may be suggested by an undefined coast line shown by Bianco's map of 1448, previously mentioned, and, like Behaim's island, set opposite the Cape Verde group. In Venetian Italian it bears an obscure inscription, which calls it an "authentic island" and is variously interpreted as saying that this coast is fifteen hundred miles long or fifteen hundred miles distant. The map of Juan de la Cosa (1500) exhibits off the coast of Brazil, and with an outline similar to Behaim's, "the island which the Portuguese found." His date is too late to have influenced Behaim, too early to have been prompted by Cabral's accidental discovery of that very year. It is more likely that he and Behaim were both acquainted with Bianco's work or that all three drew from the same report of discovery.

Later Maps

From this time on there is never more than one island for St. Brendan, but it indulges in wide wanderings. Especially as the attention of men was attracted to the more northern and western waters, the map makers shifted the island thither. Thus the map of 1544 purporting to be the work of Sebastian Cabot and probably prepared more or less under his influence,​18 places the island San Brandan not far from the scene of his father's explorations and his own. It lies well out to sea in about the latitude of the Straits of Belle Isle. The Ortelius map of 1560 repeats the showing with no great amount of change. In short, the final judgment of navigators and cartographers, before it quite vanished from the maps, made choice of the waste of the North Atlantic as its most probable hiding place. Perhaps this westward tendency in rather high latitudes may be partly responsible for the hypotheses in recent times which have taken the explorer quite across to interior North America on a missionary errand.  p46 There is certainly nothing to prohibit any one from believing them, if he can and if it pleases him.


In general review it appears likely that St. Brendan in the sixth century wandered widely over the seas in quest of some warm island, concerning which wonderful accounts had been brought to him, and found several such isles, the Madeira group receiving his special approval, according to the prevailing opinion of the fourteenth and fifteenth centuries. But this judgment of those centuries is the only item as to which we can speak with any positiveness and confidence.

The Author's Notes:

1 Anecdota Exoniensia: Lives of the Saints, from the Book of Lismore. Edited, with a translation, notes, and indices, by Whitley Stokes. Clarendon Press, Oxford, 1890. Reference on p252.

[decorative delimiter]

2 T. J. Westropp: Brasil and the Legendary Islands of the North Atlantic: Their History and Fable, Proc. Royal Irish Acad., Vol. 30, Section C, pp223‑260, Dublin, 1912. Reference on p250.

[decorative delimiter]

3 T. J. Westropp, op. cit., p229.

[decorative delimiter]

4 The Anglo-Norman Trouveres of the 12th and 13th Centuries, Blackwood's Edinburgh Mag., Vol. 39, 1836, pp806‑820: reference on p808.

[decorative delimiter]

5 Alexander von Humboldt: Examen critique de l'histoire de la géographie du nouveau continent, 5 vols., Paris, 1836‑1837; reference in Vol. 2, p166.

[decorative delimiter]

6 R. G. Benedict: The Hereford Map and the Legend of St. Brandan.

[decorative delimiter]

7 The Guanches of Tenerife: The Holy Image of Our Lady of Candelaria and the Spanish Conquest and Settlement. By the Friar Alonso de Espinosa of the Order of Preachers. Translated and edited, with notes and an introduction, by Sir Clements Markham. Hakluyt Soc. Publs., 2nd Series, Vol. 21, London, 1907. Reference on p29.

[decorative delimiter]

8 A. E. Nordenskiöld: Periplus: An Essay on the Early History of Charts and Sailing Directions, Stockholm, 1897, Pl. 8.

[decorative delimiter]

9 The Geography of Strabo. Literally translated with notes. The first six books by H. C. Hamilton, the remainder by W. Falconer. 3 vols., H. C. Bohn, London, 1854‑57. Reference in Vol. I, p226.

Thayer's Note: The passage is II.13; the link is to a more recent edition.
[decorative delimiter]

10 The Historical Library of Diodorus the Sicilian, in 15 Books. To which are added the fragments of Diodorus, and those published by H. Valesius, J. Rhodomannus, and F. Ursinus. Translated by G. Booth, Esq., in two vols., London, 1814. Reference in Vol. I, Bk. 5, Ch. 2, pp308‑309.

Thayer's Note: The passage is V.19; the link is to a more recent edition.
[decorative delimiter]

11 [E. F.] Jomard: Les monuments de la géographie, ou recueil d'anciennes cartes européennes et orientales. . . ., Paris, [1842‑62]: Pl. X, 1. Also W. H. Babcock: Early Norse Visits to North America, Smithsonian Misc. Colls., Vol. 59, No. 19, Washington, D. C., 1913, Pls. 1 and 2.

[decorative delimiter]

12 Theobald Fischer: Sammlung mittelalterlicher Welt- und Seekarten italienischen Ursprungs, 1 vol. of text and 17 portfolios containing photographs of maps, Venice, 1877‑86; reference in Portfolio V (Facsimile del Portolano Laurenziano-Gaddiano dell' anno 1351), Pl. 4.

[decorative delimiter]

13 Book of the Knowledge of All the Kingdoms, Lands, and Lordships That Are in the World, and the Arms and Devices of Each Land and Lordship, or of the Kings and Lords who Possess Them. Written by a Spanish Franciscan in the middle of the 14th century. Published for the first time with notes by Marcos Jiménez de la Espada in 1877. Translated and edited by Sir Clements Markham. Hakluyt Soc. Publs., 2nd Ser., Vol. 29, London, 1912. Reference on p29.

[decorative delimiter]

14 Gustavo Uzielli: Mappamondi, carte nautiche e portolani del medioevo e dei secoli delle grandi scoperte marittime construiti da Italiani o trovati nelle biblioteche d'Italia. Part II (pp280‑390) of "Studi Bibliografici e Biografici sulla Storia della Geografia in Italia," published on the occasion of the Second International Geographical Congress, Paris, 1875, by the Società Geografica Italiana, Rome, 1875; reference on Pl. 8 (the second edition, Rome, 1882, does not contain the plates). Also W. H. Babcock, work cited in footnote 11, Pl. 4.

[decorative delimiter]

15 Theobald Fischer, op. cit., Portfolio XI, Pls. 3 and 4.

[decorative delimiter]

16 Ibid., Portfolio XIII, pl. 5.

[decorative delimiter]

17 E. G. Ravenstein: Martin Behaim, His Life and His Globe, London, 1908, p59.

[decorative delimiter]

18 S. E. Dawson: The Voyages of the Cabots in 1497 and 1498; With an Attempt to Determine Their Landfall and to Identify Their Island of St. John, Trans. Royal Soc. of Canada, Vol. 12, Section II, 1894; map on p86. The map is also reproduced by Jomard, in the work cited in footnote 11.

Thayer's Notes:

a For an altogether different approach and different conclusions, see "The Legend of St. Brendan", Celtic Review I.135‑147.

b There are many, many medieval versions of St. Brendan's Legend, in at least ten European languages: Latin, Dutch, French, Italian, German, Occitan, Catalan, Norse, Anglo-Norman, English. Some of these were in turn translated into other languages, also in the Middle Ages. Caxton's telling of it in English, one of the later ones since first printed by him in the Golden Legend in 1484, will be found at Elfinspell (in partially modernized English).

c Al‑Idrisi, 1st Climate, 1st Section. It is, however, quite impossible from that passage to determine whether he had the Canaries in mind.

[image ALT: Valid HTML 4.01.]

Page updated: 26 Oct 18