Rangel’s Island and the Missing Horizon: The Conceição wreck and the retrospective identification of Peros Banhos

Research Paper | Pen for Rights – Research Papers
Published by Pen for Rights – Republic of Maldives. A PDF version is available for download.

Abstract

The Portuguese nau Conceição was wrecked on 22 August 1555. The 1735 relation of Manoel Rangel’s account names the site as the baixos de Pêro dos Banhos (shoals of Pêro dos Banhos). Later writers identify these shoals with the modern atoll of Peros Banhos in Folhlhavai (Chagos Archipelago). That equation is retrospective. On the day before the wreck, Rangel places the ship six degrees from the Line. He gives no longitude. A latitude near six degrees south fits several island groups east of Madagascar. His description of the wreck islet mentions no coconut palms. He reports no other land in view. The coconuts, huts, and people appear only at a later landfall, reached after more than a month at sea. Every legible printed European map in the register from 1596 to 1851 keeps the baixos de Pêro dos Banhos apart from the Maldivian names Candú and Polvara. No candidate in Folhlhavai fits the combined textual, cartographic, and geographic evidence. The attribution of the wreck to Folhlhavai is a later label, not a finding from Rangel’s account. The account neither identifies the rescuers’ nationality nor fixes the wreck to a modern coordinate.

Keywords: Manoel Rangel; Conceição (1555); Peros Banhos; Folhlhavai; Candú; Polvara; Maldives; historical cartography; hydrography; Lazare Picault; Mascarene region; retrospective identification

Rangel as witness, survivor and author

Manoel Rangel is known chiefly through his relation of the wreck. Its title identifies him as someone aboard the Conceição. It also says that he reached Cochin in January 1557. Modern scholarship describes him as a passenger. No secure biographical record establishes his birth, occupation, or social rank. Rangel writes as a participant, but the title and later transmission do not supply a biography.¹

The earliest presently traceable printed version of Rangel’s relation appears in Bernardo Gomes de Brito’s História Trágico-Marítima, volume 1 (Lisbon, 1735), pages 169–217. C. R. Boxer could trace no earlier edition. A shorter relation sometimes blended with Rangel’s is a separate, later Jesuit text. It was probably composed around 1620 from second-hand material. It is not another eyewitness account.² These texts must be kept distinct. The received narrative has sometimes combined details from different stages or versions, and even from other wreck accounts in the same collection.

How the name became a modern location

The 1735 relation transmits the wreck under the name baixos de Pêro dos Banhos. Its narrative calls the island from which the survivors departed the Ilha de Pêro dos Banhos (island of Pêro dos Banhos). The name may be Rangel’s own wording, or it may have been supplied in the printed heading. Either way, it is the name under which the earliest surviving text records the loss. It is also the name against which the charts must be read. The relation gives no longitude and makes no reference to the Chagos or Folhlhavai. The clearest explicit link to present-day Peros Banhos appears in Abel Fontoura da Costa’s 1940 edition of a Portuguese roteiro. In an editorial note, Costa equates the named shoal with the group of small islands and reefs that the English call “Peros Banhos.” He also gives a modern position for Île du Coin. This is Costa’s retrospective gloss, not a coordinate from the sixteenth-century route text.³

Later writers repeat or debate Costa’s identification. James Duffy’s Shipwreck and Empire (1955) places the wreck in the Peros Banhos narrative. Paulo Guinote, Eduardo Frutuoso, and António Lopes list it in their 1998 catalogue of losses on the India route. Humberto Leitão (1958) argued that later charts transposed the relative positions assigned to Pêro dos Banhos and Chagas in sixteenth-century sailing directions. His argument makes the match between the old name and the modern atoll a question for investigation. W. Ph. Coolhaas (1960) examined the earliest Dutch notices of the same shoals. Donald Taylor’s survey of Chagos charting also warns that early records and island positions are uncertain. Nigel Wenban-Smith begins from the conventional Peros Banhos location but questions it. He proposes Nelson Island as a candidate and later returned to the wreck in a book-length study with Ullattil Manmadhan. Philippe Sands retells Peros Banhos as settled fact.⁴

The charts of the period preserve two distinct registers. Pierre Desceliers’s world map of 1550, five years before the wreck, draws islands on the Maldives ridge south of Addu. Their positions correspond to those that later maps label Candú, the Kandhoo of Maldivian usage, and Polvara, Foalhavahi (Diego Garcia). The map also draws islands in the Mascarene region. The labels are illegible in available reproductions, so the map itself does not supply those names. João de Barros describes the southern extent of the chain within the same decade. In the third Década da Ásia (1563), he traces it for three hundred leagues. He begins with Mamále, at twelve and a half degrees north. He places the last islands, Candú and Adú, at seven degrees south, “chamadas Candú & Adú, estam em sete graos da parte do sul” (called Candú and Adú, [which] lie at seven degrees on the southern side). Barros thus places Candú at the southern end of the Maldives within a decade of the wreck. He knows it by its own name, distinct from the Pêro dos Banhos of the route books.⁵ Read as a count of completed degrees from the Line, Barros’s seven degrees south covers the band from 7° to 8°. Foalhavahi (Diego Garcia), the Polvara of the charts, lies within that band. Barros’s Maldive chain therefore reaches Polvara’s latitude, although he does not name it.

The Adú that Barros names with Candú should not be assumed to be present-day Addu Atoll. Later charts set an Adu beside Candu. Houtman and Wright (1598) draw Pomoluco, Adu, Candu and I. do Poluoro as a separate chain south of the Maldives. Elwe (1792) labels “Addou et Poüé Molucque” apart from “Adu et Caudu”. Thomson (1817) still draws “Adu” and “Candy” east of the Chagos group. These charts support reading Adú as a feature distinct from Addu and associated with Candú. They do not establish that Adú was an island within Kandhoo Atoll. That identification remains a working hypothesis requiring corroboration. The same rule applies here as to Pêro dos Banhos. No historical name is identified with a modern place without independent geographical evidence.⁵

Later legible maps keep the two registers apart. On the 1598 chart of Cornelis de Houtman and Benjamin Wright, I. de Pº dos Banhos appears in a western group with S. Pº de Mascarenhas, I. de Dº Garcia, and Roquepiz. Pomoluco, Adu, Candu, and I. do Poluoro form a separate chain south of the Maldives. Hondius (1606), Blaeu (1617), Allard (1660), Loots (1707), and Valentijn (1726) place Pêro dos Banhos with the Chagas and Diego Graciosa group in the west. They place Candu, Adu, Pumoluque, and Polvara in the east. A second tradition begins with Linschoten (1596) and Langenes (1598). Gerritsz (1632), Visscher (1658), and van Keulen (1680) continue it. These maps place the Baixos de Pêro dos Banhos west of the northern Maldive atolls and north of the Equator. Across the forty-five printed European geographic maps examined from 1502 to 1851, Candú and Polvara remain in the Maldivian register. The maps keep them separate from the Pêro dos Banhos label.⁶

The maps do not place the shoal consistently. Some put it in the western Mascarene grouping. Another tradition places its label west of the northern Maldives. These are different chart placements. They do not establish a separate island group northwest of the Maldives. No map in the register associates Pêro dos Banhos with the Candú and Polvara register south of Addu. Polvara appears there from the Venetian woodcut of 1595. Candu appears from 1598. Both names recur on maps through Tallis in 1851. John Thomson’s 1817 map places “Peros Banhos or Borde I.” among the Chagos islands, beside Eagle, Danger, and Diego Garcia. It still draws “Adu” and “Candy” separately to the east. Penelope Steel (1809), James Wyld (1837), and John Tallis (1851) likewise keep Adu and Candu apart from the Chagos group.

State hydrographic records form a separate documentary sequence. The earliest presently identified French hydrographic sheets associating Peros Banhos and Chagas are two manuscripts in the French navy’s hydrographic collection, now at the Bibliothèque nationale de France. One is attributed to Lazare Picault and dated 16 April 1744. It is catalogued as Plan des isles de Peros Banhos decouverte par la tartane l’Élizabeth (plan of the islands of Peros Banhos discovered by the tartane Élizabeth). The other sheet is marked “Copié sur le Plan original” (copied from the original plan) and is dated by the library to after 1744. It is titled Plan des Iles de Chagas (plan of the islands of Chagas). Fleurieu’s signature appears on its reverse.

The sheets carry the same drawing. Both show the same arc of islands, detached bank, soundings, track, landing mark, descendu à terre (went ashore), and latitude of 5°30′ south. Their titles also follow the same formula. Each names the tartane, Captain Lazare Picault, the date of 16 April 1744, and the latitude. They differ in the island name. In the cartouche of the sheet attributed to Picault, that name has been struck through. “Peros Banhos” is written above it in a smaller hand. The struck word is hatched over and cannot be read directly in the scan examined. Its length, a descending stroke where a g would fall, and the copy’s otherwise identical title all point to Chagas. On that reading, both sheets were first titled Chagas. Peros Banhos was added to the original sheet later, by correction, at an unrecorded date. The copy also adds a longitude of 92°6′ without naming a meridian. It preserves the title as it stood before the correction.

The drawing corresponds in its principal features to the Peros Banhos plan on British Admiralty chart 725. A western chain of long islands is broken by a single pass. The plan marks a landing there, matching the Passe de Île Poule between Petite Soeur and Île Poule. A row of small islands runs along the north, corresponding to Île de la Passe, Moresby, Parasol, Longue, and the Bois Mangue islets. Two islands stand apart on the east, like the Coquillage pair. A pair at the southern end of the western chain corresponds to Île du Coin and its neighbour. A single island inside the ring lies where Île Vache Marine appears. The detached bank drawn to the south, with soundings from eleven to forty-five, is consistent with the northern edge of the Great Chagos Bank. Both sheets give a latitude of 5°30′ south, which falls at the atoll’s southern rim on chart 725. The drawing’s own scale overstates the atoll’s dimensions. The correspondence therefore rests on configuration, not measurement. On that comparison, the sheets depict the atoll now called Peros Banhos, the Candú of the printed maps. They document a history of naming in French hydrographic work. The attribution of the survey to Picault on the date given in the title is open to serious doubt.

Fauvel’s 1909 extract from Picault’s journal places him eight leagues west of the Amirantes on 16 April 1744. That position is about nineteen degrees of longitude west of the atoll shown in the plans. The journal extract makes the date and Picault attribution untenable as a record of a survey he made there that day. Fauvel’s extract is itself a third-person report of an archival abstract whose compiler is unknown. The conflict therefore rests on the best available witness to Picault’s movements, not on his own log. Fleurieu’s signature places the copy within the document’s later transmission. It does not identify him as surveyor or draughtsman. Who first surveyed the atoll, and when, remains unestablished.⁷

The Chagas sheet is a retrospective copy. It is not evidence that Picault reached Folhlhavai in 1744. The library catalogues the copy as made after 1744. Fleurieu’s signature and that catalogue date place it within the later history of French naval hydrography. Fleurieu was born in 1738. He became deputy inspector of the navy’s charts dépôt in the mid-1770s and director of ports and arsenals in 1777. If his signature dates from his service at the dépôt, the sheet may belong to the mid-1770s. That would place it before the French reasserted their claim at Diego Garcia in 1786. The reassertion followed an East India Company attempt to settle there. It would also precede the settlement initiatives of the late 1780s, when Le Normand was authorised to settle the atoll, and the copra works Lapotaire established in 1793. The precise date of the copy remains unknown.⁸

The plans remained in manuscript. In print, Thomson’s 1817 map places Peros Banhos among the Chagos islands while showing Adu and Candy separately to the east. British Admiralty charting followed. Robert Moresby surveyed the Maldives from 1834 to 1836 without the Sultan’s explicit consent. He and Powell surveyed the Chagos in 1837. The Admiralty chart of the Chagos Archipelago, published in 1839, standardised that arrangement in British hydrographic charting. The Maldivian names Candú and Polvara were thereafter dropped from the Admiralty chart series. Commercial maps such as Wyld’s and Tallis’s continued to show Adu and Candu for a time.⁹ The geographic-map tradition retained the Maldivian names. State hydrographic records reorganised the region’s navigational nomenclature. Costa’s 1940 equation projects that hydrographic arrangement back onto 1555. It joins two names that three centuries of geographic maps had kept apart.

The distinction is fundamental. The Conceição was a Portuguese vessel, and the name Pêro dos Banhos belongs to the Portuguese documentary tradition. The identification of that name with Kandhoo Atoll, however, cannot be attributed to the Portuguese navigators of 1555 on the evidence examined. João de Barros, writing in 1563, identified Candú independently within his geographical account of the southern Maldives, without equating it with the shoals of Pêro dos Banhos. The surviving cartographic evidence examined here instead points to French hydrographic manuscripts as an important stage in the reassignment of the name, followed by its consolidation in British Admiralty charting. When Abel Fontoura da Costa associated the Portuguese wreck narrative with the islands the English called Peros Banhos in 1940, he employed an identification already established in later European cartography. His retrospective identification cannot, without independent contemporary corroboration, be treated as evidence that the Portuguese themselves located the wreck at Kandhoo Atoll.

What Rangel records before the wreck

The fleet chose to sail outside the Ilha de São Lourenço (island of Saint Lawrence), east of Madagascar. It sighted the Cape of Good Hope on 18 July. On 21 August, Rangel places the ship “em seis graus da linha da Índia” (at six degrees of the Line of India). He does not say how the figure was obtained. In Portuguese navigation, the Line normally meant the Equator. The reading that best fits the voyage is a latitude of about six degrees south. The phrase does not name a hemisphere, however, and the qualifier “da Índia” (of India) needs analysis. The master then put the shoals twenty-four leagues away. By the first watch, he placed them about four leagues distant. The officers therefore believed they were approaching a charted hazard. No one in that exchange names it Pêro dos Banhos. The ship struck on 22 August, thirty-five days after sighting the Cape.¹⁰

Latitude without longitude defines a band, not a place. The account does not say whether the six-degree figure was observed or reckoned. This study therefore treats four to eight degrees south as a search band, not as a demonstrated error range. Along a route east of Madagascar, that band includes islands of Folhlhavai, including Candú and Foalhavahi, as well as low outer islands of the Mascarene Plateau. Latitude alone can neither place the wreck in Folhlhavai nor exclude it. Barros places Candú near seven degrees south, so latitude cannot distinguish Candú from the shoal the officers had in mind. The names provide the distinction. Barros, the route books, and the charts of the period keep them apart. Each candidate must also be tested against Rangel’s description of the islet and the positions assigned to Pêro dos Banhos on period charts.

What Rangel describes at the wreck

The ship grounded at night. As dawn approached, the survivors made out low land covered with white birds. Rangel describes a small, low island beside reefs. He does not describe an atoll survey. He gives the sandy beach’s dimensions as 300 by 160 passos (paces). Ordinary steps of about 0.825 metres would make it roughly 248 by 132 metres. Geometric paces of 1.65 metres would make it about 495 by 264 metres. Rangel does not identify the standard. The beach formed only part of the island, so its dimensions set a minimum scale for the dry land. They do not give the island’s full extent.¹¹

The survivors dug into the soil and found it wet. Rangel later says rainwater collected in the island’s lower centre, below its edges. He reports seabirds, turtle eggs, and turtles. When the boat could not go fishing, the survivors ate “lobo ou tartaruga” (sea wolf or turtle). He later says that a “lobo marinho” (sea wolf) and a turtle were dried in the sun. Herbs grew on the island and were eaten. Rangel names no trees or coconut palms there. He says the island had little firewood. From January onward, most survivors received raw fish because no firewood remained. This suggests a shortage of usable fuel, but it does not prove that woody plants were absent.¹²

Rangel does not describe the animal he calls a lobo marinho in enough detail to identify it. Stoddart’s survey of pinniped and sirenian reports in the western Indian Ocean argues that some early accounts of dugongs may describe seals. The term alone cannot distinguish a seal, dugong, or another marine animal.¹³ Later place names such as Seacow Island, given after 1770, provide no evidence for 1555. Wenban-Smith proposed Nelson Island partly on the evidence of seals. Yet no seal has been recorded anywhere in Folhlhavai. If Rangel’s animal was a seal, that would weigh against every island there, including Nelson Island.

The missing horizon

Rangel does not report seeing another island from the wreck site. This is an argument from silence, not an explicit statement that no other land existed. The silence is still geographically important. Wenban-Smith calls the absence of any visible island one of the narrative’s sharpest difficulties for the Peros Banhos hypothesis.¹⁴ From a low islet, a standing observer would see bare sand on a neighbouring islet at about twelve kilometres. A wooded skyline would be visible at about twenty-two kilometres. These estimates assume an eye height of about 2.7 metres, sand rising about two metres, and trees about fifteen metres high. They also assume standard atmospheric refraction. The modern atoll contains numerous reef islets around a lagoon roughly twenty kilometres across. If the wreck occurred on a rim islet, nearby land might have been visible. Rangel’s failure to mention it therefore needs an explanation before the identification can be treated as established.

Silence alone would not settle the matter. Night, haze, and the observer’s position affect visibility. The silence also coincides with the map evidence. Printed geographic maps keep Pêro dos Banhos apart from Candú, the atoll now called Peros Banhos. These two lines of evidence reinforce each other. The missing horizon weighs against the clustered atolls. The charts also separate the named shoal from Candú. The later name is therefore the only support for placing the wreck there.

Testing the candidates

Within the four-to-eight-degree band, the possible wreck islands fall into two groups. In the Mascarene region, the granitic Seychelles are high islands and do not fit Rangel’s description of low land. The smallest cays are too small for the reported beach. The Amirantes include clustered islands such as the D’Arros, Poivre, and Marie Louise groups. The Alphonse group is also clustered. These groups raise a serious difficulty for the horizon test. Distances and sightlines from each possible wreck beach would need to settle it. Île Platte, Desroches, and Coëtivy are isolated and pass that test most readily. Kaplin and Pirazzoli describe swampy depressions along Coëtivy’s central axis. These invite comparison with Rangel’s lower centre, where rainwater collected. They do not show that the island had the same form in 1555.¹⁵

No candidate in Folhlhavai fits the combined textual, cartographic, and geographic evidence. The test has two parts. The rim islets of Candú and Salomon, the Three Brothers, Egmont, and Foalhavahi each lie within a few kilometres of other islets in their groups. Those distances fall well within the visibility estimates above. The locations are therefore strongly inconsistent with Rangel’s silence about other land. This objection rests on his silence and the general spacing of the rim islets. It is strong, but it is not a completed test. Measured sightlines from each beach would be needed to demonstrate it.

Nelson Island and Danger Island are isolated. They may fit the physical setting, so the horizon test alone does not rule them out. The name presents a separate problem. The 1735 relation transmits the wreck under the name baixos de Pêro dos Banhos. No map in the register of printed geographic maps from 1502 to 1851 places that shoal in the Candú and Polvara region, where Nelson and Danger lie. The Folhlhavai identification rests only on the later transfer of the name Peros Banhos to the atoll that sixteenth-century charts call Candú.

The islands seen after departure

The coconuts and people belong to a later episode. Rangel’s group left the wreck island in a vessel they had built and embarked on 1 April 1556.¹⁶ They steered northeast for about twenty days. They then believed they were near the coast of India. The wind forced them southeast. After a further thirteen days, they sighted two reef-fringed islands about four leagues apart. Rangel says the islands were called Mameluco and places them at the latitude of Melinde. He gives no longitude. The survivors landed on one island, which had coconut palms and huts but no people. The people who used these islands came there only to fish and make coir. When they arrived, they took the survivors to a populated island ruled by a Moorish king. He gave them a boat for Cananor. On the way, the survivors reached a second island with another king. Rangel does not describe that king as Moorish. The final passage to Cochin took ten days.¹⁷

The account describes one seasonally used landfall and two inhabited islands with kings. It identifies the first king as Moorish. It names none of the islands. The coconut trees, huts, and visitors cannot be transferred back to the wreck site. Mameluco has no secure modern identification. Rangel gives a latitude, not a location near Melinde. The name alone cannot settle the question. Any identification must account for the stated latitude, the distance between the islands, and their seasonal use.

The two kings can be distinguished. The second is the better place to begin because the final passage’s duration bears on his island. One hypothesis is that his island belonged to Mamále. The name resembles Rangel’s Mameluco. But the Mameluco landfall cannot be the second king’s island. The survivors reached that island only after leaving Mameluco and passing the first king. Barros says Mamále took its name from “hum Mouro de Cananor” (a Moor of Cananor), who was lord of the first islands. He places them forty leagues off the Malabar coast at twelve and a half degrees north.⁵

Rangel gives ten days for the passage from the second island to Cochin. He says nothing about the boat, winds, or route. Ten days is compatible with an island near the Malabar coast, which Barros’s distance of forty leagues would allow. It does not identify the island. Barros places Mamále at the northern end of the chain he calls the Maldives. He also says the first islands were under the lordship of a Moor of Cananor. This establishes a political connection with Cananor. It does not record a cession by a named Maldivian sultan, or identify Rangel’s Mameluco with Mamále.

Distance weighs against the southern atolls, though it does not exclude them. Candú lies about 950 nautical miles from Cochin. Ten days of continuous progress over that distance requires an average of 3.96 knots. Such a passage would be demanding for a small sailing vessel, although not impossible under favourable conditions. Rangel says nothing of the vessel, the winds or the route. Malé lies about 385 nautical miles away.¹⁸ Distance alone does not decide the case for Malé. The question of the central Maldives remains open until the political evidence is set out from a sixteenth-century source. The ten-day passage makes Mamále plausible and a southern Maldivian origin less persuasive, without excluding it. The first king can be approached from the other end of the sequence. If the second island lay in Mamále, the first lay farther from the Malabar coast. The evidence for his identity is discussed below. No record yet names his island. Both identifications remain hypotheses. Rangel gives neither island a name or position.

Nationality and location

Rangel calls the visitors negros (Black people) and describes their ruler as Moorish. Negros refers to appearance. Mouro (Moor) was a period confessional or ethnographic label. Neither term identifies geographic origin. East African or Comorian visitors are compatible with the account, but the terms do not identify them. Duarte Barbosa describes inhabited islands off the Swahili coast, including Pemba, Mafia, and Zanzibar, under Moorish rulers and engaged in maritime trade. Coir rope was also a substantial East African industry in the sixteenth century. In 1506, Pedro Ferreira Fogaça reported from Kilwa that six islands, including Maotoe (Mayotte), had Moorish kings and supplied provisions to Kilwa and Mombasa.¹⁹

These records establish Moorish rulers and maritime exchange in the western Indian Ocean. They also show that fishing and coir were not uniquely Maldivian practices. The Swahili islands lie close to the latitude of Melinde that Rangel gives for Mameluco. They do not prove that people from those islands visited Mameluco. Any such hypothesis must account for Rangel’s voyage. He gives northeast and southeast as the courses steered. Both carried the boat away from the African coast. An East African identification must therefore explain how visitors from those islands reached Mameluco. The roteiro’s references to Saia da Malha, Pêro dos Banhos, and the Ilhas Sete Irmãs deserve reconstruction against contemporary charts. Costa’s later gloss cannot turn Rangel’s account into coordinates.²⁰

Conclusion

The 1735 relation records a wreck on a named shoal. On the previous day, Rangel reported the ship’s position as six degrees from the Line. The account describes survival on a small, low island with no coconut palms and no other land reported in view. It then describes a separate voyage to seasonally used islands. The survivors later passed two island kings, the first explicitly Moorish, before reaching Cananor and Cochin.

The reported latitude defines a band that includes Folhlhavai and other island groups. Rangel gives no longitude. No candidate in Folhlhavai fits the combined evidence. The clustered island groups conflict with his failure to report other land in view. Nelson and Danger Islands may fit the physical description. Yet printed geographic maps keep the named shoal apart from the Candú and Polvara register in which those islands lie. The attribution of the wreck to Peros Banhos is therefore a later label, not a finding from Rangel’s account. It rests on a later reorganisation of the region’s nomenclature, first documented, in the material identified in this research, in French hydrographic manuscripts and fixed in nineteenth-century charting, read back into the sixteenth century.

Within the Mascarene region, Coëtivy, Île Platte, and Desroches are the candidates most worth testing against Rangel’s description. The ten-day passage to Cochin is compatible with Mamále as the second king’s island. Barros places its first islands under a Moor of Cananor. The passage makes a southern Maldivian origin less persuasive without excluding it. An origin for the first king in the Moorish island kingdoms of the western Indian Ocean is compatible with the account, though the evidence does not prove it. The precise island, the identity of Mameluco, and the rescuers’ nationality remain unresolved. The episode cannot, on this evidence, prove that Portuguese sailors discovered Folhlhavai in 1555.

Notes

1. Manoel Rangel, “Relação do naufrágio da nau Conceyção,” in Bernardo Gomes de Brito, História Trágico-Marítima, vol. 1 (Lisbon, 1735), 169–217. Nigel Wenban-Smith, “The Conception, A 450-Year-Old Mystery,” Chagos News 29 (2007), 9–12.
2. C. R. Boxer, “An Introduction to the História Trágico-Marítima,” Revista da Faculdade de Letras (Lisbon), 3rd ser., no. 1 (1957), 48–99, esp. 55–56. Idem, “Some Corrections and Clarifications,” Quaderni Portoghesi 5 (1979), 99–112.
3. “Roteiro para a Índia e Oriente,” in Abel Fontoura da Costa, ed., Roteiros Portugueses Inéditos da Carreira da Índia do Século XVI (Lisbon, 1940), 37–38, 78 n. 54.
4. James Duffy, Shipwreck and Empire (Cambridge, MA, Harvard University Press, 1955). Paulo Guinote, Eduardo Frutuoso, and António Lopes, Naufrágios e outras perdas da Carreira da Índia, séculos XVI e XVII (Lisbon, 1998), 49–50. Humberto Leitão, “Identificação dos Baixos de Pero dos Banhos e das Chagas,” Studia 1 (1958), 118–22. W. Ph. Coolhaas, “As notícias neerlandesas mais antigas acerca dos Baixos de Pero dos Banhos e das Chagas,” Studia 5 (1960), 7–19. Donald Taylor, “The Gradual Discovery of the Islands of the Chagos Archipelago,” Chagos News 17 (2001), 4–8. Wenban-Smith, “Conception,” 9–12. Ullattil Manmadhan and Nigel Wenban-Smith, The Sad Story of the Conceição (Chagos Conservation Trust, 2023). Philippe Sands, “Chagos, The Last British Colony in Africa,” 10th Whitehall Lecture (2022), 14–15.
5. João de Barros, Terceira Década da Ásia (Lisbon, 1563), livro III, capítulo VII, “Em que se descreve o sítio das ilhas de Maldiva” (in which the situation of the Maldive islands is described), fol. 71. The passage reads “começando nas a que chamam de Mamále. Nome de hum mouro de Cananor que era senhor das primeyras: que estam apartadas da costa Malabar per espaço de quorenta legoas em altura de doze graos & meyo da parte do norte. E as derradeyras nesta distancia de trezentas legoas chamadas Candú & Adú, estam em sete graos da parte do sul” (beginning at those they call Mamále, the name of a Moor of Cananor who was lord of the first, which lie forty leagues off the Malabar coast in twelve and a half degrees north, and the last, at this distance of three hundred leagues, called Candú and Adú, lie in seven degrees south). On the recording of latitude, Amândio J. M. Barros, “A latitude como instrumento: uma lista de alturas e a navegação no Estado da Índia (século XVI),” População e Sociedade 44 (2025), 28–52, https://doi.org/10.52224/21845263/rev44a2, section 3 and note 13. Section 3 notes that the instruments “tinham as escalas marcadas em graus sendo que os minutos, quando referidos, resultam de estimativas dos pilotos” (had their scales marked in degrees, the minutes, when given, resulting from the pilots’ estimates). Note 13 gives the same whole-degree phrasing in a period latitude, “Moçambique esta em 15 graus da banda do sul” (Mozambique is in fifteen degrees on the southern side), beside chart values of 14½. On Adú, Cornelis de Houtman and Benjamin Wright (1598), Jan Barend Elwe, Partie de la nouvelle grande carte des Indes Orientales (Amsterdam, 1792), and John Thomson (1817), in the cartographic register at note 6.
6. Cartographic register analysed for this research, forty-five printed geographic maps, 1502 to 1851, https://drive.google.com/drive/folders/1kuBPB3WfgArYLui4Tf7LDhfi19WJYgZx. For this survey, a map counted as legible when the names needed for comparison were readable in the copy examined.
7. Lazare Picault (attrib.), Plan des isles de Peros Banhos decouverte par la tartane l’Élizabeth, 16 April 1744, manuscript, Bibliothèque nationale de France, GE SH 18 PF 215 DIV 13 P 1, https://catalogue.bnf.fr/ark:/12148/cb44373387f. Plan des isles de Chagas découverte par la tartane l’Élizabeth, Capitaine Lazare Picault, le 16 avril 1744, manuscript copy, after 1744, signed on the reverse “Le Chev[alier] de Fleurieu,” Bibliothèque nationale de France, GE SH 18 PF 215 DIV 13 P 1/1, https://catalogue.bnf.fr/ark:/12148/cb443734135. For the comparison, United Kingdom Hydrographic Office, British Admiralty chart 725, British Indian Ocean Territory, Plans in Chagos Archipelago, plan of Peros Banhos. A. A. Fauvel, Unpublished Documents on the History of the Seychelles Islands Anterior to 1810 (Mahé, Government Printing Office, 1909), 42–43.
8. D. R. Stoddart, “Settlement and Development of Diego Garcia,” Atoll Research Bulletin 149 (1971), 210–211. Bibliothèque nationale de France, “Fleurieu, Charles-Pierre Claret de (1738–1810),” Catalogue général, notice de personne, https://catalogue.bnf.fr/ark:/12148/cb121786370 (accessed 17 October 2026), which dates his appointment as inspecteur en second du dépôt des cartes et plans de la Marine (deputy inspector of the navy’s charts and plans dépôt) to 1776. Projet Lapérouse, “Charles Fleurieu,” CUFR Champollion, Valorisation du Patrimoine project (2013–2014), https://blogs.univ-jfc.fr/projetlaperouse/le-voyage/journaux-et-correspondance/charles-flerieu/ (accessed 17 October 2026), gives February 1775.
9. Robert Moresby and F. T. Powell, Admiralty Chart No. 3, Chagos Archipelago, surveyed 1837, published 2 December 1839, National Library of Australia, nla.obj-574968554. Idem, India Trigonometrical Survey of the Maldivh or Maldiva Islands, three sheets, published 1839, Royal Museums Greenwich, G251:21/1B, sheet 3 also issued as Admiralty Chart 66c. H. C. P. Bell, The Máldive Islands (Colombo, 1883), 36–37. Robert Moresby, “Reports on the Maldivas,” Transactions of the Bombay Geographical Society 1 (1844), 102–108.
10. Rangel, “Relação,” 172–175.
11. Wenban-Smith, “Conception,” 10–11. Beatriz Piccolotto Siqueira Bueno, “Decifrando mapas, sobre o conceito de ‘território’ e suas vinculações com a cartografia,” Anais do Museu Paulista 12, no. 1 (2004), 193–234, table of historical measures (ordinary passo, 0.825 m. Geometric pace, 1.65 m).
12. Rangel, “Relação,” 186, 201–205.
13. D. R. Stoddart, “Pinnipeds or sirenians at western Indian Ocean islands?,” Journal of Zoology 167, no. 2 (1972), 207–217.
14. Wenban-Smith, “Conception,” 12. The visibility distances follow the standard formula for the distance of the horizon, about 1.17 nautical miles multiplied by the square root of the height in feet, summed for observer and target and allowing for normal refraction. An eye height of 2.7 metres gives about 3.5 nautical miles, sand of two metres about 3.0, and trees of fifteen metres about 8.2, so sand becomes visible at about 6.5 nautical miles (12 kilometres) and a wooded skyline at about 11.7 nautical miles (22 kilometres). Nathaniel Bowditch, The American Practical Navigator, Pub. No. 9 (Bethesda, MD, National Geospatial-Intelligence Agency, 2019), Table 12, Distance of the Horizon.
15. P. A. Kaplin and P. A. Pirazzoli, “A Geomorphological Reconnaissance of Coëtivy Atoll, Seychelles, Indian Ocean,” Atoll Research Bulletin 319 (1988), 1–8.
16. Rangel, “Relação,” 208.
17. Rangel, “Relação,” 196–217. Wenban-Smith, “Conception,” 11–12.
18. Distances are great-circle measurements from Cochin (9.97° N, 76.27° E) to the centre of the atoll now called Peros Banhos (5.30° S, 71.85° E), about 954 nautical miles, and to Malé (4.18° N, 73.51° E), about 385 nautical miles. Ten days of continuous progress over 954 nautical miles requires an average of 3.96 knots.
19. The Book of Duarte Barbosa, ed. Mansel Longworth Dames, vol. 1 (London, Hakluyt Society, 1918). Malyn Newitt, A History of Mozambique (London, Hurst, 1995). Pedro Ferreira Fogaça to the King, Kilwa, 1506, in Documentos sobre os portugueses em Moçambique e na África Central, 1497–1840, vol. 1 (Lisbon, 1962), 761, as cited in Iain Walker, Islands in a Cosmopolitan Sea, A History of the Comoros (London, Hurst, 2019).
20. Costa, Roteiros, 37–38, 78. Leitão, “Identificação”. Coolhaas, “Notícias neerlandesas”.