Columbus's First Landfall: Which Island Did He Actually Reach?
The Guanahani Debate

The Virgin Islands lie approximately 1,000 miles southeast of the Bahamas, east of Puerto Rico. Any route placing Columbus's first landfall here would require him to have sailed past Hispaniola — an island roughly the size of Ireland — without recording it.

The theory relies solely on the transatlantic track, specifically a disputed reading of a September 13, 1492 Diario entry made in the middle of the Atlantic. No backward inter-island reconstruction from Cuba has ever been attempted, because none is possible.

Dr. Luis M. Coin Cuenca proposed the Virgin Islands as the landfall in 1989, based on his interpretation of Atlantic current patterns. He has attracted no subsequent supporters.

A single ambiguous Diario entry from mid-Atlantic is the sole basis for this theory.

The theory requires Columbus to have sailed past Hispaniola without observing or recording it, and to have crossed several hundred additional miles beyond what his log records. No inter-island route from the Virgin Islands to Cuba is consistent with the log.

Lignum Vitae Cay (also called Great Harbour Cay) is located in the Berry Islands at the extreme northern end of the Bahamian chain, approximately 500 miles northwest of where all serious theories place the landfall.

The theory relies on a claimed physical fit between Lignum Vitae Cay and Columbus's Island I descriptions. No viable transatlantic track or backward inter-island reconstruction supports this location.

John H. Winslow, a professor at Florida Atlantic University, proposed this theory in 1989 after personally visiting Lignum Vitae Cay and Great Harbour Cay. He has attracted no subsequent supporters. Sun-Sentinel, 1990.

The physical description of the island — particularly the central lake and multicolored stones — has some superficial correspondence with Columbus's Island I descriptions. Winslow at least visited the island, which is more than most theorists did.

Reaching Lignum Vitae Cay requires Columbus's ships to have passed through or near Treasure Cay or Harbour Island, neither of which appears in the log. The proposed route requires transiting the Middle Bight of Andros Island, a shallow tidal area approximately three feet deep. No island or distance descriptions in the Diario support a northern Bahamas landfall.

Cat Island is a long, narrow island in the central Bahamas, approximately 48 miles long and one to four miles wide, running roughly north-northwest to south-southeast. It was labeled "San Salvador" or "Guanahani" on 18th-century charts before anyone had seriously analyzed the Diario — the theory's origin is cartographic tradition, not navigational evidence.

The Cat Island theory predates systematic analysis of the Diario and is based primarily on historical cartographic tradition, not navigation reconstruction.

Mark Catesby identified Cat Island as Guanahani in his Natural History of Carolina (1731). The theory gained its most influential support when Washington Irving asked U.S. Navy Commander Alexander Slidell Mackenzie to evaluate the two leading theories of the day. Mackenzie favored Cat; Irving published the analysis in 1828. Alexander von Humboldt supported it in 1837. The theory lost traction after the Diario was published in full in 1875–76.

Historical cartographic tradition placed Guanahani at Cat Island on many 18th-century charts. The origin of this convention is itself unresolved. Also, its Taino name, Guanima, is phonetically similar enough to Guanahani that it passes my phonics criterion.

Cat Island has no reef completely surrounding it — a clear Columbus criterion. There is no enclosed large harbor. The proposed Island II (Conception) is far too small: less than two leagues north-south and east-west against Columbus's five and ten leagues respectively. The proposed Islas de Arena are the non-existent "Mucara" islands.

Egg Island is a tiny landmass of approximately 0.2 square miles off the western tip of Eleuthera in the northern Bahamas, far from any transatlantic track placing the landfall in the central or southern Bahamas.

The theory relies on a transatlantic track requiring Columbus to have been a celestial navigator — something U.S. Navy Rear Admiral Bob McNitt's analysis of the log demonstrates he was not. The inter-island route requires crossing the Great Bahama Bank.

Arne B. Molander, a retired civil engineer, proposed Egg Island in 1981 and has been its sole serious advocate since. His defense of the theory's many problems frequently invokes what Pickering calls the "All Purpose Indian Map Excuse" — when the log contradicts the theory, Molander asserts Columbus was misled by an inaccurate native map. But Columbus himself wrote that he could not obtain cartographic information from the natives until he had learned their language.

Egg and Royal Island together have a surrounding reef and a reasonably sized harbor at Royal Island.

The proposed route from Island III to Island IV crosses approximately 90 miles of the extremely shallow Great Bahama Bank at speeds requiring Force 6 winds — conditions under which that passage would be impossible. James E. Kelley Jr. (1992) made a convincing and unrefuted argument from Columbus's own report of the canoes that visited on October 13: the minimum population of San Salvador (Guanahani) was between 634 and 1,115. Egg Island is far too small to have supported that population. Pickering lists 28 unresolved problems.

Conception Island is a small, uninhabited island in the central Bahamas, southeast of Cat Island and northwest of Rum Cay. Its position within the central Bahamas gives it more geographic plausibility than the northern and southern outliers.

The theory uses both transatlantic track analysis and inter-island reconstruction, as well as archaeological and paleogeographic analysis of Bahamian shorelines as they existed in 1492.

Rupert Gould proposed Conception Island in 1943. Geologist Steven Mitchell and archaeologist William Keegan later used paleogeographic reconstruction of 1492 shorelines to argue for Conception — a methodological approach worth acknowledging even if unconvincing to most analysts.

Conception Island has a surrounding reef and reasonable harbor. The island is flat and green. Archaeological evidence suggests Lucayan settlement. The paleogeographic reconstruction by Mitchell and Keegan is a valid methodological contribution.

The inter-island route from Conception requires Island II to have a north-south coast facing Island I — but no island facing Conception from the south matches this description. The proposed island chain does not map cleanly onto the log's distances and directions.

Watlings Island was officially renamed San Salvador Island by the Bahamian parliament in 1926, legally cementing its standing as the landfall. Most world maps still identify it as Columbus's first landfall — an authority that has outlasted the scholarly consensus that once supported it.

The quintessential transatlantic track theory. Samuel Eliot Morison spent five months in 1939–40 personally retracing Columbus's route under sail. John W. McElroy's 1941 technical study of ship performance provided the underpinning Morison incorporated into his 1942 biography.

Juan Bautista Muñoz first suggested Watlings in 1793. A. B. Becher revived it in 1856; Lieutenant Joseph B. Murdock provided the most detailed 19th-century defense in 1884. The theory's dominant 20th-century position rests almost entirely on Samuel Eliot Morison's Pulitzer Prize-winning Admiral of the Ocean Sea (1942). Philip Richardson and Robert Goldsmith of the Woods Hole Oceanographic Institution reaffirmed Watlings in 1987 using Atlantic current modeling.

Watlings Island has a large central lake — the best match among all candidates for Columbus's "very large laguna in the middle." The transatlantic track reconstructions of McElroy and Richardson-Goldsmith both end near Watlings. Archaeological excavation at the Long Bay site confirmed Lucayan habitation consistent with Columbus's population descriptions. Morison's personal sailing experience gave his conclusions empirical grounding absent from purely theoretical analyses. Also, the Taino name — Waanhani, though I have also seen it recorded as Guanahani, one of two islands supposedly given that name — is phonetically close, or better, to Guanahani.

Support for Watlings collapsed in the scholarly community after 1996 when Pickering resolved the leagues-versus-miles question. Once established that the Bahamian descriptions are in leagues, the route unravels: Rum Cay's north-south coast measures less than two leagues against Columbus's five, and its east-west coast less than four leagues against Columbus's ten. The surrounding reef at Watlings does not completely encircle the island. The reef entrance is approximately 1,400 meters wide — not "very narrow." One of the biggest detractions for me personally is that it would have been impossible for Columbus to have seen a light from any nearby island hours before landfall — a requirement of the October 11 evidence. Pickering lists 28 unresolved problems.

East Caicos is the largest of the Turks and Caicos Islands, lying at the southern end of the Bahamian archipelago just north of the Windward Passage. It is approximately 18 miles long and was heavily settled by the Lucayan people.

The theory uses inter-island reconstruction, working forward from Caicos as Island I through Mayaguana, Acklins, and Great Inagua to Cuba.

Pieter Verhoog, a Dutch sea captain, proposed Caicos in 1947. His theory gained support from Edwin and Marion Link in 1958 and Robert Fuson in 1961. Verhoog's confident declaration — "I have never found a single serious objection against Caicos" — suggests he had not scrutinized the inter-island track rigorously enough.

East Caicos is large enough to match Columbus's "fairly large" description and has a lagoon and considerable Lucayan archaeological evidence. Its position in the southern Bahamas is consistent with transatlantic track analysis using Portuguese leagues. Oviedo's Historia general (1535) notes that Guanahani was adjacent to an island called "Caicos." Also, its Taino name was Guana, which is close to Guanahani, and even closer if combined with the Middle Caicos name, Aniana, to make Guananiana.

Caicos comprises five islands, not one. The pre-contact Lucayan population was almost certainly several times larger than the 500–1,100 range Columbus's descriptions imply. The Island II distance is severe: approximately 40 nautical miles from Caicos to Mayaguana — far outside Columbus's five to seven leagues even using Portuguese leagues. Pickering lists 19 unresolved problems.

Grand Turk is the easternmost island in the Lucayan archipelago, now the capital of the Turks and Caicos Islands. It is a small, low island surrounded by small sandy islets of exceptional whiteness.

The Grand Turk theory relies primarily on transatlantic track analysis combined with a selective reading of early historical sources. The inter-island reconstruction is its weakest element.

Martín Fernández de Navarrete proposed Grand Turk in 1825. The theory was periodically revived — by H. E. Sadler in 1980, Robert H. Power in 1983, and Josiah Marvel in 2016. Power acknowledged his theory's most serious problem: his proposed Island III requires Columbus to have mistaken open ocean for a coastline.

Grand Turk is surrounded by small sandy islets of exceptional whiteness, matching Oviedo's 1535 description of the islands near Guanahani. The October 11 light fits well — potentially coming from another island to the east. Pickering gives Grand Turk a perfect 3 on the October 11 light criterion. Its Taino name, Amuana, is the weakest match of any island that still passes my phonetic criterion.

The inter-island route requires Columbus to have sailed 58 nautical miles in six hours at ten knots — faster than his ships could sail. The theory requires him to have interpreted open ocean as a coastline for Island III. The Island I to Island IV distance is approximately five times what Columbus states. Pickering lists over fifteen unresolved problems.

Samana Cay is a small, uninhabited island lying north of Acklins Island in the central-southern Bahamas, approximately 60 miles east-southeast of Long Island. It is roughly oval with an east-west orientation, low-lying, and surrounded by reefs.

Backward inter-island reconstruction from Cuba, placing Crooked-Acklins as Island II, Long Island as Island III, and Fortune Island as Island IV — the same route used by Plana Cays and Mayaguana.

Gustavus V. Fox, former Undersecretary of the U.S. Navy, first proposed Samana Cay in 1882. Criticized by James B. Murdock in 1884 — Fox had died in 1883 and could not respond — the theory lay dormant for a century. Joseph Judge revived it dramatically in National Geographic (November 1986). Alejandro Perez subsequently found evidence in Las Casas's Historia de las Indias that refuted Murdock's objection.

Samana Cay is completely surrounded by a reef — one of Columbus's clearest Island I criteria. The shared route through Crooked-Acklins, Long Island, and Fortune Island is the strongest inter-island sequence in the debate. The De la Cosa map shows Guanahani with small islets, and Samana Cay has one distinct islet to the east and one to the south — a configuration that fits the map reasonably well. Numerous Lucayan archaeological sites have been identified on the island, consistent with the population range Columbus describes. Samana is one of two islands whose geographic form is similar to the depiction of Guanahani on the De la Cosa map.

Samana Cay has three unresolved problems. First, there is no obvious island to the east that could have produced the October 11 light. Second, the coast of Crooked-Acklins that faces Samana runs east-west, not north-south as Columbus describes for Island II. Third, Samana Cay has no central pond matching the "laguna muy grande." Samana Cay, at 30 miles from Island II, is too far — even using Portuguese leagues and a 15% margin of error — to have been 5 to 7 leagues away. Lastly, Samana is simply not close enough phonetically to Guanahani.

"It is impossible to explain the facts at our disposal in any other way. The solution to the mystery is Samana Cay."Joseph Judge, Senior Editor, National Geographic

The Plana Cays consist of two islands — North Plana and South Plana — in the southeastern Bahamas, east of Acklins Island and west of Mayaguana. This two-island geography is both the theory's central strength in Pickering's analysis and, in my view, its central weakness.

Backward inter-island reconstruction from Cuba, following the same Crooked-Acklins to Long Island to Fortune Island route as Samana Cay and Mayaguana. Pickering additionally applies transatlantic track analysis using updated 2004 geomagnetic field models.

The Plana Cays theory was first proposed by Ramón Julio Didiez Burgos in Guanahaní y Mayaguain (1974). Pickering revived it in a 1994 article in the journal DIO and has been its most vigorous proponent since. At a 1992 U.S. Naval Institute debate, moderator William F. Buckley awarded a TKO to the Plana theory after Pickering demonstrated it was the only candidate whose Island II had a north-south coast facing Island I.

The critical directional clue — the north-south coast of Island II (Crooked-Acklins) faces Island I — is satisfied by Plana (and Mayaguana). A surrounding reef is present. The De la Cosa map potentially shows Guanahani as two islands, which matches Plana's two-island geography. The 2004 geomagnetic track analysis is consistent with a central Bahamas landfall. The October 11 light fits well: it would have come from Mayaguana, directly east of Plana. Its modeled population is roughly in line with Columbus's description.

Plana Cays consists of two islands. Columbus never says Guanahani is two islands, and his description — "a very large laguna in the middle" — describes a single unified landmass. As Joseph Judge wrote to Pickering: "If the Plana Cays were one island... it would absolutely be the landfall." (Worth noting: this was based on Italian leagues, which I dispute.) But it is not one island. The ponds at Plana are not in the middle of the island. There is no peninsula with a narrow neck. Acklins Island is marginally visible from the Plana anchorage — Columbus does not mention other visible islands until after leaving Island I. Additionally, it is unlikely the Lucayan population of Plana was sufficient to correspond to the number of people and canoes Columbus describes.

On the De la Cosa map: as described above, I examined the original at the Museo Naval in Madrid in July 2026. What Pickering reads as two islands separated by a channel appears to me to be a single island with pigment loss — a pattern I photographed repeated elsewhere on the same map, in places with no bearing on the landfall question. The Diario's own language supports the single-island reading. I do not accept this criterion as evidence for Plana.

On leagues: Pickering assumes the Italian league (~2.67 nautical miles). I believe Columbus used the Portuguese Imperial League (~3.2 nautical miles). Using Italian leagues, Guanahani was 13.3 to 16 modern miles from Island II — a range that only Plana Cays (about 15 miles from Crooked Island) fits. Using Portuguese leagues, that range extends to 18.7 to 22 modern miles, which brings Mayaguana (about 25 miles from Crooked Island) within Pickering's own ±15% margin of error as well. Under Italian leagues, only Plana fits; under Portuguese leagues, both islands do.

Lastly, the Taino name, Amawayo, is not phonetically close to Guanahani.

"If the Plana Cays were one island... it would absolutely be the landfall."Joseph Judge, former National Geographic Senior Editor, in a letter to Keith Pickering, August 10, 1992

While other factors such as Portuguese leagues and phonetics reduce the certainty of Judge's statement, two islands are still two islands.

Mayaguana is the easternmost island in the Bahamian archipelago proper, lying west of the Caicos Islands and east of the Plana Cays. At approximately 24 miles long and 6 miles wide, it is the largest island in the southeastern Bahamian cluster. Mayaguana is a single, unified island with a surrounding reef, a large bay (Abraham's Bay) on its southern coast enclosed by a reef line, and small islets on its eastern side. The island has no mountains, is flat, and is covered in green vegetation — matching Columbus's description precisely.

I followed Pickering's backward inter-island reconstruction methodology. Starting from Cuba, working backward through the Ragged Islands, Island IV (Fortune Island), Island III (Long Island), and Island II (Crooked-Acklins), the route leads to the southeastern Bahamian cluster. Within that cluster, I depart from Pickering's conclusion for the reasons below.

Mayaguana was first proposed by Francisco Adolfo de Varnhagen in 1864 in La verdadera Guanahani de Colón. Varnhagen was a Brazilian historian and diplomat who conducted serious primary source analysis — his identification of Acklins-Crooked as Island II and Fortune Island as Island IV were methodological innovations that Pickering later incorporated into his own Plana Cays theory. The Marquis de Belloy supported Varnhagen in 1877. The theory has had few serious advocates since.

Mayaguana is a single island, consistent with Columbus's first-person description of Guanahani as a unified landmass with "a very large laguna in the middle." The two-island reading that favors Plana on the De la Cosa map is, in my view, an artifact of pigment loss rather than original cartography.

The name correspondence is striking and has been almost entirely ignored in the literature: Mayaguana is more phonetically similar to Guanahani than any other original Taíno island name in the Bahamas. The inversion of syllables — Guanahani / Mayaguana — is a recognized pattern in oral transmission of place names across language boundaries. I am surprised this has not been given more weight.

Mayaguana has small islets on its eastern side as well as a reef line along the southern boundary of Abraham's Bay — a single large island with associated islets, consistent with the De la Cosa map. The De la Cosa map equally supports Samana and Mayaguana on the islet criterion.

The distance from Mayaguana to Crooked Island is approximately 25 modern miles. Columbus describes the journey as "farther than five leagues, closer to seven," delayed by contrary tide and not completed until noon. Plana Cays is approximately 15 miles from Crooked Island; Mayaguana approximately 25. Both fall within Pickering's ±15% margin when Portuguese leagues are used — but the qualitative feel of the passage (tidal delay, noon arrival) fits a 25-mile crossing more naturally than a 15-mile one.

Pickering himself acknowledges that Mayaguana has "by far the best fortifiable peninsula at Island I" of any candidate. Abraham's Bay provides a large harbor between the reef and the island with a narrow entrance — matching Columbus's description closely. If Mayaguana is Island I, the October 11 light could have come from the Turks and Caicos, such as Providenciales. Plana Cay, which consists of multiple islands, is within five Portuguese leagues of Mayaguana — addressing Pickering's concern about visible islands after departure from Island I.

The historical record attributes a larger pre-contact population to Mayaguana than the 500–1,100 range Columbus's descriptions imply. However, archaeological population estimates carry significant uncertainty, and Columbus likely did not see the entire population. Some researchers suggest a native population of approximately 3,000 would be necessary to produce the visible signs Columbus observed of 500–1,100. My own population modeling estimates a population of 3,162 on Mayaguana in 1492.

Lastly, Mayaguana is phonetically close to Guanahani and, along with Samana Cay, has the closest geographic shape to the island of Guanahani on the De la Cosa map.

Pickering raises several legitimate concerns. The distance from Mayaguana's likely west coast anchorage to the fortifiable peninsula on the east is approximately 25 nautical miles, making a round-trip boat journey in less than a day unlikely under normal conditions. However, Mayaguana is large, and a mariner would naturally explore it by boat rather than on foot — which actually supports the island's larger size.

The Island IV distance problem — Fortune Island is roughly twenty leagues from Mayaguana while Columbus states Island IV was eight leagues from Island I — is the same problem that affects Plana Cays and Samana Cay, and may reflect a navigational estimate rather than a precise measurement.