Coconuts in the Americas.
Coconut is the iconic palm of beaches throughout the tropics. It was a major plantation crop for much of the 19th and 20th centuries, and still provides an income to millions of small farmers. It is an invaluable plant providing many of the basic necessities for survival in traditional societies, especially in the southern and western Pacific, where it will have a 21st-century role as a biofuel. It was possibly the first pan-tropical crop plant, yet where it came from and how it was dispersed has long fascinated scholars. The most hotly contested part of this discussion concerns the coconut reported by the first European explorers on the Pacific coast of Panama in the early 16th century. Over the past century the discussion has waxed and waned, and now new genetic evidence suggests a relationship between the current tall population on the Pacific coast of Panama and coconuts in the Philippines (Baudouin & Lebrun, 2009; Gunn et al., 2011). The genetic analyses are compelling, but a single line of evidence is seldom sufficient to convince the gamut of scholars. Thus it is worthwhile to review all available evidence to try to answer the questions of how coconut traveled to Panama, where it came from and when it arrived, or to suggest new hypotheses for future research.
The renowned 19th century Swiss botanist, Alphonse de Candolle, is considered the founder of modern crop biogeography because he proposed a multidisciplinary methodology to identify crop origins and trace crop diffusions (1883). At a minimum, this requires information from botany, especially patterns of variation, history, linguistics and archaeology. During the early 20th century, the famous Russian geneticist, Nicolay I. Vavilov, expanded the patterns of variation to include genetics (Vavilov, 1951), initially Mendelian and quantitative, and more recently molecular genetics (Zeder et al., 2006).
This methodology, or parts of it, has been used to examine other candidates for trans-Pacific human dispersal, notably sweet potato (Ipomoea batatas) and more recently chicken (Gallus gallus). Several lines of evidence suggest that Polynesians made voyages to and from the west coast of the Americas, carrying sweet potato westwards to western Polynesia and chickens eastwards to coastal Chile (Jones et al., 2011).
The evidence for sweet potatoes includes archaeology, ethnobotany, linguistics and genetics, and seems reasonably complete. The sweet potatoes in Oceania appear to have originated from two separate introductions from America (Ballard et al., 2005). The earliest was to Yen's "ellipse" area of eastern Polynesia by 900-1,000 years before present (BP) (Yen, 1974; Green, 2005: 46-7; Lebot, 2009: 94) and thence to New Zealand and Hawaii. Linguistic and other evidence is cited by Scaglion (2005: 35-41) for the Gulf of Guayaquil, Ecuador, as the origin of these sweet potatoes, which corresponds with the South American center of sweet potato domestication (Roullier et al., 2011). The second wave of introductions came to Papua New Guinea and western Melanesia from Indonesia after the Spanish voyages from Mexico; these introductions came from the Mesoamerican center of sweet potato domestication (Roullier et al., 2011). A study by Zhang et al. (2004) questioned whether the Ecuadorian origin was not representative of both origins and dispersals, and concluded that Mesoamerican sweet potatoes dominate Oceania. Roullier et al. (2013) have now confirmed Yen's (1974) hypothesis, making this the first plant evidence for a Neotropical to Oceania dispersal.
A combination of archaeological, genetic and some linguistic evidence was recently used by Storey et al. (2007, 2011a) to propose that the chicken had been introduced into El Arenal, Chile, from Polynesia a century before European conquest. They counter earlier questioning of the validity of the archaeological results of the chicken remains in Chile (e.g., Gongora et al., 2008a, b). Although Acosta (1940) mentions them in the Caribbean and Mexico, Patino (1970: 34, 35) considers that no reliable historical record exists before Columbus took chickens on his second expedition and disembarked them in 1493 as presents for local chiefs, which thus explains Acosta's record, but is post-conquest in that region. Storey et al. (2007) affirm that Francisco Pizarro found chickens in Peru, but Patino's (1970) documental analysis does not agree.
Skeletal, artifact and non-material cultural evidence also suggests Polynesian contacts with the Chilean coast (Matisoo-Smith & Ramirez, 2010; Ramirez-Aliaga, 2011). Sewn plank canoe-building technology, characteristic of Polynesia and Micronesia, found in small areas in southern California and southern Chile, also suggests contact with the Pacific Islands (Jones et al., 2011; Ramirez-Aliaga, 2011).
The information available for coconut is much less complete, with new genetic information and an historical record that has been questioned and defended by different authors. No information on the archaeology and linguistics of coconut in Panama has yet been published. Only a fraction of the coconut genetic variation in the Philippines is present in Panama, suggesting a single small founder event (Baudouin & Lebrun, 2009), while populations on the Pacific coast of Mexico contain considerable Philippine variation, but also only a fraction of the Panama genetic variation (Guun et al., 2011). This lack of variation raises two possibilities: a direct introduction from the Philippines to Panama (Baudouin & Lebrun, 2009), or an introduction via Mexico after European conquest (Harries, 1978). Previous work has already shown that this founder event could not be explained by coconuts drifting on ocean currents (Ward & Brookfield, 1992). We thus seem to have a reasonable answer to the "where it came from" question, but the "how it got there" remains an open question, although a trans-Pacific route is clear. An associated question is "when," which raises contrasting possibilities: a very early direct introduction from the Philippines to Panama, e.g., 2,250 BP, as postulated by Baudouin and Lebmn (2009); at a date immediately before European contact (suggested by the extremely limited endemic distribution reported at the time of European conquest); or over the Manila-Acapulco route less than 450 years ago (Small, 1929; Hill, 1929; Merrill, 1954; Harries, 1971, 1978). These hypotheses can be examined with linguistic and archaeological evidence, as well as with evidence from ethnobotany and history, which might help with the "how" question.
We apply de Candolle's and Vavilov's multidisciplinary methodology to the question of coconut on the Pacific coast of Panama at the time of European conquest. As our starting point, we summarize in greater detail the new genetic information concerning its origin. We then examine the historical record for the period between the discovery of the Pacific Ocean by Spaniards and the trans-Pacific trade organized by the Spanish crown, and investigate the archaeological record, linguistic and ethnobotanical evidence. Finding that these latter lines of investigation are not fruitful, we re-examine the genetic and historical evidence, as well as five hypotheses concerning the origin of coconut in the Americas that have appeared in this century or so of debate: coconuts 1) originated in the Americas; 2) floated on ocean currents from Oceania; 3) were carried by Polynesians; 4) were carried by unknown mariners; and 5) were carried by Spanish galleons returning from the Philippines.
The Genetic Evidence
Two recent studies based on DNA evidence suggest that the origin of the reputedly pre-Columbian coconuts in Panama is in Southeast Asia, most probably in the Philippines (Baudouin & Lebrun, 2009; Gunn et al., 2011). Although the coconuts of the Pacific coast of Mexico and of Panama came from the same general region, they are readily distinguished using microsatellite markers (Baudouin & Lebrun, 2009). The origin of the Mexican coconuts is known: They were available directly from the Philippines (as well as some from other locations) starting in 1565 for a period of 250 years, when the Spanish crown inaugurated and managed the Manila-Acapulco trade route (Fig. 1). This trade route did not necessarily involve a large number of introductions nor a large number of nuts at any one time, as the coconuts were carried for consumption by the crew and passengers during a three-month journey. Any early germinators not used could be planted, but once an initial planting came into bearing, it would soon supply all future planting material on a year-round basis and no further introductions were needed until commercial demand for planting material occurred at the start of the 20th century. By contrast, the Panama Tall on the Pacific coast is characterized by markedly reduced diversity (expected heterozygosity=0.324 in this Panama Tall versus 0.579 for all Tall coconuts combined; observed heterozygosity=0.230 in this Panama Tall versus 0.480 for all Tall coconuts; Gunn et al., 2011). Such a reduction is the signature of a bottleneck, i.e., a strong reduction of population size at some moment in the history of a population, as is typical of a small founder event.
This founder event is clearly seen by examining individual microsatellite loci, such as locus CnCir 2 (Fig. 2). Only two of the seven alleles found in Southeast Asia exist in Panama and allele 222 represents 96 % of the total variation present in Panama (Baudouin & Lebrun, 2009). Allele 222 is also the most frequent in Southeast Asia, and its frequency is especially high in the Philippines (65 %). Its frequency decreases progressively from Southeast Asia to Melanesia, Micronesia and Polynesia. Such a tendency is observed at most of the 30 loci studied (data not shown; see Baudouin & Lebrun, 2009; Gunn et al., 2011).
Baudouin and Lebrun (2009) created a similarity index to integrate the allelic information across 30 microsatellite loci to compare different sets of possible source regions for the alleles found in the Pacific coast Panama Tall (Fig. 2). The Philippines variety is most similar to both this Panama Tall and to the Mexican Pacific coast coconuts, while the Polynesian island sources are the least similar, although they are geographically closer. This is the genetic basis for the affirmation that both the Panama and Mexican Pacific coast coconuts were introduced from the Philippines.
The limited similarity between the Philippines and this Panama Tall (0.475) is due to the extremely small founder population of the Panama coconut, and possibly different selection pressures since introduction (Baudouin & Lebrun, 2009). However, the initial selection pressures may have been uniform and closely related to trans-Pacific travel. The reduction in population size probably occurred because most of the nuts were consumed before arrival and the selection pressure was for the early germinators that survived to be planted, since these would have been less preferred for consumption. Thus, only a small fraction of the allelic diversity found in the Philippines is found in this Panama Tall, whereas the allelic diversity found in Mexico is quite similar to that found in the Philippines (Fig. 3). The similarity of Mexican and Philippine varieties is probably due to several introductions after 1565, especially to supply the toddy market in the 17th century (Zizumbo-Villarreal, 1996).
A Time Line of Spanish Contact with Coconut
The journal of Christopher Columbus (Cristobal Colombo) records that, on the 17th of November 1492, when sailing near Puerto del Principe on the north coast of Cuba: "There was a beautiful meadow and many very tall palms. They found a very large nut of the kind belonging to India, ..." (Columbus, 1893: 80). Columbus thought he had reached Asia and found coconuts, but he had misidentified the Royal palm (Roystonia spp). Some 30 years later, palms also thought to be coconuts were reported on the Pacific coast of Panama by Gonzalo Fernandez de Oviedo y Valdes, who was appointed as one of the official historians of the Indies (Spanish American colonies) in 1523, after his return from Santo Domingo (today's Dominican Republic). He published a first Summary entitled "Natural History of the Indies" in 1526, before returning to the Americas. Part of his "General and Natural History of the Indies" was published in 1535, but the whole General History was only published in 1851-1855, edited by J.A. de los Rios y Serrano.
Although Oviedo was not a naturalist, his "Natural History" is considered an important reference about the natural resources of the Americas. Throughout he makes clear that only part of the information was obtained personally, and part was obtained by interviews and correspondence, as the Spanish crown demanded information from all explorers who received crown support. Before and during the period that Oviedo was active as historian, information was flowing into the Iberian Peninsula at a great rate, as Spanish and Portuguese explorers visited new parts of the New and Old Worlds. It is important to remember that they did not explore the same areas because of the Treaty of Tordesillas, which effectively divided the non-European world between Spain and Portugal.
The case of coconut is emblematic of this flow of information, as Vasco de Gama was the first to bring coconut to Portugal in 1499, after leaving seedlings in the Cape Verde islands, where they grew to reproductive age and were distributed to the Americas (Harries, 1977). The conquistadors who sailed to the New World in the first decades of the 16th century knew that coconuts were special because in 1501 King Manuel of Portugal had written a letter to Ferdinand and Isabella of Spain, extolling the prime value of coconuts as a source of drinking water and cordage for sailing ships (cited by Harries, 1977; Patino, 1963, 2002). However, the fast complete information available in Iberia about coconut uses in Asia was published by Ludovico di Varthema (1510), based on his travels in Asia between 1501 and 1508. His report would have been read by Magellan before setting out in 1519 to navigate the globe and by Oviedo before going to Darien (Caribbean coast of Panama) in 1521. Hence, as Patino (1963, 2002) emphasizes, Oviedo and others of the period used information from Asia and the Americas in their reports without specifying the origin of each detail, a common occurrence before modern citation methods were developed. What has not previously been well recognized is that no Spanish explorers or historians had ever seen growing and fruiting coconut palms because the Treaty of Tordesillas prevented them from sailing to Asia, although some may have seen the nuts in Lisbon or other Iberian cities.
Zizumbo-Villarreal & Quero (1998) analyzed Oviedo's Summary and General History to determine if coconut was present in Panama at the time of European conquest. Patino (1963; 2002) analyzed both of these documents and the origins of the information that Oviedo used to prepare his Summary and General History, although he recognized that many documents have been lost and some were unavailable to him even after 40 years of effort. Both Patino and Zizumbo & Quero cautiously accept that Oviedo's reports are correct. Based on Patino's analysis, this is the time line for the early Spanish references to coconut in Panama.
1492--Discovery of New World
1494--Treaty of Tordesillas divided the world outside Europe between Portugal (eastwards) and Spain (westwards), which effectively prevented Spanish mariners from sailing to Asia (until 1580).
1501--King Manuel of Portugal wrote about coconut in detail in a letter to Queen Isabella and King Ferdinand of Spain.
1513--Vasco Nunez de Balboa crossed the Isthmus of Panama and discovered the Pacific Ocean. The expedition explored only a small area and returned without mentioning coconut, although numerous documents of this expedition have been lost.
1514-1515--Alonso de la Puente (royal treasurer in Darien, on Panama's Caribbean coast--heard of coconut) and Rodrigo Colmenares (ship pilot and interpreter--saw coconut) visited the Pacific coast of Panama and were interviewed by Pedro Martir de Angleria, another official crown historian, in late 1516 in Spain. Angleria's report was first published in 1530 (translated in 1944), but Oviedo knew Angleria and had access to his report, which explains why this information is in the Summary. The explorers Puente and Colmenares reported on the coconut (that they thought to be the same fruit as cultivated in Calicut--the common term for India in the first decades after European contact) and go on to state that it was cultivated on some of the islands in the Gulf of Panama. The fact that they thought it to be the same as the coconut from India is significant, but does not mean that it was indeed coconut.
Angleria states that coconuts were observed in three places along the Pacific coast: Chiman, to the southeast of the current City of Panama; Nata, to the west of the City of Panama; and Burica, further to the west at the frontier with modern Costa Rica (Fig. 1, Panama insert). Chiman is thought to be the first place that Europeans saw coconut and Patino even suggests that Balboa's expedition may have seen it there also. Nata was reputed to have an abundance of coconuts, reported by the natives to genninate and grow spontaneously along the shore, although they also reported young plants were transplanted. Angleria also reports that "Some think that sea currents bring the seeds of these trees from unknown regions."
1519--Gaspar de Espinosa y Luna (explorer, participated in the founding of the City of Panama in 1519) explored west of the City of Panama as far as the point of Burica, where he reported "Many beautiful and large mameys (Pouteria zapota) and many palms with the large coconuts, ..." It is not clear exactly when Angleria obtained this information from Espinosa. Patino suggests that Espinosa may only have realized the importance of his observation after learning about coconut in Asia.
1524--Francisco Pizarro Gonzalez (mayor of the City of Panama 1519-1524) led an expedition that explored the South American coast from the City of Panama to the northern limits of Colombia. Near the Cape of Corrientes (Fig. 1), they reported "a large quantity of coconuts" at La Candelaria Bay, which are reported in Oviedo's General History, but not in the Summary.
1526--Juan de Cabezas (pilot) discovered Cocos Island (Fig. 1) and provided information used by Oviedo in the full General History.
1539--Alvaro de Guijo (resident in the City of Panama) sent nuts to Hernan Cortes de Monroy y Pizarro (conqueror of Mexico), along with advice about how to plant them. In his letter that accompanied them, Guijo wrote: "I have heard that you do not have this fruit we call cocos, so I sent a boat to a place along the coast to collect some, and I send you two dozens of ripe fruits. Some of the riper ones can be sown by placing half the fruit in the ground. If you like them, I can send more another time, as well as some already germinated." This letter is cited in Bruman (1947), but Zizumbo-Villarreal & Quero (1998) caution that this possible introduction to Mexico may not have been successful. In fact, Hernan Cortes left for Spain in 1541 and never returned to Mexico.
1549--Coconuts from the Cape Verde islands (Portuguese) planted in Puerto Rico (Spanish), attributed to Diego Lorenzo, canon of Cape Verde (Bruman, 1944).
1553--Coconuts from the Cape Verde islands planted at Bahia, Brazil (Bruman, 1944), possibly the first coconuts on continental America.
1565--Philippines to New Spain (Mexico) return trip was accomplished independently by Alonso de Arellano and Andres Urdaneta, and the Manila-Acapulco commercial route was established by 1571 and continued until 1815 (Fig. 1). Coconuts, carried for consumption by passengers and crew, became a regular item but were not always mentioned in the ships' manifests.
1569--Alvaro de Mendana introduced coconuts from the Solomon Islands to Colima, Mexico (Sevilla de Rio, 1974).
1580--Treaty of Tordesillas became invalid when Portugal was ruled by Philip II of Spain.
The Archaeological Evidence
The central Pacific coast of Panama has a long cultural history, with human activities present in the archaeological record since 11,000 years before the present (BP) (Piperno & Pearsall, 1998). By the late Pre-Ceramic Period (7000-5000 BP), horticulture started to become important, expanding rapidly through the Early Ceramic Period (5000-2500 BP) until most subsistence was horticultural by the Middle Ceramic Period (2500-1500 BP). The Ceramic Period extended until European Conquest (500 BP). Nonetheless, subsistence still included gathering, especially of palm fruits and other useful plants.
The macro-and micro-archaeological records show that palms were very important to the subsistence of the Native Americans in the Gran Chocle (central Panama) and Gran Chiriqui (western Panama) regions. The macro-archaeobotanic remains from these regions include Acrocomia aculeata (Jacq.) Lodd. ex Mart., Attalea butyracea (Mutis ex L.f.) Wess. Boer, Bactris major Jacq., Elaeis oleifera (Kunth) Cortes and Astrocaryum sp. (Dickau, 2010). All of these species are members of the Cocoseae tribe within the Arecaceae, but Acrocomia, Astrocaryum and Bactris are spiny and like Elaeis have fruits that are morphologically clearly distinct from coconut. The Attalea butyracea is especially important to our discussion because it has relatively large fruits (4.5-8.5 cm long by 3-4.5 cm wide) that look somewhat like very small dried coconuts, and the palm has a similar stature to that of coconut (Henderson et al., 1995), permitting the supposition that some of the first reports may have confused this Attalea with coconut. A. butyracea grows on the slopes above beaches and river courses along the Pacific coast of Panama in many places today.
Large numbers of endocarp fragments of Attalea butyracea are reported at numerous locations in the Gran Chocle and Gran Chiriqui regions of western Panama from 7000 BP to European conquest (Smith, 1980; Cooke and Ranere, 1992; Dickau, 2010). Phytoliths, which are microscopic silica bodies found within and between cells in many plants, of palms are also abundant in the same region during the same period (Dickau, 2010).
It is important to mention that most of this work was carried out in the central and western parts of Panama. In terms of the Pacific coast, the Gran Chocle and Gran Chiriqui regions of Panama extend from near Nata on the coast of Parita Bay west of Panama City to near Burica on the coast of the Gulf of Chiriqui, at the frontier with Costa Rica (Fig. 1, Panama insert). Specifically, Dickau (2010) mentions finds at La Pitaya, a small coastal island about 50 km east of Point Burica, and several coastal sites about 50 km southwest of Nata (Vampiros, La Mula Sarigua, Monagrillo, Cerro Juan Diaz), but she does not report coconut. In comparison, the Gran Darien region of eastern Panama is much less well studied. Hence, none of the archaeological sites reported are close to Chiman or La Candelaria Bay.
In summary, no archaeological evidence exists for coconut in Panama before European conquest, as observed by Morcote-Rios and Bernal (2001: 342), even though "the woody endocarp of the coconut is an appropriate material to be preserved at archaeological sites in wet environments." However, the absence of evidence is not necessarily proof of absence, as the archaeological record tends to be determined by sampling intensity, artifact preservation, and luck. Several archaeological sites are close to historically mentioned sites, but none are exactly at those historically mentioned sites, so future work may reveal new information.
The Ethnobotanical Evidence
Although many large palms have been important in Native American subsistence in Panama since well before European Conquest, there is no record of use of coconut by native Panamanians at the time of conquest (Patino, 2002), although the native Panamanians apparently were familiar with its propagation (see Time Line 1514-1515 above). Patino (2002: 253) writes: "There is no doubt that the inhabitants lacked the tradition about the use of this plant; ... Would one have to conclude that these Nations met this palm at least four generations ago, but did not use it?" Patino states that there would have been no reason for coconut not to be culturally assimilated by the native population quite rapidly, even if it were growing in relatively remote locations. This is a remarkable anomaly given Native American uses of a wide variety of palm species, including endosperm, and even the domestication of peach palm (Bactris gasipaes Kunth; Clement, 1995). This dramatic lack of evidence is a strong reason to doubt the accuracy of the first Spanish reports, or at least to assume coconuts could not have arrived in Pacific Panama more than a few years before European Conquest.
The Linguistic Evidence
Paleobiolinguistics uses the comparative method of historical linguistics to reconstruct the biodiversity known to human groups of the unrecorded past (Brown, 2006a, b, 2010). By comparing words for biological species in languages of the same language family, paleobiolinguistics facilitates reconstruction of terms for plants and animals in the vocabularies of their ancestral or proto-language. Reconstructed words for species are indicative of their substantial significance to speakers of protolanguages. The approach, then, has the potential to contribute to the discussion of the antiquity of coconut in Panama by comparing words for the species in genetically related modern languages spoken there and in adjacent areas.
The modern languages of Panama include Spanish, Creole, and Amerindian languages affiliated with two language families, Chibchan and Choco (Fig. 4). The Chibchan languages of Panama include (from west to east) Teribe, Ngabare, Buglere, and Kuna. Other Chibchan languages are spoken in Honduras, Nicaragua, Costa Rica, and Colombia. The smaller Choco language family includes Embera and Woun Weu, both having dialects spoken in Panama and Colombia.
A large portion of Panama is empty of contemporary Native American languages (white areas of Fig. 4), notably the western Pacific coastal area and central parts of the country. This does not mean that no indigenous languages were spoken there in the past. Archaeological sites in parts of the area attest to Native American occupation, perhaps as late as the 19th century (Locascio, 2010). However, very little, if any, evidence bears on the language or languages of these archaeological populations. Kaufman (1994) published a time-of-contact language-distribution map for the Caribbean region that includes southern Central America. On that map much of the contemporary empty regions of Panama (Fig. 4) is identified as being filled by speakers of Mobe, Doraske, and Bokota (in the west) and Kuna (in the east), all Chibchan languages.
For this study, we undertook a paleobiolinguistic investigation of domesticated and useful plants, including coconut, in both Proto-Chibchan and Proto-Choco. Plants that reconstruct for Proto-Chibchan, spoken at the latest around 4400 BP (Holman et al., 2011), include cacao (Theobroma cacao), cotton (Gossypium hirsutum), hog plum (Spondias spp.), maize (Zea mays), manioc (Manihot esculenta), sweet potato (Ipomoea batatas), and tobacco (Nicotiana tabacum) (cf. Constenla Umana, 1981). Those reconstructed for Proto-Choco (c. 2258 BP) include Banisteriopsis caapi, bottle gourd (Lagenaria siceraria), guava (Psidium guajava), maize, manioc, and sweet potato. No words for coconut reconstruct for these prehistoric proto-languages, suggesting that the species was not known to their speakers.
Comparative evidence is such that a word for coconut is not reconstructable for any chronological stages of Chibchan and Choco language families, including those closest to the time of the European conquest. For example, a term for coconut cannot be retrieved for Proto-Embera, a daughter language of Proto-Choco spoken at the latest around 875 BE This is due to the fact that all contemporary offspring languages of Proto-Chibchan and Proto-Choco for which lexical sources are available fail to show native terms for coconut that are cognate. In fact, all of the terms for the referent in available lexicons are of non-native origin. All languages except one or possibly two have borrowed words for the plant and its fruit from a European language, either Spanish (coco) or English (coconut).
Twelve terms for coconut, presented in original orthography, were extracted from lexical sources available to us for Chibchan and Choco languages (Table 1). All but two terms in the 12 languages are unambiguous loans from European languages. The two exceptions are Boruea siahua and Kuna ogoba, both of which are so phonologically dissimilar to one another as to exclude the possibility of cognation and, thus, the possibility that a word for coconut with such reflexes pertained to Proto-Chibchan. In fact, the phonology of the Kuna word is such that its status as a loan based on Spanish coco cannot be ruled out as a strong possibility. Of the remaining 10 terms, nine are loans based on coco, and one (in Rama) is a loan from English (coconut).
Words for coconut in the 12 languages robustly suggest that the plant was introduced by Europeans into southern Central America. Speakers of native languages of Latin America typically have named newly encountered items, including plants, animals, and artifacts, by borrowing words for these things from languages of the people that introduced them, in this case from Spanish and Portuguese (Brown, 1994, 1999). On the other hand, only rarely have native terms for indigenous things been replaced by Spanish or Portuguese loanwords (Brown, 1999: 92-104).
However, occasionally Latin America Indians have coined words for introduced items by using the lexical resources of their native languages rather than by borrowing a term from a European language. This practice has been very common in languages spoken by Amerindians influenced by English, French, and Russian intruders, but comparatively rare among native languages of Latin America (Brown, 1994, 1999). For example, in the Bachajon dialect of Tzeltal, a Mayan language of southern Mexico, the introduced sheep is tumin cix, literally "cotton deer," a usage almost certainly motivated by the resemblance of the European sheep to the native deer, the most salient mammalian herbivore known to Tzeltal speakers (Witkowski & Brown, 1983).
The only word for coconut of the 12 languages that clearly is not a European loan is Boruca siahua. Plausibly, this term originally denoted a native palm similar to the introduced coconut. Evidence for this comes from Cabecar, which, like Boruca, is a Chibchan language of Costa Rica. Cabecar contains a word similar to the Boruca term, i.e., sera, designating Acrocomia aculeata, a palm whose fruit is fed to cattle and occasionally consumed by humans in the Cabecar region. The phonological similarity of Boruca and Cabecar words may be due to cognation or, if not, to borrowing. In either case, this suggests the original referent of the Boruca term was A. aculeata, a word that later became referentially extended to the introduced coconut. No term for coconut is listed in Margery's (1989) exceptionally thorough dictionary of Cabecar, and no term for A. aculeata is found in Quesada Pacheco and Rojas Chaves's (1999) comprehensive dictionary of Boruca.
In conclusion, a pre-Columbian presence of coconut in Panama and surrounding areas is not attested by paleobiolinguistic evidence. Indeed, this evidence strongly suggests that the modern occurrence of the plant in the region is accountable to European introduction in historical times.
We have also undertaken a preliminary paleobiolinguistic survey of many language families of Latin America for evidence of pre-Columbian coconut, mainly focusing on the reconstructibility of terms for the plant in proto-languages. With only one possible exception, no such terms are apparent. The possible exception is Proto-Chinantecan (c. 1935 BP), whose contemporary offspring languages are spoken in northern Oaxaca state, Mexico. For this ancestral language, Rensch (1989:78) reconstructs *[ha:.sup.H] ([sup.H]=high tone), assigning to it the gloss coconut. However, a survey of reflexes of this hypothetical word in various Chinantecan languages shows that some of these denote Acrocomia aculeata rather than, or in addition to, coconut. This suggests that, like the Boruca term discussed above, the Proto-Chinantecan word designated A. aculeata, and that its reflexes were referentially extended to the introduced coconut. If so, paleobiolinguistics evidence assembled to date fails to provide support for the prehistoric occurrence of coconut for any region of Latin America. This finding concurs with the work of Merrill (1937), who contrasted abundant linguistic evidence for a long human association with coconut in Southeast Asia and Oceania with the lack of anything similar for the Americas.
A Preliminary Summary of the Evidence
The genetic evidence concerns modern coconuts, so says nothing about the historical presence of coconut in Panama. The historical record had previously been accepted with caution (Patino, 1963, 2002; Zizumbo-Villarreal & Quero 1998), but the clear absence of archaeological, ethnobotanical or linguistic evidence suggests that the caution was warranted. This lack of interacting evidence is an unexpected result when applying de Candolle's and Vavilov's methodology, and suggests either that coconut arrived immediately before European Conquest, rather than 2,250 years BP as suggested by Baudouin and Lebrun (2009), or it arrived after European Conquest. Either way, both the genetic evidence and historical records need further study.
Genetic Sampling and Expanded Analysis
In any study of the genetic relationships among populations, such as coconuts in Panama, Mexico and the Philippines, the sample used will affect results. Hence, we examine the samples used in the genetic studies, especially those from Mexico and Panama. The coconut data set was developed within the framework of the Generation Challenge Program (GCP) of the Consultative Group on International Agricultural Research by Luc Baudouin and Patricia Lebrun, of CIRAD, the French Agricultural Research Center for International Development, Montpellier, France. This data set was not designed primarily to identify the origin of the Pacific coast Panama Tall, although Baudouin and Lebrun (2009) used it this way with interesting results. It is worth mentioning, however, that sampling in Oceania, including the Philippines, does not yet represent the variability that is present in the region either (Fig. 1, contrast between shaded areas and the Pacific Ocean). Even the Philippines, where coconut is a major crop today, is not well represented (Table 2).
The varieties in the GCP data set that interest us most are those from the Philippines, Mexico and Panama (Table 2). Dwarf varieties from the Philippines are not included in the table, as these are unlikely to have contributed to this Panama Tall's genetic composition, nor is the Mexican Atlantic Tall, as this variety originated in the Indo-Atlantic group of varieties (Harries, 1977; Gunn et al., 2011).
The sample of Mexican Pacific Tall coconut varieties used by Baudouin and Lebrun (2009) and Gunn et al. (2011) includes several with known origins and dates of introduction, but some plants have been joined into "state-level" varieties without due consideration of their morphology (Table 2). The Pacific Tall of Colima is a mixture of 11 plants that have Rennell Tall fruit morphology and three that appear to be from the Philippines. The 11 Rennell Tall-type plants are derived from a plantation established in 1890-1900 at the margins of the Coauhuayana River in Tecoman, Colima state. It is represented in the vast majority of plantations in Colima, the western portion of Michoacan and eastern Jalisco states. The seed originated in the environs of the city of Colima, from orchards established in the 18th century. They may be progeny of the introduction made by Alvaro de Mendana from the Solomon Islands in 1569 (Zizumbo-Villarreal & Colunga-GarciaMarin, 2001). The Pacific Tall of Guerrero includes eight coconuts that are morphologically similar to the San Ramon variety and three that are similar to the Baybay Tall variety. No historical record of these introductions exists, but they agree with the Philippine location of probable early introductions (see below). The Pacific Tall of Michoacan was introduced by President Lazaro Cardenas, apparently from the Philippines, and was initially established in plantations in 1937-1938. This is the most representative variety between Acapulco and Lazaro Cardenas Port, Guerrero state (Zizumbo-Villarreal & Colunga-GarciaMarin, 2001). The Nuxco plantation was established in the 1950s, possibly from progeny of the early introductions to Acapulco from Philippines. All of these varieties have clearly different microsatellite profiles when compared to the Pacific coast Panama Tall (Gunn et al., 2011: Table S1).
Numerous samples of Panama Tall from the Pacific coast were taken into the Caribbean during the 20th century and were then distributed elsewhere. When the Maypan hybrid (a cross between the Malayan Dwarf variety and the Panama Tall) was produced in Jamaica (Harries & Romney, 1974) it was resistant to lethal yellowing, the most important coconut disease in the Americas, and the Panama Tall became very important in the Caribbean (Harries, 1995). However, not all coconut populations along the Pacific coast of Panama were sampled. For example, the sample used by Baudouin and Lebrun (2009) and Gunn et al. (2011) did not include plants from the Gran Darien nor the Gran Choco (present day Pacific coast of Colombia), where historic sources place the pre-conquest presence of coconut at Chiman and Cape Corrientes, respectively. Thus, the sample is biased towards the middle-western portion, covering the Gran Chocle and Gran Chiriqui regions. Note also that these middle-western regions are precisely those where Native American populations have disappeared (Fig. 4), so one might expect coconut introductions from other localities during the colonial and modern periods. Unfortunately, we have not found a history of these introductions.
Additionally, the various samples of Panama Tall from the Pacific coast listed above (Table 2) were not collected directly in Panama for the GCP study. Rather, they had been collected years earlier and taken to various countries before having their DNA extracted (Luc Baudouin, pers. com., 2011), which explains part of the Indo-Atlantic alleles found in some plants that is evidently due to introgression (see Gunn et al., 2011: Table S1).
The two genetic studies were both well executed for the questions that they asked, but neither took the opportunity to look more closely at the variability within the countries involved. Part of this is due to the very small sample sizes (Table 2), especially in the Philippines, but also Mexico and to some extent even Panama, because this will determine the reliability of the relationships found. Nonetheless, a look at these relationships can offer ideas for new studies. Hence, we extracted the microsatellite genotypes for the Philippines, Mexico and Panama from the Global Challenge Program dataset, and used Nei et al.'s (1983) genetic distance, the Neighbor Joining algorithm, and 1,000 bootstrap iterations to obtain a preliminary idea of relationships. Because sample sizes are small and all bootstrap confidence levels are weak, we do not present the dendrogram. However, there seems to be a relationship between the Philippine San Ramon variety and the Mexican Pacific Tall of Colima, which is curious because this Mexican variety may have originated in the Solomon Islands (Table 2). All the other Mexican varieties appear to be derived from this relationship and are a sister group to the Panamanian varieties. The relationship with all the other Mexican varieties may be due to the close relationship of San Ramon with a small set of Philippine varieties (Baybay, Pandan, Tagnanan), one of which (Baybay) is morphologically similar to the Pacific Tall of Guerrero (Table 2). The modern Pacific coast Panama Tall is also morphologically similar to San Ramon (Vargas & Blanco, 2000). Hence, this preliminary analysis suggests an introduction of coconut to Mexico that then influenced other introductions and was also introduced to Panama, with the very small founder event detected by Baudouin and Lebrun (2009).
To look even more closely at how the Panama Tall may be related to the Philippine and Mexican varieties, we extracted the Structure assignments at K=5 from Gunn et al.'s (2011) Table S1 (Fig. 5). Four Philippine varieties (San Ramon, Baybay, Pandan, Tagnanan) have numerous plants with considerable proportions of Panama assignment, so even though San Ramon is identified as the most probable ancestor of the Pacific coast Panama Tall, other varieties could have contributed and would not be easily detected because of the extremely reduced size of the Panama Tall founder event. From the Structure analysis for the Mexico Pacific Tall of Colima, it is possible to hypothesize that the proportions of Panama and Papua New Guinea assignments explain its similarity with San Ramon, even though the high proportion of South Pacific confirms the origin of Colima in the Solomon Islands. The Mexican varieties also have considerable proportions of Panama assignment, although less than the Philippines overall. Nonetheless, numerous plants in this small sample have enough Panama assignment to suggest that the Panama Tall is derived from the Philippines via Mexico. We then took the GCP dataset and analyzed the Philippine, Mexican and Panama varieties with Structure 2.3.3 (Pritchard et al., 2000; Hubisz et al., 2009), following Gunn et al.'s (2011) parameters. This generated three groups (data not shown), with all Philippine and Mexican varieties except the Mexican Atlantic Tall in one group, all Panama varieties in one group, and the Mexican Atlantic Tall in the third group. The San Ramon variety had some Atlantic assignment, as is also visible in Fig. 5, and there is some admixture of Atlantic with Panama. This new analysis strongly supports the Philippine--Mexico connection, but is less clear about showing relationships with Panama, given the fact that only the Philippine varieties represented the whole Pacific Ocean.
The similarity of Pacific coast Panama Tall and named varieties in Southeast Asia has long been noted (Harries, 1978; Zizumbo-Villarreal et al. 1998; Zizumbo-Villarreal & Quero, 2005), in particular with the San Ramon type in the Philippines (Vargas & Blanco, 2000). Considering the lack of archaeological, ethnobotanical and linguistic evidence for coconut in Panama at conquest, it seems reasonable to ask how the San Ramon type might have arrived in Panama. The Manila-Acapulco galleon route (Fig. 1) that had been suggested by Safford (Small, 1929; Hill, 1929), although post-conquest, is the obvious candidate. Although it had been accepted by Merrill (1954) and others (Harries, 1971, 1978; Zizumbo-Villarreal et al. 2005), it has not been closely examined until now (see also Harries, 2012).
In 1564 an expedition to establish a Spanish settlement set out from Puerto de la Navidad (Barra de Navidad), Jalisco, Mexico to avoid sailing through Portuguese waters, setting a course for Cebu in the Visayas, where Magellan had landed 41 years previously. Manila subsequently became the premier city of the Philippines, because it already had trade links with China and Japan and, as the northernmost harbor, it became the departure point for the galleon route. However, on the first occasion in 1565, Alonso de Arellano and Andres Urdaneta, returning independently, both sailed from Cebu and not from Manila, which is significant because there were superior coconuts in that region. An agricultural observer in the seventeenth century, Father Francisco Ignacio Alzina, who resided in the Visayas from 1634 to 1667, wrote: "There are very big ones [coconuts] which would measure more than one azumbre" (~21) (Alzina 1668). At the head of the Sulu Sea, Cebu in the Visayas was not far from Mindanao and, when the production of copra became commercially important in the Philippines at the beginning of the 20th century, the San Ramon coconuts in Mindanao were highly regarded. This was because "... there are no records from any other part of the world of plantation averages showing such size of nut as those of San Ramon ..." (Copeland, 1914). However, similar coconuts were reported from Colombia at the same time: "Gorgona Island between 3rd and 5th parallel N of Equator 24 miles off Colombia ... is famous for producing coconuts of immense size and are of great use to planters as seed nuts ..." (Bardy, 1914).
The possibility that coconuts were carried to Mexico in 1565 has previously been discounted (Bruman, 1945), because coconuts were not recorded on the list of provisions. However, Arellano's account of cooking oil solidifying (literally "freezing") is strong circumstantial evidence of coconuts as deck cargo for the crew to drink or use when preparing food (Harries, 2012). The early germination of this type--more than 75 % in 105 days (Harries, 1981)--would have meant that, in August 1565, at the end of a four-month voyage to Puerto de la Navidad, there would be seedlings ready for planting. The Barra de Navidad lagoon borders Colima province, which became, and is still today, the center for coconuts in western Mexico. It also seems reasonable to suppose that the subsequent Manila-Acapulco galleons would carry the same sort of coconuts. This could be done, either by arranging for them to be collected in Mindanao or the Visayas for transshipment or, more easily, by plantings near to Manila. This would explain Copeland's note that San Ramon coconuts were in general cultivation in the coastal district of Pangasinan province, Luzon and a report that "the largest nuts in the world are produced around Lingayen Gulf, Luzon Island" (O.W. Barrett cited by Smith & Pape, 1914: 537). These locations, to the north of Manila, would have been the most convenient for taking deck cargo on board. Confirmation that seedlings were planted in Mexico comes from the activity of skilled Filipino toddy tappers in 1580, who tapped "tuba" for fermenting to "coconut spirits" for consumption in Mexico (Zizumbo-Villarreal 1996; Zizumbo-Villarreal & Colunga-GarciaMarin, 2008). As first the tuba market and in the 20th century the copra market expanded, later introductions of larger numbers of seednuts intended only for planting would have been made.
Thus, the Panama Tall coconuts could have come from one or more locations, but still represent the San Ramon and closely related varieties. It is even likely that the seedlings from the post-t565 introductions into Mexico were fully grown and in bearing within five to seven years and the year-round production of seednuts, rather than the once yearly supply from Manila, would be disseminated from Colima southwards as far as Peru, including Gorgona Island, Colombia (Fig. 1), for example. A single, small sample from this source might explain the extremely narrow genetic base of the Panama Tall reported by Baudouin and Lebrun (2009). Zizumbo-Villarreal and colleagues in Mexico are collecting and analyzing the Pacific coast Tall varieties there, as well as in other locations along the Pacific coast southwards. Further DNA analysis of the San Ramon and similar varieties in the Philippines and their comparison with the Pacific coast Tails of Mexico and Panama will permit these relationships to be refined with much better precision than is possible with the current dataset.
Botanical and Historical Questions
Given the doubts about the historical record, we revisit Oviedo's account because, although it was previously analyzed by Patino (1963, 2002), Allen (1965) and Zizumbo-Villarreal & Quero, (1998), the DNA data match between the Panama and the Philippines coconuts raises flesh concerns. Oviedo's account of coconut has been questioned on a number of occasions because the illustration does not match the description, parts of the description can match other palm genera and parts of the information may have come from Asia rather than the Americas, as indicated above. Patino (1963, 2002) discards the first problem by pointing out that Oviedo finished describing the cane palm, Bactris major, on the page where he started describing coconut. While Patino and Zizumbo-Villarreal & Quero looked at the similarities and cautiously accepted that Oviedo had seen coconuts, Allen saw differences and suggested that Oviedo may have regarded them as aberrant and inferior sorts. We consider their opinions and identify two properties, found only in genuine coconuts, which Oviedo did not mention and apparently did not know about. The following sections of Oviedo's account (as translated for use by Allen, 1965) are worth transcribing and commenting on.
"There are other pall trees whose fruit are called cocos, this being a genus of large palm trees, and whose leaf is of the same kind as that of the date palls, ..." This is the first sentence of the account and suggests that numerous Neotropical palm fruits were called cocos, which is at odds with our linguistic analysis. Hence, what we may be seeing in this sentence is the Spanish naming of native American pall species. Over the next centuries, 131 palm species or sub-species were attributed to the genus Cocos. These other palms are now assigned to different genera (Beccari, 1917; Henderson et al., 1995) and Cocos nucifera is monotypic. Some of them have statures similar to coconut, with pinnate leaves, although many have spines that are hard to miss (Acrocomia, Astrocaryum, Bactris). The spineless ones are now in Attalea, which has the appropriate stature, fruit that look somewhat like small coconuts (see archaeological evidence above), and today grow on the hills above the beaches of Panama.
"These trees or pall trees put forth a fruit which is called coco, ... Altogether, such as it is on the tree, it has a much greater bulk than a man's head; ..." A century before Columbus crossed the Atlantic, the Arab Ibn Battuta (1929) visited East Africa and India, and reported that coconut is the size of a man's head, so this kind of information was available to historians in Iberia at the time that Oviedo wrote, as was Varthema's (1510) account.
"But there in these our Indies the Indians do not trouble to cure these cords and cloths which can be made from the wool or burlap of these cocos, such as in the Levant, for here there is much cotton and henequen and cabuya to supply such necessity for cords." These comments by Oviedo are with respect to the fibrous mesocarp of coconuts and a clear indication of the lack of use by Native Americans, as noted above with respect to the lack of ethnobotanical evidence. The description appears to be a justification for lack of use, since the Native Americans had numerous other good fibers.
"This fruit which is within that burlap, is the coco, as big as the fist of a closed hand, and some as big as two fists, and more or less, it is a sort of round nut, and some are elongated. The crust is hard, and as thick as the width of the inscription [title] on a Castillian silver real [coin]. Inside, attached to the crust of that nut or coco, is a meaty part in width like half the thickness of the small finger of the hand, or as thick as a writing quill of the kind common to geese." This part of the description is somewhat at odds with the original description of the size of the fruit, so if these were coconuts they had small nuts, quite unlike the Panama Pacific Tall or San Ramon varieties.
"This is the fruit proper of the coco and what is edible, and it is as white as a cleaned almond and better tasting than almonds, and of smooth taste to the palate, It is eaten the same way peeled almonds might be eaten, ..." This is a good description of the endosperm of a coconut, but will be contrasted below with another description.
"By way of pith or marrow of this fruit which is in its middle, ..., is a place taking up the remaining part or entire quantity of the coco, full of a most clear and excellent water, and as much as would fill the shell of the egg of a hen, and more or less, in proportion to the bigness or size of the coco: ..." Again, this is a good description of coconut water in an immature coconut. However, Oviedo does not mention the sound of the water splashing in the cavity of a mature coconut when shaken, as the water is naturally absorbed when the nuts mature.
"After I wrote the report I have mentioned, I was in the province and headland of Borica, and I ate some of these cocos and carried many with me to Nicaragua, and came to loathe them, and others did as I did and said the same thing as well." This statement was not cited in Patino (1963, 2002) nor in Zizumbo-Villarreal & Quero, (1998), although both read it since they mention Oviedo going from Burica to Nicaragua. Some individuals do find the kernel indigestible and may become tired of it, as Allen pointed out, but that is not usually a group phenomenon. Note also Oviedo's choice of "ate" rather than "drank," as it suggests that they were fruit with little water. Oviedo's use of the word aborreci (loathe) is a surprising reaction--more people would agree with Charles Darwin (1860: 407): "After walking under a burning sun, I do not know anything more delicious than the milk of a young cocoa-nut." Perhaps the water had been absorbed by early germination, typical of Panama Tall (Harries, 1981), but Oviedo does not mention the soft, sweet and very edible haustorium (or "apple") that would immediately identify a real coconut. Nor does he say if they were being taken to Nicaragua for planting. So, the question is: What palm was this? The answer to this question has implications for the entire time line presented above and for the initial Spanish contacts with coconut in the Americas. It is also possible that this is a mixture of information from Asia (size, endosperm, water, flavor, etc.) and the Americas.
The list of palms near the Pacific coast of Panama with large enough stature to be confused with coconut by non-specialists was presented in the archaeological evidence (above). Even non-specialists would note and comment on spines on the trunk and leaf petioles, as pointed out by Allen (1965), so the spiny species can be discarded. That leaves Attalea butyracea, whose fruits can superficially be confused with very small coconuts. The seeds of A. butyracea do not have liquid endosperm when ripe, but they are perfectly edible. Like coconut, some people may not like the flavor of the seed, and any palm seed, including coconut, can go rancid if stored in conditions that do not allow germination but do allow respiration, but Oviedo's account does not suggest this. Coconuts on the open deck of a boat would germinate rather than rot and, if they had been present in Panama at this time, they would be regularly carried for refreshment on any coastal craft.
Moving on from Oviedo, there is a logical question related to the 1539 letter to Hernan Cortes (see Timeline above): If coconuts were interesting enough to send from Panama to Mexico, why weren't they also sent to Madrid--at least to show at court? Given the lack of this kind of display, it seems likely that 20 years after the first report there still were no real coconuts on the Pacific coast of Panama.
A Second Summary of the Evidence
The genetic and morphological evidence clearly shows that the Panama Tall is closely related to varieties from the Philippines and the San Ramon variety is a likely candidate, but the microsatellite information suggests that the original sample was very small and did not capture the full San Ramon genetic profile. The San Ramon variety was probably introduced into Mexico early via the Manila-Acapulco route. The historical evidence is not as clear as might be hoped for and appears to mix information from Asia about real coconuts with information from the Americas about other palms. With this summary in mind, we look at the hypotheses about how and when coconut arrived along the Pacific coast of Panama.
Hypotheses about Coconuts in Pre-conquest Panama
Five hypotheses have been presented to account for the historical observations, three of which (2-4) are not mutually exclusive. These are:
Hypothesis 1 Coconuts originated in the Americas. This was held by de Candolle at first (1855), but discarded later (1883). The hypothesis of American origin was reinstated by Cook (1910), strongly criticized by Beccari (1917) and Merrill (1937), and thereby decisively debunked. Gunn (2004) placed a final nail in its coffin with a molecular genetic phylogeny of the Cocoseae.
Hypothesis 2 Coconuts floated from mid-Pacific islands on one or many occasions from ancient times to the present day. This hypothesis is based on the fact that coconuts can disperse by floating over some distance, washing onto a suitable shore, striking root and growing. But, given the genetic relationship between the Pacific coast Panama Tails and coconuts in the Philippines, two questions arise: Could they have floated across the Pacific Ocean (perhaps via intermediate islands), or must they have been carried? Both cases have been argued by many scholars. For example, Bruman (1944) and Purseglove (1972) argue for natural dispersal, while Dennis and Gunn (1971) argue for carriage by man. A crucial question is the period for which coconuts will remain viable when floating in the sea. In two experiments to test the viability of nuts floating in the sea, the longest period any nut floated and remained viable was 110 days (Edmondson, 1941; Ward & Allen, 1980). This suggests that the maximum flotation period may be on the order of four months. In an experiment at the Coconut Experimental Station in Sulawesi in 1931 (Reyne, 1948), nuts were floated in barrels of sea water, but the water temperature would have been significantly higher than actual sea temperatures. The lower temperatures of floating coconuts delay germination compared with nuts at average ground level temperatures (Ward & Allen, 1980).
Computer simulations of Pacific wind and surface current directions and speeds, islands and coasts allowed tests of the possibility of coconuts drifting across the Pacific from several possible starting places (Levison et al., 1973; Ward & Brookfield, 1992). It is important to note that, as a coconut floats, the husk absorbs water, the nut's weight increases, it floats deeper in the water and the effect of wind relative to current decreases. In the simulation experiments, variations were used to cover current only, wind only, and combined wind and current forces. Experiments were conducted for drift periods of 4, 6 and 8 months. The last two periods far exceed any known example of the time a coconut might retain viability when floating.
Initial simulation experiments from Ducie and Reao, in the extreme east of French Polynesia (Fig. 1), showed that nuts floating from Polynesia had no chance of drifting to the Americas from that part of Polynesia, even within 8 months (Ward & Brookfield, 1992: 473-4). The Equatorial Counter Current offers the most likely possibility for coconuts drifting from west to east and therefore starting points used for the remaining simulated eastward drifts from Micronesia were Christmas Island (in eastern Kiribati) and Palmyra in the Line Islands, and Motuiti in the northern Marquesas, Polynesia (Fig. 1). No coconuts drifted to the American coast in any of the 6,588 simulated drifts from these islands, even in those of eight month's duration. In the "current only" experiment, 37 % of drifting nuts did reach the Galapagos group, the shortest crossing being in 178 days from Christmas Island, with a mean crossing time of 207 days. Both lengths of time far exceed the known viability period for floating coconuts and, in any case, coconuts did not grow in the Galapagos at the time of early European contacts. In
"wind only" experiments, no nuts made the crossing, and in "wind and current" experiments, of the 732 nuts started from Christmas Island, only 1% reached the Galapagos, in a mean time of 225 days. The conclusion must be that coconuts cannot drift to the Americas within any reasonable period of viability.
Hypothesis 3 Coconuts were carried in canoes from mid-Pacific islands by Polynesians on one or many occasions before discovery by Europeans. This hypothesis was ruled out by Baudouin and Lebrun (2009), based on the lack of similarity between the microsatellite profiles of different Polynesian coconut varieties and the Panama Tall variety. However, voyagers throughout the Pacific Islands regularly carried green and mature coconuts on their journeys for drink and food, and would be expected to plant some coconuts in situations where they were not already growing, or plentiful. Computer simulations of eastbound voyages from Polynesian islands towards America show that such voyages are possible from such starting points as Samoa, Tonga, eastern French Polynesia and Rapa Nui (Easter Island) (Fig. 1) (Levison et al., 1973; Irwin, 1992:163-4; Fitzpatrick and Callaghan, 2009). The voyage from Rapa Nui to the American coast would take about one month, and from the Marquesas two months (Irwin, 1992: 214; Fitzpatrick & Callaghan, 2009: 218). From Tonga or Samoa the journey might take between 66 and 128 days (Fitzpatrick and Callaghan, 2009: 218). Simulated voyages from eastern Polynesia approaching the American coast tend to be carried northwards by winds and the Peru Current in the latter part of their journey, and some make landfall as far north as the northern Ecuador coast. Those from Hawai'i reach the coast of Nicaragua and Costa Rica (Fitzpatrick & Callaghan, 2009). The distance and likely duration of voyages from eastern Micronesia (Kiribati) are similar to those from Samoa or Tonga, and such voyages would be more likely to reach the Panama coast.
Hypothesis 4 Coconuts were carried in canoes directly from the Philippines, bypassing any island where Polynesian coconuts grew at that time. This hypothesis is based on Bandouin and Lebrun's (2009) analysis that showed the close genetic relationship between coconuts in the Philippines and Panama, and has never previously been analyzed. As coconuts cannot have drifted to the Americas within any reasonable period of viability, one must conclude that they were carried there by voyagers from the western Pacific Islands. If coconuts from the Philippines reached Panama before the 17th century and did not come from Polynesia, we need to consider whether voyagers from Micronesia, or elsewhere in the western Pacific, may have had the capacity to carry coconuts to Panama. There is no record to suggest that Philippine mariners had the necessary technologies and knowledge.
Micronesian communities, however, have striking traditions of long-distance voyaging in the western Pacific Islands. Long-distance two-way voyaging by Micronesian people and their sophisticated methods of navigation are well documented (Lewis, 1972; Gladwin, 1970; Finney, 1979; Hezel, 1983; Thomas, 1987; Rainbird, 2004). The sea lanes between islands extending over more than 1,600 km were named (D'Arcy, 2006:154-5) and maps constructed of sticks and shells were used. Sophisticated concepts of estimating distances sailed and directions followed were taught. Two-way voyages in large outrigger canoes were conducted annually for economic and socio-cultural reasons over the 2,400 km length of the Caroline archipelago (Fig. 1). Anson reported in the 1740s that Marianas canoes (with their asymmetrical hulls) were "designed to sail as close as possible to the wind" and that they could reach a speed of 20 knots (quoted by Horridge, 1995: 148).
It is clear that Micronesians had the seamanship, navigational skills and canoes to make long voyages of exploration. They did so by settling groups of islands extending over 5,600 km from west to east and over 1,600 km north--south, and maintained regular links between these islands for many centuries. Irwin has pointed out that "the general trajectory of Pacific colonization was first upwind" and that this "implies pragmatic strategies of exploration" as "it is safest to search in the direction from which one can most easily return in the event of not finding new land" (1992:81). Micronesians following such strategies could have made long easterly voyages, for example from Kiribati or Christmas Island (Fig. 1). Finney (1985) also points out that eastward journeys would be easier in El Nino years. Once east of about 125[degrees] West longitude, they could readily take advantage of south-westerly winds to reach Panama (see Irwin, 1992: 9-16). Using such equatorial mutes they would not encounter any Polynesian islands en route so that any coconuts they were carrying would have come from stock in their home islands. Although the Micronesian sample used by Baudouin and Lebrun (2009) and Gunn et al. (2011) is only distantly related to the Pacific coast Panama Tall, Micronesia borders on Melanesia, which has more closely related coconuts in New Britain and northern Papua New Guinea (Fig. 1), as well as being closer to the Philippines.
However, until more intensive genetic sampling of coconuts is done in Micronesia, and clear archaeological evidence is found of Micronesian contacts with America, we cannot claim that Micronesians were the trans-Pacific carders. As Storey et al. (2011b) affirm, there is currently no such evidence. However, if Polynesia is ruled out as a source by the wide genetic gap between the sampled Polynesian and Panamanian coconuts, the history and capacities for navigation of Micronesians suggest they may be candidates for any pre-Spanish carriage.
Hypothesis 5 Coconuts were not present until carried by Spanish galleons returning from the Philippines. This hypothesis is supported by the lack of archaeological, ethnobotanical and linguistic evidence for the viability of the other four hypotheses, and by the doubts raised about the historical record. Baudouin and Lebrun (2009) appear to have assumed that the great number of coconuts carried between the Philippines and Mexico over a 250 year period was inconsistent with the limited similarity between the Philippines and the Panama Tall, which they regarded as typical of an extremely small founder population, while different selection pressures accounted for observed differences. However, regardless of the number of coconuts carried for consumption from Cebu to Navidad and then from Navidad to Panama, the reduction in population size each time was due to most of the nuts being consumed before arrival and the selection pressure every time was for the early germinators that survived to be planted. The genetic evidence suggests a relationship that is amenable to future study.
In order to determine if coconut was on the Pacific coast of Panama at the time of European conquest, it is essential to collect a truly representative sample of the coconuts along that coast, considering that the historical record suggests that not all the coconuts reported may have been similar to the San Ramon variety, even though the modern Pacific coast Panama Tall is indeed so. This type of collection has already started, led by Daniel Zizumbo-Villarreal, and will be analyzed with the same microsatellite markers used by Baudouin and Lebrun (2009) and Gunn et al. (2011). A larger sample of the relevant Philippine, Melanesian and Micronesian varieties is also needed to allow a more precise genetic analysis with the Pacific coast Panama Tall and Mexican Tall varieties.
The new genetic evidence is quite clear that modern coconut varieties from the Pacific coast of Panama are closely related to known modern Philippine varieties, as previously shown by morphometric analysis. There is, however, no archaeological, ethnobotanical or linguistic evidence that supports a pre-Columbian origin of these Pacific coast Panama Tails. A reanalysis of the historical record strongly suggests that early explorers made honest mistakes in identification. Hence, the most parsimonious explanation is that the Panama coconuts were introduced after Spanish conquest. The Manila-Acapulco galleon trade route that was active between 1565 and 1815 is very probably the means by which the Spanish introduced Philippine varieties of coconut to the Pacific coasts of the Americas. This is supported by the DNA analysis, and history records Spanish voyages with coconuts. The very small founder event that gave rise to the Pacific coast Panama Tall variety probably came from Mexico soon after the first Mexican plantations were established. New collections along the Mexican to Colombian Pacific coasts are improving the sampling for genetic analysis, and new work in the Philippines is suggested to confirm precise origins. Unless new archaeological remains are found to prove otherwise, this hypothesis can direct new research on the origins of American Pacific coast coconuts.
Acknowledgements Our special thanks to Luc Baudouin, CIRAD, Montpellier, France, for information about the samples chosen for coconut genetic analysis and for stimulating our reconsideration of pre-Columbian coconuts; to Madhavan Nayar, formerly Director, Central Plantation Crops Research Institute, Kasaragod, India, for providing bibliographic material; to Michael G. Price, Michigan Center, Michigan, Kenneth M. Olsen, Washington University, St. Louis, Missouri, Lalith Perera, Coconut Research Institute, Sri Lanka, for reading the manuscript in draft and making numerous useful suggestions, about the Philippines, the genetic analysis and the presentation; and to Pamela Brown, Pensacola, Florida, for careful review of the language.
Acosta, J. 1940. Historia natural y moral de las Indias. Fondo Cultural Economica, Imprensa Manuel Leon Sanchez, Mexieo. [Revised by Edmundo O'Gorman from the 1st Edition published in Sevilla, Spain]
Allen, P. H. 1965. Oviedo, on "Cocos". Principes 9: 62-66.
Alzina, F. I. 1668. On the palms which are called Cocos and their great usefulness (Translated by L. B. Uichanco, 1931). Philippine Agriculturalist 20: 435-446.
Angleria, P. M. 1944. Decadas del Nuevo Mundo. Editorial Bajel, Buenas Aires.
Ballard, C., P. Brown, R. M. Bourke & T. Harwood (eds.). 2005. The sweet potato in Oceania: a reappraisal. Ethnology Monographs 19, Pittsburgh; Oceania Monograph, 56, Sydney.
Bardy, W. 1914. Prospectus for Gorgona Island Coconut Estate Ltd. BT 31/22185/135025 UK Public Record Office, London.
Battuta, I. 1929. Travels in Asia and Africa 1325-1354 (Translated and edited by H. A. R. Gibb). Broadway House, London. Online: http://www.furdham.edu/halsall/source/1354_ibnbattuta.asp
Baudouin, L. & P. Lebrun. 2009. Coconut (Cocos nucifera L.) DNA studies support the hypothesis of an ancient Austronesian migration from Southeast Asia to America. Genetic Resources and Crop Evolution 56: 257-262.
Beccari, O. 1917. The origin and dispersal of Cocos nucifera. Philippines Journal of Science, C. Botany 12: 27-43.
Brown, C. H. 1994. Lexical acculturation in Native American languages. Current Anthropology 35:95-117.
-- 1999. Lexical acculturation in Native American languages. Oxford University Press, New York.
-- 2006a. Glottochronology and the chronology of maize in the Americas. Pp 647,663. In: J. E. Staller, R. H. Tykot, & B. F. Benz (eds). Histories of maize: Multidisciplinary approaches to the prehistory, biogeography, domestication, and evolutions of maize. Elsevier, San Diego.
-- 2006b. Prehistoric chronology of the common bean in the New World: The linguistic evidence. American Anthropologist 108:507-516.
-- 2010. The development of agriculture in prehistoric Mesoamerica: The linguistic evidence. Pp 71-107. In: J. E. Staller & M. Carrasco (eds). Pro-Columbian foodways: Interdisciplinary approaches to food, culture, and markets in Mesoamerica. Springer, Berlin.
Bruman, H. J. 1944. Some observations on the early history of the coconut in the New World. Acta Americana 2: 220-43.
-- 1945. Early coconut culture in western Mexico. The Hispanic American History Review 25: 212-223.
-- 1947. A further note on coconuts in Colima. The Hispanic American History Review 27: 572-573.
de Candolle, A. 1855. Geographic botanique raisonnee, vols. 1-2. Masson, Paris.
-- 1883. Origine des plantes cultivees. Germer Bailliere, Paris [first English translation: 1884. Origin of cultivated plants. Kegan, London].
Clement, C. R. 1995. Pejibaye (Bactris gasipaes). Pp 383-388. In: J. Smartt & N. W. Simmonds (eds). Evolution of crop plants, ed. 2nd. Longman, London.
Columbus, C. 1893. The Journal of Christopher Columbus (during his first voyage 1492-1493). Translation by C.R. Markham, The Hakluyt Society, London. On-line: http://www.archive.org/details/ journalchristop01markgoog
Constenla Umana, A. 1981. Comparative Chibchan phonology. Ph.D. diss. University of Pennsylvania, Philadelphia.
Cook, D. F. 1910. History of the coconut palm in America. Contributions to the US National Herbarium 14: 271-342.
Cooke, R. G. & A. J. Ranere. 1992. Prehistoric human adaptations to the seasonally dry forest of Panama. World Archaeology 24:114-133.
Copeland, E. B. 1914. The coconut. Macmillan & Co., London.
D'Arcy, P. 2006. The people of the sea: Environment, identity and history in Oceania. University of Hawai'i Press, Honolulu.
Darwin, C. 1860 Journal of researches into the natural history and geology of the countries visited during the voyage of H.M.S. Beagle round the world under the command of Capt. Fitz Roy, R.N. [2nd ed.] J. Murray, London.
Dennis, J. V. & C. R. Gunn. 1971. The case against trans-Pacific dispersal of the coconut by ocean currents. Economic Botany 25: 407-413.
Dickau, R. 2010. Microbotanical and macrobotanical evidence of plant use and the transition to agriculture in Panama. Pp 99-134. In: A. M. VanDerwarker & T. M. Peres (eds). Integrating zooarchacology and paleoethnobotany: A consideration of issues, methods, and cases. Springer, New York.
Edmondson, C. H. 1941. Viability of coconut after floating in sea. Occasional Papers Bernice P. Bishop Museum, Honolulu, Hawaii 16: 293-304.
Finney, B. R. 1979. Hokule'a: The way to Tahiti. Dodd, Mead and Co, New York.
-- 1985. Anomalous westerlies, El Nino, and the colonization of Polynesia. American Anthropologist 87:9-26.
Fitzpatrick, S. M. & R. Callaghan. 2009. Examining dispersal mechanisms for the translocation of chicken (Gallus gallus) from Polynesia to South America. Journal of Archaeological Science 36:214-223.
Gladwin, T. 1970. East is a big bird: Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge.
Gongora, J., N. J. Rawlence, V. A. Mobegi, H. Jianlin, J. A. Alcalde, J. T. Matus, O. Hanutte, C. Moran, J. J. Austin, S. Ulm, A. J. Anderson, G. Larson & A. Cooper. 2008. Indo-European and Asian origins for Chilean and Pacific chickens revealed by mtDNA. Proceedings of the National Academy of Science of the USA 105:10308-10313.
--, et al. 2008b. Reply to Storey et al.: More DNA and dating studies needed for ancient El Arenal-1 chickens. Proceedings of the National Academy of Sciences of the USA 105: E100.
Green, R. C. 2005. Sweet potato transfers in Polynesian prehistory. Pp. 43-62. In: C. Ballard, P. Brown, R. M. Bourke & T. Harwood (eds.). The sweet potato in Oceania: a reappraisal. Ethnology Monographs 19, Pittsburgh; Oceania Monograph, 56, Sydney.
Gunn, B. F. 2004. The phylogeny of Cococae (Arecaceae) with an emphasis on Cocos nucifera. Annals of the Missouri Botanical Garden 91:505-522.
--, L. Bauduuin & K. M. Olsen. 2011. Independent origins of cultivated coconut (Cocos nucifera L.) in the Old World Tropics. PLuS ONE 6: e21143.
Harries, H. C. 1971. Coconut varieties in America. Oleagineux 26: 235-242.
-- 1977. The Cape Verde region (1499 to 1549): The key to coconut culture in the Western Hemisphere? Turrialba 27:227-231.
-- 1978. The evolution, dissemination and classification of Cocos nucifera. Botanical Review 44: 265-320.
-- 1981. Germination and taxonomy of the coconut palm. Annals of Botany 48: 873-883.
-- 1995. Coconut. Cocos nucifera L. (Palmae). Pp 389-395. In: J. Smartt & N. W. Simmonds (eds). Evolution of crop plants, ed. 2nd. Longman, New York.
--. Key to coconut cultivation on the Pacific coast of America: the Manila-Acapulco galleon route (1565-1815). Palms (J. Int. Palm. Soc.) 56(2): 72-77.
Harries, H. C. & D. H. Romney. 1974. Maypan: an F1 hybrid coconut variety for commercial production in Jamaica. World Crops 26:110-111.
Henderson, A., G. Galena & R. Bernal. 1995. Field guide to the palms of the Americas. Princeton University Press, Princeton.
Hezel, F. X. 1983. The first taint of civilization. A history of the Caroline and Marshall Islands in pre-colonial Days, 1521-1885. University of Hawaii Press, Honolulu.
Hill, A. W. 1929. The original home and mode of dispersal of the coconut. Nature 124(3117): 133-134 & 151-153.
Holman, E. W., C. H. Brown, S. Wichmann, A. Muller, V. Velupillai, H. Hammarstrom, S. Sauppe, H. Jung, D. Bakker, P. Brown, O. Belyaev, M. Urban, R. Mailhammer, J.-M. List & D. Egorov. 2011. Automated dating of the World's language families based on lexical similarity. Current Anthropology 52: 841-875.
Horridge, A. 1995. The Austronesian conquest of the sea--upwind. Pp 143-160. In: P. Bellwood, J. J. Fox, & D. Tryon (eds). The Austronesians: Historical and comparative perspectives. Department of Anthropology, Research School of Pacific and Asian Studies, ANU, Canberra.
Hubisz, M. J., D. Falush, M. Stephens & J. K. Pritchard. 2009. Inferring weak population structure with the assistance of sample group information. Molecular Ecology Resources 9:1322-1332.
Irwin, G. 1992. The prehistoric exploration and colonisation of the Pacific. Cambridge University Press, Cambridge.
Jones, T. L., A. A. Storey, E. A. Matisoo-Smith & J. M. Ramirez-Aliago (eds). 2011. Polynesians in America: Pro-Columbian contacts with the New World. AltaMira Press, Lanham.
Kaufman, T. 1994. The languages of South America. Pp 46-76. In: C. Moseley & R. E. Asher (cds). Atlas of the world's languages. Routledge, New York.
Lebot, V. 2009. Tropical roots and tuber crops: Cassava, sweet potato, yams and aroids. CAB International, Oxford, UK.
Levison, M., R. G. Ward & J. W. Webb. 1973. The settlement of Polynesia: A computer simulation. University of Minnesota Press, Minneapolis.
Lewis, D. 1972. We, the navigators: The ancient art of landfinding in the Pacific. University of Hawaii Press, Honolulu.
Lewis, M. P. 2009. Ethnologue: Languages of the World, 16th edition. SIL International, Dallas. Online version: http://www.ethnologue.com/
Locascio, W. A. 2010. Communal tradition and the nature of social inequality among the prehistoric households of El Hatillo (HE-4). Panama. Ph.D. dissertation, University of Pittsburgh, Pittsburgh.
Margery, E. 1989. Diecionario Cabecar- Espanol Espanol-Cabdcar. Editorial de la Universidad de Costa Rica, San Jose.
Matisoo-Smith, E. A. & J. M. Ramirez. 2010. Human skeletal evidence of Polynesian presence in South America? Metric analyses of six crania from Mocha Island, Chile. Journal of Pacific Archaeology 1: 76-88.
Merrill, E. D. 1937. On the significance of certain oriental plant names in relation to introduced species. Proceedings of the American Philosophical Society 78:112-146.
-- 1954. The botany of Cook's voyages. Chronica Botanica 14: 161-384.
Morcote-Rios, G. & R. Bernal. 2001. Remains of palms (Palmae) at archaeological sites in the New World: A review. The Botanical Review 67: 309-350.
Nei, M., F. Tajima & Y. Tateno. 1983. Accuracy of estimated phylogenetic trees from molecular data. Journal of Molecular Evolution 19: 153-170.
Oviedo y Valdes, G. F. 1526. Sumario de la natural historia de los Indias. Remor de Petras, Toledo. [1995. Edicion de N. del Castillo Mathieu, Universidad de Bogota "Jorge Tadeo Lozano", Santafe de Bogota, Colombia.
-- 1535. La historia general de las Indias. Juan Cromberger, Sevilla [1851-1855. J. A. de los Rios y Serrano (ed.). Imprenta de la Real Academia de Historia, Madrid]
Patino, V. M. 1963. Plantas cultivadas e animales domesticos en America Equinoccial. Tomo II. Frutales. Imprenta Departamental, Cali, Colombia.
-- 1970. Plantas cultivadas e animals domesticos en America Equinoccial. Tomo V. Animales domesticos introducidos. Imprenta Departamental, Cali, Colombia.
-- 2002. Historia y dispersion de los frutales nativos del Neotropico. Centro Internacional de Agricultura Tropical, Call.
Piperno, D. R. & D. M. Pearsall. 1998. The origins of agriculture in the lowland Neotropies. Academic, New York.
Pritchard, J. K., M. Stephens & P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945-959.
Purseglove, J. W. 1972. Tropical crops: Monocotyledons. Longman, London.
Quesada Pacheco, M. A. & C. Rojas Chaves. 1999. Diccionario Boruca-Espanol Espanol-Boruca. Editorial de la Universidad de Costa. Rico, San Jose.
Rainbird, P. 2004. The archaeology of Micronesia. Cambridge University Press, Cambridge.
Ramirez-Aliaga, J. M. 2011. The Mapuche connection. Pp 95-110. In: T. L. Jones, A. A. Storey, E. A. Matisoo-Smith, & J. M. Ramirez-Aliago (eds). Polynesians in America: Pre-Columbian contacts with the New World. AltaMira Press, Lanham, MD.
Rensch, C. R. 1989. An etymological dictionary of the Chinantecan languages. The Summer Institute of Linguistics. Arlington, Texas.
Reyne, A. 1948. De Cocospalm. Pp 427-525. In: C. J. J. van Hall & C. van de Koppel (eds). De Landbouw in den Indischen Archipel, Vol. 2A. S-Gravenhage, Van Hoeve.
Roullier, C., G. Rossel, D. Toy, D. McKey & V. Lebot. 2011. Combining chloroplast and nuclear microsatellites to investigate origin and dispersal of New World sweet potato landraces. Molecular Ecology 20: 3963-3977.
Roullier, C., L. Benoit, D. B. McKey & V. Lebot. 2013. Historical collections reveal patterns of diffusion of sweet potato in Oceania obscured by modern plant movements and recombination. Proceedings of the National Academy of Sciences of the USA 110(6): 2205-2210.
Scaglion, R. 2005. Kumara in the Ecuadorian Gulf of Guayaquil? Pp. 35-41. In: C. Ballard, P. Brown, R. M. Bourke & T. Harwood (eds.). The sweet potato in Oceania: a reappraisal. Ethnology Monographs 19, Pittsburgh; Oceania Monograph, 56, Sydney.
Sevilla del Rio, 17. 1977. Provanca de la Villa de Colima: Documentos originales del Amhivo Municipal de la Ciudad de Colima 1612-1724. Textos paleograficos de F. Sevilla del Rio. Ed. Jus, Ciudad de Mexico, D.E.
Small, J. K. 1929. The Coconut Palm--Cocos nucifera. Journal of the New York Botanical Garden 30(355): 153-161. and (356): 194-203.
Smith, C. E., Jr. 1980. Plant remains from the Chiriqui sites and ancient vegetational patterns. Pp 151-174. In: O. F. Linares & A. J. Ranere (eds). Adaptive radiations in prehistoric Panama, Peabody Museum Monographs, Vol. 5. Harvard University, Cambridge.
Smith, H. H. & F. A. G. Pope. 1914. Coco-nuts: The consols of the East, ed. 2nd. Tropical Life Publ, London.
Storey, A. A., J. M. Ramirez, D. Quiroz, D. V. Burley, D. J. Addison, R. Walter, A. J. Anderson, T. L. Hunt, J. S. Athens, L. Huynen & E. A. Matisoo-Smith. 2007. Radiocarbon and DNA evidence for a pre-Columbian introduction of Polynesian chickens to Chile. Proceedings of the National Academy of Sciences of the USA 104: 10335-10339.
--, D. Quiroz & E. A. Matisoo-Smith. 2011a. A reappraisal of the evidence for pre-Columbian introduction of chickens to the Americas. Pp 139-170. In: T. L. Jones, A. A. Storey, E. A. Matisoo-Smith, & J. M. Ramirez-Aliago (eds). Polynesians in America: Pre-Columbian contacts with the New Word. AltaMira Press, Lanham.
--, A. C. Clarke & E. A. Matisoo-Smith. 2011b. Identifying contact with the Americas: A commensal-based approach. Pp 111-138. In: T. L. Jones, A. A. Storey, E. A. Matisoo-Smith, & J. M. Ramirez-Aliago (eds). Polynesians in America: Pre-Columbian contacts with the New World. AltaMira Press, Lanham.
Thomas, S. D. 1987. The last navigator. Henry Holt & Co., New York.
Vargas, A. & F. A. Blanco. 2000. Fruit characterization of Cocos nucifera L. (Arecaceae) cultivars from the Pacific coast of Costa Rico and the Philippines. Genetic Resources and Crop Evolution 47: 483-487.
Varthema, L. 1510. The itinerary of Ludovico di Varthema of Bologna from 1502 to 1508. Translated from the original Italian edition of 1510 by John Winter Jones in 1863. Reprinted by Asian Educational Services, New Delhi.
Vavilov, N. I. 1951. The origin, variation, immunity and breeding of cultivated plants. Chronica Botanica, Waltham, MA.
Ward, R. G. & B. J. Allen. 1980. The viability of floating coconuts. Science in New Guinea 7:69-72.
-- & M. Brookfield. 1992. The dispersal of the coconut: Did it float or was it carried to Panama? Journal of Biogeography 19: 467-480.
Witkowski, S. R. & C. H. Brown. 1983. Marking-reversals and cultural importance. Language 59: 569-582.
Yen, D. E. 1974. The sweet potato in Oceania: An essay in ethnobotany. Bishop Museum Press, Honolulu, Hawaii.
Zeder, M. A., D. G. Bradley, E. Emschwiller & B. D. Smith (eds). 2006. Documenting domestication--New genetic and archaeological paradigms. University of California Press, Berkeley, CA.
Zhang, D., G. Rossel, A. Kriegner & R. Hijmans. 2004. AFLP assessment of diversity in sweet potato from Latin America and the Pacific region: Its implications on the dispersal of the crop. Genetic Resources and Crop Evolution 51: 115-120.
Zizumbo-Villarreal, D. 1996. History of coconut (Cocos nucifera L.) in Mexico: 1539-1810. Genetic Resources and Crop Evolution 43:505-515.
-- & P. Colunga-GarciaMarin. 2001. Morpho-physiological variation and phenotypic plasticity in Mexican populations of coconut (Cocos nucifera L.). Genetic Resources and Crop Evolution 48: 547-554.
--, -- & H. J. Quero. 1998. Re-evaluation of early observations on coconut in the New World. Economic Botany 52:68-77.
--, M. Fernandez-Barrera, N. Torres-Hernandez & p. Colunga-GarciaMarin. 2005. Morphological variation of fruit in Mexican populations of Cocos nucifera L. (Arecaccae) under in situ and ex situ conditions. Genetic Resources and Crop Evolution 52: 421-434.
-- & P. Colunga-GarciaMarin. 2008. Early coconut distillation and the origins of mescal and tequila liquor in western Mexico. Genetic Resources and Crop Evolution 55:493-510.
Charles R. Clement (1,8) * Daniel Zizumbo-Villarreal (2) * Cecil H. Brown (3,4) * R. Gerard Ward (5) * Alessandro Alves-Pereira (6) * Hugh C. Harries (7,8)
(1) Instituto Nacional de Pesquisas da Amazrnia, Av. Andre Araujo, 2936, 69067-375 Manaus, AM, Brazil
(2) Unidad de Recursos Naturalcs, Centro de Investigacion Cientifica de Yucatan A.C., Merida, Yucatan, Mexico
(3) Northern Illinois University, DeKalb, IL 60115, USA
(4) University of West Florida, Pensacola, FL 32514, USA
(5) Research School of Asia and the Pacific, Australian National University, Canberra, Australia
(6) Generics and Plant Breeding Post-graduate Program, Universidade de Sao Paulo, Escola Superior de Agricultura Luiz de Queiroz, Piracicaba, SP, Brazil
(7) Coconut Time Line, Weymouth, Dorset DT3 5NP, UK
(8) Author for Correspondence; e-mail: firstname.lastname@example.org; email@example.com
Published online: 17 August 2013
Table 1 Terms for coconut in Chibchan and Choco languages Coconut Term Language Family Location siahua Boruca Chibchan Costa Rica koko Bribri " Costa Rica ko ko Chimila " Colombia koko Dorasque (extinct) " Panama Coco Guatuso " Costa Rica ogoba Kuna " Panama koko Ngdbare " Panama kukunup Rama " Nicaragua kokoha Paya (Pcch) " Honduras koko Northern Embera Choco Panama k'ok'o Epena " Panama kok Woun Weu " Panama Table 2 Coconut varieties included in the Generation Challenge Program data set that were analyzed with microsatellite markers for the Baudouin and Lebrun (2009) and Gunn et al. (2011) studies. Variety names follow GCP/CIRAD nomenclature Variety n Comments The Philippines 46 Ballesteros Tall Tarraq 7 Baybay Tall 8 Macapuno Tall 5 Pandan Tall 6 San Ramon Tall 6 Fruit similar to Panama Tall Tagnanan Tall 14 Mexico 43 Pacific Tall Colima 14 11 plants possibly introduced from Rennell Island (Solomon Islands, Melanesia) in 1569--fruit similar to Rennell Tall; 3 plants with fruit similar to Philippine varieties Pacific Tall Guerrero 11 8 plants with fruit similar to San Ramon; 3 plants with fruit similar to Baybay Tall Pacific Tall Michoacan 14 Unnamed Philippine variety introduced in the 1930s Pacific Tall Nuxco 4 Possibly introduced from the Philippines into Acapulco shortly after 1572 Panama 105 Panama Tall 44 Some introgression with Indo-Atlantic; includes 5 from Oxtapacab, Yucatan, Mexico Panama Tall Agua Dulce 13 West of Nata, inland, taken to Jamaica, then Ivory Coast--considerable introgression with Indo-Attantic Panama Tall Bowden 10 Taken to Jamaica Panama Tall Costa Rica 19 Costa Rica--minor introgression with Indo-Atlantic Panama Tall Monagre 19 West of Nata along coast--minor introgression with Indo-Atlantic
|Printer friendly Cite/link Email Feedback|
|Author:||Clement, Charles R.; Zizumbo-Villarreal, Daniel; Brown, Cecil H.; Ward, R. Gerard; Alves-Pereira, Al|
|Publication:||The Botanical Review|
|Date:||Sep 1, 2013|
|Previous Article:||Histones and plant hormones: new evidence for an interesting interplay.|
|Next Article:||The multifunctional role of ectomycorrhizal associations in forest ecosystem processes.|