Research Article

Independent Origins of Cultivated Coconut (Cocos nucifera L.) in the Old World Tropics

  • Bee F. Gunn,

    Affiliation: Division of Evolution, Ecology and Genetics, Research School of Biology, The Australian National University, Canberra, Australia

  • Luc Baudouin,

    Affiliation: Centre International de Recherches en Agronomie pour le Développement (CIRAD), Montpellier, France

  • Kenneth M. Olsen mail

    Affiliation: Biology Department, Washington University, St. Louis, Missouri, United States of America

  • Published: June 22, 2011
  • DOI: 10.1371/journal.pone.0021143

Reader Comments (4)

Post a new comment on this article

About pre-Columbian coconuts

Posted by Dr_Luc_Baudouin on 29 Jun 2011 at 08:18 GMT

to "Mikegprice", who "Greatly enjoyed and appreciated the paper, but wish[es] to flag one weak claim, that coconuts were brought across the Pacific to Panama in pre-Columbian times. There's no credible evidence to support that belief. It just ain't so".

You would certainly enjoy reading the paper by Zizumbo and Quero who comment on Oviedo's description of large numbers of coconut when the Spaniards reached the Pacific coast of Panama. My contribution was to show that their origin is south-east asiatic rather than polynesian. This rules out the hypothesis of natural dissemination. I am not dogmatic about the date of arrival. I simply take note of a striking coincidence between genetic and archaeological findings.

Baudouin, L. and 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(2): 257-262. DOI 10.1007/s10722-008-9362-6

Zizumbo Villarreal, D. and H. J. Quero (1998). "Re-evaluation of early observations on coconut in the new world." Economic Botany 52(1): 68-77.

Competing interests declared: I am the second author of the present paper

RE: About pre-Columbian coconuts

harrieshc replied to Dr_Luc_Baudouin on 01 Jul 2011 at 09:34 GMT

I agree with Mike Price that there is no credible evidence that coconuts were brought across the Pacific in pre-Columbian times. The description by Oviedo of palms that looked like coconuts can equally apply to other endemic genera such as Attalea or Bactris. The coconut palms on the Pacific coast that Luc attributes to ancient Austronesians can more realistically be due to the historically recent and documented activities of Spanish explorers. Furthermore, the map in this paper shows unrealistic routes both across the Pacific and into the Atlantic. Sailing ships, like floating coconuts, go with winds and tides. The introduction of coconut into the Atlantic and Caribbean by the Portuguese was via the Cape Verde islands not the Gulf of Guinea and annually, for more than 200 years, the Spanish carried hundreds of people and thousands of coconuts from the Philippines to the west coat of America (from Mexico to Peru) by a north Pacific route avoiding any islands. The Austronesian and Poynesian distributions and the two centres of diversity have been identified previously, before DNA analysis was available, and it is satisfying to have this confirmation.

Competing interests declared: I have published a number of papers on the subject of coconut evolution and dissemination.

RE: RE: About pre-Columbian coconuts

Dr_Luc_Baudouin replied to harrieshc on 04 Jul 2011 at 09:18 GMT

Dear Hugh,

"Inside, attached to the shell there is a meaty part in width like half the thickness of the small finger of a hand [...]. In the interior part, surrounded by the meaty part, can be found a clear excellent water in the same way that the white of an egg is enclosed in the shell but proportionate to the shell of the coconut". Please, could you precise whether these excerpt from Oviedo apply to Bactris or to Attalea? (the translation is due to Zizumbo and Quero who discuss about some apparent inconsistencies like in the drawings that were not in Oviedo's original.) . The fruit size and morphology are unmistakably those of a coconut .

Of course, coconut could have reached Costa Rica, Panama, Colombia, Ecuador and Peru by the same route and at the same time as the Mexican Pacific coconuts. But this didn't happen : in that case, they would have had about the same allelic frequencies at microsatellite markers as in Mexico and in the Philippines). Instead, coconuts from this vast region have a surprisingly low genetic diversity for Tall coconuts, which indicates that all of them descend from a handful of founding individuals. Unlike the case of Mexico, the spaniards didn't have to introduce large number of coconut from the Philippines to Panama, simply because the pre-columbian population was already there.

The arrows on our map never intended to represent actual nautical routes. They simply connect the point of origin to the point of introduction of populations.

Hugh, you are right on the following point: it is amazing that Vavilov already identified both centres of origins as early as in 1935 without the help of molecular markers. However, it would be exagerated to say that the notion of an indian centre was widely popularized by further authors...

Competing interests declared: I am the second author of the present paper

RE: About pre-Columbian coconuts

cclement replied to Dr_Luc_Baudouin on 09 Jul 2011 at 14:22 GMT

Dear Luc and Hugh,

Let's agree to discard Bactris from the discussion, as no Bactris has fruits large enough to fit with Oviedo's description of the size of a head (cited by Zizumbo & Quero, 1998). Both Patiño (1963) and Zizumbo & Quero (1998) suggest that the figure in Oviedo is probably that of Bactris major, the cane palm, which is described by Oviedo on the same page that he starts describing coconut.

There are also no Attaleas in that area with fruit large enough to meet Oviedo's description. A. butyracea is a large enough palm, but the fruit are only 8.5 cm long by 4.5 cm diameter (Henderson, 1995, Palms of the Americas), which is the wrong shape. Hence, we can discard Attalea also.

That leaves coconut as the most like "cocos" mentioned by Oviedo. Patiño (1963) discusses the origins of Oviedo's knowledge about cocos and shows that he probably learned about the species and its importance while in Spain, where numerous reports were coming in from India and Southeast Asia about its importance there. Many of the uses he describes may be from those reports, as Native Americans apparently had much less use for it (Patiño, 1963), which is contrary to Zizumbo & Quero's (1998) interpretation. There are also very few indigenous names for coconut in that area (Patiño, 1963), again suggesting lack of importance and possible a recent arrival, rather than the 2,250 y proposal in Gunn et al. (2011). Luc already commented in a private e-mail exchange that he is not convinced about the 2,250 y date, which is probably too early to be due to humans.

I find the microsatelite analysis by Gunn et al. (2011) to be quite convincing, and the historical documental analysis by Patiño (1963) and Zizumbo & Quero (1998) strongly suggests that the Panama coconuts were present before the Spanish started introducing more coconut from the Philippines. As I wrote to Luc, 2,250 y is a long shot, but the idea of a Chinese introduction in 1421 (Gavin Menzies, The year China discovered America, Harper Perennial) is dicey and may be too late for dispersal from Burica to Candelaria.

We need some archaeological information from that region of Panama to dig into this better.

No competing interests declared.

RE: RE: RE: About pre-Columbian coconuts

harrieshc replied to Dr_Luc_Baudouin on 10 Jul 2011 at 02:55 GMT


In 1975, in a book on crop plant evolution, the chapter on coconut had a map showing an arrow across continental Africa from east to west. The eminent editor was willing to believe that coconuts had reached West Africa by an overland route. Nobody, before or after, has ever provided any evidence for such a claim.. On the contrary, other authorities had already suggested introduction by the Portuguese and since then research has accumulated evidence that the Cape Verde Islands were the point of first introduction and subsequent dissemination to the Caribbean and the Atlantic coasts of America and Africa. Can molecular markers shed light on the subject?

I accept the description of a coconut fruit by Oviedo but we should not forget that he was not a botanist. He was a Spanish colonial administrator and, at the time that he was writing the world had been divided in half between the Portuguese and Spanish by the Treaty of Tordesillas. From the letter written by King Manuel of Portugal in 1501 it is clear that coconuts were recognized as a vital commodity for sailing ships and by chance (or possibly by design) the Portuguese had the half with most coconuts. It was important to Oviedo to persuade the Spanish court that there were coconuts in America. He may have found a few (see Paitino, 1973, 2002) and simply assumed that the all palms were coconuts. Many years later a botanist admitted to the same mistake (Cook, O.F. (1940) An endemic palm on Cocos Island near Panama mistaken for the coconut palm. Science 91, 140-142).

I also accept your own identification of low genetic diversity in the Pacific coast coconut populations south from Mexico. But your "handful of founding individuals" is still much more likely to have arrived from the Philippines in the 16th century with the first settlers (Spaniards & Filipinos). Any earlier Austronesian introduction would probably not have survived for centuries - the coconut cannot only compete without human assistance

Competing interests declared: I have published a number of papers on the subject of coconut evolution and dissemination.

RE: RE: RE: RE: About pre-Columbian coconuts

Dr_Luc_Baudouin replied to harrieshc on 11 Jul 2011 at 10:45 GMT

Dear Hugh,

Up to now, we didn't analyze coconuts from Cape Verde. Reaching a firm conclusion would presumably require analysing quite a lot of palms.

Regarding Oviedo, "He may have found a few (see Paitino, 1973, 2002) and simply assumed that the all palms were coconuts", this seems to imply that coconut was introduced prior to the Spanish conquest after all...

I assume you meant Patiño. Thank you for refering to his posthumous book, which happens to be available online.


Competing interests declared: I am the second author of the present paper

RE: RE: RE: RE: RE: About pre-Columbian coconuts

harrieshc replied to Dr_Luc_Baudouin on 11 Jul 2011 at 11:13 GMT


Indeed it would be desirable to analyze samples from the Cape Verdes. Also the Cocos-Keelings, Palmyra and many other places where "wild type" characteristics may predominate.

Perhaps this is a good reason for a Global Coconut Genome Consortium to be established?

Oviedo could have written his description of a coconut fruit from general knowledge, available even before 1501, without necessarily seeing one at all. Many other writers have done so since.

Thank you for correcting my reference to Patino.

Would your co-authors care to comment?


Competing interests declared: I want a Global Coconut Genome Consortium to be established.

RE: RE: RE: RE: RE: RE: About pre-Columbian coconuts

BeeGunn replied to harrieshc on 17 Jul 2011 at 23:25 GMT

Yes, it is a good idea to have a GCGC as it would help us understand the coconut's genomic composition much better. Currently whole genome sequencing technology is available and in the Arecales, only the Phoenix dactylifera's genome has been sequenced. It is time to add the coconut's genome to a list that includes other globally important economic plants. It will involve a network of institutions from various countries to work on different levels of the project.
I think that it would be great to focus more on the pressing need to identify the candidate genes and QTLs of the East African coconut populations, some of which show resistance to lethal yellowing disease but not others.

Competing interests declared: I am the first author of the current paper.

Global coconut Genome Consortium

Dr_Luc_Baudouin replied to harrieshc on 18 Jul 2011 at 08:56 GMT

Dear Hugh,

A Coconut Genome Consortium would certainly be useful. I think that the main aim should be to try and identify genes subjected to natural selection due to Lethal Yellowing. The East African populations should be a good model due to their unique genetic structure: their genome is made up of about 2 parts of "Indian" genes for 1 part of South-East Asian (SEA) genes. The only African populations that appear to have resistance factors are found in the North of Tanzania.

An analysis of worldwide results of field testing suggests that (partial) resistance factors are more frequent in South-East Asia and absent in the Indo-Atlantic populations. This general tendency seems to apply irrespective of the variation of the pathogen)

This is an opportunity to identify at least chromosomal regions involved in resistance to the disease.


Competing interests declared: I am the second author of the present paper