Coconut originated in South America, how could it not have?

[Dimovi]

I read everywhere that the coconut is from the Indo Pacific, but this makes no phylogenetic sense. All closely related genera to Cocos are in South America, like Syagrus, Jubaea, Parajubaea, Butia. Clearly the coconut started from a common ancestor there and once it evolved a large floating seeds it spread around the tropical shores of the world almost instantaneously by the evolutionary timescale. I assume since the Indo Pacific was "discovered" by Europeans long before America, it seemed that it was the place of origin of the Coconut, but by now we should know better.

[Silas_Sancona]

1 hour ago, Dimovi said:

I read everywhere that the coconut is from the Indo Pacific, but this makes no phylogenetic sense. All closely related genera to Cocos are in South America, like Syagrus, Jubaea, Parajubaea, Butia. Clearly the coconut started from a common ancestor there and once it evolved a large floating seeds it spread around the tropical shores of the world almost instantaneously by the evolutionary timescale. I assume since the Indo Pacific was "discovered" by Europeans long before America, it seemed that it was the place of origin of the Coconut, but by now we should know better.

This is always a fun discussion to which i can see both theories having clout on..

Many ( ..Many, lol  ) years ago, i'd mentioned how some think the genus related to Cocos evolved  ..The " Seeds of Cocos / it's direct ancestor got smaller as they were carried further and further up hill / away from prime Cocos habitat, over long periods of time " theory in another " Where did Coconuts come from / how are Jubea and Queen Palms related " - type of discussion here..

Understanding how plants evolve, that idea has substance to it -my own opinion on that of course..

As you mention ..who was the common ancestor? and did it evolve in the Americas( Personally think this makes the most sense ),  then drift off to new homes on distant horizons?  ..Or, is it possible ( Can see this happening too ) that the ancestor evolved in say the Indo Pacific, then drift over to ...and eventually diversify in... South America / the rest of the lower Americas where they could survive??..
...While not being able to colonize and diversify on the other side of the world? ( which they should have been able to, imo )

Could Becarriophoenix be some sort of link?  ...a convergent evolutionary  " mirror " to Cocos that appeared around the same time as it's ancestor?  (  ..using the " Cactus evolved in the Americas / Euphorbia evolved in Africa to fill the same niches " theory? )  

 Probably not but,  has every single tiny detail of how Cocos evolved been discovered yet, ...or could there still be some surprises  awaiting future discovery?  I honestly don't think the story is completely settled, no matter what some others might think.

Just have to wait and see i suppose, ..and enjoy the on going story of how these ...and everything else... came about.

As far as what Europeans " discovered " in their travels,  i can go out into the desert and " discover " a population of some plant which may be disconnected from the rest of the species in it's Genus / near - related genus, sometimes by 100s of miles.

Doesn't mean it's origin lies where i " discovered " it,  ...just because i want that to be true. 

[Merlyn]

@Dimovi that sure does make a lot of sense, given the recent genetic tree:

 

The only two in that "branch" that aren't South American are Jubaeopsis (Africa) and Beccariophoenix (Madagascar).  Maybe it's just a "myth" that it originated in Southern Asia.  Perhaps the only reason that story came about is that it was first heavily cultivated there by humans?

Or maybe it did originate in Southern Asia/Indo-Pacific and there were cocos that landed in South America 50 million years ago and mutated? 

Or maybe it originated in South America, and the only seeds to make it to Southern Asia were the biggest ones?  The other smaller seeded relatives couldn't float and survive that far away. 

Of course, 50-100 million years ago the continents looked very different, and Antarctica was fairly tropical.  So the common ancestor palm wouldn't have had to float all that far from South America to Madagascar or Southern Asia:

[Xenon]

An anthropological argument would strongly favor somewhere around maritime southeast Asia and New Guinea. Coconut is deeply integrated into the cuisine of southeast Asia (especially south of ~15 degrees latitude) and also southern India and Sri Lanka...seemingly since the beginning of recorded history. 

Coconut is not featured prominently in any current or recorded historical indigenous American cuisine unlike other useful new world crops like corn, squash, cassava, etc. Coconut isn't even common in traditional West African cuisine which also supports a recent introduction there from elsewhere. 

[Dimovi]

48 minutes ago, Xenon said:

Coconut is deeply integrated into the cuisine of southeast Asia (especially south of ~15 degrees latitude) and also southern India and Sri Lanka...seemingly since the beginning of recorded history. 

Tomatoes are synonymous with Italian food and potatoes with Irish and both came from the new world just a few hundred years ago. Cocos evolved millions of years ago, so there is not an anthropological argument to be made. There is no doubt that Cocos is huge part of the culture from that part of the world and it may have been cultivated in Indo Pacific first, but the origin of the wild Cocos there is extremely improbable.

[Dimovi]

1 hour ago, Silas_Sancona said:

Or, is it possible ( Can see this happening too ) that the ancestor evolved in say the Indo Pacific, then drift over to ...and eventually diversify in... South America / the rest of the lower Americas where they could survive??..

This would have been a possible hypothesis if there was more or at least just as much speciation in happening in Indo Pacific. It appears highly improbable since all closely related genera are in South America, and the more distant ones in Africa, possibly then the continents were in contact or at least very close. In phylogenetic studies (like the one that @Merlyn posted) Parajubaea is closest to Cocos but less close to Syagrus and Jubaea. If they all evolved from Cocos, then they would all share a common branch in the tree, but Cocos is a limb.

[Xenon]

Just now, Dimovi said:

Tomatoes are synonymous with Italian food and potatoes with Irish and both came from the new world just a few hundred years ago. Cocos evolved millions of years ago, so there is not an anthropological argument to be made. There is no doubt that Cocos is huge part of the culture from that part of the world and it may have been cultivated in Indo Pacific first, but the origin of the wild Cocos there is extremely improbable.

Tomatoes and potatoes are indisputably known to be introductions from the new world. Comparing tomatoes in Italy with coconut in southeast Asia is nonsense. The root word for coconut in that part of the world is thousands of years old itself. Afaik, there's nothing to suggest that coconut was brought there from another part of the world. 

The center of genetic diversity for coconut is also centered around Indonesia/New Guinea. The New World comparatively has much lower genetic diversity that can be traced back to a few introductions. Yeah not buying this new world coconut conspiracy 😆

[Merlyn]

One confusing part about the "genetic diversity in Indonesia" argument is that there's no actual evidence that "genetic diversity = origin of species."  It's just a hypothesis.  It might be completely wrong.  Maybe Cocos originated in South America but floated over and flourished and diversified in Indonesia. 

Or...my favorite new hypothesis...Beccariophoenix was the original species, and it dispersed (from a centralized location in Madagascar) to East and West and brought about the rest of them.  Of course, I might be biased because I really like Beccariophoenix. 

[Silas_Sancona]

11 minutes ago, Merlyn said:

One confusing part about the "genetic diversity in Indonesia" argument is that there's no actual evidence that "genetic diversity = origin of species."  It's just a hypothesis.  It might be completely wrong.  Maybe Cocos originated in South America but floated over and flourished and diversified in Indonesia. 

Or...my favorite new hypothesis...Beccariophoenix was the original species, and it dispersed (from a centralized location in Madagascar) to East and West and brought about the rest of them.  Of course, I might be biased because I really like Beccariophoenix. 

As mentioned, a link between crossed my mind as well..  

Physical geographical proximity to one another at that time aside, would sea levels at the time of evolution / dispersal of a common ancestor influenced things / made it easier, ...or more difficult.. for " Ancestor X "  to circulate around the globe?


Swaying a little for a moment, On a related note, anyone see this?  ..May have missed any past discussion of it here ..but found this article while looking for some other info on a similar subject.  Correct or proven wrong, found it interesting.

Hopefully the link below works.

https://www.journals.uchicago.edu/doi/10.1086/681605
 

[Dimovi]

14 minutes ago, Merlyn said:

Or...my favorite new hypothesis...Beccariophoenix was the original species, and it dispersed (from a centralized location in Madagascar) to East and West and brought about the rest of them.  Of course, I might be biased because I really like Beccariophoenix. 

Sure, they certainly had a common ancestor with Beccariophoenix maybe from when Africa was closer, but according to your phylogenetic tree there are several last common ancestor before the Beccariophoenix branch. If Cocos branches off from the last common ancestor with Beccariophoenix, then it would be just as distantly related to Parajubaea, Syagrus and Butia as Beccariophoenix, but that is not the case morphologically and phylogenetically.

[awkonradi]

This article describes a fossil coconut in India, from near the K-Pg boundary, 66 million years ago.

https://www.researchgate.net/profile/Dr-Rashmi-Srivastava/publication/277265170_Cocos_2014_Acta_Pal/links/556581bb08ae89e758fda339/Cocos-2014-Acta-Pal.pdf

[Dimovi]

27 minutes ago, awkonradi said:

This article describes a fossil coconut in India, from near the K-Pg boundary, 66 million years ago.

This article states that the place of origin on coconut is stil in dispute. Regardless there are a lot of inconsistencies with the article. The oldest fossil record of a palm tree is about 80 million years ago, that would make Cocos one of the easiest phylogenetic branches and yet it is just a leaf node. Species that genetically distant cannot create an offspring (fertile or not), yet people are crossing Cocos with Butia and Parajubaea. Fossil record and even morphology is a much weaker argument to genetics.

[Xenon]

Or the common ancestor of coconut and Syagrus originated in the Indo-Pacific and a progenitor of Syagrus arrived and radiated in the Americas sometime after the Cocos lineage had already diverged from the last common ancestor between Cocos and the new world cocoides. 

Again, why would something so useful like coconut be unknown/go unnoticed to indigenous peoples of the Americas when every other arguably less useful palm is utilized in its native range i.e Euterpe, Oenocarpus, Bactris, Mauritia in the Amazon; Borassus, Arenga, Nypa in SE Asia; Elaeis in west Africa etc. 

[Dimovi]

5 minutes ago, Xenon said:

Or the common ancestor of coconut and Syagrus originated in the Indo-Pacific and a progenitor of Syagrus arrived and radiated in the Americas sometime after the Cocos lineage had already diverged from the last common ancestor between Cocos and the new world cocoides. 

Again, why would something so useful like coconut be unknown/go unnoticed to indigenous peoples of the Americas when every other arguably less useful palm is utilized in its native range i.e Euterpe, Oenocarpus, Bactris, Mauritia in the Amazon; Borassus, Arenga, Nypa in SE Asia; Elaeis in west Africa etc. 

When humans left Africa less than 100 thousand years ago to colonize the world there were already coconuts all over the tropical world. What humans did with coconuts is irrelevant. All recent ancestors are in South America and more distant ones are in Africa.

[Dimovi]

As I am researching this, it seems that genetic studies do point to American origin, but I guess it takes time to change the status quo.

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0007353

[Harry’s Palms]

This is all very interesting stuff and certainly beyond my level of education. I have read about human migration “Wilderness At Dawn” , good reading. Plants as well have moved around the planet by various means. My mind goes to the scene in the movie “Holy Grail” where the man is using coconut shells to imitate the clacking of horses hooves . “Where did you get the coconut shells? “If you haven’t seen it , it is hilarious. Sorry I just couldn’t help it! Carry on….. Harry……. it’s a British thing!

[Merlyn]

1 hour ago, Dimovi said:

Species that genetically distant cannot create an offspring (fertile or not), yet people are crossing Cocos with Butia and Parajubaea.

There are no crosses of Cocos with anything, to my knowledge.  Two people on here felt that they had a Cocos x Butia cross, but it has yet to be proven true, or in the most recent try, to have survived seedling stage.  OTOH there are plenty of people who have painstakingly but successfully crossed Parajubaea with Butia or Syagrus, and other varieties.  This points to Cocos being substantially "more different" from the close neighbors on the genetic tree.  In that study there might have been some correlation numbers that weren't in the graphic.  That would say a lot more about which ones were most closely related.  Also keep in mind that the genetic tree is not infallible.  It only uses a certain number of genetic markers to distinguish species. 

[PlantMorePalms]

There's a webinar on palms.org by Dr. Rodrigo Bernal about the palms of Columbia where he talks about this, and in addition to the genetic evidence he mentioned that what we know about ocean currents since the separation of Africa from South America would support the idea of coconuts floating from the coast of South America to polynesia and beyond. The question comes up at 59:23 in this video (you have to be an IPS member to watch):

https://palms.org/webinar-the-palms-of-colombia-with-dr-rodrigo-bernal/

 

[tim_brissy_13]

I know next to nothing on the topic but thought I’d just add to the above that’s there’s also Voanioala from Madagascar and the now extinct Paschalococos disperta from Easter Island as the other Cocoid genera. Not sure if this sways the argument one way or the other. 

[Merlyn]

10 minutes ago, tim_brissy_13 said:

I know next to nothing on the topic but thought I’d just add to the above that’s there’s also Voanioala from Madagascar and the now extinct Paschalococos disperta from Easter Island as the other Cocoid genera. Not sure if this sways the argument one way or the other. 

Yep, I thought that Voaniola was South American, so I didn't look up the origin earlier.  Whups.

Looking at the others nearby, it is interesting that Elaeis Guineensis is African (Guinea) and Olerifera is South American.  So closely related palms in the same branch being on different continents may be more common than we might expect.  Especially since many diverged 50 plus million years ago when the continents were much, much closer together.

[SeanK]

15 hours ago, Dimovi said:

Tomatoes are synonymous with Italian food and potatoes with Irish and both came from the new world just a few hundred years ago. Cocos evolved millions of years ago, so there is not an anthropological argument to be made. There is no doubt that Cocos is huge part of the culture from that part of the world and it may have been cultivated in Indo Pacific first, but the origin of the wild Cocos there is extremely improbable.

I'm inclined to agree. Chefs are a recent addition to Earth's history and many non-indigenous foods are used in "traditional" meals.

[bubba]

Very interesting discussion that I hope continues and expands.

[Johnny Palmseed]

On 5/6/2024 at 7:26 PM, Harry’s Palms said:

This is all very interesting stuff and certainly beyond my level of education. I have read about human migration “Wilderness At Dawn” , good reading. Plants as well have moved around the planet by various means. My mind goes to the scene in the movie “Holy Grail” where the man is using coconut shells to imitate the clacking of horses hooves . “Where did you get the coconut shells? “If you haven’t seen it , it is hilarious. Sorry I just couldn’t help it! Carry on….. Harry……. it’s a British thing!

Carried by a Swallow! Gripped by the husk…😆

[Jonathan]

An African swallow maybe but not a European one...lol

[Tyrone]

I read an article that the Spanish took the coconut to South America from the Phillipines. The coconut was an excellent source of clean water and nutrients for such a long journey across the Pacific which back then would have been more akin to  travelling to the moon nowadays. Once they got to South America they planted a few so that in a few years time they had coconuts for the return journey. I seem to remember reading that the DNA evidence backed that story up too. 
 

I personally believe that the coconut is originally from the Indian subcontinent. The district of Goa has its history filled deep with the coconut. Ocean currents and people moved it around the globe. 

[Merlyn]

@Tyrone a brief search finds this article about 66 million year old coconut fossils in Colombia with beetle predation.  While it's clear that humans moved coconuts around for food and water, coconuts clearly were in South America waaaay before humans existed.

https://www.sci.news/paleontology/seed-beetle-coconut-fossil-10762.html

[Tyrone]

On 5/7/2024 at 8:41 AM, tim_brissy_13 said:

I know next to nothing on the topic but thought I’d just add to the above that’s there’s also Voanioala from Madagascar and the now extinct Paschalococos disperta from Easter Island as the other Cocoid genera. Not sure if this sways the argument one way or the other. 

There was also an extinct coconut palm species from NZ too. The cocosoids seem to have been everywhere. Also nobody has mentioned Beccariophoenix. 

[Merlyn]

5 hours ago, Tyrone said:

There was also an extinct coconut palm species from NZ too. The cocosoids seem to have been everywhere. Also nobody has mentioned Beccariophoenix. 

Er...we also talked about Beccariophoenix starting with post #2.  I posted the semi-joking theory that Beccariophoenix was actually the original split and source of all the cocosoids.   

In a more serious note, the genetic tree really does say that they all originated in the vicinity of what's now South America.  If you look at the distribution of all the closely related species, every single one in the "lower tree splits" are in South America:

Ockham's Razor suggests that the simplest explanation is usually the correct one.  I can see four arguments:

1. So back 50+ million years ago seeds of a common palm ancestor (like the 66 million year old fossil in Colombia) floated off the South American continent and landed up the East coast of Africa.  These developed into Beccariophoenix, Jubaeopsis, and Voanioala among other extinct species.  This explains why those species are in the upper splits of genetic similarity.  The remaining 8 South American species developed locally.  Cocos probably developed near the coastline where evolutionary pressure favored salt tolerance and large seeds with thick husks.  Coincidentally that same feature allowed the seeds to float in the ocean for thousands of miles...thus popping up along the coastline of virtually every continent and island on the planet.  The "genetic diversity in the Indo-Pacific" is a result of Cocos arriving in the area and flourishing.
2. Cocosoids originated in India and spread to the other continents.  This would require the Cocos to be the first branch in the evolutionary tree, or for the other species to exist (or become extinct) in the Indo-Pacific area.  Or it would require some fossil record indicating the oldest Cocosoids were in India.  None of these three things exist.
3. Beccariophoenix (or a pre-Beccario) was the original species in Madagascar, and seeds floated off of the Madagascar/India island during the late Cretaceous period, landing in South Africa and South America.  This is reasonable from a geography, but doesn't explain why the oldest cocosoid fossils are in Colombia and at the same age in India.  And it doesn't explain why the "higher splits" in the genetic tree are also all South American. 
4. Cocosoids originated in Pangea around 200 million years ago, and our notion of "South America," "Africa" and "India" is irrelevant to the geologic time scale.  There are fossils of palms in what's now New Jersey and France, dating back around 100 million to 150 million years and even back as far as 180-230 million years.  Fossils of cocosoid relatives have been found in Colombia ~66 million years old and in India ~60-70 million years old.  That suggests that the genetic split in the tree is well before that time, and they could have been all growing in the Southern region of Pangea.  Keep in mind that the map during the evolution of the cocosoid branch looked really different from today. "Floating across vast oceans" simply wasn't a thing...  This also explains why most species in South America are genetically similar, as the sudden existence of the Atlantic Ocean split off those species from the rest of today's continents.  And it explains why Elaeis has a South American and an African species that are closely related but distinct.

[Xenon]

23 minutes ago, Merlyn said:

 

1. So back 50+ million years ago seeds of a common palm ancestor (like the 66 million year old fossil in Colombia) floated off the South American continent and landed up the East coast of Africa.  These developed into Beccariophoenix, Jubaeopsis, and Voanioala among other extinct species.  This explains why those species are in the upper splits of genetic similarity.  The remaining 8 South American species developed locally.  Cocos probably developed near the coastline where evolutionary pressure favored salt tolerance and large seeds with thick husks.  Coincidentally that same feature allowed the seeds to float in the ocean for thousands of miles...thus popping up along the coastline of virtually every continent and island on the planet.  The "genetic diversity in the Indo-Pacific" is a result of Cocos arriving in the area and flourishing.

 

The big problem with this is the lack of extant pre-Columbian coconuts in South America. Even the one isolated population in Panama has been linked with relatively recent Pacific trade with Austronesian seafarers. Every other coconut population in the Americas is linked to colonial trade routes i.e Philippines -> Mexico ("Pacific Tall"), West Africa -> Brazil & the Caribbean ("Jamaican Tall").  There are tons of historical accounts and indigenous artefacts of things like cacao, avocado, pineapple in South America etc but none whatsoever of coconuts (very hard to miss). I think it's pretty safe to say Columbus did not see coconuts lining the coast of Hispaniola when he landed there. 

It also seems unlikely that something with such a specialized niche like coconut would go extinct on the South American coast and then flourish when "reintroduced". But something like that happened with horses, so it's very possible. 

[Dimovi]

1 hour ago, Merlyn said:

So back 50+ million years ago seeds of a common palm ancestor (like the 66 million year old fossil in Colombia) floated off the South American continent and landed up the East coast of Africa.  These developed into Beccariophoenix, Jubaeopsis, and Voanioala among other extinct species.  This explains why those species are in the upper splits of genetic similarity.  The remaining 8 South American species developed locally.  Cocos probably developed near the coastline where evolutionary pressure favored salt tolerance and large seeds with thick husks.  Coincidentally that same feature allowed the seeds to float in the ocean for thousands of miles...thus popping up along the coastline of virtually every continent and island on the planet.  The "genetic diversity in the Indo-Pacific" is a result of Cocos arriving in the area and flourishing.

I'm thinking this is the most likely hypothesis. A less likely one is that the last common ancestor floated from Africa to South America, but that still does not implicate the Coconut as it evolved way after that split.

[ChicagoPalma]

A little new to the topic, but it seems to me there was some ancestor of the cocos palm that seemed to have originated somewhere in south america or india and not just the cocos palm itself. I think it would be better to look for a more older species that is related or has been related to the cocoid palm and related to most likely the cocos palm. I don't think the cocos palm was the original palm of where everything branched off from.

[Merlyn]

1 hour ago, Xenon said:

The big problem with this is the lack of extant pre-Columbian coconuts in South America. Even the one isolated population in Panama has been linked with relatively recent Pacific trade with Austronesian seafarers. Every other coconut population in the Americas is linked to colonial trade routes i.e Philippines -> Mexico ("Pacific Tall"), West Africa -> Brazil & the Caribbean ("Jamaican Tall").  There are tons of historical accounts and indigenous artefacts of things like cacao, avocado, pineapple in South America etc but none whatsoever of coconuts (very hard to miss). I think it's pretty safe to say Columbus did not see coconuts lining the coast of Hispaniola when he landed there. 

It also seems unlikely that something with such a specialized niche like coconut would go extinct on the South American coast and then flourish when "reintroduced". But something like that happened with horses, so it's very possible. 

Yet there are fossils of coconuts in New Zealand, India, Colombia, Australia, etc.  For example, here is a paper suggesting that the coconut fossil found in ~50+ million year old sediment in New Zealand is closely related to high altitude Parajubaea from South America. 

https://www.researchgate.net/publication/256767883_Fossil_Oligocene_coconut_from_Northland

Though certainly not conclusive, that suggests that Parajubaea-like palms were found all across the Southern end of Pangea prior to, and during the splitting up of the supercontinent.  Those species went extinct in New Zealand.

It's also likely that Cocos was wiped out in the Northern Hemisphere during the last ice age, if it existed there.  A 10F drop in regional temps in Florida would eliminate Cocos from the state, and maybe also from Cuba and Hispaniola.  And there's disagreement on whether the early European explorers found Cocos in Panama and other places.  For example, this paper stub suggests that Martir de Angleria found coconuts in Panama around 1500, and that they were closely related to Philippines coconuts:

https://academic.oup.com/aob/article/113/1/1/171140

[awkonradi]

I think it's entirely possible that coconuts, or "pre-coconuts", were floating around the world for much of the Cenozoic Era (last 66 million years).  Most of the Cenozoic was much warmer than today, and the Isthmus of Panama closed only 3 million years ago.  Why coconuts were not known in the Atlantic basin during the historical period, prior to European exploration of the Indian and Pacific oceans, would be a mystery.  I note that sea snakes also were limited to the Indian and Pacific oceans, despite apparently suitable habitat in the Caribbean.

[Xenon]

1 hour ago, Merlyn said:

 

It's also likely that Cocos was wiped out in the Northern Hemisphere during the last ice age, if it existed there.  A 10F drop in regional temps in Florida would eliminate Cocos from the state, and maybe also from Cuba and Hispaniola.  And there's disagreement on whether the early European explorers found Cocos in Panama and other places.  For example, this paper stub suggests that Martir de Angleria found coconuts in Panama around 1500, and that they were closely related to Philippines coconuts:

https://academic.oup.com/aob/article/113/1/1/171140

I'm sure the Pacific coast of Colombia remained warm enough for coconuts during any ice age, the coastal rainforest there has served as a refugia for eons. 

The one bottlenecked population of coconut is the only case ever discussed of pre-Columbian coconuts and the running theory is that they were introduced there by seafaring peoples from the other side of the Pacific.  

Sure we can talk about tens of millions of years ago, but there's nothing to suggest coconuts were ever present in 99% of the New World in the period spanning human arrival/Ice Age in the Americas and colonial times. In all likelihood, the extant Cocos nucifera hadn't been present in the Americas for millions of years if it ever was at all.

[Xenon]

26 minutes ago, awkonradi said:

I think it's entirely possible that coconuts, or "pre-coconuts", were floating around the world for much of the Cenozoic Era (last 66 million years).  Most of the Cenozoic was much warmer than today, and the Isthmus of Panama closed only 3 million years ago.  Why coconuts were not known in the Atlantic basin during the historical period, prior to European exploration of the Indian and Pacific oceans, would be a mystery.  I note that sea snakes also were limited to the Indian and Pacific oceans, despite apparently suitable habitat in the Caribbean.

The discussion is confusing when cocoid fossils or coconut-like palms from 60 million years ago are used as a proxy for +/- genetically modern coconuts. Wherever its ancient ancestry lies, it does seem like the center of origin and "natural" dispersal of extant coconuts is the equatorial Indo-Pacific and not the Americas. 

[awkonradi]

1 hour ago, Xenon said:

The discussion is confusing when cocoid fossils or coconut-like palms from 60 million years ago are used as a proxy for +/- genetically modern coconuts. Wherever its ancient ancestry lies, it does seem like the center of origin and "natural" dispersal of extant coconuts is the equatorial Indo-Pacific and not the Americas. 

I think there is more than one interesting question.

(1) Where lived the last common ancestor of Cocos and its nearest living relative, which MAY be Parajubaea?  The diversity of Cocoids in South America, and morphology, and genetics, MAY suggest South America.  I think this is probable, but not certain.  I note, the occurence of ancient and quite coconut-like fossils outside South America MAY be reason to question that South America is where Cocoids originated.

(2) Where did modern Cocos emerge, or last pass through a geographic bottleneck?  Historical records, and again genetics, MAY suggest the western Pacific, Indonesian archipelago, or Indian ocean.  Again, I think this is probable, but not certain.

I think we should be careful not to have too great faith in others' interpretations of fossil or genetic data.  I know from my own career that interpretations of genetic data often are revised profoundly over time.

[Merlyn]

7 hours ago, Xenon said:

The discussion is confusing when cocoid fossils or coconut-like palms from 60 million years ago are used as a proxy for +/- genetically modern coconuts. Wherever its ancient ancestry lies, it does seem like the center of origin and "natural" dispersal of extant coconuts is the equatorial Indo-Pacific and not the Americas. 

Well, there are really two questions to be answered.

1. Where did Cocosoids originate?  That's the genetic tree question that I was looking at.  The answer is probably Southern Pangea, given the widespread fossil and modern ancestors across the Southern Hemisphere from New Zealand, India, Madagascar, South Africa, and the rest of South America.  And of course other things like related (on a nearby branch in the diagram) Hyophorbes off the coast of Madagascar.  I was just originally thinking of it being distributed by ocean currents.  But given the age of fossils and the timeline of the separation of Pangea...it seems more likely that there were a whole batch of widely spread and related species ~100+ million years ago that got split up with the supercontinent.  It seems reasonable that Cocos' most closely related ancestors could also be widely distributed, like apparent Parajubaea ancestors in New Zealand.  It wouldn't surprise me that some populations went extinct.  One disease, predator, bug, or volcano could easily wipe out an entire island's (or continent's) population.  Just look at modern LY or LB diseases.
2. Where did modern Cocos Nucifera originate, and how did it spread across the globe?  It looks like genetic similarities on the Atlantic Ocean coasts suggest spanish distribution, and Pacific coast populations suggest Polynesian or some other Indo-Pacific origin.  The fact that Cocos can't be crossed with any modern relatives also means that the ancestors in question 1 were probably split off geographically from any close relatives for many millions of years.  So from that perspective a single Indian/Southern Asia origin seems totally reasonable.

[Xenon]

9 hours ago, Merlyn said:

Well, there are really two questions to be answered.

1. Where did Cocosoids originate?  That's the genetic tree question that I was looking at.  The answer is probably Southern Pangea, given the widespread fossil and modern ancestors across the Southern Hemisphere from New Zealand, India, Madagascar, South Africa, and the rest of South America.  And of course other things like related (on a nearby branch in the diagram) Hyophorbes off the coast of Madagascar.  I was just originally thinking of it being distributed by ocean currents.  But given the age of fossils and the timeline of the separation of Pangea...it seems more likely that there were a whole batch of widely spread and related species ~100+ million years ago that got split up with the supercontinent.  It seems reasonable that Cocos' most closely related ancestors could also be widely distributed, like apparent Parajubaea ancestors in New Zealand.  It wouldn't surprise me that some populations went extinct.  One disease, predator, bug, or volcano could easily wipe out an entire island's (or continent's) population.  Just look at modern LY or LB diseases.

It might be interesting to draw comparisons with another primitive palm with a well documented fossil record: Nypa. It was pantropical/present in the Americas in the late Cretaceous but disappeared from the fossil record there sometime after. The extant species is native to the Indo-Pacific but has naturalized in West Africa and to some extent in parts of the Caribbean. 

[Blueman]

There's maps that show the primary origins of coconuts based on DNA

1) Niu Vai type: round, thin husk, early germinating  - In the Philippines, Indonesia, & SE Asia which where taken to Madagascar by Austronesians, and across the Pacific to Panama by pre-Columbian people, and later to Mexico by the Spanish.  

2) Niu Kafa type: long fruit, thick husk, late germining (floating longer)  - In India/Sri Lanka which were taken to East Africa by Arab/Persian traders, & West Africa by the Portuguese, and later from Africa to the Caribbean.  

3) Niu Leka type: have characteristics of both Talls and Dwarves - In Fiji/Samoa.