A quick google search will show you that bananas are the most popular fruit in the world according to multiple sources (excluding trickery such as switching the culinary definition of fruit for the botanical one). Since 2019, approximately 20 million tons of bananas were grown and exported to the global market yearly, with Central America in the lead for producers (Guatemala alone exports 6 million tons per year!), while the EU and the USA lead the world in gobbling up all those bananas for domestic consumption (26.7% and 21.7% of global banana imports in 2022, respectively).1
This delicious fruit is well-known for its high content in vital potassium and magnesium, as well as its relatively high proportion of carbs. While most people are aware that bananas grow on trees, and that they require a tropical climate to thrive, it’s likely that few would be able to point out their area of domestication on a map, or tell you who domesticated them. As it turns out, even the scientists aren’t entirely sure of all the details themselves. Let’s take a look at some of the latest litterature on the topic to try and understand the murky origins of the world’s favourite fruit.
The inspiration for writing this post came from my recent read-through of Stone Age Herbalist’s first book “Berserkers, Cannibals and Shamans”, more specifically the essay looking into the impact of the Austronesian expansion on Australia.
Botanical Fundamentals
As usual, we’ll start by an overview of the plant which will help us understand the context of its domestication and cultivation later on. Bananas belong to the order Zingiberales, within the family Musaceae, which makes them relatively close relatives of edible ginger, from the family Zingiberaceae found within the same order. There are two main species whose genes are found in cultivated banana, these being Musa acuminata, the most common one, and M. balbisiana. The former has also been divided into 9 subspecies, ranging in their distribution from Northern India to New Guinea. Indeed, the crop originates from the region of Southern/South-East Asia, but also underwent domestication in some Melanesian islands.
These plants are monocotylones, one of the two great clades of flowering plants, which means they share fundamental physiological traits with such diverse species as grasses and orchids. I go a little bit further into the details of how monocots differ from their dicot cousins in part of a previous twitter thread. Like many monocots, bananas never actually figured out the trick to becoming a true tree, but use trickery instead, making them giant herbaceous plants. While dicot trees have trunks and branches which give them their rigidity and structure, monocot “trees” are in fact tight bundles of hardened leaf sheaths. There is no bark or wood involved.
Essentially, the plant can be divided into three components. Underground, we find the “corm”. This tuber-like underground structure produces the roots of the plant, and will eventually clone itself via the production of “suckers” (genetically identical “child” plants that’ll grow close to the original). This corm is the actual true stem of the plant, even though it is entirely subterranean. It carries between 10 and 15 buds which might form suckers, of which there are four types.
The plant’s height is formed by a “pseudostem”, which as mentioned above is a tightly rolled bundle of leaf sheaths, pushed upwards by growth of new leaves initiated from the corm at ground level, carrying the rolled up leaves at the top where they unfold progressively as the plant reaches maturity. Under good conditions a new leaf appears every week, poking out of the pseudostem just as the previous new leaf finishes spreading for good. Finally, the very top of the plant eventually produces a large dangling “inflorescence” (bundle of flowers) from whence the bananas themselves will come. The size of the plant varies between person-sized and several metres in height, depending on the cultivar.2
The inflorescence is an impressive beast. Growing out in lowland tropical areas after close to 30-40 leaves have been produced by the plant, it sprouts from the centre of the corm’s true stem (the aerial stem) until it pokes out of the pseudostem. Botanically speaking it is a highly specialised floral spike, although unlike most spikes observed in the plant kingdom, instead of pointing upwards when it flowers, it sags downards because of its large weight. The female flowers which will form into bananas are located at the base of the inflorescence, while the few rows of male flowers (rich in nectar) are located at the tip, covered by thick purplish bracts which eventually fall off. Between these two groups are “neutral” flowers which won’t develop into fruits. The entire inflorescence is originally covered by purple bracts in a spiral arrangement which successively fall off to reveal the flowers as the rachis elongates.3
As for the fruits of banana plants, a description of the domesticated type is certainly unnecessary for most people, but it may still be good to remind people that it is botanically classified as a berry: a simple fruit type, commonly found among plants, which is derived from a single fertilised ovary, doesn’t have differentiated internal tissues, and contains many seeds.4 Indeed, it’s easy to forget that like all plants wild bananas reproduce in this way, since domestication rendered most varieties functionally infertile with regards to sexual reproduction.
Ecologically speaking, wild bananas are associated with disturbed or cleared areas in tropical and subtropical climes. This means places where the vegetal cover was disturbed so that bare soil, or at least the shorter grassy vegetation at ground level, is fully exposed to sunlight. Once implanted there through seed dispersal (likely by a member of the bewilderingly large range of animals that feed on the fruit) it can quickly come to dominate the area through the production of a large patch of suckers, to the point they compete with each other for light. This is true across many species of wild bananas, regardless of whether they contributed to the domesticated genome or not. There are however still variations regarding soil type, level of moisture and compaction, etc.5
This type of ecological preference may be significant, as humans have been, at least since the neolithic, a major source of land clearance through deforestation, settlement construction, or various techniques such as slash-and-burn agriculture. I’ve already speculated in the past, within the context of cannabis, that some plants may have attracted human attention thanks to their preference for establishing themselves in areas affected by frequent human activity.
Setting up the pieces
Cultivated banana is broadly split into two categories: the most familiar to Western people may be “dessert” banana, large and very sweet, only eaten when ripe, usually raw or minimally processed. The other kind, “plantain” banana, is richer in starch and has a milder taste. It is typically only used for cooking rather than eaten raw and can be prepared even in an unripe state, its starchiness making it a tropical food staple prepared much in the same way as we cook potatoes. Genetically it is characterised by the presence of two genome elements: “A” originates from M. acuminata, while “B” comes from crosses with M. balbisiana.
In a wild state, M. acuminata subspecies are widely spread out over Melanesia, Indonesia and South-East Asia, while M. balbisiana is thought to have originally been restricted to the mountainous regions of continental South-East Asia, with some incertitude over whether the Northern Philipines population is truly indigenous or not. Meanwhile, current research indicates that the genesis of the domesticated banana took place in New Guinea. Considering M. balbisiana isn’t native to the area, how did the two meet to intermingle?
The answer, of course, is human activity. Musa balbisiana is known to have long been cultivated as a source of banana fiber, as both its pseudostem and its leaf veins can be harvested to obtain long and very strong fibers which were used for rope-making, mat-weaving and potentially clothing.6 As an aside, there is a species of wild banana called M. textilis, claimed to be used for fiber production, but investigations at least in the Philipinnes showed that this was often just a misidentified M. balbisiana. There is no archaeological evidence for contact between Southern China and the Philipinnes before 500BC that could have led to the plant’s introduction there, although some linguistic inquiries would place its introduction back to 1500BC at least, before domesticated banana types differentiated.7
The domestication process
Now we have our two culprits in the same area, how do we make domesticated banana out of them? Well, most cultivated banana types are actually triploids, meaning, instead of having just two copies of their DNA, like us animals, they have three. This is usually the result of interbreeding between two species: one diploid and normal, and one tetraploid (4 copies of the genome). Tetraploids are usually themselves hybrids, although completely viable in the wild, and it’s believed the process may somehow enhance their capacities to adapt to new environments during colonisation.
A triploid plant will appear morphologically normal, although it can occasionally result in even more vigour than normal specimens, but because of the odd number of gene copies meiosis is unable to complete itself, resulting in a lack of gametes. The most obvious consequence of this is parthenocarpy: the lack (or extreme reduction) of seeds inside the fruit, with bananas being an archetypal example of the process.8 The plants are thus exclusively spread vegetatively (clonally), most commonly by replanting suckers and roots tock, since they produce no viable seed.
However there are also edible “semi cultivated” or “cultiwild” pure strain M. acuminata individuals, some parthenocarpic, others simply selected for increased fruit pulp production which were achieved throughout S-E Asia through the crossing of different subspecies rather than species. In this case M. acuminata subsp. banksii from Papua is the main genome contributor, although as one moves away from New-Guinea its influence wanes as crosses take place with the other subspecies.[7]
It should be noted that beyond these original hybridations between subspecies of M. acuminata (and occasionally involving M. balbisiana), recent genetic analyses recovered unknown contributors to the Cavendish dessert banana variety subgroup, accounting for over 50% of all global production. Based on our current genetic knowledge, it’s been proposed that M. schizocarpa, a very under-examined species, could be involved. Besides that, 2 other sub-species of M. acuminata play a more or less minor role. It’s believed that classical banana breeding technique for forming sterile but productive triploids was a one-step hybridation between parent species/subspecies, but nowadays the preferred technique is to separately improve the normal (diploid or tetraploid) parent types through selective breeding before crossing them to form a comercially viable triploid.9
The differentiation between dessert and plantain bananas comes down to their pedigrees. Plantain is heavily influenced by M. balbisiana which is also a very starchy plant (this is described as the “AAB or ABB group”), while dessert bananas typically involve only subspecies of M. acuminata or other species of Musa besides M. balbisiana (the “AAA” group). As we’ll see below, these crosses likely took place in different areas of the islands of S-E Asia at different times.
Chronology
This is the least understood aspect of banana domestication, as it took place in tropical contexts which aren’t always kind to biological remains and under-prospected, spread out over a thousand years in different steps across different areas of S-E Asia and Melanesia. For a quick summary, Musa-type phytoliths were found in Sri-Lanka (Beli-Lena) from 11500—9500BC but presumed to be from exploitation of wild specimens. More are found in the Papuan highlands dating to 8000BC, but some hypothesize they resulted from fiber banana exploitation and not edible banana. There is more definite proof of edible banana cultivation (the banksii type) in the Kuk swamps of the latter region, from 5000-4500BC, the earliest date the species was cultivated. However, some people still believe diploid M. acuminata banksii banana crosses appeared in the early Holocene, up to 11000 years ago. As mentioned earlier, linguistic work seems to favour a pre-Austronesian timeframe for the domestication of the banana, meaning it must have largely taken place before 2000-1500 BC, the general era of their arrival in New Guinea.10
Once the initial domestication process took place, there were two broad movements centered on New Guinea: a spread of “plantain” bananas Westwards towards Africa, and one Eastwards towards the Pacific islands. The latest date at which bananas could have been started being cultivated in Subsaharan Africa would be approx. 725-75 BC based on volcanic phytolyths and charred pottery remains (they must have been cultivated since bananas of the Musa genus aren’t found in the wild there) while less certain phytolyths broadly assigned to the Musa genus were found in Pakistan in 2000BC (also outside the genus’ native range) and in 1500BC from the Lapita culture in the Western Pacific.
Paradoxically, the triploid varieties most associated with banana cultivation are mainly grown outside the original hybridation zone of South Guinea: African lines such as “Mutika lujugira” (broadly East African) or “African Plantain” (central and West African) as well as the “Pacific Plantain” types. Let’s of course not forget the “Cavendish” dessert banana, heavily cultivated in Central America. It’s interesting to note that the two main African varieties come from different sub-areas of the Island S-E Asia hybridation zone, as they have different parent (sub)species. In any case both linguistics and archaeology agree that this initial East coast introduction took place pre Bantu-expansion (a phenomenon that a twitter mutual of mine wrote about here), although no one really seems to know how the Pakistani phytolyths fit into this, besides suggesting a land trade route could be possible.11 Let’s not forget that the Indus Valley Civilisation of Western India had been in contact with Eastern Africa as early as the 3rd millenium BC. It’s also generally believed currently that despite the existence of ancient Pakistani banana phytolyths, banana dispersion to the Middle-East, North-Africa (MENA) region post-dates the Islamic conquests of the 7th century AD.
Worldwide spread
The long-scale translocation of crop plants by humans is a phenomenon almost as old as agriculture itself, and observed worldwide. I’ve talked about the spread of cereals of various kinds before, and even my recent posts on salvia and tobacco point in the same direction. Let’s see how these same dynamics played out with bananas.
The long scale transmission of bananas would be quite doable to someone with the know-how. Indeed, even without seeds, the plants can be proliferated by digging up, dividing, and drying out the root-stock, which can survive in this dormant state for at least six months and probably far more, starting a new growth once replanted in favourable conditions. It’s believed that through this means, the Austronesians carried them to all the major islands of the Pacific ocean, since, indeed, bananas have been found on all of these by European explorers, even on the then-uninhabited Pitcairn island, which holds the title of “most remote island in the world” (halfway between Chile and New-Zealand, for more info I recommend this utterly charming amateur documentary free on youtube).
The transmission to America is more contentious, with the orthodox position being that the monk Thomas de Berlanga introduced the species to Hispaniola, taking root-stock from the Canary Islands that was introduced there by the Portuguese in the early 15th century, beginning a wave of introductions throughout Spanish and Portuguese controlled lands. However, a second theory is that the plant was known in tropical regions through introduction on the coast of Equador by Polynesian navigators at least a few centuries before contact with Europeans took place. Its critics argue that it relies too much on theoretical trans-Pacific crossings.12 More modern publications don’t even mention this theory, although I couldn’t find an in-depth analysis of American banana plant genomes that would definitely put the question to rest (the banana plants taken Eastwards by ancient mariners into the Pacific being a different genotype than the ones taken Westwards towards Africa and the Middle-East). It’s worth noting that some authors are quite dead set that contact did take place somewhere along the Ecuadorian coast prior to 1000AD, based on the presence of South American sweet potato in Eastern Polynesian islands, and linguistic similarities between the word for it in both languages.13
Despite the overall harshness of the climate, the Arab world, especially the Arabian peninsula, is still a player in global banana production (with 2% of global production). Thanks to the wide-ranging trade that took place from the peninsula even in ancient times, it may have served as a hub of banana transmission. Most notably, the same banana genotypes are found in Oman and Egypt, indicating a transfer along the Red Sea. However, no banana remains were found that pre-dated their earliest known African cultivation, so a “central Indian Ocean corridor” hypothesis is preferred to the idea of a coastal road of dispersion from South Asia to the Middle East then East Africa.
Once arrived in East Africa, it’s believed the banana was transmitted westwards via human movement along the Northern fringes of the equatorial forests into West Africa, where they may have provided the easily processed nutrients required to fuel population growth and help kickstart the Bantu expansion itself, or at least its latter phases. However, because of the scarcity of evidence for banana cultivation in East Africa and the intervening dry lands unsuitable for its cultivation between the East and West coasts, an even more ambitious hypothesis has been proposed of an Austronesian circumnavigation of Africa, bringing bananas directly to the West. This is of course extremely speculative.14
This is not to say that bananas bypassed the Middle East completely: in his conquests, Alexander the Great’s army came upon banana plants in western India, leading to the oldest depiction of the species, in Greek texts dating from 327 BC. One academic also claims that banana is cited in the Quran as the “tree of paradise”, which would contradict the currently held belief that the plant’s dispersal into the Arab world post-dated the Umayyad conquests of Persia and North-Western India. There’d be nothing surprising about a much earlier familiarity of the Arabs with bananas, as traces of extensive trade and especially crop exchanges between Gujarat, the Arabian Peninsula and East Africa are known as far back as the 3nd millenium BC.15 But remember that African plantain and East-African cooking bananas can’t have come from India, because their genotypes only match those from the South-East-Asian islands!16
Much needed recap and conclusion
As you’ve been able to glean by now, to be able to form a full picture of the spread of the banana would require at least a conversational knowledge of most of the pre-European contact history of the entire global South: the mainland East Asian and Island South-East Asian agricultural domestication hotspots, the prehistoric Indian Ocean trade (between the latter region, India, the Arabic Peninsula and East Africa), the Austronesian expansion, the agriculture of Southern China, the peopling of Madagascar, the conquests of the Umayyad Caliphate and its internal trade network, the Bantu expansion, and the population dynamics of post-conquest South and Meso America in the 16th century. A history poster of the caliber and thoroughness of Nemets could likely pull it off, but certainly not someone with as narrow a focus as me!
In short(ish), bananas were domesticated in the islands of S-E Asia especially around New Guinea, where the native Musa acuminata subspecies was mixed with other subspecies from surrounding islands to form an edible diploid hybrid, which (possibly accidentally) combined with fiber bananas of the species Musa balbisiana brought South from the Asian mainland to form the triploid seedless edible bananas we know today. Some of these were spread Westwards, probably through sea trade, via a purely oceanic shipping lane making use of seasonal winds, or a coastal route going around India and the Middle-East up to the Red Sea (but in this case, without the bananas spreading anywhere else, or at least not leaving any traces), while others were later carried Eastwards by the Austronesian people into the farthest reaches of the Pacific ocean.
The Westward travelling bananas eventually made it to the East African coast, before the occupation of Madagascar or the Bantu expansion, through at least two waves of introduction of genetically distinct types from different islands of South-East Asia. Some are only found in central and West Africa, while others are found in East Africa. They may have spread westwards from the point of contact through human transportation, or even been carried directly to the Western coast by extremely adept Austronesian navigators. In any case, the oldest finds of banana phytolyths in Africa being in Cameroon, near the heart of the Bantu expansion, and dated to a time during which this event was occuring, leads some to believe they played an integral role in providing a starchy, nutritious food staple that this agricultural society required to grow its population.
Currently archaeology seems to mostly consider the Arab world a dead end which was bypassed by ancient banana cultivation, this could be due to a lack of contact with the plant, or a failure to thrive of bananas in the area at the time, for climatic or human reasons. In any case the plant seems to have been at the very least known to some of the inhabitants by the time of the Islamic conquests, and a couple of centuries later the Umayyads came across them firsthand in India.
America is also an unknown, although the general consensus is that the fruit was brought to Hispaniola by the Portuguese in the early 16th century AD, and rapidly spread throughout the tropical regions because of its high nutritional value and ease of clonal propagation, so that it was well established by the 17th century.
I hope you found this very abridged version of the current state of the art on banana domestication to have been instructive, or at least entertaining.
FAO. 2023. Banana Market Review 2022. Rome.
Swennen & Ortiz 1997. Morphology and growth of plantain and banana. International institute of tropical agriculture (IITA) research guide 6.
Karamura & Karamura 1995. Banana morphology - part II:the aerial shoot. Bananas & Plantain, chapter 7.II, edited by Gowen.
https://www.britannica.com/science/berry-plant-reproductive-body
Vu, T.D., Vu, D.T., Janssens, S.B. et al. The description, distribution and habitat of wild banana species in northern Viet Nam. Genet Resour Crop Evol 70, 479–504 (2023). https://doi.org/10.1007/s10722-022-01442-2
https://textileengineering.net/types-properties-and-uses-of-banana-fibre/
De Langhe et al. 2015.The Original Banana Split: Multi-disciplinary implications of the generation of African and Pacific Plantains in Island Southeast Asia http://dx.doi.org/10.17348/era.14.0.299-312
Sellars, 2013. Triploidy. Brenner’s encyclopedia of genetics: second edition. doi:10.1016/B978-0-12-374984-0.01582-5
Martin et al. 2020. Genome ancestry mosaics reveal multiple and cryptic contributors to cultivated banana. The Plant Journal, 102: 1008-1025.
Donohue & Denham, 2009. Banana (Musa spp.) Domestication in the AsiaPacific Region: Linguistic and archaeobotanical perspectives www.ethnobotanyjournal.org/vol7/i1547-3465-07-293.pdf
Pettier et al. 2011. Multidisciplinary perspectives on Musa (spp.) domestication. www.pnas.org/cgi/doi/10.1073/pnas.1102001108
Langdon, 1993. The banana as a key to early American and Polynesian history. http://dx.doi.org/10.1080/00223349308572723
Scaglione & Cordero, 2013. Did Ancient Polynesians Reach the New World? In: Polynesians in America: Pre-Columbian Contacts with the New World, chapter 9.
Fuller et al. 2011. Across the Indian Ocean: The Prehistoric Movement of Plants and Animals. http://dx.doi.org/10.1017/S0003598X00067934
Al-Busaidi, 2013. Banana domestication on the Arabian Peninsula: A review of their domestication history. DOI 10.5897/JHF2013.0325
Mbida et al, 2005. The initial history of bananas in Africa. A reply to Jan Vansina, Azania, 2003. http://www.tandfonline.com/action/showCitFormats?doi=10.1080/00672700509480419
Great post so far. Noticed a typo early on: litterature
Lan, you've written another banger🏆