How Linnaeus inadvertently muddled the taxonomy of bananas when he gave Latin binomials to two edible bananas. This is the first of a two-part series on the classification of edible bananas. Part two is on modernizing the nomenclature system specific to bananas that were domesticated for edibility.
"Surely there is no other science in which the errors of the past are preserved by law to confuse the present", said Normand Simmonds in his 1991 address to the Society for Economic Botany. The author of two influential books—Bananas, and The evolution of the bananas—was referring to the nomenclature system formalized by Carl Linnaeus that consists in giving to species Latin binomials (Latinized two-part names: the first one indicating the genus, and the second one the species). The problem Simmonds was alluding to was their application to crops, that like bananas, potatoes and sugarcane have complex genetic histories involving hybridization between species and polyploidy (see The domestication of bananas).
The domestication of bananasBananas are native to the humid tropic belt that extends from India to the Solomon Islands. Like human beings, wild bananas are diploid, that is they have two copies of each gene-bearing chromosome. When they are pollinated, the flowers of wild bananas produce small fleshy fruits full of seeds1 . A genetic predisposition to parthenocarpy (the ability to produce a fruit in the absence of pollination) in at least one species that we know of, Musa acuminata, set the stage for the domestication of most sterile and seedless edible bananas.
The sterility of edible bananas is attributed to a combination of structural and genetic factors. The structural factors are linked to matings between more or less distant relatives (different subspecies of Musa acuminata, or when the latter hybridized with Musa balbisiana)2 . This is because inheriting from each parent chromosomes that are mismatched makes it difficult for the progeny to produce fertile ovules and pollen. There is also evidence that farmers preferentially propagating plants producing fruits with less seeds unwittingly selected for genes that contribute to sterility3 .
The first edible bananas were diploid, but nowadays farmers mostly grow the more productive triploid bananas (three copies of each chromosome). Triploidy arose when one of the diploid parents normally passed on one copy of its genome, while the other contributed both copies (a phenomenon called meiotic restitution).
Although edible bananas are sterile, except for residual fertility that allows them to be used in breeding schemes, their characters are not fixed. Mutations occur. Diversity is produced by farmers propagating mutant plants that exhibit desirable traits.
One consequence of their untidy histories, simmonds noted is "that conventional Latin-based, rule-bound nomenclatures tend to collapse under the strain of identification, priority and typification." He argued that the problems created by applying Latin binomials to these crops disappear when "Latin tags and formal rules are abandoned and informal genetically-based groups and cultivar names are adopted instead".
If Simmonds sounded confident, it’s because he thought he had made the problems disappear for bananas when the scoring method he published in 1955 with his colleague Kenneth Shepherd later became an informal nomenclature system to classify cultivars on the basis of the relative contribution of their wild ancestors to their genetic make-up. While the new system has spurred scientific progress, it is not without its own drawbacks and unintended consequences (see part two). It also did not make Latin binomials disappear.
But let's start with the Latin binomials that opened the door for the hundreds that followed: Musa paradisiaca and Musa sapientum, coined by Linnaeus himself. Unlike the Latin binomials assigned to wild plants, which refer to individual species, these names do not hark back to any cultivar in particular when they are still used in the scientific literature. That was not originally the case. To their credit, Linnaeus and the botanists who followed his lead were trying to give unique names to banana cultivars.
The muddied waters of Linnean names
Because it is the first Linnean name given to a banana, Musa paradisiaca is, according to the rules of the botanical code, the type species4 for the genus Musa. Linnaeus included it in the first edition of Species Plantarum, the 1753 publication that marks the boundary between Linnean and pre-Linnean names. His model was an edible banana growing in the glasshouse of a Dutch banker. It was famous for being the first banana to flower in Europe.
For the 10th edition of Systema Naturae, which came out in 1758, Linnaeus distinguished a second edible banana, which he knew only from drawings by George Ehret in Christoph Jacob Trew’s Planta selectea (plates 21, 22 and 23). He called it Musa sapientum.
The paradox is that Linnaeus did not consider cultivated plants to be species since they are the product of a process mediated by people. That’s why he usually described them as varieties of wild species, a practice for which he was later criticized because it dragged cultivated plants into the classification system for wild plants5 .
Evidently Linnaeus could not classify his banana models as varieties since he didn’t know from which species they derived. Whatever his reasons for describing these two cultivars, thereby elevating them to the rank of species, the problem is that it led to a disproportionate emphasis being put on cultivated bananas by the botanists who followed in his footsteps. One consequence is that for a long time the taxonomy of the genus was a confusing mess, especially since the application of Latin binomials to cultivars varied widely among botanists.
Many described the edible bananas they came across as varieties of Musa paradisiaca or Musa sapientum. The World Checklist of Selected Plant Families curated by the Royal Botanic Gardens, Kew, lists 58 varieties of Musa paradisiaca and 64 varieties of Musa sapientum. Other botanists gave them species names, like Musa cavendishii (now known as 'Dwarf Cavendish') and Musa emasculata (to reflect the absence of a male bud in a Horn Plantain cultivar).
It didn't help that the European botanists handing out Latin names left and right were usually just passing through. As Frank Howes, a botanist at the Royal Botanic Gardens, Kew, commented after collecting wild and cultivated bananas in southeast Asia in the 1920s: "a botanical classification…can only be attempted by one who is resident for some time in one of these countries and is able carefully to observe the different varieties in their different stages of development".
That’s exactly what K. Cherian Jacob did for 15 years in the province of Madras in India. His conclusion, which he published in 1952, in Madras bananas, was that the traits supposedly characteristic of Musa paradisiaca were also found in typical Musa sapientum varieties. Hence his solution, which was to rename 64 cultivars as varieties of Musa sapidisiaca. Not that it matters anymore, but for reasons of priority the mash-up name should have been Musa paradientum.
Attempts were also made to squeeze wild species in a framework based on Musa paradisiaca and Musa sapientum. Because the latter produced seedless fruits, Musa sapientum subsp. seminifera and Musa paradisiaca subsp. seminifera were created for seedy bananas, to which varietal names were sometimes added to precisely identify a plant. As David Constantine, the compiler of an annotated list of Musa names, remarked about Musa paradisiaca subsp. seminifera var. hookeri: "A nomenclature system that gives to a seed-bearing diploid species (Musa sikkimensis Kurz) the status of a variety of a subspecies of a seedless triploid is obviously in trouble although, of course, this could be recognised only in hindsight."
Cutting through the fog
As far back as 1865, Wilhem S. Kurz—a German botanist who worked in India, Indonesia, Burma (now Myanmar) and Malaysia—had come to the conclusion that Linnaeus’s bananas were hybrids, and even suspected the identity of the two species involved. However, his work, cut short by his untimely death, was mostly ignored. It was not until the 1940s that Ernest E. Cheesman, a British botanist who had spent decades studying bananas—at the Imperial College of Tropical Agriculture in Trinidad and Tobago, and at the Royal Botanic Gardens, Kew, in England—made progress in sorting out the confusion. By ignoring the cultivars, he was able to tease out from the scientific literature some of the most important species in the genus Musa and give them modern definitions. Among his achievements are proper definitions of Musa acuminata and Musa balbisiana, which by then had been recognized to be involved, albeit not always together, in the domestication of most edible bananas.
Cheesman also criticized the practice of applying Musa paradisiaca to "plantains" and Musa sapientum to "bananas", based on the resemblance of Linnaeus’s bananas with, respectively, a Plantain cultivar and a dessert banana, most likely a Silk cultivar. He even researched the origin of opposing the words bananas and plantains because he wanted to break down the association with the Linnean names. The association, he wrote, "must be broken down before we can arrive at any approximately correct nomenclature for the cultivated bananas, and, as old and familiar usage should not be set aside without reasoned justification, the position may be made clearer by showing that the association arose quite accidentally in the first place and never had good authority behind it" (see The origin of "bananas and plantains").
The origin of "bananas and plantains"
The modern usage of calling the sweet bananas that are eaten raw, banana, and the starchy ones that are cooked, plantain, arose in English by chance, according to E.E. Cheesman. He pointed out that both the Spanish word platano, from which the word plantain is derived, and the word banana6 , a word spread by the Portuguese7 , originally applied to all types of bananas. He also noted that in countries where bananas are indigenous, the local word for banana is used for cooking and dessert bananas alike.
Cheesman tried to pinpoint the time when authors started making a false distinction between bananas and plantains. The earliest references he found are the 1657 History of the Barbados and the 1696 Catalogue of Jamaican plants. These early works suggested to him that in the British West Indies only two types of banana were commonly planted. He hypothesized that one of them had a Spanish source and the other a Portuguese one. "In any event", he wrote, "it seems to have been quite by chance that, out of all the many banana varieties that exist, the two at first generally established in the West Indies belonged to different genetic groups so that the culinary difference in the quality of the fruit was reinforced by genuine botanical differences."
By 1947, the British botanist had become convinced that it was impossible to reconcile the wild and cultivated bananas within the Linnean nomenclature system. "The classification of the cultivated varieties", he wrote "is almost a separate problem from the taxonomy of the genus, needing a different technique for its solution, and confusion of the two makes both almost impossible".
Cheesman did not propose a strategy for attacking the problem of how to classify edible bananas. The solution came from his younger peers, Norman Simmonds and Kenneth Shepherd, who both worked on bananas at the Imperial College of Tropical Agriculture in Trinidad and Tobago in the 1950s. (Continued in part two)
This is the first of a two-part series on the classification of edible bananas. See part two on modernizing the nomenclature system specific to edible bananas.