Nomenclature system for cultivated bananas
The nomenclature system used to classify banana cultivars was developed by Norman Simmonds and Kenneth Shepherd in 1955. It classifies cultivated bananas into genome groups, according to the relative contribution of their ancestral wild species, and into subgroups, sets of cultivars that are related to each other through somatic mutations. This system eliminates almost all the difficulties and inconsistencies of a taxonomy based on Musa paradisiaca and Musa sapientum. However, due to difficulties in assigning bananas to their corresponding groups and subgroups, there are inconsistencies in the way the system is applied. Adding to the confusion is the continued use of Latin binomials to classify cultivated bananas. Cultivated bananas are unusual in not having a Latin scientific name.
Simmonds and Shepherd's genome-based system
In this system, bananas, at least the ones that are related to Musa acuminata and Musa balbisiana, are classified according to the relative contribution of these species designated by the letter A, for acuminata, and B, for balbisiana. A cultivar is assigned to a genome group according to the number of chromosome sets in its genome (its ploidy) and the species that donated them. Diploid cultivars can belong to the AA or AB genome group, while triploid cultivars fall into three genome groups: AAA, AAB and ABB (see the cultivar diversity portal).
Some taxonomists recognize a BBB genome group, but its existence has not been conclusively demonstrated. Tetraploid cultivars are mostly hybrids produced by breeders.
Genome groups are further divided into subgroups usually defined as a set of cultivars derived from each other through somatic mutations. On the basis of this system, cultivar names are put between inverted commas and preceded by the name of the genus and when known, the name of the group and subgroup. For example: Musa AAA (Cavendish subgroup) 'Robusta'.
The system is based on 15 characters that were chosen because they are different in Musa acuminata and Musa balbisiana. Each character is scored on a scale from one (typical Musa acuminata) to five (typical Musa balbisiana). The possible total scores range from a minimum of 15 to a maximum of 75. The expected scores are 15 for AA and AAA, 35 for AAB, 45 for AB, 55 for ABB and 75 for BB.
|Character||Musa acuminata||Musa balbisiana|
|Pseudostem colour||More or less heavily marked with brown or black blotches||Blotches very slight or absent|
|Petiole canal||Margin erect or spreading, with scarious wings below, not clasping pseudostem||Margin inclosed, not winged but clasping pseudostem|
|Peduncle||Usually downy or hairy||Glabrous|
|Ovules||Two regular rows in each loculus||Four irregular rows in each loculus|
|Bract shoulder*||Usually high (ratio<0.28)||Usually low (ratio>0.30)|
|Bract curling||Bracts reflex and roll back after opening||Bracts do not reflex|
|Bract shape||Lanceolate or narrowly ovate, tapering sharply from the shoulder||Broadly ovate, not tapering sharply|
|Bract colour||Red, dull purple or yellow outside; pink, dull purple or yellow inside||Distinctive brownish-purple outside; bright crimson inside|
|Colour fading||Inside bract colour usually fades to yellow towards the base||Inside bract colour usually continuous to base|
|Bract scars||Prominent||Scarcely prominent|
|Free tepal of male flower||Variably corrugated below tip||Rarely corrugated|
|Male flower colour||Creamy white||Variably flushed with pink|
|Stigma colour||Orange or rich yellow||Cream, pale yellow or pale pink|
Previous nomenclature system
Before Simmonds and Shepherd's system, cultivated bananas were classified using the binomial nomenclature system developed by Carl Linneaus that is used to this day to name species. In fact Linneaus is the one who gave the name Musa paradisiaca to the banana. Being the first Linnean name given to a banana, Musa paradisiaca is technically the "type species" for the genus Musa. Except that the banana Linneaus described is a cultivar which, as specialists later noted, does not represent a species in any reasonable sense of the word. Nevertheless, this name and others that were proposed in its wake, continue to be used to designate cultivars despite the existence of the nomenclature system developed by Simmonds and Shepherd.
It is Ernest Cheesman who noticed that the model for Musa paradisiaca was in fact a type of Plantain. When it was later realised that Musa paradisiaca and Musa sapientum — a Silk type that Linnaeus had also added to the genus – were hybrids between Musa acuminata and Musa balbisiana, some authors began to use the form Musa x paradisiaca and Musa x sapientum to emphasise that fact.
Over the years several authors based the taxonomy of bananas on Musa paradisiaca and Musa sapientum. Sometimes Musa sapientum was treated as a subspecies of Musa paradisiaca, but at other times botanical priority was ignored and Musa paradisiaca was treated as a subspecies of Musa sapientum. Moreover, since Musa paradisiaca is seedless, the subspecies seminifera was created in order to accommodate the wild seeded forms. Giving a seed-bearing wild species the status of subspecies to a seedless cultivar is a good example of the stultifying effect formal nomenclature has had on the crop's taxonomy.
Indeed, as Cheesman noted in 1948, "Some botanists have regarded the seedless forms as ranking with the fertile species and have bestowed Latin binomials upon them. Others have preferred to regard them as varieties of one mythical “ species” (usually called Musa sapientum) which is supposed to exist somewhere in the wild and fertile condition … Such mistakes... are not peculiar to the genus Musa, but they are unusually conspicuous in this group".
It was eventually recognised that most cultivars (except for certain types such as the Fei bananas) are derived from either Musa acuminata alone or hybridized with Musa balbisiana. Some of these cultivars are, like their wild relatives, diploids, i.e. they have two sets of chromosomes (one inherited from each parent). The majority, however, are triploids, i.e they have three sets. This means that at one point, the reproductive cells of one of the parents did not undergo the normal halving of its genome and produced unreduced gametes. The other parent contributed a normal haploid genome.
This complexity made it difficult to devise a Latin name-based taxonomy that could cope with all possible permutations. As Cheesman noted, "the classification of the cultivated varieties is almost a separate problem from the general taxonomy of the genus, needing a different technique for its solution, and confusion of the two makes both almost impossible". He realised that the use of Latin names for cultivars would have to be abandoned. The challenge of coming up with an alternative was taken up by two of his young colleagues, Norman Simmonds and Kenneth Shepherd.