Truth about Musa basjoo (Constantine)

From PlantUse English
Jump to: navigation, search
The truth about Musa basjoo
David R. Constantine

This needs editing!!

This article has originally been uploaded at Globalnet, 2007. Due to its importance, I decided to duplicate it here, although I have been unable to locate the author and obtain permission. Other articles are available here.


Ever since its introduction and throughout western horticulture Musa basjoo has been known as the Japanese or Japanese Fibre Banana. The paper by Turner et al (2002) is the most recent to report that M. basjoo is native to the Ryukyu Islands of southern Japan and to draw attention to its useful fibre. However, Japan has no native Musa and M. basjoo was introduced from China. Further, M. basjoo has no particular utility as a fibre plant. The banana that is cultivated for fibre on the Ryukyus, also introduced, is M. balbisiana. It is time now formally to correct and explain these persistent mistakes in the western literature on M. basjoo.

Historical notes

To set the story of M. basjoo in context it should be noted that the genus Musa got off to a bad start and has a chequered taxonomic history. The type species, Musa paradisiaca, named by Linnaeus in his Species Plantarum of 1752, is not a species at all but a natural, sterile, triploid hybrid derived from two true Musa species, Musa acuminata Colla and M. balbisiana Colla. The intra-specific hybrid nature of Musa paradisiaca L., and many other bananas cultivated for fruit, was not generally recognised until the 1940's. In the intervening two hundred years a large number of Musa "species" were described, the majority of which were cultivars of complex origin. As a result of the confusion the taxonomic literature of Musa, a modest genus with about 55 species currently recognised, is littered with mistakes and more than 250 redundant taxa.

The situation began to change in the 1940's only after a banana breeding programme was started at the Imperial College of Tropical Agriculture (ICTA) in Trinidad. Under the leadership of Prof. E. E. Cheesman, a research programme was instituted to support the breeding work with three major consequences for Musa taxonomy. First, the monopodial Ensete was separated from the sympodial Musa (Cheesman 1947). Secondly, the true nature and complexity of the edible bananas was recognised (Cheesman 1948a). Thirdly, two separate nomenclature systems were adopted formally to name the bananas. The Linnaean binomial system was retained for the wild, seeded Musa species but, unaccountably little used in horticulture except by tropical pomologists, a genome-based system was adopted for parthenocarpic, seedless, edible bananas (Simmonds & Shepherd 1955). In a series of "critical notes on species" dating from 1947, Cheesman gave proper definition to some of the most important species and laid the foundations of modern Musa taxonomy.

For all his insights into the genus, Cheesman was at a major disadvantage. He was working in the New World whereas Musa is exclusively an Old World and specifically an Asian-Pacific genus. All of the plant material that Cheesman and his team worked on had to be imported. With British representatives scattered all over Asia and the Far East this was not difficult to arrange. Nevertheless, there were risks, of which, to be fair, Cheesman was aware, in defining species, often for the first time, based on a limited sampling of material often of less than ideal provenance. There were British representatives in Japan who might have furnished Cheesman with material of the Japanese Fibre Banana. Instead, Cheesman's description of M. basjoo was based on plant material that came, not from the Ryukyu Islands, not even from Japan, but from Devon (Cheesman 1948c). For by the time Cheesman came to review it, M. basjoo was already well known and well established in horticulture, its remarkable cold tolerance having long been recognised.

Musa basjoo was first introduced into cultivation in the west by James Veitch & Sons, the 1863 London offshoot of the great Veitch nursery company founded in Devon in 1808 (Webber 1968). Charles Maries, a foreman at the nursery, collected the plant during a three-year trip to the Far East (1887 - 1889) during which he spent most of his time in Japan. Through collectors such as Maries, Veitch were responsible for the introduction of many significant plants into UK horticulture but they do not seem to have counted M. basjoo among them and information on the plant in their extant records is scanty indeed. James Veitch himself does not mention M. basjoo (despite two reports to the contrary in Gardeners’ Chronicle) in his account of his own travels in Japan from 1891 to 1893 (Veitch 1896). Musa basjoo receives only the briefest of mentions in Hortus Veitchii (Veitch 1906) where the name M. japonica is given as a synonym. J. G. Baker cites the name M. japonica giving the reference Hort. Veitch and clearly implying use of the name by Veitch (Baker 1891). The reference Hort. Veitch must here refer to a catalogue and not Hortus Veitchii but no mention of M. basjoo or M. japonica could be found in James Veitch & Sons’ catalogues in the Lindley Library. There do not appear to be any catalogues of Robert Veitch & Sons extant. In the absence of information direct from Veitch we must turn to secondary sources to try to pinpoint the provenance and timing of Musa basjoo’s introduction.

Maries travelled in Hokkaido and Honshu (Heriz Smith 1992) so we can be sure M. basjoo was introduced to the UK from one of those islands. The French nurseryman Sallier says that it was from the northernmost island of Hokkaido and although he does not state the source of this information Sallier was in contact with Veitch at the time and it may have been direct from them (Sallier 1896). Similarly, a correspondent writing in the Gardeners’ Chronicle of 22nd December 1900 states that the species is “native” to the island of Yesso (modern Hokkaido) in the neighbourhood of Hakkodate.

Maries’ trip lasted from 1887 to 1889 and M. basjoo must have been in the first batch of plants sent home. According to J. G. Baker the plant flowered for the first time in the west in the Temperate House at Kew in the summer of 1890. Baker notes that prior to its flowering at Kew M. basjoo had been grown “for several years in the open air” at James Veitch’s Coombe Wood nursery on Kingston Hill in Surrey. Similarly, an item in Revue Horticole dated 1st November 1889 states that “this year again [M. basjoo] has passed the winter in the open ground at Plessis-Piquet at [the nursery of] Thibaut & Keteleer”. These items strongly suggest 1887 as the date of introduction although 1888 is the date usually given in the literature. Based on its flowering at Kew the plant was illustrated in Curtis’s Botanical Magazine along with a description (Baker 1891).

From the above it seems clear that M. basjoo was introduced to UK horticulture in 1887 or 1888 from material collected by Maries on the island of Hokkaido.

The first record of the plant in the general gardening literature is in the Gardeners’ Chronicle of 15th September 1894. This records a letter received from Canon Ellacombe, rector of Bitton, near Bristol, and an eminent Victorian gardener, who “says [Musa basjoo] dies down in the winter, but the summer growth is 8 ft. in height, whilst each leaf is more than 4 ft. long and nearly 2 ft. broad. It is a grand thing for a sheltered corner. Our first acquaintance with this species was in the nursery of Messrs. Veitch at Combe Wood nr. Kingston” (Anon 1894).

Although it was introduced by James Veitch & Sons material of M. basjoo was sent to the founding Exeter nursery, then known as Robert Veitch & Sons. The first record of the plant in the proceedings of the RHS is mention of a display of plants by Robert Veitch & Sons at the 1899 Cornwall Daffodil & Spring Flower Society Show at Truro. The display, “a collection of rare shrubs suitable for outdoor cultivation in Devon & Cornwall etc.” included M. basjoo and received a sliver-gilt medal (Anon. 1899).

Robert Veitch & Sons are a possible source of the material from Devon sent to Cheesman in Trinidad in the late 1930’s or early 1940’s, James Veitch & Sons having been dissolved in 1914 (Webber 1968). However, St. Bridget Nurseries Ltd. the successors to Robert Veitch & Sons have no record of this (Mrs J. M. Flinn pers. com.). Another possible source in Devon is Otto Overbeck whose remarkable collection of plants is largely intact at Overbecks Garden & Museum, the National Trust property at Salcombe. Overbecks have photographs of the gardens dating to around 1900 including, even then, obviously long established clumps of M. basjoo. Like Canon Ellacombe, Otto Overbeck presumably obtained the plant from Veitch more or less as soon as it became commercially available. Unfortunately, Otto Overbeck’s papers were destroyed on his instructions at his death and the National Trust has no information on this matter (Nick Stewart pers. com.). There is no information in RBG Kew’s Outwards Records regarding the transfer of M. basjoo to Trinidad (Kate Pickard pers. comm.). It seems unlikely but it is possible that Mr Coutts, Curator at Kew, referred to by Cheesman as the provider of the material may have been acting in a private capacity. Whatever the source, with the cultivated material from Devon in front of him, Cheesman based his treatment of the plant on the accounts of P. F. B. von Siebold (1830) and J. G. Baker (1891).

Musa basjoo was first named by von Siebold in his Synopsis Plantarum Oeconomicarum universi regni Japonici of 1830. Philip Franz Balthasar von Siebold was born in Wurzburg, Bavaria in 1796 and after training in the family tradition as a physician joined the Dutch East India Company (VOC) as a surgeon-major in 1822. He was posted first to Batavia where he quickly impressed his superiors through his facility with languages and conscientious attitude to his work. In 1823 he was chosen for an important posting in Japan, still an essentially closed society, where the VOC had its sole trading post on the artificial island of Dejima in Nagasaki harbour on the southern island of Kyushu. During his time in Japan von Siebold avidly collected flora, fauna, ethnography and maps, the latter of which got him into serious trouble. In 1828 von Siebold undertook the traditional journey north for an audience at the Emperor’s court at Edo (modern Tokyo). He amassed such a collection en route that he was forced to send it home to Dejima. When the shipment arrived, the Japanese authorities discovered maps that were considered restricted. Von Siebold was accused of spying for the Russians and expelled from Japan in late 1829. He eventually settled in Leiden and his collections passed to various public institutions there. His substantial herbarium and paintings commissioned for a Flora Japonica are housed at the National Herbarium of the Netherlands. Unfortunately, M. basjoo did not feature in either of these collections and von Siebold’s comments on the plant are restricted to a short paragraph in the Synopsis Plantarum Oeconomicarum compiled and published in 1830 during a short stay in Batavia on his way back to Holland. Cheesman (1948c) quotes von Siebold's Latin text, or at least a fair copy of it he found at Kew, but did not find it necessary to provide a translation, which I now do:

Ex insulis Liukiu introducta, vix asperitati hiemis resistens. Ex folii linteum, praesertium in insulis Liukiu ac quibusdam insulis provinciae Satzuma conficitur. Est haud dubie linteum, quod Philippinis incolis audit Nippis.
"It was introduced from the islands of Ryukyu, and can scarcely withstand the bitterness of the winter. From its leaves is made linen, especially on the islands of Ryukyu and certain islands in the province of Satzuma. It is without doubt linen, of a kind which is called Japanese by the inhabitants of the Philippines"

In 1830, the Ryukyu Islands were not part of Japan and thus the Kyushu-based von Siebold wrote of the plant being introduced from them. When in 1895 the Ryukyu Islands were incorporated into Japan M. basjoo ostensibly became a Japanese species. The province of Satzuma (sic) mentioned by von Siebold is modern Kagoshima, the most southerly prefecture on Kyushu.

Before its incorporation into Japan as the Ryukyu Islands, the Kingdom of Ryukyu, centred on the island of Okinawa, engaged in a flourishing entrepot trade with China, Korea, Japan and southeast Asia and one of the items of trade was banana cloth. The weaving of banana cloth was not and is not unique to the Ryukyus but it is a local speciality and the best cloth is of particularly fine quality. The cloth, basho-fu (banana-cloth), was produced from the banana plant, basho, or more specifically the thread banana, ito-basho. Basho-fu is perfect for hot weather; like linen, it is light and airy and tends not to cling to the skin. In the Ryukyu kingdom, royalty and commoners alike wore basho-fu. As a local speciality, bolts of the finest cloth were given in tribute to the rulers of the various centres with which the Ryukyus traded. Today, basho-fu is a luxury cloth made only in the village of Kijoka, on Okinawa.

Von Siebold clearly thought his plant was the ito-basho used as a source of fibre in the Ryukyu Islands and named it accordingly. The specific epithet basjoo is derived directly from the Japanese basho. Subsequently, influential commentators on Musa, notably Sulpiz Kurz in Calcutta (Kurz 1866, 1877) and J. G. Baker at Kew (Baker 1891), followed von Siebold. Cheesman certainly saw no reason to demure, and his “critical notes” on M. basjoo seem rather uncritical in comparison to his work with other species (Cheesman 1948c). Comments that crop up in Cheesman's diagnoses of other species indicating a certain caution or suggesting that more field work might be necessary are absent from his paper on M. basjoo (Cheesman 1948c). The only doubt that can be detected in Cheesman's paper is his final comment on the plant:

"Though said to be grown in Southern Japan for its fibre, it does not seem to have been investigated as a potential fibre plant elsewhere, nor does our material seem to have any value in that respect."

Cheesman's comments have a particular resonance since throughout the time of the British Empire there was a very active search for commercially useful fibre crops and several bananas had been so investigated; but not M. basjoo.

Cheesman did not himself travel in search of bananas. Although there was a large collection of banana species at ICTA it was incomplete and Cheesman's work was often what might be called forensic taxonomy. He often “discovered” species as it were in the herbarium and literature, sorting out the true species from a mass of synonyms, and not from a working knowledge of plants in the field. In 1954-5, Cheesman's young colleague N. W. Simmonds undertook a major expedition in Asia looking for wild banana species (Simmonds 1956). As a result of Simmonds’ travels a broad picture emerged of the phytogeography of Musa. The genus spreads from the eastern part of the Indian subcontinent in a north easterly arc into southern China, down in parallel strands through Indo-China and the Philippines to Malaysia, the strands uniting in the Indonesian archipelago and continuing eastwards to New Guinea and neighbouring islands and, just, south into north-eastern Australia. As this picture emerged, the Ryukyu Islands’ M. basjoo could be seen as a distinctly odd outlier and it is perhaps surprising that it did not attract a little more attention as a result. In his classic book on the Evolution of the Bananas, Norman Simmonds even presented a map (p. 28) of the geographical distribution of the wild bananas that excluded physical representation of M. basjoo in Japan (Simmonds 1962). In the accompanying text Simmonds included M. basjoo in an area with south China. The lack of attention to M. basjoo can perhaps be explained because the focus of work at ICTA was to understand the origin of edible bananas to facilitate a breeding programme. Although it was used in remote hybridisation work as part of cytological studies (Simmonds 1962, Shepherd 1999), M. basjoo had nothing to contribute to the main effort.

The standard, Western horticultural references (e.g. Moore 1952; Bailey 1957; Simmonds 1962; Argent 1984; Huxley 1992; Mabberley 1997) and more recent works on or mentioning M. basjoo (Turner et al (2002); Pollefys et al (2004)) all essentially follow von Siebold, given the imprimatur of modern authority by Cheesman (1948c). Musa basjoo is indigenous to the Ryukyu Islands of Japan and is the Japanese Fibre Banana. But neither of these “facts” is true and the Japanese have known this for years.


Von Siebold did not provide a description or herbarium material of the plant and so M. basjoo was first published in 1830 nomen nudum. Although not mentioned by Cheesman (1948c), the first valid publication of the name M. basjoo is considered to have been by Yokusai Iinuma in the second edition of his “Illustrations and Descriptions of Plants” (Iinuma 1874). Although described in Japanese and illustrated in Iinuma’s first edition of 1852 it was only in the second edition that the illustration was annotated in Roman text “Basho. Musa basjoo S et Z. (Musaceae)” thus constituting a valid publication of the name. Iinuma’s work was first recognised in the west by Franchet and Savatier in their enumeration of Japanese plants (1879).

Iinuma provides a description of the plant and a pen drawing of the inflorescence, flower and young fruit, the latter two hand coloured in folio. Although somewhat stylised, Iinuma’s drawing is sufficiently recognisable to have been used by Baker (1891) to identify Veitch’s plant. Baker notes that to identify Veitch’s plant “I rely mainly upon [Iinuma’s] figure” and indeed, even if Iinuma’s rather obscure text was accessible to Baker, it does not provide much diagnostically useful information. Moore (1957) also had Iinuma’s text translated but does not seem to have used any of the information in compiling his own description of the plant.

Iinuma does not give any specific uses for M. basjoo but cites Linnaeus [Hayashi] as mentioning “several ways in which the flowers and leaves [of bananas] can be used for medicinal purposes and the fruit for food”. Even by the time of the second edition of his “Illustrations” in 1874 the Ryukyu Islands were still not part of Japan and did not therefore figure in Iinuma’s work. Rather, Iinuma’s work shows us that M. basjoo was a familiar plant on the main islands of Japan.

The first treatment of the flora of the Ryukyus per se came just two years after they were incorporated into Japan. With first-hand knowledge of the plants of the islands, Jinzô Matsumura published a set of notes on some Ryukyu plants in 1897. In this publication, Matsumura formally described just one banana, which he named Musa sapientum var. liu-kiuensis. It is quite clear from Matsumura’s description of this glaucous plant with red, pruinous bracts that it is not M. basjoo. It is also clear that Matsumura knew M. basjoo for his most telling comment is his observation on M. basjoo itself.

Secundum auctores nonnullos Musa basjoo in insula Liu-kiu crescit, sed non vidi.
“According to several authors Musa basjoo thrives on the Ryukyu Islands, but I didn’t see it.”

After Matsumura (loc. cit.), the most notable feature of M. basjoo in the floras of the Ryukyu Islands is that it is absent. Sonohara et al (1952), Hatusima (1971), Hatusima (1975), Walker (1976), Hatusima & Nackejima (1979) do not mention M. basjoo at all. It is only in Hatusima & Amano (1994) that M. basjoo is listed as part of the flora of the Ryukyus, along with 21 other introduced taxa in the Musaceae.

Matsumura (loc. cit.) does not mention any use for M. sapientum var. liu-kiuensis but that it is the ito-basho is confirmed by the continuity of references to the taxon in the literature. In 1934, Matsumura re-classified M. sapientum var. liu-kiuensis as M. textilis var. liukiuensis (Walker, 1976). Musa textilis or abaça is an important fibre banana from the Philippines and the source of Manila hemp, still used today for such diverse uses as marine cordage and tea bags. It might seem reasonable to associate two significant fibre bananas but in truth they are very different plants. To take just one character, with its sulcate, polished bracts M. textilis is a member of section Australimusa (recently absorbed into the Callimusa by Wong et al (2002)) whereas the plant described by Matsumura has the glaucous bracts of section Musa. In 1900 the combination Musa liukiuensis (Matsum.) Makino appears in a paper by Kuroiwa (Kuroiwa 1900). It seems that this can be read as an expression of intent on the part of Makino to elevate the plant to species status but it was another twelve years before he formalised this (Makino 1912).

The true identity of the Japanese Fibre Banana became known only in the aftermath of Japan’s defeat in WWII when the Ryukyu Islands came under the control of the United States Civil Administration of the Ryukyu Islands (USCAR). USCAR brought to Okinawa Egbert H. Walker, a staff member of the Department of Botany, Smithsonian Institution who was in charge of the Serviceman’s Collecting Program (SCP) in which US forces members were encouraged to collect and submit botanical and other specimens. But Walker was no desk-bound administrator of the SCP. He was an accomplished field botanist and developed a thorough knowledge of the Ryukyu flora. His work, during which he personally collected over 7,000 plant specimens on Okinawa and neighbouring islands, culminated in his Flora of Okinawa (Walker 1976). As mentioned above, Walker made no mention of M. basjoo in this work but did include the ito-basho, Musa liukiuensis, after the treatment of which he commented:

“Seeds from plants [of Musa liukiuensis] in Oku [village in northern Okinawa] were grown in Kingston, Jamaica by the Banana Breeding Scheme of the Banana Board. The seed, seedlings and flowers were reported in 1973 to be identical with those of Musa balbisiana Colla.”

One can only speculate why having made such a comment Walker refrained formally from reducing Musa liukiuensis. Hatusima (1975) adds a little to Walker’s comments and does reduce M. liukiuensis:

"Recently, as a result of Walker's agricultural experiments, it has been understood that Ryukyu basho is the same as Musa balbisiana Colla - M. liukiuensis Mak. (1900) syn. nov. fide Walker. Distribution: Java. We have come to understand that Ryukyu basho is the type indigenous to Java named above. It is likely that it was introduced into Ryukyu in the distant past as a result of searching for fibres."

Further confirmation of the synonymy comes from Jarret (1987) who showed that “enzyme polymorphisms at eight enzyme-coding loci failed to reveal any differences between M. balbisiana and M. liukiuensis”.

The identity of the ito-basho, the true Japanese Fibre Banana, is thus firmly established as Musa balbisiana L. A. Colla (Cheesman 1948b).

Having established that M. basjoo is not the Japanese Fibre Banana we turn to a consideration of its true place of origin.

The absence of M. basjoo from the floras of the Ryukyus is telling since a true endemic, and such a physically large herb, would surely have received substantial attention whether it was a useful fibre plant or not. Indeed, it is widely known in Japan that M. basjoo is not native to any part of the country but an introduction from China.

The first formal record of this appears to be by Tomitaro Makino in his New Illustrated Flora of Japan, 1940 [?]. Makino was the first Japanese botanist to work on the Japanese flora using the standard botanical classification system common in the western literature.

Makino’s text on M. basjoo begins matter of factly “a large herbaceous perennial from the temperate regions of China”. More specifically the ideograms used by Makino translated as “temperate regions” (Makino 1948, 1979) indicate a zone in China from 23° 24’ N to 66° 33’ N i.e. all of China north of a line with Guangzhou.

[Need translation of Makino's 1940 & 1979 description here]

The precise origin of M. basjoo in China has continued to intrigue Japanese botanists to this day.

[Need to incorporate Amano et al 1991a and 1991 b here]

It is clear from the Japanese literature that i) the true ito-basho is an introduced banana, Musa balbisiana probably from the Philippines and ii) that M. basjoo is another introduced banana and from China. How then did von Siebold come to make his mistake?

The answer is most probably extremely simple. In Japanese, the same word basho has different meanings depending on context. In a general sense the word means "belonging to the Musa family" or "a banana". In a specific sense the word refers to the banana cultivated for fibre in Okinawa. The ideogram for each is different but the pronunciation is the same and therein lies the possibility for confusion to the western ear. There is no evidence that von Siebold travelled to the Ryukyu Islands, which were not then even part of Japan, or even to the offshore islands of Satsuma. So von Siebold almost certainly did not see the ito-basho in its supposed native place. It is likely that von Siebold did see the banana he named M. basjoo growing as an ornamental in gardens on Kyushu and Honshu and noted that it was called basho by the locals. From his interest in ethnography von Siebold knew that basho was cultivated in the Ryukyu Islands as a source of fibre. It is most likely that he simply assumed that the two basho were one and the same plant (Amano et al 1991b).

The original distribution of most banana species is uncertain and the picture is clouded due to human activity. Nevertheless, it can be said with some confidence that there are no Musa species native to Japan sensu lato; all have been introduced. The ito-basho, Musa balbisiana, was introduced to the Ryukyus as an economically useful fibre plant, and presumably the associated technology to produce cloth was transferred at the same time. But it is not known from where. Hatusima (1975) suggested Java as the source; the weaving of banana cloth is practised there also and certain resist dyeing techniques originating in southeast Asia are used in the Ryukyus (Amanda Stinchecum pers. comm.). Java is a possible source although M. balbisiana is perhaps not wild there, as suggested by Hatusima, but introduced (Simmonds 1962). Geographically closer, the Philippines is another possible source. The Philippines has a long history of cultivation of bananas for fibre and of weaving banana cloth [ ] . Musa balbisiana is native to the Philippines and is associated there in a complex with Musa textilis that is still cultivated today as a source of fibre (Valmayor et al 2002). ‘Canton’ fibre, for example, is derived from a Musa balbisiana x Musa textilis hybrid although it is considered inferior to that of “pure” Musa textilis or abaça for most purposes. Further, it is known that the Philippines exported its banana cloth weaving technology to [Java? and] Oceania (Takuya Nagaoka pers. comm.) and it may similarly have reached Okinawa. However, southern China seems the most likely source. Musa balbisiana is native there, a cloth called basho-fu has been woven there since at least the 3rd century and the yarn-making process used on the Ryukyus, different to that used in the Philippines, is the same as that in southern China (Hendrickx pers. comm.). Southern China is also physically closest to the Ryukyus and had strong trading links with them.

The presence of M. balbisiana on the Ryukyus is well documented. We cannot yet be certain of its origin but the purpose for which Musa balbisiana was introduced to the Ryukyus seems clear; it was introduced as a fibre plant. In sharp contrast there is no evidence that M. basjoo occurred in or was introduced to the Ryukyus in historical times. Absence of evidence is not, of course, evidence of absence but the paucity of evidence that M. basjoo even occurs on the Ryukyus at all is striking. Matsumura (1897) noted that he could not find M. basjoo in the Ryukyus and it is absent from all floras until 1994. Amano et al (1991a) reported that while M. basjoo is widely planted in northern regions of Japan it is much less common in the south. Makino (1979) suggests that M. basjoo does not occur in the subtropical south of China and anecdotal evidence from the USA suggests that M. basjoo does not thrive in the heat and humidity of the southern states preferring cooler conditions further north.

It is now unclear why M. basjoo should have been introduced to the Ryukyus if it is indeed useless for fibre. Might it have been as a fruit plant? This seems unlikely. Musa basjoo does not produce seedless fruit parthenocarpically, it produces ripe fruit only if pollinated and the fruit is then full of seeds. Although it is possible to eat ripe M. basjoo fruit [ ] the palaver of de-seeding it is hardly worth it. Indeed, the lack of esteem in which seedy bananas are held may perhaps be gauged by the Samoans’ name for the seedy fruit of another Musa species, bowdlerised by Balick & Cox as “animal excrement” (Balick & Cox 1996).

It seems to be M. basjoo’s value as an ornamental plant that explains its appearance in the Floras of Japan from an early date. [Valder 1999] mentions the significance of Musa in Japanese gardens … [? More here from Andy ?] and one of Japan’s most famous Haikku poets took his pen name Basho from the plant.

In the light of this it seems reasonable to propose that M. basjoo was first introduced from China in historical times not to the Ryukyus at all but to Japan proper as a cold tolerant ornamental banana. It then spread north and south finding particular favour in the north where most bananas do not succeed. The association of M. basjoo with the Ryukyus is completely spurious.


China is geographically contiguous with a major centre of diversity of Musa in south east Asia and has a native flora currently comprising about 10 species of Musaceae although more await formal description. The Flora of China shows that M. basjoo is rather widespread but the original distribution of the plant has been obscured by human activity. Makino says that M. basjoo is native to the temperate regions of China implying a zone from 23° 24’ N to 66° 33’ N (Makino 1979). Wild populations of M. basjoo are now apparently confined to Sichuan (Liu et al 2002) but it is thought to have had a much wider natural distribution in the past. Without extensive fieldwork it would be futile to speculate on the true origin of the plant in China.

In China as elsewhere M. basjoo has no mainstream economic value. There seem to be no primary reports of it being used there as a fibre plant. There is one report in the literature of M. basjoo being cultivated as a fruit plant in an agroforestry system (Tao & Fang 1991) but in view of its seedy fruits this seems unlikely and is most probably based on misidentification of the plant.

Remarkably, until very recently, even the Chinese literature has been dominated by the hegemony of the notion that M. basjoo is a Japanese plant. For example, [Wu & Kress (2002)] in the Flora of China note that it is native only to Japan & Korea. It is only recently as a result of the writings of Liu Aizhong, citing the work of Amano and colleagues in Japan (Amano et al 1991 a & b), that it has been formally reported in the Chinese literature that M. basjoo is a Chinese native (Liu et al 2002).

The Chinese origin of M. basjoo clears up etymological, phytogeographical and horticultural enigmas. It takes us one step back etymologically since the Chinese ba jiao is the root of the Japanese basho, itself the root of basjoo (Makino 1979). It eliminates the requirement to explain the island origin of the plant. Musa basjoo is now revealed not as an isolated island endemic but as part of a contiguous Musaceae flora in mainland Asia. It helps to explain M. basjoo’s remarkable cold tolerance. Cold tolerance is not an unexpected trait for a plant originating in China and with wild populations in Sichuan whereas it would be very unexpected if the plant had truly originated in the sub-tropical Ryukyu Islands.

Taxonomic notes

Musa basjoo P. F. von Siebold Synopsis Plantarum Oeconomicarum universi regni Japonici. Verhandelingen van het Bataviaasch Genootschap van Kunsten en Wetenschappen [Transactions of the Batavian Society of Arts and Sciences], 12e dl. (1830) ex Y. Iinuma, Sōmoku-Dusetsu [Illustrations and Descriptions of Plants] 3: pl. 1 (1874) et J. G. Baker, Botanical Magazine t. 7182 (1891).

Joseph Gerhard Zuccarini is often credited as co-author of M. basjoo with von Siebold but this is a mistake. Von Siebold did indeed collaborate with Zuccarini on Flora Japonica, the first instalments of which were published in 1835 and on Florae Japonicae Familiae Naturales published in 1846. However, the first mention of M. basjoo was in von Siebold’s Synopsis Plantarum Oeconomicarum of 1830 complied, printed and published, independently of Zuccarini, in Batavia while von Siebold was on his way back to Holland having been expelled from Japan.

Fortunately, von Siebold’s mistaken 1830 identification of the ito-basho as M. basjoo instead of Musa balbisiana need have no taxonomic consequences. We might argue that von Siebold’s banana was the true ito-basho and that, in ignorance of Musa balbisiana Colla, von Siebold assumed it must be a new species and named it M. basjoo. Musa balbisiana dates from 1820 (Cheesman 1948a), comfortably pre-dating M. basjoo. This argument would force us to reduce M. basjoo under M. balbisiana and propose a new name for our familiar “hardy” banana. However, in the absence of any evidence that von Siebold had really seen the true ito-basho this would be stretching speculation too far. Also, there is no trace of confusion in the Japanese literature; M. basjoo and the ito-basho, M. balbisiana, have been treated as distinct entities since 1891 (Anon. 1891, Matsumura 1897). It is reasonable to assume that von Siebold simply did not know that the basho he knew from Kyushu and Honshu was different from the basho he had heard of from the islands of Ryukyu. Von Siebold’s M. basjoo is thus the same plant we know today under that name. This allows us to maintain the taxonomic status quo. It is only the trivial name of M. basjoo, the Japanese Fibre Banana, that is wrong. This name is so ingrained that it will no doubt take years to change but the name Siebold’s Banana is surely more appropriate. Despite his mistaken association of M. basjoo with the ito-basho of the Ryukyus, von Siebold did coin the name and was the first to bring this plant to the attention of the west.

Although there are no taxonomic consequences from the uncovering of von Siebold’s mistake, there are certainly labelling consequences for botanic gardens identifying M. basjoo as the Japanese Fibre Banana, an appellation properly belonging to a strain of M. balbisiana. There are also no doubt consequences for economic botany collections containing Japanese artefacts labelled as being made from M. basjoo which are almost certainly made from M. balbisiana or possibly M. textilis fibre, which was imported into Japan from the Philippines. The literature on banana fibre usually (Dodge 1897, Jarman et al 1997, Turner et al 2002) but not invariably (Kirby 1963) mentions M. basjoo as a source. This literature will also need revision. Reports in that literature of M. basjoo being used as a source of fibre seem all to be based on von Siebold’s mistaken assertion rather than the author’s direct experience. For example, Richardson cites Baker (Baker 1893) as the source of the information that M. basjoo is used for fibre in Japan (Richardson 1957). Richardson thus included M. basjoo fibres in study that mainly concerned abacá (M. textilis). Although Richardson commented that M. basjoo had strong but very fine fibre it was the weakest fibre of those tested with a mean breaking load of 405g as compared to M. textilis at 1,000 to 1,305g. However a relatively weak fibre strength that might disqualify the fibre for industrial use may not preclude its use in textiles; Richardson found that the mean breaking load of M. balbisiana fibre was 410g (Richardson loc. cit.).

It is just possible that on a small scale M. basjoo fibre was used in Japan. There is one tantalising reference in The Useful Plants of Japan Described and Illustrated to fibre being obtained from M. basjoo (Anon. 1891). However, this could be said of any banana and the entry goes on, “in the Okinawa Islands, a different species of Musa grows plentifully and from its fibre the natives weave a cloth called Basho-fu (Musa linen)”.

The relationship of M. basjoo with other banana species requires re-evaluation and further work. The genus Musa is incompletely known and new species are still being described. The genus has long been sorted into five sections (Argent 1976) but it has recently been proposed that these be reduced to three (Wong et al 2002) and that chromosome number rather than plant morphology should define them. Section Ingentimusa (x = 7) has just one species, the world’s tallest herbaceous plant, Musa ingens Simmonds from Papua New Guinea. This is not in cultivation. Section Callimusa now incorporates the Australimusa with a chromosome number based on x = 10. The Callimusa is currently the largest section with about 31 species although Shepherd (1999) has argued that many of the species in the “old” Australimusa should be reduced to subspecific status. Section Musa (x = 11) now includes section Rhodochlamys (Wong et al loc. cit.) and has about 23 species including M. basjoo.

The relationship of M. basjoo with other species in section Musa is not well understood and reports are contradictory. Based on work at ICTA, Simmonds (1962) keyed out M. basjoo and M. itinerans together and commented that while they crossed with difficulty they nevertheless appeared closely allied and rather distant from the rest of section Musa. However, in a numerical taxonomic analysis Simmonds & Weatherup (1991) found that M. basjoo clustered, but not closely, with a group that included M. acuminata, M. flaviflora, M. itinerans, M. sikkimensis and M. schizocarpa. In this study, M. itinerans clustered most closely to M. sikkimensis whereas M. basjoo was rather remote. Unfortunately, M. basjoo was not included in recent AFLP studies by Wong et al (2001, 2002) that are revealing new taxonomic alliances within Musa. The understanding that M. basjoo is part of the mainstream and not an isolated island endemic makes it all the more important that it should be included in further studies. Simmonds’ (1962) comment that M. basjoo and M. itinerans are allopatric can now be discounted, as they are sympatric in China. This raises the intriguing prospect, taxonomically as well as horticulturally, of natural hybrids from China.

For a recent description of M. basjoo the reader is referred to Turner et al (2002) although with reservations since the basis of their paper is that M. basjoo is the Japanese Fibre Banana. The fine painting by Mieko Ishikawa shows a much greener inflorescence than is commonly seen in plants growing in the UK, which have bracts brownish-yellow externally and more yellow internally. The contrast with the illustration in Baker 1891 is striking in this respect. On the other hand, Ishikawa’s drawing of the inflorescence in Turner et al (2002) hardly looks like the same species with its strangely elongated male bud and exaggerated imbrication. Such morphological and colour differences may reflect growing conditions or possibly intra-species variation.

There is reportedly one botanical variety of M. basjoo from Taiwan, M. basjoo var. formosana (Warb.) S. S. Ying (Ying 1985). The variety seems to differ from M. basjoo in having purple-red bracts as opposed to brownish-yellow and purple-red ripe fruit rather than yellow. It is not yet in cultivation.

The published synonymy of M. basjoo is short but confused and based on mistakes rather than formal taxonomic revision.

Musa japonica

1. Musa japonica Hort. Thiébaut & Keteleer ex Revue Horticole 61 : 491-492 (1889).

J. G. Baker (1891) indicated that the name Musa japonica was initially used by Veitch but it has not proved possible to find a mention of the taxon in their publications until Hortus Veitchii of 1906 where it is listed as a synonym of M. basjoo. Musa japonica is thus attributed to C. Thiébaut & J. B. Keteleer, nurserymen at Plessis-Piquet (= Plessis-Robinson) near Paris, who were responsible for the first introduction of the plant into France from James Veitch. The anonymous item in Revue Horticole is dated 1st November 1889 and states that “this year again [M. japonica] has passed the winter in the open ground at Plessi-Piquet”. Thiébaut & Keteleer must have obtained the plant from Veitch almost as soon as it was received from Japan. Although Thiébaut & Keteleer introduced the plant, its initial propagation and distribution in France was due to J. Sallier who took over Thiébaut & Keteleer’s business, and augmented their material with fresh supplies from Veitch (Sallier 1896).

Baker (loc. cit.) formally reduced Musa japonica under M. basjoo but the French obviously had a problem with the admittedly odd-sounding epithet. “Is this the name the natives call the plant in Japan?” wondered an incredulous Sallier in Revue Horticole (Sallier loc. cit.). Unable to countenance the adoption of such a clumsy epithet the name Musa japonica was conserved for a period in France because it was “simple” and had “the advantage of indicating the country of origin” (Sallier loc. cit.). De Wildeman (1913) noted disapprovingly that Thiébaut & Keteleer had abandoned the laws of priority in coining the name Musa japonica and he too reduced it under M. basjoo and so the name passed out of use in French horticulture.

Musa martinii

2. Musa martinii A. Van Geert, Revue d'Horticulture Belge et Étrangère 18: 107, fig. 12 (1892).

This puzzling taxon, initially published as Musa Martini, is normally attributed to Auguste Van Geert, the Belgian nurseryman who was most probably the author “A.” of the 1892 article in the magazine he published. In 1893, Elie-Abel Carrière, the French horticulturist and editor of Revue Horticole wrote a brief note on M. Martini that made no mention Van Geert (professional rivalry perhaps?) but instead quoted a description of the plant taken from the catalogue of a M. Lille of Lyon. M. Lille’s description is almost word-for-word a copy of Van Geert.

Van Geert describes the plant that he received as seedlings from Tenerife as having a reddish stem and midribs, leaves glaucous above and green below with a thick lamina, so that they were resistant to tearing in the wind, and bearing pretty, pink flowers. He also mentions that M. Martini is more cold tolerant than Musa Ensete (= Ensete ventricosum (Welw.) Cheesman). The accompanying illustration is too poor to give any clue to the identity of the plant but from Van Geert’s brief description it is obviously not M. basjoo nor does Van Geert even mention this taxon in his discussion. Nevertheless, Musa Martini has been associated with M. basjoo following Baker (1893) who concludes his note on M. basjoo as follows:

"[Musa basjoo] is as hardy as M. Ensete, and is grown in Southern Japan for its fibre. M. Martini [ ] has the habit of M. sapientum, and is said to be more hardy than M. Ensete, with bright rose-red flowers. The leaves are oblong, long-petioled, firm in texture, bright green above, glaucous beneath, with reddish veins. It was brought from the Canary Islands."

Baker inverts the leaf characters described by Van Geert, which were probably anyway a printer’s error inverting dessous and dessus. Baker's comments are distinctly odd because while they are appended to his note on M. basjoo they seem to have nothing to do with it but are rather the continuation of a train of thought on hardiness relative to “M. Ensete”. This may have arisen simply because of Van Geert’s comment but, coincidentally, when M. basjoo was first flowered in the Temperate House at Kew in 1890 it was grown “side by side with Musa ensete” (Baker 1891). The association of M. basjoo and M. Martini established in Baker’s 1893 text was repeated without alteration in subsequent papers on the bananas emanating from Kew (Anon. 1894, Anon. 1906).

Later authors were certainly confused by Baker. De Wildeman (1912) tentatively reduced M. Martini under M. basjoo. Fawcett (1939) seemed to think there had been a printer's error whereby Musa Martini, instead of being listed as a separate entity, was absorbed into the text of M. basjoo. Fawcett thus inserted M. Martini between M. basjoo and M. textilis in his treatment of the genus and did not associate it with M. basjoo. In his treatment of M. basjoo, Cheesman (1948) referred to Baker’s 1893 paper but, probably wisely, ignored Musa Martini.

Musa Martini Van Geert is the same plant as that described in the RHS Dictionary (1956) as M. Martinii but here it is definitely linked with M. basjoo and described as "like M. basjoo but taller and [flowers] rose-red. Canary Islands. [Introduced to U.K.] 1892". If it was cultivated in the UK it seems to have disappeared, as it did in Europe (de Wildeman 1912).

Moore (1956) recommended that the name Musa martinii Van Geert should be rejected under provisions of the ICBN and it would appear that its association with M. basjoo is quite spurious. Ironically, and to compound the fictitious relationship between them, it is possible that M. martinii rather than being a form of Musa basjoo was actually Musa balbisiana. Van Geert’s description is consistent with this suggestion and Cheesman (1948) received seed labelled Musa martinii, which grew into Musa balbisiana at ICTA.

To add to the confusion, the name M. Martini was published for a second time in Revue Horticole by Raphaël de Noter in 1895. De Noter’s plant is an apparently solitary banana from Vietnam with a description suggestive of an Ensete rather than a Musa. In apparent ignorance of Van Geert (1892) and Carrière (1893), de Noter named his plant after M. J. Martin, then Director of the Botanical Garden in Hanoi. Whatever it is, this taxon has nothing to do with M. basjoo.

Musa dechangensis, Musa lushanensis, Musa luteola

3. Musa dechangensis J. L. Liu & M. G. Liu, Acta Botanica Yunnanica 9 (2) : 163 (1987). 4. Musa lushanensis J. L. Liu, Acta Botanica Yunnanica. 11 (2) : 171 (1989). 5. Musa luteola J. L. Liu, Investigatio et Studium Naturae 10 : 41, f. 1 (1990).

These three taxa can be treated together. From 1987 to 1990, Liu J. L. published descriptions of three new Chinese Musa from Yunnan. In 1997, Wu Te-Lin reduced all of these under M. balbisiana. Most recently Liu Aizhong et al (2002) have re-evaluated the herbarium material of these species and transferred them all under M. basjoo. This situation is difficult to explain. One might well get the impression from this and from preceding discussion that it is easy to confuse Musa balbisiana and M. basjoo. In reality, they are very distinct in their vegetative, floral and seed characters. If nothing else, this episode is perhaps suggestive of interesting variation in M. basjoo in China.


Musa basjoo has been in continuous cultivation in the UK and mainland Europe since its introduction in 1887 or 1888 and propagated vegetatively by separation of suckers derived from leaf-opposed lateral buds; the Musaceae do not have axillary buds. Nearly all the M. basjoo plants offered for sale today are micropropagated, a process that essentially accelerates natural suckering through exposure of shoots to synthetic cytokinin in vitro. In highlighting the supposed relationship between M. basjoo and M. itinerans, Turner et al (2002) follow Simmonds (1962) and comment that both have long, stoloniferous rhizomes. This statement cannot readily be reconciled with horticultural experience as can be seen, for example, by examining the several clumps of M. basjoo growing at RBG Kew; Musa basjoo is a relatively close suckering species. In contrast, M. itinerans is said to produce suckers 2m away from the parent stem (Cheesman 1949). Cheesman (1948c) commented that M. basjoo is shy suckering and this can be seen in old established plants such as at Overbecks Garden at Sharpitor, Devon that have only ever been subject to conventional propagation from suckers. On the other hand, M. basjoo produced from micropropagation tend to sucker profusely, a probable carryover effect of the synthetic cytokinin used in the micropropagation process.

Musa basjoo produces viable seed in Japan where the winter temperature is no lower than -3.5°C (Amano), and presumably also in China, but so far seed has not been reported on plants in cultivation in the UK. The absence of seed has confused some people but has a simple explanation. In a typical banana inflorescence, the basal flowers are functionally female and the apical flowers functionally male; emphasis is placed on function rather than structure since all Musa flowers have male and female elements. There are a few Musa species that have functionally hermaphrodite basal flowers but M. basjoo is not one of these. By the time male flowers are produced on a M. basjoo inflorescence, the females are no longer receptive. An isolated plant producing a single inflorescence will not therefore set seed. However, where two nearby plants or two stems on the same “mat” flower asynchronously so that the first to flower can pollinate the second, there is no reason why viable seed should not be produced. Although the temporal separation of female and male flowers is designed to ensure outcrossing there is no self-incompatibility mechanism in Musa. However, in the U.K. another limiting factor is that a long, warm autumn would be required to develop and ripen the fruit. This has so far precluded seed production even in otherwise favourable microclimates such as at Overbecks where there are several flowering clumps of M. basjoo. Recently, there have been reports of viable M. basjoo seed being produced in the U.S. (Wagner pers. com. 2003) and there must be large areas of that country with a climate suitable for production of seed.

Seed of more and more banana species is becoming available commercially but for some reason seed of M. basjoo has not so far been offered, or, at least, not under that name. Chinese seed offered by Carl Sandeman in 2000 as Musa wilsonii turned out to be M. basjoo but with very low viability and seed of that provenance has not been offered again. With the increased interest in the Musaceae as ornamentals it is presumably only a matter of time before M. basjoo seed from China, or Japan, becomes routinely available.

Because of the lack of seed and the fact that it has been vegetatively propagated for so long there has been little variation in the M. basjoo offered for sale in Europe. Most plants represent the Veitch clone, more or less unaltered since its introduction. It is possible that some clonal variation has occurred through sporting, to which bananas are especially liable, or the accumulation of systemic pathogens. Anecdotally there do appear to be “good” and “poor” strains of M. basjoo in cultivation, but this requires further investigation. It is also possible that there were other introductions of M. basjoo by nurseries such as Vanhoutte, Verschaffelt & Donckelaer who were active in importing plants from China and Japan in the latter part of the 19th century. Although there are no contemporary reports of such introductions in Revue d'Horticulture Belge et Étrangère or Revue Horticole this also requires further investigation.

Different forms of M. basjoo are now becoming available. The upsurge of interest in plants for tropical effect has encouraged seedsmen and nurserymen to look for “new” material and several named forms of M. basjoo are now available. The most active nurseryman in this respect is Jean-Luc Penninckx of Enghien (Edingen) in Belgium and some of his material is available in the U.K.

1. Musa basjoo 'Fuji-Yama'.

Jean-Luc Penninckx uses this cultivar name merely to distinguish the clone from other selections he offers. The plant has no connection with Mt. Fuji-Yama and in fact is the same as any undifferentiated M. basjoo available in Europe. Penninckx obtained starter material for this clone, probably the Veitch clone, from the nursery Xotus in Delft, Holland. In the context of Penninckx’s own collection there is a need to differentiate the clone but there is no reason why the cultivar name 'Fuji-Yama' should be widely applied.

2. Musa basjoo x hybridum.

This plant is featured at It is an invalid name applied to a form of M. basjoo said by Tropicaflore to be hardy to - 12ºC to - 15ºC as opposed to "normal" M. basjoo which, in contrast, is said to be hardy only to - 10ºC.

3. Musa basjoo ex Sandeman Seeds’ M. wilsonii.

In 2000, the author raised a single plant from seed labelled as Musa wilsonii from the seedsman Carl Sandeman (re-packaged by Norfolk Tropicals) and subsequently built up a small clone by separation of suckers. In 2003, the author has been given a sucker of a second plant of M. basjoo raised from the same seedlot by Mrs D. Newell in Norfolk. Although it has not yet flowered this material is obviously M. basjoo. When a sucker was supplied by the author to Jean-Luc Penninckx he named it ‘Wilsonii’ and noted that it had somewhat salmon-coloured pseudostems in comparison to other material of M. basjoo in his collection. The application of the name 'Wilsonii' perpetuates the mistaken association of the name with this material and is likely to cause confusion in the trade. For no rational reason, banana enthusiasts tend to treat Musa wilsonii as if it was the holy grail of hardy bananas. There is no evidence for this and in fact, M. wilsonii probably doesn’t exist at all as a distinct taxon.

Musa wilsonii was published by Tutcher (1902) from material grown in Hong Kong from seed collected by E. H. Wilson in Yunnan. Tutcher commented that it appeared to be closely allied to Musa glauca a taxon that was transferred to the genus Ensete as E. glaucum by Cheesman (1947). The illustration that accompanies Tutcher’s description shows an extraordinary plant with a typical Ensete inflorescence, but bolt upright, a character that no known Ensete has. Musa wilsonii is also a confused entity in the Chinese literature and has been applied to herbarium specimens recently identified as Musa itinerans (Liu 2001). Following Cheesman (1947) but contrary to Simmonds (1960), Wu & Kress include E. wilsonii in the Flora of China as a Yunnan endemic. However, no botanist has seen this plant other than as herbarium specimens in China and it seems probable that it is no more than a high altitude provenance of E. glaucum. If it is to be used at all, the cultivar name ‘Wilsonii’ should be reserved for a good, cold tolerant provenance of Ensete glaucum from high altitude in Yunnan. Such material is not yet in cultivation.

Material of this seed-raised Musa basjoo is available to interested parties. It does not appear to be any hardier than “normal” M. basjoo but because of its salmon-coloured pseudostem it can at least be distinguished visually from other forms of M. basjoo, which is more than can be said for next plant.

4. Musa basjoo 'Sakhalin'.

This name was first applied to clonal material by Jean-Luc Penninckx imported by him in the spring of 1996, if not directly, from Sakhalin Island on the Pacific coast of Russia. This selection has been available for some time in the UK and an amount of "urban myth" has built up around it that needs to be cleared away.

The Russian territory of Sakhalin Island is 3,000 km from the Ryukyu Islands and is the northernmost large island of the archipelago otherwise constituting Japan. The belief that M. basjoo is native to the Ryukyu Islands has led many erroneously to assume that a cline must exist the length of Japanese archipelago and that 'Sakhalin' must surely represent the most northerly and therefore the most cold tolerant provenance of M. basjoo. The knowledge that M. basjoo is Chinese not Japanese undermines the logic that inevitably equates 'Sakhalin' with the ultimate in banana hardiness. The plant was introduced to Sakhalin too although whether from Japan or direct from China is not known. The mode of ‘Sakhalin’’s introduction to cultivation in Europe is also claimed by some to be mysterious. Simply put, Jean-Luc Penninckx was responsible for its introduction, from a Russian contact, and for its initial commercialisation in Europe.

Penninckx makes no great claims for ‘Sakhalin’s hardiness compared with the Veitch clone of M. basjoo. It does seem to re-grow with greater vigour in spring but the foliage of ‘Sakhalin’ is no more cold tolerant than the Veitch clone, being burned off by a degree or so of air frost. Whether or not it is significantly hardier than the Veitch clone, 'Sakhalin' does appear to be a somewhat more robust plant, stockier and with thicker leaves and petiole wings. Compared side-by-side and en masse with the Veitch clone of M. basjoo on Jean-Luc Penninckx’s nursery, 'Sakhalin' is darker green in colour. Using AFLP analysis 'Sakhalin' can be distinguished from the Veitch clone of M. basjoo (Engelborghs pers. com. 2002). However, there is no unambiguous feature or set of features by which to determine with certainty whether an isolated plant one encounters is 'Sakhalin' or "normal" M. basjoo. The lack of a set of clear, visual distinguishing features has already caused confusion in the trade. In the circumstances it is doubtful whether 'Sakhalin' deserves cultivar status; it may be uniform and stable but it is insufficiently horticulturally distinct.

5. Musa basjoo 'Saporro'.

Sometimes erroneously referred to as Musa saporro this plant was featured for a time at and is still listed at It was imported by Dr Michael Lorek from Saporro, Hokkaido, and was said to survive -19ºC "with standard protection methods". Rather in the manner of 'Sakhalin', the belief that M. basjoo is native to the Ryukyu Islands of Japan led Dr. Lorek to assume that 'Saporro' represented the most northerly native provenance of M. basjoo in Japan and must therefore be exceptionally cold tolerant. That Musa basjoo occurs at all in Hokkaido is due to human intervention since it is well north of the region of Japan where M. basjoo can reproduce itself by seed. Ironically, it is probable that Maries collected M. basjoo from Hakkodate in Hokkaido in the first place.

6. Musa basjoo 'Chechnya'.

This name or ‘Tchechenia’ is applied by Jean-Luc Penninckx to a clone he obtained from a contact in the war-torn Russian republic. Despite its exotic name this selection is doubtfully distinct from the Veitch clone of M. basjoo found in the trade in Europe.

7. Musa basjoo 'Variegata'.

This is the oldest and most enigmatic of the forms of M. basjoo to be found in the literature.

This plant was mentioned in the 1956 RHS Dictionary and has been included in recent books on plants for tropical effect (e.g. Riffle 2000). The M. basjoo 'Variegata' reported in 1956 has long since disappeared from cultivation. But it may be about to make a comeback. When plants are micropropagated it is not unusual to find some aberrant plants. Bananas are notoriously prone to such abnormalities that are sometimes manifest as variegation. On the nursery of Jos van der Palen in Holland in 1999 a variegated plant appeared in a batch of micropropagated M. basjoo. This plant has maintained its variegation so far, which is by no means certain as most variegated bananas from micropropagation quickly revert to green. However, one isolated plant does not constitute a cultivar and it remains to be seen whether van der Palen's plant is sufficiently stable to be propagated even on a small scale. In the light of this, Jean-Luc Penninckx’s naming of van der Palen's plant M. basjoo 'Fuji-Yama Variegata' is somewhat premature. Even if van der Palen's plant cannot be propagated, all may not be lost, sporadic variegated plants have continued to turn up in batches of micropropagated plants and perhaps one of these will be sufficiently stable to reconstitute the “lost” cultivar.

Claims about plant hardiness are notoriously anecdotal and the relative hardiness of these named forms of M. basjoo is not properly known. Unless and until these forms are collected together in one environment and compared side by side it is not possible to draw any conclusions as to the comparative hardiness claimed by the suppliers. No doubt there is variation in M. basjoo but the regrettable tendency for nurserymen to give cultivar names to material of doubtful provenance exaggerates the variation available. Only two available forms seem at all distinct form the Veitch clone, 'Sakhalin' and the plants derived from Carl Sandeman’s seed. Despite their modest distinctions they at least open our eyes to the fact that we have hitherto sampled but a tiny fraction of the germplasm of M. basjoo. Of course, any search for new germplasm should focus on China and not on Japan. There may be other exciting horticultural introductions of the plant, especially from Sichuan and Hebei. The availability of seed of M. basjoo from China, surely an imminent prospect, should also facilitate the development of interesting new clones.

Acknowledgements: Special thanks are due to Christine Smithee (U.K.) and Rob Wagner (U.S.A.). I would like to thank the following at RBG Kew; John Flanagan, Head of Library & Archives and his unfailingly helpful library staff; Mrs Kate Pickard, Archivist; Julia Steele, Economic Botany Collection. Thanks are similarly due to Brent Elliot and his staff at the RHS Lindley Library and to Izumi Tytler, Bodleian Japanese Library, Oxford. Thanks are also due to the following individuals: Takuya Nagaoka (New Zealand), Katrien Hendrickx (Okinawa), R. A. Dorc (U.K.), Iris Engelborghs (Belgium), J. M. Flinn (U.K.), Hidenobu Funakoshi (Japan), Shirley Heriz-Smith (U.K.), W. J. Kress (U.S.A.), Aizhong Liu (China), Jean-Luc Penninckx (Belgium), Nick Stewart (U.K.), Amanda Stinchecum (U.S.A.), Genta Tanaka (Japan), S. S. Ying (Taiwan), J. van Dooren (Belgium). Translations from Latin were by and from the Japanese by Andrew Blair-Smith. Lastly, I should like to thank Tim Berners-Lee for the Internet that facilitated communications without which this paper would simply never have been conceived.


  • Amano, M., Y. Sawada, T. Motohashi, M. Miyata, T. Yoshizawa, and S. Masuda. (1991-1992a). Cultivation area of Musa basjoo Sieb. et Zucc. Japan. J. Agric. Sci. Tokyo Nogyo Daigaku 36: 58 - 65.
  • Amano, M., Y. Sawada, T. Motohashi, M. Miyata, T. Yoshizawa, and S. Masuda. (1991-1992b). A consideration to the original home of Musa basjoo Sieb. et Zucc. J. Agric. Sci. Tokyo Nogyo Daigaku. 36: 66 - 70.
  • Anon. (1891). Musa basjoo Sieb. In: Useful Plants of Japan Described and Illustrated. Agricultural Society of Japan, Tokyo. p. 93. no. 337. (English text published 1895).
  • Anon. (1894). Musa basjoo. Gardeners Chronicle. 16 (403): 316.
  • Anon. (1894). Species and Principal Varieties of Musa. Bulletin of Miscellaneous Information. No. 92. Royal Gardens, Kew, London.
  • Anon. (1900). Musa japonica syn. M. basjoo. Gardeners’ Chronicle. 28 : 456.
  • Anon. (1906). Species and Principal Varieties of Musa. Kew Bulletin, Additional Series VI. II. Royal Gardens, Kew, London.
  • Argent, G. C. G. (1976). The Wild Bananas of Papua New Guinea. Notes Roy. Bot. Gard. Edinb. 35: 77 – 114.
  • Argent, G. C. G. (1984). Musaceae. In: Walters, S. M. et al (Eds.) The European Garden Flora. II. Cambridge University Press.
  • Bailey, L. H. (1937). The Standard Cyclopedia of Horticulture. Macmillan, New York.
  • Baker, J. G. (1891). Musa basjoo - Native of the Liu-Kiu Archipelago. Bot. Mag. t. 7182.
  • Baker, J. G. (1893). A synopsis of the genera and species of Museae. Ann. Bot. 7:189-229.
  • Balick, M. J. & Cox, P.A. (1996). Plants, People, and Culture – The Science of Ethnobotany. Scientific American Library. New York.
  • Carrière, E.-A. (1893). Musa Martini. Rev. Hort. : 53 – 54.
  • Cheesman, E. E. (1947). Classification of the Bananas. I. The Genus Ensete Horan. Kew Bulletin. 2 (2): 97 – 106.
  • Cheesman, E. E. (1948a). Classification of the Bananas. III. Critical Notes on Species. c. Musa paradisiaca and Musa sapientum. Kew Bulletin. 3 (2): 145 – 153.

Cheesman, E. E. (1948b). Classification of the Bananas. III. Critical Notes on Species. a. Musa balbisiana. Kew Bulletin. 3 (1): 11 - 17.

  • Cheesman, E. E. (1948c). Classification of the Bananas. III. Critical Notes on Species. e. Musa basjoo. Kew Bulletin. 3 (3): 323 – 325.
  • Cheesman, E. E. (1949). Classification of the Bananas. III. Critical Notes on Species. g. Musa itinerans. Kew Bulletin. 4 (1): 23 – 24.
  • Fawcett, W. (1913). The Banana. Its cultivation, distribution and commercial uses. Duckworth & Co., London.
  • Franchet, A. & Savatier, L. (1879). Enumeratio Plantarum in Japonia Sponte Crescentium. 2: 20-21.
  • van Geert, A. (1892). Revue d'Horticulture Belge et Étrangère 18: 107, fig. 12.
  • Hatusima, S. (1971). The Flora of the Ryukyus (Ryukyu shokubutsu-shi). Study Group of Biological Education of Okinawa. Okinawa, Japan.
  • Hatusima, S. (1975). Flora of the Ryukyus, Okinawa Society of Biological Education and Research, Naha, Okinawa, Japan. NOT SEEN [from Amanda Stichecum]
  • Hatusima, S. & Nackejima, C. (1979). Flowers of the Ryukyu Islands. Kodansha, Tokyo.
  • Hatusima, S, & Amano T. (1994). Flora of the Ryukyus south of Amani Island. The Biological Society of Okinawa. Okinawa, Japan.
  • Heriz-Smith, S. (1992). James Veitch & Sons, Chelsea: Harry Veitch’s Reign, 1870 – 1890. Garden History. 20 (1) 57 – 70.
  • Huxley, A. (1992). The New Royal Horticultural Society Dictionary of Gardening. Macmillan, London.
  • Iinuma, Y. (1852). Somoku zusetsu [Illustrations and Descriptions of Plants]. III: pl. 1.
  • Iinuma, Y. (1874). Somoku zusetsu [Illustrations and Descriptions of Plants]. 2nd. Ed. III: pl. 1.
  • Jarret, R. L. (1987). Biochemical/Genetic Markers and Their Uses in the Genus Musa. In: Persley G. J. and De Langhe E. A. (1987). Banana and Plantain Breeding Strategies; proceedings of an international workshop held at Cairns, Australia, 13-17 October, 1986. Australian Centre for International Agricultural Research Proceedings No. 21, 187p.
  • Kuroiwa, H. (1900). A list of Phanerogams collected in the Southern Part of Isl. Okinawa one of the Loochoo Chain. Bot. Mag., Tokyo, 14: 141.
  • Kurz, S. (1866). Note on the Plantains of the Indian Archipelago. J. Ag. Hort. Soc. Ind. 14: 295 – 301.
  • Kurz, S. (1877). The Banana : a Pomological contribution. J. Ag. Hort. Soc. Ind. [ ]: 112 – 168.
  • Liu, A-Z et al 2002
  • Mabberley, D. J. (1997). The Plant-Book. Cambridge University Press.
  • Makino, T. (1912). Observations on the Flora of Japan. Bot. Mag., Tokyo 26: 180 - 182.
  • Makino, T. (1912).
  • Makino, T. and Nemoto, K. (1931). Flora of Japan [Nippon-Shokubutsu-Soran]. 2nd ed. Shunyodo Shoten, Tokyo.
  • Matsumura, J. (1897). Notes on some Liu-kiu plants. Bot. Mag., Tokyo 11: [69]
  • Moore, H. E. (1952). Musa and Ensete – The Cultivated Bananas. Baileya. 5: 167 – 194.
  • de Noter, R. (1895). Le Musa Martini. Rev. Hort. : 290 – 291.
  • Pollefeys, P., Sharrock S. and Arnaud E. (2004) Preliminary analysis of the literature on the distribution of wild Musa species using MGIS and DIVA-GIS. INIBAP. Montpelier
  • Sallier, J. (1896). Un Bananier Rustique – Musa Japonica. Revue Horticole. 68: 202-204.
  • Shepherd, K. (1999). Cytogenetics of the genus Musa. INIBAP. Montpellier. France.
  • Siebold, P. F. (1830). Synopsis Plantarum Oeconomicarum universi regni Japonici, Bataviaasch Genootschap van Kunsten en Wetenschapen, Batavia, Java.
  • Simmonds, N. W. (1956). Botanical results of the Banana Collecting Expedition, 1954 – 5. Kew Bulletin. 11 (3): 463 – 489.
  • Simmonds, N. W. (1962). The Evolution of the Bananas. Longmans. London.
  • Simmonds, N. W. & Shepherd, K. (1955). The taxonomy and origins of the cultivated bananas. J. Linn. Soc. (Bot.) 55: 302 – 312.
  • Simmonds, N. W. & Weatherup, S. T. C., (1990). Numerical taxonomy of the wild bananas (Musa). New Phytologist 115 (3): 567 - 571.
  • Sonohara, S., Tawada, S. & Amano, T. (1952). Flora of Okinawa. USCAR. Okinawa, Japan.
  • Stinchecum Amanda Mayer, “Textiles of Okinawa,” in William Jay Rathbun, ed., Beyond the Tanabata. Bridge: Traditional Japanese Textiles (Hong Kong and Seattle: Thames and Hudson, 1993), 83.
  • Tao A. & Fang Z. (1991). The Characteristics and Benefits of Forest Network Farmland in the Zhujiang River Delta. In: Agroforestry Systems in China, Editors Zhu Zhaohua et al Eds.) Published Jointly by: Chinese Academy of Forestry, People's Republic of China and International Development Research Centre. Canada.
  • Turner 2002
  • Ying, S. S. (1985). Miscellaneous Notes on the Flora of Taiwan (I). Mem. Coll. Agric. Natl. Taiwan Univ. 25 (2): 98 – 106.
  • Valmayor, R. V., Espino, R. R. C. & Pascua, O. C. (2002). The Wild and Cultivated Bananas of the Philippines. PAARFI. Los Baños, Laguna, Philippines.
  • Veitch, J. H. (1896). A Traveller’s Notes. Notes of a tour through India, Malaysia, Japan, Corea, the Australian Colonies and New Zealand during the years 1891 – 1893. J. Veitch & Sons, Chelsea.
  • Veitch, J. H. (1906). Hortus Veitchii. A history of the rise and progress of the nurseries of Messrs. James Veitch and Sons, together with an account of the botanical collectors and hybridists employed by them. J. Veitch & Sons, Chelsea.
  • Walker, E. H. (1976). Flora of Okinawa and the southern Ryukyu Islands. Smithsonian Institution Press. Washington DC.
  • Webber, R. (1968). The Early Horticulturists. David & Charles.
  • Wu Delin & W. John Kress. 2000. Musaceae in Flora of China 24: 297–313. Missouri Botanical Garden Press.