The ancestral type of taro

Conflicting views on the original distribution of taro, and its long history of cultivation, have resulted in different views on which (if any) of the current wild forms correspond with the original species. It has often been assumed that var. antiquorum most closely resembles the ancestral stock, with var. esculenta representing a long-term selection for larger corms. However, Ivancic and Lebot (1999) identified dasheen (var. esculenta) type plants as probably indigenous to New Caledonia. Somewhat similar wild plants found in Nepal and Yunnan (China) by Yoshino (2002) were thought to be hybrids arising from interspecific or intergeneric hybridisation within Colocasia or between Colocasia and Alocasia.

Yoshino (2002) suggested that Colocasia esculenta var. aquatilis was probably a descendent of the pre-cultivation species. Vinning (2003) suggested that Colocasia esculenta var. aquatilis should be referred to as ‘wild type’ Colocasia esculenta. Colocasia esculenta is considered native in the Northern Territory of Australia (Cowie and Albrecht 2005), and these stoloniferous plants would be placed in var. aquatilis in an infraspecific classification.

Ivancic and Lebot (2000) considered that there was a ‘true’ wild type (e.g. in the Solomon Islands) characterised by extremely long stolons, fast leaf regeneration, continuous growth, small elongated corms and, usually, high calcium oxalate concentrations. The corms of these plants are almost never eaten, but the leaves may be. In addition these authors recognised wild taro in the Solomon Islands and Papua New Guinea with some characteristics of cultivated types. They also recognised wild taro that they thought may have originated as hybrids between cultivated and true wild type taro, and plants that were obvious and hardly-altered escapes from cultivation.

A study by Kreike et al. (2004) using AFLP DNA fingerprinting techniques found there were two distinct gene pools in diploid taro, one found in the Pacific plus Thailand, and the other found only in SE Asia. Two major groups of triploids were found, each internally diverse, suggesting that triploids had arisen at least twice, and that this was an old, rather than recent, event. They found no connection between ploidy and dasheen/eddoe type of corm.

Classification of taro

There is no clear consensus on the classification of taro. Names are applied in three main contexts: botanical (taxonomic), agronomic, and cultivar, and there is only limited congruence between the three.

Formal taxonomic names

The formal taxonomy of this group is very complex and the subject of continued scientific debate. The taxonomy of the genus as a whole has not been formally revised for at least 30 years. The standard reference for taxonomic treatment of the group is Purseglove (1972), although Hay (1996) provided a synopsis of Colocasia for Australia and Papua New Guinea, in which he discussed variation within Colocasia esculenta. Traditionally, taxonomists have recognised two main taxa, but there is considerable difference of opinion on the ranks to be applied to them. The two taxa are:

Synonyms: Colocasia esculenta var. typica A.F. Hill; Colocasia antiquorum var. esculenta (L.) Schott ex Engl.

This is the plant commonly known as large corm taro, or dasheen type. The variety is characterised by its large cylindrical central corm, lacking hairs, and with few cormels. The central corm provides the main crop.

Colocasia esculenta var. antiquorum (Schott) Hubbard & Rehd

Synonyms: Colocasia antiquorum Schott; Colocasia esculentum subsp. antiquorum Haudricourt; Colocasia esculenta var. globulifera (Engl. & K. Krause) Young; Colocasia antiquorum var. typica Engl.

This is the plant commonly known as small corm taro, or eddoe type. The variety is characterised by having a small, more or less globular, central corm, bearing long shaggy hairs in the upper part, with several relatively large cormels arising laterally. The lateral cormels provide the main crop.

Important works recognising these varieties include Purseglove (1972), Onwueme (1999), Bown (2000), Ivancic and Lebot (2000), Chay-Prove and Goebel (2004), Kreike et al. (2004), James et al. (2006) and Orchard (2007). A few maintain the above two taxa as distinct species (Colocasia esculenta and Colocasia antiquorum) rather than varieties of Colocasia esculenta (e.g. Onyilagha et al. 1987).

In addition to the above two taxa, others are sometimes recognised, including the following:

Colocasia esculenta var. aquatilis Hassk.

Synonyms: Colocasia esculenta ‘Aquatilis’; Colocasia antiquorum var. aquatilis (Hassk.) Hassk. ex Engl.

This name is applied to wild types of Colocasia esculenta that are found in SW and SE Asia, China, the Ryukyu Islands, Australia and some oceanic islands (and as introduced in USA – see NRCS (2009)).

The variety is characterised by having small, poorly developed, acrid, often inedible corms, long lateral stolons, and an absence of daughter corms. It is mainly used for pig food (Yoshino 2002). Exceptionally, the stolons are pickled as a delicacy in Yunnan (Xu et al. 2001).

Works accepting this taxon include Bailey and Bailey (1976), NRCS (2009), Xu et al. (2001), and Orchard (2007).

Colocasia esculenta var. nymphaeifolia (Vent.) A.F. Hill

Synonyms: Caladium nymphaeifolium Vent.; Arum nymphaeifolium (Vent.) Bull; Colocasia nymphaeifolia (Vent.) Kunth (sometimes incorrectly spelled nymphiifolia).

This variety is found in Japan, under the cultivar name Hasubaimo, and has been introduced to the USA (NRCS 2009). It is characterised by an absence of bending between petiole and leaf blade.

Works accepting this taxon include Hirai et al. (1989) and NRCS (2009).

Synonyms: Colocasia esculenta ‘Euchlora’; Colocasia euchlora K.Koch & Linden; Colocasia antiquorum var. euchlora (K.Koch & Linden) Engl.

This variety is characterised by deep green leaves with purple margins and petioles, but is now usually only recognised as distinct in gardening books.

In some recent taxonomic works, including Floras and handbooks, there is a tendency to ignore formal varieties, and treat taro as a unitary species with the name Colocasia esculenta (e.g. Smith 1979; Hay 1990; Nasir 1978, Thompson 2000, and Hay 1996). This position was supported by Vinning (2003), quoting the results of the TANSAO Project. The recognition of varieties is, however, supported by others, including Orchard (2007) and others cited above.

For the purposes of this report the following taxa are recognised: Colocasia esculenta (L.) Schott var. esculenta, Colocasia esculenta var. antiquorum (Schott) Hubbard & Rehd. and Colocasia esculenta var. aquatilis Hassk.

Agronomic classifications

Of agronomic papers that distinguish different types of taro, almost all use the distinction of dasheen (large corm) and eddoe (small corm) types, rather than the botanical varieties. Most do not define these types, but the consensus seems to be as follows:

Dasheen type (large corm taro) (Figure D.1)

Dasheen is the main type grown in Africa, the West Indies, southern Asia and Oceania (Purseglove 1972). The corm is large and cylindrical, up to 30 cm in length and 15 cm in diameter, lacking hairs, with few cormels, and only very rarely, stolons. Sprouting arises from the petiole base, and if this is excised, the remaining corm will usually not sprout, i.e. the corm has a terminal meristem. The central corm is harvested as the main crop, the lateral cormels being discarded or used as planting stock. A portion of the spadix bearing male flower protrudes above the neck of the spathe, and has a very short sterile appendage free of the spathe.

Eddoe type (small corm taro) (Figure D.2)

The eddoe type is the favoured taro in northern Asia, including Japan and China. It is also grown in the West Indies and, to a limited extent, in southern USA (Purseglove 1972). The central corm is small, globoid, and surrounded by cormels (stem tubers) and daughter corms. It has long shaggy hairs in the upper part. If the apex of the corm is removed, the remaining corm will sprout readily from lateral buds, i.e. the corm has copious lateral meristems. The lateral cormels and corms constitute the main crop as well as the new planting stock: the central corm is generally considered inedible. The sterile appendage of the spadix is long, three times the length of that in the dasheen type, and is retained within the inrolled tip of the spathe.

Figure D.2: A typical eddoe type small corm taro

Although Figures D.1 and D.2 depict typical large and small taro corms, it should be noted that due to the large number of cultivars recorded, commercially cultivated corms are likely to vary in size and shape, in addition to colour, flavour and texture.

Cutting across the botanical and agronomic classifications is a broad classification of many locally recognised cultivars. Some local guides to these cultivars have been produced, as for example, those for 50 cultivars in peninsular Malaya (Ghani 1984), those of Japan (Hirai et al. 1989; Matsuda 2002), Japan and Taiwan (Tanimoto 1990), those of E and SE Asia and Oceania (Yoshino 2002; Lebot et al. 2004), and those of Hawaii (Whitney et al. 1939; James et al. 2006). Cho et al. (2007) discussed the approximately 200 cultivars that may once have existed in Hawaii, compared with only about 75 surviving in official living collections there today.

These informal classifications, at least the most recent ones, tend not to recognise the traditional division between var. esculenta and var. antiquorum, but instead name informal groups of morphologically and/or genetically similar cultivars. Thus, Ghani (1984) recognised 50 cultivars identified by local names, in four groups and several subgroups in peninsular Malaya. Hirai et al. (1989) recognised 80 cultivars from Japan, divided into nine major groups. Xu et al. (2001) studied 42 samples (with local names) representing 20 cultivars in five major morphotypes in China. Rodríguez Manzano et al. (2001) grouped 42 Cuban clones into six informal groups and several subgroups. Matsuda (2002) recognised 80 cultivars in eleven main cultivar groups in Japan. Lebot et al. (2004) worked on 2298 accessions representing, inter alia, at least 378 cultivars from Vanuatu, 54 from Indonesia, 15 from Malaysia, 19 from the Philippines, 35 from Thailand, and 29 from Vietnam.

Maemouri (2003) reported on a collection of 824 taro accessions representing 594 landraces from just four out of nine provinces of the Solomon Islands. She noted that there was a constant turnover of cultivars in the Solomon Islands, with new forms being continually selected and trialled. In Hawaii, MacCaughey and Emerson (1913, 1914) suggested that there might once have been 150–175 distinct named cultivars of taro in the islands, of which many are now lost. Whitney et al. (1939) recognised 84 distinct types of taro in Hawaii, of which 69 were derived from local plants. James et al. (2006) provided a key to 97 cultivars of taro in the Hawaiian Islands. Of these Hawaiian cultivars, 95 were assigned to var. esculenta, one (Tsuronoko (Araimo)) to var. antiquorum, and one (Zuiki) to Colocasia gigantea.

A number of Colocasia esculenta cultivars are grown as ornamental horticultural subjects under the common name Elephant Ears. There are many of these, but some of the better known ones (Christman 2006) are:

• 
  Colocasia esculenta ‘Black Magic’ – dark purple leaves
  
• 
  Colocasia esculenta ‘Jet Black Gold’ – dark purple leaves
  
• 
  Colocasia esculenta ‘Jet Black Wonder’ – dark purple leaves with light coloured veins
  
• 
  Colocasia esculenta ‘Portadora’ – a vigorous form with ruffled leaves
  
• 
  Colocasia esculenta ‘Hilo’ – a dwarf form with variegated leaves
  
• 
  Colocasia esculenta ‘Fontanesia’ (syn. Colocasia esculenta var. fontanesii (Schott ex Engl.) A.F.Hill) – with violet stems and leaves with wine red veins, margins and petioles
  
• 
  Colocasia esculenta ‘Illustris’ (syn. Colocasia esculenta var. illustris (Bull) A.F.Hill) – purple markings between the leaf veins
  

Wherever taro is grown there are a multitude of locally recognised cultivars, distinguished by the colour of the tuber flesh (white, pink or red), the colour of the leaf lamina and veins (shades of green, with or without a purple spot on the upper surface above the insertion of the petiole), the colour of the petiole (various shades of green, pinkish purple, almost black, or streaked), and by taste (acridity) of the tubers and leaves. There are many hundreds of cultivars, or perhaps even 15 000 (Ivancic and Lebot 2000), and their nomenclature is complex and inconsistent, with the same or very similar cultivars being grown under different names in different countries. For example, Purseglove (1972) suggested that ‘Common Eddoe’ of the West Indies was the same as ‘Trinidad’ in the USA.

In the Pacific some of the cultivars grown commercially are:

• 
  Akado – Hawaii (Valenzuela and Sato 2004)
  
• 
  Dirratengadik – Palau (Brooks 2000b)
  
• 
  Lehua Maoli (Maui Lehua) (for poi) – Hawaii (Trujillo et al. 2002; Hamasaki et al. 2000; Valenzuela and Sato 2004)
  
• 
  Manu’a – American Samoa (Navarro et al. 1986)
  
• 
  Meltalt – Palau (Brooks 2000b)
  
• 
  Miyako – Hawaii (Valenzuela and Sato 2004)
  
• 
  Ngeruuch – Palau (Trujillo et al. 2002)
  
• 
  Palau – Samoa (Aregheore and Perera 2003)
  
• 
  Pa'lehua, Pa'akala, and Pauakea – Hawaii (Trujillo et al. 2002)
  
• 
  Pink Samoan – Fiji (Daniells et al. 2004)
  
• 
  Pwetepwet – Samoa (Aregheore and Perera 2003)
  
• 
  Rota (Antiguo) – Northern Marianas (Brooks 2000b)
  
• 
  Samoa and Samoa hybrid – Samoa (Jackson et al. 2001)
  
• 
  Taro Niue (Tau(s)ala ni Samoa) – Fiji (Daniells et al. 2004); Hawaii (Valenzuela and Sato 2004); American Samoa (Navarro et al. 1986); Samoa (Jackson et al. 2001)
  
• 
  Toakulu – Samoa (Jackson et al. 2001)
  
• 
  Tsurunoko – Hawaii (Valenzuela and Sato 2004)
  

The ‘piko’ group of taro in Hawaii are unusual in that they have hastate leaves (lamina attached by their margin to the petiole) rather than peltate leaves (petiole joins the lamina near the middle) (Onwueme 1999). In Hawaii, 27 main groups of taro cultivars were recognised in 1880. By 1935 only eight main groups were recognised, and many cultivars within each group had been lost (Cho et al. 2007).

In Japan the numerous minor cultivars have been divided into nine main groups on morphological and storage protein characters by Hirai et al. (1989): Eguimo, Dodare, Ishikawa-wase, Tonoimo, Akame, Migashiki, Binroshin, Yatsugashira and Takenokoimo. These cultivar groups included both dasheen and eddoe types. Matsuda (2002) recognised eleven cultivar groups for Japan, based on restriction fragment length polymorphism (RFLP) in rDNA. Most of these corresponded with those of Hirai et al. (1989), but some were different.

In China Xu et al. (2001) grouped 20 traditional cultivars from Yunnan according to ethnobotanical, agro-morphological and genetic characteristics into six major morphotypes:

• 
  Inflorescence morphotype (sometimes referred to as Colocasia tonoimo and grown for its edible inflorescence) – Kaihuayu or kaihua-yutou (Yoshino 2002)
  
• 
  Single-corm morphotype (more or less equivalent to dasheen type) – Daziyu, Huanggengyu, Lugengdayutou, Shuiyu, Pobule, Byongmanizong
  
• 
  Multicormel morphotype (equivalent, in part, to eddoe type) – Bulena, Bulece, Bulene, Byongmasongho, Byongmane, Byongmazabyong, Byongmaayo, Bigeana, Bikuobine, Achichibiu, Byongmaayopiu, Qingu, Zhongyu
  
• 
  Multicorm morphotype (equivalent, in part, to eddoe type) – Gouzhuayu, Mianhuayu, Mibiu
  
• 
  Petiole morphotype (grown for its edible petioles, corms are of poor quality) – Caiyu
  
• 
  Stolon morphotype (more or less equivalent to Colocasia esculenta var. aquatilis) – Wangenyu, Yingjiangyeyu
  

Most taro produced in Australia is broadly categorised as Taro Pacific, the dasheen type, which is also imported from Fiji and Tonga (Vinning 2003). Within this category, the most common cultivars are:

• 
  Bun-long or Pan Long Wu, a soft cooking type believed to originate in China (Bown 2000) with white flesh and purple flecks, favoured particularly by the Asian community (Chay-Prove and Goebel 2004; Daniells et al. 2004; Horsburgh and Noller 2005)
  
• 
  Taro Nuie or Tausala ni Samoa, a pink taro sourced from Samoa, favoured particularly by Pacific Islanders (Chay-Prove and Goebel 2004; Daniells et al. 2004)
  
• 
  Samoan Pink, a white fleshed variety with pink flecks and firm texture from Samoa , favoured particularly by Pacific Islanders (Chay-Prove and Goebel 2004; Daniells et al. 2004)
  
• 
  Taro Vietnam, a small dasheen type (clear skinned, no hairs) grown mainly as a backyard crop in Queensland and sometimes marketed through Vietnamese grocery shops (Vinning 2003).
  

Since about 1998 there has been increasing interest in cultivation of Taro Supreme, the eddoe type of taro, tentatively identified with the Japanese cultivar Ishikawa wase (Vinning 2003). The local cultivar of this form is mainly that known as NORADA 1 (Midmore et al. 2006). Taro Supreme is cultivated particularly in the Northern Rivers district of New South Wales, with smaller plantations in northern Queensland and south-west of Darwin (Vinning 2003; Hicks and Nguyen 2004). Trials are also being undertaken in Western Australia. Processed Taro Supreme is imported into Australia from China as frozen, peeled or brined products, either alone or as mixed vegetables (Vinning 2003). Import of fresh eddoe type taro is currently prohibited (Public Quarantine Alert 0482, 23 May 2006), as detailed in the background to this pest risk analysis.

Taro weed potential

Taro propagation, both in cultivation and in the wild, is predominantly by vegetative means (Purseglove 1972; Onwueme 1999). Flowering and seeding in cultivars is rare, with the notable exception of one cultivar (known as Colocasia esculenta ‘Inflorescence morphotype’, Kaihuayu, kaihua-yutou or Colocasia tonoimo) in Yunnan, China, which is grown predominantly for its edible inflorescences (Xu et al. 2001; Yoshino 2002). However, Kreike et al. (2004) noted that flowering and seeding was common in wild type Colocasia esculenta in Indonesia and Papua New Guinea.

The lateral corms of the dasheen type of taro can be removed as propagating material, and in the wild they may give rise to their own crop of daughter corms. A number of wild-type dasheen taro can also form long stolons and propagate and spread from these (Ivancic and Lebot 1999). One selection in Yunnan is cultivated for these stolons (Colocasia esculenta ‘Stolon morphotype’, Wangenyu, Yingjiangyeyu) (Xu et al. 2001).

The eddoe type of taro has a smaller central corm and produces many more (up to 200) daughter corms laterally. These daughter corms are all capable of growing if detached, and each will potentially produce a large number of daughter corms of their own in the next season.