Colocasia Esculenta C. Formosana

Matthews, Nguyen, Tandang, Agoo and Madulid, 2015 Taxonomy and ethnobotany of Colocasia...
Taxonomy and ethnobotany of Colocasia

esculenta and C. formosana (Araceae):
implications for the evolution, natural range, and
domestication of taro
Peter J. Matthews
National Museum of Ethnology
Osaka, Japan
Nguyen, Van Dzu,

Institute for Ecology and Biological Resources
Hanoi, Vietnam
Daniel Tandang,
National Museum of the Philippines,
Manila, Philippines
E. Maribel Agoo,
De La Salle University,
Taft, Manila, Philippines
Domingo A. Madulid,
De La Salle University,
Taft, Manila, Philippines
ABSTRACT vegetable for human and animal

A critical problem for the taxonomy of consumption (as food and fodder). The
taro (C. esculenta), and for understanding the example of C. formosana Hayata is
evolution and domestication of this species, introduced here because our observations
is that there is no way to recognise, by so far indicate that this is a naturally-
simple visual inspection, a wild population distributed wild species throughout its
of taro as part of a natural distribution. known range, despite its close phenotypic
This is because people throughout similarity to C. esculenta. To learn about the
Southeast Asia have long used wild taro as a evolution, natural range, and domestication
Aroideana VOL 38E NO 1, 2015 153

Figure 1 The “Water Kelady” (Caladium aquatile) illustrated by Rumphius 1741–50 (2011) in
The Ambonese Herbal. Hasskarl (1848) referred to this work to establish the name C. esculenta
Schott var. aquatilis (Rumph). The parent plant is flowering, and a new shoot has sprouted
from the stolon at left.

of taro, closer study of C. formosana is ‘water kelady’ by Rumphius) (Figure 1).

recommended. The ambiguity of its status as a wild plant
seen growing in and around villages, and
along rivers, is already apparent in the
INTRODUCTION naming and description provided by
The distribution, use, ecology, and genetics Rumphius.
of wild forms of taro, Colocasia esculenta (L.)
Schott, have been explored over many years Formal descriptions of natural botanical
in an attempt to discover the genetic and varieties or sub-species should, in principle,
geographical origins of cultivated taro, a be based on wildtypes present in naturally-
starchy root crop that is found in tropical to distributed wild populations. According to
temperate regions of the world (Matthews, the international codes of nomenclature for
2014). The current species name, C. esculenta wild and cultivated plants (Spencer et al.,
(L.) Schott, is generally understood to refer 2007), a natural sub-species is ‘generally
to a single highly-polymorphic species, with understood as having defining
several varieties, but historically there has characteristics that are usually
been great uncertainty about definition of geographically separated, although they may
the varieties and their taxonomic status as occupy different ecological niches’, while a
botanical varieties or distinct species ‘variety’ is ‘often understood as having
(Plucknett, 1983; Orchard 2006). The characters that differ in a minor way’ in
original descriptions for the species, and for plants that ‘do not have a clearly defined
two commonly recognised varieties, var. geographical or ecological distribution’. A
esculenta and var. antiquorum Hubbard & critical problem for the taxonomy of C.
Rehder (1932) are all based on cultivated esculenta, and for understanding the
plants, while a third variety, var. aquatilis evolution and domestication of this species,
Hasskarl (1848) is based on an illustration is that there is no way to recognise, by
and written description by Rumphius 1741– simple visual inspection, a wild population
50 (2011) of an Indonesian wildtype or of taro as part of a natural distribution. We
possibly naturalised (feral) wild taro (Hay, can assume neither that a particular patch
1998). As a result, the polymorphism of wild taro represents a natural variety or
apparent in previous taxonomic descriptions sub-species, nor that it is derived from
reflects variation produced by human cultivated plants. This is because people
selection during domestication, as well as throughout Southeast Asia have long used
variation produced during natural evolution wild taro as a vegetable for human and
of the species. Rumphius gave two animal consumption (as food and fodder).
alternative Latin names for the same plant, The example of C. formosana Hayata 1919 is
based on where it could be seen growing: introduced here because our observations
Caladium vicorum (‘taro of the villages’), and so far indicate that this is a naturally-
Caladium aquatile (‘aquatic taro’; also called distributed wild species throughout its

known range, despite its close phenotypic taro was most commonly known as a
similarity to C. esculenta. For understanding vegetatively propagated crop, and even
the evolution and domestication of C. today, it is still a surprise for many people to
esculenta, there may be much to learn from learn that wild breeding populations exist in
closer study of C. formosana. many countries. In Queensland, Australia,
inflorescences and fruit with seeds were
Previous theories of natural range and seen in a large wild taro patch, in a region
the origins of cultivated taro where taro is rarely cultivated. Many leaf
samples were collected in transects through
Wild breeding populations of taro are the patch, the leaves were frozen in liquid
distributed in tropical and subtropical nitrogen, carried to a laboratory for DNA
regions from South Asia to East Asia, and was extraction, and the extracts were
through Southeast Asia to northern preserved in freezers from 1987 to the
Australia and Papua New Guinea. In all present day. Now it is known that taro
these regions, there are many local uses for DNA can be more easily sampled by simply
wild taro, which has served as a food source drying the leaves in bags with silica gel. The
(corms, stolons, leaves, inflorescences, and - 1987 samples were recently sent to the
though rarely reported - fruit), medicinal McDonald Institute for Archaeological
plant, and as a fodder for domestic pig Research in the United Kingdom for
(Matthews et al., 2012; Masuno et al., 2012). microsatellite DNA analysis, and our results
Since wild forms of taro are useful plants, provided the first genetic evidence of
and are known from field interviews to be breeding in a wild taro population, while
transplanted, it is not easy to distinguish also indicating the predominance of one
naturally-distributed wild taro populations genetic clone in the taro patch studied
from populations that may arise directly or (Hunt et al., 2013).
indirectly through human activities.
In Japan, and other regions of Northeast
Following a survey of records of taro in Asia, taro is cultivated at high latitudes that
botanical literature and European herbarium experience cold winters with snow. Under
collections, and a survey of wild and such conditions, farmers must give special
apparently breeding populations of taro in care to ensure the survival of living plants
northern Australia and Papua New Guinea from autumn until the next spring, despite
(Matthews, 1987; 1990), maps were the fact that the plants appear adapted to
published showing the global distribution the cool and seasonal environment. The
of taro in cultivation, the likely natural northern cultivars typically produce many
range of wild taro, and the distribution of side-corms that can be easily stored during
two better-known wild species (C. affinis, C. winter, using storage methods that vary
fallax) (Matthews, 1991; 2006). Previously, according to latitude and altitude (Matthews,
2002). The parent corms and child side-

Figure 2. The possible restricted natural range of taro (C. esculenta), representing a single area
of origin for cultivated taro in Southeast Asia. Diversity in the morphology of temperate
and tropical cultivars is indicated outside the map boundary. In this scheme, wild populations
in northern Australia and Papua New Guinea may have been derived from wild or cultivated
forms introduced by people. The large area outlined, encompassing the region from India to
northern Australia, is the maximum likely natural range (see F3) (from Matthews, 1990;
2014).

Figure 3. The maximum likely natural range for taro (C. esculenta), in Asia and the Pacific, and
the possible multiple origins of cultivated forms throughout this range. In this scheme, wild
taro in northern Australia and Papua New Guinea arrived first through natural dispersal.
Diversity in the morphology of cultivated taro is indicated outside the map boundary. The
possible area of natural origin for the species is unchanged from F2 (from Matthews, 1990;
2014).

corms become dormant during winter, and scattered records of wild taro in botanical
carry many buds for regrowth in spring. literature and herbarium collections, for
Even with cold damage, and the death of these and other regions (Matthews, 1991).
some buds or shoots, the northern cultivars
will sprout again in spring, and are easy to In the first scheme (Figure 2), the
maintain. selection and dispersal of taro by humans
began with starchy wildtypes that evolved
The observations in tropical Australia and and dispersed naturally within a range
Papua New Guinea, and in temperate Japan, restricted to mainland Southeast Asia. This
led to suggestion that there had been two scheme followed the suggestion by Hotta
main sources and directions for the (1983) (see also Hutterer, 1983), that root
domestication of taro (Matthews, 1990; crops developed where species hibernated
2014): (i) cool-adapted wild forms of taro naturally in response to either seasonal dry
the Himalayan mountains (somewhere in periods or seasonal cold periods. It was thus
the region from northeast India to China) assumed that starch was a target for the
leading to the temperate adapted cultivars initial selection, use, dispersal, and
of taro (mainly triploids), and (ii) warm- cultivation of wild forms. In the second
adapted forms of taro in a tropical lowlands scheme (Figure 3), selection and dispersal
area (somewhere in the region from India to of taro by humans began with non-starchy
northern Australia and Papua New Guinea) wildtypes of wide natural occurrence, in
leading to tropical cultivars. tropical and warm temperate zones,
respectively. In this case, it was suggested
Since little was known about the uses and that starchiness increased in multiple areas
dispersal of wild forms of taro, it was of domestication, after breeding and
considered possible that the natural range selection in early farming systems, and
of taro: (i) is restricted to a region of specifically in early swidden systems or
mainland Southeast Asia where mountain nursery plantings (Matthews, 1995). Such
and lowland environments are present, systems and plantings provide obvious
within the vicinity of other wild Colocasia opportunities for breeding among cultivar
species known at that time (Figure 2), or assemblages. Fallow gardens within the
(ii) extends throughout the region from swidden cycle often include remnant plants
India to Southeast Asia, northern Australia that can be later reincorporated into active
and Papua New Guinea (Figure 3). gardens, and during the fallow period, taro
may breed, out of sight of the farmer.
In these figures, the maximum likely
natural range of taro was identified In the many years since these two
according to the observations in northern contrasting theories were proposed, new
Australia and Papua New Guinea, and evidence has emerged concerning the
economic uses of wild taro, the diversity

and distribution of wild species of Colocasia, been studied in detail under controlled
the ecological requirements for wild conditions.
breeding populations of taro, and genetic
relationships among Colocasia species. Although the corms of wild taros are
generally considered by local people to be
Below, we highlight some of the recent more acrid than those of cultivars,
evidence, leaving genetic data aside, and differences in acridity have also not been
recommend a focus on commensal wild studied systematically. Acridity and starch
populations and naturalisation in order to production in wild taros are likely to vary
define natural range limits, and learn about according to both genotype and
the evolution and domestication of C. environmental conditions. Such variation
esculenta. may also be reflected in the different local
uses of wild taros. Wild taro corms have
The uses of wild taro (C. esculenta) been recorded historically as a useful but
minor food, or as a famine food, in
Among cultivated taros, two main northern Australia (Scarlett, 1985;
morphotypes are recognised: var. esculenta, Matthews, 2014), Myanmar (Matthews &
with a large edible main corm and few Naing, 2005), and southern Japan
cormels, and var. antiquorum, with a small or (Matthews et al., 1992), but not yet in Papua
medium-sized main corm and a large New Guinea (Matthews, 2014). In Australia
number of small cormels (Plucknett, 1983). and Papua New Guinea, the leaves and
A wide range of intermediate forms is also stolons of wild taro have not been reported
known, as well as forms in which stolons as edible, but these parts are commonly
appear to have become shortened and eaten in China (Yang & Xu, 2000),
enlarged. In contrast to the great diversity Myanmar (Matthews & Naing, 2005),
of cultivated forms, the vegetative Philippines (Matthews et al., 2012), and
morphology of wild taros in warm regions Vietnam. Although the use of taro stolons
of Southeast Asia to Australia and Papua is widespread, little has been reported
New Guinea is relatively uniform (Ivancic & regarding their variation, production,
Lebot, 2000; Matthews, 1991;1997; selection, utilisation, and nutritional value as
Matthews & Naing, 2005; Matthews et al., vegetables.
2012). Most wild taros do not display
swelling of the corm relative to the leaf- In southern Japan (Matthews et al., 1992),
base diameter at the corm apex (thus Myanmar (Matthews & Naing, 2005), the
leading to elongate shapes in larger plants), Philippines (Matthews et al., 2012), and
and the corms are often fibrous and watery northern Vietnam (Matsuda & Nawata,
(less-starchy). However, such differences in 2002, Masuno et al., 2012), wild taro leaves
corm dimensions and starchiness have not are commonly cooked with other
ingredients as a fodder for pigs. Use of wild

taro (Figure 1) as ‘a plain potherb for slaves the main factor that determines whether or
and common people’, and as pig fodder, not wild taros in a particular location or
was reported in eastern Indonesia in the region are used and transplanted.
18th century (Rumphius, 2011). Since pig Informants very often express concern
husbandry is ancient in Southeast Asia, it is about acridity when using wild taro corms,
likely that the use of wild taro as a fodder leaves, or stolons, and give various reasons
for pigs is an ancient and widespread for the occasional experience of acridity
practice. even after wild taro parts have been cooked
(e.g., Matthews et al., 2012). Other less
In the Philippines, wild taro was previously obvious anti-nutritional factors may also
reported as absent or derived from have contributed to whether or not wild
cultivated populations. Wild taro today is taro plants were used in the past, and
common in warmer and wetter regions of eventually domesticated (Matthews, 2010).
the archipelago, and is used extensively and
intensively as leaf vegetable for human The widespread use of wild taro as a
consumption (Matthews et al., 2012). This vegetable for human consumption, and as a
may largely explain the general absence of fodder for pigs, was not known when the
flowering during our surveys, since cutting two models of domestication (above) were
the plant tops prevents inflorescences from first proposed. The main difference
developing. A wild breeding population has between the two models is whether or not
been found in only one area, in wild taros were widespread in Southeast
northwestern Luzon, where wild plants are Asia and the western Pacific before
mainly used as a source of edible stolons. domestication, regardless of the manner of
These can be harvested without obviously dispersal (natural, human transfer, or both).
interfering with the flowering cycle (there It is now clear that people can and do
may be non-obvious physiological transplant both wild and domesticated taro
consequences). into wild or disturbed habitats, for
economic purposes, with or without interest
In Australia and Papua New Guinea, the in the use of corms for starch. At the same
use of wild taro leaves as a vegetable for time, it is also clear that wild taros in
human consumption has not been reported. tropical Asia and the Pacific can easily
This may explain why flowering, fruiting spread by natural means, through the
and seed production can be easily observed production of sweet fruit that are attractive
in wild taro patches (Matthews, 1990; 1995; to birds and other animals.
Price et al., 2008; Hunt et al., 2013).
For human consumption especially, the

risk of encountering strong acridity may be

Figure 4. Colocasia formosana Hayata from roadside at edge of forest, Mt Polis, Ifugao,
Philippines, in 2011. The leaves display typical rounded shape with a relatively shallow sinus.
Note fruiting heads at left, and young stolon emerging at right. The overall green colour and
indeterminate stolon growth of C. formosana are also common in wild populations of C.
esculenta (L.) Schott.

Figure 5. Fruit of wild Colocasia formosana Hayata, in Taiwan, 2008. The fruit of this species
can be used as a bait to catch birds in simple spring-noose snares (Yuasa, 2000).
Hutterer (1983) and Matthews (1996) their natural range, and then become
emphasized that the common distinction naturalised. Nevertheless, in regions where
between ‘wild’ and ‘cultivated’ is not the wild breeding populations are present, an
same as the distinction between a wildtype individual plant found in a wild breeding
(natural genotype) and domesticate population can be regarded as a possible
(genetically modified through human wildtype, and the likelihood of this being
selection). Cultivars can be wildtypes, and true increases when local people regard the
domesticates can enter wild habitats and plant as wild, or naturally-occurring, or as
become naturalised. Pollen and seed may inedible, and if they give the wild plant a
disperse more-or-less freely between the different name from the cultivated forms
different habitats. Wildtype plants can also they are familiar with.
be introduced to geographical areas outside

In our fieldwork in Southeast Asia and the forests), and are native to the Himalayan
Pacific, we have focused on collecting region, though their distributions may range
possible wildtype taros, and recording further (Bown, 2000). The following list of
information on their uses, in order to: wild Colocasia species (not complete) is
illustrative for the present discussion. More
(i) Compare genetic variation in different than half the known species of Colocasia
wild taro populations, have been described within the last 20 years.
(ii) Compare genetic variation in wild and 1. C. affinis Schott 1859, wild in Southeast
cultivated taros, Asia (including Northeast India), lower
altitudes.
(iii) Look for correlations between usage
and genotype among wild taros. 2. C. boyceana Gogoi and Borah 2013, wild in
Arunachal Pradesh, Northeast India, at
Wild populations that are not used, or are altitudes of 1200–1600 m.
little-used, and that are genetically distinct
from cultivated taros, are more likely to be 3. C. debangensis Gogoi and Borah 2013, wild
naturally-occurring wild populations. in Arunachal Pradesh, Northeast India, at
Combining ethnobotanical and genetic data altitudes of 1800–2200 m.
to characterise different wild taro
populations, may eventually make it possible 4. C. fallax, Schott 1859, wild in Northeast
to distinguish naturally-distributed India.
populations and those derived from
cultivars through introduction and 5. C. formosana Hayata 1919, abundant and
naturalisation. widespread in Taiwan, where it forms a
morphologically homogeneous wild
Wild (non-cultivated) species of population and is not known as a
Colocasia are more diverse and domesticate; it is morphologically distinct,
widespread than previously thought but minimally so, from wild C. esculenta.
The genus Colocasia Schott is now believed 6. C. gaoligongensis H. Li & C.-L. Long 1999,
to contain at least twelve and perhaps many wild in high mountains, Yunnan, at 3,700 m
more distinct species, all of which are found asl.
in humid to semi-aquatic habitats in
Southeast Asia to southern China. The 7. C. lihengiae Long & Liu 2001 wild in
closely related genera Ariopsis, Steudnera and mountains, northern Vietnam to Yunnan,
Remusatia include mainly shade-loving China. The known range of this species was
species (in the understorey of tropical

recently extended to Arunachal Pradesh in environments, in contrast to the cooler,

Northeast India (Gogoi & Borah, 2013). shadier, drier, and more elevated
environments occupied by most of its wild
8. C. menglaensis Yin, Li & Xu 2004, wild in relatives. Although the full diversity of
mountains, northern Vietnam to Yunnan, species in Colocasia is not yet known, recent
China. fieldwork and genetic analyses (Ahmed,
2013; Ahmed et al,. 2013) suggests that
9. C. oresbia Hay 1996, wild on Mt Kinabalu, three wild species may be of special
Borneo, Malaysia (also in Bangladesh, significance for the evolution and
according to Ara & Hassan, 2005). domestication of C. esculenta, namely: C.
lihengiae and C. yunnanensis (both sister
10. C. yunnanensis Long & Cai 2006, wild in species?) and C. formosana (a recently
mountains, Yunnan, China. evolved species derived from taro?). In this
paper, we give particular attention to C.
formosana.
The evolutionary origin of C. esculenta was
previously considered to be somewhere in
the vicinity of Northeast India, because that Wild breeding populations require warm
is where the greatest diversity of other wild and wet conditions, and insect
Colocasia species was known (Figures 2 and pollinators
3). Now we can see that wild Colocasia
species are spread over a huge region, from In experimental living taro collections that
India to southern China and Malaysia. are not subject to harvesting, flowering is
Within the expanded region of known common. In temperate regions such as
diversity in Colocasia species, further central Japan and northern New Zealand,
unknown species are likely to be located in floral development is quickly terminated by
steep and inaccessible mountain valleys at the arrival of cold winter conditions
altitudes between 400 and 4,000 metres. (Matthews, 1985; 1995; 2014), and require
glasshouse heating and hand pollination for
The diversification of Colocasia species in breeding experiments in central Japan (cf.
Himalaya may reflect gradual uplift of the Yoshino, 2002). The present authors have
mountain ranges, and repeated cycles of observed wild breeding populations, with
warmer and cooler climate, over many fruit and seed production, in Papua New
millions of years. As mountain populations Guinea, northern Australia, Myanmar,
became increasingly isolated, they may have northern Vietnam, northern Philippines,
given rise to new species. Somewhere in this and northern India. In Papua New Guinea,
mountain zone of speciation, C. esculenta seedlings are common on wet ground
may have evolved as a low-elevation plant around the fruiting parent plants. When
adapted to warm, sunny, and constantly wet collected from ripe fruit, and planted
immediately, taro seeds germinate after

about two weeks at room temperature, of more-or-less specialised insect

when sown on wet soil in a covered pollinators. When Colocasiomyia spp.
container. (Drosophilidae) were first observed on taro
in Papua New Guinea and Southeast Asia, it
It is likely that the most sensitive stage in was believed that they were very host-
the life cycle is when young seedlings are specific (Carson & Okada, 1980; Matthews,
first establishing roots, as the seedlings 2014). This may be true where wild plant
quickly die or become stunted if they populations and the insects have co-existed
experience drought under laboratory in a stable environment for long periods.
conditions. In tropical to subtropical However, when the host plants are moved
regions, warm temperatures and into new environments by transplantation,
continuously damp or wet ground or when the host plants naturally disperse
conditions appear to be ideal for wild and mix with other aroid taxa in disturbed
breeding populations. Young plants may habitats, ecological relationships between
grow slowly in shaded micro-environments host plants and insect pollinators may
near the parent plants, or wherever change rapidly. It is now known that
deposited by animal agents, but once Colocasiomyia spp. are able to move between
established, it is likely that they can quickly Alocasia spp. as host plants (Miyake &
take advantage of any available sunlight (for Yafuso, 2005), and here (Table 1) we report
example, when a canopy gap exists or is the presence on Colocasia spp. of
created by vegetation disturbance). In Colocasiomyia steudnerae, which was first
laboratory experiments with taro seeds and observed on Steudnera (Takenaka et al.,
seedlings, the rate of growth and size of 2006). When pollen are moved between
taro seedlings can easily be controlled by different genera, sterile hybrid offspring
restricting or increasing available nutrients may be produced like those described by
and light (unpublished notes, Field Sciences Yoshino (2002). When pollen is carried
Laboratory, National Museum of between previously-isolated but closely-
Ethnology). The young seedlings require related species in the same genus,
only water and little light, but are very hybridisation might occur easily. If hybrid
tolerant of low or high nutrient conditions, individuals become isolated from their
and low or high light conditions. When parent species, and form a new breeding
placed in fertile soil, in larger pots, with population, then a new species of hybrid
more light, the plants quickly increase in origin may emerge. The causes and
size. In wild taro patches, the flowering consequences of hybridisation are thus
individuals are usually larger plants growing closely linked to pollen flow, plant dispersal,
in sunny positions with deep soil. and any changes in environment that affect
the relative fitness of parent and hybrid
Another important requirement for populations.
successful breeding by taro is the presence

In northern Vietnam, while travelling by existing wild Colocasia species, and also
road from the lowlands to the mountain introducing C. esculenta as a cultigen into
area of Sapa, and in the hills around Sapa, close proximity with those species. The full
we collected Colocasiomyia samples at various range of possible genetic interactions
altitudes from the inflorescences of C. between the different host taxa, and
esculenta, C. lihengiae, C. menglaensis, and C. between the different pollinator taxa, is
yunnanensis. The associations found (Table 1) unknown. With the insect pollinators
suggest that the diversity of pollinators is present and able to follow the floral scents
higher in the mountain zone where multiple of different host plants, there may be a
wild species of Colocasia are present. double-synergy of diversification in host
However, the collecting region is a region plant populations and diversification in the
where there has been both ancient and pollinators. This is a clearly a region where
modern intrusion of farmers and much can be learned about the ecology and
settlements into the mountains, creating genetic potentials of wild and cultivated
new opportunities for the expansion of Colocasia species.
Table 1. Multiple species of Colocasiomyia (Drosophilidae) associated with Colocasia species in

northern Vietnam. Samples collected by Matthews and Nguyen Van Dzu in 2012. Summary
data based on initial identifications provided by M. J. Toda, Hokkaido University Museum.
Host Colocasiomyia
Colocasia species species
identified
xenalocasieae alocasiae steudnerae other spp.
C. esculenta + + + +
C. yunnanensis + - + +
C. lihengieae + - + +
C. sp. Mau Son - - - sp. 2 aff. iskandari
C. gigantea + + - +
Is Colocasia formosana Hayata a with no wild breeding population of taro,

distinct species? and a cultivar assemblage dominated by
cool-adapted, triploid forms of C. esculenta.
Japanese botanists who observed,
Most cultivars known to them in central or
collected, and first described Colocasia
northern Japan would have displayed leaves
formosana Hayata (1919) in Taiwan in the
with a dark-green upper surface and pale
early 20th century came from a country

lower surface, together with starchy mother formosana is not clearly distinct from C.
corms and many side-corms (i.e., the esculenta. In our experience, all traits
morphotype of C. esculenta var. antiquorum). described in the original species description
From the cultivar assemblages of southern (Hayata, 1919) lie within the range of
Japan and Taiwan, they may also have been variation of C. esculenta.
familiar with plants producing large starchy
corms and few side-corms (i.e. the Recently, we have discovered a small
morphotype of C. esculenta var. esculenta) number of apparently isolated populations
(Plucknett, 1983). For them, wild C. of C. formosana in wet mountainous regions
formosana must have appeared very different of Luzon, in the northern and central
from all the known cultivated forms of taro. Philippines (Figure 4). In both Taiwan and
the Philippines, C. formosana is known to be
In recent years, the diversity of Japanese very acrid, and is generally not regarded as
taro cultivars has been quite thoroughly an edible wild vegetable, though methods
described in morphological and genetic for preparing the plant to make it edible are
terms (Hirai et al., 1989; Matsuda 2002), and known. Detailed studies of the ethnobotany
a wild form of taro, C. esculenta var. aquatilis, of this species are needed, but already it
has been reported in Okinawa, southern appears to be a much less useful plant than
Japan, where it is not known to breed wild C. esculenta, which is so widely used as a
(Hotta, 1970; Matthews et al., 1992). This vegetable and fodder plant in Southeast Asia
plant is likely to have been introduced as an (see above).
edible wild vegetable or as a fodder plant
from Southeast Asia. Phenotypically similar The populations of C. formosana found
wild taros are common throughout until now, in Taiwan and the Philippines,
Southeast Asia, New Guinea, and across appear homogeneous in colour and general
northern Australia, but are not uniform in morphology, and are readily recognised in
their chloroplast DNA sequences (Ahmed, the field because of a very rounded leaf
2013). These wild taros often form breeding blade with shallow sinus. All inflorescences
populations in wet open habitats, and are seen on plants in the Philippines and
most abundant in the vicinity of human Taiwan have a long sterile appendage, like
settlements. In terms of overall plant colour that found on the original type specimen
(green), and vegetative morphology from northern Taiwan, and the proportions
(relatively small corms and long stolons), of different parts of the spadix are quite
they are generally more similar to C. uniform. Fruit colour in wild C. esculenta
formosana in appearance than tropical and varies from bright orange/red in Myanmar
temperate cultivars, but their acridity and (Matthews & Naing, 2005) to a yellowish-
morphology do vary, across Asia and the brown colour in Papua New Guinea (Price
Pacific, and remain to be compared et al., 2008). Fruit colour in C. esculenta
systematically. In morphological terms, C. generally (including wild and cultivated

forms) ranges from green to orange and encompassed by the current taxonomic
purple (Ivancic & Lebot, 2000). Thus, the concept of C. esculenta? Most wild species of
bright orange/red fruit colour of C. Colocasia, and especially those that have been
formosana (Figure 5) is not a distinguishing described only recently, are known from
trait, though it is consistent within the relatively few locations. Future exploration
species. to map these species in wet monsoonal
regions of Asia might also reveal further
The northern distribution of C. formosana new species of Colocasia, and these might
in island Southeast Asia hints at a natural include new candidates for involvement in
presence of the species in southern China, the domestication of taro. The optimal
and dispersal eastwards from there, search space for Colocasia as a genus is
following or before a split from C. esculenta, gaining definition, and predictive mapping
or within C. esculenta. The northern methods can be developed and used for the
distribution also lends support to the model genus as a whole, and for particular known
of wide natural range (for C. esculenta) as species.
suggested in Figure 3, assuming that C.
formosana is derived from C. esculenta. A On morphological grounds it is difficult to
genetic bottleneck, in C. esculenta or C. support C. formosana as a species separate
formosana, could have arisen during long from C. esculenta, but since C. esculenta is
distance dispersal from a mainland source itself poorly circumscribed, we cannot reject
region, and/or increasing isolation in C. formosana without further study. In the
Taiwan as a result of rising sea level and future, when wild populations of C. esculenta
changing climate (during the late are better known, it may be necessary and
Pleistocene, for example). Genetic possible to split C. esculenta into multiple
comparisons of the known populations in species, or subspecies, based on wildtypes
Taiwan and the Philippines might provide of defined natural range rather than basing
evidence for movement of C. formosana the taxonomy on a mix of wild and
from west to east, from Taiwan to the cultivated forms with poorly known
Philippines. distributions. This brings us back to the
problem of how to recognise wildtypes in a
Naturalisation and crop domestication species, or species-complex, when the plants
can easily naturalise and perhaps hybridise.
It is difficult to investigate the natural A broad answer to this conundrum is that
history, domestication, and dispersal of taro, we must integrate ethnobotanical,
when so many uncertainties remain in the ecological, and genetic approaches to the
basic description of closely-related species study of the genus and its species.
and their distributions. Where did C.
esculenta originate as a natural biological
In particular, our priorities should be: (i)
species, and how many species are really
Systematic comparison of different wild

populations of taro (C. esculenta), within the history, however, this problem must be
full range of wild species of Colocasia. Of faced directly. Plant taxonomists have
special interest are the genetic relationships nevertheless helped circumscribe the
between wild and cultivated taro in lowland problem through their studies on aroid
to mountain habitats in Asia and the taxonomy and the genus Colocasia in
western Pacific. (ii) Genetic testing for particular. Knowing more about the wild
introgression and hybridisation among wild species of Colocasia makes it easier to
Colocasia species, including C. esculenta, recognise the outer limits of diversity within
during the evolution and domestication of C. esculenta, which in turn makes it possible
taro. (iii) Ecological study of seed to define geographical limits for the species
production, dispersal and germination, and as a whole.
also the distribution, host range, habitat
requirements, and diversity of insect To recognise and locate the genetic and
pollinators, in order to understand how seed geographical origins of cultivated taro, we
dispersal and pollination may limit or must continue looking for ways to recognise
promote hybridisation and speciation in the and map natural wild populations in this
genus Colocasia. (iv) Ethnobotanical study of species. This work is difficult because it
the uses and management of wild taro (C. requires the combined efforts of many
esculenta) and other wild Colocasia species that researchers, in collaboration with diverse
may have been involved in the evolution local communities, using diverse local
and domestication of taro. This, along with languages. The work is also difficult because
the ecological approaches, will help us to natural wild populations can also have
understand the role of commensal and economic significance, which makes it
naturalised populations in the historical impossible to exclude the possibility of a
development and domestication of the human role in their biological history.
crop.
In Southeast Asia, the uses of wild taro as
It may be no exaggeration to say that for food and fodder are significant
99% of living crop species, there has been contemporary economic activities. They are
little or no consideration of how human likely to have started long before
activities might have extended the domestication of the crop. They also link
geographical range of wild populations, or the history of taro to the history and
of naturalisation as a long-term historical domestication of pig, itself an important
process. Ignoring the problem of how to component of economic history in
recognise natural range is convenient for Southeast Asia. Range extension to new
biological studies, because it simplifies the habitats and new regions, and naturalisation
interpretation of genetic variation in wild in new habitats (e.g. ruderal habitats in
populations, and comparison between wild cleared land) and new regions (outside the
forms and cultivars. For understanding crop true natural range), could be central to

domestication in taro, in terms of the Naturalisation as a central focus for

process, without necessarily being central in research
terms of the geography.
In many areas of Southeast Asia, including
the Philippines and Vietnam, naturalised
Naturalisation is a process that can create populations derived from cultivated taro are
new breeding populations that are to some likely to be the main sources of wild taro as
degree isolated from natural wild a food and fodder plant. Since ancient
populations with undesirable traits. In times, such populations might have been
crosses between selected (favoured) important for the generation of new
cultivars and wild taros in any given area, cultivars, including cultivars resistant to
the offspring are likely to be poor and not pests and diseases. Commensal wild
favoured, unless people are in the habit of populations present in and around human
using the particular wild parent plants settlements may or may not be naturalised,
involved, or existing cultivars are not much or deliberately spread. This uncertainty is
different from the wild plants (as might be not a reason to ignore them as ‘invasive
the case during an early period of weeds’, or too difficult to study, or of
domestication), or the wild plants belong to secondary historical importance. Over
a population derived from cultivated forms, thousands of years, naturalised populations
through naturalisation. may have had central and continuing roles
in the primary and secondary domestication
In the last case, the naturalised population of taro. Defining the natural range of taro
might consist of a mixture of clones (Figures 2 and 3) is a necessary challenge
derived from cultivation, and their sexual for evolutionary study of the species, and
offspring, thus retaining the favoured traits has helped bring the process of
of parent cultivars, or producing new naturalisation into focus. To develop more
combinations that are potentially superior to realistic models for the evolution and
those of the parents. In modern breeding domestication of taro, and many crops, it
trials in Papua New Guinea, using wild taro may help to make naturalisation a central
as a source of disease resistance was very focus of research. Further taxonomic and
difficult because the genetic basis of ethnobotanical studies will also help in the
resistance was not easily separated from development of models that embrace
undesirable qualities introduced from wild commensal wild populations and
parents that may have been natural naturalisation. The phenomenon of natural
wildtypes. Even with deliberate wild populations becoming ‘denaturalised’
backcrossing, assessment, and selection, it or commensalized, as they spread by natural
was difficult to produce acceptable new dispersal into human environments, is also
cultivars using unscreened wild parent of interest. This can be seen in Taiwan,
plants (Ivancic & Lebot, 2000). where C. formosana has been found (in just

one location, to date) in an agricultural Ahmed, I., P. J. Matthews, P. J. Biggs, M.

lowland near forested mountains. Naeem, P. A. McLenachan, P. J.
Lockhart. 2013. Identification of
chloroplast genome loci suitable for
ACKNOWLEDGMENTS high-resolution phylogeographic
Special thanks to M. J. Toda, Hokkaido studies of Colocasia esculenta (L.)
University Museum, Japan, for insect Schott (Araceae) and closely related
identifications, I. Ahmed, for research taxa. Mol. Ecol. Resour. 13: 929– 937.
discussions, W. Wilfredo, Department of
Botany, National Museum, for field support Bown, D. 2000. Aroids: Plants of the Arum
in the Philippines, T. Masuno, Graduate Family. Timber Press: Portland.
University of Advanced Studies, Japan, for
field support in Vietnam, and E. Tabuchi,
Carson, H. L. & T. Okada. 1980.
National Museum of Ethnology, for
Drosophilidae associated with
technical assistance. Recent fieldwork in
flowers in Papua New Guinea: I.
Taiwan was assisted by K. C. Tsai. This
Colocasia esculenta. Kontyu, Tokyo 48:
paper was prepared under academic
15–29.
agreements between the National Museum
of Ethnology, Japan, and the National
Museum, Philippines (2012–2017), and the Hasskarl, J. K. 1848. Plantae javanicae rariores,
Institute for Ecology and Biological adjectis nonnullis exoticis, in Javae hortis
Resources, Vietnam (2012–2017). The cultis descriptae, Berolini: Sumptibus A.
research was supported in part by JSPS Foerstner.
KAKENHI Grant No. 23405004, Japan.
Hay, A. 1998. Botanical varieties in taro,
Colocasia esculenta: leaving old baggage
REFERENCES behind. Commissioned report.
C02C/98/038. Sydney.
Ahmed, I. 2013. Evolutionary Dynamics in
Taro (Colocasia esculenta L.). Ph.D. Hayata, B. 1919. Colocasia formosana Icones
thesis, Massey University, Palmerston plantarum formosanarum nec non et
North. contributiones ad floram formosanam 8:
133–134, f. 60.

Hirai, M., T. Sato & K. Takayanagi. 1989. Hutterer, K. L. 1983. The Natural and
Classification of Japanese cultivars of Cultural History of Southeast Asian
taro (Colocasia esculenta (L.) Schott Agriculture. Anthropos 78: 169–212.
based on electrophoretic pattern of
tuber proteins and morphological Ivancic, A. & V. Lebot. 2000. The genetics and
characters. Jap. J. Breed. 39: 307–317. breeding of taro. CIRAD, Monpellier.
Hotta, M. 1970. A system of the Family Masuno, T., D. N. A. Luu, V. D. Nguyen, &
Araceae in Japan and adjacent areas, P. J. Matthews. 2012. Fodder sources
Part I. Memoirs of the Faculty of Science, and backyard pig husbandry in BaVi,
Kyoto University Series of Biology 4(1): Hanoi province, northern Vietnam.
72–96. In A.-N. 2012 (Ed.), Proc. of the 1st
International Conference on Animal
Hotta, M. 1983. Taxonomy and origin of Nutrition and Environment, Khon
root crops in East Asia. The Search for Kaen University Press, Khon Kaen,
Japanese Culture: the Origins of Japanese pp. 657– 660.
Agriculture. K. Sasaki. NHK
Publication Association, Tokyo: pp. Matsuda, M. 2002. Taro in Japan, and its
18–42 (in Japanese). Dispersal in East and Southeast Asia.
In S. Yoshida & P. J. Matthews (eds)
Hubbard, F. T. & A. Rehder. 1932. Vegeculture in Eastern Asia and Oceania,
Nomenclatorial notes on plants Japan Center for Area Studies, Osaka,
growing in the botanical garden of pp. 117–134.
the Atkins Institution of the Arnold
Arboretum at Soledad, Cienfuegos, Matsuda, M. & E. Nawata. 2002. Taro in
Cuba. Botanical Museum Leaflets 1: 1– northern Vietnam: its uses,
11. cultivation, and genetic variation.
Japanese Journal of Tropical Agriculture
Hunt, H. V., H. M. Moots, & P. J. Matthews. 46: 247–258.
2013. Genetic data confirms field
evidence for natural breeding in a Matthews, P. J. 1985. Nga taro o Aotearoa.
wild taro population (Colocasia Journal of the Polynesian Society 94:
esculenta) in northern Queensland, 253–272.
Australia. Gen. Resour. Crop Evo. 60:
1695–1707.
Matthews, P. J. 1987. Wild taro and the
context of cultivation. Aroideana 10:
9– 13.

Matthews, P. J. 1990. The Origins, Dispersal Matthews, P. J. 2002. Taro storage systems.
and Domestication of Taro. In S. Yoshida & P. J. Matthews (eds.)
Unpublished PhD thesis, The Vegeculture in Eastern Asia and Oceania
Australian National University, JCAS Symposium Series 16. Japan
Canberra. Center for Area Studies, Osaka, pp.
135–163.
Matthews, P. J. 1991. A possible tropical
wildtype taro: Colocasia esculenta var. Matthews, P. J. 2003. Taro planthoppers
aquatilis. Bulletin of the Indo-Pacific (Tarophagus spp.) in Australia and the
Prehistory Association 11: 69–81. origins of taro (Colocasia esculenta) in
Oceania. Archaeology in Oceania 38:
Matthews, P. J., E. Takei, & T. Kawahara. 192–202.
1992. Colocasia esculenta var. aquatilis
on Okinawa Island, southern Japan: Matthews, P. J. 2006. Written records of
the distribution and possible origins Taro in the Eastern Mediterranean.
of a wild diploid taro. Man and In: Z. F. Ertug (ed) Proceedings of the
Culture in Oceania 8: 19–34. Fourth International Congress of
Ethnobotany (ICEB 2005), Yayinlari,
Matthews, P. J. 1995. Aroids and the Istanbul, pp. 419–426.
Austronesians. Tropics 4: 105–126.
Matthews, P. J. 2010. An introduction to the
Matthews, P. J. 1996. Ethnobotany, and the history of taro as a food. In V. R.
origins of Broussonetia papyrifera in Rao, P. J. Matthews, P. B. Eyzaguirre
Polynesia (an essay on tapa and D. Hunter (eds) The Global
prehistory). In: J. M. Davidson, G. Diversity of Taro: Ethnobotany and
Irwin, B. F. Leach, A. Pawley, D. Conservation. Bioversity International,
Brown (eds) Oceanic Culture History: Rome, pp. 6–30.
Essays in Honour of Roger Green. New
Zealand Journal of Archaeology Matthews, P. J. 2014. On the Trail of Taro: An
Special Publication, Wellington, pp. Exploration of Natural and Cultural
117–132. History. National Museum of
Ethnology, Osaka.
Matthews, P. J. 1997. Field guide for wild-
type taro, Colocasia esculenta (L.)
Schott. Plant Genetic Resources
Newsletter No. 110: 41–48.

Matthews, P. J., E. Takei and T. Kawahara Plucknett, D. L. 1983. Taxonomy of the

1992. Colocasia esculenta var. aquatilis Genus Colocasia. In J.-K. Wang (ed)
on Okinawa Island, southern Japan: Taro: A Review of Colocasia esculenta
the distribution and possible origins and its Potentials. University of Hawaii
of a wild diploid taro. Man and Press, Honolulu, pp. 14–33.
Culture in Oceania 8: 19– 34.
Price, T. V., K. Poka & G. Bogarei.(2008.
Matthews, P. J. & K. W. Naing. 2005. Notes Observations on the liberation,
on the provenance and providence dispersal and germination of taro
of wildtype taros (Colocasia esculenta) seed (Colocasia esculenta) under natural
in Myanmar. Bulletin of the National conditions in Papua New Guinea.
Museum of Ethnology 29: 587–615. Aroideana 31: 134–143.
Matthews, P. J., E. M. G. Agoo, D. N. Rumphius, G. E. (1741–50) [2011] The

Tandang & D. A. Madulid. 2012. Ambonese Herbal. Translated,
Ethnobotany and Ecology of Wild annotated and introduced by E. M.
Taro (Colocasia esculenta) in the Beekman. Yale University Press:
Philippines: Implications for New Haven, CT [originally published
Domestication and Dispersal. In M. in Amsterdam by Apud Fransicum
Spriggs, D. Addison, & P. J. Changuion, Joannem Catuffe,
Matthews (eds) Irrigated Taro Hermannum Uytwerf].
(Colocasia esculenta) in t he Indo-Pacific.
Senri Ethnological Studies 78. Scarlett, N. H. 1985. A preliminary account of
National Museum of Ethnology, the ethnobotany of the Kija people of
Osaka, pp. 307–340. Bungle Bungle outcamp. East Kimberley
Working Paper No. 6. Botany
Miyake, T. & M. Yafuso. 2005. Pollination Department, La Trobe University,
of Alocasia cucullata (Araceae) by two Bundoora.
Colocasiomyia flies known to be
specific pollinators for Alocasia odora. Spencer, R., R. Cross & P. Lumley. 2007.
Plant Species Biology 20: 201–208. Plant names: A guide to botanical
nomenclature. CSIRO Publishing,
Orchard A. E. 2006. Infra-specific variation Collingwood.
in Colocasia esculenta (L.) Schott
(Araceae). Australian Systematic Botany
Society Newsletter 129: 2-5.

Takenaka, K., J.-T. Yin, S.-Y. Wen & M. J. Yoshino, H. 2002. Morphological and
Toda. 2006. Pollination mutualism Genetic Variation in Cultivated and
between a new species of the genus Wild taro. In S. Yoshida and P. J.
Colocasiomyia de Meijere (Diptera: Matthews (eds) Vegeculture in Eastern
Drosophilidae) and Steudnera Asia and Oceania. Japan Center for
colocasiifolia (Araceae) in Yunnan, Area Studies, Osaka, pp. 95–116.
China. Entomological Science 9: 79–91.
Yuasa, H. 2000. Segawa's Illustrated
Yang Y. and J. Xu, 2000. Ethnobotanical Ethnography of Indigenous Formosan
studies of taro, Colocasia esculenta (L.) People: The Tsou. Moritani and Co.,
Schott, in Yunnan, China. In: D. Zhu, Tokyo.
P. B. Eyzaguirre, M. Zhou, L. Sears,
G. Liu (eds) Ethnobotany and genetic
diversity of Asian taro: focus on China.
IPGRI Office for East Asia, Beijing,
pp. 6–11.

Colocasia Esculenta C. Formosana

Uploaded by

Copyright:

Available Formats

Colocasia Esculenta C. Formosana

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Colocasia Esculenta C. Formosana

Uploaded by

Copyright:

Available Formats

Matthews, Nguyen, Tandang, Agoo and Madulid, 2015 Taxonomy and ethnobotany of Colocasia...

Taxonomy and ethnobotany of Colocasia

Nguyen, Van Dzu,

ABSTRACT vegetable for human and animal

Aroideana VOL 38E NO 1, 2015 153

Aroideana VOL 38E NO 1, 2015 154

of taro, closer study of C. formosana is ‘water kelady’ by Rumphius) (Figure 1).

Aroideana VOL 38E NO 1, 2015 155

Aroideana VOL 38E NO 1, 2015 156

Aroideana VOL 38E NO 1, 2015 157

Aroideana VOL 38E NO 1, 2015 158

Aroideana VOL 38E NO 1, 2015 159

Aroideana VOL 38E NO 1, 2015 160

For human consumption especially, the

Aroideana VOL 38E NO 1, 2015 161

Aroideana VOL 38E NO 1, 2015 162

Aroideana VOL 38E NO 1, 2015 163

Aroideana VOL 38E NO 1, 2015 164

recently extended to Arunachal Pradesh in environments, in contrast to the cooler,

Aroideana VOL 38E NO 1, 2015 165

about two weeks at room temperature, of more-or-less specialised insect

Aroideana VOL 38E NO 1, 2015 166

Table 1. Multiple species of Colocasiomyia (Drosophilidae) associated with Colocasia species in

Is Colocasia formosana Hayata a with no wild breeding population of taro,

Aroideana VOL 38E NO 1, 2015 167

Aroideana VOL 38E NO 1, 2015 168

Aroideana VOL 38E NO 1, 2015 169

Aroideana VOL 38E NO 1, 2015 170

domestication in taro, in terms of the Naturalisation as a central focus for

Aroideana VOL 38E NO 1, 2015 171

one location, to date) in an agricultural Ahmed, I., P. J. Matthews, P. J. Biggs, M.

Aroideana VOL 38E NO 1, 2015 172

Aroideana VOL 38E NO 1, 2015 173

Aroideana VOL 38E NO 1, 2015 174

Matthews, P. J., E. Takei and T. Kawahara Plucknett, D. L. 1983. Taxonomy of the

Matthews, P. J., E. M. G. Agoo, D. N. Rumphius, G. E. (1741–50) [2011] The

Aroideana VOL 38E NO 1, 2015 175

Aroideana VOL 38E NO 1, 2015 176

You might also like