Genet Resour Crop Evol (2011) 58:1263–1274

DOI 10.1007/s10722-011-9752-z

NOTES ON NEGLECTED AND UNDERUTILIZED CROPS

Morphological variability in 17 wild elephant foot yam

(Amorphophallus paeoniifolius) collections from southwest
India
Shirly Raichal Anil • E. A. Siril • S. Suhara Beevy

Received: 22 February 2011 / Accepted: 31 August 2011 / Published online: 15 September 2011
Ó Springer Science+Business Media B.V. 2011

Abstract Amorphophallus paeoniifolius (Dennst.) genetic advance (GA) as a percent of the mean
Nicolson is a tuberous herb occurring in the wild and assessed for 18 accessions revealed high heritability
cultivated state. Vegetative morphological characters estimates. A highly significant (P \ 0.01) correlation
were studied at full foliage stage in 17 accessions of coefficient between circumference of petiole base and
A. paeoniifolius which include two cultivars, 15 wild corm diameter, corm height, corm weight, east west
accessions and a related species, A. dubius Blume. The spread and north south spread suggests that circum-
first principal component (PC) accounted for 42.32% ference of petiole and canopy spread are indicators to
of phenotypic variance followed by second for another corm weight and size.
18.38%. Major traits that accounted for more vari-
ability in both PC1 and PC2 include offset shape, Keywords Amorphophallus paeoniifolius

cormel weight per corm, corm fresh weight, petiole Genetic advance
Genotypic coefficient of variation

surface pattern and canopy spread. The unweighted Multivariate analysis
Neglected and underutilized
pair- group method with mathematical averaging crops
Phenotypic coefficient of variation

(UPGMA) clustering method revealed three principal Plant genetic resources
Wild accessions
clusters which separated all the accessions between
Euclidean distances of 0.3–1.6. Both cluster analysis
and principal co-ordinate analysis revealed T10, a
morphotype of A. paeoniifolius var. campanulatus Introduction
(Decne.) Sivad. cultivated in Tamil Nadu and P19
accession of A. paeoniifolius var. paeoniifolius from The genus Amorphophallus Blume ex Decne., a
Karnataka as morphologically distinct, which needs tuberous herb belonging to the family Araceae and a
further validation on the basis of floral characters or paleotropical aroid comprising of more than 200
molecular markers and G3 from Gujarat as an species (Grob et al. 2002; Van der Ham et al. 2005;
immediate ancestor of cultivated elephant foot yam. Sedayu et al. 2010; Jaleel Abdul et al. 2011) occurring
The genotypic (GCV) and phenotypic (PCV) coeffi- in Africa, Madagascar, India, continental South East
cients of variation, broad sense heritability (h2B) and Asia, Malesia and North East Australia (Mayo et al.
1997) is one of the true marvels of nature and a true
treasure of variation (Hetterscheid and Ittenbach
S. R. Anil
E. A. Siril (&)
S. S. Beevy 1996). Striking variations can be accounted in the
Department of Botany, University of Kerala, Kariavattom,
Thiruvananthapuram 695 581, India spathe characters, appendix, tuber characters, petiole
e-mail: easiril@yahoo.com and individual flowers (Hetterscheid and Ittenbach

123
1264 Genet Resour Crop Evol (2011) 58:1263–1274

1996). In India, sixteen species were reported depressant action on central nervous system (Jayawe-
(Sivadasan and Jaleel 2000). In the present study, era 1981; Dey et al. 2011). The crushed seed relieves
A. paeoniifolius (Dennst.) Nicolson, alone was con- tooth ache and the seed paste applied to relieve
sidered for a detailed morphological characterization rheumatic swellings (Anonymous 1998). It is used in
mainly due to its economic importance. A. paeoniifo- the folk medicine to cease tumor growth, lung
lius is chiefly valued for its edible tuber and profound swelling asthma, vomiting, abdominal pain, piles,
medicinal properties (Hanelt and IPK 2001; Ochse and hemophilic conditions, skin diseases, obesity, dyspep-
Van der Brink 1980). A. paeoniifolius has its distri- sia, debility and control intestinal worms (Kirthikar
bution in Madagascar eastwards via India to Malesia, and Basu 1987; Siraj and Balachandran 1994; Nadk-
southern China, Indo-China, Polynesia, northern Aus- arni and Nadkarni 2000; Prajapathi et al. 2004).
tralia (Hetterscheid and Ittenbach 1996; Lebot 2009). Elsewhere this taxon was referred as A. paeoniifolius f.
Nicolson (1987) has identified the cultivated form as sylvestris Backer with rough petiole (Hanelt and IPK
A. paeoniifolius Nicolson var. campanulatus (Decne.) 2001).
Sivad., commonly called elephant foot yam, with Wild relatives of crop plants are considered as
smooth bright green petiole with white blotches and valuable resources in the crop improvement pro-
leaflet bases not decurrent to junction of the three main grammes. Traditionally morphological and phenolog-
petiolules and a round corm with or without very small ical characteristics are used to develop quantitative
cormels. The corms, young petiole (Dey 1896; Raghu estimates of genetic similarities and relationships
et al. 1999) and unopened inflorescence of this between the wild and cultivated relatives (Mac Key
cultivated form are used as vegetable. Elephant foot 1988). Despite the custom use of molecular markers in
yam is a rich source of carbohydrate, protein, minerals the recent years, morphological descriptor in genetic
like calcium, iron, phosphorous, vitamin A, B, C, diversity analysis is still worthwhile. The multivariate
flavanoids, and fibre (Kay 1987; Shilpi et al. 2005). analysis and in particular principal component and
Important medicinal attributes of Amorphophallus cluster analysis have been an important strategy for
include hepatoprotective, antioxidant, uterus stimu- characterization, evaluation and classification of plant
lating agent (Singh et al. 2011; Laderman1983). In genetic resources, especially when large numbers of
India, A. paeoniifolius var. campanulatus is exten- accessions are to be assessed for several characters
sively cultivated in the states of Gujarat, Maharashtra, (Peeters and Martineelli 1989).
Kerala, Tamil Nadu and Andhra Pradesh (Anonymous Earlier classifications of family Araceae was based
1998). Wild congenerics of the cultivated A. pae- on striking floral characters (Nicolson 1982). How-
oniifolius known as A. paeoniifolius Nicolson var. ever, other workers (Mondal 2005; Sharma 2009)
paeoniifolius (Nicolson 1987) is characterized by argued for equal emphasis to both vegetative and
strongly muricate dark green to purplish petiole, reproductive characters as both of them are subjected
mottled and leaflet bases strongly decurrent to the to different types of selection pressures. Nevertheless,
junction of the three main petiolules. The wild form, no serious attempts have been made till now in
A. paeoniifolius var. paeoniifolius possess powerful designing a valid key for identifying various species
therapeutic action against piles and gastro-intestinal of Amorphophallus based on vegetative morphological
disorders (Raghu et al. 1999). The corms are aperient, characters with few exceptions (Hetterscheid and
carminative and expectorant. The corm extract applied Ittenbach 1996). The present study aimed to analyze
externally as an irritant to treat acute rheumatism, the vegetative morphological variation in A. paeoniifo-
administered internally in the treatment of dysentery, lius and to derive their evolutionary relationships.
diarrhoea, piles, hemorrhoids and in the formulation of
indigenous medicines to cure inflammatory conditions
and ophthalmia (Drury 1873; Watt 1889; Bala et al. Materials and methods
2007; Purwal et al. 2011). The wild form is having
emmenagogue, antibacterial properties and the roots A comprehensive collection of A. paeoniifolius germ-
extracts are used to cure abdominal colic, elephanti- plasm was made due to the continuous field explora-
asis, skin and blood diseases, fistula, glandular swell- tion in Western Ghats and also through National
ings in the neck, urinary diseases and dropsy and has Bureau of Plant Genetic Resources (NBPGR, New

123
Genet Resour Crop Evol (2011) 58:1263–1274 1265

Delhi, India). The collections were maintained in the followed by construction of scatter plots based on
Botanic Garden, Department of Botany, Kariavattom, ordination of accessions was performed. Morphomet-
Thiruvananthapuram, Kerala, India. The collections ric analyses were performed using Multivariate Sta-
consisted of 15 accessions of A. paeoniifolius var. tistical Package MVSP Version 3.1 (Kovach
paeoniifolius, two accessions of A. paeoniifolius var. computing Services, Wales, UK).
campanulatus and one accession of A. dubius Blume. In addition to the above 14 quantitative traits,
(Table 1). An important common feature noticed in lengths of three main petioles (LP1, LP2 and LP3)
two A. paeoniifolius varieties and two different species were considered for analysis of variance (ANOVA)
selected in the present study are long styled flowers using SPSS 7.5 (SPSS Inc. Chicago, IL, USA). In
(Nicolson 1987). On the basis of features like petiole order to determine relatedness of various traits con-
surface pattern, habit and presence of large number of sidered in the present investigation, Pearson’s multiple
globose cormels, A. dubius is regarded as closely correlation matrix was plotted. In addition to these by
allied to A. paeoniifolius var. campanulatus (Hooker using data on quantitative traits, variance components,
1894; Gamble 1967) thus, included in the present coefficient of variation, heritability and genetic
study. Further A. dubius was elevated as synonym of advance (GA) were determined (Johnson et al. 1955).
A. paeoniifolius elsewhere (Hetterscheid and Itten-
bach 1996). Among cultivated form (A. paeoniifolius
var. campanulatus), ‘Gajendra’—a popular high Results and discussion
yielding non-acrid variety (Shirly and Palaniswami
2008) released for the cultivation in India and another Principal component analysis
local cultivar namely ‘karunaikizhangu’ were consid-
ered for the present study. Specific codes were The first principal component (PC) accounted maxi-
allocated to each accession (Table 1). mum variation (Table 3). The first PC had cormel
Observations on 30 vegetative qualitative traits development, cormel shape, petiole surface pattern,
(Table 2) and 14 quantitative traits were scored at full leaflet margin, cormel number, cormel weight per corm,
foliage stage of all the accessions mainly based on cormel length, corm fresh weight, East–West spread and
descriptors developed by National Bureau of Plant petiole length as traits with the highest loadings.
Genetic Resources for elephant foot yam (Abraham However, cormel development, cormel shape, cormel
et al. 2006) with few additional characters as appro- weight per corm, cormel length and cormel number had
priate to wild collections. The quantitative traits were positive values. Therefore PC1 distinguished accessions
recorded on the following; breadth of largest leaflet based on distinct corm and cormel characters.
(BLLFT), breadth of smallest leaflet (BSLFT), cormel The second component accounted for 18.38% of the
weight per corm (CRLWTC), corm diameter (CDM), variability with dominance of corm characters such as
cormel length (CRLL), cormel number (CRLN), corm corm fresh weight, corm diameter and corm height as
fresh weight (CFW), corm height (CHT), east west well as foliar characters like petiole surface pattern,
spread of canopy (EWSPD), north south spread east–west spread, north–south spread, circumference
(NSSPD), length of petiole (LOP), circumference of of petiole and length of petiole. The third principal
petiole base(GOP), length of largest leaflet (LLLFT) component accounted for 11.84% and contributed by
and length of smallest leaflet (LSLFT). cormel development, corm colour (abaxial and adax-
The data on both qualitative and quantitative traits ial), corm flesh colour, petiole surface pattern, corm
were recorded on ten plants in each accession. The fresh weight. Major traits that accounted for more
observations in a complied form encompasses mean of variability in PC1, PC2 and PC3 include cormel shape,
3 years data from 2008–2010. Multivariate analysis corm fresh weight and petiole surface pattern. The
was performed by numeric taxonomic techniques highest loaded variables in PC1, PC2 and PC3 were
using the procedure of principal component analysis cormel weight per corm (0.66), corm fresh weight
(Sneath and Sokal 1973). To bring out the patterns of (0.73) petiole surface pattern (0.36), respectively.
similarity and dissimilarity, data was subjected to Hence corm and cormel characters and foliar charac-
cluster analysis based on UPGMA method to group the ters especially petiole surface pattern are important in
18 accessions. Principal coordinate analysis (PCoA) distinguishing various accessions of A. paeoniifolius.

123
1266 Genet Resour Crop Evol (2011) 58:1263–1274

Table 1 Details of collection of A. paeoniifolius from various localities

Sl. no Accession Place of Genetic group/species Status wild
code collection*/origin or cultivated

1. V1 Vittal, DK A. paeoniifolius var. paeoniifolius Wild

2. V4-P Vittal, DK A. paeoniifolius var. paeoniifolius Wild
3. P1 Puttur, DK A. paeoniifolius var. paeoniifolius Wild
4. P5 Puttur, DK A. paeoniifolius var. paeoniifolius Wild
5. K4 Thrissur, KL A. paeoniifolius var. paeoniifolius Wild
6. P17 Puttur, DK A. paeoniifolius var. paeoniifolius Wild
7. P18 Puttur, DK A. paeoniifolius var. paeoniifolius Wild
8. P19 Puttur, DK A. paeoniifolius var. paeoniifolius Wild
9. T1 Vithura, TVM, KL A. paeoniifolius var. paeoniifolius Wild
10. T10 Nagercoil, TN A. paeoniifolius var. campanulatus Cultivated
11. A.d NBPGR A. dubius Wild
12. K3-1 Thrissur, KL A. paeoniifolius var. paeoniifolius Wild
13. K3-2 Thrissur, KL A. paeoniifolius var. paeoniifolius Wild
14. K3-3 Thrissur, KL A. paeoniifolius var. paeoniifolius Wild
15. K5 Thrissur, KL A. paeoniifolius var. paeoniifolius Wild
16. G3 Navasari, GUJ A. paeoniifolius var. paeoniifolius wild
17. GJ (Gajendra) Kovvur, AP A. paeoniifolius var. campanulatus (National variety) cultivated
18. M1 Moozhiyar, PTA, KL A. paeoniifolius var. paeoniifolius Wild
* DK Dakshina Kannada, Karnataka, PTA Pathanamthitta, Kerala, GUJ Gujarat, AP Andhra Pradesh, TVM Thiruvananthapuram, TN
Tamilnadu

Cluster analysis distinct cormels and smooth petioles with identical

petiole surface pattern (light green with white
UPGMA cluster analysis with 44 variables revealed blotches), but are distant by a Euclidean distance of
three principal clusters which separated all the acces- 1.19. The local cultivar T10 had long cormels on a
sions at a Euclidean distance of 1.6 (Fig. 1). In the first small corm. This might have evolved from A. dubius
principal cluster a morphotype of A. paeoniifolius var. due to homology in corm and cormel pattern. A. dubius
paeoniifolius from Navasari, Gujarat (G3)—a wild has distinct globose cormels. However, accession T10
accession, clustered with A. paeoniifolius var. cam- with many long cormels having medicinal properties
panulatus cv. ‘Gajendra’ at a distance of 1.5. These was thus selected for cultivation. G3 and A. dubius in
were the most distantly placed among all the acces- the above two clusters may be the primitive ones as
sions. Clustering of these accessions was due to large they are found in the wild state and cultivated ones
corm size and canopy spread. The cultivar ‘Gajendra’ might have evolved from that after many cycles of
has smooth petiole and massive habit with large corm selection.
with few cormels where as G3 accession had highly All the remaining wild accessions constituted the
muricate or scabrous petiole, massive habit, and large third cluster which confirms the varietal status of
corm with many cormels (5–20) which indicates that A. paeoniiflius var. paeoniifolius. The third principal
G3 may be an escape from cultivated type or the cluster may be divided into two groups, the first group
former may be a wild ancestor. Both had light green being T1, K3-1, K3-2 and M1 with K3-2 and T1
petiole with white blotches. G3 may be an immediate standing well apart. They clustered together due to
ancestor of cultivated elephant foot yam and the large rare occurrence of cormels in these accessions, but K3-
distance between GJ and G3 may be due to the long 2 has a distinct petiole pattern (Table 2) and T1 having
period of evolution through selection . another pattern (Table 2). Accession T1 remained
In the second principal cluster A. dubius clustered apart within this cluster at a Euclidean distance of 0.9.
with T10 (‘karunaikizhangu’) due to the presence of T1 is distinguished from the other accessions of this

123
Genet Resour Crop Evol (2011) 58:1263–1274 1267

Table 2 List of qualitative variables evaluated for the systematics of A. paeoniifolius

Ref. Character description Descriptor states
no

1. Habit 1-small, 2-massive

2. Size of corm and cormel 1-cormels very small, 2-cormels more prominent
3. Corm shape 1-sub globose, 2-flat globose, 3-depressed globose, 4-saucer-shaped,
5-vertically elongate, 6-irregularly subglobose
4. Offset development 1-absent, 2-seasonal, 3-gradual
5. Offset shape 1-spindle shaped, 2-globose, 3-conical, 4-rhizomatous, 5-elongate,
6-spindleshaped elongate
6. Corm surface 1-rough, 2-smooth
7. Colour of corm surface (upper) 1-black, 2-dark brown, 3-white
8. Colour of corm surface (lower) 1-black, 2-dark brown, 3-white
9. Hairiness of corm 1-pubescent 2-glabrous
10. Colour of corm flesh 1-pale greenish yellow, 2-light yellowish green, 3-light yellow, 4-yellowish
white
11. Foliar phenophase 1-lasting one growing season, 2-rarely long lasting
12. Petiole shape 1-terete, 2-angular
13. CS of the upper portion of petiole near the 1-trigonal, 2-triangular, 3-terete
junction of primary partition
14. Petiole nature 1-smooth, 2-scaberulous, 3-scabrid, 4-scabrous, 5-hairy
15. Petiole colour 1-unicolorous, 2-variously blotched
16. Petiole surface pattern 1-light green with white blotches, 2-dark green with white blotches, 3-Blackish
or blackish green (unicolorous), 4-grey coloured or flesh coloured with dark
green or black small patches and white dots or various combinations of these,
5-Green with dark green long dots which are fused or separate to give a
blackish green appearance, 6-Green with dark green long dots which are
fused or separate to give a blackish green appearance and white irregular
small spots which are not fused, 7-Flesh coloured with dark green long dots,
8-Green with brown spots sometimes fused and separate, 9-Light brownish
with dark brownish patches, 10-Blackish with grayish white tinch or irregular
mixing in between giving a blackish overall appearance, 11-Light green with
white dots, 12-Ivy green with white stripes and/or clear cut white patches.
17. Rachis nature 1-naked, 2-narrowly winged with supernumerary leaflets, 3-broadly winged
with supernumerary leaflets
18. Extent of decurrence 1-decurrent to halfway to the junction of rachises, 2-decurrent to the junction of
rachises, 3-not decurrent
19 Leaflet shape 1-obovate, 2-elliptic, 3-mostly elliptic-few obovate, 4-orbicular, 5-oblong-
elliptic-lanceolate, 6-lanceolate, 7-elliptic-ovate, 8-elliptic-obovate,
9-lanceolate-elliptic-ovate, 10-elliptic-oblong-obovate
20. Leaflet apex 1-acute, 2-acuminate, 3-acute-acuminate, 4-cuspidate, 5-acuminate-cuspidate
21. Leaflet margin 1-entire, 2-erose, 3-undulate (wavy), 4-slightly wavy
22. Leaflet appearance (abaxial) 1-shiny, 2-not shiny (glabrous)
23. Leaflet appearance (adaxial) 1-shiny, 2-not shiny
24. Leaflet colour (abaxial) 0-vivid yellowish green, 1-strong olive green, 2-deep bluish green, 3-strong
olive green with reddish margin, 4-deep bluish green dark green with reddish
margin, 5-brilliant yellowish green
25. Leaflet texture 1-chartaceous, 2-choreaceous
26. Corm emergence 1-regular yearly emergence, 2-shy yearly emergence or delayed
27. Petiole partition 1-primary partition alone, 2-primary and secondary partition, 3-up to tertiary
partition, 4-quarternary partition

123
1268 Genet Resour Crop Evol (2011) 58:1263–1274

Table 2 continued
Ref. Character description Descriptor states
no

28. Canopy spread 0-sparse, 1-medium, 1-thick

29. Cormel number grouping 0-No cormel or rare occurrence of cormel;1- \ 4 cormels; 2-[ 4 cormels
30. Phenology-vegetative phase 1-normal season (April–Sep), 2-late (Aug–Nov)

Table 3 Principal component analysis in 18 accessions of to the presence of identical petiole pattern. The
A. paeoniifolius, Eigen values, Eigen vectors and percent of accessions P17 and K4 are least distant among all
variation accounted by the first two principal components accessions (0.3), though they are from different
(highly loaded variables in combined analysis given and sig-
localities. Both are similar in their petiole characters
nificant PC values indicated in boldface) (log10 transformed)
(Table 2), and canopy spread. The accessions V1, V4-
Variables PC1 PC2 PC3 P, K5 and K3-3 with thick canopy grouped together
Cormel development 0.260 0.040 0.087 with significant intra-clusteral distance. V4-P, K5 and
Cormel shape 0.250 0.125 20.164 K3-3 are having identical petiole pattern and thick
Colour of corm (abaxial) 0.010 -0.032 0.101 canopy whereas V1 has a separate pattern. The intra
Colour of corm (adaxial) 0.018 -0.092 0.225 clusteral distance observed within this cluster may be
Corm flesh colour -0.051 0.051 -0.160
indicative of the different stages in the evolution of
Petiole surface pattern -0.117 -0.223 0.363
A. paeoniifolius.
In the above analysis, accession T10, (‘karunaikizh-
Extent of decurrence 0.019 -0.108 -0.010
angu’) a popular cultivar in Tamil Nadu, is placed in a
Leaflet margin -0.151 -0.040 -0.010
different cluster from that of ‘Gajendra’. Both T10 and
Cormel no 0.230 0.018 -0.006
GJ are cultivated forms of A. paeoniifolius var.
CRLWTC 0.668 0.287 0.029
campanulatus and show similarity in petiole surface
CDM -0.060 0.239 0.035
pattern but T10 stands distinct from all the remaining
CRLL 0.184 0.084 -0.087
accessions due to the presence of long cormels and
CRLN 0.388 -0.007 0.233
small corm. T1 and P19 were the outliers in the third
CFW -0.209 0.735 0.261
principal cluster. The voucher specimens of these
CHT -0.049 0.212 0.050
accessions are being deposited in the Herbarium of the
EWSPD -0.134 0.166 -0.012
Department of Botany, University of Kerala .
GOP -0.103 0.203 -0.072
LLLFT -0.112 0.003 -0.016
NSSPD -0.136 0.167 -0.028 Principal co-ordinate analysis (PCoA)
LOP -0.112 0.186 -0.061
Eigen values 0.432 0.188 0.121 PCoA involving both qualitative and quantitative
Percent variation 42.32 18.38 11.84 characters considered all together (Fig. 2) gave true
Cumulative percentage 42.32 60.70 72.54 species coordination for A. dubius and cultivated
elephant foot yam ‘Gajendra’ (GJ) indicating them as
cluster due to the presence of light yellowish green separate species. The accession from Navasari, Gujarat
corm flesh colour. (G3) also stood apart as it is distinguishable from all
The second group within this principal cluster was accessions in the study due to the presence of the
the largest with 10 accessions with P19 placed well highest number of small globose cormels (16–20) in a
apart within this group at a distance of 1.1. P19 is fully matured corm. Accession T10 stood far apart
distinct from all accessions in having a massive habit due to the presence of more number of long cormels
and delayed or shy or inconsistent emergence of and the findings are in tune with the plotted cluster
vegetative phase. P1, P17, K4, P18 and P5 clustered diagram.
together because of their medium canopy with P1 The cluster and principal component analysis on
standing apart. P18 and P5 are close to each other due vegetative morphological characters revealed the

123
Genet Resour Crop Evol (2011) 58:1263–1274 1269

Fig. 1 UPGMA
phenogram based on
qualitative and quantitative
characters

Fig. 2 PCoA based on qualitative and quantitative characters

existence of variability among the investigated acces- Cormel shape, number, length, cormel weight per
sions. The multivariate analysis recorded considerable corm had positive values in PC1. At the same time
variation and similarities among the accessions, petiole surface pattern and other leaf characters were
principally due to corm or cormel characters and highlighted in the PC2. However, they had negative
petiole surface pattern. The greater importance of such values. In sweet potato Veasey et al. (2007) had
characteristics was revealed in cluster and PCoA. reported the shape of storage root and secondary flesh

123
1270 Genet Resour Crop Evol (2011) 58:1263–1274

colour as important characters contributing to diver- Analysis of variance

sity. In Smallanthus sonchifolius (Poepp. et Endl.) H.
Robins (yacon) also tuber shape is the most important Analysis of variance performed on various quantita-
character contributing to diversity and it was proved to tive data showed significant (P \ 0.05) variation
be genetically determined Valentovaä et al. (2006). In among the 18 accessions studied (Table 4). Breadth
the present study corm flesh colour was an important of largest leaflet was significantly (P \ 0.05) high in
trait which accounted for variability in PC3. Tuber K5. Cormel length was significantly high in V4
flesh colour is an important character contributing to followed by K3-3. Cormel number per corm was
diversity in genus Dioscorea (Dansi et al. 1999). In significantly (P \ 0.05) high in G3 followed by T10.
potato, Andean folk classification system is based on Corm diameter, corm fresh weight, corm height,
tuber characteristics along with other characters. circumference of petiole base, EW spread, NS spread
Petiole pigmentation has been an important aerial and petiole length was significantly high (P \ 0.05) in
vegetative descriptor in sweet potato (Veasey et al. cultivar ‘Gajendra’, indicating the large size of the
2007). As per PCoA, petiole surface pattern was an plant. Length of largest leaflet was maximum in K3-1,
important character in distinguishing the various M1 and K3-2 and was significant (P \ 0.05).
accessions as evident due to PC1, PC2 and PC3. The highest coefficient of variation (Table 4)
Nicolson (1987) has grouped long styled species estimated for cormel weight per corm (213.3), cormel
such as A. dubius, A. campanulatus, A. paeoniifolius length (112.0), cormel number (178.6) and corm fresh
under Amorphophallus sect. Candarum. Several char- weight (104.6) signify the existence of high degree of
acters have been used to segregate the species in variability with regard to these traits and the same can
various parts of Asia. Unfortunately these characters be used for distinguishing the accessions. The coef-
segregate separately rather than in groups making it ficient of variation was lowest for breadth and length
difficult to define the taxa (Nicolson 1987) and of largest and smallest leaflet and petiole length
suggests that muricate petiole and strongly decurrent indicating that leaf characters are least important in
leaflet bases are prioritized vegetative characters distinguishing the accessions. However, qualitative
distinguishing wild A. paeoniifolius from the culti- foliar characters have some role in distinguishing the
vated ones (Hanelt and IPK 2001). In the present study accessions as evident by PCA where petiole surface
two varieties viz., paeoniifolius and campanulatus pattern was one of the highly loaded variables in the
were clustered in two different principal clusters first three principal components.
except T10 which clustered with A. dubius in a The phenotypic coefficient of variation (PCV) was
different principal cluster. In a previous systematic slightly higher than the genotypic coefficient of
analysis (Nicolson 1987) also these two varieties were variation (GCV) for all the characters studied signi-
placed into different groups. Major characters attrib- fying that genotypic factors exerted a reasonable effect
uted to this grouping in the present study was in estimating the variation. Though environment
variations in petiole surface pattern, corm and cormel influenced the expression of these characters to some
characters, corm flesh colour, leaflet characters and extent (Table 5) as is the case in most of the crops e.g.
canopy spread. An intra-specific variability based on West African rice (Okelola et al. 2007), the magnitude
morphological characters was highlighted and they of error variance was relatively lower than the
clustered at distance between 0.26 and 1.2 which genotypic variance for all traits. Lower level of error
indicated intermediate stages in the evolution of variance indicates that the genotypic component
cultivated elephant foot yam. But the morphotypes drives the total variance for these characters. It can
P19, T10 may be considered as distinct taxonomic be concluded that most of the variability observed in
units at least at varietal level if this morphological the phenotype for different characters has more of a
characterization remains constant (Pissard et al. 2008) genetic than a non-genetic basis. The variability due to
and those morphological marker systems allowed genotypic variance indicates considerable scope for
classifying the accessions and morphotypes (Arbizu selection. The high heritability estimates for most of
et al. 1997) but needs validation with floral characters the characters indicates that environmental factors did
and molecular markers. not greatly affect phenotypic variation for the traits;

123
Genet Resour Crop Evol (2011) 58:1263–1274 1271

Table 4 Ranges, means,

Characters Range Mean SE (±) CV F values
and standard error for 17
for accessions
characters in 15 accessions
of A. paeoniifolius var. BL leaflet (cm) 3.2–8.4 5.7 0.08 20.95 45.86*
paeoniifolius, two
accessions of Leaflet (cm) 0.9–3.5 2.2 0.04 27.55 11.48*
A. paeoniifolius var. Cormel wt per corm (g) 0–90 6.0 0.94 213.3 7.99*
campanulatus and one Corm dm (cm) 3.4–22 8.4 0.23 36.79 32.67*
accession of A. dubius
Cormel length (cm) 0–10 0.8 0.09 112.0 16.01*
Cormel no. 0–20 1.9 0.24 178.6 15.76*
Corm fw (g) 15–2,000* 297.5 23.17 104.6 33.78*
Corm ht (cm) 2–12.5 5.4 0.13 32.93 32.65*
Ewspread (cm) 30–135 75.9 1.85 32.71 27.95*
GOP (cm) 3.5–17 8.6 0.22 34.86 54.42*
Llleaflet (cm) 9.9–26 15.7 0.33 28.69 50.25*
Lsleaflet (cm) 1.9–7.3 4.2 0.07 25.45 13.19*
NSSpread (cm) 30–135 75.9 1.82 32.33 31.05*
Petiole length (cm) 22–127 54.3 1.52 37.55 43.75*
LP1 (cm) 4–17 8.8 0.23 36.35 33.89*
LP2 (cm) 4–119.5 9.9 0.66 90.05 3.99*
* Significant at P \ 0.05
LP3 (cm) 5–28 12.1 0.34 38.06 28.82*
level

Table 5 Summary statistics and estimates of phenotypic coefficient of variation (PCV) and genotypic coefficient of variation
(GCV), broad sense heritability and genetic advance in 18 accessions of A. paeoniifolius for 17 quantitative traits
Characters Mean Vp Ve Vg PCV (%) GCV (%) H2B GA GA (% of mean)

BL leaflet 5.7 1.2 0.0 1.2 0.47 0.46 98 2.2 39.4

Bsleaflet 2.2 0.2 0.0 0.2 0.31 0.30 91 0.9 39.3
Cormel wt per corm 6.0 77.7 9.7 68.0 3.61 3.38 87 15.9 266.6
Corm dm 8.4 7.8 0.2 7.6 0.96 0.95 97 5.6 66.3
Cormel length 0.8 0.5 0.0 0.4 0.78 0.75 93 1.3 169.2
Cormel no. 1.9 7.3 0.5 6.8 1.97 1.91 94 5.2 279.2
Corm fw 297.5 79,528.0 2,350.4 77,177.5 16.35 16.11 97 563.8 189.5
Corm ht 5.4 2.6 0.1 2.5 0.69 0.68 97 3.2 59.4
Ewspread 75.9 484.6 17.3 467.3 2.53 2.48 96 43.7 57.6
GOP 8.6 8.0 0.1 7.8 0.97 0.96 98 5.7 66.7
Llleaflet 15.7 17.9 0.4 17.6 1.07 1.06 98 8.6 54.5
Lsleaflet 4.2 0.7 0.1 0.6 0.41 0.39 92 1.6 37.9
NSSpread 75.9 484.7 15.6 469.1 2.53 2.49 97 43.9 57.9
Petiole length 54.3 359.0 8.2 350.8 2.57 2.54 98 38.1 70.3
LP1 8.8 8.3 0.2 8.1 0.98 0.96 97 5.8 65.9
LP2 9.9 24.6 6.2 18.4 1.58 1.37 75 7.7 77.5
LP3 12.1 16.8 0.6 16.2 1.18 1.16 97 8.1 67.3

rather genetic constitution of the accessions was by cormel weight/corm, east–west spread, north -south
responsible for the variation. Maximum genotypic spread and petiole length and minimum for breadth of
variation was recorded in corm fresh weight followed smallest leaflet. Phenotypic variation also showed a

123
1272 Genet Resour Crop Evol (2011) 58:1263–1274

similar trend. GCV and PCV were maximum for corm resources of potential value, especially as the culti-
fresh weight and cormel weight per corm and mini- vated species are losing their ability to reproduce
mum for breadth of smallest leaflet. However, GA as sexually because of selective pressures to preferen-
percent of mean was high for corm fresh weight, tially allocate more photosynthate towards tuber
cormel length, cormel number and cormel weight/ production (Singh and Gadgil 1995) and extreme
corm. The high value for GCV and heritability along protogyny (Arakeri 1956; Sreekumari 2000). Hence
with high GA% suggest that improvement is effective conservation of this rich diversity of wild relatives of
through phenotypic selection and indicative of the cultivated plants is important.
effect of additive genes (Pansey and Sukhatme 1985)
for these characters. High heritability associated with Acknowledgments The authors are thankful to Dr. (Mrs.)
Ashalatha S Nair, Professor and Head, Department of Botany,
moderate or low GA% observed for the remaining
University of Kerala for providing facilities. Thanks are due to
characters may be attributed to the non-additive gene NBPGR, New Delhi for providing A. dubius germplasm. SRA is
action and improvement is possible through hybrid- thankful to the Director, Central Tuber Crops Research Institute
ization (Kamruzzahan et al. 2000). (Indian Council of Agricultural Research), Thiruvananthapuram,
Kerala for granting study leave.
Pearson’s correlation coefficient indicate a signif-
icant (P \ 0.01) correlation between circumference of
petiole base (GOP) and corm diameter (r = 0.793),
corm height (r = 0.753), corm weight (r = 0.773), References
East–West spread (r = 0.750) and North–South
spread (r = 0.752). Similarly East–West spread Abraham Z, Latha M, Asha KI, Varghese C, Lakhminarayan S,
(r = 0.579) and North–South spread (r = 0.577) had Pareek SK (2006) Minimal descriptors of agri-horticultural
a highly significant correlation to corm weight. This crops Part V: spices, tubers and plantation crops, NBPGR,
Regional station, Thrissur, pp 102
indicates that GOP and canopy spread are indicators Abraham Z, Latha M, Brinda R (2008) Character association
for the corm weight and size. Abraham et al. (2008) studies in elephant foot yam. J Root Crops 34:70–72
has also reported thickness of petiole, corm diameter Anonymous (1998) The wealth of India–raw materials, vol 1- A.
and number of leaflets as yield attributing characters in Council of Scientific and Industrial Research, New Delhi,
pp 233–234
cultivated form of A. paeoniifolius var. campanulatus Arakeri HR (1956) A note on the storage and germination
based on GCV, PCV, heritability, genetic advance and requirements of seeds of Suran (A. campanulatus). Indian J
association studies. Genet 1:27–29
In several occasions taxon identification is hard or Arbizu C, Blas R, Holle M, Vivanco F, Ghislain M (1997)
Advances in the morphological characterization of oca,
even impossible when only vegetative plant parts are ulluco, mashua and arracacha. Program Report 1995–1996.
present Jaleel Abdul et al. (2011), which is often the International Potato Centre, Lima
case during collection. Thus, present study suggests Bala N, Wilson B, Manju VS, Sundaresan S (2007) Pharma-
that morphological characters such as cormel shape, cological properties of Amorphophallus paeoniifolius. In:
Padmaja G, Premkumar T, Edison S, Bala N (eds) Root and
cormel weight per corm, corm fresh weight, corm flesh tuber crops: proceedings of the national seminar on
colour and petiole surface pattern can be used for the achievements and opportunities in post harvest manage-
identification of various morphotypes of A. paeoniifo- ment and value addition in root and tuber crops (NSTRC
lius. The multivariate analysis separates two acces- 2). Central Tuber Crops Research Institute, Thiruvanant-
hapuram, pp 325–328
sions T10 and P19 as distinct morphotypes with at Dansi A, Mignouna HD, Zoundjihekpon J, Sangare A, Asiedu R,
least a varietal status. Morphological characterization Quin FM (1999) Morphological diversity, cultivar groups
is an easily manageable tool that can be used for and possible descent in the cultivated yams (Dioscorea
delineation of wild germplasm. On the other hand cayenensis-Dioscorea rotundata complex) of Benin
Republic. Genet Resour Crop Evol 46:371–388
additional information about the genotype of the plant Dey KL (1896) The indigenous drugs of India, 2nd edn. Thacker
based on molecular markers and breeding experiments Spink and Co., Calcutta
can supplement to resolve taxonomical problems in Dey YN, De S, Ghosh AK, Gaidhani S, Suman K, Jamal M
this genus. Further, additional traits like medicinal (2011) Synergistic depressant activity of A. paeoniifolius in
Swiss albino mice. J Pharmacol Pharmacother 2:121–123
properties, acridity, resistance to collar rot disease Drury CH (1873) The useful plants of India with notices of their
needs to be evaluated in wild germplasm. The wild chief medicinal value in commerce, medicine and the arts.
congenerics of the cultivated species harbour genetic Higginbotham and Co., Madras

123
Genet Resour Crop Evol (2011) 58:1263–1274 1273

Gamble JS (1967) Flora of the presidency of Madras, vol III. Peeters JP, Martineelli JA (1989) Hierarchical cluster analysis
Botanical survey of India, Calcutta as a tool to manage variation in germplasm collection.
Grob GBJ, Gravendeel B, Eurlings MCM, Hetterscheid WLA Theor Appl Genet 78:42–48
(2002) Phylogeny of tribe Thomsonieae (Araceae) based Pissard A, Arbizu C, Ghislain M, Faux AM, Paulet S, Bertin P
on chloroplast matK and trnL intron sequences. System Bot (2008) Congruence between morphological and molecular
27:453–467 markers inferred from the analysis of the intra-morphotype
Hanelt P, Institute of Plant Genetics and Crop Plant Research genetic diversity and spatial structure of Oxalis tuberosa.
(2001) Mansfeld’s encyclopedia of agricultural and horti- Genetica 132:71–85
cultural crops, vol 6. Springer-Verlag, Berlin, Heidelberg, Prajapathi ND, Purohit SS, Sharma AK, Kumar T (2004) A
New York, pp 2317–2340 hand book of medicinal plants—a complete source book.
Hetterscheid WLA, Ittenbach S (1996) Everything you always Jodhpur, India
wanted to know about Amorphophallus but were afraid to Purwal L, Shrivastava V, Jain UK (2011) Studies on antidiar-
stick your nose into. Aroideana 19:7–151 rhoeal activity of leaves of Amorphophallus paeoniifolius
Hooker JD (1894) Flora of British India, vol V. I L Reeve & Co., in experimental animals. Int J Pharmaceutical Sci Res
London 2:468–471
Jaleel Abdul V, Sivadasan M, Alfarhan AH, Thomas J, Alatar Raghu RV, Deepa C, Sundaran K (1999) A study of Soorana
AA (2011) Revision of Amorphophallus Blume ex Decne. (Amorphophallus paeoniifolius) the king of tubers. In: Bal-
Sect. Rhaphiophallus (Schott) Engl. (Araceae) in India. agopalan C, Nayar TVR, Sundaresan S, Lakshmi KR (eds)
Bangladesh J Plant Taxon 18:1–26 Tropical tuber crops in food security and nutrition. Oxford
Jayaweera DMA (1981) Medicinal plants used in ceylon Part I. and IBH Publishing Co. Pvt. Ltd, Calcutta, pp 10–14
National Science Council of Sri Lanka, Colombo Sedayu A, Eurlings MCM, Gravendeel B, Hettescheid WLA
Johnson HW, Robinson HF, Comstock RE (1955) Estimates of (2010) Morphological character evolution of Amorpho-
genetic and environmental variability in soybeans. Agron J phallus (Araceae) based on a combined phylogenetic
47:314–318 analysis of trnL, rbcL and LEAFY second intron sequences.
Kamruzzahan MM, Islam HR, Alam MF (2000) Variability and Bot Stud 51:473–490
correlation studies in methods for tomato (Lycopersicon Sharma OP (2009) Plant taxonomy, 2nd edn. Tata Mc Graw Hill
esculentum Mill.). Bangladesh J Genet Biotech 1:21–26 Education Pvt. Ltd., New Delhi
Kay DE (1987) Root crops—crop and product digest 2. Tropical Shilpi JA, Ray RK, Sarder SJ, Vddin SJ (2005) Analgesic
Development and Research Institute, London, p 380 activity of Amorphophallus campanulatus tubers. Fitoter-
Kirthikar KR, Basu BD (1987) Indian medicinal plants, vol IV. apia 76:367–370
International Book Distributors, Dehradun Shirly RA, Palaniswami MS (2008) Performance of Gajendra in
Laderman C (1983) Wives and midwives childbirth and nutri- different states of India. In: Palaniswami MS, Shirly RA,
tion in rural Malaysia. University of California Press, Sajeev MS, Unnikrishnan M, Singh PP, Choudhury BC
California, p 247 (eds) National Seminar on Amorphophallus: innovative
Lebot V (2009) Tropical root and tuber crops: cassava, sweet technologies—abstract book, status papers and extended
potato, yams and aroids. Crop production science in hor- summary. Central Tuber Crops Research Institute, Thir-
ticulture (117). CAB Books, CABI, Wallingford uvanathapuram, p 103
Mac Key J (1988) A plant breeder’s aspect on the taxonomy Singh SN, Gadgil M (1995) Ecology of Amorphophallus species
of cultivated plants. Biologisches Zentralblatt 107: of Uttara Kannada district of the Karnataka state, India:
369–379 implications for conservation. Aroideana 18:5–18
Mayo SJ, Bogner J, Boyce PC (1997) The genera of Araceae. Singh SK, Rajasekar N, Vinod Raj AN, Paramaguru R (2011)
Royal Botanical Gardens, Kew Hepatoprotective and antioxidant effects of Amorpho-
Mondal AK (2005) Advanced plant taxonomy. Central Book phallus campanulatus against acetaminophen–induced
Agency, Kolkotta hepatotoxicity in rats. Int J Pharm Pharm Sci 3:202–205
Nadkarni KM, Nadkarni AK (2000) Indian materia medica, vol Siraj VV, Balachandran I (1994) Ayurvedic drugs and their
I. Popular Prakashan, Mumbai plant sources. Oxford and IBH Publishing Co. Pvt. Ltd.,
Nicolson DH (1982) Translation of Engler’s classification of New Delhi
Araceae with updating. Aroideana 5:67–88 Sivadasan M, Jaleel AV (2000) Amorphophallus hirsutus Te-
Nicolson DH (1987) Araceae. In: Dassanayake MD, Fosberg FR ysm. et Binn (Araceae): a new report from India. Rheedea
(eds) A revised handbook to the flora of Ceylon, vol. VI. 10:143–147
Amerind Publishing Co. Pvt. Ltd, New Delhi, pp 17–101 Sneath PHA, Sokal RR (1973) Numerical taxonomy. WH
Ochse JJ,Van der Brink RCB (1980) Vegetables of the Dutch Freeman and Company, San Francisco
East Indies. Amsterdam Sreekumari MT (2000) Flowering, intervarietal hybridization,
Okelola FS, Adebisi MA, Kehinde OB, Olewole AM (2007) selfing and seed production in elephant foot yam (Amor-
Genotypic and phenotypic variability for seed vigour traits phophallus paeoniifolius (Dennst.) Nicolson). J Root
and seed yield in West African rice (Oryza sativa L.) Crops 26:18–22
genotypes. J Am Sci 3:34–41 Valentovaä K, Lebeda A, Dolezÿalovaä I, Jirovskyä D, Simo-
Pansey VG, Sukhatme PV (1985) Statistical methods for agri- novska B, Vovk I, Kosina P, Gasmanovaä N, Dziechciark-
cultural workers. Indian Council of Agricultural Research, ovaä M, Ulrichovaä J (2006) The biological and chemical
New Delhi variability of yacon. J Agric Food Chem 54:1347–1352

123
1274 Genet Resour Crop Evol (2011) 58:1263–1274

Van der Ham R, Grob G, Hetterscheid WLA, Star W, Van potato (Ipomoea batatas) landraces of the Vale do Ribeira.
Heuven B (2005) Notes on the genus Amorphophallus Sci Agric 64:416–427
(Araceae)-13. Evolution of pollen ornamentation and ul- Watt G (1889) Dictionary of economic plants of India, vol 2.
rastructure in Amorphophallus and Pseudodracontium. Supt. Govt. Printing, Calcutta, India
Grana 44:252–265
Veasey EA, Silva JRQ, Rosa MS, Borges A, Bressan EA, Peroni
N (2007) Phenology and morphological diversity of sweet

123