Content uploaded by Budi Waluyo
Author content
All content in this area was uploaded by Budi Waluyo on Jun 24, 2020
Content may be subject to copyright.
B I O D I V E R S I T A S
ISSN: 1412-033X
Volume 21, Number 7, July 2020 E-ISSN: 2085-4722
Pages: 2991-3000 DOI: 10.13057/biodiv/d210716
Variation in morphological traits of a selection of Indonesian winged
bean accessions (Psophocarpus tetragonolobus) and its analysis to assess
genetic diversity among accessions
IZMI YULIANAH1,2, BUDI WALUYO2, SUMERU ASHARI2, KUSWANTO2,
1Graduate Program, Faculty of Agriculture, Universitas Brawijaya. Jl. Veteran, Malang 65145 East Java, Indonesia
2Department of Agronomy, Faculty of Agriculture, Universitas Brawijaya. Jl. Veteran, Malang 65145, East Java, Indonesia. Tel.: + 62-341-569984,
♥ email: kuswantoas@ub.ac.id
Manuscript received: 22 April 2020. Revision accepted: 8 June 2020.
Abstract. Yulianah I, Waluyo B, Ashari S, Kuswanto. 2020. Variation in morphological traits of a selection of Indonesian winged bean
accessions (Psophocarpus tetragonolobus) and its analysis to assess genetic diversity among accessions. Biodiversitas 21: 2991-3000. In
Indonesia, winged bean (Psophocarpus tetragonolobus (L.) DC.) is a traditional vegetable crop grown mainly for its edible green pods.
Plant breeding programs aim to produce cultivars with high production and good nutritional qualities. The objective of this present study
was to assess genetic diversity among 21 selected Indonesian winged bean lines based on observation of morphological characters. This
was the first step in determining an appropriate breeding program for the development of improved vegetable cultivars. Twelve
qualitative characters and eight quantitative variables were assessed for each of the 21 lines. Categorical differences among lines were
observed in characters such as leaflet, pod and seed shape, pod surface texture, anthocyanin pigmentation of stem, flowers and pods.
Several of these characters are useful as genetic markers, and cluster analysis of the 21 lines on the basis of qualitative characters
enabled two distinct groupings to be identified. Quantitative variation across line means was also high for several of the quantitative
variables (a coefficient of variation > 25% for pod length, number of pods per plant, and total pod weight per plant). Principal
component analysis applied to the eight variables accounted for 86% of the total variation in just three components with eigenvalues > 1.
On Component 1, the characters number of days to first open flower, pod length and pod weight were closely aligned with total weight
of pods per plant. Number of pods per plant was not closely aligned with weight of pods per plant. This study has enabled broad
differences between groups of lines to be categorized and has identified particular lines with characteristics that recommend them for
inclusion as parents in inheritance studies designed to elucidate the contribution that individual characters make to overall productivity,
attractiveness, and nutrition of this useful, high protein, vegetable species.
Keywords: genetic diversity, Indonesian, morphological characters, plant breeding, winged bean
INTRODUCTION
Indonesia is recognized as one of the genetic diversity
centers for winged bean. However, the area planted to
winged bean in Indonesia appears not to have expanded
beyond what it was when Thompson and Haryono (1980)
wrote their optimistic report. Cultivation still uses
traditional low input methods, and currently, winged bean
is best considered as a backyard plant, its utilization being
limited to household consumption (Handayani 2013).
For the most part, winged bean landraces in Indonesia
are facultative perennials. They can be planted both in
rainy and dry seasons, with a capacity to produce vegetable
pods in low to moderate quantities over several seasons if
allowed. Winged beans planted in the rainy season produce
more leaf but fewer flowers and pods (Kuswanto et al.
2016). The abundance of leaves during the vegetative
phase appears to inhibit the transition to the generative
phase, resulting in a long, drawn out harvest time. The time
from planting to first harvestable pod in winged bean is
often almost twice as long as for other popular leguminous
vegetable species; i.e. 13-14 weeks for winged bean,
compared to just 8-9 weeks for crops like common bean
(Phaseolus vulgaris L.) and yardlong bean (Vigna
unguiculata L. Walp. subsp. sesquipedalis (Eagleton
2019). Several approaches have been carried out with the
aim of increasing winged bean pod and seed production.
Throughout the traditional growing areas in tropical Asia
and Melanesia, winged beans are indeterminate climbers in
growth habit, so one avenue of research has been to attempt
to manipulate biomass production by such things as
adjustments to trellising support and vegetative pruning
with the aim of stimulating flower and pod production.
Using two-metre high trellising, ratooning, and multiple
harvesting over an extended time, period Wong (1980) and
(Motior et al (1998b) have reported obtaining very high
green pod yields and seed yields of 35 t/ha and 6.26 t/ha
respectively, but the cost of the trellising is high. Another
approach has been to use plant breeding methods to
generate varieties with modified plant architecture,
particularly branching pattern, with a view to maximizing
pod set (Tanzi et al. 2019). Approaches using plant
breeding have included classification of genetic diversity
across the species (Khan 1976; Eagleton et al. 1980;
Mohamadali and Madalageri 2015; Handayani et al. 2015);
expansion of genetic diversity through the induction of
B I O D I V E R S I T A S
21 (7): 2991-3000, July 2020
2992
genetic mutations (Klu et al. 1997); baseline inheritance
studies (Erskine and Khan 1977; Erskine 1980; Erskine
1981); hybridisation followed by screening and selection in
the segregating generations of crosses (Erskine and
Kesavan 1982; Eagleton 1983; Mohamadali and
Madalageri, 2004; Kuswanto et al. 2016; identification of
disease resistant genotypes (Thompson and Haryono
1979); and genomic sequencing studies (Mukhopadhyay
2000; Sharma et al. 1996).
Plant breeding research on winged bean has been
ongoing since the 1970-80’s era, but it has to be admitted
that there have been few developments that have resulted in
significant improvements in winged been productivity, and
the way it is grown and utilized for human consumption.
The research task in Indonesia today is for plant breeders to
develop superior cultivars of winged bean that combine
high productivity of vegetable pods with high nutritional
value under relatively low input conditions; the kind of
variety that would assist traditional horticulturists produce
winged bean not only for their own needs but also enable
them to benefit from expanding urban market opportunities
Genetic diversity within a population is a prerequisite
for a successful plant breeding program (Borojevic 1990).
Adequate genetic diversity provides opportunities for
adaptive genotypic responses to environmental change. The
results of research on clover (Trifolium subterraneum L.)
show that there was a correlation between several agro-
morphological characters with environmental variables.
These can be considered highly adaptive to environmental
pressures (Abdi et al. 2020). Furthermore, it provides
opportunity for plant breeders to manipulate the gene pool
to generate new genotypes expressing desired combinations
of traits. Germplasm evaluation studies over many years
have recorded morphological variations in a wide range of
characters including in leaves, flowers, pods, seeds, stems
and tubers (Khan 1975; Erskine and Kesavan 1982;
Mohamadali and Madalageri 2004; Fatihah et al. 2012;
Mohanty et al. 2013; Handayani et al. 2015).
Principal component analysis, as applied to data
obtained from the measurement of a number of observable
characteristics on a large collection of plant accessions, is a
technique used for objectively reducing the dimensionality
of the dataset to a limited number of uncorrelated, new
variables. It can be used to identify those particular
characteristics that contribute most to the total observed
variation within the collection of evaluated accessions. The
output from cluster analysis consists of coefficients of
similarity or dissimilarity used to measure overall distance
between accessions and between identified groups of
accessions. Genetic distance information derived from the
analysis can be used in decision-making about which
accessions to include as parent lines in on-going plant
breeding programs. Accessions with a wide distance
between them, if included in a breeding program, expand
the selection options available for developing new
cultivars.
There have been few studies on the genetic diversity of
Indonesian winged bean, and none that have attempted to
estimate the genetic distance between accessions in
germplasm collections.. The aim of this study is to glean
information on genetic variation and distance between the
accessions as a preliminary step in the selection of parents
for the eventual development of new productive winged
bean cultivars.
MATERIALS AND METHOD
Study area
The experiment was carried out between November
2017 and June 2018 in Dadaprejo Village, Junrejo District,
Batu City, East Java Province. The field site was located at
an altitude of 600 m asl., at longitude 112.572947E and
latitude -7.908635S and. The daily temperature at the site
ranged between 17C and 30.8C. The average humidity
ranged from 62% to 94% (BMKG 2018).
Materials
The 21 winged beans (Psophocarpus tetragonolobus)
lines evaluated in the study (Table 1) derived from a
collection of accessions assembled from five provinces of
Indonesia (Bengkulu; Central Java; East Java; West Nusa
Tenggara; and Southeast Sulawesi). Nineteen of the winged
bean accessions originated from farmers and/or
communities who plant winged bean in small amounts as a
backyard garden crop or in fringing hedges around fields or
household compounds. These accessions underwent
preliminary screening and selection to identify lines with
interesting, potentially useful characteristics indicative of
possible underlying genotypic differences (Table 1). Seed
from these was multiplied by line-breeding to form 19 of
the lines evaluated in this study. The other two lines
(UB1.1 and UB 1.2) included in this study came from a
collection held in the Faculty of Agriculture at Universitas
Brawijaya. Thus the 21 lines evaluated in the study were a
sampling of Indonesian winged bean germplasm with wide
potential for genotypic variation.
Method
The research method used a randomized block design to
evaluate the 21 winged bean lines in three replicate blocks.
Each block was 3.6 m x 27 m. Individual plots consisted of
beds 1 m wide by 3.6 m in length with an intended plant
spacing within the beds of 0.6 m x 1 m. Each plant was
provided by thick bamboo poles that are separated from
other plants. Bamboo poles height of about 2 m, consisting
of three bamboos and tied at the top. To foster germination,
seeds were scarified using sandpaper to abrade the back of
the seeds and then soaked in warm water for 24 hours. The
21 seed batches were then sown into trays containing a 1: 1
mixture of soil and manure, and grown in the trays for two
weeks before planting out into the field at the two trifoliate
leaf stage. The field planting was carried out in the
afternoon. Each bed plot consisted of six planting holes
with each planting hole filled by a single winged bean
seedling. Re-planting has carried out for two weeks to gap-
fill any missing planting points. Fertilization was
conducted on three occasions: (i) one week after planting
(7 DAP), (ii) during flowering, and (iii) during pod
formation. At 7 days after planting, the fertilizer applied
YULIANAH et al. – Morphological variation of Indonesian winged bean accessions
2993
was Urea 14 g, SP36 (superphosphate containing 36%
P2O5) 20 g; and KCl 14 g per plant. At the flowering stage,
the dosage of fertilizer was Urea10 g, SP36 15 g, and KCl
10 g per plant. During pod formation, NPK fertilizer (15:
15: 15) was applied at a rate of 10 g per plant. Irrigation
was carried out (by opening a water channel once a week)
only when there was no rain. Pests and plant diseases were
controlled mechanically if the intensity was low, and using
pesticides if the pest attack was moderate to severe.
Observations were made on both quantitative and
qualitative characters.
The observed quantitative characters were: (i) days to
first flowering (i.e. the number of days from planting to the
date when the first open flower appeared in the plant); (ii)
days to first green pod harvest (i.e. the number of days
from planting until the date when the vegetable pods in a
plant were judged to first reach harvestable stage); (iii) pod
length based on the mean of ten pods per plant; (iv) pod
width, based on the mean of ten pods per plant; (v) the
weight of a pod, based on the mean of ten pods per plant;
(vi) total number of pods per plant, calculated from the first
to the 10th harvest; (vii) total weight of pods per plant;
calculated from the number of pods multiplied by the
weight of pod; (viii) mean number of seeds per pod;
calculated from a random sample of ten pods. The samples
used for in measurement were three plants in each replicate
plot.
The observed qualitative characters were twelve in
number, namely: stem color; leaflet shape; pod color; pod
wing color; pod surface texture; pod shape; presence of pod
specks; the anthocyanin color intensity in pods; calyx
color; corolla color; seed color; and seed shape (see Table
2). These observations on qualitative characters were
carried out using the IBPGR (1982) descriptor list for
winged bean. In Indonesia, this descriptor list has been
published as a modified Test Implementation Guide (PPU)
to ensure uniqueness, uniformity, and stability in
describing plant characteristics (PVTPP 2014).
Data analysis
For each qualitative character, the number of lines (n)
expressing a particular trait as listed in Table 2 was
converted into a percentage (n x 100/21).
For quantitative characters, mean plot values were used
as input to calculate the minimum, maximum, mean,
standard deviation, and coefficient of variation for each
line across its three replicate blocks, using the 19th edition
of the Genstat software package. In addition, the XLSTAT
version 2014 software package was used to construct
boxplots for each quantitative character that illustrated in
graphical terms the minimum, 1st quartile, median, third
quartile, and maximum values across the 21 lines.
Table 1. The origins of the twenty-one winged bean lines
evaluated in this study
Winged bean
lines
Origin
SKB 1.5
Bengkulu, West Sumatra
PLB 1.1
Brebes, Central Java
PLB 2.3
Brebes, Central Java
KPN 3.2.1
Nganjuk, East Java
BNN 1.1
Nganjuk, East Java
MNN 1.1
Nganjuk, East Java
NSM 2.1
Malang, East Java
CKM 1.1.1
Malang, East Java
KePM 1.2.3
Malang, East Java
SWM 1.1
Malang, East Java
KaPM 2.1
Malang, East Java
MDM 1.2
Malang, East Java
KePM 2.5
Malang, East Java
KePM 2.2
Malang, East Java
DJB 1
Batu, East Java
DJB 2
Batu, East Java
KPJ 1.1.1
Jember, East Java
PTL 2.1
Lombok Utara, West Nusa Tenggara
MML 1.4
Luwu Timur, Southeast Sulawesi
UB 1.1
Plant Breeding Laboratory Collection, Faculty of
Agriculture, Univ. Brawijaya, Malang, East Java
UB 1.2
Plant Breeding Laboratory Collection, Faculty of
Agriculture, Univ. Brawijaya, Malang, East Java
Table 2. List of observed qualitative characters based on the terminology of the modified (IBPGR 1982) descriptor list incorporated in
PVTPP (2014).
Characters
Expression
Stem color
Green (1), Greenish purple (2), Purple (3), Other (4)
Leaflet shape
Ovate (1), Deltoid (2), Ovate lanceolate (3), Lanceolate (4), Long lanceolate (5)
Pod color
Cream (1), Green (2), Pink (3), Purple (4), Other (5)
Color of wing pods
Green (1), Purple (2), Other (3)
Pod surface texture
Smooth (3), Medium (5), Rough (7)
Pod shape
Rectangular (1), Semiflat (2), Flat on side (3), Flat on suture (4)
Presence of pod specks
Absent (1), Present (9)
Anthocyanin color intensity in pods*
None (1), Weak (3), Medium (5), Strong (7), Very strong (9)
Corolla color
White (1), Blue (2), Purplish blue (3), Purple (4), Reddish purple (5)
Calyx color
Green (1), Greenish purple (2), Purple (3), Other (4)
Seed color
Cream (1), Light Brown (2), Brown (3), Purple (4), Black (5), Brown-black (6),
Purplish black (7), White (8).
Seed shape
Round (1), Oval (2), Oblong (3)
Note: * indicates additional characters not listed in Descriptors for winged bean (IBPGR 1982) and PVTPP (2014)
B I O D I V E R S I T A S
21 (7): 2991-3000, July 2020
2994
A multivariate analysis was then carried out on the data
set, employing the Multivariate Statistical Program
(MVSP) software package version 3.1. The matrix of
pairwise Pearson's correlation coefficients between the
eight quantitative variables across the 21 lines provided the
input into a principal component analysis (PCA). The PCA
was used to determine to what extent each of the eight
variables contributed to the total measured variation among
the eight variables measured on the 21 lines. Of the eight
orthogonal components generated by the analysis, only
three components with an eigenvalue >1 were accepted as
accounting for significant proportion of the total variation
(Mattjik and Sumertajaya 2011). On these three principal
components, a character was considered as making a
significant contribution if its loading coefficient ≥ 0.6,
(Peres-Neto et al. 2003). The combined data set was then
subjected to a cluster analysis using an agglomerative
hierarchical clustering procedure applied to Euclidean
distance and a UPGMA (Unweighted Pair Group Method
Using Arithmetic Average) strategy to group the 21 lines
and to represent these groupings in the form of
dendrograms (Mohammadi and Prasanna 2003)
RESULTS AND DISCUSSION
Observations on qualitative characters
Investigations of the genetic variation within
underutilized crop species are indispensable to planning
useful breeding strategies to improve them (Cooper et al.
2001). In addition, comprehensive characterization
provides a foundation for efficient germplasm conservation
that enables future genetic improvement (Fitriana and
Susandarini 2019).
Genetically based phenotypic variation in winged bean
in Indonesia has been observed among local collections of
accessions from several regions. A previous investigation
characterized several local accessions of Indonesian
winged bean and documented observed variation in several
morphological characters e.g. corolla color, appearance of
anthocyanin in pods, pod length, and days to first flowering
(Permatasari et al. 2018).
In the present study, morphological observations were
conducted on 12 qualitative and 8 quantitative characters
among 21 Indonesian winged bean lines. The results are
summarised in Table 3 for the qualitative characters and in
Table 4 for the quantitative characters.
Among the qualitative characters, stem color was
predominantly green (19 lines i.e. 90.47%) while greenish-
purple and purple colors respectively were observed in just
one line each (i.e. 4.76% each). This result is similar to that
of Khan (1976) in Papua New Guinea who observed that
the basic colors of winged bean stem were either green or
purple, with green being the most common. The intensity
of purple on the stem in our study varied from purple tinge
with green-base to purple. Leaflet shapes were of three
types (deltoid 85.71%, ovate-lanceolate 4.76%, and long
lanceolate 9.52%). Three pod shapes were found
(rectangular 61.90%, semi-flat 33.33%, and flat on the side
4.76%). Pod shape was observed by cross-cutting the pods.
Pod surface texture was of two types (smooth 33.33% and
medium 66.67%). Pod color was all green (100%). The pod
wing color had two types ( green 90.48%, and purple
9.52%). The pods had three basic colors i.e. green, pink,
and pale yellow. On green pods, purple specks with various
degrees of color intensity were observed, to the point that
some pods could be considered as uniformly purple in
color (Khan 1976). Pods without any anthocyanin were the
most common (90.47%) while pods with anthocyanin were
found in only two lines (9.52%). Of these two lines, one
had pods with weak anthocyanin color intensity (4.76%)
while the other line had a moderate anthocyanin color
intensity (4.76%) (Figure 1). The pod wing color was
generally green with variant variation from light purple to
dark purple (Khan 1976). There were four corolla colors:
blue (28.57%), purplish-blue (57.14%), purple (9.52%),
and reddish-purple (4.76%) (Figure 2). The corolla base-
color was purple and blue in which blue was more typically
found than purple. In some cases, there were white and
bright blue colors. The most common calyx color was
green (85.71%), while two lines (9.52%) had greenish-
purple calyces, and one line (4.76%) had dark purple
calyces. Seed color also varied, namely light brown
(9.52%), brown (80.95%), and black (9.52%). Seed shape
was predominantly round (18 lines; 85.71%), while three
lines had oval-shaped seed (14.29%).
In this study, purple coloration was found in several of
our defined characters: calyx color; corolla color; pod
specks; pod wing color; and anthocyanin intensity in pods.
Some lines had purple or reddish-purple corollas, while the
calyx color appeared dark purple. Thompson and Haryono
(1980), in their study in Central Java of Indonesian winged
bean germplasm, found evidence that dark purple corolla
color was related to pigmentation in the calyx, pod specks,
and the pod wings. In Papua New Guinea germplasm,
Erskine and Khan (1977) carried out genetic analysis of
five qualitative characters finding each of them to be
strongly influenced by singe gene loci; purple stem color
was found to be completely dominant over green stem
color, purple calyx over green calyx, purple over green pod
wing color, purple specks over uniformly green pod color,
and rectangular pod shape over flat pod shape.
Furthermore, significant genetic linkage was observed
between stem color and calyx color and between wing
color and pod specking. In the present study, it was found
that there was a correspondence between purple or reddish-
purple corolla and the appearance in pods of anthocyanin
with weak to moderate intensity, and with purple color in
the wings of pods. Two lines that had those character traits
were KePM 2.2 and MDM 1.2. It needs to be emphasized
that the phenotypic expression of a simple inherited
character may differ depending on the genetic background
within the particular population in which the character is
observed. Even so, qualitative characters such as stem
color, leaflet shape, calyx/corolla color, pod shape, and pod
surface texture have proved to be very useful, relatively
stable, markers for genetic studies (Erskine and Khan
1977).
YULIANAH et al. – Morphological variation of Indonesian winged bean accessions
2995
A
B
C
Figure 1. Anthocyanin color intensity of winged bean pods. A. None; B. Weak; C. Moderate
A
B
C
D
Figure 2. Color variations of the corollas. A. Blue; B. Purplish blue; C. Purple; D. Reddish purple
Observations on quantitative characters
The results of observations on eight quantitative
characters measured on the 21 winged bean lines are
presented in Table 4, in which the values for the minimum,
maximum, mean, and coefficient of variation for each
quantitative character are listed.
On average, the 21 lines took 68 days from the time of
planting to reach the first open flower stage. The line PTL
2.1 was the quickest to flower (58 days) while KPJ 1.1 was
the slowest (81 days). The average time from planting to
first harvestable green pods was 123 days with MML1.4 b
being the fastest (99 days) and SKB 1.5 the slowest (151
days). The long time required by winged bean, with its viny
plant architecture, to reach flowering and pod maturity has
been reported in many studies (Eagleton 2019; Tanzi et al.
2019). Daylength and temperature have significant
influences on growth and phenology of winged bean
(Schiavianto and Valio 1996). Some Indonesian accessions
adapted to high rainfall and luxuriant growing conditions
such as those that occur in Bogor, West Java, have a
particularly extended vegetative growth phase before they
reach flowering, followed by a long reproductive phase
(Eagleton 2019). On the other hand, Erskine (1981)
reported Papua New Guinea accessions that reached first
open flower in as few as 44 days. In our study, we
identified Indonesian lines in which the number of days to
first flower was as short as 58 days, compared with the
study in Bogor in which no flowering occurred in under 70
days.
Tanzi et al. (2019) have suggested that the interaction
between morphology, growth, and yield is a potential target
in breeding programs aimed at developing productive
winged bean ideotypes. In this study, the line with shortest
pod length (13.7 cm), line PLB 1.1 also had the lowest pod
weight, 8.05 g/pod. while the line with the longest pods
(38.4 cm), line NSM 2.1 also had the highest pod weight
B I O D I V E R S I T A S
21 (7): 2991-3000, July 2020
2996
(25.65 g/pod). The average number of pods per plant
among the 21 lines ranged from 29 in line DJB 1 to 55 in
line SWM 1.1. The total pod weight per plant among the 21
lines ranged from 372.7 g to 1169.5 g, with an overall
mean of 673.1 g. The number of seeds per pod ranged from
11 in line PLB 1.1 to 18 in line NSM 2.1, with an average
of 14.2 per pod across all 21 lines. For the character pod
length, the variation across the 21 winged bean lines
appeared to fall into three natural groupings, lines with
short pods (8.0 cm - 15.0 cm), those with moderate pod
length (15.1 cm - 24.0 cm), and those with long pods
(>24.1 cm). Genotype NSM 2.1, with long pods (> 30 cm)
and high number of seed per pod (> 15) produced the
greatest weight of pods per plant (> 1000g). Kuswanto et
al. (2016) have suggested that pod length is a pivotal
character in the selection and improvement of winged bean
varieties despite its higher yield, the long pods of NSM 2.1
are irregular in shape, thus lowering its aesthetic appeal as
a vegetable variety. Careful consideration will need to be
given to various pod characteristics in developing
vegetable ideotypes for winged bean.
In planning plant breeding programs, a critical issue is
an extent to which phenotypic variability in characteristics
is a reflection of underlying genotypic variability, and thus
is exploitable in genetic terms. One indicator of this
underlying genotypic variability, in a germplasm
evaluation of the kind undertaken in this study, is the
coefficients of variation for the characters assessed (Table
4). Another is the evenness of distribution of this variation
across the range for each character best expressed
graphically in the form of box plots. Figure 3 presents box-
plot graphs for each of the eight quantitative characters, in
which the spread of values for each character is depicted in
terms of the first quartile, median and third quartile values.
Coefficients of variation (CV) for the two characters
number of days to first open flower and for pod width were
less than 10%. For number of days to first harvestable
green pods, for number of pods per plant, and for number
of seed per pod, the CV lay between 13.9% and 20.7%;
while for pod length, pod weight, and pod weight per plant
CV exceeded 25%.
The box plot analyses give an indication not only of the
spread but also the skew in the distribution of the 21 line
means for each character. For number of days to first open
flower, for pod length, and for individual pod weight, there
were more lines with values below the overall mean than
above it. On the other hand, for number of days to first
harvestable green pods and for number of pods per plant
there were more lines above the mean than below it.
Extreme values, such as in pod length for line NSM 2.1, are
a cause of skewness, and draw attention to possible
exploitable genetic variants.
Table 3. Variations in 12 qualitative characters based on the
modified Psophocarpus tetragonolobus descriptor list (PVTPP
2014, derived from IBPGR 1982).
Characters
Description
No. of
lines
% of
lines
Stem color
Green
Greenish purple
Purple
19
1
1
90.47
4.76
4.76
Leaflet shape
Deltoid
Ovate lanceolate
Long lanceolate
18
1
2
85.71
4.76
9.52
Pod color
Green
21
100.00
Color of wing pods
Green
Purple
19
2
90.48
9.52
Pod surface texture
Smooth
Medium
7
14
33.33
66.67
Pod shape
Rectangular
Semiflat
Flat on side
13
7
1
61.90
33.33
4.76
Presence of pod
specks
Absent
Present
19
2
90.47
9.52
Anthocyanin color
intensity in pods
None
Weak
Medium
19
1
1
90.47
4.76
4.76
Corolla color
Blue
Purplish blue
Purple
Reddish purple
6
12
2
1
28.57
57.14
9.52
4.76
Calyx color
Green
Greenish purple
Dark purple
18
2
i
85.71
9.52
4.76
Seed color
Light brown
Brown
Black
2
17
2
9.52
80.95
9.52
Seed shape
Round
Oval
18
3
85.71
14.29
Table 4. Minimum, maximum, mean, and coefficient of variation for eight quantitative characters of 21 winged bean (Psophocarpus
tetragonolobus) lines.
Characters
Code
Min.
Lines
Max.
Lines
Mean
CV (%)
Days to first open flowers
Days to first harvestable green pods
Pod length (cm)
Pod width (cm)
Individual pod weight (g)
Number of pods per plant
Total weight of pods per plant (g)
Number of seeds per pod
DFF
DHGP
PL
PW
IPW
NPP
TWPP
NSP
58
99
13.7
1.30
8.05
29
372.7
11.1
PTL 2.1
MML 1.4
PLB 1.1
KPN 3.2.1
PLB 1.1
DJB 1
UB 1.1
PLB 1.1
81
151
38.4
1.97
25.65
55
1169.5
17.6
KPJ 1.1
SKB 1.5
NSM 2.1
BNN 1.1
NSM 2.1
SWM 1.1
NSM 2.1
NSM 2.1
68
123
22.6
1.47
16.09
43
673.1
14.2
9.5
14.1
27.8
9.6
20.7
32.1
33.4
13.9
YULIANAH et al. – Morphological variation of Indonesian winged bean accessions
2997
Principal components analysis
The results of a principal components analysis (PCA)
applied to the quantitative characters, are presented in
Table 5. Three components with eigenvalues > 1,
accounting in total for 85.57% of the data on the original
eight characters, were considered adequate to describe the
data and to allow for useful inferences about the structure
of the complete data. Component 1, with an eigenvalue of
3.917, accounted for 48.97% of the variation in the original
data set. Component 2 with an eigenvalue of 1.574,
accounted for an additional 19.67% of the original
variability, and component 3 accounted for another 16.9%.
Several characters were closely aligned on component 1.
These were days to first open flower and days to first
harvestable green pod, along with pod length, pod weight,
number of seeds per pod and the total weight of pods per
plant. On the other hand, the two characters number of
pods per plant and pod width, contributed significantly
only to component 3, and their contribution was in opposite
directions.
According to Afuape et al. (2011), principal component
analysis is a technique useful for identifying plant
characters that contribute the most to the observed variation
in genotypic groups. It has practical applications in
exploring germplasm for the identification of parents for
possible inclusion in breeding purposes and for inferring
hypotheses about the contribution that individual characters
might be expected to make to breeding objectives. The
results of our principal components analysis appear to be in
general agreement with those of Erskine and Kesavan
(1982) who found a strong heritable correlation between
pod length and pod weight, but little relationship of these
variables with pod number per plant. The results also
suggest that the quest for an early flowering winged bean
genotype might be associated with a lower accumulative
vegetable pod yield within the Javanese environment
(Eagleton 2019).
Cluster analysis of winged bean lines
The cluster analysis of multivariate dissimilarity
measures between the 21 winged bean lines based on
observations of the 12 qualitative characters is shown in
Figure 4 in the form of a dendrogram, with Euclidean
dissimilarity coefficients ranging from 0 to 9.46. Low
dissimilarity coefficients indicate close relationships
among lines, and between clusters of lines. Eight lines had
dissimilarity coefficients of 0 between them i.e. KaPM 2.1,
MML 1.4, KPN 3.2.1, UB 1.2, KPJ 1.1. 1, KePM 1.2.3,
BNN 1.1, SKB 1.5. This indicated they were identical in
terms of the observed set of qualitative characters. Besides
those eight lines, the two lines (PLB 1.1 and PLB 2.3) had
a dissimilarity coefficient of 0 between them. The lines
PLB 1.1 and PLB 2.3, both originated from Brebes, Central
Java.
In the dendrogram in Figure 4, a Euclidean dissimilarity
coefficient of 4.8 was used as a reference point in
identifying two main clusters among the 21 lines based on
qualitative characters. The similarity of characters in lines
in clusters and the distinguishing of character between
clusters in Table 6.
Figure 3. Boxplots of eight quantitative characters measured across 21 winged bean lines: DFF: days to first Flower; DHGP: days to
first harvestable green pods; PL: pod length (cm); PW: pod width (cm); IPW: individual pod weight (g); NPP: number of pods per plant;
TWPP: total weight of pods per plant (g); NSP: number of seeds per pod. In the box plot, the cross, is the mean value, the blue dots are
the minimum and maximum values, the lower edge of the box is the first quartile value, the central line is the median value, and the
upper edge of the box is the third quartile. The whiskers are the Tukey limits (lower and upper respectively)
B I O D I V E R S I T A S
21 (7): 2991-3000, July 2020
2998
Table 5. The coefficients of loading of quantitative character on
the three principal components as the result of PCA
Characters
PC1
PC2
PC3
Days to first flowering
0.728
-0.565
0.233
Days to green pod harvesting
0.684
-0.606
0.137
Pod length
0.827
0.352
-0.329
Pod width
-0.381
0.290
-0.644
Weight per pod
0.917
0.265
-0.156
Number of pod per plant
-0.277
0.529
0.757
Weight of pods per plant
0.735
0.541
0.356
Seed per pod
0.795
0.191
-0.182
Eigenvalue
Variance (%)
Cumulative (%)
3.917
48.970
1.574
19.670
1.354
16.930
85.570
Appropriate breeding methods based on genetic
variation and distance
Hybridization between distinctive genotypes followed
by judicious selection in the segregating generations is a
classical plant breeding method. If the prospective parent
genotypes are very different from one another then it is to
be expected that there will be a high level of genetic
variation among the segregating generations that ensue
from a cross between them. In our study, the evaluation of
qualitative and quantitative characters on the 21 lines was
an important step in determining the genetic diversity
among our collection of Indonesian winged bean
germplasm and in identifying potential parents for breeding
improved cultivars.
Table 6. The similarity of characters in clusters and the distinguishing character between clusters
The similarity of
character in cluster 1
Lines in
cluster 1
The similarity of
character in cluster
2
Lines in cluster 2
The distinguishing of character
between clusters
Green stem color
Deltoid leaflet shape
Green pod color
Purple pod wing color
Presence of pod specks
Black seed color
MDM 1.2
KePM 2.2
Green pod color
Green pod wing color
Absence of pod specks
NSM 2.1, CKM 1.1, MNN
1.1, PLB 2.3, PLB 1.1, SKB
1.5,
BNN 1.1, KePM 1.2, KPJ
1.1.1, UB 1.2, KPN 3.2.1,
MML 1.4, KaPM 2.1, KePM
2.5, PTL 2.1, DJB 2,
SWM 1.1, DJB 1,
UB 1.1
Pod wing color
Presence of pod specks
Anthocyanin color intensity in pods
Seed color
UPGMA
Euclidean
UB 1.1
DJB 1
SWM 1.1
DJB 2
PTL 2.1
KePM 2.5
KaPM 2.1
MML 1.4
KPN 3.2.1
UB 1.2
KPJ 1.1.1
KePM 1.2.3
BNN 1.1
SKB 1.5
PLB 1.1
PLB 2.3
MNN 1.1
CKM 1.1
NSM 2.1
KePM 2.2
MDM 1.2
9.6 8 6.4 4.8 3.2 1.6 0
Figure 4. Genetic dendrogram of 21 winged bean genotypes analyzed by Unweighted pair group method (UPGMA) based on 12
qualitative characters
YULIANAH et al. – Morphological variation of Indonesian winged bean accessions
2999
Differences in qualitative characteristics in clusters of
lines are useful as markers in validating the effectiveness of
crosses in a breeding program. In some cases, a particular
qualitative characteristic may be linked as a marker to
desired characteristics in a commercial cultivar. In other
cases, particular qualitative characteristics or combinations
of characteristics may be desired in their own right. Thus,
for example, the genotype MDM 1.2 could be used as a
candidate for introducing anthocyanin color with moderate
color intensity, in order to generate novel, purple-winged
pods into cultivars. Among the quantitative characters in
our study, pod length and weight had high coefficients of
variation, with line NSM 2.1 having an exceptional pod
length of 38.4 cm (Table 4). Genetic analysis, involving
crosses between lines like NSM 2.1 and MDM 1.2,
followed by evaluation in F1, F2 and perhaps backcross
generations, would help to elucidate the extent to which
pod length and anthocyanin color are under genetic control
and what influence they have on important productive
parameters such as total yield of vegetable pods (Erskine
and Kesavan 1982).
ACKNOWLEDGEMENTS
The author would like to thank LPDP (Indonesian
Endowment Fund for Education) for the granting of a 2016
Indonesian domestic lecturer scholarship (BUDI DN)
which funded doctoral study research.
REFERENCES
Abdi AI, Nichols PGH, Kaur P, Wintle BJ, Erskine W. 2020.
Morphological diversity within a core collection of subterranean
clover (Trifolium subterraneum L.): Lessons in pasture adaptation
from the wild. PLoS One. 15(1):1-26.
Afuape SO, Okocha PI, Njoku D. 2011. Multivariate assessment of the
agromorphological variability and yield components among
sweetpotato ( Ipomoea batatas ( L .) Lam) landraces. Afr J Plant Sci 5
(2): 123-132.
BMKG. 2018. Data iklim harian. Badan Meteorologi Klimatologi dan
Geofisika. http//dataonline.bmkg.go.id/akses data.
Borojevic S. 1990. Principles and Methods of Plant Breeding. Elsevier
Science Publishers B.V., Netherlands.
Cooper HD, Spillane C, Hodgkin T. 2001. Broadening the genetic base of
crops an Overview. In: Cooper H, Spillane C, Hodgkin T, editors.
Broadening Genetic Base of Crop Production. CABI Publishing,
London.
Eagleton G, Halim A, Chai N. 1980. Towards the effective utilisation of
winged bean germplasm. In: Proceeding of Legumes in the Tropics.:
Universiti Pertanian Malaysia, Selangor, Malaysia.
Eagleton GE. 1983. Evaluation of genetic resources in the winged bean
(Psophocarpus tetragonolobus (L.) DC.) and their utilisation in the
development of cultivars for higher latitudes. [Dissertation]. Institute
of Agriculture, University of Western Australia, Perth.
Eagleton GE. 2019. Prospects for developing an early maturing variety of
winged bean (Psophocarpus tetragonolobus) in Bogor, Indonesia.
Biodiversitas. 20 (11): 3142-3152.
Erskine W, Kesavan V. 1982. Genetic variability in the green pod
production of winged bean. J Hortic Sci 57 (2): 209-213.
Erskine W, Khan TN. 1977. Inheritance of pigmentation and pod shape in
winged bean. Euphytica 26 (3): 829-831.
Erskine W. 1980. Measurement of the cross-pollination of winged bean in
Papua New Guinea. Sabrao J 12 (1): 11-14.
Erskine W. 1981. Heritability and combining ability of vegetative and
phenological characters of winged beans (Psophocarpus
tetragonolobus (L) DC.). J Agric Sci 96: 503-508.
Fatihah HNN, Maxted N, Rico Arce L. 2012. Cladistic analysis of
Psophocarpus Neck. ex DC. (Leguminosae, Papilionoideae) based on
morphological characters. S A J Bot 83: 78-88.
Fitriana N, Susandarini R. 2019. Morphology and taxonomic relationships
of shallot (Allium cepa L. group aggregatum) cultivars from
Indonesia. Biodiversitas 20 (10): 2809-2814.
Handayani T. 2013. Kecipir (Psophocarpus tetragonolobus L .), Potensi
Lokal yang Terpinggirkan. IPTEK Tanaman Sayuran 1: 1-8.
[Indonesian]
Handayani, Kusmana T, Liferdi, Hidayat I. 2015. Karakterisasi Morfologi
dan Evaluasi Daya Hasil Sayuran Polong Kecipir (Psophocarpus
tetragonolobus (L.) DC). Jurnal Hortikultura 25 (2): 126-132.
[Indonesian]
IBPGR. 1982. Revised Winged Bean Descriptors. Southeast Asia
Regional Committee of International Board for Plant Genetic
Resources. FAO, Rome, Italy.
Khan TN. 1975. Variation in winged bean in Papua New Guinea. South
East Asian Genetic Resources. Lembaga Biologi Nasional LIPI.
Bogor.
Khan TN. 1976. Papua New Guinea : A centre genetic diversity in winged
bean (Psophocarpus tetragonolobus (L.) DC). Euphytica 25: 693-
706.
Klu GYP, Jacobsen E, Van Harten AM. 1997. Induced mutations in
winged bean (Psophocarpus tetragonolobus L. DC) with low tannin
content. Euphytica 98: 99-107.
Kuswanto, Ardiarini NR, Saptadi D, Waluyo B. 2016. Evaluation and
selection on local strains of winged bean in Brawijaya University
Indonesia. In: Osman M, Abdullah S, Muhammad N, Milan A,
Sulaiman Z, Rokman Z (eds.). Proceeding Transactions Persatuan
Genetic Malaysia 3: 51-55.
Mattjik AA, Sumertajaya IM. 2011. Sidik Peubah Ganda dengan
Menggunakan SAS. IPB Press, Bogor. [Indonesian]
Mohamadali A, Madalageri M. 2004. Correlation and path analysis
studies for green vegetable pod yield and yield attributes in winged
bean (Psophocarpus tetragonolobus (L.) DC). Indian J Hortic. 61 (4):
335-337.
Mohamadali A, Madalageri MB. 2015. Analysis of genetic variability for
seed yield and its component characters in winged bean
(Psophocarpus Tetragonolobus (L.) Dc). Legum Res Intl J 30 (4):
290-291.
Mohammadi SA, Prasanna BM. 2003. Analysis of genetic diversity in
crop plants — salient statistical tools. Crop Sci 43: 1235-1248.
Mohanty CS, Verma S, Singh V, Khan S, Gaur P, Gupta P, Nizar MA,
Dikshit N, Pattanayak R, Shukla A, et al. 2013. Characterization of
winged bean (Psophocarpus tetragonolobus (L.) DC.) based on
molecular, chemical and physiological parameters. Am J Mol Biol 03
(4): 187-197.
Motior MR, Wan Mohamad WO, Wong KC, Shamsuddin ZH. 1998b.
Physiological studies of ratooned and non-ratooned winged bean
grown with and without support. In: Proceeding 9th Australian
Agronomy Conference. Wagga, NSW.
http;//www.rwgional.org.au/au/asa/1984/4/214motior.htm
Mukhopadhyay D. 2000. The molecular evolutionary history of a winged
bean α-chymotrypsin inhibitor and modeling of its mutations through
structural analyses. J Mol Evol 50 (3): 214-223.
Peres-Neto PR, Jackson DA, Somers KM. 2003. Giving meaningful
interpretation to ordination axes: Assessing loading significance in
principal component analysis. Ecology 84 (9): 2347-2363.
Permatasari N., Adiredjo A., Kuswanto. 2018. Evaluasi keragaman dan
potensi galur kecipir (Psopocarpus tetragonolobus L .) lokal hasil
koleksi. Jurnal Produksi Tanaman 6 (12): 3022-3031. [Indonesian]
PVTPP. 2014. Panduan pelaksanaan uji (PPU) keunikan, keseragaman
dan kestabilan kecipir winged bean (Psophocarpus tetragonolobus L).
Pusat Perlindungan Varietas Tanaman dan Perizinan Pertanian,
Kementerian Pertanian, Jakarta [Indonesian].
Schiavianto MA, Valio IFM. 1996. Influence of staking on development
of winged bean plants. R Bras Fisiol Veg 8 (2): 99-103.
Sharma V, Srinivas VR, Surolia A. 1996. Cloning and sequencing of
winged bean (Psophocarpus tetragonolobus) basic agglutinin (WBA
I): Presence of second glycosylation site and its implications in
quaternary structure. FEBS Lett. 389 (3): 289-292.
B I O D I V E R S I T A S
21 (7): 2991-3000, July 2020
3000
Tanzi AS, Eagleton GE, Ho WK, Wong QN, Mayes S, Massawe F. 2019.
Winged bean (Psophocarpus tetragonolobus (L.) DC.) for food and
nutritional security: synthesis of past research and future direction.
Planta. 250 (3): 911-931.
Tanzi AS, Ho WK, Massawe F, Mayes S. 2019. Development and
interaction between plant architecture and yield-related traits in
winged bean (Psophocarpus tetragonolobus (L.) DC.). Euphytica.
215 (2): 36. DOI: 10.1007/s10681-019-2359-8
Thompson AE, Haryono SK. 1979. Sources of resistance to two important
diseases of winged bean, Psophocarpus tetragonolobus (L.) DC.
HortScience. 14 (4): 532-533.
Thompson AE, Haryono SK. 1980. Winged bean: unexploited tropical
food crop. HortScience 15 (3): 233-237.
Wong KC. 1980. Agronomy of the winged bean in Malaysia. In: The
Winged Bean. Papers Presented in the 1st International Symposium
on Developing the Potentials of the Winged Bean, January 1978,
Manila. Philippine Council for Agriculture Resources Research, Los
Banos, Laguna, Philippines.
Wong KC. 1983. Effects of Daylength, Temperature, Light Intensity and
Applied Growth Substances on the Growth, Flowering and
Tuberization of Winged Bean (Psophocarpus tetragonolobus (L.)
DC). [Dissertation]. Universiti Putra Malaysia, Selangor.
