Yield Stability and Genotype × Environment Interaction of Faba Bean (Vicia faba L.)

Yield Stability and Genotype × Environment Interaction of Faba Bean (Vicia faba L.)
Yield Stability and Genotype × Environment
Interaction of Faba Bean (Vicia faba L.)
*1Mesfin Tadele, 2Wassu Mohammed and 1Mussa Jarso
1Holetta Agricultural Research Center, Ethiopian Institute of Agricultural Research, P. O. Box 2003, Addis Ababa, Ethiopia
2School of Plant Sciences, College of Agriculture and Environmental Sciences, Haramaya University, P.O. Box 138, Dire-
Dawa, Ethiopia
The present research was conducted to assess the effect of genotype × environment interaction
(GEI) on grain yield and determine yield stability of faba bean genotypes using 50 genotypes in
randomized complete block design with three replications tested at Holetta, Watebecha Minjaro
and Jeldu with and without lime application in 2017. The grain yield performances of genotypes
were varied across environments which indicate the existence of GEI. The mean grain yields of
genotypes were ranged between 51.16g (Wayu) and 96.40g (CS20DK) with an overall mean value
of 78.02g/5plants. The AMMI ANOVA showed that environment, genotype and GEI contributed
58.05, 16.08 and 14.28% of total variation in grain yield, respectively. The significant differences
among genotypes, environments and interaction effect of the two way interactions on grain yield
showed the differential response of genotypes over locations and managements and the test
environments were different each other. Based on mean grain yield, stability parameters from
AMMI and GGE-biplot, Tumsa, Cool-0034, EH07015-7 and EKLS/CSR02019-2-4 were identified as
the four most stable/relatively stable and productive genotypes whereas Wolki, Numan, EH09004-
2 and CS20DK had high grain yield and dynamic response to environments. Therefore, this
experiment has to be repeated for one more season for reliable recommendation.
Keywords: AMMI model, Faba bean, GGE-biplot, Grain yield, Stability
INTRODUCTION
Faba bean (Vicia faba L.) is the leading pulse crop in
Ethiopia in terms of area coverage and total production
(Mesfin et al., 2019). It mainly cultivated in the mid to high
altitude areas of the country, with elevations of 1800- 3000
m.a.s.l. (Mussa and Gemechu, 2006). As faba bean
genotypes grow in different environmental conditions they
response differently which is known as genotype ×
environment interaction (GEI), and it is important for
breeding program because it brings about differences in
the performance of a test material in several locations. The
GEI determines whether a genotype is widely adapted or
has specific adaptation. Differential responses of crop
varieties to variable environmental conditions limit
accurate yield estimates and identification of high yielding
stable ones. In order to identify stable genotypes, the GEI
can be evaluated using stability statistics that are
assignable to each genotype evaluated across a range of
environments (Fernandez, 1991). As environment is the
sum total of physical, chemical and biological factors that
influence the development of an organism (Dabholkar,
1992); and GEI is the difference between the phenotypic
value and the value expected from the corresponding
genotypic and environmental values (Baker, 1988).
Accordingly, lime application as acid soil management
affects the growth and performance of faba bean
genotypes which leads to lime treated and untreated
environment to be considered as separate environments.
*Corresponding Author: Mesfin Tadele, Holetta
Agricultural Research Center, Ethiopian Institute of
Agricultural Research, P. O. Box 2003, Addis Ababa,
Ethiopia
E-mail: mesfintadele64@gmail.com
Co-Author Email2
: wasmoha@yahoo.com
Co-Author E-mail1
: mjarso@yahoo.com
International Journal of Plant Breeding and Crop Science
Vol. 7(2), pp. 833-846, September, 2020. © www.premierpublishers.org, ISSN: 2167-
0449
Research Article

Mesfin T. 834
Conducting multi-environment trial enables to identify
adaptability of a crop variety and it is an important feature
of crop improvement that needs to be considered in a
breeding program to develop crop varieties for multi-
environments (Fekadu et al., 2012). The presence of a
significant GEI for quantitative traits, such as grain yield,
can seriously limit genetic gains in selecting superior
genotypes for the development of improved varieties
because environmental variation causes differential
genotypic responses that result in rank changes of
genotypes (Kang, 1990). The presence of significant GEI
reduces the association between genotype and phenotype
and thereby reduces the genetic advance; the best option
is either to exploit it by selecting superior genotypes for
specific environments or to avoid it by selecting widely
adapted and stable genotypes across a wide range of
environments (Ceccarelli, 1989). The determination of
grain yield stability of genotypes enables breeders for
cultivar recommendations despite the variable
environmental conditions (Yan et al., 2007). Therefore,
information on GEI is important to plant breeders for the
development, selection and recommendation of cultivars
that are suitable for different growth environments.
Various analysis methods have been used to explore GEI
and to identify superior genotypes with wide or specific
adaptation to different environments. The additive main
effects and multiplicative interaction (AMMI) model and the
genotype main effects and genotype × environment
interaction effects (GGE-biplot) model are the two most
frequently used models for statistical analyses of multi-
environment trials (Gauch et al., 2008). However, it is better
to use more than one stability statistics model to provide an
accurate picture because of the genotype’s multivariate
response to varying environments (Lin et al., 1986).
Where environmental differences are great, as in drained
and un-drained waterlogged Vertisols, it may be expected
that genotype by environment interaction effect is also high.
In such cases, care must be taken not to use statistical
models that omit inclusion of the GEI effects unless
otherwise proven that the performance of the given crop or
the given trait is not considerably influenced by
environmental fluctuations (Gemechu and Mussa, 2009).
Since faba bean is grown across a wide range of
environments in the highlands of Ethiopia, it is exposed to
the effect of GEI. However, there is little information
concerning the GEI and cultivar stability on faba bean in
Ethiopia (Fekadu et al., 2012). Hence, it is very essential to
study the nature and magnitude of GEI and stability of faba
bean genotypes in Ethiopia. The investigation of GEI in
multi-environment trials is thus important in the
development of soil acidity tolerant, high yielding and stable
faba bean genotypes.
The national and regional variety trials have been part of
the faba bean breeding program in national research
system for many years in Ethiopia. However, the genotype
by environment interaction and stability of faba bean
genotypes for yield under different levels of acid soils has
not been much studied and documented in Ethiopia.
Similarly, study on soil acidity problems and responses to
lime application have been done in some part of the
country, however, information on the stability of genotypes
under soil management across locations is scanty.
Therefore, identifying stable genotypes under acid soil
stresses and non-stress environments is of a paramount
importance for breeding faba bean genotypes adaptable to
acidic soils. Hence, this study was initiated to identify high
yielding and stable faba bean genotypes managed with and
without lime at different locations.
MATERIALS AND METHODS
Description of Experimental Sites
The experiment was conducted at six environments (Table
1) (three locations with and without lime application) at
Jeldu, Holetta and Watabecha Minjaro during 2017 main
cropping season under rain fed condition.
Table 1: Description of experimental areas
Locatio
n
Geographical
position
Altitud
e
(m.a.s
.l.)
Annu
al
rain
fall
(mm)
Temperat
ure (°C)
Soil pH
Longitu
de
Latitu
de
Min Max Befor
e
lime
Afte
r
lime
Jeldu 090
16'N
380
05'E
2800 1200 2.06 16.9 4.66 5.0
3
Holetta 090
00'N
380
30'E
2400 1072 6.6 24.1 4.49 4.8
0
Watebe
cha
Minjaro
090
05'N
38036'
E
2565 1100 8.7 23.3 4.94 5.0
8
Experimental Materials and Design
Fifty (50) faba bean genotypes including 22 released
varieties and 28 pipe line materials were used in this study.
The genotypes were collected from Holetta Agricultural
Research Center (HARC) and Kulumsa Agricultural
Research Centers (KARC) (Table 2). Additionally, a
product of Derba Cement Factory limestone (CaO)
collected from HARC were applied at the rate of 1.91, 1.65
and 5.29 t/ha at Holetta, Watebecha Minjaro and Jeldu,
respectively, one month ahead of planting on lime treated
sub-blocks at all the three locations depending the lime rate
requirement for each location. The experiment was
arranged in Randomized Complete Block Design with three
replications using adjacent block technique (growing the
two sets (lime treated and untreated) adjacent to each
other). Each block was divided into two adjacent sub-blocks
to accommodate both with and without lime plots. The
agronomic practices were carried out uniformly to all

genotypes as per the recommendations made by the
national research system for faba bean.
Table 2: Description of 50 faba bean genotypes used in the study
No. Genotypes Code Year of release Origin Seed source
1 Cool-0030 G1 --- Collection HARC
2 Wolki¥ G2 2008 Hybridization HARC
3 EKLS/CSR02012-2-3 G3 --- Hybridization KARC
4 Obse G4 2007 Hybridization HARC
5 NC58 G5 1978 Collection HARC
6 Ashebeka¥ G6 2015 Hybridization KARC
7 Hachalu¥ G7 2010 Hybridization HARC
8 Degaga G8 2002 Introduction HARC
9 EH09031-4 G9 --- Hybridization HARC
10 Holetta-2 G10 2001 Introduction HARC
13 EK05006-3 G13 --- Hybridization KARC
15 Numan G15 2016 Hybridization KARC
16 Bulga 70 G16 1994 Collection HARC
17 EK05001-1 G17 --- Hybridization KARC
18 Dosha G18 2008 Collection HARC
19 Gora G19 2012 Hybridization KARC
21 Wayu G21 2002 Collection HARC
23 Mesay G23 1995 Hybridization HARC
27 Kasa G27 1980 Collection HARC
34 EK 05037-4 G34 --- Hybridization KARC
36 KUSE2-27-33 G36 1979 Introduction HARC
39 Selale¥ G39 2002 Collection HARC
40 Moti G40 2006 Hybridization HARC
44 Tumsa G44 2010 Hybridization HARC
45 Gebelcho G45 2006 Hybridization HARC
46 EK05037-5 G46 --- Hybridization HARC
47 Didi’a¥ G47 2014 Hybridization KARC
49 CS20DK G49 1977 Collection HARC
50 Tesfa G50 1995 Introduction HARC
“---’’ = pipeline genotypes, ¥ =Varieties released for areas with waterlogging problems, HARC and KARC= Holeta and
Kulumsa Agricultural Research Center, respectively.

Mesfin T. 836
Data Collection and Analysis
The grain yield data were collected on faba bean genotypes
and adjusted to standard moisture content for pluses (10%).
The SAS computer package version 9.3 statistical software
(SAS Institute, 2010) was used to test for presence of
outliers and normality of residuals. All data were subjected
to analysis of variance (ANOVA) for RCBD as per the
procedure indicated by Gomez and Gomez (1984) using
SAS software (SAS Institute, 2010). The SAS GLM
(General Linear Model) procedure was employed for the
analysis of variance. Analysis of variance was conducted for
data collected from each location and management level
(with and without lime application) separately and
combined. For combined ANOVA, the homogeneity of error
variance was tested using the F-max method, which is
based on the ratio of the larger mean square of error (MSE)
from the separate analysis of variance to the smaller mean
square of error. When the ratio of larger error mean square
over the smaller error mean square is less than or equal to
3, the error variance is considered as homogeneous
(Gomez and Gomez, 1984).
F − ratio =
Larger MSE
Smaller MSE
…………………(1)
The grain yield data were analyzed using statistical
windows software GenStat 15th edition (Genstat, 2012).
AMMI and GGE-biplot stability parameters were
employed. Since AMMI model does not make provision for
a quantitative stability measure, AMMI stability value
(ASV) (Purchase, 1997) measure was computed in order
to quantify and rank genotypes according to their yield by
using Microsoft office excel 2010. The AMMI model
presented as follows:
𝑌𝑖𝑗 = 𝑚 + 𝐺𝑖 + 𝐸𝑗 + Σ λ 𝑘α𝑖𝑘 𝛾𝑗𝑘 + έ𝑖𝑗……………… (2)
Where 𝑌𝑖𝑗 is the yield of the ith genotype in the jth
environment; m is the grand mean; 𝐺𝑖 and 𝐸𝑗 are the
genotype and environment deviations from the grand mean,
respectively; λ 𝑘 is the eigenvalue of the PCA axis k; α𝑖𝑘 and
𝛾𝑗𝑘 are the genotype and environment principal component
scores for axis k; n is the number of principal components
retained in the model and έ𝑖𝑗 is the error term. According to
Gollob (1968) the df for the PC axis can be calculated as: df
= G + E -1- 2n
Where: G= genotype, E=environment and n=number of
IPCA axis Because the IPCA1 score contributes more to the
GEI sum of squares, a weighted value (ASV) was needed,
which was calculated according to the relative contribution
to the interaction by IPCA1 as compared to IPCA2
(Purchase, 1997):
ASV = √[
SSIPCA1
SSIPCA2
(IPCA1score)]
2
+ (IPCA2score)2 ……..…..(3)
Where SSIPCA1/SS IPCA2 = is the weight given to the IPCA1
value by dividing the IPCA1 sum of squares by the IPCA2
sum of squares, IPCA1 score is the IPCA1 score for that
specific genotype, and IPCA2 score is the IPCA2 score for
that specific genotype.
The GGE-biplot was used for analyzing GEI and stability of
the genotypes (Yan, 2001). The GGE-biplot approach is
preferred to AMMI since only G and GEI are important and
E is not important, and therefore only these components
must be simultaneously considered (Yan et al., 2007).
GGE-biplot best identifies GEI pattern of data and clearly
shows which variety performs best in which environments,
and thus facilitates mega-environment identification than
AMMI. Otherwise, both GGE and AMMI models are
equivalent as far as their accuracy is concerned (Fekadu et
al., 2012). The GGE-biplot model based on singular value
decomposition (SVD) of t principal components is given as
follows:
Yij - µi - βj= ∑k = 1λk αik γjk +↋ij……………(4)
Where, Yij is the performance of genotype i in environment j, µ is the
grand mean, βj the main effect of environment j, k is the number of
principal components (PC), λk is singular value of the kth PC, αik and
γjk are the score of ith genotype and jth environment, respectively for
PCk, ↋ij is the residual associated with genotype i in environment j.
Usually only the first two PCs are used especially if they account for
the major portion of the GEI..
RESULTS AND DISCUSSION
Mean Grain Yield Performances of Genotypes
The mean grain yield performances of 50 faba bean
genotypes over locations and managements (3 × 2 = 6
environments) indicated the presence of significant
variations among genotypes, locations and managements
for grain yield (Table 3). The result partially agrees with the
findings of Abebe and Tolera (2014) who reported
significant difference for grain yield and other traits as a
result of lime application on acid soils of western highlands
of Ethiopia. Many reports also showed the presence of
significant effects of G × E interaction on grain yield in faba
bean in different sets of environments in Ethiopia (Gemechu
and Mussa, 2009, Million and Habtamu, 2012; Tamene et
al., 2015).
The genotypes had overall mean grain yield (g/5plants) of
78.02g with the range between 51.16g (Wayu) and 96.40g
(CS20DK) over locations and managements. The
genotypes CS20DK and Moti had higher while Wayu had
lower grain yield (GY) over locations under two
managements with significant difference among the mean
values of genotypes. Among the evaluated genotypes, 56%
of them were yielded greater than the overall grand mean
(Table 3). The grain yield of Wayu was the least under each
management level, location and their interaction due to its
smaller hundred seeds weight. In agreement with this result
the older varieties (Kuse2-27-33, NC-58, Wayu and Selale)
were reported as consistently low yielder genotypes over
environments (Tamene, 2008) and CS20DK was high
yielder genotype over locations under optimum environments
(Tamene et al., 2015).

Regarding the locations, Jeldu was the lowest and highest
yielder environment without and with lime application,
respectively. The variation in the highest and lowest yield at
each location was a result of significant genotype by
management interaction. The variety Wayu was the least
yielding at separate and over environments. The GY varied
within genotypes with different managements, location and
their interaction. The highest average relative reduction of
grain yield was recorded on genotype G30 (43.03%) and
G36 (43.04%) while the smallest for G10 (16.87%) and it
varies 24.44 - 46.69% across test locations due to soil
acidity problem of test locations and lime application
improved GY for all genotypes over lime free condition
(Table 3). The grain yield difference with and without lime
application at each location indicated that the growing
environments were diverse and contributed for GY in
addition to genotypes. Similarly, previously reported that
liming significantly increased grain yield (Ouertatani et al.,
2011) and 32% yield increment as a result of lime
application reported in faba bean (Endalkachew et al.,
2018), 26% in common bean (Hirpa et al., 2013). In contrary
to the current result CS20DK was reported as the lowest
yielder variety as compared to Gora, Walki and Geblecho
(Degife and Kiya, 2016). Improvement in grain yield in low
acidity may be related to reduction of toxic levels of soil Al3+
and H+ ions by lime addition (Fageria et al., 2012) whereas,
low yields in acid soil could mainly be either due to the
deficiency of phosphorus, calcium and magnesium and
toxicity of aluminium, iron and manganese (Dodd and
Mallarino, 2005; Endalkachew et al., 2018).
Table 3: Mean grain yield (g/5plants) performance of 50 faba bean genotypes evaluated without and with lime,
combined over locations and managements at three locations in 2017 main season
No. Genotypes
Holetta Watebecha Minjaro Jeldu Mean
Without With without with Without with
1 Cool-0030 60.20h-o 97.12c-k 69.27e-m 78.23j-n 48.47f-l 112.43a-d 77.62e-m
2 Wolki 96.51a 101.65b-g 89.40ab 102.27a-f 40.80klm 87.20g-o 86.30bc
3 EKLS/CSR02012-2-3 66.78f-n 99.08c-i 61.27k-s 87.37f-m 56.43b-j 109.17a-f 80.02c-i
4 Obse 80.80a-g 100.37b-g 78.70a-f 84.60g-n 73.17a 108.20a-g 87.64b
5 NC58 64.73g-n 82.65g-m 69.90d-m 84.13g-n 50.30d-l 79.87k-q 71.93lmn
6 Ashebeka 80.60a-g 99.25c-i 63.20i-s 94.53d-i 64.23a-e 102.63a-j 84.08b-e
7 Hachalu 80.23a-g 90.74d-l 75.60c-h 95.97c-i 62.90a-f 82.45j-q 81.32b-h
8 Degaga 72.78c-k 84.68f-l 60.93k-s 92.43e-j 45.13i-l 84.67h-p 73.44i-n
9 EH09031-4 68.93d-m 89.43d-l 65.50h-r 88.07f-m 50.37d-l 88.30f-o 75.10g-m
10 Holetta-2 59.99h-o 65.15mno 54.20rs 63.27o 45.57i-l 63.60qr 58.63q
11 EH09007-4 56.41k-o 85.38e-l 66.10g-r 86.37g-m 41.80j-m 96.27b-l 72.06lmn
12 EH07023-3 68.06e-n 91.62d-l 67.73e-o 93.53e-i 61.43a-h 122.73a 84.19b-e
13 EK05006-3 93.37ab 89.33d-l 54.18rs 90.77e-k 62.47a-g 117.07ab 84.53b-e
14 EKLS/CSR02014-2-4 56.86k-o 82.78g-m 72.00c-l 81.17i-n 59.43a-i 103.03a-j 75.88f-m
15 Numan 86.65a-d 84.85f-l 72.67c-k 78.20j-n 58.93a-i 110.97a-e 82.05b-g
16 Bulga 70 66.07g-n 86.32e-l 69.90d-m 90.53e-l 30.10mn 90.53e-n 72.24k-n
17 EK05001-1 66.83f-n 80.05i-n 71.33c-l 75.70l-o 49.63e-l 107.50a-g 75.17g-m
18 Dosha 85.21a-e 89.71d-l 66.93f-q 93.47e-i 70.00ab 111.03a-e 86.06bc
19 Gora 69.21d-m 96.66d-k 78.00b-g 92.2e-k 65.77abc 110.97a-e 85.47bcd
20 EH08035-1 72.81c-k 104.63a-f 56.33o-s 103.33a-e 43.90i-m 90.03e-o 78.51d-l
21 Wayu 43.59o 58.80o 53.27s 73.97mno 25.30n 52.03r 51.16r
22 EKLS/CSR02023-2-1 58.35j-o 85.28e-l 55.17p-s 95.38d-i 47.77f-l 93.77d-m 72.62j-n
23 Mesay 60.58h-o 75.65l-o 60.13l-s 77.33k-n 44.80i-m 83.50i-p 67.00nop
24 EH09004-2 75.38c-j 122.95a 67.87e-o 96.37c-i 37.23lmn 78.27l-q 79.68c-j
25 EH06088-6 52.69l-o 96.89d-k 61.43k-s 87.33f-m 65.90abc 103.40a-j 77.94e-m
26 EKLS/CSR02017-3-4 68.53d-n 94.17d-l 61.53k-s 95.23d-i 49.03e-l 113.83a-d 80.39c-i
27 Kasa 50.51no 80.23h-n 58.27m-s 74.80mno 45.50i-l 72.17n-q 63.58pq
28 Cool-0025 72.95c-k 101.37b-g 65.50h-r 90.33e-l 49.00e-l 103.80a-i 80.49b-i
29 EH06070-3 70.23d-l 85.55e-l 56.90n-s 73.47mno 44.63i-m 94.40d-m 70.86mno
30 EKLS/CSR02010-4-3 70.60d-l 109.16a-d 67.60e-o 112.70ab 47.90f-l 103.57a-i 85.25bcd
31 Cool-0031 59.10i-o 97.57c-k 72.10c-l 101.93a-f 48.60f-l 95.77c-l 79.18c-k
32 Cool-0018 70.34d-l 100.98b-g 68.13e-o 82.50h-n 45.90h-l 109.57a-e 79.57c-j
33 EKLS/CSR02028-1-1 77.09b-i 94.85d-l 72.33c-l 87.87f-m 44.90i-m 113.40a-d 81.74b-g
34 EK 05037-4 69.13d-m 104.96a-e 79.67a-e 97.17c-h 51.50c-l 100.50b-k 83.82b-e

35 Cool-0035 66.55f-n 98.27c-i 83.30abc 95.63d-i 54.77b-k 98.17b-l 82.78b-f
36 KUSE2-27-33 65.09g-n 94.28d-l 58.30m-s 95.53d-i 39.97klm 97.90b-l 75.18g-m
37 EH07015-7 79.17a-g 107.09a-d 54.77qrs 92.97e-j 65.07a-d 101.87a-j 83.49b-e
38 Cool-0024 79.49a-g 100.15b-h 74.00c-j 109.27a-d 50.57c-l 100.33b-k 85.64bcd
39 Selale 51.53mno 62.52no 60.73k-s 97.33c-h 37.97lmn 64.87pqr 62.49pq
40 Moti 84.53a-f 119.24ab 75.40c-i 110.50abc 58.30a-i 115.57abc 93.92a
41 EH06027-2 72.05d-k 83.06g-m 67.27f-p 92.90e-j 47.00g-l 84.30i-p 74.42h-m
42 EKLS/CSR02019-2-4 73.56c-k 107.17a-d 69.07e-n 89.90e-l 57.33b-j 96.07b-l 82.18b-g
43 EH09002-1 76.79b-i 77.60k-n 61.97j-s 84.93g-n 43.83i-m 97.77b-l 73.82i-n
44 Tumsa 77.51b-h 97.49c-k 78.07b-g 99.23b-g 57.17b-j 106.93a-g 86.07bc
45 Gebelcho 66.59f-n 95.03d-l 71.83c-l 91.33e-k 57.23b-j 74.40m-q 76.07f-m
46 EK05037-5 55.92k-o 78.14j-n 60.50k-s 81.37i-n 42.47j-m 85.73h-o 67.36nop
47 Didi’a 90.60abc 97.93c-j 81.97a-d 110.37abc 50.90c-l 94.23d-m 87.67b
48 Cool-0034 64.18g-n 94.49d-l 79.77a-e 99.17b-g 57.17b-j 105.60a-h 83.40b-e
49 CS20DK 85.81a-e 116.57abc 89.97a 115.63a 62.90a-f 107.53a-g 96.40a
50 Tesfa 57.44j-o 92.17d-l 52.87s 71.20no 46.63h-l 70.07o-r 65.06op
Mean 69.98c 92.62b 67.66c 90.80b 51.16d 95.96a 78.02
CV 12.82 10.63 9.03 8.25 15.01 18.09 13.15
R2 0.73 0.74 0.78 0.77 0.72 0.36 0.85
CV, R2 = coefficient of variation and determination, Mean values followed by similar letter(s) in each column had non-
significant difference at P<0.05.
Genotype × Environment Interaction and Yield Stability
of Genotypes
The genotypes showed differential performance across all
environments of the testing sites, which means the
genotypes reacted differently to different environmental
conditions resulted in performance variation of the
genotypes thereby showed Genotype by Environment
Interaction (GEI). When the expression of the genetic
potential of the genotype is influenced by the environmental
factors, screening of genotypes with higher stability is a very
important breeding strategy.
Analysis of Variance from AMMI Model
The analysis of variance from AMMI model showed that
environments (E), genotypes (G) and genotype ×
environment interaction (GEI) had significant effect on grain
yield of 50 faba bean genotypes. This model has been
regarded as a powerful analytical tool while dealing with
large GEI data sets and it provide the relative contribution
of factors to the total sum squares (Gauch, 1992).
Accordingly, the environment accounted for 58.05% of the
total sum of squares while the genotype and GEI accounted
for 16.08% and 14.28%, respectively (Table 4). The
magnitude of environment was 4.1 times greater than the
GEI. From this result, the large sum of squares for
environments indicated that the environments were diverse,
with large differences among environmental means causing
most of the variation in seed yield and influence faba bean
production. Likewise, high environmental contributions (48
to 88%) to grain yield variability of faba bean in Ethiopia
have been reported by other authors (Mulusew et al., 2008;
Tamene, 2015; Teklay et al., 2015; Asnakech et al., 2017).
Therefore, different genotypes need to be evaluated over
locations to determine their performance across
environments.
The AMMI analysis of variance indicated that the mean
squares of the first four IPCA scores and residual were
significant (P≤0.01). The first principal component axis
(IPCA1) of the interaction captured 37.81% of the
interaction sum of squares. The first two IPCAs (IPCA1 and
IPCA2) together contributed 60.07% of the total GEI sum
of squares (Table 4). It has been reported that 5 0 t o 77%
of the first IPCA score contribution in faba bean
genotypes (Mulusew et al., 2008; Teklay et al., 2015;
Asnakech et al., 2017). Tamene (2015) also reported 66.6%
of contribution of the first two IPCAs to GEI sum square.
The highly significant (P ≤ 0.01) mean squares of GEI
indicated that the grain yield of tested genotypes varied
across environments. Therefore, apart to the effects of
environments and genotypes the GEI also attributed for the
differential yield performance of genotypes over
environments. It was reported in grain yield of chickpea that
AMMI analysis showed significant (p≤0.01) GEI indicating
the presence of genetic variation and possible selection of
stable entries (Rashidi et al., 2013).
Table 4;AMMI analysis of variance for grain yield
(g/5plants) of 50 faba bean genotypes tested at six
environments (three locations with and without lime
applications) in 2017
Source of
variation
DF SS MS %Variance
Explained
%
Cumulative
Total 899 411475 458
Treatment 299 363728 1216**
Genotype 49 66124 1349** 16.08
Environment 5 238846 47769** 58.05
Block 12 4836 403**
Interaction 245 58758 240** 14.28
IPCA1 53 22214 419** 37.81
IPCA2 51 13082 257** 22.26 60.07
IPCA3 49 9191 188** 15.64 75.71
IPCA4 47 8219 175** 13.99 89.70

Residuals 45 6052 134
Error 588 42911 73
** Significant difference at (P≤0.01), DF= degree of
freedom, SS= sum of square, MS= mean squares, IPCA=
Interaction principal component axis
Yield Stability of Genotypes from AMMI Model
According to AMMI model, genotypes with large IPCA
scores regardless of their positive or negative sign is an
indication of instability of the genotypes or their specific
adaptibility whereas the small scores close to zero have
small interactions and are stable (Zobel et al., 1988).
Accordingly, ten genotypes (G48, G44, G9, G37, G8, G42,
G3, G39, G45 and G1) that had small IPCA scores close to
zero identified more stable for yield whereas seven
genotypes (G2, G32, G24, G20, G4, G30 and G15) (Table
5) had high IPCA scores showing their instability over
locations and soil acidity managements. A previous report
indicated that, genotypes with means greater than grand
mean and IPCA score nearly zero are considered as
generally adaptable to all environment whereas genotypes
with larger IPCA scores are adapted to specific
environments (Rashidi et al., 2013).
Genotypes with least ASV or have smallest distance from
the origin are considered as the most stable, where as those
which have highest ASV are considered as unstable
(Purchase, 1997). The finding in this research indicated that
different genotypes showed different stability results based
on their ASV values for different locations with different soil
managements. Accordingly, G44, G48, G9, G39, G3, G37,
G45, G42, G1 and G8 had smaller ASV values ranked as
the 1st to 10th indicating these genotypes were most stable
whereas G49, G2, G17, G13, G24, G6, G15, G23, G18 and
G25 with their higher ASV considered as the least stable for
grain yield performance across the testing environments
(Table 5). Based on genotypes that had grain yield greater
than the overall mean and lower values of IPCAs and ASV
values G44, G48, G3, G37 and G42 were stable while G2,
G15 G24 and G49 found unstable. Stable genotypes show
more or less similar yield performance over environments
whereas unstable genotypes perform differently.
Table 5: Stability parameters for grain yield (g/5plants) of 50 faba bean genotypes from AMMI model analyses at
six environments (three locations with and without lime) in 2017
Geno Pooled mean AMMI model stability Geno Pooled mean AMMI model stability
IPCA1 IPCA2 ASV IPCA1 IPCA2 ASV
G1 77.62(30) -0.61 0.42 0.61(42) G26 80.39(23) -0.30 1.06 1.05(30)
G2 86.30(5) 4.38 -2.08 6.00(2) G27 63.58(47) -0.59 -0.32 0.88(36)
G3 80.02(24) -0.40 0.26 0.43(46) G28 80.49(22) -0.20 0.95 0.95(34)
G4 87.64(4) -1.54 -1.44 2.15(15) G29 70.86(43) -0.99 -0.22 2.13(16)
G5 71.93(42) -0.06 -1.01 1.01(31) G30 85.25(10) 1.13 2.35 2.48(12)
G6 84.08(13) -1.10 0.09 3.78(6) G31 79.18(27) 1.16 0.83 1.61(24)
G7 81.32(21) 1.15 -2.30 2.15(14) G32 79.57(26) -1.55 1.72 0.88(37)
G8 73.44(38) 0.27 -0.05 0.64(40) G33 81.74(20) 0.57 -0.19 0.96(33)
G9 75.10(35) -0.16 -0.17 0.23(48) G34 83.82(14) 0.21 1.01 1.01(32)
G10 58.63(49) -1.19 -2.10 2.28(13) G35 82.78(17) 0.59 -0.23 0.92(35)
G11 72.06(41) -0.18 0.65 0.64(41) G36 75.18(33) 1.61 0.70 2.53(11)
G12 84.19(12) 0.53 -1.24 1.19(26) G37 83.49(15) -0.26 0.06 0.51(44)
G13 84.53(11) -1.96 -0.34 4.73(4) G38 85.64(8) 0.84 0.94 1.23(25)
G14 75.88(32) -0.41 -1.65 1.67(20) G39 62.49(48) -0.32 0.46 0.37(47)
G15 82.05(19) 0.31 -3.19 3.19(7) G40 93.92(2) 0.98 1.43 1.65(22)
G16 72.24(40) 0.01 1.65 1.65(23) G41 74.42(36) -0.94 0.50 1.19(27)
G17 75.17(34) -2.02 -0.37 4.74(3) G42 82.18(18) 0.29 0.08 0.57(43)
G18 86.06(7) -1.89 -0.90 2.88(9) G43 73.82(37) 0.14 -1.07 1.07(29)
G19 85.47(9) 1.10 -2.06 1.90(18) G44 86.07(6) 0.12 -0.14 0.08(50)
G20 78.51(28) 1.33 1.70 2.07(17) G45 76.07(31) 0.24 -0.54 0.51(45)
G21 51.16(50) -0.53 0.25 0.73(39) G46 67.36(44) -1.25 0.60 1.70(19)
G22 72.62(39) -0.70 1.28 1.17(28) G47 87.67(3) 0.51 0.55 0.74(38)
G23 67.00(45) -1.30 -0.22 3.17(8) G48 83.40(16) 0.09 0.04 0.15(49)
G24 79.68(25) 2.94 1.69 4.23(5) G49 96.40(1) 2.4 -0.28 7.02(1)
G25 77.94(29) -1.22 0.22 2.84(10) G50 65.06(46) -1.22 0.59 1.66(21)
Geno = genotype, AMMI= Additive main effect and multiplicative interaction, IPCA= interaction principal component axis,
ASV= AMMI stability value, number in parenthesis are ranks of genotypes.

Mesfin T. 840
Regarding the environments the average grain yield for the
genotypes over environments ranged from 51.15g at E5 to
95.95g at E6. The G49 was selected as 1st ranking genotype
at E3 and E4 and 3rd ranking at E2. The G2 also identified
as 1st ranking genotype at E1 2nd and 3rd ranking at E4 and
E3, respectively. The G40 also identified as 1st and 3rd
ranking genotype at three environments (E2, E4 and E6)
and G47 identified as 2nd, 3rd and 4th ranking genotype at
E3, E1 and E4, respectively (Table 6). The AMMI first four
ranking selections over six environments indicated that
G49, G2, G40 and G47 were better performing genotype at
half of the environments than others and these genotypes
were performed good both under lime and without lime
application except G40.
Table 6: The first four AMMI selections for grain yield
(g/5plants) per environment
Environment
ID
Environment Mean
GY
1st
2nd
3rd
4th
E1 Holetta
unlimed
69.98 G2 G13 G47 G15
E2 Holetta limed 92.62 G40 G24 G49 G30
E3 Watebecha
Minjaro
unlimed
67.64 G49 G47 G2 G48
E4 Watebecha
Minjaro limed
90.79 G49 G2 G40 G47
E5 Jeldu unlimed 51.15 G4 G18 G19 G25
E6 Jeldu limed 95.95 G12 G13 G40 G26
GY= grain yield
The stability of genotypes for grain yield over locations
alone may not allow making a decision as the genotypes
are worth for production. For instances, only genotype G44
had six ranking mean grain yield among the first 10 stable
genotypes identified by low IPCA1 and IPCA2 scores, and
among the first 10 stable genotypes identified by low ASV.
Therefore, it is necessary to consider all the stability
parameters along with the mean grain yield. Considering
the IPCAs scores, ASV and the first four AMMI selections
along with mean grain yield, G49, G40, G47 and G2 had
high mean grain yield with 1-5th ranks and identified as the
first four selections of AMMI each at three environments
with contrasting managements (with and without lime
applications) except G40 . In addition, G48 and G44 had
16th and 6th ranking mean yield and identified as the first two
most stable genotypes by ASV, IPCA1 and IPCA2 scores.
G18 was the 7th ranking for mean grain yield and 9th stable
genotype identified by ASV. G48 was identified among the
first four AMMI selections at one environment; however,
G44 was not identified among the first four AMMI selections.
The genotype G4 had the 4th ranking mean grain yield,
identified as one of the first four AMMI selections over six
environments at one environment but not identified among
the first 10 stable genotypes by IPCAs scores and ASV. In
contrast, the genotypes G9, G3 and G45 identified as 3rd,
5th and 6th stable genotypes, respectively, by ASV and
IPCAs scores while G37 and G42 identified as 7th and 8th
ranking stable genotypes by ASV. However, the genotypes
had mean grain yield in the rank between 15 and 35 among
50 faba bean genotypes (Table 5 and Table 6).
In choosing superior genotypes, a low or minimal genotype
× environment interaction must exist (Cotes et al., 2002).
However, if genotypes had varied stability for grain yield, it
is necessary to consider the stability parameters along with
high performance but the varieties can be responsive to
changing environments (dynamic stability) (Yan and Kang,
2003). Therefore, G49, G40, G47, G2, G48, G44, G18 and
G4 may be considered as worthy genotypes for high grain
yield over environments than G9, G3, G45, G37 and G42
identified as stable genotypes with lower grain yield than the
former genotypes.
“Which-Won-Where” Patterns and Stability of
Genotypes
A polygon view of the GGE-biplot for grain yield resulted in
eight vertex genotypes with both positive (high yielding) and
negative (low yielding) PCA1 scores. Eight genotypes G21,
G24, G2, G49, G40, G12, G17 and G10 that located on the
vertices of the polygon performed either the best or the
poorest in one or more environments (Figure 1). The GGE-
biplot analysis provided a visual interpretation of the
relationship among the genotypes and test environments.
As stated by Asnakech et al. (2017) when the environments
fell in different sectors of the polygon view of the GGE-biplot
it indicated the best genotype at one environment may not
perform best at another environment. Likewise, the
environments fell in two different sectors of the polygon
view.
Environments with large PC1 scores are those
environments that better discriminate among genotypes
and those with PC2 scores near zero are more
representative of an average environment (Yan, 2001). In
this study the environments, E2 and E6 had larger PC1
scores and well discriminated among the genotypes.
According to Yan (2001) genotypes at the apex of each
sector are the best performing at environments included in
that sector if the GGE is sufficiently approximated by PC1
and PC2. As shown in Figure 1, PC1 and PC2 accounted
73.53% of the variation of the total PCs for grain yield over
six environments showing that they had sufficiently
explained the GGE. Accordingly, genotypes G40, G49 and
G2 were the best performing genotypes at Holetta and
Watebecha Minjaro at both lime levels and G12 was the
best performer at Jeldu whereas the other four vertex
genotypes G24, G21, G10 and G17 fell in sectors with no
environment markers. So this indicates these genotypes are
low yielding genotypes in at least one environment.

Int. J. Plant. Breed Crop Sci. 841
Therefore, genotypes with environmental markers can be
recommended for adaptation to those specific
environments. However, first the stability of genotypes over
environments should be considered.
Large positive PC1 scores for genotypes indicated that
those genotypes had higher average yield and PC2 scores
near zero indicated that those genotypes were more stable
(Yan, 2001; Fekadu et al. 2012). Accordingly, genotypes
G49, G40, G44, G48, G47, G30, G42, G6, G37 and G38
were high yielding genotypes. On the other hand,
genotypes G21, G10, G39, G27, G50, G23 and G46 were
with large negative PC1 scores and they were low yielding
genotypes. Genotypes with relatively low PC2 scores such
as G44, G48, G28, G42, G6, G37, G22, G36, G9, G46, G27,
G50 and G10 can be considered as relatively stable.
However, among these genotypes, only G44, G48, G28,
G42, G6 and G37 were high yielding and should be
considered as stable for recommendation (Figure 1).
Figure 1: GGE-biplot showing environments and their respective faba bean genotypes. (G= genotype and G1-G50
listed in Table 2, E= environment and E1-E6 listed in Table 6)
Evaluation of Environments and Ranking of Genotypes
over Environments
The GGE-biplots showing the discriminating ability and
representativeness of the test environments were
presented in Figure 2. The average environment is
represented by the small circle at the end of the arrow and
contains the average coordinates of all test environments
(Yan and Tinker, 2005). As suggested by Yan (2001)
environments with short vector length have the smallest
angle with the “Average-Environment Coordinate” (AEC)
indicating that it is more representative of the test
environments. The concentric circles aid in the visualization
of the length of the environment vectors. E6 had the longest
vector from the biplot origin indicating it was the most
discriminating of the test environments In general based on
the representativeness of test environments, E1 was the
most representative of the environments for grain yield
followed by E2, E3 and E4. In contrast, E5 and E6 were the
least representative environments. However, E6 was the
best discriminating (informative) of the genotypes and it was
the least representative of the test environments. An ideal
test environment should effectively discriminate genotypes
and represent the environments (Yan and Kang, 2002).
According to Yan and Tinker (2005), environments that give
little information on genotypes (Non-discriminating) should
not be used as test environments. Thus, in this study
among all the six environments, E2 represented the ideal
test environment (in the first concentric circle) with
moderate discriminating ability of the genotypes and
representativeness of the test environment for faba bean
grain yield. This environment can be used for selecting
generally adapted genotypes (Figure 2).

Mesfin T. 842
Figure 2: Comparison of environments based on discriminating ability and representativeness. (G= genotype and
G1-G50 listed in Table 2, E= environment and E1-E6 listed in Table 6)
The line that passes through the biplot origin is called the
average environment coordinate (AEC), and it shows the
stability of the genotypes (Yan, 2001). The stability of the
genotypes is measured by their projection to the AEC y-axis
(A line). That means, the greater the absolute length of the
projection of a genotype, the less stable it is or the shorter
the absolute length, the more stable it is (Yan, 2001). The A
line separates genotypes with grain yield below the grand
mean and above the grand mean. Those genotypes to the
right of this line were high yielders while those to the left
were low yielders. The single-arrow on the AEC points to
higher mean yield. Accordingly, G49 had the highest yield,
followed by G40 while G21 is the poorest genotype for grain
yield. The double-arrowed line is the AEC ordinate that
points in either direction to greater variability (least stability).
Accordingly, G44, G6, G37, G28, G48, G31, G35, G42,
G34, G9, G36, G22, G46, G27, G50 and G10 were the most
stable genotypes but only G44, G6, G37, G28, G48, G31,
G35, G42 and G34 with above average performance while
genotypes G24, G2, G47, G15, G12 and G13 were the least
stable but high yielding (Figure 3).
Figure 3.: Mean yield performance and stability of genotypes over environments. (G= genotype and G1-G50 listed
in Table 2, E= environment and E1-E6 listed in Table 6)

Stability and Mean Grain Yield of Genotypes
A total of 28 genotypes had mean yield greater than overall
mean yield of genotypes (>78.02 g/5plants) of which 18
genotypes had yield advantage of 0.16 to 17.49% over the
recently released variety (Numan, G15) of which eight were
released varieties. The two varieties CS20DK (G49) and
Moti (G40) released in 1977 and 2006, respectively, had the
highest grain yield advantages of 17.49 and 14.47%,
respectively (Table 7). The genotypes yield performance
and yield stability judged by AMMI and GGE-biplot
parameters match each other. The genotypes, G44 and
G49 were identified as most stable and unstable,
respectively, by ASV and GGE-biplot parameters, but G49
and G44 had first and sixth ranking high mean yield with
17.49 and 4.9% advantages over the recently released
variety, respectively. In addition, G49 was identified among
the first four AMMI selections at three environments both
under lime and without lime applications over two locations
but G44 not identified among the first four AMMI selections.
The other two genotypes, G48 and G2 were identified as 2nd
most stable and unstable genotypes, respectively, by ASV
and GGE-biplot parameters, however, G2 and G48 had 5th
and 16th ranking mean yield with 5.18 and 1.65% yield
advantages over the recently released variety, respectively.
In addition, G2 was identified among the first four AMMI
selections at three environments both under lime and
without lime applications over two locations but G48 was
identified among the first four AMMI selections at one
environment E3 (Table 6). The yield rank performances of
genotypes were inconsistent across environments.
Similarly, it was reported that no single cultivar that showed
superior performance over all environments (Tamene,
2015).
The three genotypes, G40, G47 and G4 were not identified
among the first 10 stable genotypes by AMMI and GGE-
biplot parameters, but had mean yield advantages of 14.47,
6.85 and 6.81%, respectively, over recently released variety
(Numan, 2016). The mean yield rank of G40 and G47 was
2nd and 3rd, respectively, whereas G4 had the 4th rank
among 50 genotypes. G40 and G47 are identified among
the first four AMMI selections at three environments under
lime and both under lime and without lime applications over
three and two environments, respectively, while G4
identified among the first four AMMI selections at one
environment E5. The two genotypes, G37 and G42 were
identified as 7th and 8th stable genotypes, respectively, by
ASV and had relatively low value of IPC2 of GGE-biplot
indicating the genotypes were stable for yield (Table 7).
Low or minimal genotype × environment interaction must
exist to identify superior genotypes (Cotes et al., 2002),
however, if genotypes yield stability and mean yield rank
varied, it is necessary to consider both the stability
parameters and high performance over environments but
they can be responsive to changing environments (Yan and
Kang, 2003). The main problem with stability statistics is
that a single model cannot provide an accurate picture
because of the genotype’s multivariate response to varying
environments (Lin et al., 1986). Therefore, it is necessary
simultaneous consideration of mean yield, stability of
genotypes evaluated by two or more stability parameters of
different models. Accordingly, G49, G40 G47 and G2 could
be recommended as high yielding genotypes over locations
and soil acidity managements whereas G4, G18 and G19
as better performing genotypes at specific locations (E5 and
E6) and G48 at E3 could be considered as high yielding.
The two genotypes, G30 and G12 also could be considered
as good yielder at E2 and E6, respectively (Table 6).
Generally, based on AMMI and GGE-biplot stability
parameters among the evaluated 50 faba bean genotypes
Tumsa (G44), Cool-0034 (G48), EH07015-7 (G37) and
EKLS/CSR02019-2-4 (G42) were identified as the four most
stable or relatively stable and productive genotypes. These
genotypes may perform more or less similarly across
environments thereby considered as wide adaptable
whereas G2, G15, G24 and G49 had a combination of high
yield, dynamic response to environments. Therefore, the
pipeline genotypes G48, G37 and G42 needs to be
evaluated for one more season to recommend for
commercial release considering their stability, and better
yield performance (Table 7). In line with this result, based
on AMMI stability model Tumsa (G44) was reported as high
yielding and the most stable variety across environments
(Tamene, 2015).
Table 7: List of 28 genotypes with mean yield greater the overall mean yield of genotypes, and stability by AMMI
and GGE models
No. Geno Yield
% advantage over AMMI GGE-biplot
Mean Numan IPCA1 IPCA2 ASV AMMI
Selec.
PC2 double-
arrow
1 G49 96.40(1) 23.56 17.49 2.4 -0.28 7.02(1) 3 relatively high Above
2 G40 93.92(2) 20.38 14.47 0.98 1.43 1.65(22) 3 relatively low Above
3 G47 87.67(3) 12.37 6.85 0.51 0.55 0.74(38) 3 relatively high Above
4 G4 87.64(4) 12.33 6.81 -1.54 -1.44 2.15(15) 1 relatively low Above
5 G2 86.30(5) 10.61 5.18 4.38 -2.08 6.00(2) 3 relatively high Below
6 G44 86.07(6) 10.32 4.90 0.12 -0.14 0.08(50) No relatively low Above

7 G18 86.06(7) 10.31 4.89 -1.89 -0.9 2.88(9) 1 relatively low Below
8 G38 85.64(8) 9.77 4.38 0.84 0.94 1.23(25) No relatively low Above
9 G19 85.47(9) 9.55 4.17 1.1 -2.06 1.90(18) 1 relatively low Below
10 G30 85.25(10) 9.27 3.90 1.13 2.35 2.48(12) 1 relatively low Above
11 G13 84.53(11) 8.34 3.02 -1.96 -0.34 4.73(4) 2 relatively high Below
13 G6 84.08(13) 7.77 2.47 -1.1 0.09 3.78(6) No relatively low Above
15 G37 83.49(15) 7.01 1.76 -0.26 0.06 0.51(44) No relatively low Below
16 G48 83.40(16) 6.90 1.65 0.09 0.04 0.15(49) 1 relatively low Below
17 G35 82.78(17) 6.10 0.89 0.59 -0.23 0.92(35) No relatively low Above
20 G33 81.74(20) 4.77 -0.38 0.57 -0.19 0.96(33) No relatively low Above
21 G7 81.32(21) 4.23 -0.89 1.15 -2.3 2.15(14) No relatively low Above
22 G28 80.49(22) 3.17 -1.90 -0.2 0.95 0.95(34) No relatively low Below
23 G26 80.39(23) 3.04 -2.02 -0.3 1.06 1.05(30) 1 relatively low Below
25 G24 79.68(25) 2.13 -2.89 2.94 1.69 4.23(5) 1 relatively high Above
27 G31 79.18(27) 1.49 -3.50 1.16 0.83 1.61(24) No relatively low Above
28 G20 78.51(28) 0.63 -4.31 1.33 1.7 2.07(17) No relatively low Above
AMMI= Additive main effect and multiplicative interaction, ASV= Ammi stability value, IPCA= interaction principal component
axis
CONCLUSION
The research results, allowed to concluded that based on
AMMI and GGE-biplot stability parameters among the
evaluated 50 faba bean genotypes Tumsa (G44), Cool-
0034 (G48), EH07015-7 (G37) and EKLS/CSR02019-2-4
(G42) were identified as the four most stable or relatively
stable and productive genotypes that may perform more or
less similarly across environments thereby considered as
wide adaptable contrary to this genotypes G2, G15, G24
and G49 had high grain yield, dynamic response to
environments. The differential performance of genotypes
over locations and managements suggested the evaluation
of genotypes over locations with and without lime
application in a future breeding activity to identify genotypes
tolerant to acid soils. However, the experiment was
conducted for one season and the possible effect of season
on yield of genotypes was not assessed. Therefore, it
needs to evaluate at least for one more seasons to make a
reliable and conclusive recommendation for commercial
release considering their stability, and better yield
performance.
ACKNOWLEDGMENTS
The authors would like to express their deepest gratitude to
staff members of Holetta Agricultural Research Center,
particularly highland pulse breeding program and soil
laboratory for their valuable contribution for successful
accomplishment of this research. We are thankful to the
Ethiopian Institute of Agricultural Research for financial
support.
CONFLICT OF INTEREST
The authors declare that they have no conflict of interests.
REFERENCES
Abebe Z, Tolera A. (2014). Yield response of faba bean to
fertilizer rate, rhizobium inoculation and lime rate at
Gedo highland, western Ethiopia. Glob. J. Crop Soil Sci.
Plant Breed, 2 (1): 134-139.
Asnakech T, Julia S, John D, Asnake F. (2017). Analysis of
genotype x environment interaction and stability for
grain yield and chocolate spot (Botrytis fabae) disease
resistance in faba bean (Vicia faba L). Australian Journal
of Crop Science, 11 (10): 1228-1235.
Baker RJ. (1988). Differential response to environmental
stress. pp. 492-504, In: Weir BS, Eisen EJ, Goodman
MM, Namkoong G (eds). The Second International
Conference on Quantitative Genetics Proceedings.
Sinauer Sunderland, Massachusettes.
Ceccarelli S. (1989). Wide adaptation: How wide?
Euphytica, 40: 197-205.
Cotes JM, Nustez CE, Martinex R, Estrada N. (2002).
Analyzing genotype by environment interaction in potato

using yield-stability index. American Journal of Potato
Research, 79: 211-218.
Dabholkar AR. (1992). Elements of biometrical genetics.
New Delhi 110059: Concept Publishing company.
Degife AZ, Kiya AT. (2016). Evaluation of Faba Bean (Vicia
faba L.) Varieties for yield at Gircha Research Center,
Gamo Gofa Zone, Southern Ethiopia. Scholarly Journal
of Agricultural Science, 6 (6): 169-176.
Dodd JR, Mallarino AP. (2005). Soil-test phosphorus and
crop grain yield responses to long-term phosphorus
fertilization for corn-soybean rotations. Soil Science
Society of American Journal, 69:1118–1128.
Endalkachew F, Kibebew K, Asmare M, Bobe B. (2018).
Yield of faba bean (Vicia faba L.) as affected by lime,
mineral P, farmyard manure, compost and rhizobium in
acid soil of lay gayint district, northwestern highlands of
Ethiopia. Agriculture and Food Security, 7 (16): 1-11.
Fageria NK, Baligar VC, Melo LC, de Oliveira JP. (2012).
Differential Soil Acidity Tolerance of Dry Bean
Genotypes. Communications in Soil Science and Plant
Analysis 43(11): 1523-153.
Fekadu G, Ersulo L, Asrat A, Fitsum A, Yeyis R, Daniel A.
(2012). GGE-biplot Analysis of Grain Yield of Faba Bean
Genotypes in Southern Ethiopia. Electronic Journal of
Plant Breeding, 3 (3): 898-907.
Fernandez GCJ. (1991). Analysis of genotype x
environment interaction by stability estimates.
Horticultural Science, 26: 947-950.
Gauch HG. (1992). Statistical Analysis of Regional Trials:
AMMI Analysis of Factorial Design. Elsevier,
Amsterdam.
Gauch HG, Piepho HP, Annicchiarico P. (2008). Statistical
analysis of yield trials by AMMI and GGE: Further
considerations. Crop Science, 48: 866–889.
Gemechu K, Mussa J. (2009). Comparison of Two
Approaches for Estimation of Genetic Variation for Two
Economic Traits in Faba Bean Genotypes Grown under
Waterlogged Verisols. East African Journal of Sciences,
3 (1): 95-101.
GenStat, (2012). GenStat Procedure Library Release. 15th
edition. VSN International Ltd.
Gollob HF. (1968). A statistical model which combines
features of factor analytic and analysis of variance
techniques. Psychometrika, 33: 73-115.
Gomez KA, Gomez A. (1984). Statistical Procedures for
Agricultural Research, 2nd Edition. John Wiley & Sons,
New York.
Hirpa L, Nigussie D, Setegn G, Geremew B, Firew M.
(2013). Response to Soil Acidity of Common Bean
Genotypes (Phaseolus vulgaris L.) Under Field
Conditions at Nedjo, Western Ethiopia. Science,
Technology and Arts Research Journal, 2 (3): 03-15.
Kang MS. (1990). Genotype-by-environment interaction
and plant breeding. Louisiana State University, Baton
Rouge, LA.
Lin CS, Binns MR, Lefkovitch LP. (1986). Stability analysis:
Where do we stand? Crop Science, 26: 894-900.
Mesfin T, Wassu M, Mussa J. (2019). Genetic Variability on
Grain Yield and Related Agronomic Traits of Faba Bean
(Vicia faba L.) Genotypes under Soil Acidity Stress in the
Central Highlands of Ethiopia. Chemical and
Biomolecular Engineering, 4(4): 52-58. doi:
10.11648/j.cbe.20190404.12
Million F, Habtam S. (2012). Genetic Variability on Seed
Yield and Related Traits of Elite Faba Bean (Vicia faba
L.) Genotypes. Pakistan Journal of Biological Sciences,
15 (8): 380-385.
Mulusew F, Suso MJ, Tadele T, Legesse T. (2008).
Analysis of multi-environment yield performance of faba
bean (Vacia faba L.) genotypes using AMMI model.
Journal of Genetics and Breeding, 62: 25-30.
Mussa J, Gemechu K. (2006). Vicia faba L. In: Brink, M.
and Belay, G. (eds.). Plant Resources of Tropical Africa
1: Cereals and Pulses. Wageningen,
Netherlands/Backhuys: PROTA Foundation.
Ouertatani S, Regaya K, Ryan J, Gharbi A. (2011). Soil
liming and mineral fertilization for root nodulation and
growth of faba beans in an acid soil in Tunisia. Journal
of Plant Nutrition, 34: 850–860.
Purchase JL. (1997). Parametric analysis to describe
genotype x environment interaction and yield stability in
winter wheat. A PhD. Thesis, Department of Agronomy,
Faculty of Agriculture of the Orange Free State,
Bloemfontein, South Africa.
Rashidi M, Farshadfar E, Jowkar MM. (2013). AMMI
analysis of phenotypic stability in chickpea genotypes
over stress and non-stress environments. International
Journal of Agriculture and Crop Sciences, 5 (3): 253-
260.
SAS Institute. (2010). SAS/STAT guide for personal
computers, version 9.3 edition. Cary, NC: SAS Institute
Inc.
Tamene T. (2008). Genetic gain and morpho-agronomic
basis of genetic improvement in grain yield potential
achieved by faba bean (Vicia faba L.) breeding in
Ethiopia. MSc Thesis, Hawassa University, Hawassa,
Ethiopia.
Tamene T. (2015). Application of AMMI and Tai’s stability
statistics for yield stability analysis in faba bean (Vicia
faba L.) cultivars grown in central highlands of Ethiopia.
Journal of Plant Sciences, 3 (4): 197-206.
Tamene T, Gemechu K, Hussein M. (2015). Genetic
progresses from over three decades of faba bean (Vicia
faba L.) breeding in Ethiopia. Australian Journal of Crop
Science, 9 (1): 41-48.
Teklay A, Yemane N, Muez M, Adhiena M, Assefa W.
Hadas B. (2015). Genotype by environment interaction
of some faba bean genotypes under diverse broomrape
environments of Tigray, Ethiopia. Journal of Plant
Breeding and Crop Science, 7 (3): 79-86.
Yan W. (2001). GGE Biplot: A Windows Application for
Graphical Analysis of Multi-Environment Trial Data and
Other Types of Two-way Data. Agronomy Journal 93:
1111-1118.

Mesfin T. 846
Yan W, Kang MS. (2002). A graphical tool for breeders,
geneticists, and agronomists. CRC Press, UK.
Yan W, Kang MS. (2003). GGE Biplot Analysis. CRC Press,
New York.
Yan W, Tinker NA. (2005). An integrated biplot dnalysis
system for displaying, interpreting and exploring
genotype x environment interaction. Crop Science, 45:
1004-1016.
Yan W, Kang MS, Ma B, Woods S, Cornelius PL. (2007).
GGE biplot vs. AMMI analysis of genotype-by
environment data. Crop Scicence, 47: 643-655.
Zobel RW, Wright MS, Gauch HG. (1988). Statistical
analysis of a yield trial. Agronomy Journal, 80: 388-393.
Accepted 18 September 2020
Citation: Mesfin Tadele, Wassu Mohammed and Mussa
Jarso (2020). Yield Stability and Genotype × Environment
Interaction of Faba Bean (Vicia faba L.). International
Journal of Plant Breeding and Crop Science, 7(2): 833-846.
Copyright: © 2020: Mesfin T. This is an open-access
article distributed under the terms of the Creative Commons
Attribution License, which permits unrestricted use,
distribution, and reproduction in any medium, provided the
original author and source are cited.

Yield Stability and Genotype × Environment Interaction of Faba Bean (Vicia faba L.)

More Related Content

What's hot (20)

Similar to Yield Stability and Genotype × Environment Interaction of Faba Bean (Vicia faba L.) (20)

More from Premier Publishers (20)

Recently uploaded (20)

Yield Stability and Genotype × Environment Interaction of Faba Bean (Vicia faba L.)