© The Authors, 2023, Published by the Universidad del Zulia*Corresponding author: cmaldonado@utb.edu.ec
Keywords:
Interspecic crosses
Phenotypic stability
Adaptability
Phenotypic stability of forty advanced lines of rice at Babahoyo, Ecuador
Estabilidad fenotípica de cuarenta líneas avanzadas de arroz en Babahoyo, Ecuador
Estabilidade fenotípica de quarenta linhagens avançadas de arroz na área de Babahoyo, Ecuador
Cristina Evangelina Maldonado Camposano
1,3*
Walter Oswaldo Reyes Borja
1
Luis Alberto Duicela Guambi
2
Fernando Javier Cobos Mora
1
Rev. Fac. Agron. (LUZ). 2023, 40(3): e234031
ISSN 2477-9407
DOI: https://doi.org/10.47280/RevFacAgron(LUZ).v40.n3.09
Crop Production
Associate editor: Dra. Evelyn Peréz-Peréz
University of Zulia, Faculty of Agronomy
Bolivarian Republic of Venezuela
1
Universidad Técnica de Babahoyo, Ecuador.
2
Escuela Superior Politécnica Agropecuaria de Manabí,
Ecuador.
3
Doctorado en Ciencias Agrarias de la Facultad de Agronomía
Universidad del Zulia, Venezuela.
Received: 29
-05-2023
Accepted: 31-08-2023
Published: 20-09-2023
Abstract
The crosses between Oryza sativa L. and O. rupogon Gri., create a
high genetic diversity to develop rice varieties with high yield and phenotypic
stability. In the present investigation, forty advanced lines of rice were
evaluated in subsidiaries F
5
(dry season) and F
6
(rainy season), together with
three commercial controls in the town of Babahoyo, Ecuador. A Randomized
Complete Block Design (DBCA) was applied with three repetitions,
recording morphoagronomic and productive characters. Statistical analyzes
were applied and phenotypic stability was determined using the Eberhart
and Russell, AMMI, Lin and Binns, PROMVAR models. The average
morphoagronomic results were: days to owering (72), vegetative cycle (98
days), plant height (111 cm), panicle sterility (6 %); the productive variables
the results were: tillers per plant (32), panicles per plant (31), panicle length
(27 cm), grains per panicle (168) and yield (8,100 kg.ha
-1
). The stable lines
identied by the models: Eberhart and Russell were 1, 2, 10, 11, 13, 18, 25,
26, 30 and 37; AMMI identied lines 8 and 22; Lin and Binns to lines 2, 12,
18, 27, 37 and 40; and PROMOTE lines 2, 10, 13, 18, 25, 30, 38, 40 and 43;
concluding that seven lines (2, 10, 13, 18, 25, 30 and 40) coincided with the
applied models except AMMI. The average yield of the lines mentioned in
the two seasons was 7,797 kg.ha
-1
, higher than the average of the commercial
controls that obtained 6,809 kg.ha
-1
.
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Rev. Fac. Agron. (LUZ). 2023, 40(3): e234031. July-September. ISSN 2477-9407.2-7 |
Resumen
Los cruces entre Oryza sativa L. y O. rupogon Gri., crean una
alta diversidad genética para desarrollar variedades de arroz con alto
rendimiento y estabilidad fenotípica. En la presente investigación se
evaluó cuarenta líneas avanzadas de arroz en liales F
5
(época seca) y
F
6
(época lluviosa), junto con tres testigos comerciales en la localidad
de Babahoyo, Ecuador. Se aplicó un Diseño de Bloques Completos
al Azar (DBCA) con tres repeticiones, registrando caracteres
morfoagronómicos y productivos. Análisis estadísticos fueron
aplicados y la estabilidad fenotípica fue determinada utilizando
los modelos Eberhart y Russell, AMMI, Lin y Binns, PROMVAR.
Los resultados morfoagronómicos en promedios fueron: días a la
oración (72), ciclo vegetativo (98 días), altura de planta (111 cm),
esterilidad de panícula (6 %); las variables productivas los resultados
fueron: macollos por planta (32), panículas por planta (31), longitud
de panícula (27 cm), granos por panícula (168) y rendimiento (8.100
kg.ha
-1
). Las líneas estables identicadas por los modelos: Eberhart y
Russell fueron 1, 2, 10, 11, 13, 18, 25, 26, 30 y 37; AMMI identicó
las líneas 8 y 22; Lin y Binns a las líneas 2, 12, 18, 27, 37 y 40; y
PROMVAR las líneas 2, 10, 13, 18, 25, 30, 38, 40 y 43; concluyéndose
que siete líneas (2, 10, 13, 18, 25, 30 y 40) coincidieron con los
modelos aplicados excepto AMMI. El rendimiento promedio de las
líneas mencionadas en las dos épocas fue de 7.797 kg.ha
-1
, superior
al promedio de los testigos comerciales que obtuvieron 6.809 kg.ha
-1
.
Palabras clave: cruces interespecícos, estabilidad fenotípica,
adaptabilidad.
Resumo
Os cruzamentos entre Oryza sativa L. e O. rupogon Gri.,
criam uma alta diversidade genética para desenvolver variedades
de arroz com alto rendimento e estabilidade fenotípica. Na presente
investigação, foram avaliadas quarenta linhas avançadas de arroz nas
subsidiárias F
5
(estação seca) e F
6
(estação chuvosa), juntamente com
três controles comerciais na cidade de Babahoyo, Equador. Foi aplicado
o Delineamento em Blocos Completos Randomizados (DBCA) com
três repetições, registrando-se os caracteres morfoagronômicos e
produtivos. Análises estatísticas foram aplicadas e a estabilidade
fenotípica foi determinada usando os modelos Eberhart e Russell,
AMMI, Lin e Binns, PROMVAR. Os resultados morfoagronômicos
médios foram: dias para oração (72), ciclo vegetativo (98 dias),
altura da planta (111 cm), esterilidade da panícula (6 %); nas
variáveis produtivas os resultados foram: perlhos por planta (32),
panículas por planta (31), comprimento da panícula (27 cm), grãos
por panícula (168) e produtividade (8.100 kg.ha
-1
). As linhas estáveis
identicadas pelos modelos: Eberhart e Russell foram 1, 2, 10, 11,
13, 18, 25, 26, 30 e 37; A AMMI identicou as linhas 8 e 22; Lin e
Binns para as linhas 2, 12, 18, 27, 37 e 40; e PROMOVER as linhas
2, 10, 13, 18, 25, 30, 38, 40 e 43; concluindo que sete linhas (2, 10,
13, 18, 25, 30 e 40) coincidiam com os modelos aplicados exceto
AMMI. A produtividade média das linhagens citadas nas duas safras
foi de 7.797 kg.ha
-1
, superior à média das testemunhas comerciais que
obtiveram 6.809 kg.ha
-1
.
Palavras-chave: cruzamentos interespecícos, estabilida de
fenotípica, adaptabilidade.
Introduction
The geographical origin of rice (Oryza sativa L.) was probably
in northeastern India, on the slopes of the Himalayas; the expansion
of this crop started from Southeast Asia to China 3 000 years (B.C.);
later to Korea, Japan in the rst century (B.C.) (Pinciroli et al., 2015).
The economic importance of the agricultural sector in Ecuador is
also supported by rice cultivation, as it is one of the main products of
the basic food basket, with only 4 % of the production destined for
export to Colombia and Peru (Poveda and Andrade, 2018).
In the selection process, stable and high-yielding genotypes
should be evaluated in dierent locations and during several cycles,
to identify those with outstanding adaptability potential, before being
recommended for cultivation in any location or region (Sánchez-
Ramírez et al., 2016).
Crosses between Oryza sativa L. and O. rupogon Gri. generate
abundant genetic diversity for the development of high-yielding rice
varieties. On the other hand, the introgression of certain specic
alleles of wild-type rice can contribute positively, even to increase
resistance to stress (Martínez et al., 1998).
There are statistical models that lead to determine the genetic
stability of commercial cultivars. The Eberhart and Russell Model
provides estimates of stability parameters and characterizes the
predictability of a genotype (Jiménez, 2006).
The Lin and Binns model describes genotype x environment
interaction (GEI) eects and determines adaptability in a general
sense (Acevedo et al., 2010).
The AMMI model (Additive main eects and multiplicative
interaction method), allows a more detailed analysis of IGA, ensuring
the selection of more productive genotypes, providing more accurate
estimates of genotypic response (Acevedo et al., 2010).
The PROMVAR model is based on the relationship between
genotype yield averages and their respective variances (Benítez et al.,
1988).
Due to the low genetic diversity in rice cultivation, it is possible
to create new germplasm adapted to the dierent agroecological
conditions of Ecuador. The lines characterized in this study are
derived from interspecic crosses of japonica rice (Oryza sativa
L. ssp. japonica) with wild rice (Oryza rupogon Gri.) known in
Ecuador as black rice or Puyón, with potential genetic contribution in
agronomic, phytosanitary and production terms. For this reason, the
objective was to determine the phenotypic stability of forty advanced
rice lines in Babahoyo, Ecuador.
Materials and methods
The research was conducted in the project area of the Study
Commission for the Development of the Guayas River Basin
(CEDEGE), Babahoyo, Los Ríos province, Ecuador (01º50’1.85”
South Latitude and 79º26’47.26” West Longitude, 7 masl), with an
annual average precipitation of 1,909 mm; 85 % relative humidity;
590.9 hours of heliophany and maximum and minimum temperature
of 30 and 21.9 °C, respectively [Instituto Nacional de Meteorología e
Hidrología (INAMHI, 2017)].
In table 1, the advanced line codes in F
5
(dry season) and F
6
(rainy
season) subsidiaries and commercial cultivars (controls) studied are
detailed.
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Maldonado et al. Rev. Fac. Agron. (LUZ). 2023 40(3): e234031
3-7 |
Table 1. Codes of advanced lines F
5
and F
6
and commercial
cultivars (controls) of rice. Babahoyo 2018-2019.
N° Line/cultivar codes N
o
Line/cultivar codes
1 PUYÓN/JP002 P8-30-P55-2 23 PUYÓN/JP002 P8-31-P41-4
2 PUYÓN/JP002 P8-30-P23-12 24 PUYÓN/JP002 P8-31-P7-4
3 PUYÓN/JP002 P8-30-P84-19 25 PUYÓN/JP002 P8-31-P7-27
4 PUYÓN/JP002 P8-30-P94-1 26 PUYÓN/JP002 P8-32-P97-5
5 PUYÓN/JP002 P8-30-P60-25 27 PUYÓN/JP002 P8-32-P97-13
6 PUYÓN/JP002 P8-30-P68-1 28 PUYÓN/JP002 P8-32-P8-16
7 PUYÓN/JP002 P8-30-P13-24 29 PUYÓN/JP002 P8-32-P87-26
8 PUYÓN/JP002 P11-10-P87-11 30 PUYÓN/JP002 P8-32-P35-20
9 PUYÓN/JP002 P11-10-P40-24 31 PUYÓN/JP002 P8-32-P109-24
10 PUYÓN/JP003 P11-10-P62-32 32 PUYÓN/JP002 P8-32-P40-22
11 PUYÓN/JP002 P8-28-P7-7 33 PUYÓN/JP002 P8-29-P60-1
12 PUYÓN/JP002 P8-28-P81-32 34 PUYÓN/JP002 P8-29-P60-6
13 PUYÓN/JP002 P8-28-P47-6 35 PUYÓN/JP002 P8-29-P8-16
14 PUYÓN/JP002 P8-20-P1-6 36 PUYÓN/JP002 P8-29-P8-5
15 PUYÓN/JP002 P8-20-P94-27 37 PUYÓN/JP002 P8-29-P49-30
16 PUYÓN/JP002 P8-20-P61-3 38 PUYÓN/JP002 P8-29-P32-1
17 PUYÓN/JP002 P8-20-P72-4 39 PUYÓN/JP002 P8-29-P65-5
18 PUYÓN/JP002 P8-31-P25-2 40 PUYÓN/JP002 P8-29-P66-30
19 PUYÓN/JP002 P8-31-P45-1 41 INIAP FL1480 Cristalino (testigo 1)
20 PUYÓN/JP002 P8-31-P45-28 42 SFL-11(testigo 2)
21 PUYÓN/JP002 P8-31-P30-18 43 INIAP FL-Arenillas (testigo 3)
22 PUYÓN/JP002 P8-31-P63-5
The experimental unit consisted of a useful area of 1 m
2
, excluding
the borders, where the 16 central plants were evaluated.
Morphoagronomic variables
Days to owering (DTF), when there were more than 50 % of
owering plants in the useful plot. Vegetative cycle (VC), days from
transplanting to physiological maturity of the plants (harvest). Plant
height (PH), evaluated one week before harvest by measuring from
the ground to the apex of the most protruding panicle. Number of
tillers per plant (NTP), recorded at harvest.
Production variables
Number of panicles per plant (NPP), the number of panicles
per plant was counted at the time of harvest. Panicle length (PL),
considered from the ciliate node to the panicle apex. Grains per
panicle (NGP), the total number of grains per panicle was recorded
from three random panicles per plant of the 16 plants in the useful area.
Shelled kernel length (SKL), the length in millimeters was measured
in ten randomly shelled kernels with a vernier caliper using the CIAT
shelling category scale (CIAT, 1993). Sterility (PEP), sterile kernels
were counted and the percentage of sterility was determined from
the total number of kernels. Yield in kg.ha
-1
, the weight of the grains
from the useful plot was determined, adjusted to 13 % moisture; to
uniformize the weight, the formula used by Cedeño et al. (2018) was
used.
Statistical analyses were performed using the following computer
programs INFOSTAT (UNC, 2018), INFOGEN (Balzarini and Di
Rienzo, 2016), and Microsoft Oce EXCEL. The experimental
design used was randomized complete blocks (DBCA) with three
replications. The linear analysis of variance model was applied in
DBCA with the following equation:
Y
ijk
(Phenotypic value of genotype i, evaluated in k replicates and
j environments) = µ (Overall mean) + g
i
(Genotype eect) + b(a)
k(j)
(Within-environment repeat eect) + a
j
(Environment eect) + (ga)
ij
(Genotype-environment interaction eect) + E
ijk
(Experimental error
associated with the ijk-th observation).
The scheme of the combined variance analysis was also
determined: lines by epoch: Sum of squares (SC), Degrees of freedom
(DF), Mean squares (MS) and calculated F. In the analysis process,
parametric tests were applied based on the verication of compliance
with the assumptions of normality and homoscedasticity of the data
(Corral, 2019).
In addition, the following statistics were determined: mean (Ȳ),
median (Md), variance (S
2
), standard deviation (S), standard or
typical error (EE = SȲ), coecient of variation (CV %), condence
interval of µ (95 %) (LIC and LSC).
The “Bigger is better” trait dierentiation criterion was used for
the variables number of tillers per plant (TPP), number of panicle per
plant (NPP), panicle length (PL), number of grains per panicle (NGP),
yield (kg), shelled grain length (SGL), and the “Smaller is better”
criterion for the variables days to owering (DTF), life cycle in days
(LCD), plant height (PH), and percent sterility (PEP). In addition, a
bivariate linear r correlation analysis was performed.
The stability of the lines was determined by applying four models:
The Eberhart and Russell Model (1966), AMMI (Additive main
eects and multiplicative interaction method) proposed by Zobel et
al. (1988), Lin and Binns (1988), as well as PROMVAR (ratio of
averages and relative variability) proposed by Benítez et al. (1988).
As for the Eberhart and Russell model, the stability parameters (b,
β and S
2
) were determined. The selection priority (∑MS) for the
most stable lines, according to this model, taking the yield variable
(kg.ha
-1
), was determined with the following parameters: Ḡi (k.ha
-1
)
and was transformed by assigning the value of 1 to those lines that
presented yields below the general mean and the value of 3 to those
lines that presented values above the general mean; the coecient
of determination R
2
was transformed, qualifying with 1 the values
observed below the general mean and with 2 the values above the
general mean; for the variance of the deviations (S
2
d) was qualied
with 3 the values below the general mean and 1 to the values above
the general mean. The sum of the three parameters described (Ḡi, R
2
,
S
2
d) allowed the calculation of the selection priority, where the lines
with higher scores were determined as priority 1 of higher stability.
Results and discussion
The general average of the traits by dry (E1) and rainy (E2)
seasons in the rice lines/varieties, dierentiated with the criterion
“Less is better”, were the variables DTF, VC, AP and PEP. The best
performance was in the rainy season DAF (E2)= 72 days, VC (E2)=
98 days, PH (E2)= 111 cm, PEP (E1)= 6 %. The criterion “Bigger
is better” was observed in the variables NTP, NPP, PL, NGP, kg and
SGL; however, the best performance was observed in the rainy season
in the variables, NTP (E2)= 32, NPP (E2)= 31 and PL (E2)= 27 cm,
in contrast to the dry season, where the best results were presented by
the variables NGP (E1)= 168 and yield (E1)= 8,100 kg.ha
-1
(table 2).
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Rev. Fac. Agron. (LUZ). 2023, 40(3): e234031. July-September. ISSN 2477-9407.4-7 |
Table 2. Descriptive statistics of morpho agronomic and productive characteristics in Babahoyo during the dry (E1) and rainy (E2)
seasons.
Variables
Statistics
Ῡ Md S
2
S SE CV RV Min Max N LIC LSC
DTF (E1) 74 74 3,69 1,92 0,29 2,60 0,39 70 79 43 74 75
DTF (E2) 72 72 0,94 0,97 0,15 1,30 0,20 70 75 43 72 73
VC (E1) 114 114 3,69 1,92 0,29 1,70 0,26 110 119 43 114 115
VC (E2) 98 98 0,94 0,97 0,15 0,99 0,15 96 101 43 98 99
PH (E1) 113 112 22,85 4,78 0,73 4,20 0,64 109 137 43 112 11 5
PH (E2) 111 111 28,02 5,29 0,81 4,75 0,72 104 130 43 110 113
NTP (E1) 26 26 2,15 1,46 0,22 5,70 0,86 23 29 43 25 26
NTP (E2) 32 32 2,58 1,60 0,24 5,03 0,77 27 35 43 31 32
NPP (E1) 25 25 2,27 1,51 0,23 6,00 0,91 22 28 43 25 26
NPP (E2) 31 31 2,73 1,65 0,25 5,30 0,81 26 35 43 31 32
PL (E1) 26,00 26,00 0 0 0,06 1,40 0,22 26 28 43 26 27
PL (E2) 26,50 26,40 0,58 0,76 0,12 2,88 0,44 25 29 43 26 27
NGP (E1) 168 168 23 4,80 0,73 2,90 0,44 152 177 43 166 169
NGP (E2) 140 139 22,18 4,71 0,72 3,37 0,51 133 154 43 139 141
PEP (E1) 6,20 6,30 0,87 0,93 0,14 15 2,28 4,00 9,00 43 6,00 7,00
PEP (E2) 5,70 5,80 0,44 0,66 0,10 11,60 1,77 4,00 7,00 43 5,53 5,93
kg.ha
-1
(E1) 8100 8105 755106 869 133 10,73 1,64 6271 9884 43 7840 8359
kg.ha
-1
(E2) 6608 6593 278854 528 81 7,99 1,22 5440 7777 43 6450 6766
SGL (E1) 7,20 7,10 0,01 0,11 0,02 1,59 0,24 7,00 8,00 43 7,00 7,00
SGL (E2) 7,23 7,21 0,01 0,12 0,02 1,66 0,25 6,94 7,58 43 7,20 7,30
DTF: Days to owering. VC: Vegetative cycle. PH: Plant height. NTP: Number of tillers per plant. NPP: Number of panicles per plant. PL: Panicle length. NGP: Number
of grains per panicle. PEP: percent sterility. kg.ha
-1
= Yield. SGL: shelled grain length. Ῡ= Mean; Md= Median; S
2
= Variance; S= Standard deviation; SE= Standard error;
CV= Coecient of variation %; RV= Relative variability %; Min= Minimum; Max= Maximum; n= Number of observations; LIC= Lower condence limit; LSC= Upper
condence limit.
Table 3. Combined analysis of variance of yield (kg.ha
-1
) of forty lines and three commercial rice varieties in two seasons (dry and rainy)
at Babahoyo location 2018-2019.
Source of variation Degrees of freedom Sum of squares Mean square F calculated p-value
Seasons 1 143528913 143528913 106.63 <0.0001
Lines/variety 42 69126597 1645871 1.22 0.1871
Repetition 2 10108696 5054348 3.76 0.0254
Epochs x lines/variety 42 61161261 1456221 1.08 0.3543
Error 170 228824957 1346029
Total 257 512750424
CV (%) 15.78
According to the mentioned criteria “Higher is better” and
“Lower is better” in the two seasons studied, the traits that stood out
in each season were dierentiated, where it was observed that there
is an eect of the environment on the lines and varieties, through the
evaluated traits. The best performance of the lines was observed in
the dry season; these results agree with Chloupek et al. (2004), who
mentioned that IGA is inuenced by the dierent crop management
technologies in the producing areas. The increase in yield over
time is supported by genetic gains from the substitution of cultivars
traditionally planted in the locality by cultivars from promising
lines, in which genetic improvement programs have contributed with
relevant modications in characters of agronomic, productive and
phytosanitary importance.
To decide the relevance of the parametric analysis of variance,
homoscedasticity and normality tests were performed. In the rst
case, the F test was used and in the second case the Jarque and Bera
(1987) test was used, contrasting with Chi-square. The F test for yield
(kg.ha
-1
) showed that most of the lines presented homoscedasticity
and normality in the two periods; therefore, the assumption required
for the application of parametric tests such as analysis of variance
was met.
In this study it was demonstrated that most of the lines showed
homoscedasticity between the season; therefore, the required
assumption is fullled. Siegel (1990) mentioned that the conditions
of normality, homoscedasticity and independence of quantitative data
series in agricultural research are generally not veried; therefore, the
risk of committing type I errors in statistical decisions is elevated.
These selection decisions in rice lines have usually been made in
descriptive analyses, mainly of means ± standard deviation.
In this experiment, the combined analysis of variance for yield
(kg.ha
-1
) (CV= 15.78 %), showed highly signicant dierences in
sowing seasons (dry and rainy), lines/variety, repetition and seasons
x lines (table 3).
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Maldonado et al. Rev. Fac. Agron. (LUZ). 2023 40(3): e234031
5-7 |
Table 5. Results of the Eberhart and Russell model, using the variable yield (kg.ha
-1
) with the rice lines determined as selection priority 1.
Lines
G
i
Gi R
2
S
2
d
Σ MC
Priority of
Selection
(kg.ha
-1
) b β S
2
d R
2
(1-3) (1-2) (1-3)
1 7357 1.86 0.86 0.506 0.832 3 2 3 8 1
2 8301 1.20 0.20 0.594 0.637 3 2 3 8 1
10 7546 1.64 0.64 0.682 0.458 3 2 3 8 1
11 7793 1.99 0.99 0.651 0.815 3 2 3 8 1
13 7605 0.99 -0.01 0.353 0.667 3 2 3 8 1
18 8240 0.80 -0.20 0.033 0.678 3 2 3 8 1
32 7388 0.85 -0.15 0.899 0.868 3 2 3 8 1
Ȳ (kg.ha
-1
) 7354 1.02 0.02 1.08 0.35
Ḡi kg.ha
-1
= Yield, (b= Slope, β= B prime, S
2
d= Variance of deviations, R
2
= Coecient of determination)= Eberhart and Russell stability parameters, (Gi (1-3), R
2
(1-2), S
2
d
(1-3))= Ordinal scaled category grouping of average yield statistics of genotypes, Σ MC= Multicredit value, Selection priority.
The Pearson’s product moment correlation analysis (r) determined
the statistical association between the morphoagricultural variables
as shown in table 4. The VC variable was observed correlated with
greater consistency with the following variables: NTP (r = -0.751**);
NPP (-0.740**); NGP (0.801**). Likewise, the variables NTP and
NPP (0.988**) showed the same trend of consistency, since they
approached values of 1 or -1.
Table 5 shows lines 1, 2, 10, 11, 13, 18 and 32, which were the ones
that occupied selection priority 1 due to their high stability; however,
the other lines/cultivars were distributed in priorities from 2 to 6.
Table 4. Linear correlation matrix r between morphological and productive characters of forty lines and three commercial rice varieties
in two seasons (dry and rainy), in the locality of Babahoyo 2018-2019.
Variables DAF CVD AP LHB NMP NPP LP NGP kg.ha-
1
LGD
DTF 1
VC 0.594** 1
PH 0.072 0.163 1
FLL 0.108 -0.178 0.585** 1
NTP -0.355** -0.751** -0.208** 0.140 1
NPP -0.356** -0.740** -0.197 0.148 0.988** 1
PL 0.102 0.001 0.471** 0.455** -0.058 -0.066 1
NGP 0.231** 0.801** 0.208 ** -0.169 -0.615** -0.594** 0.130 1
kg.ha
-1
0.113 0.484** 0.015 -0.152 -0.256** -0.268 ** 0.044 0.490** 1
LGD 0.154 -0.066 0.285** 0.337** 0.017 0.018 0.314** -0.106 -0.117 1
*correlation at 95 % condence, **correlation at 99 % condence where p<0.01, and GL=256 critical values of r 0.05 = 0.151 and r 0.01 = 0.197. DTF= Days to owering, VC=
Vegetative cycle, PH= Plant height, FLL= Flag leaf length, NTP= Number of tillers per plant, NPP= Number of panicle per plant, PL= Panicle length, NGP= Number of grains per
panicle, kg.ha
-1
= Yield, LGD= Shelled grain length.
The AMMI model combines additive components for the main
eects (ANDEVA) and multiplicative components for the genotype x
environment interaction (IGA), using principal components analysis
(PCA) (Jiménez, 2006).
The results of the application of the AMMI model, classify and
determine the magnitude of the interaction of the lines with the
environment using the production variable, where CP1 represented
100% of the total variance in the two periods evaluated in Babahoyo.
The most stable lines were: 3, 6, 8, 10, 13, 13, 16, 17, 18, 22, 25, 29,
31 and 32. The most stable and highest yielding lines were 8 and 22.
The lines with the highest IGA were: 1, 5, 11, 12, 14, 24, 26, 27, 35,
36 and 37 (gure 1).
Lozano-del Río et al. (2009), using the AMMI model, identied
the locations of high IGA in 14 environments, where the production
of twenty-two forage genotypes of triticale (x triticosecale Wittm.)
was evaluated. This same situation was evidenced for the highest IGA
with lines 37, 24, 11 and 35 in the dry season at the Babahoyo locality.
Figure 1. Results of the AMMI model on yield stability (kg.ha
-1
) of forty
lines and three commercial rice cultivars in two seasons (dry
and rainy), in the locality of Babahoyo 2018-2019.
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Maldonado et al. Rev. Fac. Agron. (LUZ). 2023 40(3): e234031
6-7 |
Acevedo et al. (2010), applying the AMMI analysis to establish
the magnitude of IGA in yield (t.ha
-1
) and to evaluate the adaptability
and phenotypic stability of rice genotypes, stated that the relative
stability values depend fundamentally on the representativeness of
the available environments and the number of locations evaluated.
The environmental index determined by the Lin and Binns model
(1988), allowed the identication of rice lines 2, 12, 18, 27, 37 and 40
of high adaptability (gure 2). The results of the Lin and Binns model
in the present research, corroborate what was studied by Regitano et
al. (2013), who evaluated the behavior of rainfed rice genotypes in
the state of Sao Paulo, to determine the stability and adaptability of
the yield variable, using the methodologies proposed by Eberhart and
Russel (1966) and Lin and Binns (1988), where they identied three
genotypes that showed the lowest values of environmental index Pi,
which indicated that they were the closest to maximum productivity.
In this experiment, four models (Eberhart and Russel, Lin and
Binns, AMMI and PROMVAR) were used to determine the phenotypic
stability of forty lines and three commercial controls, with only seven
lines coinciding in the three models applied. This study corroborates
the data obtained by Vargas et al. (2016), who using the Eberhart and
Russel, Lin and Binns, and AMMI models evaluated maize hybrids
developed by the National Federation of Cereal and Legume Growers
(Fenalce) with germplasm provided by the International Maize and
Wheat Improvement Center (CIMMYT), in 17 environments in six
agroecological zones of Colombia.
As for the PROMVAR model, the values were located in the
lower right grid (gure 3), showing general average values in yield
and relative variability of 7,354 kg.ha
-1
and 7.54 %, respectively;
where there were advanced lines with high yield behavior and high
stability, such as lines 2, 10, 13, 18, 18, 25, 30, 38, 40 and INIAP FL-
Arenillas (control 3), and values ranged from 6,166 to 8,301 kg.ha
-1
,
with relative variability values of 3.37 to 11.84 %. Unlike control
41 (INIAP FL-1480), which showed low yield and low stability, and
control 42 (SFL-11), which showed low yield and high stability in the
application of the PROMVAR model.
This study agrees with Velasco (2019) who used seven rice lines
in F4 in the area of Babahoyo, Ecuador, reporting that the yield
per plant in the seven lines were in a range of 51.7-61.9 g.plant
-1
,
compared with the control SFL-11 that obtained a lower value of 50.9
g. plant
-1
, inferring that the lines were superior to the control; in the
present study, the results presented the same tendency when it was
observed that the forty lines obtained average yields of 7,395 kg.ha
-1
in the two seasons, compared with the control commercial varieties
that reached 6,809 kg.ha
-1
.
Conclusions
The best morphological, agronomic and productive performance
was obtained in the dry season. The Eberhart and Russell model
allowed the selection of lines 1, 2, 10, 11, 13, 13, 18, 25, 26, 30, 32 and
37 with priority score 1. According to the Lin and Binns model, lines
2, 12, 18, 27, 37 and 40 showed high stability and higher production
as a function of the yield superiority index. The PROMVAR model
identied lines 2, 10, 13, 18, 25, 30, 38, 40 and 43 as having high
yield kg.ha
-1
and high stability (reduced relative variability).
The models that best t to identify the stability of the rice lines
during the two seasons (dry and rainy) in the locality of Babahoyo
were Eberhart and Russell and PROMVAR, followed by the Lin and
Binns model, due to the coincidence of the lines selected by the three
models to determine stability.
Figure 2. Environmental indices of the Lin and Binns model in
two growing seasons (dry and rainy) of forty lines
and three commercial rice cultivars in the locality of
Babahoyo 2018-2019.
This scientic publication in digital format is a continuation of the Printed Review: Legal Deposit pp 196802ZU42, ISSN 0378-7818.
Maldonado et al. Rev. Fac. Agron. (LUZ). 2023 40(3): e234031
7-7 |
Chloupek, O., Hrstkova, P., & Schweigert, P. (2004). Yield and its stability, crop
diversity, adaptability and response to climate change, weather and
fertilization over 75 years in the Czech Republic in comparison to some
European countries. Field Crops Research, 85(2-3), 167-190. https://
doi:10.1016/S0378-4290(03)00162-X
CIAT. (1993). Descriptores Varietales Arroz, frijol, maíz, sorgo. Cali, Colombia:
Publicación CIAT. https://cgspace.cgiar.org/handle/10568/54651
Corral Dávalos, L. (2019). Estadística y técnicas experimentales para la
investigación biológica. Editorial Universitaria Abya-Yala, 83, 237-242.
https://dspace.ups.edu.ec/handle/123456789/21027
Eberhart, S., & Russel, W. (1966). Stability Parameters for Comparing
Varieties. Crop Science, 6(1), 36-40. https://doi:10.2135/
cropsci1966.0011183X000600010011x
Instituto Nacional de Meteorología e Hidrología. (2017). Anuario Meteorológico.
165. https://www.inamhi.gob.ec/docum_institucion/anuarios/
meteorologicos/Am_2013.pdf
Jarque, C. M., & Bera, A. K. (1987). A test for normality of observations and
regression residuals. International Stadistical Institute, 55(2), 163-172.
https://doi.org/10.2307/1403192
Jiménez Contreras, J. (2006). Determinación y aplicación de métodos estadísticos
para medir estabilidad genética en vegetales. Caso banano. [Tesis de
Bachiller]. Repositorio Institucional de la Escuela Superior Politécnica
del Litoral http://www.dspace.espol.edu.ec/handle/123456789/5721
Lin, C., & Binns, M. (1988). A superiority measure of cultivar performance for
cultivar X location data. Canadian Journal of Plant Science, 68, 193-198.
https://doi:10.4141/cjps88-018
Lozano-del Río, A., Zamora-Villa, V., Ibarra-Jiménez, L., Rodríguez-Herrera, S.,
Cruz-Lázaro, E., & Rosa-Ibarra, M. (2009). Análisis de la interacción
genotipo-ambiente mediante el modelo AMMI y potencial de
producción de triticales forrageros (Triticosecale Wittm.). Universidad
y Ciencia Tropico Húmedo, 25(31), 81-82. https://www.redalyc.org/
pdf/154/15416335006.pdf
Martínez, C. P., Tohme, J., López, J., Borrero, J., McCouch, S., Roca, W., &
Guimaraes, E. (1998). Estado actual del mejoramiento del arroz mediante
la utilización de especies silvestre de arroz. Agronomía Mesoamericana,
9(1), 10-17. https://doi.org/10.15517/am.v9i1.24609
Pinciroli, M., Ponzio, N., & Salsamendi, M. (2015). El arroz alimento de
millones (1 ed.). (M. Pinciroli, Ed.) Buenos Aires: Universidad Nacional
del Centro de la Provincia de Buenos Aires. http://sedici.unlp.edu.ar/
handle/10915/46744
Poveda Burgos, G., & Andrade Garófalo, C. (2018). Producción sostenible de
arroz en la provincia del Guayas. Contribuciones a las Ciencias Sociales,
(marzo, 2018). https://www.eumed.net/rev/cccss/2018/03/produccion-
arroz-ecuador.html
Regitano Neto, A., Ramos Junior, E., Boller Gallo, P., Guilherme de Freitas, J., &
Azzini, L. (2013). Comportamento de genótipos de arroz de terras altas
no estado de São Paulo. Revista Ciência Agronômica, 44(3), 512-519.
https://doi.org/10.1590/S1806-66902013000300013
Sánchez-Ramírez, F., Mendoza-Castillo, M., & Mendoza-Mendoza, C. (2016).
Estabilidad fenotípica de cruzas simples e híbridos comerciales de maíz
(Zea mays L.). Fitotecnia Mexicana, 3, 269-275. https://www.redalyc.
org/pdf/610/61046936012.pdf
Siegel, S. (1990). Estadística no paramétrica aplicada a las ciencias de la conducta.
3ra ed., México. Trillas.
Universidad Nacional de Córdoba. (s.f.). INFOSTAT versión 2018. Programa
estadístico. Programa analítico. Brujas, Argentina. https://www.infostat.
com.ar/index.php?mod=page&id=46
Vargas Escobar, E. A., Vargas Sánchez, J., & Baena García, D. (2016). Análisis de
estabilidad y adaptabilidad de híbridos de maíz de alta calidad proteica en
diferentes zonas Agroecológicas de Colombia. Acta Agronómica, 65(1),
72-79. https://doi.org/10.15446/acag.v65n1.43417
Velasco, C. (2019). Producción y agronomía de siete líneas F4 de arroz, derivadas
de cruces. 129. https://dspace.utb.edu.ec/handle/49000/5995
Zobel, R. W., Wright, M. J., & Gauch Jr., H. G. (1988). Statistical analysis of a
yield trial. Revista de Agronomía, 80(3), 388-393. https://doi:10.2134/agr
onj1988.00021962008000030002x
Figure 3. Four-cell plot of the PROMVAR model of the variable
Yield (kg.ha
-1
) in the locality of Babahoyo in two seasons
(dry and rainy) 2018-2019.
Lines 2, 10, 13, 18, 25, 30 and 40 converge in the Eberhart and
Russell, Lin and Binns and PROMVAR models, also showed high
stability and the highest production in the two seasons (dry and rainy)
evaluated.
The average yield of lines 2, 10, 13, 18, 25, 30 and 40 determined
in the two seasons was 7,797 kg ha
-1
, higher than the average of the
commercial controls, which obtained 6,809 kg ha
-1
.
The commercial controls INIAP FL-1480, SFL-11 and INIAP-FL
Arenillas, showed low productivity and low stability, with respect to
the aforementioned lines in the two seasons (dry and rainy) evaluated
in the locality of Babahoyo.
Literature cited
Acevedo, M., Reyes, E., Castrillo, W., Torres, O., Marín, C., Álvarez, R., &Torres,
E. (2010). Estabilidad fenotípica de arroz de riego en Venezuela utilizando
los modelos Lin-Binns y AMMI. Agronomía Tropical, Maracay 60 (2), 131-
138. https:// https://dialnet.unirioja.es/servlet/articulo?codigo=5238317
Balzarini, M., & Di Rienzo Julio, A. (2016). Info-Gen versión 2016. Facultad de
Ciencias Agropecuarias, Universidad Nacional de Córdoba. Argentina.
https://www.info-gen.com.ar (software estadístico aplicado a genética).
Benítez E., A., Valencia A., J., Estrada S., E., & Baena, D. (1988). Caracterización
fenotípica de algunas introducciones del banco de germoplasma
del Lulo. Solanum quitoensis Lam. Acta Agronómica, 38(3-4), 34-
49. https://revistas.unal.edu.co/index.php/acta_agronomica/article/
view/86536/74610
Cedeño Dueñas, J., Cedeño García, G., Alcivar Alcivar, J., Cargua Chávez, J.,
Cedeño Sacón, F., Cedeño García, G., & Constante Tubay, G. (2018).
Incremento del rendimiento y calidad nutricional del arroz con fertilización
NPK complementada con micronutrientes. Scientia Agropecuaria, 9(4),
503-509. https://doi: 10.17268/sci.agropecu.2018.04.05
Line 2
Line 10
Line 13
Line 18
Line 25
Line 30
Line 32
Line 38
Line 40
Line 41
Line 42
Line 43
0
2
4
6
8
10
12
14
6000 6300 6600 6900 7200 7500 7800 8100 8400 8700
Relative Variability ( %)
Yielding kg.ha
-1
Low yielding
Low stability
High yielding
Low stability
Low yielding
High stability
High yielding
High stability