(go to
KOMUGI Home) (go
to WIS List) (go to NO.80
Contents)
Results and
discussion
The analysis of variance separately conducted for each environment
indicated significant differences among the varieties for gain
yield, number of grains per ear, 1000 grain weight and number of
ear-bearing tillers. 'Me pooled analysis (Table
1) based on
Eberhart and Russell (1966) revealed significant differences among
the varieties for all the characters. The linear effect of the
environment was also highly significant for all the characters,
indicating that salt stress widely influenced the performance of the
varieties. 'Me predictable linear variety x environment (salt stress)
interaction component was also significant for all the
characters.
The estimates of the three stability parameters mu (mean),
bi (linear regression coefficient) and
S2di (deviation from linear regression) for
grain yield and its component characters for each variety are given
in Table
2 and
Table
3, respectively.
The varieties were classified into the following groups
(Table
4).
(A) Absence of G x E interactions: This group comprises of varieties
with non-significant bi and S2di , thereby
suggesting their wide adaptability. However none of the varieties
fell in this group for any of the characters studied, thereby
indicating the predominance of genotype x salt stress
interactions.
(B) Presence of G x E interactions: In this group three types of
varieties must be considered. (i) The major portion of G x E
interactions was explained by the linear environmental change, i.e.,
a significant bi and a non- significant S2di.
(ii) Along with a significant bi the estimates of
S2di were also significant, suggesting a
significant contribution of linear and non-linear components to the
performance of the varieties. (iii) The major portion of G x E
interactions was explained by the nonlinear component, indicating
high unpredicatability hence instability of the varieties.
Finlay and Wilkinson (1963) considered linear regression coefficient
as a measure of stability, whereas Eberhart and Russell (1966)
emphasized that both linear (bi) and non-linear
(S2di) components of the interaction should be
considered for judging the phenotypic stability of a genotype. Breese
(1969) and Paroda and Hayes (1971) underlined that linear regression
coefficient should simply be regarded as a measure of response of a
particular genotype, whereas the deviation which could not be
explained by the linear regression should be considered as a measure
of stability; a genotype with the lowest deviation being the most
stable and vice-versa. Keeping these in view the varieties classified
to the first group are of more importance.
Varietal classification given in Table
4 indicated that
for a majority of varieties a significant portion of G x E
interactions was explained by the linear environmental change, i.e.,
a significant bi and a non-significant
S2di. In this group, varieties Hybrid 65 and
PBW 65 showed significantly higher mean yield along with bi nearly
equal to unity. As shown in Table
2 and
Fig.
1, varieties like
Hybrid 65 and PBW 65 had high grain yield/ row under normal
environment and the reduction of yield was relatively small under
salinity stress of ECiw 20 and 30 dSm-1. These
varieties also gave moderately better yields under alkalinity stress
of pH 9.2 and 9.4. Similar result was also seen in varieties Sujata
and Hybrid 65 for number of grains per ear, and in varieties HDR 77,
Sujata, Hybrid 65, PBW 65 and C 306 for 1000 grain yield, while no
variety was stable for the number of ear-bearing tillers. It can be
thus inferred that the variety Hybrid 65 was the most suitable
variety under salt stress conditions.
<--Back | -->Next
(go to
KOMUGI Home) (go
to WIS List) (go to NO.80
Contents)