Introduction
An organized and prosperous hybridization program depends upon an understanding of genetic architecture and inheritance of quantitative characters of prominent cultivars. For such studies various workers have used different methods but method of Hayman (1954a) and Jinks (1954) has most commonly been used. Even though genetic analysis of quantitative characters has widely been reported yet evidences of some partial dissimilarity among research workers have been observed in certain traits. Such divergence could be credited to variation in genetic material, environmental conditions and methods engaged for such studies.
In cotton different authors, Bhatade (1981), Azhar et al. (1994), Khan et al. (1995), Saeed et al. (1996), Ahmad et al. (1997), Murtaza et al. (1992 b), Busharat et al. (1998 b), Busharat et al. (1999 ) and Subhan et al. (2001), detected significant additive effects with partial dominance for yield of seed cotton plant-1 and lint percentage. Others like Rehman et al. (1988), Tariq et al. (1992) and Khan et al. (1993), have reported over dominance type of gene action.
This contradiction in results may be due to different breeding material utilized under different environmental conditions. The current strategy involves high potential commercial varieties being widely used in crossing programs. Such type of studies is expected to provide practical information for scheduling specific and precise breeding programs.
Materials and Methods
All possible F1 crosses, including reciprocals, were made among eight cultivars of cotton viz. CYTO 9/91, B-496, SLS-1, Niab-78, NIAB-313/12, B-622 and Niab-78, NIAB-313/12, B-622, NIAB-92 and CYTO-11/91. Fifty six crosses and eight parents were sown at Faculty of Agriculture, Gomal University, D.I.Khan, during 1996-99 in Randomized Complete Block Design with three replications. Row to row and plant-to-plant spacing were kept as 75 cm and 30 cm, respectively. F1 hybrids plants from each cross and parents were self pollinated to raise F2 progeny. F2 population was grown during May 1997 in a triplicate progeny row trial employing the said design. The plot size for each cross was 3 x 6 m2. The data were recorded on ten randomly selected plants from each entry for two characters viz. yield of seed cotton plant-1 and lint percentage in both the generations. The plot means were used for analysis of variance and where the means were significant, data were further subjected to diallel analysis technique developed by Mather and Jinks (1977).
Results and Discussion
The analysis of variance showed highly significant differences among progenies
for yield seed cotton plant-1 and lint percentage in F1
and F2 generations. This indicated presence of adequate genetic variability
which could be exploited in different crossing programs (Table
1). There are indications that additive genetic effects were present for
lint percentage in both the F1 and F2 generations as the
mean squares due to male and female items in the basic diallel ANOVA were significant,
while interaction and reciprocal mean squares were concluded as non significant
(Table 2 a, 3 a).
| Table 1: |
Estimates of mean squares and F ratios of different quantitative
traits of cotton (Gossypium hirsutum L.) in F1 and F2
generation |
|
| **, Highly significant |
| Table 2a: |
Analysis of variance of F1 of diallel data for
lint percentage |
|
| NS, non-significant; *, significant (P<0.05) and **, highly
significant (P<0.01) |
| Table 2b: |
Mean data over replications and reciprocals |
|
| Regression coefficient (b) = 0.9517±0.0888, difference
of b from zero (b0) = 10.7164** different for b from unity (b1) = 0.5443NS
NS, non-significant **, highly significant |
| Table 2c: |
Analysis of variance for arrays |
|
| NS, non-significant |
| Table 3a: |
Analysis of variance of F2 of diallel data for
lint percentage |
|
| NS, non-significant,*, significant (P<0.05) and**, highly
significant (P<0.01) |
| Table 3b: |
Mean data over replications and reciprocals |
|
| Regression coefficient (b) = 0.9663±0.081, difference
of b from zero (b0) = 11.9629** different for b from unity (b1) =0.4175NS
NS, non-significant **, highly significant |
| Table 3c: |
Analysis of variance for arrays |
|
| **, Highly significant |
| Table 4a: |
Analysis of variance of F1 of diallel data for
yield of seed cotton per plant |
|
| NS, non-significant, *, significant (P<0.05) and **, highly
significant (P<0.01) #Tested against reciprocal mean square. |
| Table 4b: |
Mean data over replications and reciprocals |
|
| Regression coefficient (b) = 0.8782± 0.1256, difference
of b from zero (b0) = 6.992** different for b from unity (b1) = 0.9701NS
NS, non-significant **, highly significant |
| Table 4c: |
Analysis of variance for arrays |
|
| **, Highly significant and *,significant |
| Table 5a: |
Analysis of variance of F2 of diallel data for
yield of seed cotton per plant |
|
| NS, non-significant, *, significant (P<0.05) and **, highly
significant (P<0.01) |
| Table 5b: |
Mean data over replications and reciprocals |
|
| Regression coefficient (b) = 0.9882±0.0759, difference
of b from zero (b0) = 13.0197** different for b from unity (b1) =0.1546NS
NS, non-significant **, highly significant |
| Table 5c: |
Analysis of variance for arrays |
|
| **, Highly significant and *,significant |
|
| Fig. 1: |
Wr/Wr graph for lint percentage (F1) |
|
| Fig. 2: |
Wr/Wr graph for lint percentage (F2) |
In so far as the character like seed cotton in F1 is concerned,
all the four items were significant, consequently, additive genetic effects
along with non additive genetic effects and maternal effects were present (Table
4a).
|
| Fig. 3: |
Wr/Wr graph for seed cotton yield per plant (F1) |
|
| Fig. 4: |
Wr/Wr graph for seed cotton yield per plant (F2) |
Further more, all the three items were exposed as significant while the fourth
item like reciprocal was uncovered as non significant depicted thereby both
additive and non additive effects but nothing like maternal effects were present
in case of yield of seed cotton plant-1 in F2 generation (Table
5a). The regression coefficients of seed cotton in F1 (b) = 0.8782±0.1256,
F2 (b) = 0.9882 ± 0.07590 and as well as of lint percentage
in F1 (b) = 0.0888±0.9517, F2 (b) = 0.9663±0.081,
did not differ significantly from unity and of course differed significantly
from zero, suggesting the absence of epistasis and revealed the fulfilment of
the assumptions for diallel analysis (Table 2b, 3b,
4b, 5b). The non-significant differences
for Wr+Vr indicated the absence of dominance genetic effects, whereas the non
significant differences for Wr-Vr showed the absence of non allelic interaction
and thus additive dominance model is fully adequate in case of lint percentage
in F1 population (Table 2c). While both the Wr
+ Vr and Wr – Vr reflected the significant variance ratio suggesting the
presence of epistasis and thus additive dominance model is molded partially
adequate for lint percentage in F2 population (Table
3c). At the same time as significant variance ratio (Table
4c, 5c) for both the analysis of Wr + Vr and Wr –
Vr, were concluded, thus there was present evidence for non allelic interaction
which turned the model partially adequate with reference to seed cotton plant-1
in both F1 and F2 populations.
Graphic position (Fig. 1, 2 and 4)
revealed that regression lines for the said characters touched Wr axis above
the origin hence signifying thereby additive type of gene action with partial
dominance type of gene action except yield of seed cotton plant-1
(Fig. 3) in F1 generation which depicted over dominance
type of gene action as the regression line intercepted the Wr axis below the
origin. Such type of situation is represented in a diallel analysis when F1
hybrids score more than either of the parents, which is nothing but the reflection
of hybrid vigor or heterotic effects in this connection. The distribution of
array points along the regression line showed that genotypes Niab-313/12 (Fig.
1), Niab-313/12 (Fig. 2) in case of lint percentage while
Niab-313/12 (Fig. 3) and Niab-313/12 (Fig. 4)
in case of seed cotton contained the maximum number of dominant genes in contrast
to Cyto-9/91 in all the characters carried the maximum recessive alleles. Similarly,
several workers like Murtaza et al. (1992), Azhar et al. (1994),
Khan et al. (1995), Saeed et al. (1996), Busharat et al.
(1998), Busharat et al. (1999 b) and Subhan et al. (2001) have
reported additive type of gene action with partial dominance for these characters.
However, findings of some others like, Rehman et al. (1988), Tariq et
al. (1992), Khan and Khan (1993) and Murtaza et al. (1995) do not
agree with these observations, which might be due to different genotypes used
under different agro-ecological factors.
Additive type of gene action with partial dominance was accomplished which discloses beneficial outlines for selection of superior hybrids for the characteristics under study in both the generations. In case of yield of seed cotton-1, over dominance type of gene action was also exposed specially in F1 generation, which reflected the manifestation of heterotic effects.