Heritability and Correlation for Nitrogen (N₂) Fixation and Related Traits in Peanut (Arachis hypogaea L.)

MATERIALS AND METHODS

Plant materials: Three peanut lines (PI 268770, PI 269109 and PI 152133) selected for high top dry weight, number of nodules and nodule weight from previous evaluation (Toomsan et al., 1991) and two high yielding cultivars (KKU 1 and KKU 72-1) in Thailand were crossed in 5x5 diallel design without reciprocals. The resulting F₁ hybrids were grown for seed increase. The F₂ crosses were further evaluated for combining ability and heritability for the characters related to N₂ fixation and agronomic traits and the results were reported elsewhere (Phudenpa et al., 2004).

Six crosses (KKU 1xPI 152133, PI 268770xPI 269109, KKU 1xKKU 72-1, PI 268770xKKU 72-1, KKU 1xPI 269109 and KKU 72-1x PI 269109) with high general performance were selected and the randomly chosen 20 F₂ families of each cross were planted for seed increase in F₃ generation. The F₄ seeds were evaluated for heritability within crosses and the results were reported previously (Pimratch et al., 2004b).

Based on the evaluation in the F₄ generation, the F₄ families were subjected to divergence selection for both high and low performance (bi-directional selection). The numbers of selected families for all crosses were not exactly equal, ranging from 10 to 14 families with the total number of 74 families and then they were available for evaluation in the F₅ generation.

Experimental procedures: The experiment was conducted at the agronomy farm of Khon Kaen University in the rainy season 2003 with supplemental irrigation under limited soil nitrogen conditions. Lime (CaO) at the rate of 625 kg ha^-1 was incorporated into the soil during soil preparation. The 74 families were laid out in a Randomized Complete Block Design (RCBD) with four replications. A non-nodulating line (referred to as non-nod) was also included as a reference plant for nitrogen determination. The plot size was three-row plot with 2 m in length and spacing of 50 cm between rows and 20 cm between plants within a row. Seeds were treated with captan at the rate of 10 g kg ^-1 seed to control seedling blight disease. The seeds were over-planted and later thinned to one plant per hill at seven days after emergence. Chemical fertilizers as P₂O₅ and K₂O at the rates of 56.25 and 37.5 kg ha^-1, respectively, were applied at 20 days after planting and gypsum (CaSO₄) at the rate of 312.5 kg ha^-1 was applied at 45 days after planting. Neither nitrogen fertilizer nor rhizobium inoculation was applied to the crop because the experimental site had history of peanut cultivation in previous years and peanut did not respond to the inoculation (Toomsan et al., 1988). Weeds, insects and diseases were adequately controlled during crop growth.

Ten plants in the middle row of each plot were harvested at maturity. Care had been taken to remove nodules from the soil as many as possible. Marketable mature pods were removed and air-dried. Plants were cut at crown level, roots after removal of the nodules were discarded and above ground parts (stems and leaves)and nodules were oven-dried at 80°C for 48 h. Dried pods were counted, shelled and then data were recorded for top dry weight, nodule dry weight, pod number, seed number and 100 seed weight on per plant basis. The derived data were also calculated for harvest index and shelling percentage.

To determine fixed nitrogen, dry aerial parts (stems and leaves) were chopped into small pieces and mixed thoroughly for individual samples. Small amounts of samples were taken from the bulked samples of 10 plants and ground for nitrogen determination by micro-Kjeldahl method (Black, 1965). Fixed nitrogen was calculated as;

Fixed nitrogen = total nitrogen of each genotype-nitrogen in reference plant (non-nod)

Data analysis: The data were subjected to analysis of variance using random statistical model as follows;

Y_ij = μ + r_i + F_j + E _ij,

where Y_ij, μ, r_i, F_j and E_ij represent observed value, overall mean, replication effect, family effect and error associated with observed value, respectively.

Duncan’s multiple range test was used to compare means (Gomez and Gomez, 1984).

Heritability in broad sense (H) was computed as suggested by Fehr (1987).

where Ó²_G and Ó²_E represent the variance components due to family and experimental error, respectively.

Heritability estimates were grouped as high (>0.50), moderate (0.20-0.50) and low (<0.20) as suggested by Stansfield (1986). Simple correlation coefficients were calculated using observed values of the plots. Most statistical analyses were performed using MSTAT-C package (Bricker, 1989), except for the partitioning of variance components and the calculation of heritability estimates which were performed manually using Microsoft Excel program.

RESULTS AND DISCUSSION

Heritability: Genotypes differences were highly significant (p<0.01) for all N₂ fixation and related traits and all agronomic traits, confirming the presence of variability in the genetic materials (data not reported). Previous studies also revealed significant differences in their parents for these traits (Pimratch et al., 2004a). The significant differences among progenies allowed us to further process the data to determine the heritability. Estimates of heritability for N₂ fixation and related traits ranged from 0.33 to 1.00 (Table 1). Most heritability estimates were high (>0.50) except one moderate heritability estimate (0.33) for leaf color score in the cross PI 268770xKKU 72-1. As the progenies evaluated were previously subjected to selection for high and low performance for nitrogen fixation and pod yield in stead of random selection, the more precise term used should be repeatability. Heritability is referred to either random segregating or naturally available population (Fehr, 1987).

Similar results of relatively high broad-sense heritability estimates have been reported for nitrogenase activity (0.60 to 0.82) and shoot dry weight (0.63 to 0.85) (Arrendell et al., 1986). However, Arrendell et al. (1988) reported very low realized heritability estimates for fresh plant weight (0.01), nodule number (0.05) and dry weight (0.06) and acetylene reduction (0.31) because of the insufficient control of environmental variation. The contrasting results might be due to differences in experimental conditions and materials used. Denison et al. (1985) found in soybean that variations in N₂ fixation and related traits were affected by stresses and availability of soil nitrogen. In high available soil nitrogen plants were less dependent on fixed nitrogen and available soil nitrogen can alter relationships between nitrogen fixation and related traits. Present results would support the conclusion that evaluation of nitrogen fixation requires low soil nitrogen conditions and well-managed experiment.

Predominance of high heritability estimates for all traits in all crosses indicated that genetic variability is conserved in the F₅ generation and divergence selection in the F₄ generation had no significant effect on heritability estimates. These traits were highly heritable and less affected by environmental conditions under insufficient soil nitrate. Selection for these traits would be more effective and successful in advanced generations or even in early generation without going for further progeny testing.

Table 1:	Heritability estimates for Leaf Color Score (LCS), Nodule Dry Weight (NDW), Shoot Dry Weight (SDW), Total Nitrogen (TN) and Fixed Nitrogen (FN) of six peanut crosses in the F5
1/ = Pale yellow ; 4 = Dark green

Table 2:	Heritability estimates for Pod No. (PN), Pod Weight (PW), Seed No. (SN), Seed Weight (SW), Seed Size (SS), Shelling Percentage (SP) and Harvest Index (HI) of six peanut crosses in the F5 generation

Table 3:	Correlation coefficients between pairs of characters related to nitrogen fixation of six peanut crosses in the F5 generation
*,** significant at 0.05 and 0.01 probability levels, respectively; Leaf Color Score (LCS), Nodule Dry Weight (NDW), Shoot Dry Weight (SDW) and Fixed Nitrogen (FN)

Similar to those for N₂ fixation and related traits, the heritability estimates for all agronomic characters were moderate to high ranging from 0.37 to 1.00 (Table 2). Most of them were high (>0.50) except for shelling percentage in the cross KKU 1xKKU 72-1 (0.37). Earlier studies pertaining to heritability for agronomic traits in peanut have indicated variable results. Relatively high heritability estimates for both broad sense and narrow-sense for and pod length but relatively low for pod yield per plant were previously reported (Ali and Wynne, 1994; Ali et al., 1995). Green et al. (1983) reported moderate heritability estimates for yield and fruit traits, suggesting selection in later generations. The relatively high heritability estimates in this study should be resulted from genetic divergence in parental materials and the conservation of genetic variability after divergence selection in the F₄ generation.

Correlation between traits: Significantly moderate to high correlation coefficients were found for all correlation combinations in the crosses PI 268770xPI 269109 (r = 0.68* to 0.92**) and KKU 1xKKU 72-1 (r = 0.65* to 0.94**) (Table 3). Significant correlation coefficient was not found in the cross KKU 72-1xPI 269109. Most correlation coefficients were highly significant and positive in the cross KKU 1xPI 269109 except for the correlation between nodule dry weight and leaf color score. Correlation coefficient between leaf color score and nodule dry weight was significant in the cross KKU 1xPI 152133. Correlation coefficients between fixed nitrogen and nodule dry weight and between fixed nitrogen and top dry weight were also significant and positive. In the cross PI 268770xKKU 72-1, Leaf color score was positively correlated with nodule dry weight. Nodule dry weight was also positively correlated with top dry weight and top dry weight and fixed nitrogen were well associated positively. In general, fixed nitrogen had strong, positive and consistent relationship with top dry weight across the crosses except in the cross KKU 72-1xPI 269109.

All the traits were generally well associated, it is interesting to note here that degrees of associations among N₂ fixation and its components varied depending on crosses. This might be due to the differences in genetic backgrounds of their parents. Stalker et al. (1994) found that total dry matter accumulation was independent of root nodule or rate of fixation. Thus, correlations for these characters were not consistent across the crosses. However, the correlations between N₂ fixation and its surrogate traits especially shoot dry weight were high in most crosses. This might be due to the close association between leaf nitrogen and carbon assimilation (Sinclair et al., 1993). Shoot dry weight is a good indicator for N₂ fixation and should be use as a selection criterion. Because larger biomass crops require more nitrogen, N₂ fixation will be increased with as biomass yield is increased (Herridge and Rose, 2000). However, this approach might be not necessary to increase pod yield, but truly improves capacity of N₂ fixation sufficient for nitrogen requirement of larger crops.

Breeding for improved N₂ fixation has not proven the end unto itself and improved yield and agronomic traits should be considered. Associations between traits related to nitrogen fixation and traits related to pod yield were studied to determine the consequence of selection.

Table 4:	Correlation coefficients between pairs of agronomic characters and characters related to nitrogen fixation of six peanut crosses in the F5 generation
*,** significant at 0.05 and 0.01 probability levels, respectively; Pod No. (PN), Pod Weight (PW), Seed Weight (SW) and Seed Size (SS), Shelling Percentage (SP), Harvest Index (HI), Leaf Color Score (LCS), Nodule Dry Weight (NDW), Total Dry Weight (TDW) and Fixed Nitrogen (FN)

Correlations between N₂ fixations parameters and agronomic characters varied, depending on crosses (Table 4). In the cross PI 268770xPI 269109, most correlation coefficients were not significant and significant and positive correlation coefficients were found between leaf color score and pod number (r = 0.69*) and between leaf color score and seed size (r = 0.67*). Similar results were observed in the cross KKU 72-1xPI 269109 in which only correlation coefficient between top dry weight and shelling percentage (r = 0.56*) was significant and positive. High and positive associations were found between seed size and all nitrogen fixation parameters (r = 0.83** to 0.94**) in the cross KKU 1xKKU 72-1 and nodule dry weight was also correlated positively with pod weight (r = 0.67*) and seed weight (r = 0.71*). In the cross KKU 1xPI 269109, pod weight and seed weight were both positively correlated with all traits related to nitrogen fixation. Nodule dry weight and top dry weight were also positively correlated with pod number and seed size. In the cross KKU 1xPI 152133, seed size was positively correlated with nodule dry weight, top dry weight and fixed nitrogen and nodule dry weight was also positively correlated with pod number, pod weight and seed weight. In the cross PI 268770xKKU 72-1 pod number was positively correlated with all nitrogen fixation parameters. Pod weight and seed weight were positively correlated with leaf color score and top dry weight. Seed size was positively correlated with nodule dry weight and harvest index was positively correlated with leaf color score.

Implications for breeding for N₂ fixation: The associations between traits related to N₂ fixation and traits related to pod yield were not as high as the associations among each group of characters and the associations were low for shelling percentage and harvest index. This might be due to the fact that higher portion of fixed nitrogen was partitioned to shoot rather than to seed. Residual benefits of the stover after the removal of marketable pods on soil N fertility should result. However, N₂ fixation may be low in high nitrate soils and should be future priority of breeding programs. Selection for nitrate tolerance should increase percent nitrogen under non-limited soil nitrogen conditions because it may be impossible to produce a legume solely dependent upon N₂ fixation for growth (Herridge and Rose, 2000). Burias and Planchon (1990) found in soybean that seed yields of F₄ derived lines were significantly correlated with N₂ fixation of their parents and also correlated with nodule volume and nodule weight particularly at each cross level, suggesting possibility to improve yield.

Present results indicate that it is possible and effective to improve N₂ fixation in this peanut population. Shoot dry weight is the best choice of characters alternative to direct N₂ fixation for screening of large samples under limited soil nitrogen conditions. Simultaneous improvement of N₂ fixation and pod yield is, in lesser extent, also possible and selection for both N₂ fixation and pod yield is essential because low association between traits and the associations varied depending on crosses. This information should help breeders to formulate appropriate breeding strategies for improving N₂ fixation and pod yield.