Analysis of Genetic Parameters and Trait Relationship for Seed Yield and its Attributing Components in Lentil (Lens culinaris Medik.)

Lentil (Lens culinaris Medik. ssp. culinaris) 2n=2x=14 is the oldest domesticated grain legume that belongs to the family Fabaceae. It is grown extensively for its nutritious lens-shaped seed in marginal and rainfed areas of temperate countries and most tropical highlands. Roots of lentil go as back as 8000-9000 BC (Harlan, 1992) when it was originated from L. culinaris Medik. ssp. orientalis in Fertile Cresent of Eastern Asia (Cubero, 1981). After soybean and hemp, lentil is having the third-highest seed protein content among plant-based food; making it a vital source of inexpensive protein (Bhatty, 1988). India ranks second in lentil production in the world after Canada with a total production of 1.10 million tons yielding out of 1.30 million hectares (INDIASTAT, 2020). The average productivity of lentil is 847 kg/ha in India which is quite low as compared to average productivity of the world of 1195 kg/ha (FAOSTAT, 2019). Therefore, to address the issue of low productivity, there is a need to tailor large number of superior cultivars using different breeding strategies. Genetic variability is essential in any crop improvement programme. Plant breeders need to know about the variability in primary gene pool and relationship between characteristics and yield to improve complicated traits like seed yield. Seed yield is an economically important trait in virtually all crops. It is complex in nature and under polygenic control and it exhibits multiplicative interactions with its attributing traits and environments (Sharma et al., 2020). Therefore, to yield gains in productivity, information on genetic variability and traits association of seed yield with other attributing traits is highly valuable as it helps in an appropriate selection approach. The pace and magnitude of genetic improvement through selection or hybridization can be determined by the magnitude of genetic variability. Furthermore, selection for seed yield could be made more effective and productive when contribution of each causal effect to yield is quantified through path analysis (Dewey and Lu, 1959). Thus, the current study was carried out to get better insights into genetic variability, trait association between seed yield and its attributing components and to assess the direct and indirect effects of various attributing components on seed yield.

Forty-three diverse genotypes of lentil (Lens culinaris Medik.) were evaluated in field trials for eleven quantitative traits (Table 1). The accessions were grown in a randomized block design with three replications, with a plot size of 7.2 m² (0.30 cm × 4 m × 6 rows) on research farms of Pulses Section, Department of Genetics and Plant Breeding, CCS HAU, Hisar, during 2019-20 and 2020-21 Rabi cropping seasons. The experimental site is subtropical region, located on latitude of 29°10’ North and longitude of 75°46’ East and 215.2 m above mean sea level. The crop was raised by adopting all recommended package of practices. The data was observed on 11 quantitative traits viz., days to 50% flowering, days to maturity, plant height (cm), number of pods per plant, number of primary branches, number of fruiting branches, seeds per pod, 100-seed weight (g), biological yield per plot (kg), harvest index (%) and seed yield per plot (kg). Data were pooled across two seasons to carry out statistical analyses. The analysis of variance was carried out as suggested by Fisher (1925). The genetic parameters viz., genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) were calculated as per standard procedure given by Burton and De vane (1953). The heritability in the broad sense and genetic advance were determined by using methodology of Johnson et al., (1955). The path coefficients were obtained by following the method of Dewey and Lu (1959). Statistical analyses were done using R studio and INDOSTAT software.
 

Variation among the pooled mean over two years of 43 diverse genotypes for 11 quantitative traits was highly significant (p<0.01) for all the eleven traits which validated further statistical and genetic analysis (Table 2). The scrupulous analysis of variance, mean, standard errors of mean and critical difference (CD) revealed highly significant differences among the genotypes for all 11 quantitative traits studied. This connoted the presence of ample amount of genetic variability among the genotypes under study. Significant genetic variability in lentil has been reported by several researchers for different traits (Kumar et al., 2015; Kumar et al., 2020; Sharma et al., 2020).
 

       
Genotypic coefficient of variation (GCV) and phenotypic coefficient of variation (PCV) is a potent tool for estimation of variation in breeding material and determining to what extent environment alterations have their role in output of a trait. The values of variance, coefficient of variation, heritability and genetic advance for different traits are presented in Table 3. Phenotypic variance denoting total variance was maximum for number of pods per plant (172.92) followed by plant height (36.15) and harvest index (14.02). The PCV for all the traits was higher than GCV though the differences were less. The narrow gap between PCV and GCV revealed low influence of environment in the expression of these characters. Al-Aysh (2014), Hussan et al., (2018) and Kumar et al., (2020) also found PCV values to be slightly higher than that of GCV for all the characters., The GCV and PCV varied from 2.57% to 17.56% and 2.62% to 17.78%, respectively. According to Deshmukh et al., (1986), GCV and PCV values less than 10% are regarded as low, whereas values greater than 20% are considered as high and values between 10 and 20% to be medium. Based on this classification, moderate GCV and PCV values were observed for 100-seed weight (17.56% and 17.78%, respectively), seed yield per plot (16.14% and 16.62%, respectively), biological yield per plot (14.93% and 15.61%, respectively), plant height (13.45% and 13.67%, respectively), number of fruiting branches (12.27% and 12.94%, respectively), number of pods per plant (11.24% and 11.61%, respectively), seeds per pod (10.36% and 10.92%, respectively) and harvest index (9.97% and 10.83%, respectively), suggesting that selection for these traits would be amenable for genetic improvement. However, low GCV and PCV values were observed for days to maturity (2.57% and 2.62%, respectively), days to 50% flowering (4.02% and 4.08%, respectively) and number of primary branches (8.49% and 9.18%, respectively).
 

       
The estimates of genetic coefficient of variations along with heritability and genetic advance would be beneficial in predicting gain under selection (Assefa et al., 1999; Sahu et al., 2015). The heritable portion of phenotypic variance considered by value of σ²g relative to σ²p expressed as h²_BS was very high (89.9% to 97.5%) for all 11 traits except harvest index (84.7%) and number of primary branches (85.4%) that showed most of the traits had moderate heritability (Table 3). Thus, all the traits except harvest index and number of primary branches were affected by environmental fluctuation to minimal extent. Furthermore, high heritability estimates (h²_BS) with corresponding high genetic advance as percent of mean (GAM) is more efficient and pragmatic approach for selection than that with low GAM. The estimates of heritability in broad sense (h²_BS) and corresponding GAM, both were high for 100-seed weight, seed yield per plot, biological yield per plot, plant height, number of fruiting branches and number of pods per plant. Therefore, expression of these traits was controlled by additive gene action and direct selection would be highly fruitful for their genetic improvement over short span of time. Similar findings were observed for 100-seed weight, seed yield, pods per plant and plant height by Tyagi and Khan (2010), Abdipur et al., (2011), Hussan et al., (2018) and Kumar et al., (2020).
       
Variability and heritability data provide opportunities for genetic improvement in different traits, but they do not suggest any kind of association between them. As a result, understanding the relationships between the traits is helpful in the indirect selection and improvement of economically important traits for a successful breeding program for any crop (Shabanimofrad et al., 2013). In the present study, seed yield per plot was positively correlated with biological yield per plot, number of pods per plant, plant height, days to maturity, number of primary branches, harvest index, days to 50% flowering and number of fruiting branches (Table 4). Thus, selection for these positively associated yield  attributing traits could bring about sufficient gain in seed yield. However, seed yield per plot was found negatively correlated with 100-seed weight (Table 4).
 

       
As the number of factors in correlation studies increases, Pearson’s correlation coefficient may not provide exact representation of association between yield and its contributing traits. In such perplexing situations, path coefficient analysis allows a more in-depth study of specific direct and indirect efforts of trait and thorough investigation of the precise forces acting and quantifies the relative importance of each causal effect (Khan et al., 2016). In the present study, path coefficient analysis has been conducted taking seed yield per plot as dependent variable. The direct and indirect effects of various traits on seed yield are provided in Table 4. The highest positive direct effect on seed yield was exerted by biological yield (0.854) followed by harvest index (0.579), days to maturity (0.078), number of primary branches (0.069), number of pods per plant (0.030) and days to 50% flowering (0.018). Hence, these traits should be given high weightage and positive selection should be done to improve seed yield. However, 100-seed weight (-0.062), seeds per pod (-0.049), plant height (-0.046), number of fruiting branches (-0.045) had negative direct effect on seed yield per plot. These results are in accordance with the findings of previous studies of Younis et al., (2008), Aghili et al., (2012), Dalbeer et al., (2015). Furthermore, Latif et al., (2010) found negative direct effect of 100-seed weight on seed yield whereas Kumar et al., (2020) found negative direct effect of seeds per pod on seed yield. However, days to flowering, number of pods per plant, plant height, number of fruiting branches and seeds per pod had positive indirect effect on seed yield via biological yield. These indirect effects had not only validated the low magnitude direct effect but also explained highly significant positive association of these traits with seed yield. The apparent inconsistency between Pearson’s correlation and path analysis was most likely due to the fact that the former only evaluates mutual association without taking into account the cause, whereas the latter defines the causes and assesses their relative importance (Bhatt, 1973). The presence of a low residual effect (0.124) showed that the independent characters made a significant contribution to the dependent trait i.e. seed yield and the characters selected for path analysis were acceptable and appropriate.
       
Crop yield is mostly determined by biological yield and partitioning accumulated biomass to reproductive structures (Andrade et al., 1999). The fraction of total biomass devoted to reproductive tissues is known as reproductive partitioning (Hay, 1995; Sinclair, 1998). Any crop plant’s productivity is determined not only by its photosynthetic efficiency, but also by the successful translocation of assimilates to the seeds, as evaluated by the harvest index. Thus, selection for biological yield, harvest index and seed yield per se could bring about significant genetic gain as these traits have relatively high coefficient of variation along with high heritability and genetic advance as percent of mean. These traits were further reinforced by positive direct effect on seed yield. Therefore, biological yield, harvest index and seed yield may be used as better selection indices for lentil crop improvement. The genotype LL 931 (4.096 kg) is highest biomass yielder followed by DPL 15 (3.996 kg) and IPL 316 (3.972 kg) and for harvest index L 4717 (44.23%) followed by LH 18-04 (43.53%) and LH 17-19 (40.40%). High seed yield potential per se is undoubtedly an important consideration when it comes to selection of genotypes. Amongst the genotypes, IPL 316 (1.373 kg) followed by LH 18-04 (1.346 kg) and LH 17-19 (1.301 kg) were highest seed yielders (Table 5).
 

Genetic improvement and tailoring of high yielding cultivars require knowledge of amount and nature of genetic variability that is present in the primary gene pool. Therefore, we examined genetic variation for 11 quantitative traits of 43 diverse lentil genotypes in order to have better insight into genetic variability, character association and direct and indirect effects of various yield components in lentil. High PCV and GCV coupled with high heritability and genetic advance were observed for 100-seed weight, seed yield per plot and biological yield suggesting that expression of these traits is controlled by additive gene action and selection would be highly fruitful. Most of the quantitative traits were positively correlated with seed yield except 100-seed weight. It is evident from path coefficient analysis that biological yield, harvest index, days to maturity, number of primary branches, number of pods per plant and days to 50% flowering were the highest direct contributor to seed yield. Therefore, to obtain high yielding transgressive segregants, the genotypes LL 931, L 4717, IPL 316, LH 18-04, LH 17-19 and DPL 15 should be used in hybridization programmes to exploit genetic variability present in lentil stocks.

None.