Correlation and path coefficient analysis of grain yield and its contributing traits in salt-tolerant rice genotypes under sodic soil conditions

Introduction

Oryza sativa L. (rice) sustains more than 3.5 billion people globally and supplies ~20% of dietary energy (FAO, 2022). India is the world's second largest producer with annual output exceeding 116 million tonnes. Productivity in salt-affected and sodic soils affecting ~7 million ha of the Indo-Gangetic Plain remains a critical challenge (Nayak et al., 2016). Grain yield is a complex polygenic trait shaped by multiple interrelated component characters. Correlation studies quantify trait associations but cannot distinguish direct causal effects from indirect influences. Path coefficient analysis, developed by Wright (1921) and standardised for plant breeding by Dewey and Lu (1959), partitions the correlation coefficient into direct and indirect effects the sum of which equals the total phenotypic correlation with grain yield (Singh and Chaudhary, 1977;Milligan et al., 1990;Saleh et al., 2020). This approach is critical because a trait may show a high correlation with grain yield yet have a negligible or even negative direct effect, implying that its association operates entirely through indirect pathways (Saleh et al., 2020). Salt-affected soils, including sodic and saline types, are the major abiotic constraints limiting rice productivity, particularly in the Indo-Gangetic Plain of eastern India where pH often exceeds 9.0 (Kumar et al., 2022). The development of high-yielding salt-tolerant varieties requires a thorough understanding of trait-yield relationships. Despite extensive studies in rice on correlation and path analysis (Kumar et [10, 25, 40, 7, 38, 23, 34, 34, 19, 20, 16, 6, 34, 39, 18, 17, 35] ; Saketh et al., 2023;Alshugeairy et al., 2023;Mahalakshmi et al., 2024;Shajan et al., 2024), information on newly developed salttolerant germplasm lines under multi-environment sodic soil conditions of eastern Uttar Pradesh is limited. The present investigation was therefore undertaken with the objectives to: (i) estimate genotypic and phenotypic correlation coefficients among yield-contributing characters; (ii) determine direct and indirect effects through path analysis; (iii) define reliable selection criteria; and (iv) identify promising genotypes for hybridisation programmes [33, 1, 22, 37] .

Materials and methods

Experimental site

The field experiment was conducted during the Kharif season of 2019 at three experimental sites of Acharya Narendra Deva University of Agriculture and Technology (ANDUAT), Narendra Nagar (Kumarganj), Ayodhya, Uttar Pradesh, India (26

Experimental design

The experiment was laid out in a randomized complete block design (RCBD) with three replications at each experimental environment. Each plot consisted of five rows, each 5 m in length, with a spacing of 25 cm between rows and 15 cm between plants. Recommended fertilizers were applied at the rate of 120:60:40 kg ha⁻¹ of N:P₂O₅:K₂O, respectively. Five irrigations were applied at critical crop growth stages, viz., active tillering, early booting, late booting, heading, and flowering.

Characters studied

Eleven quantitative and yield-related traits were recorded from five randomly selected plants in each plot: (i) days to 50% flowering (DF), (ii) plant height (PH, cm), (iii) productive tillers per plant (PT/P), (iv) panicle length (PL, cm), (v) leaf area index (LAI), (vi) number of fertile spikelets per panicle (NFS/P), (vii) spikelet fertility (SF,%), (viii) biological yield per plant (BYP, g), (ix) harvest index (HI,%), (x) leaf-to-breadth ratio (LBR), and (xi) grain yield per plant (GYP, g).

Statistical analysis

The analysis of variance (ANOVA) was performed following the procedure described by Panse and Sukhatme (1967) using SPSS software version 21.0 (IBM, 2012). Variance components were estimated according to Johnson et al. (1955). Phenotypic correlation coefficients were calculated following Searle (1961), and their significance was tested against the tabulated values of the correlation coefficient (r) as described by Fisher and Yates (1963). Path coefficient analysis was performed according to Dewey and Lu (1959) and Saleh et al. (2020), using the relationship (2020): rᵧᵢ=Pᵧᵢ+Σrᵢᵢ′Pᵧᵢ′, where Pᵧᵢ=direct effect (path coefficient), rᵢᵢ′Pᵧᵢ′=indirect effect of the ith causal variable via the i′th variable, and the sum of direct and all indirect effects equals the total phenotypic correlation with grain yield. Residual: e=√(1-R²). Effect scale Lenka and Mishra (1973): neglected (<0.10), weak (0.10-0.19), moderate (0.20-0.29), strong (0.30-0.99), very strong (>1.00) [27, 13, 14, 36, 11, 7, 34, 21] .

Results and discussion

Mean performance and ANOVA

The analysis of variance (Table 2) revealed highly significant differences (p ≤ 0.01) among the genotypes for all eleven traits, indicating the presence of substantial genetic variability within the evaluated germplasm. Significant genotype × environment (G × E) interactions were also observed for all traits, demonstrating differential genotypic responses across the experimental environments. Grain yield per plant varied from 21.20 g in CSRRIL-01-IR 165 to 33.50 g in the check variety CSR 43, with an overall mean of 27.85 g. The highest grain yield was recorded for CSR 43 (33.50 g), followed by CSR-2748-4441-193 (31.40 g), CSR 2016-IR 18-10 (31.20 g), CSR 27 (30.70 g), CR 2851-S-B-1-2B-1 (30.40 g), and CSR-C27 SM-117 (30.00 g). Plant height ranged from 94.0 to 147.3 cm, with a mean of 125.28 cm, whereas days to 50% flowering varied from 75 to 92 days, with a mean of 80.54 days. The substantial variation observed among the genotypes is consistent with previous reports of significant genetic variability in salt-tolerant rice genotypes evaluated under Uttar Pradesh environments (Pratap et al., 2018;Singh et al., 2020;Kumar et al., 2021). More recently, Rani et al. (2024) evaluated the World Rice Mini Core Collection under sodic soil conditions at Kanpur, Uttar Pradesh, and reported grain yield per plant ranging from 18.6 to 36.2 g, which is comparable with the range observed in the present study [30, 39, 18, 31] .

Correlation analysis

Phenotypic correlation coefficients among the eleven traits are presented in Table 4 and Fig. 1. Grain yield per plant exhibited a highly significant positive correlation with plant height (r = 0.590) and days to 50% flowering (r = 0.478**), while significant positive associations were observed with biological yield per plant (r = 0.418**) and harvest index (r = 0.250*). The positive association between plant height and grain yield suggests that taller genotypes may possess greater biomass accumulation, which could contribute to enhanced yield formation through improved assimilate production and partitioning (Chhangte and Devi, 2019; Babu et al., 2012). Days to 50% flowering exhibited a highly significant negative association with productive tillers per plant (r = -0.519**), suggesting a possible trade-off between crop duration and tiller production under the evaluated conditions (Kahani and Hittalmani, 2015). Grain yield per plant showed a significant negative correlation with productive tillers per plant (r = -0.304*), which may reflect competition for assimilates among tillers within the plant (Dhurai et al., 2016). Leaf area index exhibited a highly significant positive association with the number of fertile spikelets per panicle (r = 0.570**), while the number of fertile spikelets per panicle was also positively associated with spikelet fertility (r = 0.487**). These associations are broadly consistent with earlier reports in rice (Kumar et Pradesh, and reported a similar negative association, suggesting that excessive tiller production may increase intra-plant competition for assimilates and consequently limit grain yield per plant.

In contrast, Saleh et al. (2020), in an evaluation of 22 Egyptian and exotic rice genotypes over two growing seasons (2017-2018), reported highly significant positive correlations of grain yield with panicle number per plant (r = 0.791), 1000-grain weight (r = 0.766**), and number of filled grains per panicle (r = 0.670**).

The differences in the traits most strongly associated with grain yield between the present study and that of Saleh et al. (2020) may be attributed to differences in germplasm composition and environmental conditions, particularly the salt-tolerant CSRderived material evaluated under sodic soil conditions in the present study versus Egyptian and exotic germplasm evaluated under irrigated conditions. Nevertheless, correlation coefficients describe the degree of association between traits and do not necessarily establish their causal contribution to grain yield. Therefore, path coefficient analysis was subsequently employed to partition the observed correlations into direct and indirect effects and to identify traits with greater potential as selection criteria [5, 2, 15, 8, 20, 16, 6, 26, 39, 22, 17, 37, 34] .

Path coefficient analysis

Path coefficient analysis was employed to partition the phenotypic correlations of the component traits with grain yield per plant into direct and indirect effects (Table 5; Fig. 2). The coefficient of determination (R² = 0.5244) indicated that the ten predictor traits collectively accounted for 52.44% of the variation in grain yield per plant, while the residual effect was 0.6896. The relatively high residual effect indicated that additional traits not included in the present analysis may also contribute to variation in grain yield, warranting further investigation of traits such as root architecture and chlorophyll content. Plant height exhibited the highest positive direct effect on grain yield per plant (β = 0.3874), categorized as a strong effect, followed by days to 50% flowering (β = 0.2061) and biological yield per plant (β = 0.2037), both showing moderate positive direct effects. Plant height also exhibited positive indirect effects through days to 50% flowering (0.0915) and biological yield per plant (0.0656). The relatively high direct effect of plant height contributed substantially to its strong positive overall correlation with grain yield (r = 0.590), indicating that plant height may have , who, in an association study involving 24 elite salt-tolerant rice lines under direct-seeded conditions in Andhra Pradesh, reported positive associations of biological yield and harvest index with grain yield at both phenotypic and genotypic levels, highlighting their potential utility as selection criteria under salt-stress conditions [2, 12, 34, 39, 32, 35, 33, 29] . , who observed that although several traits showed significant positive associations with grain yield, some traits had negligible direct effects. In their study, plant height (β = 0.0501) and panicle length (β = 0.0726) exhibited negligible direct effects despite their associations with other yield-related traits. Similarly, the present study revealed negligible direct effects for spikelet fertility, fertile spikelets per panicle, and LBR [37, 1, 24, 6, 34] . Overall, the path analysis indicated that plant height, days to 50% flowering, and biological yield per plant made comparatively larger positive direct contributions to grain yield than the other traits evaluated. Therefore, these traits may be considered promising selection criteria for grain yield improvement in the salt-tolerant rice genotypes evaluated under sodic soil conditions. However, the relatively high residual effect also indicates that grain yield is influenced by additional traits not included in the present analysis, and further multi-environment evaluation would be valuable for validating the stability of these selection criteria.

Acknowledgement

AcknowledgementThe authors gratefully acknowledge the Department of Genetics and Plant Breeding, Acharya Narendra Deva University of Agriculture and Technology (ANDUAT), Ayodhya, for providing the necessary facilities and support for conducting the present study. The authors also acknowledge the Central Soil Salinity Research Institute (CSSRI), Karnal, and other collaborating institutes under the All India Coordinated Rice Improvement Project (AICRIP) for providing the germplasm used in this investigation.

References

1. Alshugeairy ZK, Alogaidi FF, Hadi BH, Hassan WA. Path coefficient analyses of introduced rice. SABRAO J Breed Genet. 2023;55(2):516-524.
2. Babu VR, Shreya K, Dangi KS, Usharani G, Shankar AS. Correlation and path analysis studies in popular rice hybrids. Int J Sci Res. 2012;2:1-5.
3. Bhutta MA, Ahmad HM, Ali Q, Raza A. Correlation and path analysis contributing to yield under drought in rice. Nat Sci Biol Med. 2019;10:38-44.
4. Chandra BS, Kumar SS, Srinivasa Rao V, Rao VT. Genetic variability and character association in rice. J Plant Sci Res. 2009;25:27-31.
5. Chhangte L, Devi TR. Correlation and path analysis in aromatic rice of north-east India. Pharm Innov J. 2019;8:1-4.
6. Devi AS, Hemant BC, Venkataramana G. Path analysis in rice for yield and quality traits. Int J Curr Microbiol App Sci. 2017;6(5):1660-1666.
7. Dewey DR, Lu KH. Correlation and path coefficient analysis of crested wheatgrass. Agron J. 1959;51:515-518.
8. Dhurai SY, Wankhade PS, Lal GM. Correlation and path coefficient analysis in rice. Bioscan. 2016;11(3):2087-2091.
9. Ezeaku IK, Mohammed SG. Character association and path analysis in grain sorghum. Afr J Biotechnol. 2006;5:337-340.
10. FAO. FAOSTAT Database. Rome: Food and Agriculture Organization; 2022. http://www.fao.org/faostat.
11. Fisher RA, Yates F. Statistical Tables for Biological, Agricultural and Medical Research. 6th ed. Edinburgh: Oliver and Boyd; 1963.
12. Herawati R, Alnopri M. Correlations and path analysis for selection in upland rice. Bio Sci Res. 2019;16:1492-1499.
13. IBM Corporation. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp.; 2012.
14. Johnson HW, Robinson HF, Comstock RE. Estimates of genetic and environmental variability in soybeans. Agron J. 1955;47:314-318.
15. Kahani F, Hittalmani S. Association and path coefficient analysis in rice under aerobic condition. J Rice Res. 2015;3:142.
16. Karim D, Siddique MNA, Sarkar U, Hasnat MZ, Sultana J. Correlation and character association of aromatic rice. J Bios Agric Res. 2014;1:34-46.
17. Kayastha P, Chand H, KC B, Pandey B, Magar BR, Bhandari J, et al. Correlation coefficient and path analysis of yield and yield attributing characters of rice (Oryza sativa L.) under reproductive drought stress. Arch Agric Environ Sci. 2022;7(4):564-570.
18. Kumar A, Nanda A, Sethi SK, Bastia DN, Das TR. Path coefficient analysis for yield traits in rice. Int J Bio-resource Stress Manag. 2021;12(4):403-409.
19. Kumar D, Yadav RK, Sheoran RK, Sharma PC. Development of salt-tolerant rice varieties to enhance productivity in salt-affected environments. Environ Sci Proc. 2022;16(1):30. doi:10.3390/environsciproc2022016030.
20. Kumar Y, Singh BN, Verma OP, Tripathi S, Dwivedi DK. Correlation and path coefficient analysis in scented rice. Environ Ecol. 2011;29(3B):1550-1556.
21. Lenka D, Mishra B. Path coefficient analysis of yield in rice varieties. Indian J Agric Sci. 1973;43:376-379.
22. Mahalakshmi PJ, Satish Y, Ramana JV, Raju MRB. Correlation and path coefficient analysis for yield and its component traits in advanced breeding lines of rice (Oryza sativa L.). J Adv Biol Biotechnol. 2024;27(9):769-777.
23. Milligan SB, Gravois KA, Bischoff KP, Martin FA. Crop effect on genetic relationship among sugarcane traits. Crop Sci. 1990;30:927-931.
24. Minnie SM, Kannan R, Arulmozhiselvan K. Path analysis for yield components in rice. Indian J Appl Res. 2013;3(11):283284.
25. Nayak AR, Chaudhary D, Reddy JN. Genetic variability, character association and path analysis in rice. Electron J Plant Breed. 2016;7(2):276-283.
26. Panda GH, Pandey ID, Tripathy SK. Correlation and path analysis among CMS lines and maintainers. Int J Curr Microbiol App Sci. 2019;8:609-614.
27. Panse VG, Sukhatme PV. Statistical Methods for Agricultural Workers. 2nd ed. New Delhi: ICAR; 1967.
28. Patel JR, Saiyad MR, Prajapati KN, Patel RA, Bhavani RT. Genetic variability and inter-relationships in rainfed upland rice. Electron J Plant Breed. 2014;5(3):455-460.
29. Prasanth V, Haritha T, Girija Rani M, Usha Rani I. Association studies in elite salt tolerant lines under direct seeded conditions in rice (Oryza sativa L.). J Res ANGRAU. 2026;54(1):01-10.
30. Pratap N, Singh PK, Shrivastava AK, Shukla DK, Verma OP. Genetic variability and character association in rice germplasm. J Pharmacogn Phytochem. 2018;7(1):17091714.
31. Rani A, Usha Rani I, Haritha T. Correlation coefficient analysis of grain yield and component traits in world mini core rice collections under sodic soils. Int J Adv Biochem Res. 2024;8(3):919-922.
32. Saha S, Begum HA, Mitra S, Das K, Halder T, Ghosh B. Path coefficient analysis in rice. Int J Bio-resource Stress Manag. 2019;10(3):241-247.
33. Saketh T, Gouri Shankar V, Srinivas B, Hari Y. Correlation and path coefficient studies in rice. Int J Plant Soil Sci. 2023;35(19):1549-1558.
34. Saleh MM, Salem KFM, Elabd AB. Definition of selection criterion using correlation and path coefficient analysis in rice (Oryza sativa L.) genotypes. Bull Natl Res Cent. 2020;44:143. doi:10.1186/s42269-020-00403-y.
35. Saran D, Gauraha D, Sao A, Sandilya VK, Kumar R. Correlation and path analysis in rice. Int J Plant Soil Sci. 2023;35(18):79-89.
36. Searle SR. Phenotypic, genotypic and environmental correlations. Biometrics. 1961;17:474-480.
37. Shajan KT, Lal GM, Bara BM. Evaluation of genetic variability, correlation and path analysis in yield and yield contributing characters in rice (Oryza sativa L.) genotypes in Prayagraj, Uttar Pradesh. Electron J Plant Breed. 2024;15(2):412-421.
38. Singh BK, Chaudhary BD. Biometrical Methods in Quantitative Genetic Analysis. Ludhiana: Kalyani Publishers; 1977.
39. Singh P, Singh PK, Singh V, Verma OP. Correlation and path analysis in rice under sodic soil. Int J Curr Microbiol App Sci. 2020;9(11):2896-2903.
40. Wright S. Correlation and causation. J Agric Res. 1921;20:557-585.