​Stability of Advanced Medium Duration Genotypes Across Seasons for Yield in Pigeonpea [Cajanus cajan (L.) Millsp.]

Pigeonpea [Cajanus cajan (L.) Millsp.]  is commonly known as redgram or arhar or tur or thogari in India. It is an important grain legume that originated in the India (Varshney et al., 2017). It is grown in many parts of the world including Southern Africa particularly the region encompassing Kenya, Mozambique, Malawi and Southern Tanzania (Hogh-Jensen et al., 2007). It ranks sixth in global grain legume production and worldwide it is cultivated in about more than 5.0 m ha area. India is the largest producer and consumer of pigeonpea with an area of 4.4 m ha, with annual production of 3.68 m t and productivity of 832 kg/ha (Anonymous 2019).
       
Pigeonpea breeders look forward for widely adapted genotypes which are responsive to input intensive as well as input deficient agriculture in order to enhance production and productivity of the crop. Selection and yield testing are the two major phases of varietal development and the later one is highly influenced by the locations and years of testing. Yield is polygenically inherited and highly influenced by G × E interaction higher magnitude of G × E interaction among genotypes invalidates the fitness of genotypes across the environments. Hence, quantification G × E interaction and identification of stable genotypes across environments will enhance production and productivity of pigeonpea. With this back ground the present study was undertaken under rainfed situation in three seasons/ years to identify stable genotypes of pigeonpea for seed yield.

The present experiment comprised of thirty three advanced medium duration lines of pigeonpea along with six check varieties received from seven research institute viz., (a) ICRISAT, Hyderabad, (b) Regional Agricultural Research Station (RARS), Lam, Andhra Pradesh, (c) Agricultural Research Station (ARS), Badnapur, Maharashtra, (d) RAKCA, Sehore, Madhya Pradesh, (e) Indian Institute of Pulse Research (IIPR), Kanpur, Uttar Pradesh,  (f) PJTSAU, Tandur, Telangana and (g) ZARS, Kalaburagi, Karnataka.  The list and source of genotypes is presented in the Table 1. The trials were conducted in a randomized block design with two replications in three season/ year viz., kharif-2017 (E₁), kharif-2018 (E₂) and kharif-2019 (E₃), under rainfed condition at Zonal Agriculture Research Station (ZARS), Kalaburagi, which is situated in agro-climatic zone-2 (North Eastern Dry Zone) of Karnataka state with 17^o 20' Latitude (N), 76^o 49' Longitude (E) and at an altitude of 443.88 meters above mean sea level. The rainfall pattern of three environments is presented in Annexure-I. The plot size of four rows each with 4 m length was followed with spacing of 90 cm between rows and 25 cm between the plants. Standard agronomic practices were followed and plant protection measures were taken as and when required as per package of practices (Anonymous. 2017). Observations were recorded on five randomly selected plants in each replication in each environment with respect to days of 50% flowering, maturity, plant height (cm) at maturity, 100 seed weight and seed yield per plant and seed yield per plot. Seed yield being the economic trait, it was converted to per hectare yield and used to find out stability parameters viz., mean, regression co-efficient (bi) and mean square deviation (S²d_i) as per stability model proposed by Eberhart and Russel (1966) using statistical analysis package software Windostat 9.2.
 

Pooled ANOVA for stability of seed yield tonnes per hectare is given in Table 2. Genotype × environmental interaction as per Eberhart and Russell’s (1966) model indicated that, Environment linear component was highly significant for seed yield, whereas, G × E (linear) interaction was non-significant for the character. These findings are in agreement with the earlier findings of Ghodke (1992) obtained non significant G × E for majority of the traits. Further, higher value of mean squares due to environment (linear) as compared to genotype × environment (linear) displayed that linear response of environments accounted for the most of total variation for the trait under study. Similar findings in this regard were obtained by Kumara et al., (2015). As regard to pooled deviation (nonlinear portion of variance), which is unpredictable portion of G × E interaction was highly significant for the trait under study. This demonstrated that genotypes respond differently to variation in environmental condition, indicated that the deviation from linear regression also contributed substantially toward the differences in stability of genotypes. The results are in accordance with Balakrishna and Natarajratnam (1989); Sawargaokar et al., (2011); Pawar et al., (2013); Patel and Tikka (2014); Kumara et al., (2015); Singh et al., (2015); Meena et al., (2017); Ramesh et al., (2017) and Deepak Pal et al., (2020).
 

       
As indicated in the Table 3, the genotype GRG-177 showed highest seed yield (1.18 t/ha). While, IBTDRG-4 less seed yield (0.54 t/ha) and population mean over three environments was 0.93 t/ha. All the genotypes showed non-significant value for regression coefficient and deviation from regression. The 23 genotypes viz., GRG-177, GRG-152, ICPH-3762, TS-3R (check), ICPL-20108, Asha (check), ICPH-2671, ICPH-2740, ICPL-99050, RVSA-15-5, ICPL-20116, LRG-105, BDN-2011-1, RVSA-15-10, IBTDREG-3, ICPL-20098, BDN-2013-45, TDRG-58, IBTTDRG-5, AGL-1603-4, IBTDRG-6, ICPH-3933 and TDRG-60 were found to have higher mean value than population mean with non significant bi and S2di values.
 

       
The genotype GRG 177 though showed highest yield, it exhibited nonlinear regression indicating highly sensitiveness to different environments. The genotype GRG 152 had second highest mean yield with positive regression (0.52) and non significant deviation from regression co-efficient indicating the its specific suitability to unfavorable environments. Considering all stability parameters the genotype ICPL 20108 considered as stable and high yielding because it had high mean yield (1.07 t/ha), regression coefficient around unity (bi=0.94) and non significant deviation from S²d_i. In addition to stable performance of the variety ICPL 20108, it had ideal agronomic traits like medium maturity (160 days) and test weight (10.6 g/100 seeds). The average performance of genotypes for agronomic traits over three years is presented in Table 4. The genotypes ICPL 20098 and AGL-1603-4 (0.94 t/ha) had average yield above population mean, regression around unity and non significant deviation from S²d_i indicating stableness of these genotypes across the environments. The agronomic characters of these genotypes also found ideal (Table 4). The check entry TS 3R and the hybrid ICPH 2740 had high mean with bi value less than one (i.e around zero) and non asignificant S²d_i. Indicating their suitability to unfavorable or low input environments. Referring to the ancillary traits (Table 4), the genotypes TDRG 60, IBTTDRG-5 and BDN-2013-45 were early maturing and had average yield above population mean. The entry LRG 41 was bold seeded genotype.
 

       
Environmental index (EI) refers to a variety that has response across environments that is parallel to the mean response of all genotypes in the trial (i.e. the mean regression on the environmental index). The regression of genotypes for seed yield (t/ha) across environments and stability parameters is presented in Fig 1. As indicated by EI, the genotypes TDRG 60,  AGL 1603-4, ICPH 2671, JKM 189 (Ch), RVSA-15-5 and BDN-2013-45 had linear regression over the environments. These findings are in accordance with Shoran et al., (1981); Muthiah and Kalaimagal (2005); Vannirajan et al., (2007); Patel et al., (2009); Sreelakshmi et al., (2010); Thanki et al., (2010); Sawargaonkar et al., (2011); Niranjan Kumar (2013); Muniswamy et al., (2017); Ramesh et al., (2017) and Manish Sharma et al., (2020).
 

From the present study it can be concluded that the genotype ICPL 20108 was found to be a stable for seed yield across the environments with all desirable yield and yield component features, followed by the genotypes ICPL 20098 and AGL-1603-4. The genotype GRG 152, TS-3R and ICPH2740 found suitable for unfavorable or low input environments. The stable genotype identified could be directly released after validation of results or could be used as parents in future breeding program for developing suitable genotype with wider adaptability.

The authors thankful to the ICRISAT, Hyderabad, RARS Lam, (Andhra Pradesh), ARS, Badnapur, (Maharashtra), RAKCA, Sehore, (Madhya Pradesh), Indian Institute of Pulse Research (IIPR), Kanpur, (Uttar Pradesh,) PJTSAU and Tandur, (Telangana) for sharing seed materials under DAC (NFSM cell funded)  project.

This research was funded by the Department of Agriculture Cooperation and Farmers Welfare, Ministry of Agriculture and Farmers Welfare, the Government of India through ICRISAT, Hyderabad.

Conceptualization of research: S. Muniswamy, Rajeev K. Varshney and Rachit K. Saxena. Designing of the experiments: S. Muniswamy, Rajeev K. Varshney and Rachit K. Saxena. Contribution of experimental materials: S. Muniswamy and Rachit K. Saxena. Execution of field/lab experiments and data collection: S. Muniswamy and Praveen Kumar. Analysis of data and interpretation: S. Muniswamy, Praveen Kumar and Geeta. Preparation of the manuscript: S. Muniswamy, Praveen Kumar and Geeta.

The authors declare no conflict of interest.