Multi-environment evaluation of dual-purpose baby corn hybrids for yield stability and forage biomass in rainfed agro-ecological systems

Santosh Kumar^1*Pardeep Kumar²Banshidhar .³Bhupender Kumar²Dr Zahoor Ahmed Dar⁴Krishnan P Abhijith⁵Sravani D.⁶Aditi Eliza Tirkey⁷Yathish K. R.⁸Preeti Singh^9*Abhijit Kumar Das²Pramod Kumar Pandey¹⁰Priya Ranjan Kumar⁹K. K. Singh¹¹Vishal Nath⁹

• ¹Indian Agricultural Research Institute (Jharkhand), Hazaribagh, India
• ²ICAR - Indian Institute of Maize Research, Ludhiana, India
• ³ANDUAT, Kumarganj, Ayodhya, Uttar Pradesh, India, Ayodhya, India
• ⁴Sher-e-Kashmir University of Agricultural Sciences and Technology of Kashmir, Srinagar, India
• ⁵ICAR- Indian Agricultural Research Institute, Dhemaji, Assam, India, Assam, India
• ⁶Agricultural Research Station, PJTAU, Karimnagar, India, Karimnagar, India
• ⁷Rabindra Nath Tagore Agriculture College, BAU, Deoghar, India, Deoghar, India
• ⁸Winter Nursery Center (ICAR-Indian Institute of Maize Research), Hyderabad, India, Hyderabad, India
• ⁹ICAR - Indian Agricultural Research Institute Jharkhand, Gauria Karma, India
• ¹⁰College of Agriculture, CAU (Imphal), Kyrdemkulai, Meghalaya, India, Meghalaya, India
• ¹¹Indian Agricultural Research Institute, Regional Station, Pusa, Bihar, India, Pusa, India

Maize (Zea mays L.) is globally acknowledged as a versatile cereal crop with vital contributions to food systems, animal nutrition, and industrial applications (Kumar et al., 2022). Its inherent adaptability permits successful cultivation across diverse agro-ecological zones in India, positioning it as a strategic crop for ensuring national food and feed security, while also supporting bioresource development (Yathish et al., 2022;Kumar et al., 2021). As per FAOSTAT (2023), maize cultivation in India encompassed 10.74 million hectares, yielding 38.09 million tons, whereas at the global scale, maize covered 208.23 million hectares with production of 1.24 billion tons. With the implementation of India's National Bio-Energy Policy, maize has gained added significance for its contributions to food and renewable energy systems.Among the diversified uses of maize, baby corn-a nonfertilized cob harvested shortly after silking-has emerged as a climate-smart and dual-purpose crop. It serves the dual goals of supplying nutritious fresh vegetables and producing foragegrade biomass suitable for livestock systems. The crop's early maturity and low input requirements make it particularly wellsuited to marginal lands, peri-urban areas, and seasonal fallows, aligning with goals for sustainable intensification and urban food system resilience (Hossain et al., 2022). Baby corn cultivation offers complete biomass harvest, including stalks, husks, and foliage, which can be efficiently used as tender green fodder. This enhances overall resource-use efficiency, especially in mixed farming systems (Kumar et al., 2023). Nutritional analyses show that baby corn cobs are nutrient dense containing around 18% protein, along with fiber, minerals (Ca, Mg, P, Fe), and vitamins (ascorbic acid, β-carotene), with good starch content and high protein digestibility. Equally important, the green fodder harvested from baby corn provides 6.5-9.0% crude protein, 24-28% crude fiber, and 7.7-9.9% ash, underscoring its role as a high-quality feed resource (Hooda and Kawatra, 2013).This dual functionality-marketable vegetable cobs and highquality green fodder-makes baby corn a strategic component of a circular agricultural bioeconomy. Its cultivation supports ecosystem-based adaptation and aligns with global sustainability targets. In urban and peri-urban settings, baby corn production contributes to SDG 2 (Zero Hunger) by enhancing nutritional access, to SDG 12 (Responsible Consumption and Production) through efficient biomass utilization, and to SDG 13 (Climate Action) by promoting adaptive, land-efficient systems (Kumar et al., 2023;Singh et al., 2024). The crop's short growth period and capacity to fit between cropping windows or in small urban plots allow for multiple harvests annually, ensuring rapid returns for both smallholders and urban agricultural entrepreneurs.The rising demand for fresh vegetables, fueled by changing dietary preferences and increasing health awareness in urban populations, has created a favorable market pull for baby corn near metropolitan zones (Boraiah et al., 2022;Kumar et al., 2020;Yathish et al., 2024). In contrast to traditional field maize, baby corn offers benefits such as reduced post-harvest losses, lower water and nutrient requirements, and closer proximity to marketsfactors that cumulatively improve the sustainability and resilience of food supply chains. Nonetheless, baby corn breeding remains underdeveloped. Given maize's cross-pollinated nature, hybrid development offers an efficient approach to improve early maturity, marketable yield, and biomass production (Das et al., 2021;Singh et al., 2021;Neelam et al., 2020). However, strong genotype × environment interactions (GEI) often obscure trait expression and complicate direct selection in breeding programs (Devi et al., 2019;Rajora et al., 2017;Sah et al., 2016), particularly under rainfed or microclimatic variations typical of peri-urban agriculture.These environmental sensitivities demand robust performance testing across diverse conditions to ensure trait stability and adaptability. Such evaluations are particularly relevant in expanding urban and semi-urban agricultural zones, where variability in microclimate, soil, and water access is common. Therefore, multi-environment trials (METs) are indispensable for identifying resilient genotypes suited to diverse agro-ecological contexts (Kumar R. et al., 2024). Statistical tools like the Additive Main Effects and Multiplicative Interaction (AMMI) and Genotype plus Genotype × Environment (GGE) biplot analyses have proven effective in dissecting GEI, enabling the visualization of genotype responses, and identifying stable performers across environments (Gauch, 1992;Yan et al., 2000;Yan and Kang, 2003;Yan and Tinker, 2006;Yan et al., 2007). These models not only capture both additive and multiplicative effects but also guide breeders in defining mega-environments for precise genotype targeting.Against this backdrop, the present study evaluated 61 baby corn hybrids developed at ICAR-IARI, Jharkhand, along with two commercial checks, across four agro-ecologically distinct locations-Hazaribag, Ludhiana, Karimnagar, and Srinagar. These locations represent both conventional and emerging zones for rainfed and peri-urban agriculture. The study was guided by three major objectives: (i) to quantify the GEI for key agronomic and biomass traits such as days to first picking, baby corn weight, total green husk weight (TGHW), and fodder weight (FW); (ii) to identify high-yielding, stable genotypes with dual-purpose potential suitable for sustainable intensification; and (iii) to delineate representative testing environments that can inform future hybrid deployment and adaptive breeding. By addressing these aims, the study contributes to the development of yield-resilient, climateadapted baby corn hybrids and advances system-level resilience by promoting integrated crop-livestock production strategies, particularly in regions facing forage scarcity or seeking to optimize biomass utilization. A total of 61 experimental baby corn hybrids (New F 1 crosses made at ICAR-IARI, Jharkhand during Rabi 2023-24), were evaluated alongside two commercial check varieties, AH7043 and CMVLBC-2, to assess their dual-purpose performance in terms of baby corn yield and forage potential. The trials were conducted under rainfed conditions in a multi-location framework during kharif 2024 across four agro-climatic zones of India: Hazaribag (North Eastern Plain Zone), Ludhiana (North-Western Plain Zone), Karimnagar (Peninsular Zone), and Srinagar (North Hill Zone) (Figure 1). These locations were deliberately selected Geographical locations of multi-environment baby corn hybrid trials conducted under rainfed conditions across four agro-climatic zones of India, highlighting site-specific rainfall and soil characteristics.to represent contrasting environmental conditions-spanning gradients in altitude, rainfall, soil types, and temperature regimesthereby providing a robust platform for dissecting genotype × environment interaction (GEI), as outlined in multi-environment trial protocols (Gauch, 1992;Yan and Kang, 2003).Field experiments at each location were laid out in an alpha lattice design (seven blocks and nine genotypes/block) to reduce spatial error and enhance precision in genotype comparisons (Gauch and Zobel, 1997). Each genotype was evaluated as one plot per genotype per location in two replications. Standardized agronomic practices for baby corn cultivation were adopted at all sites, ensuring uniformity in sowing density, nutrient management, and crop care across environments. Nutrient was applied uniformly at all locations with recommended dose of 125:60:30 kg ha -1 of N:P 2 O5:K 2 O. Phosphorus and potassium were applied basally at sowing, while nitrogen was applied in three equal splits-one-third as basal, one-third at the knee-high stage, and the remaining one-third at tasseling. The following traits were recorded: Days to first picking (PD)-as a proxy for early maturity; Baby corn weight with husk (BCWH), Baby corn weight without husk (BCWoH)-as the principal yield trait; Total green husk weight (TGHW)-to estimate green biomass potential; and Fodder weight (FW)-measured at two sites (Ludhiana and Hazaribag), representing the tender above-ground biomass at harvest stage. The inclusion of FW was designed to evaluate the hybrids' potential as green fodder providers, particularly relevant in mixed or peri-urban livestock production systems. Given baby corn's early harvest window, the remaining green biomass is highly palatable and nutritionally valuable as ruminant feed (Kumar et al., 2023). For the traits PD, BCWoH, and TGHW, a combined analysis of variance (ANOVA) was performed across the four locations using R statistical software (RStudio, 2020). Since the trials were laid out in an alpha lattice design, data from each environment were first analyzed to account for block effects, and the adjusted genotype means were then used for the combined ANOVA. The agricolae package was employed to estimate the main effects of genotype (G), environment (E), and their interaction (G × E). The model applied was:FW ijk = μ + G i + E j + (GE) ij + ε ijkwhere FW ijk is the fodder yield of the i th genotype in the j th environment and k th replication, G i is the genotype effect, E j is the environment effect, and GE ij is their interaction. Due to the availability of FW data from only two environments, genotypic means and standard deviations for FW were computed to assess average performance and stability. A combined ranking index was developed using the sum of ranks for mean FW (higher = better) and standard deviation (lower = more stable). This dual-ranking approach helped identify hybrids with both high fodder yield and consistent performance. Genotypes were visually plotted using scatter plots (mean FW vs. SD) to interpret stability in relation to yield (Figure 2). To interpret GEI patterns, GGE biplot analysis was conducted using R packages metan, agricolae, and GGE Biplot GUI. Biplots were constructed using a tester-centered (centering = 2) and unscaled (scaling = 0) approach, with singular value partitioning (SVP = 1 and SVP = 2) enabling both genotype-focused (Mean vs. Stability) and environment-focused (Which-Won-Where) visualizations (Yan and Tinker, 2006). These analyses facilitated the identification of stable hybrids, delineation of mega-environments, and classification of environments based on their discriminative and representative capacities (Yan et al., 2000;Yan and Kang, 2003).In addition, environmental evaluation was conducted through vector length and angle analysis in the GGE biplot space. Environments with long vectors were considered highly discriminative, while those with smaller angles to the average environment axis (AEA) were deemed representative (Putto et al., 2008;Rao et al., 2011). These insights provide strategic guidance for future site selection in hybrid testing and cultivar deployment.To enhance the robustness of hybrid selection, two complementary stability indices were calculated: AMMI Stability Value (ASV), derived from the first two interaction principal components (IPCA1 and IPCA2) of the AMMI model, ASV quantifies genotypic stability; lower values indicate greater stability (Purchase et al., 2000) and Yield Stability Index (YSI), which combines yield and stability ranks into a single metric, enabling simultaneous selection for high yield and stability across test environments. The combined analysis of variance (ANOVA), conducted using the Additive Main Effects and Multiplicative Interaction (AMMI) model, provided critical insights into the phenotypic behavior of three principal traits-days to first picking (PD), baby corn weight without husk (BCWoH), and total green husk weight (TGHW)-under varied agro-climatic conditions. The ANOVA results revealed highly significant (p < 0.001) contributions from genotype (G), environment (E), and genotype × environment interaction (GEI), reaffirming the complex and multifactorial nature of trait expression in baby corn under rainfed multi-location systems. These findings substantiate the need for robust multi-environment testing frameworks to effectively identify stable and climate-resilient hybrids capable of sustaining performance in variable conditions. The overwhelming dominance of environmental factors-accounting for more than 99% of the total variation for all traits-underscores the influence of agroecological parameters such as temperature, elevation, soil type, and rainfall patterns in shaping crop phenotypes. This pattern aligns with established findings on maize's environmental responsiveness, particularly for biomass and flowering traits (Gauch, 1992;Yan and Tinker, 2006;Badu-Apraku et al., 2012).For the phenological trait PD, the environmental effect accounted for 99.22% of the observed variance, with genotype and GEI contributing marginally at 0.42% and 0.36%, respectively (Table 1). These results reinforce the notion that maize flowering and maturity are acutely sensitive to micro-environmental cues such as photoperiod, thermal time, and elevation-induced variations (Craufurd and Wheeler, 2009). Location-wise comparisons showed that Ludhiana (55.25 days) and Karimnagar (57.60 days) facilitated faster maturity than Srinagar (67.78 days), likely due to their warmer climates and quicker accumulation Source of variation Location wise mean (PD (days), BCWH (q/ha), BCWoH (q/ha) and TGWH (q/ha) of growing degree days. These differential maturity timelines have significant implications for the strategic deployment of hybrids, as they support the argument for region-specific cultivar recommendations rather than blanket varietal releases. This perspective directly contributes to yield resilience and system adaptation, especially in zones with limited growing windows or erratic climatic behavior. TGHW, a key trait indicative of total vegetative and reproductive biomass, exhibited a similar trend of environmental predominance, with 99.13% of its phenotypic variation attributed to site-specific factors. Genotypic and GEI effects accounted for 0.53% and negligible fractions, respectively. This high environmental sensitivity supports prior observations that maize biomass accumulation is heavily influenced by agro-climatic conditions such as vegetative period length, soil moisture dynamics, and temperature regimes (Banziger et al., 1997). In this study, TGHW progressively increased from Hazaribag to Srinagar, a pattern that may be linked to prolonged crop duration and better biomass partitioning under cooler, temperate conditions. These insights emphasize the relevance of site-adapted agronomic strategies and spatial targeting of dual-purpose hybrids in sustainable intensification programs-particularly those addressing fodder shortages or promoting crop-livestock integration.In contrast, BCWoH-a direct indicator of marketable cob yield-showed a relatively greater, albeit still modest, contribution from genotypic (0.49%) and GEI (0.31%) effects, with environmental factors remaining dominant at 99.20%. BCWH also showed similar pattern as of BCWoH. Though numerically small, the genetic and interaction components in BCWoH suggest a higher degree of heritable control than PD or TGHW, offering tangible scope for genetic improvement. This observation is consistent with the conclusions of Badu-Apraku et al. (2012), who emphasized that traits related to marketable yield, when evaluated under controlled agronomic conditions, often display moderate heritability and are thus amenable to breeder-led selection and improvement. BCWoH varied widely across environments-from 6.61 q ha -1 at Hazaribag to 25.87 q ha -1 at Srinagar-underscoring both the environmental modulation and the trait's utility as a discriminator for hybrid adaptability under differential climatic conditions. The higher mean yield at Srinagar may be attributed to the temperate highland ecology, where hybrids experienced a significantly longer growth duration (67.8 days to physiological maturity compared with 55-59 days at the tropical and subtropical locations), along with a favorable temperature regime during the crop growth period and the inherently high fertility level of the soils, thereby facilitating greater biomass accumulation and yield expression.A noteworthy aspect of the AMMI analysis was the detection of statistically significant, though numerically limited, GEI across all three traits. These interaction effects, while accounting for less than 1% of total variation, indicate crossover interactionssituations where genotype rankings shift across environments. As emphasized by Yan and Tinker (2006), even subtle GEI signals carry agronomic importance, particularly for countries like India, where micro-regional climate variability and resource constraints demand location-specific genotype deployment. Recognizing and exploiting such interactions is essential for climate-smart agriculture, allowing the tailoring of genotypes to maximize productivity and stability under context-specific constraints.Taken together, the trait-specific response patterns suggest differentiated strategies for hybrid selection and deployment. PD and TGHW, being largely environment-driven, are more suitable for agronomic optimization and zone-based management interventions, while BCWoH-with its partial genetic control and observable GEI-serves as a robust candidate trait for genotypic selection in dual-purpose breeding programs. Its balanced responsiveness ensures that selected hybrids can combine marketable yield potential, green fodder value, and resilience across environments-a trifecta that supports sustainable cropping systems, improves resource-use efficiency, and contributes to SDG-aligned goals such as enhanced food security, responsible production, and climate adaptation. The evaluation of baby corn hybrids across contrasting agroecological zones necessitates a comprehensive understanding of their performance not only in terms of marketable cob yield but also in relation to biomass accumulation, particularly under the dual-purpose cropping paradigm. This approach gains added significance in peri-urban and integrated crop-livestock systems, where land-use efficiency, short growth cycles, and nutrient recycling are central to resilience and sustainability. In the present study, a multivariate assessment of genotype behavior was conducted for four critical traits-days to first picking (PD), baby corn weight without husk (BCWoH), total green husk weight (TGHW), and fresh fodder biomass weight (FW)-to capture both productivity and phenotypic stability across environments.With respect to BCWoH, a direct indicator of marketable cob yield, five hybrids-CR168 (23.39 q/ha), CR70 (22.45 q/ha), CR87 (20.98 q/ha), CR71 (20.93 q/ha), and CR82 (19.96 q/ha)-emerged as superior across the multi-environment testing (Figure 3a). CR168 exhibited the highest mean BCWoH, though its performance was accompanied by a pronounced genotype × environment interaction (GEI), suggesting a specific adaptation to favorable microclimates. In contrast, CR70, CR71, and CR82 demonstrated both high mean yields and low GEI, indicating broad adaptability and resilient performance across heterogeneous conditions. These traits are desirable for scaling baby corn production across diverse production ecologies. The checks varied in performance: Check 1 (19.70 q/ha) exhibited moderate yield and acceptable stability, while Check 2 (19.89 q/ha) showed greater inconsistency, making it less suitable for reliable deployment in dual-purpose systems.The analysis of PD-used as a surrogate for crop earlinessrevealed considerable variability among hybrids. Through GGE biplot analysis employing the Average Environment Coordination (AEC) method, genotypes such as CR168 (61 days), CR70 (61 days), CR87 (59 days), and CR82 (60 days) were positioned near the AEC arrowhead, reflecting longer crop durations and stronger yield responses (Figure 3b). This trend supports the notion that extended vegetative growth enhances biomass accumulation and favors better photosynthate partitioning, particularly under optimal conditions (Banziger et al., 1997). Notably, CR71 (56 days) deviated from this pattern by achieving high cob productivity despite early physiological maturity, a trait indicative of enhanced resource-use efficiency and rapid growth kinetics. Such genotypes are especially suitable for peri-urban farming systems, where cropping windows are narrow, and rapid turnover aligns better with market cycles (Craufurd and Wheeler, 2009).To assess biomass potential, TGHW was analyzed as a proxy for total green mass harvested at cob maturity. Among all hybrids, CR70 (57.10 q/ha) recorded the highest TGHW (Figure 3c), which may be attributed to its longer vegetative period and robust foliage development. On the other hand, CR71 (34.92 q/ha), CR87 (30.10 q/ha), and CR82 (42.34 q/ha)-although superior in cob yield-produced relatively lower TGHW due to earlier harvesting schedules. This trade-off between biomass volume and digestibility is important in livestock systems, where tender, early-harvested fodder offers better palatability and higher protein content. These findings support the integration of early maturing genotypes into livestock-based systems requiring high-quality forage, as also observed in earlier assessments (Kumar B. et al., 2024).For a direct evaluation of forage value, fresh fodder weight (FW) was recorded at Hazaribag and Ludhiana, two representative locations with complete data availability. Mean FW and its standard deviation (SD) across these environments were calculated to determine genotype performance and stability. The resulting scatter plot (Figure 2), plotting mean FW on the x-axis against SD on the y-axis, visually captured genotype behavior across environmental gradients. Interestingly, all the top-performing and stable genotypes for BCWoH (CR168, CR70, CR87, CR71, and CR82) also recorded higher FW than both standard checks, reinforcing their dual-purpose potential.Hybrids such as CR44 (28.46), CR7 (29.68 t/ha) and CR50 (30.90 t/ha) showed the highest FW means with lower standard deviations, pointing to high biomass potential and stability across the location. It can thus serve as strong entries for peri-urban ruminant systems or resource-constrained regions needing dependable fodder supply. In contrast, CR12 (30.18 t/ha) and CR 66 (29.46 t/ha) showed high yield but were associated with high SD showing lower stability. These genotypes may be more suited to high-input or well-managed production systems, where environmental conditions are less variable. Genotypes like CR67, CR84, and CR39, despite exhibiting low variability, had limited FW outputs, and may hold utility only in niche agro-ecologies. The poorest performers in terms of both yield and stability were CR13, CR31, and Check 2, identifying them as unsuitable for integration into fodder-focused systems.The convergence of data across yield and forage traits establishes a foundation for dual-purpose baby corn breeding, where market demands for fresh cobs can be effectively paired with on-farm green biomass utilization. Genotypes such as CR70, CR71, and CR82, which consistently ranked high across multiple dimensions-yield magnitude, stability, and biomass contribution-exemplify the multi-functional ideotype necessary for climate-resilient and sustainable cropping systems. Conversely, CR168, with its high yield and specificity, may be prioritized in targeted, high-performance environments.These results offer a valuable genetic base for baby corn improvement, especially in systems where fresh food production and livestock nutrition must coexist. The performance of these hybrids aligns with broader sustainability goals: enhancing land productivity (SDG 2), promoting efficient biomass utilization (SDG 12), and supporting adaptive, low-carbon agriculture (SDG 13). Furthermore, the identified hybrids hold promise for refinement using molecular breeding and genomic selection tools, enabling further improvements in combining ability, trait resilience, and environmental responsiveness. Through such system-aware and climatesmart breeding strategies, dual-purpose baby corn hybrids can serve as a transformative solution for urban and peri-urban agri-food systems, ensuring food-feed synergy and improved agro-ecological resilience. A central objective of multi-environment trials (METs) is the identification of genotypes with superior performance and stability, alongside the delineation of agro-climatic regions most suitable for their deployment. In this regard, the GGE biplot "Which-Won-Where" framework serves as a highly effective visual tool for unraveling genotype × environment interaction (GEI) patterns. This method partitions the interaction space into distinct sectors, each defined by a leading genotype for a cluster of correlated environments (Yan and Tinker, 2006). Such partitioning is particularly instrumental in distinguishing genotypes with specific adaptation to unique agro-ecologies from those displaying broad adaptation, thereby enabling precision in genotype placement strategies.Focusing on baby corn weight without husk (BCWoH)-a commercially significant trait-the "Which-Won-Where" analysis revealed pronounced crossover interactions across the four test locations. Genotype CR71 exhibited sectoral dominance in Ludhiana (19.38 q/ha) and Hazaribag (10.92), while CR168 and CR70 emerged as the highest-yielding genotypes in Srinagar (37.90 q/ha; 38.21 q/ha) and Karimnagar (42.19 q/ha; 31.04 q/ha), respectively (Figure 4a). These interaction profiles mirrored those derived from AMMI Stability Value (ASV) and Yield Stability Index (YSI) analyses, which consistently identified CR71, CR82, and CR70 as elite hybrids combining yield potential with phenotypic stability. The environment-specific superiority of CR168 and CR70 underscores their suitability for targeted release in high-yielding environments, whereas the broad adaptability of CR71 across multiple locations enhances its value in climate-resilient, resourcevariable systems-especially those prone to unpredictable weather, soil, or input constraints.A comparable pattern was observed in the analysis of total green husk weight (TGHW), a proxy for vegetative biomass with relevance to forage output. Genotype CR70 continued to excel in high-biomass locations such as Srinagar (82.42 q/ha) and Karimnagar (72.92 q/ha) (Figure 4b), leveraging its extended vegetative growth to achieve enhanced above-ground productivity. This trait combination makes CR70 particularly valuable for dualpurpose systems that aim to integrate food and fodder outputs within a single seasonal window. In contrast, CR40 was the top performer in Ludhiana and Hazaribag, suggesting its relative advantage in shorter-season or resource-constrained environments due to earlier maturity and consistent expression across locations. These findings reinforce the notion that environment-tailored genotypes play a critical role in sustainable intensification, particularly where land availability, season length, or agro-climatic stressors constrain productivity.Complementing the genotype-centered view, the "Discriminativeness vs. Representativeness" analysis from the GGE biplot was employed to evaluate the efficacy of each environment in contributing to hybrid selection. This dual-criteria approach considers both a location's discriminative power (as indicated by vector length) and its representativeness of the target environment spectrum (as inferred by the angle with the Average Environment Axis, AEA) (Yan and Tinker, 2006). For BCWoH, the environments of Karimnagar and Srinagar showed long vectors and minimal angular deviation from the AEA, indicating both high discriminative ability and strong environmental representativeness (Figures 5a,b). These characteristics affirm their strategic importance as core testing sites for hybrid evaluation and selection. Ludhiana exhibited moderate vector length and close alignment with the AEA, suggesting a balanced contribution to genotype ranking. In contrast, Hazaribag demonstrated a short vector with a relatively wider angle from the AEA, indicating limited ability to differentiate genotypes and lower generalizability in broader breeding contexts (Kachapur et al., 2023).Taken collectively, the "Which-Won-Where" and environmental assessment perspectives generated complementary insights into genotype adaptability and site utility. Genotypes CR168 and CR70, with their sector-specific dominance, are best suited for high-yielding environments under relatively favorable management conditions. On the other hand, CR71 and CR82 exhibited superior performance across multiple, distinct agro-climatic zones, marking them as broadly adapted, multifunctional hybrids ideal for climate-resilient cropping systems. Their performance validates their use in diverse scenariosfrom peri-urban farming and low-input systems to regions with erratic rainfall or marginal soils, where stability is paramount. From a plant breeding standpoint, these genotypes represent valuable germplasm for both niche-targeted improvement and pan-regional deployment.Simultaneously, the identification of Karimnagar, Srinagar, and Ludhiana as high-performing test environments offers strategic guidance for the design and optimization of future METs. These sites can serve as discriminative core locations that improve the efficiency of genotype selection pipelines, enhance breeding precision, and support adaptive trial designs under future climate scenarios. Furthermore, the study underscores the utility of the GGE biplot approach not only as a tool for genotype evaluation but also as a means of environment characterization-a dual role that enhances breeding decisions in sustainability-focused agriculture.In an era marked by increased agro-climatic variability, land fragmentation, and the need for urban-agriculture convergence, the integration of tools like GGE biplot with eco-physiological modeling and genomic prediction platforms, as suggested by Yan and Holland (2010), presents a promising framework for next-generation baby corn breeding. By merging environmental profiles, field performance data, and genotype architecture, such multi-layered approaches can enable the development of locationsensitive, dual-purpose hybrids-supporting SDG 2 (Zero Hunger), SDG 12 (Responsible Consumption and Production), and SDG 13 (Climate Action). This integrative strategy is critical for enabling high-yielding, stable genotypes that can meet the food-fodder demands of both rural and peri-urban agricultural systems in a sustainable and climate-smart manner. The strategic identification of genotypes that simultaneously deliver high productivity and stable performance across diverse agro-climatic zones is a cornerstone in hybrid breeding programs, particularly under rainfed and climate-variable scenarios. In this study, 61 experimental baby corn hybrids developed at ICAR-IARI, Jharkhand, were rigorously evaluated for their adaptability using a combination of AMMI Stability Value (ASV) and Yield Stability Index (YSI) metrics, complementing insights from GGE biplot analysis. ASV, derived from the first two Interaction Principal Component Axes (IPCA1 and IPCA2) of the AMMI model, quantitatively ranks genotypes based on the magnitude of their genotype × environment interaction (GEI), where lower values denote greater phenotypic stability (Purchase et al., 2000). In contrast, YSI offers an integrative metric that combines average yield and stability ranks into a single selection index, allowing breeders to identify hybrids that balance productivity with consistency (Farshadfar et al., 2012).For baby corn weight without husk (BCWoH), the most direct economic trait for marketable yield, genotypes CR168, CR70, and CR87 registered the highest mean yields across the four environments tested (Table 2). However, these top performers also recorded relatively elevated ASV scores, signifying strong GEI effects and reduced performance stability. These findings are consistent with the GGE biplot outcomes, where these genotypes were positioned further from the Average Environment Axis-indicative of their sensitivity to environmental conditions (Yan and Kang, 2003). Their strong yield response under favorable conditions positions them as viable candidates for periurban, high-input cultivation systems, where controlled agronomic management can mitigate abiotic variability, thereby maximizing commercial returns.On the other hand, genotypes such as CR71, CR82, and Check 2 emerged as optimal in terms of yield-stability trade-offs. These hybrids consistently produced moderate to high average yields while maintaining low ASV scores, reflecting minimal susceptibility to GEI. Among them, CR71 ranked fourth in mean yield but eighth in stability, culminating in the highest overall YSI rank. This profile highlights its broad adaptability, reliability, and suitability for widespread cultivation, even in environments prone to climatic unpredictability or resource constraints. CR82 and Check 2, which secured the fifth and sixth YSI positions respectively, further exemplify resilient hybrid ideotypes ideal for inclusion in seed dissemination efforts targeted toward marginal or diversified farming regions.The stratification of genotypes based on their ASV and YSI profiles facilitates a functionally oriented deployment framework tailored to distinct production environments. This enables precision hybrid recommendation aligned with local agronomic practices, input levels, and risk profiles. Based on this classification, the hybrids can be grouped into three strategic categories:1. Broadly adapted and stable hybrids-CR71, CR82, and Check 2: recommended for rainfed, variable, or stress-prone agroecologies where stability and predictability are critical for food and fodder security. 2. High-yielding but moderately stable hybrids-CR70: ideal for semi-controlled environments with moderate abiotic uniformity, where farmers seek a balance between productivity and risk. 3. Highly responsive but less stable hybrids-CR168 and CR87: suited for well-managed, peri-urban, or irrigated systems, where environmental fluctuations are limited and commercial output optimization is feasible.This evidence-based framework supports the development of climate-smart cropping strategies, especially under evolving agricultural demands. It also reflects the foundational recommendation by Yan and Tinker (2006), who stressed the importance of integrating yield performance with stability measures to guide large-scale varietal deployment.Furthermore, the combined application of ASV and YSI metrics, in conjunction with multivariate tools like GGE biplot analysis, strengthens the robustness and practical utility of selection decisions in baby corn breeding pipelines. These integrative methodologies are essential under current and projected climatic uncertainties, where dual-purpose hybrids must reliably deliver both food and forage outputs. They also support key sustainability targets-enhancing food production stability (SDG 2), promoting responsible land and input use (SDG 12), and advancing climateresilient agricultural systems (SDG 13) (Farshadfar et al., 2012;Yan and Kang, 2003;Kumar B. et al., 2024). In this context, hybrids like CR71 and CR82 represent resilient genetic resources with the capacity to stabilize productivity, reinforce farming system resilience, and underpin sustainable intensification pathways in both rural and peri-urban landscapes. A deep understanding of the interrelationships among key phenological and yield-related traits is essential for designing climate-smart, resource-efficient breeding strategies aimed at enhancing baby corn productivity across varied agro-ecological contexts. This is particularly relevant for peri-urban and urban agriculture, where space, time, and resource constraints demand genotypes that offer multi-functional value. In the present study, correlation analysis was performed among four core traits-days to first picking (PD), baby corn weight with husk (BCWH), baby corn weight without husk (BCWoH), and total green husk weight (TGHW)-to identify statistically robust and biologically meaningful associations that can inform both direct and indirect selection approaches in hybrid breeding.A consistently strong and positive correlation between PD and TGHW (Figure 6) was observed, suggesting that genotypes with longer vegetative durations accumulate greater above-ground biomass. This relationship highlights the physiological basis of extended growth phases, which increase photosynthetic output and assimilate allocation, contributing to enhanced biomass development. These observations are in line with earlier findings in maize and related tropical cereals, where increased time to maturity has been linked to improved biomass accumulation and yield potential (Banziger et al., 1997;Craufurd and Wheeler, 2009).In this study, late-maturing hybrids such as CR168 and CR70 exemplified this trend, producing superior biomass and cob yield under extended growing windows. Such hybrids hold promise for use in regions with longer growing seasons, or in peri-urban systems where biomass can be harvested for secondary uses beyond the edible cob. Another highly significant positive correlation was found between BCWH and BCWoH, reflecting the intrinsic link between total cob mass and its husk-free, marketable yield component. This relationship affirms the value of using BCWH as a surrogate trait in breeding programs, especially during early-stage selection when labor and time constraints may preclude detailed husk-free measurements. As noted by Yan and Kang (2003), reliance on correlated, easily measurable traits enhances selection efficiency without compromising accuracy. This approach is particularly beneficial in high-throughput phenotyping and resource-limited conditions often encountered in public sector or community-based breeding initiatives.Furthermore, the positive correlations among BCWH, BCWoH, and TGHW point toward a coordinated accumulation of both economic and ecological traits. Genotypes that exhibit high edible yield alongside substantial green husk biomass offer a dual advantage: fulfilling food demand while contributing significantly to farm-level resource recycling. In integrated farming systems, such biomass can serve as ruminant fodder, mulch for soil conservation, or as a feedstock for composting-thus enhancing the circular bioeconomy. These synergistic trait combinations are particularly valuable in peri-urban zones, where land-use decisions must balance productivity, waste minimization, and environmental services.The clear alignment of these traits supports the formulation of multi-trait selection indices that target genotypes offering a composite of early maturity, high cob yield, and substantial biomass output. Such indices can guide selection toward resilient ideotypes capable of thriving under diverse agro-climatic conditions while contributing to system-level sustainability. Importantly, the identification of genotypes combining early phenology with respectable yield and biomass traits enhances cropping system flexibility-enabling multiple harvests, intercrop integration, and optimal use of vacant or marginal urban lands.This integrative trait-based strategy resonates with the broader vision of sustainable intensification, where productivity gains are achieved without compromising ecosystem services. It also strengthens the basis for breeding interventions tailored to the evolving needs of urban and semi-urban agriculture, where shortduration, high-output crops are crucial for food-feed security, nutritional resilience, and land-use optimization. As highlighted by Kumar B. et al. (2024), such dual-purpose baby corn hybrids can play a pivotal role in advancing urban greening initiatives, supporting multifunctional landscapes, and fulfilling Sustainable Development Goals-notably SDG 2 (Zero Hunger), SDG 12 (Responsible Consumption and Production), and SDG 13 (Climate Action). The outcomes of this multi-location evaluation highlight the significance of genotype × environment interaction (GEI) shaping phenotypic expression and yield stability in baby corn hybrids, particularly under agro-climatic conditions (Gauch, 1992;Yan and Tinker, 2006;Badu-Apraku et al., 2012). The highly significant GEI observed for traits like baby corn weight without husk (BCWoH) and days to first picking (PD) justifies the application of multi-environment trials (METs) and robust statistical models such as AMMI and GGE biplot to identify stable and high-performing genotypes (Yan et al., 2000;Yan and Kang, 2003;El-Aty et al., 2024;Demelash, 2024). The analysis showed that environmental factors explained over 99% of the total variation, while the contributions of genotype and GEI were comparatively minor. Although GEI was statistically significant, its small magnitude highlights the overriding influence of the environment. This implies that breeding for broad adaptability across diverse environments may be more effective than targeting specific mega-environments.Hybrids CR70, CR71, and CR82 emerged as broadly adapted genotypes with both high BCWoH and phenotypic stability. These hybrids ranked favorably in AMMI Stability Value (ASV) and Yield Stability Index (YSI), confirming their capacity to maintain performance across multiple environments, aligning with the recommendations by Farshadfar et al. (2012) for simultaneous selection of yield and stability. Their stable expression across Hazaribag, Ludhiana, Karimnagar, and Srinagar also positions them as strong candidates for general cultivation in climatevariable production zones (Yan and Tinker, 2006;Kumar B. et al., 2024;Kumar et al., 2025). Yield variation across environments was influenced by rainfall distribution and soil type. Locations such as Srinagar, with favorable rainfall during the later growth stages and fertile sandy clay loam soils, supported higher yields, whereas sites with erratic rainfall or lighter-textured soils, such as Ludhiana and Karimnagar, recorded comparatively lower yields.Conversely, hybrids CR168 and CR87 recorded the highest mean BCWoH but also exhibited elevated ASV and large GEI effects. The GGE biplot "Which-Won-Where" analysis identified these genotypes as sectoral winners in Srinagar and Karimnagar, indicating specific adaptation to favorable, high-input or low-stress environments (Yan and Tinker, 2006). These patterns reinforce the findings of Yan and Kang (2003) and Purchase et al. (2000), who emphasized the need to balance yield potential with GEI responsiveness when making genotype deployment decisions.The inclusion of fresh biomass yield (FW) data from two representative sites (Hazaribag and Ludhiana) further validated the dual-purpose potential of several hybrids. All top genotypes for BCWoH-CR168, CR70, CR87, CR71, and CR82-also outperformed checks in FW, confirming their utility in both food and forage systems. Additionally, hybrids like CR44, CR7, and CR50 combined moderately high FW with low standard deviation, demonstrating forage yield stability and suitability for low-input or peri-urban livestock systems (Kumar B. et al., 2024).A strong positive correlation between PD and total green husk weight (TGHW) was observed, validating the physiological principle that longer vegetative growth contributes to greater biomass accumulation (Banziger et al., 1997;Craufurd and Wheeler, 2009). Genotypes like CR168 and CR70 followed this trend, producing higher TGHW under extended durations, whereas CR71 presented a notable exception by combining early maturity with high cob yield-an ideal trait for fast-turnover systems in urban and peri-urban agriculture (Craufurd and Wheeler, 2009).Additionally, significant positive correlations between BCWH, BCWoH, and TGHW demonstrate a coordinated accumulation of marketable and ecological yield components. This is especially beneficial for integrated crop-livestock systems where baby corn residue can serve as fodder, organic mulch, or compost material, promoting a circular bioeconomy (Kumar et al., 2023). The utility of BCWH as a substitute for BCWoH also reflects the importance of using easily measurable traits in early-generation selection, enhancing breeding efficiency without compromising accuracy (Yan and Kang, 2003).The GGE biplot environment analysis further identified Karimnagar and Srinagar as highly discriminative and representative environments for genotype testing due to their long vectors and proximity to the Average Environment Axis (AEA), confirming their suitability as core locations for future hybrid trials (Yan and Tinker, 2006;Kachapur et al., 2023;Mukri et al., 2024). Ludhiana was also found to be moderately discriminative and representative, while Hazaribag offered less utility for broad hybrid discrimination. These results emphasize the value of identifying core testing environments that enhance trial efficiency and selection accuracy, particularly in diverse agro.The overall outcomes contribute significantly to sustainable intensification frameworks. Dual-purpose genotypes identified in this study support short-duration, high-efficiency cropping systems that are especially valuable in space-constrained urban, periurban, and mixed farming landscapes (Hossain et al., 2022;Kumar et al., 2023). Moreover, the use of baby corn as both a fresh vegetable and quality green fodder aligns well with the principles of integrated food-feed systems, meeting human nutritional needs while supporting livestock productivity in a climate-resilient manner.Finally, the superior hybrid combinations revealed here also suggest the presence of valuable alleles and combining ability in their parental inbred lines. These lines offer promise for further genomic characterization and marker-assisted introgression strategies, which can expedite the development of new climateadapted, dual-purpose hybrids (Singh et al., 2021;Neelam et al., 2020;Kumar B. et al., 2024). Such advancements will further enhance breeding precision and contribute to resilient agricultural systems.Overall, the findings support strategic breeding interventions targeting both food and fodder productivity, with strong implications for achieving Sustainable Development Goals-SDG 2 (Zero Hunger), SDG 12 (Responsible Consumption and Production), and SDG 13 (Climate Action). The integration of high-performing genotypes, optimized testing environments, and trait correlations within a sustainability framework offers a transformative pathway for climate-smart baby corn improvement suited to the needs of current and future agroecosystems. This multi-location study demonstrated significant genotype × environment interaction (GEI) for baby corn traits, emphasizing the need for stability-focused hybrid evaluation across agro-ecological zones. Among the 61 hybrids evaluated, CR70, CR71, and CR82 consistently showed high marketable cob yield (BCWoH) and phenotypic stability, while CR168 and CR87 exhibited high yield with specific adaptation. These genotypes represent strong candidates for broader deployment under climate-variable production systems.Incorporation of fresh biomass yield (FW) data from Hazaribag and Ludhiana further confirmed the dual-purpose potential of baby corn. All top-performing genotypes for BCWoH-CR168, CR70, CR87, CR71, and CR82-also showed superior FW compared to checks, confirming their utility for both fresh cob and fodder production. Genotypes like CR44, CR7, and CR50 combined high FW with low variability, suggesting stable forage potential under peri-urban and mixed farming systems. The use of AMMI and GGE biplot models facilitated accurate identification of stable genotypes and ideal testing sites, with Ludhiana and Karimnagar emerging as discriminative and representative environments. These results are highly relevant to breeding programs targeting shortduration, resource-efficient crops for sustainable intensification. Finally, the superior hybrids and their parental lines identified here merit further genetic and molecular characterization to support trait introgression strategies and enhance climate-resilient baby corn improvement aligned with SDG 2, SDG 12, and SDG 13.