Eight Pea Genotypes, Four Environments, One Farm
Eight field pea genotypes were tested across four sowing environments at one Indian site; the study names stable lines but reports no yield figures.
What the research shows
- Eight field pea genotypes were grown at one site in Ayodhya, India, over two years and two sowing windows, giving four environments. Meta-analysis
- The study reports significant genotype by environment interaction and names IPF 14-13, Pant P-243, Pant P-200 and Pant P-347 as high-yielding with broad adaptability. Meta-analysis
- In maize tested across five locations, environment accounted for 25.12 percent of grain yield variation, genotype 35.20 percent and their interaction 21.18 percent. UnclassifiedThe study design could not be determined automatically.
- The same abstract also describes Pant P-347 and IPF-14-13 as specifically responsive to particular environments, which sits awkwardly with the broad-adaptability finding. Meta-analysispartially supported
- The abstract describes the analysis as the Eberhart and Russell model of additive main effects and multiplicative interaction, though these are two distinct statistical methods. Meta-analysispartially supported
- A wheat study found that different stability measures disagreed about which cultivars were ideal for which sowing conditions. Unclassifiedpartially supportedThe study design could not be determined automatically.
Each of these was checked against the abstracts of the papers it cites. See the receipts.
Eight Pea Genotypes, Four Environments, One Farm
Eight field pea genotypes, grown at one site in Ayodhya, India, over two years and two sowing windows, differed in yield, and part of that difference was predictable from the sowing window. The study names IPF 14-13, Pant P-243, Pant P-200 and Pant P-347 as combining high yield with broad adaptability [6]. For a breeding programme, that is a usable shortlist. For anyone deciding what to plant, it is much less than it looks: the abstract contains no yield figure at all — not one number for any genotype in any environment.
The design is stated plainly. Eight genotypes, a randomised complete block design, two consecutive years (2023–24 and 2024–25), and two growing seasons described as early and late. That gives four environments, which the abstract calls "four different environments" [6]. The title says "sowing windows and locations," but the abstract describes work at one university farm. Whether there were other sites is not stated.
What the models actually do
Eberhart and Russell is a regression method. Take each genotype's mean yield in each environment and regress it against an environmental index — usually the average yield of all genotypes in that environment. Two numbers come out per genotype. The regression slope, often written b, measures responsiveness: near 1 means the genotype tracks the environment average, above 1 means it exploits good conditions and suffers in bad ones, below 1 means it is buffered. The second is the deviation from regression, which measures how badly the genotype departs from its own straight line. A slope near 1 with a deviation near zero is stable in the technical sense: average in its response and predictable.
AMMI works differently. It fits additive main effects for genotype and environment, then applies singular value decomposition to the residual matrix, producing interaction principal component axes that can be plotted as a biplot. The two methods can rank genotypes differently, because they answer slightly different questions about the same data.
This matters because the abstract describes the analysis as "the Eberhart and Russell model of additive main effects and multiplicative interaction (AMMI) stability models" [6]. Those are not the same model, and the paper appears to have used both — later the abstract reports that "An AMMI biplot analyses further distinguished genotypes with broad stability, such as VL-42 and HUDP-15, from those with specific environmental responsiveness, such as Pant P-347 and IPF-14-13" [6]. The phrasing blurs a distinction the analysis itself depends on.
One further wrinkle. "Eberhart" is not a unique label in plant breeding. A separate 1994 tomato study uses a "Gardner and Eberhart model III" analysis to estimate general and specific combining ability for bacterial wilt resistance [8]. That is a genetic mating design, not a stability regression. Same surname, different tool.
What the results support, and where they pull apart
The pooled analysis of variance found significant genetic diversity among the genotypes and strong genotype by environment interaction [6]. The abstract also reports that the linear component accounted for most of the variability and that the deviation from regression was non-significant, which the authors read as predictable, stable performance. Those statements are not contradictory. Interaction can be strong and still mostly linear — genotypes can differ in how steeply they respond to a better sowing window while each one's response stays consistent. That is the situation Eberhart and Russell was built to describe.
The problem is the genotype lists. IPF 14-13 and Pant P-347 are named as superior performers with broad adaptability on the stability parameters, then named again as examples of specific environmental responsiveness on the AMMI biplot [6]. Both cannot be the headline. Disagreement between stability measures is not unusual — a wheat study across ten cultivars, two locations and three seasons found AMMI and GGE useful for picking ideal genotypes but that "inconsistency was observed among some of these measurements" [1]. The pea abstract does not say which method the authors trust, and gives no slope or deviation values that would let a reader decide.
For scale, a maize study of 69 hybrids across five locations partitioned grain yield variation into 25.12 percent environment, 35.20 percent genotype and 21.18 percent genotype by environment interaction [2]. A study of 99 Vigna stipulacea accessions over two locations found genotype by location interaction highly significant but smaller in magnitude than genotypic variance [3]. Those numbers come from different crops and designs, so they are context, not a benchmark for peas. They do show that the relative sizes of these components are measured and reported in comparable work — and that this abstract does not report them.
Where the evidence runs out
The abstract gives no yield values, no environmental means, no regression slopes, no deviation mean squares, no replication count, and no list of the "yield related traits" it mentions [6]. It does not say which sowing window was the stressful one, or by how much. It reports no heritability and no measure of how much of the ranking would survive a change of site.
The environments are narrow. Two sowing dates at one farm over two years is a real test of sowing-window response, but not of soil, altitude, rainfall pattern or farming system. The trial used a randomised complete block design, a monoculture yield trial. Nothing in the abstract suggests it tested peas in mixture, under mulch, in a polyculture, or under any condition a permaculture grower would recognise as their own. The relevance to permaculture is indirect: it is about choosing genotypes that hold up when conditions shift, a real problem in variable systems, but this trial did not test variable systems.
One adjacent result is worth flagging because it bears on late sowing generally. In wheat under terminal heat stress, grain iron fell by about 32 percent and zinc by about 6 percent, with iron far more vulnerable [4]. That is wheat, not pea, and a different study. It is a reminder that shifting a sowing window changes grain quality as well as tonnage, and that this pea trial's abstract reports nothing about quality.
The bottom line: a small, single-site, two-year trial found genotype differences worth following up, and named four lines. The statistical reporting in the abstract is too thin to check the ranking, and the two analyses it cites do not agree on two of the four names. Treat the list as a hypothesis for a wider trial, not a result.
Read the paper: Stability analysis of field pea genotypes using the Eberhart and Russell model
References
- [1]Performance and Stability of Commercial Wheat Cultivars under Terminal Heat StressIbrahim Elbasyoni · 2018 · Agronomydoi:10.3390/agronomy8040037
- [2]Genotype-environment interaction for grain yield in maize (Zea mays L.) using the additive main effects and multiplicative interaction (AMMI) modelJan Bocianowski, Kamila Nowosad, Dariusz Rejek · 2024 · Journal of Applied Geneticsdoi:10.1007/s13353-024-00899-4
- [3]Understanding G × E Interaction for Nutritional and Antinutritional Factors in a Diverse Panel of Vigna stipulacea (Lam.) Kuntz Germplasm Tested Over the LocationsPadmavati G. Gore, Arpita Das, Rakesh Bhardwaj, Kuldeep Tripathi, Aditya Pratap, Harsh K. Dikshit · 2021 · Frontiers in Plant Sciencedoi:10.3389/fpls.2021.766645
- [4]Studies on effects of terminal heat stress on yield stability, grain iron and zinc contents in wheat (Triticum aestivum L.)Sourav Panigrahi, Yaswant Kumar Pankaj, Vishnu Kumar, Rajeev Kumar, Santosh Kumar Singh · 2022 · Indian Journal of Genetics and Plant Breeding (The)doi:10.31742/isgpb.82.3.3
- [5]Progression of Mineral Transformer Oils, Including Nanotechnology and Property Evaluation TechniquesAbdelrahman Sayed Tawfic, Mohamed H. Abdo, El-Sayed M. El-Refaie, Diaa-Eldin A. Mansour · 2026 · Advanced Sciences and Technology Journaldoi:10.21608/astj.2025.411961.1089
- [6]Stability analysis of field pea (Pisum sativum L.) genotypes across sowing windows and locations using the Eberhart and Russell modelK Vinay, K Shivam, Banshidhar, B Baudh, Y Shivi, S Pankaj · 2026 · Plant Science Todaydoi:10.14719/pst.14682
- [7]Productivity, Stability, and Adaptability of Onion Genotypes Using Additive Main Effects and Multiplicative Interaction and Mixed‐Model AnalysisEdson Perez Guerra, Jorge Andres Betancur Gonzalez, Patrik Marques, Juliano Galina, André Junior Ribeiro, André Dutra Silva Junior · 2026 · New Zealand Journal of Crop and Horticultural Sciencedoi:10.1002/nzc2.70266
- [8]482 PB 414 A COMPARISON OF Pseudomonas solanacearum-RESISTANT TOMATO CULTIVARS AS HYBRID PARENTSWilliam G. Gonzalez, William L. Summers · 1994 · HortSciencedoi:10.21273/hortsci.29.5.500d
- [9]Lock-Ins and Innovation by Reactivation of Neglected and Underutilised Date Varieties in TunisiaRêve Dagher, Myriam Kessari, Nicolas Faysse, Leila Temri, Faten Khamassi · 2027 · Journal of Innovation Economics & Managementdoi:10.3917/e.jie.pr2.0218
- [10]<b>Open Septorhinoplasty with Autologous Costal Cartilage Augmentation for Post-Traumatic Chronic Nasal Bone and Septal Malunion: A Case Report</b>Grace Eka Putri Rambu Hoy Anggung Praing, Agus Rudi Asthuta · 2026 · Open Access Indonesian Journal of Medical Reviewsdoi:10.37275/oaijmr.v6i6.957
- [11]Mathematical modeling and optimization of dry fractionation of wheat flour in a pneumatic screw classifierSerik Tokayev, Mukhtarbek Kakimov, Aidana Ussenova, Bakhtiyar Rzayev, Bauyrzhan Iskakov, Gulnazym Ospankulova · 2026 · Advances in Science and Technology Research Journaldoi:10.12913/22998624/231182
- [12]Development of a model for detecting diseases of crops using spectral and vegetative featuresAislu Kassekeyeva, Moldir Yessenova, Talgatbek Ayazbaev, Gulzira Abdikerimova, Gulmira Azieva, Zhanylsyn Sultanova · 2026 · Bulletin of Electrical Engineering and Informaticsdoi:10.11591/eei.v15i6.12164
- [13]Regional-scale field-level estimation and mapping of tillage practices in areas with crop diversityAmin Norouzi Kandelati, Amanda T. Stahl, Yan Yan, Siddharth Chaudhary, Steve Van Vleet, Mohammed Amine Gharsallaoui · 2027 · Remote Sensing of Environmentdoi:10.1016/j.rse.2026.115781
- [14]Impact of Agricultural Support Program on Food Security Among Vulnerable Households in Rural KyrgyzstanMeerim Esenkulova, Niraj Prakash Joshi, Ikram Ullah, Sonoko Watanabe · 2026 · The European Journal of Development Researchdoi:10.1057/s41287-026-00771-7
All 14 references resolved against Crossref on Nov 21, 2026. Full verification record.
Get the findings, not the noise
A short email when something worth knowing clears the verification gate. No takes. Unsubscribe in one click.