Breeding methodologies for mixed cropping using barley-pea mixtures as a model system

Summary

Mixed cropping is the simultaneous cultivation of different crop species in the same field at the same time. Especially combina-tions of legume with non-legume plants provide agronomic advan-tages, such as increased productivity and higher yield stability under low-inputs conditions. In mixed cropping, choosing the right cultivars is crucial for the performance of the crop, as shown for e.g. maize with bean by Hoppe (2016). As performance in pure stand can strongly diverge from performance in mixture, estimating the ability of a cultivar to be mixed with another crop is therefore of utmost importance. For this purpose, concepts of general and specific combining abili-ty from hybrid breeding have been adapted to cultivar and crop mixtures. Thus, these effects are called general mixing ability (GMA) and specific mixing ability (SMA). The harvest of most inter-specific mixtures can be separated due to different grain sizes. This fact allows to access valuable information about competition and facilitation among mixing partners, leading to diviating fraction yields. Until now, statistical deve-lopments have neglected to mobilize the additional information provided by these separated harvest lots. The concept of producer- and associate-effects (abbreviated Pr and As) describes interac-tions between varieties in this context. The Pr is the average per-formance of a cultivar grown in mixture with a companion crop, whereas As is the average effect of a cultivar on the respective mixing partner. However, in the past the Pr/As-concept has only been applied to variety mixtures of the same species, sown in alternate row trials. We used the fraction yields of a spring-pea (Pisum sativum L.) and spring-barley (Hordeum vulgare L.) mixed cropping experiment to determine Pr and As of different pea ge-notypes. The additional information provided by this approach is biologically more informative than GMA/SMA estimates, since it better reflects competition and facilitation occurring between different cultivars of the two crop-species.
Plant material comprised 28 pea- (plus 4 mixtures) and 7 barley- (plus 1 mixture) cultivars. The genotypes were chosen in a way to maximise morphological diversity and origin (breeding programs). A subset of all possible combinations was sown: 56 bi-specific pea-barley mixtures were arranged in an incomplete factorial design (Figure 1). Mixtures and pure stands were sown in 7.5 m² plots with two repetitions at two locations in Switzerland. Harvested grains were separated into pea and barley fractions. Variance components for both the GMA/SMA and the Pr/As model were estimated within a mixed model framework with best linear unbi-ased prediction. GMA of pea cultivars, SMA (interaction of pea cultivar with barley cultivar) and the error term were set as rand-om variables. The random effects were assumed of having a mean of 0 and being normally distributed. In analogy, Pr and As effects were estimated with the pea and barley component yields as de-pendent variables, respectively. Potential functional traits, such as early vigour of pea, were measured and evaluated using corre-lation analysis to relate them to GMA, Pr and As effects. When prerequisites for parametric test procedures were not fulfilled, non-parametric tests (e.g. Spearman rank-correlation) were ap-plied .
The proportion of GMA variance of pea is the variance in mixture yield explained by the presence of a given pea cultivar in mixture. We found this variance to be predominant over SMA variance (the variance due to interaction of pea and barley cultivars). Our analy-ses yielded variance components of GMA pea ≈40%, SMA ≈10%, residual ≈50%. There was a significant negative correlation between the pea Pr and it’s As with Spearman’s rs = -0.47. How-ever, few individual pea genotypes were found with positive Pr and positive As. The As of pea were correlated over locations (R²=0.48). The GMA of pea was not significantly correlated with early vigour of pea (rs = 0.21), whereas As of pea had a significant negative correlation with this trait (rs = -0.36).
The GMA approach, based on the testcross methodology from hybrid breeding is a valuable tool to determine mixing ability in pea-barley mixtures. This potential is further pronounced by our finding that pea GMA variance is predominant over SMA variance, indicating the potential for breeding for mixed cropping. The GMA approach can be extended using the Pr/As concept for under-standing trait influences on mixture behaviour. We observe a ne-gative correlation between Pr and As effects, indicating a trade-off between a cultivar’s performance and its companion-crop’s per-formance as observed also by Forst (2018) for wheat cultivar mix-tures. However, our data suggests room for genetic improvement, e.g. by selecting deviating genotypes with both positive Pr and As effects. As effects were correlated over locations, indicating an underlying heritable component. Early vigour of pea was not cor-related with GMA, however, it significantly negatively correlated with pea’s As effect (its effect on the barley yield), indicating the surplus of precision and information on trait-performance relati-onships that the Pr/As-concept gives compared to the GMA con-cept. The results allow us to seize the effects of cultivar choice in the performance of crop mixtures allow researchers to propose breeding schemes and experimental designs for improving pea-barley mixtures.
Acknowledgments
The project ReMIX “Redesigning European cropping systems based on species mixtures” is funded by the EU’s Horizon 2020 Research and Inno-vation Programme (grant agreement No 727217) and the Swiss State Sec-retariat for Education, Research and Innovation (SERI, contract number 17.00091). We thank Getreidezüchtung Peter Kunz, Agroscope Reckenholz and Stefan Rindisbacher for their contribution acknowledged.
References
Bedoussac L, Journet E-P, Hauggaard-Nielsen H, Naudin C, Corre-Hellou G, Steen Jensen E, Prieur L, Justes E (2015) Ecological principles underlying the increase of productivity achieved by cereal-grain legume intercrops in organic farming. A review. Agron Sust Dev 35: 911–935. https://doi.org/10.1007/s13593-014-0277-7
Federer W (1993) Statistical design and analysis for intercropping experi-ments, Vol I: Two crops. Springer, New York. https://doi.org/10.1007/978-1-4613-9305-4
Forst E (2018) Développement de méthodes d’estimation de l’aptitude au mélange pour la prédiction des performances et la sélection de mélanges variétaux chez le blé tendre et co-conception d’idéotypes de mélanges adaptés à l’agriculture biologique. Université Paris Saclay.
Hoppe C (2016) Entwicklung von Energiemaissorten für die Mischkultur mt Stangenbohnen. PhD Thesis, Niedersächsische Staats-und Universitätsbib-liothek Göttingen.