Root phenotypes at maturity in diverse wheat and triticale genotypes grown in three field experiments: Relationships to shoot selection, biomass, grain yield, flowering time, and environment

摘要

Modelling and limited data suggest that crops with deeper and longer roots capture more soil resources and yield more when water is available deeper in soil profiles. Interest has grown in the development of new cultivars with deeper roots. This study provides data from three field experiments to help researchers and breeders continue to assess the value of selecting for deeper roots for yield and water use efficiency gains. We asked: do genotype groups with shoot phenotypes easily selectable in pre-breeding programs express predictable root depth and length at time of grain harvest in the field? Do flowering time and shoot biomass predict deep roots measured directly in the field with coring, such that deeper roots are associated with more shoot growth and yield? Does genotype, including triticale versus wheat types, vary in roofing traits? Thirty-four wheats (Triticum aestivum L.) and two triticales ( x Triticosecale) were drawn from ten 'genotype groups'; selections from breeding programs and commercial cultivars that were distinguished on the basis of height, tillering, winter habit, and early vigour. These were grown at two independent sites and soil conditions in year 1 (experiment 1 and 2), with a subset of six wheats and two triticales repeated in year 2 at year 1 site (experiment 3). Above-ground biomass, flowering date, grain yield and root length and depth were measured with a high level of replication (four replicate plots and four soil cores per plot). Root length density was predicted from root counts obtained using the core-break method on 42 mm diameter, two m deep cores. A Bayesian multivariate mixed-effects model was used with fixed effects of the environment and random effects of genotype groups, genotypes and their interactions with the environment. Variation in roofing depth and length caused by environments was much larger than that caused by genotypes. Positive relationships between biomass, yield and root depth and length were observed across experiments and genotype groups (r = 0.62 for biomass and root depth, r = 0.61 for yield and root depth; r = 0.66 for biomass and root length, r = 0.53 for yield and root length), but the largest effects were driven by differences in soil and rainfall conditions between experiments. However, the smaller genetic effects on roofing depth and yield were positively correlated (r = 0.69). We did not find that easily selectable shoot traits like early vigour, tillering, and height reliably predicted in-field deeper roofing. Notably, the two triticales were 74 % more likely to have a deeper roofing and 82 % more likely to have less total root length, than spring wheats. We conclude that deeper and longer roots at maturity are (1) challenging to pre-select using shoot phenotype prior to field evaluation; (2) depend almost entirely on environment for expression in the field with small effects of genotype; and (3) can grow at no apparent 'cost' to shoot growth or yield and as such can remain a target for breeding.