Trifolium subterraneum is the most widely sown annual pasture legume in the P-deficient soils of southern Australia. Controlled-environment studies have demonstrated that variation exists between genotypes of this legume to acquire P and yield in low-P soils, and there appears to be a plant density effect on these traits. However, the magnitude of this effect is largely unknown. Two cultivars of T. subterraneum, that differ significantly for the aforementioned traits when using the same sowing rate, were grown to determine differences in shoot growth, P uptake and root traits with changing plant density. Microswards of both cultivars were grown at five plant densities and five P levels. Yield and P content of shoots and roots were determined after 5 weeks growth. Root samples were assessed for diameter, length and root hair length. Shoot dry mass of both cultivars increased in response to increasing P supply and increasing plant density. Differences between the cultivars for shoot yield were most pronounced at low plant densities and diminished as plant density increased. This response was particularly evident at lower soil-P levels, whereas maximum yield was relatively independent of plant density in the high-P soil. In contrast, differences between cultivars for root morphological traits such as specific root length were maintained regardless of plant density. The results demonstrate that plant density effects sward P-acquisition and hence shoot yield achieved in the P-deficient soil. Accurate screening for P-acquisition and shoot yield across the T. subterraneum genome therefore requires a uniform plant density comparable to densities observed in the field. The identification of T. subterraneum cultivars capable of improved growth in low-P soils would improve P-use efficiency in Australian soils which are often P-deficient and require annual applications of P fertiliser for high yields. This would consequently lead to greater resilience of the agricultural sector.

Aims The experiment was conducted to understand how root morphological traits contributed to the contrasting P-acquisition efficiencies of two Trifolium subterraneum cultivars.

Methods Phosphorus acquisition, root length proliferation and root acclimation to P deficiency were investigated by varying the P supply (20-300 mg P kg⁻¹) and planting density (1-12 plants pot⁻¹) of clover micro-swards.

Results The shoot yield of both cultivars increased in response to increased P supply and increased planting density. Higher planting densities increased shoot yield by reducing the time to leaf canopy closure. In P-deficient soil, the ‘less P-efficient’ cultivar acquired less P and yielded relatively poorly compared with the ‘more P-efficient’ cultivar when planting density was low (<6 plants pot⁻¹). However, the P acquisition and yields of the cultivars were equivalent when planting density was high (~6-12 plants pot⁻¹). Both cultivars proliferated nutrient-foraging roots in response to P limitation and space (i.e. lower planting densities). However, the proliferation responses of the cultivars differed, particularly in response to space.

Conclusions Differential proliferation responses, combined with differences in specific root length, explained how the cultivars differed in P-acquisition efficiency. The results indicate that inappropriate cultural conditions (e.g. planting densities) may confound attempts to define P-efficient root traits and to identify genotypes with improved P-acquisition efficiency.

Pastures grown on P-deficient soils in temperate southern Australia use mixtures of grasses and legumes. The main legumes (Trifolium and Medicago spp.) are highly productive across a wide range of environments but have high 'critical' P requirements (i.e. the P supply needed for near-maximum yield) relative to the grasses with which they grow. Improving the P-efficiency of the most important legume (T. subterraneum), or developing the agronomic merit of alternatives that are already P-efficient (e.g. Ornithopus spp.) would deliver reductions in P fertiliser inputs, improve farm incomes, and achieve better use of scarce nutrient resources. Here we describe research to improve the P efficiency of T. subterraneum. Field and controlled-environment experiments, with various pasture legume species, have demonstrated that substantial differences in the nutrient foraging potential of their roots determines their requirement for P fertiliser. Three key root morphology traits ensure efficient P acquisition from low P soil: development of high root length, high specific root length and long root hairs. Ornithopus spp. deploy an "optimal" combination of these root traits, efficiently maximising soil exploration to capture more P and to yield well in low P soils. In contrast, Trifolium subterraneum develops long roots in response to low P but has low specific root length and short root hairs which limit its ability to explore soil for P. Within T. subterraneum, variation exists in specific root length and root proliferation. These key factors determine intra-specific variation in P acquisition with the best genotypes achieving twice the yield of the worst in low P soil. The short root hairs on T. subterraneum (0.2-0.4 mm) are a major factor limiting P acquisition efficiency. Wider studies of nutrient foraging root traits among genetically-allied Trifolium species from the Section Trichocephalum revealed substantial differences in propensity for root foraging (11-35 cm root/cm3 soil) and root hair length (0.2-0.5 mm) but, like T. subterraneum, no genotypes tested to date have root foraging traits in the optimal combinations achieved by Ornithopus spp. To drive further substantive change in the P efficiency in the key pasture legume, T. subterraneum, it will be necessary to break through apparent intra-specific 'boundaries' for specific root length and root hair length by identifying radical ecotypic outliers, inter-specific introgression or directed mutagenesis.