Agrostis capillaris, Arrhenatherum elatius, Bromus willdenowii, Cynosurus cristatus, Dactylis glomerata, Elytrigia repens, Lolium multiflorum, L. perenne, Phalaris aquatica, Phleum pratense, Schedonorus phoenix, and a ryegrass selected for high surface root mass, were grown in 1 m deep · 90 mm diameter tubes of sand. Tubes were irrigated with a low ionic-strength nutrient solution and 15 N-labelled nitrate was leached down the tubes prior to harvest. Shoot weights, root weights in 10-cm depth increments, and shoot and root nitrogen concentrations were determined. Plants of L. multiflorum were the heaviest and plants of A. capillaris were the lightest. Root system shape was analysed by comparing the decay constant from an exponential model fitted to the proportion of root mass in 10-cm depth increments, and, also, by analysis of the proportion of root mass in the top 10 cm. Cynosurus cristatus was strongly surfacerooted and the perennial ryegrass, selected for high surface root mass, had more root mass between 0 and 10 cm than did the unselected perennial ryegrass cultivar. There were only small differences in root shape between the other grasses. There was a strong and positive correlation between plant dry weight and the proportion of the pulse of labelled nitrate that was intercepted. The variation in root system shape shown in this experiment had no effect on nitrate interception. Nitrate interception per unit root weight was significantly higher in A. capillaris than in the other grasses. Developing winter-active grasses that have finely divided root systems should contribute to pastures with better nitrate retention characteristics.

No abstract

Root morphological and architectural parameters were recorded for Trifolium repens L. cv. 'Grasslands Huia', inbreds of cv. 'Huia' and cv. 'Crau', and selections within 'Huia' for contrasting root morphology. The plants were grown in sand and solution culture in a glasshouse experiment. Images of roots were made using a flat-bed scanner and analysed using WinRhizo™ image analysis software. Mean root topological indices of 0.95 for solution culture and 0.93 for sand culture indicated that root media had little effect on the herringbone pattern of white clover roots. There were root media effects on five root architectural parameters, all explicable as the result of roots branching more frequently in sand. There were clover type × root media type interactions for five root parameters. These were caused by the inbred clovers being less responsive to a change in root medium than the other three clovers. Differences between roots grown in sand or solution culture were relatively small, but we recommend using sand for screening white clover for root parameters when the physical effects of a solid medium on roots is important.

A study of 386 white clover (Trifolium repens L.) mapping population F1 progeny was conducted to quantify the type and magnitude of genotypic variation for a range of root morphology traits. Clones of each of the 386 white clover progeny were grown in sand. There were significant (P < 0.05) genotypic variance components and repeatability estimates for all the root traits examined. Progeny genotypes with high expression of key traits, including number of root tips and number of root forks were identified. These types may improve phosphate uptake as their highly branched roots will explore a large volume of soil per unit root weight. A strong positive phenotypic and genotypic correlation between several root traits was identified. This suggests an opportunity for indirect selection. For example, selection for high root fork number, a trait that is relatively less complicated to measure, should result in the concurrent increase in expression of the following root traits: surface area, number of tips, volume, and dry weight. Comparison of results from the sand‐based trial with an earlier trial using hydroponic conditions, with clones of the same 386 progeny, showed similar correlations exist among the root traits in both systems. The progeny genotype‐by‐trait Best Linear Unbiased Predictor matrix generated from the sand study is currently being used for the identification of root trait quantitative trait loci.

No abstract