Some of the factors influencing the establishment and survival of temperate pasture species when sown on the soil surface were investigated. The following were the main findings from experiments conducted in the field and in a controlled environment. (1) Both the establishment and survival of seedlings sown by broadcasting on to non-arable land can be significantly improved by a prior treatment of the site with herbicide to reduce competition from existing vegetation. (2) Providing protection for seeds by retaining dead vegetative cover, or by creating some form of surface roughness, improves establishment as it reduces desiccation and provides a more favourable moisture environment in the vicinity of the seed. (3) Uptake of water by seeds resting on the soil surface is enhanced by absorbent lime or bentonite coatings, and increased germination results. (4) Failure of radicles to penetrate the soil following germination is an important factor limiting the establishment of legumes on exposed sites, and any restraint on the movement of germinating seeds under these conditions improves the penetration of the radicle. (5) Variation between species in their ability to establish on the surface suggests that the choice of more suitable species, and selection to improve their adaptation in this respect, could also contribute a great deal to the efficiency of aerial seeding.

Some of the important environmental and plant factors influencing the germination and early seedling development of a number of temperate grasses and legumes have been investigated under controlled conditions. Significant differences were found between legumes and grasses in a number of characters which appear to be important for establishment under field conditions, including quantitative aspects of water absorption, rates of germination, and early root elongation. Ryegrass was superior to all other species in its ability to germinate under conditions of moisture stress. The lower limit for germination was from 4 to 6 bars below the limiting potential found for the other species, and thus may be one of the important factors contributing to the ease of establishment of this species in the field. The early onset of autotrophic growth in young seedlings, as judged by the utilization of external nutrients and the attainment of positive net photosynthesis, occurs within 5 days of imbibition under favourable conditions, well before the exhaustion of endogenous reserves. This highlights the need for adequate water, nutrients, and light during this early stage of seedling development. It also suggests that reserves available to the seedling are probably in excess of requirements under favourable conditions, and may be of potential value under adverse conditions. These results are discussed in relation to the definition of field problems, and to some likely breeding objectives to improve establishment.

In a highly variable breeding population of P. tuberosa, marked responses were obtained to three generations of selection for high and low crude protein concentration (percentage nitrogen x 6.25) in whole tillers at heading. Total response was similar in each direction, and realized heritability estimates were h2 = 0.25 and h2 = 0.20 in the high and low directions respectively. The responses were accompanied by positively correlated changes in in vitro digestibility and in characters commonly used as indicators of nutritive value of herbage. However, there were large negatively correlated changes in seedling vigour, relative growth rates, and other morphological and physiological characters. The implications of the results for pasture plant breeding are discussed.

Nutritive quality was measured in a genetically diverse population consisting of 55 full-sib families in 16 half-sib groups. The families were grown as pure grass swards and grazed heavily at intervals during the second winter and spring. In early summer, the families differed significantly in crude protein and crude fibre contents and in in vitro digestibility. The heritabilities of these characters were 53.5, 43.0 and 60.4% respectively on a full-sib family mean basis. Phenotypic variances were such that the predicted response to selection for higher protein content was substantial, whereas those for lower fibre and higher digestibility were small. At later harvests environmental variation increased markedly, and family differences, though large, were not significant. The predicted increase in protein concentration in early summer of one percentage unit for each generation of family selection at an intensity of 10% would be accompanied by a decrease in seedling size and a delay in flowering unless countered by opposing selection pressure. Two methods for suppressing the undesirable correlated response in flowering time were found to be equally effective, and a new method was formulated to minimize the number of protein determinations required in subsequent generations. Techniques for sampling mature herbage for nutritional value assessment are discussed.

Significant responses were obtained to three cycles of directional selection for seed weight, seedling weight, leaf size, and rate of leaf appearance in a broadly based P. tuberosa population. The responses were linear in both the high and low directions, and the realized heritabilities of the characters were 0.35, 0.12, 0.61, and 0.38 respectively. Leaf size was positively correlated with seedling weight and negatively correlated with rate of leaf appearance. Selection for high seed weight resulted in marked correlated increases in seedling weight and leaf size, but selection for heavy seedlings or large leaves did not significantly increase seed weight. Selection for light seedling or small leaves, on the other hand, led to a correlated decrease in seed weight. A hybrid population obtained by crossing the parents of the high leaf size line with those of the high rate of leaf appearance line had seedlings nearly as heavy as those of the high seedling weight line and significantly heavier than the mean of the two parent lines. This effect, together with other evidence of the effect of inbreeding on all characters, indicated some degree of directional dominance, particularly for leaf size. Plants in the high seed weight line matured earlier and had fewer heads. However, because they had heavier seeds and more seeds per head, there was no apparent decrease in seed yield per plant in this line, compared with the control population. Areas of the spikelets and leaf sheaths were greater in the high seed weight line, but areas of the flag and penultimate leaves were unchanged, and there was no evidence that increases in seed weight were accompanied by increases in the amount of photosynthetic area available to each developing seed. It was concluded that selection for large leaves is a relatively efficient means of increasing seedling weight, but results in a reduced rate of leaf appearance and possibly a slower rate of tillering. Selection for seed weight, though inefficient, exploits a source of variation largely untapped by direct selection for seedling weight.