Evolution of subterranean clover in South Australia. II. Change in genetic composition of a mixed population after 19 years' grazing on a commercial farm

Abstract

Change in genetic composition of a mixed population of subterranean clover was studied at Kalangadoo in the south-east of South Australia. The seeds from an experiment in which 25 late-maturing strains were compared, were sampled in January 1977, 19 years after the plants were originally sown in 1958. Quadrats along three transects over the approximate position of the old plots were sampled and seed yields measured. The soil was analysed for several chemical and physical properties at the site of each quadrat. Thirty seeds from each sample were germinated in a glasshouse and the resulting seedlings planted out as spaced plants. These were identified on the basis of flowering time and morphological variables; in this way the seed pool of each strain at each quadrat site was determined. Seed pools were adjusted for variation in distance of original plots from the transects by estimating the real position of each plot relative to the transect and calculating the relationship between seed yield and estimated distance. Highest adjusted seed pools were produced by CPI 14454B (820 kg ha-1), CPI 12695 (626 kg ha-1), Eden Valley (502 kg ha-1), and CPI 14197A (476 kg ha-1). Cultivars Mt Barker and Tallarook were also common, but it was deduced from their distribution that they had been continuously introduced to the site from surrounding fields by livestock. All but six of the original strains were recovered in the sampling. Also present were many divergent strains, which yielded 228 kg ha-1 or 31 % of the average seed yield of 744 kg ha-1. Twenty-four of the original 25 strains and five others-Woogenellup, Yarloop, Trikkala, Clare and a divergent strain naturalized at Kalangadoo-were sown in small plots. Twenty-five agronomic variables were measured, including seasonal herbage yields, petiole lengths, leaf areas, isoflavone contents, phosphorus and sulfur uptakes, above- and below-ground seed yields, seed and seedling sizes and permeability of seed coats. The fourth component of a principal component analysis of the data was related to seed pool (P < 0.001), and it is suggested that the variables contributing to this component are important in predicting the success of a strain at Kalangadoo. The most successful tended to have high yields of buried seed, high yields of herbage and long petioles in winter, high formononetin content, early maturity and small seedlings. The results are discussed in terms of factors affecting success, generation of new genetic diversity, and field implications.