Summary. Three Lotus uliginosus populations, cv. Maku, Sharnae and the line G4704 were compared to Trifolium repens cv. Haifa on the basis of plant frequency and contribution to total above ground biomass production at 6 sites in eastern Australia (Samford, Grafton, Taree, Nowra, Moss Vale and Warragul). In addition, L. uliginosus G4703 was sown at Nowra and L. corniculatus (L.) cv. Grasslands Goldie was sown at Taree. At each site populations were sown with a grass; Setaria sphacelata cv. Narok at Samford, Grafton and Taree; Lolium perenne cv. Kangaroo Valley at Nowra and Moss Vale and cv. Ellett at Warragul. Grasslands Maku and the sown grasses were also sown alone at each site. Each of the pasture treatments was sown either with 500 kg/ha single superphosphate or without superphosphate (or with small amounts of P, if available soil Bray P The main findings were that: (i) superphosphate had little effect on lotus frequency and biomass after the first few harvests; (ii) Haifa white clover was the most successful legume in terms of plant frequency and contribution to total biomass at the subtropical sites (Taree, Grafton and Samford); (iii) Haifa was the most successful legume in the establishment year at the temperate sites, but lotus became dominant in subsequent years; (iv) Grasslands Maku was the superior greater lotus population in terms of frequency and contribution to total biomass production at all sites; and (v) at Taree, the one site where L. corniculatus cv. Grasslands Goldie was included, it was the superior population.