This study compared improved "lowfertility" grass species and rye grass with wild populations for performance under varying levels of nitrogen (N) and moisture. The spec~es studied were: Lolium perenne L.; Agrostis capillaris L.; Cynosurus cristatus L.; Festuca rubra L.; Dactylis glomerata L.; Anthoxanthum odoratum L.; Bromus mollis L.; and H oicus lanatus L. Wild-type seed was collected randomly from moderate hill country; bred genotypes were either selection lines, cultivars, or random collections. The plants were grown in simulated hill soil profiles. Each species was represented by 24 seedlings planted in a 6 x 4 block. N was applied in solution at three concentrations (0,7, and 21 kg N/ha per fortnight). Three harvests were taken. A rain shelter was placed over the plants from December until February 1991 and recovery yield noted. Throughout the N experiment, yields of wild and bred genotypes of the same species were generally similar. The response to N differed. During Harvest 2, rye grass showed an increasing yield response to N whereas the other species gave no or little response to the high N rate. At zero N, ryegrass out yielded plants normally found in low-fertility areas but equalled sweet vernal, cocks foot, and browntop. The implication of these results is discussed. The recovery from moisture stress indicated the supremacy of the bred genotypes over the wild populations. Wana cocks foot, wild red A92041 Received 26 August 1992; accepted 14 December 1992 fescue, and brown top had greater yields than ryegrass after moisture stress. The results are from single species plantings and the realisation of their potential under field conditions will be dependent on competitive ability.