This thesis examines the potential for increasing the nutritive value of Phalaris tuberosa herbage by breeding. Two important nutritive value criteria for thie species are herbage digestibility and crude protein content. One assumption made in advance of the experiments was that, for ruminants grazing P. tuberosa pastures in an environment such as south-eastern Australia, digestibility and protein content of vegetative herbage are unlikely to be major limitations to animal production. For this reason, attention was concentrated on the nutritive value of herbage at various growth stages during the reproductive phase. The hoading stage was chosen for many measurements because it is the latest of the easily recognised developmental stages of the reproductive cycle at which a reasonably large plant population may be measured for herbage quality in advance of flowering. In a plant breeding program, each cycle of selection for nutritive value at a later stage would require two years instead of one. However, the quality of more mature herbage was also examined in several experiments. The response to selection for any character can be predicted provided certain genetic parameters are known; such requirements are briefly outlined and the relevant literature on herbage digestibility and protein content is reviewed with reference to these genetic concepts. The first three experiments, which provide information on the inheritance of the two characters, are then described. In addition, the third experiment provided selection lines differing in crude protein content, and the following two experiments describe the agronomic performance of these lines under field conditions. When it became obvious that crude protein content was negatively correlated with seedling vigour and herbage yield, a final series of four experiments was undertaken to investigate the physiological nature of these adverse relationships. One of the experiments (experiment 3), conducted while the author was employed by the Division of Plant Industry, CSIRO, Canberra, Australia provides the connecting link between the genetic, agronomic and physiological aspects of thie thesis. A brief summary of the results of this experiment which are relevant to the remainder of the work is therefore included.