Ryegrass, white clover, ryegra.ss/whi te clover and lucerne herbages were used in an industrial pilot-scale study of the effects of regrowth period and season on protein extraction, separation and recovery. Herbage was harvested, adjusted to 10% tlry matter, pulped and (belt) pressed. Protein in the expressed juice was thermally coagulated and centrifugally separated to produce wet leaf protein concentrate (LPC) which was spray dried. Nitrogen recovery ratios ranged from 15 to 27%, and decreased with increasing herbage maturity. Largest decreases were for ryegrass as the herbage flowered. Summer 1977 ratios were lucerne 22%, white clover 19%, ryegrass/white clover 18% and ryegrass 17%. The higher recovery ratio for lucerne was due to a higher extractability. The grazing rotation intervals used in seasonal dairying on ryegrass/white clover pastures in New Zealand would be satisfactory for protein extraction. Estimated recoverable protein yields for irrigated and non-irrlgated pasture, and lucerne, were 979, 675 and 660 kg/ha/yr, respectively. For integrated dairying plus protein extraction, total (animal plus plant) protein yields were estimated to be 1479, 1051 and 980 kg/ha/yr, for the above respective circumstances. LPC contained from 32 to 55% crude protein which was higher for the legumes than ryegrass. Crude fat was higher in ryegrass (9%) than white clover (5%) and lucerne (7%) LPC. Amino acid content was not consistently affected by regrowth period. LPC histidine and lysine contents (g/16g N) and in vivo protein true digestibilities were: ryegrass 2.4, 5.7, 70%; white clover 2.7, 6.0, 73%; ryegrass/white clover 1.9, 5.4, 70%; lucerne 2.6, 6.1, 82%. Relative Nutritive Values (RNV) of the protein (rat growth assay, lactalbumin as reference) were: ryegrass 0.53, white clover 0.47, ryegrass/white clover 0.45 and lucerne iii ,.... 0.63. With supplemental methionine, the respective RNV were o. 79, 0.88, 0.82 and 0.89. Neither season (autumn v. spring) nor herbage age (4, 6 or 8 weeks regrowth) affected protein quality. In a second experiment LPC were prepared with and without juice treatment with metabisulphite. Protein true digestibility, available lysine . (g/16g N) and the percentage true protein in the chloroplastic fraction were: ryegrass 73%, 4.8, 73%; white clover 80%, 5.7, 56%; ryegrass/ white clover 74%, 5.1, 73% and lucerne 80%, 5.4, 49%. It is published that the chloroplastic fraction is less digestible and contains less lysine and histidine and more fat, than the cytoplasmic traction. The lower digestibility of ryegrass and ryegrass/white clover than legume protein is due to a higher proportion of the protein being in the chloroplastic fraction for the former two herbages. Metabisulphite treatment increased chemically available lysine (2%, P<0.05) and methionine (7%, P<0.05) and also increased RNV from 0.59 to 0.73. RNV further increased to 0.84 with supplemental methionine. Metabisulphite did not improve the availability of the second limiting amino acid(s) for growth. Nutritionally available methionine and cystine, without and with metabisulphite treatment, were respectively, methionine 1.57 g/16g N (81% availability) and 1.64 (85% availability) and cystine 0.26 (27% availability) and 0.70 (73% availability). The .biggest improvement in cystine availability (10% to 65%) due to metabisulphite was for ryegrass/white clover, the only herbage for which metabisulphite also improved protein digestibility (by 4 units). Cystine is more prone to nutritional damage during processing than methionine .