After allowing 0.6 pound apparently digestible crude protein (0.5 pound true protein) per 1000 pounds live weight for maintenance, an intake of crude protein equal to 1.25 times the milk protein of two weeks previous (true protein, 1.03 times the milk protein) caused a slight nitrogen loss with one cow, and a slight retention of nitrogen with the other. The intake, therefore, was close to the requirement.In the same experimental periods the total synthesized protein (milk protein plus positive or minus negative protein balances) was 0.85 to 0.90 of the amount of the crude protein available for increase (protein intake minus the maintenance requirement).It is recognized that this rate of efficiency of production of milk, in relation to the protein intake, applies less definitely to practice than to the conditions of these metabolism experiments.Reduction of the protein intake from the larger to the smaller amounts fed had the effect to reduce, the milk flow. Subsequent increase of protein intake caused a significant increase of milk production with the cow of more pronounced milk-giving tendency (no. 869), but no significant increase with the poorer cow (no. 884), while both cows responded to the increase of protein in the ration by considerable storage of nitrogen in the body.With a low protein intake the utilization of protein for milk production was more efficient than with an intake of protein sufficient to permit of more extensive use for fat and energy production.There is, with the cow, a prominent inclination to produce milk at an individual rate which is determined by inheritance. On account of the ability of the cow to deflect protein from energy production, and to withdraw protein from the tissues, and to use the same, in both cases, for milk production, the immediate effect of reduction of protein intake to reduce the production of milk is only partial in degree, the maximum effect being reached after a period which may extend to a number of weeks.The apparent digestibility of protein by the cow increases prominently with increase in the protein intake, this effect apparently being due to metabolic protein being more nearly a body constant than a variable related directly to the feed.