1. Rats were injected with [U-(14)C]glucose and after various intervals extracts of whole brain proteins (and in some cases proteins from liver, blood and heart) were prepared by high-speed centrifugation of homogenates in 0.9% sodium chloride or 0.5% sodium deoxycholate. 2. The extracts were subjected to gel filtration on columns of Sephadex G-200 equilibrated with 0.9% sodium chloride or 0.5% sodium deoxycholate. 3. Extracts prepared with both solvents displayed on gel filtration a continuous range of proteins of approximate molecular weights ranging from less than 2x10(4) to more than 8x10(5). 4. The relative amount of the large proteins (mol.wt.>8x10(5)) was conspicuously higher in brain and liver than in blood. 5. At 15min after the injection of [U-(14)C]glucose the smaller protein molecules (mol.wt.<2x10(4)) were significantly radioactive, whereas no (14)C could be detected in the larger (mol.wt.>2x10(4)) protein molecules. The labelling of all protein samples was similar within 4h after injection of [U-(14)C]glucose. Fractionation of brain proteins into distinctly different groups by the methods used in the present work yielded protein samples with a specific radioactivity comparable with that of total brain protein. 6. No evidence could be obtained by the methods used in the present and previous work to indicate the presence of a significant amount of ;metabolically inert protein' in the brain. 7. It is concluded that: (a) most or all of the brain proteins are in a dynamic state of equilibrium between continuous catabolism and anabolism; (b) the continuous conversion of glucose into protein is an important part of the maintenance of this equilibrium and of the homoeostasis of brain proteins in vivo.