R ESULTS OBTAINED IN PHYSICAL chemistry and chemical crystallography yield a rather definite picture of the native protein which fits at many points with that emerging from the various fields of the biology of the proteins, i.e. enzymology, immunology, pharmacology, virus studies, and genetics. If we attempt to sum up the general impression derived from a study of various independent sources of information on the proteins, it can be broken down into the following parts: (1) Many proteins in solution are particles, not individual molecules (21). It is useful to use the term proteon to connote a native protein unit incapable of division into subunits also having the native protein character. Every native protein may then be regarded as a proteon or a system of proteons, alike or different. That a protein particle is not a mere conglomerate of proteons but an orderly aggregate, i.e. a molecular colony, follows from the finding that in many cases where ultracentrifugal studies show that a protein is uniform in size and shape under well-defined external conditions, the protein in question is capable of dissociation into subunits when the environmental conditions are modified (21). Much information has now been assembled concerning the particle status of many proteins, ranging from the very large tobacco mosaic nucleoprotein down to horse hemoglobin. The reversible dissociation of protein particles does not imply any change in the chemical status of the proteons. It is therefore to be sharply distinguished from denaturation. (2) Interproteon associations within particles depend upon the R-substituents of the amino acid backbones (N-CQ-C) which make up the proteon skeletons. Structure chemistry provides three types of association: direct polar, mediated polar, and less polar. Horse hemoglobin, which on dilution dissociates into two subunits, provides an illustration of the first type; the Helix hemocyanin particle whose dissociations in aqueous solution are affected by Ca++ and Mg++, an illustration of the first and second types (21). That the less polar R-substituents of aliphatic and aromatic types, as well as the more polar types, emerge from protein surfaces is indicated by many data, e.g. the higher solubility of insulin in propylene glycol than in water and the affinity of human serum albumin for the paraffin chains of fatty acid molecules (13). Thus, associations of the third type are also a possibility,