Immunlological unresponsiveness to protein antigens can be induced in newborn animals with relatively small doses of proteins,l while in adult animals much larger amounts are required.2 3 It was recently reported by Dresser4 and confirmed by Claman5 and Battisto and Miller6 that all the molecules in a solution of bovine gamma globulin (BGG) were not homogeneous with respect to their ability to induce tolerance. Thus, relatively small doses of BGG could induce tolerance in adult mice4,^ and guinea pigs,6 provided the preparation had been subjected to ultracentrifugation before injection to remove the aggregated material, which appears to be the most antigenic fraction. These observations shed some light on the mechanisms of antigenicity and tolerance, since it is known that serum proteins which have been aggregated or denatured7' are very avidly cleared from the circulation by phagocytosis by macrophages of the reticulo-endothelial system. One might therefore consider that centrifuged BGG does not induce a primary response because it is taken up by macrophages less avidly than is the slightly aggregated material. The present work was undertaken to determine whether phagocytosis by macrophages is indeed a necessary step before an immunological response, by studying whether a protein solution deprived of particles is able to elicit antibody formation. For this reason, the antigen used in this study, bovine serum albumin (BSA), was first cleared of phagocytizable particles by injection intravenously into an aninmal used as a filter. The rabbit was chosen for these experiments because it is known to give a primary response to BSA under normal conditions. It was found that such a screened protein proved to be much less immunogenic than native BSA and, in a few instances, even small doses were capable of inducing tolerance. Methods.-Adult New Zealand male rabbits weighing 2-3 kg were used. Armour's crystalline bovine plasma albumin was labeled with 131 according to the method of Biozzi et al;9 the amount of bound iodine averaged 3 atoms per molecule of protein. Radioactivity was measured in a scintillation counter. Three rabbits, used as filter-animals, were each injected with 340 or 480 mg of labeled BSA. These animals were exsanguinated 48 hr later and their serum, containing 0.85-0.86 mg BSA-I131 per ml, was collected. Each of 9 rabbits received 8.0-9.3 ml of such serum intravenously, i.e., 6.8-8.0 mg of filtered BSA-IL31. Each of a control group of 7 rabbits was given an equal amount of unfiltered BSAI'l from the same preparation. A third group of six rabbits was given 8 mg of filtered BSA as above, simultaneously with 2 mg of BSA denatured by heating at 70?C for 10 min. All animals were bled from the ear at 24-48-hr intervals. The amount of radioactivity per ml sample of serum was measured, and antigen disappearance curves were studied. In order to test for tolerance, some of the animals received a challenge injection of 5 mg of alumprecipitated BSA or 10 mg native BSA on the 21st day, and a 2nd injection of 10 mg native BSA on the 35th day. For evaluation of the primary response, serum samples were taken 3 days after immune elimination of the antigen from animals showing an immune response and for evaluation of the secondary response before and 7 days after challenge injections. Antibody titrations were performed by means of passive hemagglutinationl and passive cutaneous anaphylaxis in the