The antigens of red blood cells have some advantages for studies of immune phenomena, deriving mainly from the diverse tests that may be applied to the detection of antibody populations with differing activities, and the ease with which antibodies of different specificities may be separated by the technique of antibody absorption. My brief discussion here will not attempt to survey all of the erythrocyte studies that may have bearing on tolerance phenomena, but will be largely a personal history of investigations that have not been adequately described in previous publications. Our point of departure will be the stable erythrocyte mixtures usually found in twin cattle (Owen 1945). It had been a general experience of the Wisconsin group working with immunogenetics of dairy cattle to observe an apparent identity of blood types in twin cattle much more often than would have been expected from the known frequency of identical twinning. The basis of this phenomenon was called prominently to our attention by a case of superfecundation, in which a pair of twins having sires of different breeds could each be shown to have a mixture of red cells of two different sorts, each type derived from one of the co-twins. Studies of these twins over a period of several years showed that the mixture was quantitatively stable in each twin—a clear demonstration that erythrocyte precursors from each twin foetus had become established in the other and had conferred on their new host a tolerance toward these foreign cells that lasted throughout the life of the animal. We extended our studies to about 100 additional twin pairs, and to higher multiple births, including a set of quintuplet calves (Owen, Davis & Morgan 1946). Typically, the two cell types were found in approximately equal numbers in each twin, the proportion varying in different twin-pairs from 50-50 to 40- 60, with a few pairs well outside that range. When one cell type was in excess in one twin, the same type was commonly in excess in the other as well. There was therefore no evidence that the foreign cells were ever at any selective disadvantage. Furthermore, when the two zygotic types differed in antigens controlled by several independent genetic loci, there was never evidence of intermediate cellular constitutions. This appeared to rule out any possibility that transformation at the sub-cellular level occurs in the development of erythrocytes of cattle; the cells develop with complete autonomy, according to their particular genetic constitutions.