Naturally occurring antibodies (NAbs) produced by CD5+ B-1 B cells include those with specificity for thymocytes (anti-thymocyte autoantibody, ATA). Here we describe a prototypic example, encoded by an unmutated immunoglobulin/heavy chain/light chain. Studies with ATA-(“heavy chain only”) transgenic mice demonstrated a critical requirement for self-antigen in the accumulation of B cells with this specificity and for the production of high levels of serum ATA NAb. Furthermore, analysis of B-cell development in ATA-(“heavy and light chain”) transgenic mice revealed two distinct responses by B cells to expression of this B-cell receptor (BCR).(1) Most B cells developing from bone marrow of adult mice were blocked at an immature stage in spleen and only escaped apoptosis by editing their BCR to eliminate the ATA specificity.(2) Some B cells differentiated to antibody-forming cells without altering their specificity, produced high levels of serum ATA, and many ATA-secreting plasma cells were observed in spleen. Finally, examination of B-cell development and ATA NAb production in ATA-transgenic mice with levels of Thy-1 autoantigen varying from very low to above physiologic reveals a clear relationship between BCR crosslinking by antigen and B-cell fate. Low levels of Thy-1 autoantigen resulted in diversion of ATA B cells into the marginal zone B-cell compartment, presumably because of reduced BCR signaling. Thus, our studies demonstrate a key positive selection step in the development of NAb-producing B cells and show that most of these cells in adult mice bearing such specificities fail to reach a mature stage. Importantly, because these specificities are isolated from B-1 B cells and, when expressed as transgenes, guide development into the B-1 or marginal zone B-cell pool, we identify these B cells as a major source of natural autoantibodies in mice.