The human immune system is undoubtedly one of the masterpieces of evolution. The processes of gene rearrangement and somatic hypermutation allow for the generation of an immense range of binding specificities in both antibodies and T-cell receptors. Almost any foreign material which find its way into the human body can be targeted and removed or destroyed with exquisite specificity. However, the immune system has not arisen in isolation but is the result of a biological arms race, conducted over millions of years, between man and the microorganisms that can enter the body and cause disease. As the immune system has developed, microorganisms have evolved strategies to subvert or evade its surveillance. For the microbes, one logical tactic in this battle has been the development of camouflage. If the immune system cannot distinguish an invading microbe from its own tissues, then the battle is all but won. It is therefore logical that microbial or viral pathogens should evolve surface antigens that share sequence homology with normal cellular proteins, or at least present a similar surface structure to the immune surveillance mechanisms. An inevitable consequence of this argument is that if an immune response should be mounted against the pathogen’s antigens, then damage to host tissues will occur. This tissue damage may continue even if the invading organism that triggered the immune response has been cleared completely from the system. This is the basis of the “molecular mimicry” hypothesis of autoimmune disease. An infectious agent gains access to the body and instigates an immune reaction that eventually clears the pathogen from the system. However, the antimicrobial antibodies continue to recognize native proteins bearing cross-reactive epitopes and continue to cause tissue damage. The binding to host proteins leads to cellular damage and the release of more autoantigen. This results in reamplification and spreading of the immune response. In the most fulminate cases, autoimmune destruction of the tissue results. Since this hypothesis was first proposed in the early eighties, a huge body of evidence has accumulated supporting the view that a number of diseases are caused by molecular mimicry. Over the next several months, this column will review autoimmune disorders. The coverage of this topic is due to the recent resurgence of interest among researchers, clinicians, and families in autoimmunity hypotheses and their possible involvement in the etiologies of childhood neuropsychiatric disorders. A large amount of data has demonstrated both structural similarities and antigenic cross-reaction of various bacterial, viral, and protozoan antigens with proteins in human tissues or cellular components of those tissues. Much circumstantial evidence links the onset or exacerbation of autoimmune disease with recent infections. In animal models, immunization with either viral peptides or their tissue antigen homologues can result in the development of autoimmune tissue destruction that closely mimics certain autoimmune diseases. Similarly, autoimmunity can be triggered very efficiently in certain animal models by systemic viral infections. A direct cause-and-effect relationship for infection and naturally occurring autoimmune disease in humans still eludes us. Nonetheless, a presumptive role of infectious organisms is now accepted in many autoimmune diseases, including insulin-dependent diabetes mellitus (coxsackie viruses), HLA-B27–associated ankylosing spondylitis (various Gram-negative bacteria), Guillain-Barre syndrome (Campylobacter), myasthenia gravis (herpesvirus), lyme disease arthropathies (Borrelia), and multiple sclerosis (various viruses). Multiple sclerosis (MS) is a prime example of an autoimmune disease in which the circumstantial evidence for a viral trigger is becoming almost overwhelming. As yet, however, there is still no “smoking gun.” MS is caused by the immunological destruction of myelinated nerves. It has a strong genetic element to its etiology, with almost two thirds of sufferers carrying the HLA DR2 allele. Monozygotic twins have a 30% concordance rate (10-fold higher than dizygotic twins have). The fact that the concordance rate in monozygotic twins is not 100% strongly suggests that environmental factors play an important role. There is also an intriguing finding that individuals who migrate from an area of high incidence to one of low incidence before their 15th birthday carry the lower risk factor, while those who migrate after this age retain the high risk factor. The course of the disease follows a cyclical pattern of activity and remission, with exacerbations often following viral infections. The main target of the immunological effectors in MS is the myelin basic protein (MBP) and proteolipid-protein components of the myelin sheath, although the oligodendrocyte antigen transaldolase, and other minor components, are also targeted by autoantibodies. It is likely that the minor antigens targeted in MS result from epitope spreading during the course of the disease.