The evidence reviewed indicates that proven infectious processes are capable of inducing many of the characteristic findings in connective tissue diseases, and that several mechanisms of infection are now recognized which could account for their chronicity. At lease one disease (Australia antigen-positive polyarteritis nodosa) and probably some cases of polymyositis appear to be caused by infection with specific viral agents. At the same time, it is clear that one way in which infectious processes may produce many of the manifestations of connective tissue disease is through perpetuation of immunologic stimulation and immune reactions. Hence, it should not be concluded that all connective tissue diseases are infectious, nor that infection necessarily represents a primary defect, even if the existence of infection can be demonstrated. Interpreted broadly, the data indicate that infection is one means of initiating and maintaining disease mechanisms which account for many of the characteristic features of connective tissue disease; however, other possible initiating causes are not excluded. There is reason to suspect that abnormalities in host defense mechanisms or variability in host immunologic responsiveness may be requisite for the initiation of several unusual forms of chronic infection in both man and experimental animals. Although the nature of such abnormalities has not been defined and a clear understanding of them does not appear to be imminent, great potential for significant intervention in the disease processes may be opened through attacks on infection and secondary mechanisms. If the clinical grouping of connective tissue diseases simply reflects the operation of related secondary mechanisms, (common final pathways), it seems likely that this group of diseases will be found to represent a very large number of distince entities. Hence, the syndrome of SLE might reflect atypical infection and host response to EB virus in some patients, and an unusual response to the chronic administration of hydralazine in others. Polyarteritis nodosa may result from atypical infectious hepatitis in one group, and an exaggerated response to a variety of allergens in another. Polymyositis may reflect both an unusual response to tumor antigens and a chronic paramyxovirus infection of muscle. It is reasonable to predict that as the pathogenic mechanisms of connective tissue diseases continue to unfold, a complex spectrum of inter-relationships between host responsiveness, mechanisms of chronic antigenic stimulation, aberrant forms of viral infection and host defense defects will be revealed.