It is now clear how activated T lymphocytes, through the elaboration of proinflammatory lymphokines, may orchestrate the inflammatory response seen in the asthmatic bronchial mucosa. This activation presumably follows recognition of specific antigen by the T lymphocytes. What is the nature of the putative activating antigen(s) in asthma? They might originate from an external source. Viral respiratory infections may be responsible for some exacerbations of asthma and may even initiate persistent airways obstruction, but it is unlikely that they can be implicated in every case. Aeroallergens are other obvious candidates, although it has already been argued that the atopic phenotype is not a prerequisite for the development of asthma, nor does atopy inevitably result in the development of airways hyperresponsiveness. An attractive alternative hypothesis is that the T lymphocytes may be inappropriately recognizing antigens in the bronchial mucosa; in other words asthma may be an "autoimmune" disease. An analysis of possible selective clonal expansion of T lymphocytes at the mucosal surface may provide evidence for a localized response against a limited range of antigens. It will be of particular importance to determine the amount and proportions of lymphokines secreted by T lymphocytes within the mucosal microenvironment, since properties of cells in the peripheral blood or even bronchoalveolar lavage (BAL) fluid are likely to reflect only dimly those of cells at this site. Furthermore, there is now good evidence that diapedesis of lymphocytes from the circulation to mucosal surfaces is a selective process mediated through specific "addressing" molecules on high endothelial venules at these sites. One example of such a molecule is that recognized by the monoclonal antibody Hermes-1.(ABSTRACT TRUNCATED AT 250 WORDS)