Adhesions of lymphocytes, among themselves or with other cell types, are necessary for most steps in immune responses including both induction and effector phases. Among adhesions of T cells involving specific immunological recognition, CTL-target adhesions have been the most studied. Although CTL-mediated killing is highly specific (specific/nonspecific lytic activity 50-fold), CTL-target adhesion (conjugation) is less so. In the mouse, specificity of conjugation has typically been four to eightfold. Two recent studies with cloned human CTL found much less specificity of conjugation, from one-fold (no specificity) to 1.5-fold. Thus, with cloned human CTL, adhesion may occur promiscuously with any potential target; recognition following adhesion is necessary for lethal hit delivery. The fact that antibodies to the antigen receptor (Ti or CD3) inhibit killing without inhibiting CTL-target conjugation supports this view. The ability of lymphocytes to form nonspecific adhesions, plus the dependence of even the specific mouse adhesions on temperature, metabolic energy, magnesium, and an intact cytoskeleton suggest that the bulk of the strength of T lymphocyte adhesions are not simply the sum of the bonds between antigen receptors (Ti) and antigen. Lymphocytes evidently possess separate "adhesion strengthening" mechanisms. The similarities in the properties of CTL-target adhesions and antigen-independent homotypic B lymphocyte adhesions (Table 2) suggest that at least some of these mechanisms are widely used among cells of hematopoietic origin. MoAbs to most lymphocyte surface molecules, when bound to the living lymphocyte membrane, have no evident functional effects on lymphocyte function. However, a minority can either activate or inhibit lymphocyte functions. Such antibodies identify "leukocyte (or lymphocyte) function-associated antigens," or LFAs (not all of which happen to have "LFA" in their names, Table 1). Most of the inhibitory antibodies inhibit lymphocyte adhesions, and this appears to account for their inhibitory effects on functions such as killing or proliferation. The fact that the binding of antibodies to a particular membrane glycoprotein inhibits adhesion does not guarantee that the glycoprotein in question is a direct participant in adhesion (one of the "glue" molecules). However, there is scanty evidence in support of indirect "negative signals" that may be induced by such antibodies, and direct participation of most LFAs in adhesion seems likely.(ABSTRACT TRUNCATED AT 400 WORDS)