Previous studies have shown that lymphocytic infiltrates of different mouse mammary tumors contain different proportions of the T cell subsets Lyt-1+ and Lyt-2+. These characteristic subset ratios may be established, at least in part, by differential locomotion of subsets in response to components of the tumor microenvironment, such as soluble chemotactic and chemokinetic factors, cell and stromal surfaces, oxygen tension, and pH. We describe a new in vitro assay for determining how such microenvironmental variables affect lymphocyte locomotion. Suspended lymph node cells, alone or mixed with other cell types, are sandwiched between two layers of type I collagen gel and bathed in culture medium. A halo of locomotory cells fans out around the flattened droplet. Locomotion requires energy and exogenous protein. After a 24-hr incubation at 37 degrees C, 10% CO2 in air, the cell density of the halo is analyzed optically and the subset ratios are characterized by immunofluorescence staining of the gel sandwich. Under standard culture conditions, the locomotory population is enriched 12% in Thy-1+ cells compared with the bulk population, and the ratio of Lyt-1+ to Lyt-2+ cells is significantly increased. Lymphocyte locomotion is inhibited by 1 microM PGE2, by decreased pH and oxygen tension, and by the presence of normal mammary cells or mammary adenocarcinoma cells. The Lyt-1+:2+ ratio in the locomotory population is not altered by PGE2 but is reduced by acidity and hypoxia. The ratio is also reduced by the presence of mammary cells and cells of one of the mammary adenocarcinoma cell lines tested (168) but not by two others (68H and 410). Our data support the hypothesis that the locomotion of T cell subsets is differentially responsive to the types of microenvironmental conditions that vary among tumors.