Reconstituted basement membrane matrix (Matrigel) has been utilized for in vitro assay of tumor cell invasion in recent years. In the conventional chamber for the invasion assay, however, a large number of cells passed easily through the center of the Matrigel-coated filter because the Matrigel layer could not be completely uniform by the meniscus formation. To prevent the meniscus phenomenon of the Matrigel layer, we devised a water-repellent treatment of the inside wall of the assay chamber with paraffin. Consequently, very few erythrocytes passed through the Matrigel-coated filter of this modified chamber with the erythrocyte assay, which was used to demonstrate the evenness and uniformity of the Matrigel layer on the filter. For quantitating a small number of cells which invaded through the Matrigel-coated filter by the invasion assay, a tetrazolium-based colorimetric 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide) assay was used. The invasive abilities of the eight different cells were determined by this invasion-3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium+ ++ bromide assay using the modified chamber with a filter coated with 70 microliters of the 0.2-mg/ml Matrigel. After 72 h of incubation, the malignant cell lines significantly exceeded the normal cell lines in the percentage of invasion (P < 0.01). Therefore, the modified invasion-3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium+ ++ bromide assay provides a simple, easily reproducible in vitro assay for quantitating tumor cell invasion.