When Meth-A fibrosarcoma-bearing BALB/c mice were injected subcutaneously with 10 micrograms of recombinant human interleukin 2(rIL-2) once a day for 10 days, tumour growth inhibition was in the range of 22-31% of that of the control animals. Anti-tumour effector cells against Meth-A were detected in the spleen cells of the tumour-bearing BALB/c mice injected with rIL-2, using a modified Winn-type neutralization assay with the auxiliary injection of rIL-2. To induce the strongest anti-tumour activity in this assay system, the following were necessary: 1) the effector cells were derived from tumour-bearing BALB/c mice; 2) the donors of the effector cells were injected with rIL-2; 3) the recipient mice in the Winn assay were auxiliarily injected with rIL-2 (a modified Winn assay). The anti-tumour effector activity detected in the modified Winn assay was inhibited by treatment with anti-CD8 or anti-asialo GM1 antibodies plus complement (C), but not completely. We supposed that at least two kinds of anti-Meth-A effector cells with different surface antigens, positive for CD8 and asialo GM1 antigens, were induced in the Meth-A-bearing BALB/c mice injected with rIL-2; these populations seemed to function independently and at least partly as anti-tumour effector cells in this tumour-host system. These spleen cells showed in vitro cytotoxicity against Meth-A cells, which are resistant to NK cells, if the activity was measured in a 24 h 51Cr-release assay in the presence of rIL-2.