To evaluate the effects of dietary fats on breast cancer growth and metastasis, KPL-1 human breast carcinoma cells which have a propensity for axillary lymph node metastasis when inoculated into the thoracic mammary fat pad of female nude mice were examined. The mice were fed one of three semipurified diets containing 9.5% eicosapentaenoic acid plus 0.5% linoleic acid (EPA diet), 10% linoleic acid (LA diet), or 9.5% palmitic acid plus 0.5% linoleic acid (PA diet), or commercial laboratory chow containing 8.5% fat of which 4.1% was LA, 1.1% was PA, 0.06% was EPA, and 3.24% was other (Standard diet) starting 19 days before tumor cell inoculation and continuing until the end of the experiment (43 days after tumor cell inoculation). The tumor growth was faster and at a higher incidence in the mice fed the LA diet, and much slower and at a lower incidence in the EPA diet group compared with the mice fed the PA or Standard diet; the two separate experiment demonstrated identical results. The differences in tumor weight between the LA and PA groups and between the PA and EPA groups were significant (P < 0.05, respectively) at the termination of the experiment; the differences were due to different tumor cell proliferation rates. In an in vitro MTT assay, fatty acids showed direct stimulatory or inhibitory effects on the KPL-1 cells. Lymph node metastasis was seen in the LA and Standard diet groups, whereas it was not seen in the PA or EPA groups. The body weights were significantly lighter in the LA and EPA groups compared with the PA and Standard diet groups (P < 0.05, respectively). The results indicate that the EPA diet produced a reduction in tumor cell growth and metastasis whereas the LA diet had an enhancing effect on these parameters; dietary fatty acids may thus have a direct role in the growth and metastasis of human breast carcinoma independent of their systemic effects.