AIMS: To determine whether the Eiken particle agglutination test could be modified to make it sufficiently sensitive to screen blood samples collected on Guthrie cards for the presence of antibodies to Toxoplasma gondii; to evaluate the specificity of the modified system; and to compare seroepidemiological data on the prevalence of T gondii in pregnant women. METHODS: Simulated dried blood spots were prepared from sera from pregnant women booking for antenatal care. Eluates from the simulated dried blood spot cards and sera were tested in parallel using the modified test (1 in 5 dilution of latex) and the standard assay (neat latex particles) and endpoints determined. Guthrie card eluates, from neonates in three Thames regions, were then tested using the modified test. RESULTS: The modified test produced a 4.21-fold increase in antibody titre in 85 sera when tested in parallel with the standard test. Eluates of 168/170 from simulated dried blood spots derived from seropositive patients gave a positive result in the modified test. The two eluates which gave a negative result were derived from patients with an equivocal titre of 1/16 in the standard serum test. Of the eluates derived from serum negative patients all 103 were negative at a dilution of 1 in 4 in the modified test. The seroprevalence of antibodies to T gondii in pregnancy was 21.8% using the standard test. A similar value of 20.5% was obtained when dried blood spots from neonates in a similar region of London were tested by the modified test. CONCLUSIONS: The modified Eiken Toxo-reagent test is sensitive, simple, and economic for screening large numbers of dried blood spots. The procedure could be easily semiautomated and the technique applied to the mass screening of neonatal blood samples collected on Guthrie cards to determine the seroprevalence of T gondii in pregnant women.