The diagnosis of human infection by Toxocara canis relies heavily upon serological tests, the specificity of which can be inadequate in regions of endemic helminthiasis. When different population groups of tropical Venezuela were evaluated using ELISA based upon Toxocara excretory-secretory antigen (TcESA), solid-phase adsorption of the sera with extracts of a wide variety of non-homologous parasites revealed the existence of significant cross-reactivity. This was effectively and conveniently overcome when the test sera were incubated in the presence of the soluble parasite extracts in a competitive inhibition ELISA. The mean reduction of ELISA values caused by pre-adsorption of the sera tested was 32.2%, and that caused by competitive inhibition was 42.3%, the effects of these two procedures being strongly correlated (r = 0.83). The magnitude of the reduction was inversely proportional to the actual ELISA value (r = -0.55), and ranged from a mean of 68.0% in sera from apparently healthy individuals of medium-high socio-economic level, down to 28.1% in heavily parasitized Amazon indians. Ascaris showed the greatest degree of cross-reactivity in these tests, although under conditions of competitive inhibition even sera with high levels of antibody against this parasite could be negative in Toxocara ELISA. Western blotting revealed a major 81,400 D component that was shared between Ascaris and TcESA. Our results indicate that the competitive inhibition of cross-reactivity by soluble non-homologous parasite extracts provides a convenient and economical means of increasing the specificity of ELISA for the determination of the seroprevalence of toxocariasis in tropical populations.