C-reactive protein (CRP) was assessed in pediatric serum samples using different commercial latex reagents, which were analyzed for species origin of the coating antibodies, homogeneity and density of the latex particles, and prozone agglutinating capacity. All reagents correctly agglutinated the positive and negative control sera. The antibodies coating the particles differed with regard to species origin: one was coated with rabbit, one with horse and goat, one with horse, goat, rabbit and swine, while the reference reagent had horse, goat and rabbit antibodies. Only the monospecies specific antibody-coated latex showed obvious prozoning; this reagent also had the smallest and most homogenous latex particles and showed the most clear-cut reactions. False agglutination was observed at 7-26% according to quantitation with the spot immunoprecipitate assay, which compared favorably with radial immunodiffusion measurements. The lowest percentage of false readings was noted for the rabbit antibody-coated particles; the highest for the reagent with particles coated using antibodies from 4 different species. No reagent had satisfactory precision for the low positive sera between 10 and 40 mg CRP/l.