In this study, polyclonal and monoclonal antibodies to native and denatured chicken ovalbumin (OVA) were produced to compare their dependency on continuous and three-dimensional epitopes. These antibodies were characterized with respect to reactivity to native and denatured OVA by enzyme-linked immunosorbent assay (ELISA) employing surface-bound OVA and streptavidin-capture ELISA to determine whether effects of different coating influence antibody specificity and with respect to epitope specificity by peptide ELISA, using overlapping peptides, covering the complete OVA sequence. Polyclonal antibodies to native OVA reacted strongly with native and denatured OVA in both assays, but did not react with the overlapping peptides. Polyclonal antibodies to denatured OVA reacted strongly with both OVA forms and with several of the overlapping peptides. Monoclonal antibodies to native OVA reacted preferentially with three-dimensional epitopes on native OVA and not with denatured OVA. Monoclonal antibodies to denatured OVA showed reactivity to both OVA forms. Two of these monoclonal antibodies, HYB 94-06 and 94-07, showed reactivity to overlapping peptides and their epitopes were identified as flexible structures constituting amino acids 130–135 and 136–141, respectively. Moreover, comparison of antibody reactivity to N OVA revealed that in the streptavidin-capture ELISA, antibody reactivity was notably reduced compared to ELISA employing surface-bound OVA. Collectively, immunization with native OVA preferentially generates highly specific antibodies reacting with three-dimensional epitopes, whereas immunization with denatured OVA generates antibodies occasionally reacting with continuous epitopes. Moreover, as differences in monoclonal antibody reactivity was found between the two assays, monoclonal antibodies always should be selected by an assay mimicking the desired use of the final antibodies as closely as possible.