Sixty canine serum samples were sent to 10 different diagnostic laboratories for anti-Borrelia burgdorferi antibody analysis. All laboratories knew of the study prior to receiving the samples. Agreement among all laboratories for all interpretations was 91% (546 of 600 samples). There was complete agreement among all the laboratories for only 32 (53%) of the samples. Most of the disagreements were due to differences reported by either one (15 samples) or two (7 samples) laboratories per sample. When discrepancies in interpretations existed, the interpretation reported by the majority of the laboratories was considered the standard for comparison. One laboratory had no discrepant interpretations from this standard, while the laboratory with the most discrepancies had 16. The median number of discrepancies per laboratory was five. By using pairwise comparisons between each laboratory and the majority standard, eight of the laboratories showed strong agreement and the remaining two showed fair to good agreement. The type of test used (enzyme-linked immunosorbent assay versus indirect immunofluorescence assay) did not appear to influence the number of discrepant interpretations reported. Sera considered to be positive by the majority of the laboratories usually reacted to more than five antigens in immunoblots, with at least three or more of those being intense reactions. For positive samples, reactivity was consistently present in the 60-, 41-, 31-, and 22- or 24-kDa regions. Samples considered negative usually reacted to fewer than three bands, with reactivity usually being faint. A commercially available, rapid dot blot assay showed strong agreement with the majority standard.