An indirect enzyme linked immunosorbent assay (ELISA) (Parvoscan-B19; Sweden) was compared with an in-house MACRIA for the detection of B19 specific IgM. A Parvoscan-B19 IgG test was also evaluated for its ability to detect a recent B19 infection in paired sera. Two hundred and twenty sera submitted to the laboratory for B19 serology and four MACRIA positive control sera were assayed for B19 IgM. Confirmation of the response of sera giving discordant results in the two assays was sought by the use of a "nested" polymerase chain reaction (PCR) for the detection of B19 DNA. The Parvoscan-B19 IgM test was 79% sensitive and 96% specific. Parvoscan-B19 was poor at detecting parvovirus infection in sera collected two to three months after the onset of symptoms. When sera collected more than seven weeks after the onset of symptoms were excluded from the analysis, Parvoscan-B19 IgM was 84% sensitive and 96% specific. Rubella specific IgM positive sera, rheumatoid factor positive sera, and heterophil antibody positive sera were also assayed for B19 IgM. No false positive results were encountered with these problematic sera. By using the cut off criteria for the Parvoscan-IgM test previously advocated by the manufacturers, 90% sensitivity and 87% specificity could be achieved. False positive results, however, occurred with six of the 17 rubella IgM positive sera, four of the 10 rheumatoid factor positive sera, and two of the 11 heterophil antibody positive sera tested. It is concluded that the Parvoscan-B19 was specific but insensitive when compared with in-house assays.