In a collaborative study involving four hospital laboratories Immunoglobulins G, A and M were assayed in 12 human serum specimens. The aims were to define the magnitude of inter-laboratory discrepancies in the values obtained, to identify likely reasons for these discrepancies, and to ascertain whether they were reduced when Immunoglobulins were assayed by comparison with a British Standard preparation. The results of this experiment emphasise the importance of continuing experience in Immunoglobulin assay techniques; there may be considerable variation in the results obtained when the same specimen is tested in different laboratories, or at different times in one laboratory; these discrepancies are reduced if laboratories use reagents from the same source (commercial sources in this case). When the usual laboratory Immunoglobulin standards were replaced by a British Standard preparation, there was no improvement in the results for IgG and IgA assays, but inter-laboratory variation in IgM measurements was significantly reduced.