no paraproteins could be detected by capillary electrophoresis. However one of these samples was positive in the gel and contained an IgM paraprotein type lambda of 1 g/1. When the observers looked retrospectively at this capillary electropherogram it showed an extra peak. Capillary electropherograms are easily interpreted in the case of monoclonal bands appearing in the gamma region. However monoclonal bands in other regions are more difficult to detect, which is also seen in agarose gel electrophoresis. After seeing the outcome of the agarose gel electrophoresis in the case of the missed monoclonal band, it was clear that one has to become familiar with the capillary electrophoresis patterns. This comparison study clearly shows that with both techniques interpretations are sometimes difficult to make. However from table 1 it is clearly shown that no paraproteins are missed when capillary electrophoresis is used, except one sample which has been explained. This makes capillary electrophoresis a useful technique for serum protein analysis. Comparing the results of this study with those of Wijnen & Van Dieijen-Visser (3), we conclude that serum protein electrophoresis might be performed using capillary electrophoresis instead of the widely-used agarose gel electrophoresis. The capillary electrophoresis system has the advantage of low reagent costs, and can be easily automated. In our opinion capillary electrophoresis for the analysis of serum proteins might be used as an alternative for agarose gel electrophoresis. We therefore speculate that in the future agarose gel electrophoresis for serum protein analysis will be replaced by capillary electrophoresis.