Those of us ‘of a certain age’ have seen huge changes in the way labs work during our careers. Mattingly cortisols, Lieberman Burchard reagent for cholesterol and Kober reagent for oestriols are nowhere to be found in a modern Clinical Chemistry Lab. Bence Jones protein (BJP), however, has stood the test of time. It is the earliest described ‘tumourmarker’and still provides a high index of suspicion for B-cell malignancy. Henry Bence Jones described a precipitate that was formed on the addition of nitric acid to urine. This precipitate redissolved on heating and formed again on cooling and we now know that it is a protein. It is rather humbling that all Bence Jones had was a bit of acid and a £ame to make a discovery that is still in the common phrase book of pathologists around the world. Today, we ask for serum and urine samples to investigate patients with suspected B-cell malignancy and we use sensitive and reliable electrophoretic techniques to look for the presence of monoclonal proteins, to con¢rm clonality and to identify the heavy and/or light chain types of the monoclonal proteins. However, over the last few years there have been suggestions that this investigation process needs to be updated and in particular, checking urine for BJP should be a thing of the past! In 2001, Bradwell et al. reported the development of a sensitive automated immunoassay for immunoglobulin (Ig) free light chains. The FREELITE test measures free k and free l light chains and, because it is reported that k/l ratio remains normal except in the presence of monoclonal plasma cell proliferations, Katzmann et al. use the calculated k/lratio to infer amonoclonal excess of one particular light chain. There has been considerable interest in this approach for the investigation of patients with suspected and con¢rmed B-cell dyscrasias. However, as with any new test, we have to establish that it is better than the existing tests.We need to decide, based on our knowledge and experience, whether the test can do what we ask of it.We need to thoroughly evaluate it, using the populations that would normally be investigated with that test. Finally, we need to de¢ne when and where and how it should be used. So before we decide to radically change our lab practices for the investigation of B-cell malignancy, wemust look verycarefullyat the FREELITE assays and assess its validity at both analytical and clinical levels. First, a reminder of basic physiology; we all produce small amounts of polyclonal free light chains as a result of normal B-cell function. These pass readily from the blood through the glomerulus and we can see them in normal urine by sensitive immuno¢xation. In patients with renal disease or an underlying in£ammatory process or even in the elderly, the production or excretion of free light chains can increase. In some urine samples the free light chains are seen as an oligoclonal or ladder pattern but they are not BJP. Free light chains can be monoclonal, oligoclonal or polyclonal but BJP consists of only monoclonal free light chains. Unfortunately, there is no antiserum available that has the capacity to distinguish monoclonal Igs from polyclonal Igs (or monoclonal light chains from polyclonal light chains) in serum or in urine.We know this from our routine Ig concentrations on patients with paraproteins and we see it in external quality assurance (EQA) results. Monoclonal proteins can behave strangely in immunoassay even when using highly speci¢c and puri¢ed antiserum. For example, an IgG concentration of 5 or 15 g/L or even of 50 g/L can all be due to polyclonal IgG, monoclonal IgG or oligoclonal IgG. The concentration does not tell us about the clonality of the IgG. Interpreting the IgG concentration with the IgA and IgM concentrations does not help much -the electrophoretic pattern is essential for assessing the clonality of the quantitative Ig results. Monoclonal Igs do not behave like polyclonal Igs in immunonephelometric or immunoturbidimetric assays. The reasons for this are varied but include marked diierences between the polyclonal protein used as the antigen for the antibody production, diierences between antigenicity of the monoclonal protein and the standard used to calibrate the assay and the marked variability in dilution curves for individual paraproteins. The wide range of possible concentrations of monoclonal proteins also makes the inclusion of robust antigen excess checks vital for quantitative assays where monoclonal proteins may be seen. Overall, the results given in immunoassays for monoclonal Igs generally do not provide an accurate quanti¢cation. We assess broad assay reliability by participation in EQA.The FREELITE assay is produced byonly one companyand the reagents and standards are from the same source. However, the results generated in the EQA scheme show dramatically high coe⁄cients of variation -across all the assays (and the ratio), over many distributions, across all the analytical platforms and