An interlaboratory collaborative trial was conducted on the determination of serum copper using two different methods, based on colorimetry (test combination Copper, Boehringer Mannheim, Mannheim, Germany) and flame atomic absorption spectrometry (FAAS). The general performance of the colorimetric method was below that of FAAS, except for sensitivity and linear range, as assessed by detection limit (0.44 versus 1.32 mumol/L) and upper limit of linearity (150 versus 50 mumol/L). The range of the between-run CVs and the recovery of standard additions were, respectively, 2.3-11.9% and 92-127% for the colorimetric method and 1.1-6.0% and 93-101% for the FAAS method. Interferences were minimal with both methods. The two techniques correlated satisfactorily (the correlation coefficients ranged from 0.945-0.970 among laboratories) but the colorimetric assay exhibited slightly higher results than the FAAS method. Each method was transferable among laboratories.