The multiple-wavelength anomalous dispersion technique for deriving phase information in the determination of the structures of biological macromolecules can be practised on both ID14/4 and BM14; the latter beamline will be superseded by ID29 during the course of 2000. During 1999 some 73 MAD experiments were undertaken on BM14 of which only one has been reported as unsuccessful. The edges covered (with number of structures in parentheses) include Se (44), Hg (5), Pt (3), Br (3), Pb (2), Fe (6), Zn (3), Cu (2), and one each for Ni, Gd, Eu, Ho and W. ID14/4 only commenced a full user service for MAD in May 1999, but still some 32 successful experiments were performed for Se (24), U (2), Br (3), Hg (1), Pt (1) and As (1). These impressive numbers are in addition to the routine single wavelength data collection experiments. BM1 (SNBL) and BM30A (FIP) CRG beamlines have also undertaken successful MAD experiments. The speed of MAD experiments is also astonishing with reports of several new structures being solved per day by various groups. Figure 13 shows part of an electron density map derived from a three wavelength experiment on a seleno-methionine substituted protein. The map is clearly interpretable in terms of the polypeptide chain and its production was achieved only 50 minutes after the data collection was commenced. Each data set took only 5 minutes to collect using the ADSC Quantum 4 detector system on ID14/4, whereas the determination of the Se positions (4 Se per 32.6 kDa asymmetric unit) took some 24 minutes; some of the remaining minutes were occupied by computation, but one also suspects that the experimenters had time for a cup of coffee! Such speed in optimal circumstances is an indication of our capabilities for the immediate future, and with increasing automation of both data collection methods and phase determination, the ESRF should be able to cope with future needs in this area.