The pulsed field method has been used to study the de Haas–van Alphen effect in rhodium. Using suitable filtering techniques it was possible to separate and measure over a wide range of field orientations the many frequencies present. From the data several separate sheets of Fermi surface could be identified and their dimensions determined with little ambiguity. These dimensions are compared with a Fermi surface predicted by assuming the band structure of rhodium to be like that theoretically predicted for nickel but choosing the Fermi level to agree with the different number of electrons in rhodium. This model explains satisfactorily all the frequencies observed and explains why several additional frequencies that have not been observed are below the noise level. The behaviour of the transverse magnetoresistance of rhodium is predicted and this could be used for a further test of the correctness of the model. Those values of cyclotron mass that have been measured, although necessarily only roughly, disagree significantly with this simple model but a more satisfactory agreement is obtained by considering a ‘semi rigid band model' in which the width of the
d
bands varies from metal to metal. If the
d
bands in rhodium are assumed to be of order twice the width of those in nickel both the dimensions of the Fermi surface and the cyclotron masses can be satisfactorily explained.
