We have compared microwave imaging data for a small flare with simultaneous hard X-ray spectral observations. The X-ray data suggest that the power-law index 5 of the energy distribution of the radiating electrons is 5.3 (thick-target) which differs significantly from the estimate (b = 1.4) from a homogeneous optically-thin gyrosynchrotron model which fits the radio observations well. In order to reconcile these results, we explore a number &options. We investigate a double power-law energy spectrum for the energetic electrons in the flare, as assumed by other authors: the power law is steep at low energies and much flatter at the higher energies which produce the bulk of the microwaves. The deduced break energy is about 230 keV if we tentatively ignore the X-ray emission from the radio-emitting electrons: however, the emission of soft photons by the flat tail strongly contributes to the observed hard X-ray range and would flatten the spectrum there. A thin-target model for the X-ray emission is also inconsistent with radio data. An inhomogeneous gyrosynchrotron model with a number of free parameters and containing an electron distribution given by the thick-target X-ray model could be made to fit the radio data.