Being high Q devices, low frequency underwater transducers often lack bandwidth. Variable resonance frequency transducers offer the double advantage of increased effective bandwidth and maximum response at all frequencies within the bandwidth. This paper presents and evaluates a technique to vary the first resonance frequency of some widely used underwater acoustics transducers: flexural piezoceramic bars and disks. DC bias electric fields are added to the AC driving field and used to generate in-plane tensile or compressive loads. These loads modify the flexural rigidity of the transducer, which in turn affects its resonance frequencies. Theoretical investigations show that the frequency shift per DC field is linked to the ratio of the in-plane blocked force to the critical buckling load of the transducer. This ratio depends on the type of piezoceramic material and coupling, the boundary conditions, the length to thickness ratio of the transducer, and the piezoceramic thickness and coverage. Calculations show that both significant frequency shifts per DC field and acceptable device coupling coefficients may be achievable in practice. A flexural bar transducer using k31 coupling was built and tested. The experimental frequency shift per DC field and coupling coefficient were lower than predicted. Measurements show the existence of a polarization switch due to a combined compressive stress and negative field effect at -400 V/mm DC field. This polarization switch limits the range of useful negative DC fields, therefore limiting the total frequency shift, and also results in a permanent reduction of the polarization level, therefore reducing the amount of frequency shift per DC field. In the case of k31 coupling, one must determine the safe stress and field region for the material and try to operate within the corresponding DC field range. In the case of k33 coupling, compressive stresses and negative fields do not occur simultaneously, and the available DC field range should be much higher.