Many bushcrickets produce specific song spectra for acoustic communication. Song detection and/ or recognition may make use of such specificity. Where in the nervous system are the filters for song frequency situated? A peripheral tuning for song frequency typically does not exist. Auditory receptor cells of bushcrickets connect to local and ascending neurons in the prothoracic ganglion. One of the ascending neurons (1) may function as a frequency filter in a group of four related bushcrickets (genera Ancistrura, Barbitistes). The frequency response of ascending neuron 1 is species-specific roughly corresponding to the frequency of the conspecific male song. The species-specific tuning of the neuron is not brought about by specific excitation, but by specific inhibition. By eliminating this frequency-dependent and species-specific inhibition the former filter neuron is transformed into an unspecific broad-band neuron in all four species. Its tuning then does not differ from omega neuron 1, a local neuron which is rather unspecific for frequency. Also, the supra-threshold responses of ascending neuron 1, which are different in intact animals, are similar to each other and similar to omega neuron 1 following elimination of inhibition. Only ascending neuron 1 of Ancistrura retains some species-specific features at low frequencies. In conclusion, evolution changed inhibition, not excitation of a species-specific neuron.