1. A group of 7 acoustically activated interneurones has been identified in the central nervous system of the noctuid mothHeliothis virescens. The neurones have their cell bodies and major processes in either the mesoor metathoracic neuromeres of the fused pterothoracic ganglion, and axons contralateral to the cell body (Figs. 1, 2). All neurones except one project via the contralateral connective to at least the prothoracic ganglion. 2. The frequency-response characteristics and auditory thresholds for most interneurones were similar to that of the A1 receptor (40–45 dB SPL at 16 kHz) (Fig. 3A). Interneurone intensity-response characteristics were also generally of the sigmoid type displayed by the A1 and A2 receptors (Fig. 3B). 3. Each interneurone responded to a tone with a compound EPSP of characteristic shape (Figs. 2, 4, 5). In two neurones with very similar responses, the shapes of the rising and plateau phases of their EPSPs, and the suggestion of a delayed input to one neurone, were sufficient to distinguish them (Fig. 5). 4. The subthreshold responses of all identified interneurones considerably outlasted the stimulus duration, but spiking patterns reflected stimulus duration more accurately (Figs. 4, 5, 6). Most neurones responded in a tonic or phasic/tonic manner, suggesting they might be ‘repeater’ type neurones; whereas one local (Fig. 2) and one unidentified interneurone (Fig. 9) responded like ‘pulse-coder’ neurones. 5. All responses to sound by identified interneurones were excitatory, and no sound-related IPSPs were seen (Figs. 2, 3, 4). Further, little habituation of the response occurred in any interneurone for stimulus rates up to 20/s (Fig. 7A, B). 6. The latency of the EPSP from stimulus onset in several interneurones (501, 503, 504) was in each case equivalent to a synaptic delay of less than 1 ms compared with the response latency of the A1 receptor, suggesting direct connectivity (Table 1). The input-output curves and EPSP of neurone 501 changed in such a way as to suggest a possible additional input from the A2 receptor at higher intensities (Figs. 3 B, 4). 7. Electrical stimulation of the tympanal nerve and sound stimulation of the ear evoked a similar EPSP and spike pattern in interneurone 502 (Fig. 8A, B). However, the latency of the EPSP in neurone 502 elicited by electrical stimulation was too long (9 ms) for the connection between neurone 502 and either A1 or A2 receptors to be direct. 8. The neuronal morphologies and response characteristics described above provided the basis for a simple model of the neural circuitry mediating the phonotactic response of a moth to a bat call (Fig. 10).