Tethered, flying Australian field crickets (Teleogryllus oceanicus) stimulated with ultrasound respond with a rapid, short-latency turn from the sound source. We analyzed the kinematics of two behavioral components of this acoustic startle response and recorded electromyograms from the muscles involved in producing them. The two behavior patterns studied were the swing of the metathoracic leg, which has been shown to elicit a short-latency turn, and a lateral swing of the antennae, for which a direct role in steering has not been demonstrated. The kinematic data showed that when a pulse of ultrasound was presented to one side of the animal (1) the contralateral metathoracic leg abducted and elevated, while the ipsilateral leg remained in place, (2) both antennae swung laterally, but the contralateral antenna moved farther than the ipsilateral antenna, (3) increases in stimulus intensity elicited larger movements of the leg and contralateral antenna, while the ipsilateral antenna showed little sensitivity to stimulus intensity, and (4) for the leg, the latency to the onset of the swing decreased and the duration of the movement increased with increasing stimulus intensity. Electromyograms were recorded from the leg abductor M126 and two antennal muscles: the medial scapo-pedicellar muscle M6 and the lateral scapo-pedicellar muscle M7. M7 moves the antenna laterally, M6 moves it medially. Upon stimulation with ultrasound (1) both M126 and M7 showed increasing spike activity with increasing intensity of the ultrasound stimulus, (2) M126 showed a decrease in latency to the first spike and an increase in the duration of spike activity with increasing stimulus intensity, (3) latencies for M6 and M7 were not correlated with stimulus intensity, but M7 had significantly shorter latencies than M6 and the contralateral M7 had significantly shorter latencies than the ipsilateral M7, and (4) the ipsilateral M126 spiked in response to ultrasound in 6 of the 10 animals tested. In these cases, however, latency to the first spike was substantially longer, and the spike frequency was lower than for the muscle's response to contralateral stimuli. We attempt to correlate these electromyogram data with the kinematic data and relate them to the relevance of the two behavior patterns to the execution of an escape response.