Fig. 9. Adaptation of responses to long stimuli. At a stimulus duration of 1000 ms, the sensillar potential (SP) amplitude (A), the initial slope (B), the half time of the rising phase of the SP (t_{1/2 rise}; C), the half time of the decline of the SP (t_{1/2 decline}; D) and the action potential (AP) frequency (E) and latency (F) were dose-dependent (solid lines). After adaptation by a 1000 ms stimulus applied 60 s before the test stimuli, these dose–response curves were shifted to higher stimulus intensities (broken lines). As with stimuli of 50 ms duration (Fig. 7), the shift was larger for the AP response than for the SP response. Responses in the adapted state were recorded to a maximal dose of 100 µg bombykal. The data were normalized to the highest response during each recording, which is the largest numerical value for those variables positively correlated to the stimulus intensity. The AP latency and t_{1/2 rise}, which are negatively correlated to the intensity, were inversely normalized to the smallest numerical value to focus on responses in the physiological dose range. C, control. Data represent means ± S.E.M. Sample sizes (N) range between 2 (10^–2 µg; adapted) and 22 (1 µg; non-adapted). Asterisks indicate significant differences between the adapted and non-adapted state (*P<0.05; **P<0.01; Student's t-test).