The pine caterpillar, Dendrolimus punctatus (Walker) (Lepidoptera: Lasiocampidae), is a multivoltine pest of pine trees in China, overwintering as larvae. Winter diapause was induced by short day length. The critical night length was about 10 h 40 min at 25, 28, and 31°C in the field, showing a temperaturecompensated diapause induction. Transfer experiments from a short night (L16: D8) to a long night (L12: D12) or vice versa at different times after hatching showed that sensitivity to day length was restricted to the first 14 days; the required day number for a 50% response at 25°C was about 3.5 days for short nights but 7.5 days for long nights, indicating that short nights are photoperiodically more effective. When four successive short nights (L16: D8) were used to interrupt the long-night regime (L12: D12) at different development stages and vice versa, the results showed that the highest sensitivity to photoperiod occurred on the 4th−8th day, corresponding to the second larval instar. Experiments of alternating short-night (L16:D8) and long-night (L12:D12) cycles during the larval period showed that the information of short nights as well as long nights could be accumulated. By rearing the larvae under conditions other than 24-h light–dark cycles, we clearly showed that the dark period (scotophase) played a major role in the determination of diapause. The Nanda-Hamner and Bünsow experiments failed to reveal rhythmic fluctuations with a period of about 24 h in the occurrence of diapause. Therefore, the photoperiodic clock in D. punctatus is an hourglass timer or a damped circadian oscillator.

Induction of larval diapause is a photoperiodically controlled event in the life history of the moth Pseudopidorus fasciata. In the present study, the photoperiodic counter of diapause induction has been systematically investigated. The required day number (RDN) for a 50% response was determined by transferring from a short night (LD 16:8) to a long night (LD 12:12) or vice versa at different times after hatching, The RND differed significantly between short- and long-night cycles at different temperatures. The RDN for long-night cycles at 20, 22, 25 and 28▫C was 11.5, 9.5, 7.5 and 8.5 days, respectively. The RDN for short-night cycles was 3 days at 22▫C and 5 days at 20 1C indicating that the effect of one short night was equivalent to the effect of 2–3 long nights effect. Night-interruption experiments of 24 h photoperiods by a 1 h light pulse showed that the most crucial event for the photoperiodic time measurement in this moth was whether the length of pre-interruption (D1) or the post-interruption (D2) scotophases exceeded the critical night length (10.5 h). If D1 or D2 exceeded 10.5 h diapause was induced. The diapause-averting effect of a single short-night cycle (LD 16:8) against a background of long nights (LD 12:12) showed that the photoperiodic sensitivity was greatest during the first 7 days of the larval period and the highest sensitivity was on the fourth day. Both non-24 and 24 h light-dark cycle experiments revealed that the photoperiodic counter in P. fasciata is able to accumulate both long and short nights during the photosensitive period, but in different ways. The information from short-night cycles seems to be accumulated one by one in contrast to long-night cycles where six successive cycles were necessary for about 50% diapause induction and eight cycles for about 90% diapause. These results suggest the accumulation of long-night and short-night cycles may be based on different mechanisms.

In the zygaenid moth, Pseudopidorus fasciata, both larval diapause induction and termination are under photoperiodic control. In this study, we investigated whether photoperiodic time measurement (with a 24-h light–dark cycle) in this moth is qualitative or quantitative. Photoperiodic response curves, at 22, 25, and 28 1C indicated that the incidence of diapause depended on whether the scotophases exceeded the critical night length (CNL) or not. All scotophases longer than the CNL-induced diapause; all scotophases shorter than the CNL-inhibited diapause. The CNL was 10.5 h at 25 and 28▫C, and 10 h at 22▫C. By transferring from various short photoperiods (LD 8:16, LD 9:15, LD 10:14, LD 11:13, LD 12:12, and LD 13:11) to a long photoperiod (LD 16:8) at different times, the number of light–dark cycles required for 50% diapause induction at 25 1C was 7.14 at LD 8:16, 7.2 at LD 9:15, 7.19 at LD 10:14, 7.16 at LD 11:13, and 7.13 at LD 12:12, without showing a significant difference between the treatments. Only at LD 13:11 (near the CNL), the number of light–dark cycles was significantly increased to 7.64. The intensity of diapause induced under different short photoperiods (LD 8:16, LD 9:15, LD 10:14, LD 11:13, and LD 12:12) at 25 1C was not significantly different with an average diapause duration of 36 days. The duration of diapause induced under LD 13:11 was significantly reduced to 32 days. All results indicate that the night-lengths are measured as either ‘‘long’’ or ‘‘short’’ compared with some critical value and suggest that photoperiodic time measurement for diapause induction in this moth is based on a qualitative principle.

Pseudopidorus fasciata enters diapause as fourth instar larvae at short day lengths. Using 24-h light–dark cycles, the photoperiodic response curves in this species appeared to be similar with a critical night length of 10.5 h at temperatures below 30▫C. At an average temperature of 30.5▫C, the critical night length had shifted to between 15 and 17 h. In experiments using non-24-h light-dark cycles, it was clearly demonstrated that the dark period (scotophase) was the decisive phase for a diapause determination. In night interruption experiments using 24-h light–dark cycles, a 1-h light pulse at LD12:12 completely reversed the long night effect and averted diapause in all treatments.. At LD 9:15 light pulses of 1-h, 30- or 15-min also averted diapause effectively when both the pre-interruption (D1) or the post-interruption scotophases (D2) did not exceed the critical night length. If D1 or D2 exceeded the critical night length diapause was induced. The most crucial event for the photoperiodic time measurement in this species is the length of the scotophase. A 10-min light pulse placed in the most photosensitive phase reversed diapause in over 50% of the individuals. Night interruption experiments under non-24-h light–dark cycles indicated that the photoperiodic clock measured only D1 regardless of the length of D2, suggesting that the most inductive cycles are often those in which LD are close to 24 h. In resonance experiments, this species showed a circadian periodicity at temperatures of 24.5 or 26▫C, but not at 30.5 and 23.3▫C. On the other hand, Bünsow and skeleton photoperiod experiments failed to reveal the involvement of a circadian system in this photoperiodic clock. These results suggest the photoperiodic clock in this species is a long-night measuring hourglass and the circadian effect found in the final expression of the photoperiodic response in the resonance experiments may be caused by a disturbing effect of the circadian system in unnatural regimes.