Photoperiodic clock of diapause termination in Pseudopidorus fasciata (Lepidoptera: Zygaenidae)

Photoperiodic control of diapause termination was systematically investigated in Pseudopidorus fasciata. In 24 h lightdark cycles, the rate of diapause termination in this species depended on photoperiod. The critical night length (CNL) for diapause termination was 10 h, 0.5 h shorter than that for diapause induction. Night-interruption experiments with T = 24 showed that dia­ pause was effectively terminated when the scotophases separated by light pulse were shorter than the critical night length (10 h); no developing individuals were found if the duration of the pre-interruption scotophase or the post-interruption scotophase exceeded the CNL. A 15-min light pulse was sufficient to reverse the effect of long night when it was placed 8 h after lights-off. Resonance experiments with a constant photophase of 12 h or 16 h and various scotophases of 4-80 h showed an hourglass-type photoperiodic response, where no rhythmicity was found. In another resonance experiment with constant scotophase of 8 h and various photo­ phases of 4-72 h, all individuals developed into cocoons. In the Bunsow experiment, the response curve showed two apparent peaks for diapause termination, one being 8 h after lights-off, and another 8 h before lights-on. However, there was no periodic rhythmicity, which again indicates an hourglass principle. The results lead to the conclusion that the same photoperiodic clock mechanism (a long-night measuring hourglass) is involved in both diapause induction and termination.


INTRODUCTION
Of all the physical factors that change seasonally, pho toperiod is the most regular and reliable cue to seasons.Insects have evolved numerous ways to utilize photope riod as a diapause-regulating cue (Tauber et al., 1986).So far, photoperiodic control of diapause induction has been examined in a large number of insect species.On the other hand, termination of diapause by photoperiod was reported only in about forty species (see Beck, 1980;Saunders, 1982;Koveos et al., 1993;Claret & Arpagaus, 1994).
If an insect relies upon photoperiodic information for terminating its diapause, one might logically assume that the clock system underlying termination of diapause is the same mechanism as the one involved in induction of dia pause.Thus, the methodology used for analyzing photo periodic induction of diapause is also relevant for the study of diapause termination.Brunnarius and Dumortier (1984) were the first to apply the same technique in an experiment on diapause termination in Pieris brassicae.They found that in both diapause induction and termina tion, it was actually the scotophase length that was being measured.The same conclusion was also found in Tetranychus urticae (Koveos et al., 1993).
Induction of larval diapause, an overwintering dor mancy, is a photoperiodically controlled event in the life cycle of the lepidopteran, Pseudopidorus fasciata (Xue & Kallernborn, 1998).Photoperiodic clock for diapause induction was systematically investigated by Wei et al. (2001) using various light-dark (LD) cycles, nightinterruption experiments, resonance experiments and Bunsow protocols.In the present study, these methods were used to assess the photoperiodic time measurement for diapause termination in P.fasciata.

MATERIAL AND METHODS
The population of P.fasciata used in this study was collected in the suburbs of Nanchang (28°46'N, 116°50'E), Jiangxi Prov ince, P. R. China.Full-grown larvae prior to cocooning were collected in the field.The larvae were allowed to form cocoons and to emerge under natural conditions.Eggs used in all experi ments were obtained from females reared in the insectarium.When larvae newly hatched, they were transferred to a diapause-inducing regime (12L : 12D and 25°C) and kept there for 15 days.Under this condition all individuals enter a larval diapause.Then, all diapausing larvae were transferred to various LD cycles to test termination of diapause.The criterion of dia pause termination was the forming of a cocoon; and the duration of diapause included the period of post-diapause larval develop ment.The larvae were fed with fresh foliage of Chinese sweetleaf, Symplocos chinensis (a deciduous shrub).At least 50 larvae were used for each treatment unless otherwise noted.Rearing conditions have been described elsewhere (Wei et al., 2001).

Photoperiodic response curves under 24-h light-dark cycles
Diapausing larvae were maintained at a daily mean temperature of 26.7°C in an insectarium and a constant temperature of 26 ± 1°C in the incubators (LRH-250-G) with the photoperiods ranging between 2L : 22D and 22L : 2D as well as continuous light (LL) and continuous darkness (DD).Fig. 1 shows that diapause termination in this species is controlled by photoperiod.Night lengths of 4 to 8 h (i.e. from 20L : 4D to 16L : 8D) terminated dia pause.However, a low level of diapause termination occurred in LL and a scotophase of 2 h (33.3% and 36.3% in Fig. 1A; 29.0% and 31.4% in Fig. 1B).Night lengths of more than 10 h maintained larvae in diapause.The critical night length (CNL) was 10 h.

Night-interruption in 24-h photoperiod
The night of a 9L : 15D (a most effective diapausemaintaining photoperiod) was systematically illuminated by a single 1-h pulse at 26°C and the percentage of dia pause termination was observed (Fig. 2).The result showed that the light pulses placed 7-10 h after the onset of darkness caused a complete reversal of the response.Both the lengths of pre-interruption scotophase and post interruption scotophase were shorter than the CNL.When the light pulses were placed 2-4 h and 12-14 h in the darkness, very few individuals developed.Here, the dura tions of the pre-interruption or post-interruption scotophases exceeded the critical day length.Fig. 2 also shows that the light pulses commencing 5-6 h after the onset of darkness induced development in much smaller per centage of individuals (26.1% and 62.9%) than did pulses commencing 10-11 h (43.1% and 100%), although the lengths of their scotophases are equal (10 h and 9 h).This may suggest that the late phase in the dark period is more photosensitive than the early phase.

Minimal light requirement for night interruption
To find the shortest night interruption required to termi nate diapause, light pulses ranging from 5 to 60 min were given during 8 h in the scotophase of LD 9:15 (the most sensitive part to light in diapause induction).An illumina tion of 15 min was sufficient to reverse the effect of long night, inducing development in 93% individuals (Fig. 3).

Resonance experiments (or Nanda-Hamner experiments)
In the first resonance experiment, the scotophase was serially extended from 4 to 80 h with an interval of 4 h.The unchanged photophase was 12 h (Fig. 4A) or 16 h (Fig. 4B-C).The results showed that photoperiodic responses depended on whether the night lengths exceeded the CNL or not.Development occurred at LD 12:4, 12:8, 12:10, 16:4, 16:8 and 16:10.Night lengths longer than 10 h did not result in the termination of dia pause except few individuals in some treatments.No rhythmicity was found in the photoperiodic response.
When a constant scotophase of 8 h was combined with various photophases of 4-72 h, all light regimes provoked  termination of diapause, although the process of diapause termination in the short photophases of 4 h and 8 h took longer than that in the other photophases (Fig. 5).The required mean short nights to terminate diapause at this resonance experiment varies from 27 to 51 days at 26°C (Fig. 6).The result indicates that a period of 8 h of dark ness had been measured and that measurement can take place regardless of the period of the light cycle.
These results are consistent with what is expected from an hourglass principle.

Bunsow experiments
In the Bunsow experiments, a constant photophase of 12h was combined with a scotophase of 36, 48 or 60 h, which were scanned by 1-h light pulses at 4-h intervals.The experiments were performed at a daily mean tem perature of 27.5°C in an insectarium (Fig. 7).The results showed that only the light pulse placed 8 h after the onset of darkness terminated diapause most effectively.The light pulse placed 8 h before dawn produced a relatively small peak of diapause termination, whereas others had little effect on diapause termination.

DISCUSSION
As compared with the photoperiodic response curve for diapause induction in P. fasciata (Wei et al., 2001), the response curve for diapause termination is just a reversal within the photoperiodic regimes ranging between 20L : 4D and DD at 26°C.Namely, short days induced diapause and long days terminated diapause.However, some differences occurred.Very long photo phases of 22 h and LL induced 31.4% and 29.0% dia pause termination, respectively (Fig. 1), whereas they completely prevented diapause induction.The same kind

Length of photophase (h)
of opposite response was found in Pieris brassicae (Brunnarius & Dumortier, 1984).The curves were also slightly different in the critical night length, shorter in diapause termination by about 0.5 h (10.5 h for diapause induction).However, the critical night length for both diapause induction and termination appears to be the same under natural conditions.According to field obser vations for many years, diapause rate of 50% is reached in early September at day length of about 13 h 32 min (including twilight) (Xue & Kallenborn, 1998), and ter mination of hibernation occurred around April 10 at day length of about 13 h 31 min.The differences between laboratory and field conditions may be caused by tem perature conditions, because diapause termination of overwintering larvae in the field experienced a very long period of low winter temperatures.
Comparisons of the critical daylengths for diapause induction and termination have been presented for a number of insects.A longer critical daylength for dia pause termination than for diapause induction was found for Lygus hesperus (Beards & Strong, 1966), Pyrrhocoris apterus (Saunders, 1983), P. brassicae (Brunnarius & Dumortier, 1984), Diatraea grandiosella (Takeda, 1985), and Ostrinia nubilalis (Skopik & Takeda, 1986).The effect of night interruptions on diapause inhibition has been investigated for a number of insect species (see Saunders, 1982;Koveos et al., 1993), but night interrup tion experiments on diapause termination have been done only in a few species.The first study was carried out by Hayes et al. (1970) on the oak silkworm Antheraea pernyi and the codling moth Laspeyresia pomonella.However, no clear effects of light pulses were observed.In a further study, Hayes et al. (1974) applied the same technique to four lepidopteran species under natural light and tempera ture conditions.They also did not provide any reliable evidence of the role of a light pulse in diapause termina tion in these species.Afterward, the effect of night inter ruption on diapause termination was demonstrated in five species.Brunnarius and Dumortier (1984) found one point of long day effect for diapause termination in P. brassicae, but two points (A and B) in diapause induction.Skopik et al. (1986) found bimodal peaks (A and B) of long day effect in experiments on diapause ter mination in O. nubilalis, but one broad "trough" in experiments on diapause induction.Takeda (1985) and Koveos et al. (1993) found that the response curve to light pulses in diapause termination was a mirror image of that in diapause induction in the south-western corn borer D. grandiosella and the spider mite Tetranychus urticae, respectively.Claret & Arpagaus (1994) found two points of maximal long-day effect for diapause termination in Pimpla instigator.The feature of response to night inter ruption in P. fasciata is similar to that in D. grandiosella and T. urticae.A light pulse of 1-h at 9L : 15D averted diapause most effectively when it was placed 7-10 h after lights-off (Wei et al., 2001), because the scotophases separated by these light pulses did not exceed CNL of 10.5 h.The same result was obtained in diapause termina tion, in which light pulses placed 7-10 h after lights-off completely broke diapause (Fig. 2).The results obtained from night interruption experiments strongly suggest that the same clock mechanism is involved in diapause induc tion and termination in P. fasciata.It confirms that the most crucial event for the photoperiodic time measure ment in this moth is whether the duration of the scoto phase exceeds the CNL or not.
Length requirement for night interruption in diapause induction has been investigated in several insects (Beck, 1962;Barker, 1963;Takeda, 1985), but not studied in diapause termination.In P. fasciata a 10-min light pulse falling 8 h in the darkness prevented diapause in over 50% of the individuals (Wei et al., 2001) whereas the same length of light pulse terminated diapause only in 18.8% (Fig. 3).This may suggest that the length require ment of light pulse in diapause termination is longer than that in diapause induction.In Fig. 3 a 15-min light pulse induced development in 93.7% of the individuals, indi cating that diapause termination in this species is also very photosensitive.
Resonance experiments for diapause induction have now been conducted with a number of insects and three mites (see Vaz Nunes & Saunders, 1999), whereas reso nance experiments for diapause termination have been conducted only in five species, Plodia interpunctella (Takeda & Masaki, 1976), P. brassicae (Claret, 1985), O. nubilalis (Skopik & Takeda, 1986), Pimpla instigator (Claret & Arpagaus, 1994) and the mite T. urticae (Veerman & Koveos, 1989).The first four species showed a negative resonance effect (i.e.no periodic response occurred).T. urticae showed a positive reso nance effect (i.e.response curve displayed periodic peaks).In addition, in two species (P.brassicae [Claret, 1985] and O. nubilalis [Skopik & Takeda, 1987]), dia pause termination appeared to be controlled by an hour glass, whereas induction had a circadian involvement.In the Russian strain of T. urticae, a circadian influence was found to be present on both induction and termination of diapause (Veerman & Koveos, 1989).In P. fasciata, the resonance experiments for diapause induction showed a weak circadian periodicity at temperatures of 24.5 and 26°C, but not at 30.5 and 23.3°C (Wei et al., 2001).In P. fasciata, however, the resonance experiments for dia pause termination with a constant photophase of 12 h or 16 h (Fig. 4) showed that the response curves depended completely on the measurement of CNL; all scotophases longer than the CNL effectively kept larvae in diapause.In the reverse experiment with constant scotophase (Fig. 5, 6), all photoperiods terminated diapause.No perio dicity was found.It is very clear that the photoperiodic time measurement for diapause termination in this moth is apparently accomplished according to an hourglass prin ciple.More recently, Veerman (2001) declares that the expression of a circadian rhythm found in Nanda-Hamner experiment does not mean that this rhythm reflects the operation of the photoperiodic clock.The rhythm may be the expression of a circadian disturbance, possibly by internal desynchronization in the unnatural regimes used in Nanda-Hamner experiments.He concludes that photo periodic time measurement in insects and mites is per formed by a non-circadian "hourglass" clock.
So far, Bunsow experiment for diapause termination was done only in the Indian meal moth, Plodia interpunc tella (Takeda & Masaki, 1976).It showed a negative cir cadian effect.In P. fasciata, the Bunsow experiments showed that the photoperiodic response curves for dia pause termination had two apparent peaks: First major peak occurred 8 h after lights-off, second minor peak occurred 8 h before lights-on.However, there was no rhythmicity thereafter and the photoperiodic time meas urement may function on the hourglass principle.Such two peaks locked to 8 h after dusk and 8 h before dawn were also found in diapause induction for the codling moth, Carpocapsa pomonella (Hamner, 1969) and the aphid, Megoura viciae (Lees, 1973).
All experimental results from both diapause induction and termination lead to the conclusion that the same pho toperiodic clock mechanism (a long-night measuring hourglass) is involved in both diapause induction and ter mination in P.fasciata.

Fig. 1
Fig. 1 Photoperiodic response curves for the termination of diapause in P. fasciata under 24 h light-dark cycles at a daily mean temperature of 26.7°C (A) and a constant temperature of 26°C (B).

Fig. 2
Fig. 2 Night interruption for diapause termination of P. fasciata in a 9L : 15D, 26°C regime.The scotophase was systematically scanned by 1-h light pulse.

Fig. 3
Fig. 3 Length requirement of the night interruption to terminate diapause of P. fasciata.A photoperiod of 9L : 15D was inter rupted by a light pulse of different length commencing 8h after the onset of darkness at 26 °C.

Fig. 4
Fig. 4 Diapause termination of P. fasciata at a daily mean temperature of 27.1°C (A), 29.2°C (B) and a constant tem perature of 26°C (C) in resonance experiments with constant photophase of 12 h (A) or 16 h (B, C) and a scotophase which was varied from 4 h to 80 h, in steps of 4 h.

Fig. 6
Fig. 6 Average number of short nights required for 100% dia pause termination in resonance experiments with 8h of darkness per cycle and various periods of light at 26°C.

Fig. 5
Fig.5Diapause termination of P. fasciata in a resonance experiment with a constant scotophase of 8 h and a photophase which was varied from 4h to 72h, in steps of 4h.The number of larvae that had terminated diapause, expressed as the per centage of total larvae present, was determined on the 25th day (A), 28th day (B) and 51st day (C) after diapausing larvae were transferred to the resonance experiments.

Fig. 7 .
Fig. 7. Diapause termination of P. fasciata in Bünsow experiments at a daily mean temperature of 27.5°C with a constant photo phase of 12h and an extended scotophase of 36 h (A), 48 h (B) or 60h (C) which was systematically scanned bya 1-h light pulse.