Effect of photoperiod on the duration of summer and winter diapause in the cabbage butterfly , Pieris melete ( Lepidoptera : Pieridae )

Effect of photoperiod on the duration of summer and winter diapause was investigated in the cabbage butterfly, Pieris melete. By keeping naturally induced aestivating and hibernating pupae under various photoperiods, it was shown that diapause duration of aestivating pupae was significantly longer at long than at short daylengths, whereas diapause duration of hibernating pupae was significantly shorter at long than at short daylengths, suggesting both aestivating and hibernating pupae require opposite photoperiodic signals to promote diapause development. By transferring diapausing pupae, induced under various photoperiods, to 20°C with a naturally changing summer daylength, the diapause induced by short daylengths was easier to terminate than diapause induced by long daylengths. When naturally induced aestivating and hibernating pupae were kept under natural conditions, aestivating pupae had a long diapause (mean 155 days) and wide range of emergence (90 days), whereas hibernating pupae had a short diapause (mean 105 days) and a relatively synchronized emergence (lasted 30 days). Finally, the ecological significance of photoperiodic regulation of diapause duration is discussed. 537 * Corresponding author; e-mail: fangsen@nc.jx.cn To determine whether pre-diapause photoperiod influences diapause intensity, newly hatched larvae (from eggs produced by the spring generation in an outdoor screened insectary) were transferred to round plastic boxes (7.5 × 6 cm) containing fresh radish leaves (Raphanus sativus) and then reared under one of six photoperiodic regimes ranging from 8L : 16D to 16L : 8D at 18°C until they pupated (under these conditions diapause was induced in over 80% of the pupae). Each box contained about 25 larvae. Each treatment comprised 3 boxes. The boxes were cleaned and supplied with fresh leaves daily. On the tenth day after pupation, diapausing pupae were transferred to an incubator equipped with 3 side glass windows and kept at 20°C to terminate diapause. In this incubator the diapausing pupae were exposed to a naturally changing summer daylength. The data presented in Fig. 1B is normally distributed (Skeweness/Kurtosis test, P = 0.9217). It is important to note that the mortality of aestivating pupae was very high this year (about 80%) because of high summer and autumn temperatures, especially on 25–30 September (i.e., the 153–160 day of Fig. 1B), the daily mean temperature exceeded 25°C and the highest temperature was 30°C. Thus, a lot of adults were unable to emerge successfully from pupae on these three days. Statistical analyses were conducted using STATA Version 8.0. All data were tested using one-way analysis of variance (ANOVA). For post hoc comparisons the Bonferroni test was used (Zolman, 1993). RESULTS Diapause duration of aestivating or hibernating pupae under different photoperiods At all of the temperatures used, the duration of diapause of the aestivating pupae differed significantly under different photoperiods (F = 33.56, d.f. = 2,467, P (= 0.0000) < 0.01 at 18°C; F = 15.64, d.f. = 2,543, P (= 0.0000) < 0.01 at 20°C; F = 16.43, d.f. = 2,353, P (= 0.0000) < 0.01 at 22°C) (Table 1). The duration of diapause was shortest when the aestivating pupae were kept at a photoperiod of 12L : 12D, followed by 10L : 14D and then at 14L : 10D. At all of the temperatures used, the duration of diapause of hibernating pupae also depended on the photoperiod (F = 48.87, d.f. = 2,329, P (= 0.0000) < 0.01 at 18°C; F = 21.56, d.f. = 2,323, P (= 0.0000) < 0.01 at 20°C; F = 10.16, d.f. = 2,330, P (= 0.0000) < 0.01 at 22°C) (Table 2). However, the duration of diapause was the opposite of that recorded for aestivating pupae. It was shortest when the hibernating pupae were kept at a photoperiod of 14L : 10D, followed by exposure to 12L : 12D, and 10L : 14D. Effect of pre-diapause photoperiod on diapause


INTRODUCTION
The cabbage butterfly, Pieris melete Ménétriés is a serious pest of crucifers in the mountain areas of the Jiangxi Province, PR China.It is a multivoltine species, which undergoes summer and winter diapause as a pupa (Xue et al., 1996).The seasonal life history of the cabbage butterfly comprises two different photoperiodically induced developmental arrests: aestivation at daylengths >13 h and hibernation <12 h.At intermediate daylengths (12 h to 13 h), the butterfly does not diapause (the proportion depends on temperature).The larval instars are sensitive to photoperiod (Xue et al., 1997).In the field, the spring generation of this butterfly appears between early March and early April.The larvae hatch from early April to early May and begin to pupate in late April.Almost all of the spring generation pupae enter summer diapause.During late August to early November, adult butterflies emerge from aestivating pupae.The butterflies, which emerge in late August develop without diapause and produce three generations in autumn; those emerging before mid-October produce two generations; those emerging after mid-October, produce only one generation.Thus, there are two to four generations per year.All individuals that pupate after early November enter winter diapause.

MATERIAL AND METHODS
To obtain naturally hibernating or aestivating pupae, fullgrown larvae of Pieris melete were collected from the vegetable gardens in the suburbs of Nanchang (28°46´N, 115°50´E; at an altitude of between 120-200 m a.s.l.), Jiangxi Province, in late April and mid-November 2003, respectively.They were transferred to wooden insectaries (30 × 30 × 35 cm) to pupate under natural conditions and examined daily for pupation.The pupae, which developed in the first three days, were used in this study.These pupae were divided into two groups.Those in the first group were kept in wooden insectaries under natural conditions until the adults emerged.Those in the second groups were transferred to incubators and kept under a photoperiod of 10L : 14D, 12L : 12D or 14L : 10D at 18, 20 or 22°C until the adults emerged.These experiments of the second group were done in illuminated incubators (LRH-250-GS) equipped with four fluorescent 30 W tubes controlled by a time switch.Light intensity at the level of the pupae was 500-700 lx and the variation in temperature was ± 1°C.
To determine whether pre-diapause photoperiod influences diapause intensity, newly hatched larvae (from eggs produced by the spring generation in an outdoor screened insectary) were transferred to round plastic boxes (7.5 × 6 cm) containing fresh radish leaves (Raphanus sativus) and then reared under one of six photoperiodic regimes ranging from 8L : 16D to 16L : 8D at 18°C until they pupated (under these conditions diapause was induced in over 80% of the pupae).Each box contained about 25 larvae.Each treatment comprised 3 boxes.The boxes were cleaned and supplied with fresh leaves daily.On the tenth day after pupation, diapausing pupae were transferred to an incubator equipped with 3 side glass windows and kept at 20°C to terminate diapause.In this incubator the diapausing pupae were exposed to a naturally changing summer daylength.
The data presented in Fig. 1B is normally distributed (Skeweness/Kurtosis test, P = 0.9217).It is important to note that the mortality of aestivating pupae was very high this year (about 80%) because of high summer and autumn temperatures, especially on 25-30 September (i.e., the 153-160 day of Fig. 1B), the daily mean temperature exceeded 25°C and the highest temperature was 30°C.Thus, a lot of adults were unable to emerge successfully from pupae on these three days.
Statistical analyses were conducted using STATA Version 8.0.All data were tested using one-way analysis of variance (ANOVA).For post hoc comparisons the Bonferroni test was used (Zolman, 1993).
As all diapausing pupae terminated their diapause before 19 August, they experienced photoperiods that exceeded 13 h 50 min.During the period they were in diapause, most of the aestivating pupae experienced long days for about 5 months when daylengths were > 13 h and temperatures were high (a mean of 25.3°C), whereas most of the hibernating pupae experienced short days for about 3 months when daylengths were < 12 h 30 min and temperatures were low (a mean of 9.7°C).

Diapause duration of aestivating and hibernating pupae kept under natural conditions
When the aestivating pupae of the spring generation were kept under summer conditions, diapause lasted for 120-210 days and adult emergence extended over about 90 days, whereas when hibernating pupae of the autumn generation were kept under winter conditions, diapause lasted for 90-125 days and adult emergence was relatively synchronized within one month (Fig. 1).
The difference in the diapause duration of aestivating and hibernating pupae may be due to the naturally changing temperature and photoperiod experienced by larvae and pupae.For the aestivating pupae, diapause was induced by gradually increasing daylength in April (from 13 h 20 min to 13 h 58 min, including twilight) and relatively low temperature (a mean of 19.2°C).Most of these pupae experienced long days for a long period (> 13 h for about 5 months) and high temperatures (a mean of 25.3°C) during their diapause.For the hibernating pupae, diapause was induced by gradually decreasing daylength in November (from 11 h 50 min to 11 h 22 min) and a mean temperature of 15.1°C.While in diapause most of the hibernating pupae experienced short days for a relatively short period (< 12 h 30 min for about 3 months) and low temperatures (a mean of 9.7°C) during their diapause.These results show that photoperiod and temperature have an important influence on the maintenance and termination of both summer and winter diapause.

DISCUSSION
In some insects, the duration of diapause is clearly affected by photoperiod during diapause.The pupae of S. imparilis, reared at a photoperiod of 12L : 12D as larvae required a longer period for adult emergence than those reared at 16L : 8D (Sugiki & Masaki, 1972).In C. downesi, the rate of diapause development of diapausing adults transferred from the field (on 22 December) to one of five photoperiodic regimes ranging from 12L : 12D to 16L : 8D increased with increase in photoperiod (Tauber & Tauber, 1976).In the tailed zygaenid moth, E. westwoodii, the duration of summer diapause of aestivating prepupae kept under one of five photoperiodic regimes ranging from 12L : 12D to 16L : 8D gradually increased with increase in photoperiod (Gomi & Takeda, 1992).In the case of P. melete, however, the diapause duration of aestivating pupae was significantly shorter at short daylengths of 12 and 10 h than that at a long daylength of 14 h, whereas diapause duration of hibernating pupae was significantly shorter at a long daylength of 14 h than at short daylengths of 10 and 12 h (Table 1).This result suggests that short daylengths favour the development of aestivating pupae and long daylengths the development of hibernating pupae.This may imply that aestivation and hibernation are physiologically different, as they are induced by opposite photoperiodic signals.A very similar result is recorded for the Iberian population of P. brassicae (Spieth, 2002).When the hibernating and aestivating pupae were kept at either 15L : 9D or 10L : 14D, the hibernating pupae required longer to terminate diapause than aestivating pupae at a short photoperiod (10L : 14D); and aestivating pupae required more days than hibernating pupae to terminate diapause at a long photoperiod (15L : 9D).
That the pre-diapause photoperiod affects diapause intensity has been reported for a number of insects.In the fruit flies, D. auraria, D. subauraria and D. triauraria, the photoperiods with longer scotophases induce more intense diapause than those with shorter scotophases (Kimura, 1983(Kimura, , 1990).In the green lacewing, C. ocullata and the European corn borer, O. nubilalis, a 12 h scotophase evokes a greater intensity of diapause than either longer or shorter scotophases (Nechols et al., 1987;Beck, 1989).In the Mediterranean tiger moth C. pudica, short photophases (11 or 12 h) induce a long prepupal diapause (mean 88 days) whereas long photophases (14, 16 h) induce a short diapause (mean 52 days) (Koš ál & Hodek, 1997).In the bean bug, R. clavatus, diapausing adults kept at a photoperiod of 13L : 11D as nymphs start to oviposit earlier than those kept at shorter photoperiods (Nakamura & Numata, 2000).In the present study on P. melete, short photophases 12 h induced a short diapause  (65-69 days) at 20°C, whereas long photophases 13 h induced a long diapause (80-102 days) (Table 3).This difference in the duration of diapause induced by short and long photoperiods might be due to the long summer photoperiods (> 13 h 50 min) experienced by diapausing pupae.That is, summer conditions can shorten diapause induced by short daylengths and lengthen diapause induced by long daylengths.
Under natural conditions, aestivating pupae had a long diapause (mean 155 days) and emerged over a period of 90 days, whereas hibernating pupae had a short diapause (mean 105 days) and a relatively synchronized emergence (about 30 days).This is also reported for the regulation of diapause in the leaf-mining fly, Pegomyia bicolor.In the field, the duration of diapause of aestivating pupae (> 175 days) is significantly longer than that of hibernating pupae (< 116 days) (Xue et al., 2001).
As mentioned above, in addition to the effect of temperature, photoperiod could also affect the duration of diapause of naturally aestivating and hibernating pupae of P. melete by increasing or decreasing the initial intensity of diapause and diapause development.In nature, diapause in summer is induced by a gradually increasing daylength (from 13 h 20 min to 13 h 58 min) during the development of the larvae, which may strengthen the initial intensity of diapause.This diapause was maintained by daylengths > 13 h before 17 September (for about 5 months), which may reduce the rate of diapause development in summer.In fact, 70% of the adults emerged from aestivating pupae after 17 September (Fig. 1B).This delay in development is important ecologically as it prevents adult emerging in summer and synchronizes adult emergence with the occurrence of lower temperatures and an abundance host-plants in autumn.In fact, high temperatures are unsuitable for the reproduction of the cabbage butterfly and the growth of its host-plants.In nature, winter diapause was induced by a gradually decreasing daylength (from 11 h 50 min to 11 h 22 min) in November, which may weaken the initial intensity of diapause.After that diapause was maintained by short daylengths of < 12 h before 9 February (for about 2.5 months), which may accelerate the termination of winter diapause.Therefore, adults that emerged from winter pupae have a much shorter emergence period in spring (one month).This increase in winter development ensures the spring generation of butterflies synchronizes its reproduction and offspring development with the availability of food plants, as almost all cruciferous vegetables complete their develop-ment in spring and are harvested within two months, i.e., before mid-May.

Fig 1 .
Fig 1. Duration of diapause of hibernating (A) and aestivating pupae (B) kept under natural conditions.During the period they were in diapause, most of the aestivating pupae experienced long days for about 5 months when daylengths were > 13 h and temperatures were high (a mean of 25.3°C), whereas most of the hibernating pupae experienced short days for about 3 months when daylengths were < 12 h 30 min and temperatures were low (a mean of 9.7°C).

TABLE 2 .
Diapause duration of hibernating pupae of Pieris melete kept at 18, 20 or 22°C and under different photoperiods.

TABLE 3 .
Duration of pupal diapause in Pieris melete kept at 20°C and natural changing daylengths.Diapause was induced at six different photoperiods.