Journal of Insect Physiology 57 (2011) 1249–1258

Contents lists available at ScienceDirect

Journal of Insect Physiology

journal homepage: www.elsevier .com/ locate/ j insphys

Thermoperiodic regulation of the circadian eclosion rhythm in the flesh fly,Sarcophaga crassipalpis

Yosuke Miyazaki a,b,⇑, Shin G. Goto a, Kazuhiro Tanaka c, Osamu Saito b, Yasuhiko Watari b

a Department of Biology and Geosciences, Graduate School of Science, Osaka City University, Osaka 558-8585, Japanb Laboratory of Biotechnology, Faculty of Clinical Education, Ashiya University, Hyogo 659-8511, Japanc Ecological Laboratory, General Education, Miyagi Gakuin Women’s University, Miyagi 981-8557, Japan

a r t i c l e i n f o

Article history:Received 17 February 2011Received in revised form 19 May 2011Accepted 24 May 2011Available online 15 June 2011

Keywords:Circadian clockEclosion rhythmFlesh flyOutdoor conditionThermoperiod

0022-1910/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.jinsphys.2011.05.006

⇑ Corresponding author. Present address: LaboratorClinical Education, Ashiya University, Rokurokuso-ch8511, Japan. Tel.: +81 797 23 0661; fax: +81 797 23 1

E-mail address: ymiyazaki@ashiya-u.ac.jp (Y. Miya

a b s t r a c t

We recorded the eclosion time of the flesh fly, Sarcophaga crassipalpis, at different depths in the outdoorsoil and under temperature cycles with various amplitudes in the laboratory, to examine the timingadjustment of eclosion in response to temperature cycles and their amplitudes in the pupal stage. Inthe soil, most eclosions occurred in the late morning, which was consistent with the eclosion time underpseudo-sinusoidal temperature cycles in the laboratory. The circadian clock controlling eclosion wasreset by temperature cycles and free-ran with a period close to 24 h. This clock likely helps pupae ecloseat an optimal time even when the soil temperature does not show clear daily fluctuations. The eclosionphase of the circadian clock progressively advanced as the amplitude of the pseudo-sinusoidal tempera-ture cycle decreased. This response allows pupae located at any depth in the soil to eclose at the appro-priate time despite the depth-dependent phase delay of the temperature change. In contrast, the abrupttemperature increase in square-wave temperature cycles reset the phase of the circadian clock to theincreasing time, regardless of the temperature amplitude. The rapid temperature increase may act asthe late-morning signal for the eclosion clock.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The appropriate timing of eclosion in many insects is controlledby a circadian clock. Both light–dark cycles (photoperiod) and tem-perature cycles (thermoperiod) serve as a zeitgeber to entrain thecircadian rhythm (Saunders, 2002; Myers, 2003). In some flies,pupation occurs underground where light does not penetrate.Therefore, the photoperiod is not available as the zeitgeber forthe eclosion rhythm during the underground pupal stage. It hasbeen suggested that the circadian rhythm entrained by the photo-period during the larval stage above the ground free-runs with aperiod close to 24 h after larvae burrow into the soil and thenthe flies emerge as adults at an appropriate time of the day(Winfree, 1974; Saunders, 1976, 1979; Roberts et al., 1983; Smith,1985; Joplin and Moore, 1999). However, there are different opin-ions about the effect of temperature cycles in the pupal stage onentrainment. Roberts et al. (1983) described that pupae of thebrown blow fly, Calliphora stygia, respond to temperature cyclesand pulses in the eclosion timing but considered it unlikely thattemperature cycles in the field continue to entrain the eclosion

ll rights reserved.

y of Biotechnology, Faculty ofo 13-22, Ashiya, Hyogo 659-901.zaki).

clock during the underground pupal stage, because environmentalcues are absent beneath the substrate. In contrast, Smith (1985)suggested that in the pupae of the sheep blow fly, Luciliacuprina – which occasionally remain underground up to severalmonths – daily cycles of soil temperature compensate for slightdeviations from exactly 24 h in the free-running period and main-tain entrainment with natural daily cycles. However, this has notbeen sufficiently verified experimentally.

The onion fly, Delia antiqua, also ecloses in the soil and the cir-cadian eclosion rhythm is entrained to temperature cycles as wellas light–dark cycles (Watari, 2002a,b). However, photoperiods ap-plied in the larval stage play no role in generating the eclosionrhythm and the sensitivity to zeitgebers is found in the late pupalstage (Watari, 2005). In the field, therefore, only temperature cy-cles in the late pupal stage were considered zeitgebers. Tanakaand Watari (2003) reported an interesting property in the eclosiontiming of D. antiqua under temperature cycles. Because of the lowheat conductivity of the soil, the natural daily fluctuation of tem-perature is gradually dampened and the phase of the temperaturecycle is more delayed with increasing soil depth. Therefore, if thetiming of eclosion is associated only with soil temperaturechanges, eclosion should occur later when pupae are located dee-per in the soil. Tanaka and Watari (2003) compared eclosion timesof D. antiqua kept at different depths in the outdoor soil or kept un-der artificial thermoperiods with different amplitudes in the

1250 Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258

laboratory, and found that D. antiqua compensates for the depth-dependent phase delay of the temperature change by advancingthe eclosion time as the amplitude of the temperature cycle de-creases with an increase in depth. This novel property suggeststhe importance of temperature cycles and their amplitudes forthe eclosion timing in insects, the pupae of which are located inthe soil. Such a response, however, has not been confirmed in spe-cies other than D. antiqua.

In the present study, we focused on the eclosion of the flesh fly,Sarcophaga crassipalpis Macquart. In the field, Sarcophaga pupae areburied 4–8 cm underground (Denlinger, 1981). In laboratoryexperiments, the eclosion peaks of S. crassipalpis occur shortly afterlight-on under light–dark cycles (Yocum et al., 1994), but the eclo-sion pattern under natural conditions has not been investigated.This eclosion rhythm is governed by the circadian clock, which isentrained to light–dark cycles experienced during, at least, thewandering larval stage (Joplin and Moore, 1999). The eclosion ofS. crassipalpis may be also responsive to temperature cycles duringthe pupal stage because the responsiveness to temperature pulsesis known in pharate adults of another sarcophagid fly, S. argyros-toma (Saunders, 1979). In the present study, we recorded the eclo-sion of S. crassipalpis at different depths in the outdoor soil andunder temperature cycles with various amplitudes in the labora-tory to examine whether the pupae of S. crassipalpis adjust eclosiontiming in response to temperature cycles and their amplitudes.

2. Materials and methods

2.1. Insects

A colony of S. crassipalpis Macquart (Pape et al., 2010; alsoknown as Parasarcophaga crassipalpis, Hirashima, 1989) originatingfrom adults captured in Sapporo City (43�040N) in 2004 was kindlyprovided by Mr. Shigeki Murayama of Hokkaido University. Thecolony was maintained as described (Denlinger, 1972). Newlyemerged adults under 16-h light and 8-h darkness (LD 16:8) at25 �C were transferred to LD 12:12 at 25 �C within 2 days aftereclosion and provisioned with water, sugar, and a piece of beef li-ver (day 1). The abdomen of females was dissected on day 12 andlarvae in the uterus were collected. The larvae were placed on apiece of beef liver and maintained under LD 12:12 at 20 �C. Underthese conditions, 96.5% (N = 314) of individuals entered pupal dia-pause. Diapause pupae were transferred to continuous darkness(DD) at 7.5 �C, maintained for more than 3 months to terminatediapause, and then used in the experiments.

2.2. Recording of eclosion rhythms

The method of recording eclosion under laboratory and outdoorconditions was according to the previous studies of D. antiqua (Wa-tari, 2002a; Tanaka and Watari, 2003), but the recording apparatuswas redesigned for the bigger size of S. crassipalpis. The apparatusis made of a plastic box flanked with an infrared-light emitter and adetector (GT2; Takenaka Electronic Industrial, Kyoto, Japan) and isbased on the ‘‘falling ball’’ principle (e.g., Truman, 1972; Saunders,1976; Lankinen and Lumme, 1982). Holes in the plastic plate usedto load pupae are about 7.0 mm in diameter. When a stainless-steel ball is pushed out by an eclosing fly’s head and crosses theinfrared beam, a signal is fed to a computer and the number ofeclosions is counted.

All measurements under outdoor conditions were performed atthe experimental farm of the National Agricultural Research Centerfor Hokkaido Region (NARCH) at Sapporo (43�030N) in August 2007and June–July 2008. The recording apparatus filled with post-dia-pause pupae was completely covered with a net laundry bag to

prevent invasion of small animals like ants. The apparatus toppedwith a plastic box was buried, before dusk, at 5- or 20-cm depth inthe soil near the laboratory building, where a recording computerwas placed. A corrugated PVC sheet was covered above the groundas protection against rain. The soil temperature near the recordingapparatus was monitored by using a portable data logger (Ondo-tori, TR-71S; T&D Co., Matsumoto, Japan) at 10-min intervals. Thehourly data of solar radiation were obtained from the meteorolog-ical records of NARCH (http://ss.cryo.affrc.go.jp/seisan/meteo/dailydatae.html).

Laboratory studies were conducted to assess the importance ofdaily temperature cycle as a cue to adjust the timing of adult eclo-sion. The recording apparatuses were filled with post-diapausepupae under room light and temperature within 2–3 h after DDexposure at 7.5 �C and then were kept from 16:00 to 18:00 h inan incubator (Nippon Medical and Chemical Instruments Co. Ltd.,Tokyo, Japan) in which temperature cycles or thermoperiods canbe programmed in DD. The programmed constant temperaturewas kept within a range of ±1.0 �C.

As each experimental group of flies emerged over a 2–10-dayperiod in various thermoperiods, the daily data during the wholeemergence period were pooled, unless otherwise stated, and thephase of the eclosion peak (/E) was represented by the mediantime of eclosion. The free-running period (s) of eclosion rhythmwas estimated as the mean value of the period between eclosionpeaks.

2.3. Rhythmicity

The degree of rhythmicity in eclosion was measured by theparameter R (Winfree, 1970; Saunders, 1976; Smith, 1985). Severaldays of eclosion data were pooled to calculate the total number ofeclosions for each hour of the day. The 8-h period (gate) of the daycontaining the highest number of eclosions was then determined.The measure R was calculated by dividing the number of eclosionsoutside this 8-h period by the number of eclosions within it andmultiplying by 100. The theoretical range of R is from 0, if all eclo-sions occur within the gate, to 200, if eclosions are distributed uni-formly through the day. R values of about 150 or greater showstatistically uniform eclosion (Winfree, 1970). R values of 60 or lessrepresent rhythmic eclosion, those between 60 and 90 are weaklyrhythmic, and those greater than 90 are arrhythmic (e.g., Saunders,1979; Smith, 1985; Watari, 2005).

3. Results

3.1. Eclosion under outdoor conditions

The adult eclosion of S. crassipalpis and soil temperature wererecorded under outdoor conditions at different soil depths (5 and20 cm; Fig. 1). An investigation in August 2007 showed that theaverage soil temperatures were 26.6 and 24.6 �C at 5- and 20-cmdepths, respectively. The phase and the amplitude of the tempera-ture cycle were distinctly different depending on the depth. Thesoil temperature began to increase from 08:00 h at 5-cm depth,but from 13:00 h at 20-cm depth. The range of temperature changeaveraged 3.5 �C at 5-cm depth but was only 0.6 �C at 20-cm depth(Fig. 1A and B, right panels). Daily fluctuation of soil temperaturevaried considerably from day to day (Fig. 1A and B, left panels),depending on the solar radiation (Fig. 1C). At 5-cm depth on Au-gust 12–14, for example, the soil temperature clearly showed cyc-lic fluctuations, whereas on August 15 the fluctuations attenuated.On August 16, when the solar radiation was minor, the soil temper-ature continued to decline gradually all through the day. Clearlycyclic fluctuations of the soil temperature were again observed

Fig. 1. Outdoor soil temperature and distribution of adult eclosion from pupae of Sarcophaga crassipalpis. Field experiments were performed at 5-cm (A) and 20-cm (B) depths inthe soil in August 2007 and at 5-cm (D) and 20-cm (E) depths in the soil in June–July 2008. Each left panel shows the soil temperature and the record of adult eclosions for acertain 10-day period of the experiment period (A,B,D, and E), and also the corresponding data of solar radiation in August 2007 (C) and in June–July 2008 (F). Each right panel(A,B,D, and E) shows the data for a 5-day period indicated by a double-headed arrow in the left panel, as the mean temperature at 5- or 20-cm depth in the soil and the pooledrecord of daily eclosion. Triangles in right panels represent /E (closed) and the time when the temperature cycle began to increase (open). R values show the degree ofrhythmicity. Time is Japan Standard Time. The values of /E were 9.2, 11.1, 9.8, and 11.6 h in A, B, D, and E, respectively. Average sunrise times for a 5-day record period were04:38, 04:39, 03:57, and 03:58 h in A, B, D, and E, respectively. Average sunset times for a 5-day record period were 18:40, 18:38, 19:18, and 19:18 h in A, B, D, and E, respectively.

Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258 1251

from August 19 (Fig. 1A, left panel). The timing of adult eclosionwas synchronous but differed with soil depth (Fig. 1A and B, rightpanels). The eclosion occurred earlier at 5-cm depth than at 20-cmdepth, but it is noteworthy that the difference between the medi-ans (/E) of the 2 depths was only 1.9 h, compared with the differ-

ence of 5 h in the phase of temperature cycle between therespective soil depths.

This investigation was repeated in June–July 2008. The averagesoil temperatures were 23.5 and 20.1 �C at 5- and 20-cm depths,respectively. The soil temperature began to increase from 07:00 h

1252 Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258

at 5-cm depth, but from 12:00 h at 20-cm depth. The range of tem-perature change averaged 10.5 �C at 5-cm depth but was only1.5 �C at 20-cm depth (Fig. 1D and E, right panels). The daily solarradiation was relatively stable (Fig. 1F) as compared to an investi-gation in August 2007 (Fig. 1C), and the daily fluctuation of soiltemperature did not vary considerably from day to day (Fig. 1Dand E, left panels). The timing of adult eclosion was similar to aninvestigation in August 2007 (Fig. 1D and E, right panels). Again,the difference in /E between the 2 depth groups was only 1.8 hin spite of the 5 h difference in the phase of temperature cycle be-tween the respective soil depths.

Thus, most flies emerged in the late morning in both investiga-tions and the R values were all less than 5, which indicate highlysynchronous eclosion (Fig. 1, right panels). /E at 5-cm depth wasabout 2–3 h after the lowest phase point of the soil temperaturecycle and /E at 20-cm depth occurred at about the lowest phasepoint of the temperature cycle. This means that the flies can, tosome extent, compensate for the depth-dependent phase delay ofzeitgeber, and the compensation was achieved by advancing theeclosion timing of pupae located deeper in the soil relative to thetemperature fluctuations. In addition, this advance of the eclosiontiming may be established by a circadian clock controlling eclosionrather than as a direct response to temperature change, becauseflies eclosed at the appropriate time on August 16, 2007, despitethe continuous decline of soil temperature.

3.2. Eclosion under laboratory conditions

To examine whether the timing adjustment of eclosion of S.crassipalpis is a response to the amplitude of temperature cycles,the adult eclosion was recorded under laboratory conditions withvarious amplitudes (0, 1, 2, 4, 6, or 8 �C) of temperature cycles un-der DD. First, the adult eclosion was recorded under pseudo-sinu-soidal temperature cycles where temperature was changed every1 h (Fig. 2). When the average temperature was 20 �C, the /E valueswere 10.8, 10.0, 9.5, 7.7, and 6.1 h in amplitudes of 8, 6, 4, 2, and1 �C, respectively (Fig. 2A–E). There were statistical differences inthe timing of adult eclosion among these regimes, except betweenregimes with amplitudes of 6 and 4 �C (Steel–Dwass test; P < 0.05)./E occurred a few hours after the temperature began to increaseunder the temperature cycles whose amplitude was large, whereas/E was at the lowest temperature of the temperature cycle withthe amplitude of 1 �C. Such phase relationships between /E andthe temperature cycle were coincident with those under outdoorconditions. At a constant 20 �C, the eclosion was also rhythmic(R = 54.0), but the peak was considerably broader than under pseu-do-sinusoidal temperature cycles and the /E was 3.5 h (Fig. 2F).These results show that the eclosion rhythm of S. crassipalpis wasentrained to the pseudo-sinusoidal temperature cycle, and /E

progressively advanced as the amplitude of the temperature cycledecreased. When the average temperature was 25 �C, the pseudo-sinusoidal temperature cycle also entrained the eclosion rhythmand similarly /E advanced as the temperature amplitudedecreased, although the advance response was weaker than whenthe average temperature was 20 �C (Fig. 2G–L). Thus, we observedthat the timing adjustment of eclosion under laboratory conditionswas similar to that observed under outdoor conditions. Therefore,we conclude that these advances are caused in response to theamplitude of temperature cycles.

Second, the adult eclosion was recorded under a square-wavetemperature cycle that consisted of a warm phase and a coolphase; the temperature difference between the 2 phases was var-ied (0, 1, 2, 4, 6, or 8 �C) with the same average temperature (20 �C;Fig. 3). Under a 12-h warm phase and a 12-h cool phase (WC12:12), the /E values were close to the time of the temperature in-crease in the thermoperiod, regardless of the magnitude of temper-

ature change (Fig. 3A–E), although at a constant 20 �C the eclosionpeak was unclear because eclosion was observed in only a smallnumber of individuals during the experimental period (Fig. 3F).There was no statistical difference in the timing of eclosion amongthe regimes with the amplitude of 2 �C or more (Steel–Dwass test;P > 0.05; Fig. 3A–D). In these regimes, most flies emerged within1 h after the temperature increase. Under the 1 �C amplitude, eclo-sion was less synchronous but many individuals eclosed within 1 hafter the temperature increase (Fig. 3E). Therefore, the shift of /E inresponse to the amplitude of temperature cycles was not observedunder square-wave cycles of WC 12:12, in contrast to under pseu-do-sinusoidal temperature cycles. Under WC 6:18, the /E valueswere also close to the time of temperature increase (Fig. 3G–L).Statistical differences in the timing of eclosion were found amongsome regimes (Steel–Dwass test; P < 0.05), but there was no corre-lation with the magnitude of temperature change.

Third, to examine whether differences in the amplitude of pseu-do-sinusoidal temperature cycles affect a circadian clock control-ling eclosion or merely influence more downstream eclosionprocesses, the adult eclosion was recorded at 20 �C after pupaehad experienced pseudo-sinusoidal cycles with high (8 �C) or low(1 �C) amplitude under DD; these measurements were comparedto those of pupae continuously exposed to temperature cyclesand pupae that had not experienced temperature cycles (Fig. 4).When flies were exposed to high-amplitude temperature changeand then were transferred to 20 �C, circadian rhythmicity wasshown in adult eclosion. The free-running period (s) was 24.3 h(Fig. 4B). The /E value in the pooled record was 10.9 h, whichwas close to that (10.6 h) under continuous temperature cycles(Fig. 4A and B). There was no statistical difference in eclosion tim-ing between these 2 regimes (Wilcoxon rank sum test; Z = 0.31,P = 0.754). Flies that did not experience temperature cycles beganto emerge several days later than those experiencing temperaturecycles. The rhythmicity was weak (R = 63.3) and the /E value in thepooled record was 1.4 h, which is distinctly different from the /E inflies experiencing temperature cycles (Fig. 4C). When flies were ex-posed to low-amplitude temperature change and then were trans-ferred to 20 �C, circadian rhythmicity was shown and s was 23.2 h(Fig. 4E). The /E value in the pooled record was 6.7 h, which wasclose to that (7.2 h) under continuous temperature cycles(Fig. 4D and E). There was no statistical difference in eclosion tim-ing between these 2 regimes (Wilcoxon rank sum test; Z = 0.83,P = 0.406). In flies not experiencing temperature cycles, circadianrhythmicity was also observed and s was 24.0 h. However, the /E

value in the pooled record was 20.1 h, which is far from the /E inflies experiencing temperature cycles (Fig. 4F). These results indi-cate that differences in the amplitude of pseudo-sinusoidal tem-perature cycles affect a circadian clock controlling eclosionbecause the advance response of the eclosion timing was persis-tent even in the free-running rhythm after temperature was fixedat 20 �C.

In the above exposure to low-amplitude temperature change, atthe same time, we also recorded eclosion under antiphase condi-tions, by shifting the phase of temperature fluctuations by 12 h(Fig. 5). The peak time of eclosion was also shifted by about 12 h,compared to that in Fig. 4D and E. Under continuous temperaturecycles, the eclosion peak occurred at the lowest temperature pointand the /E value in the pooled record was 18.9 h (Fig. 5A). Whenflies were transferred to 20 �C, s was 23.4 h and the /E value inthe pooled record was 17.8 h (Fig. 5B). These R values were muchless than that in flies not experiencing temperature cycles(Fig. 4F). Thus, the circadian eclosion clock was clearly entrainedto temperature cycles with the amplitude of 1 �C.

Lastly, to examine the effects of the abrupt temperature changein square-wave temperature cycles on eclosion and on the circa-dian clock controlling it, adult eclosion was recorded at 20 �C after

Fig. 2. Various pseudo-sinusoidal temperature cycles and distribution of adult eclosion from pupae of Sarcophaga crassipalpis under continuous darkness. The averagetemperature was 20 �C (A–F) or 25 �C (G–L). The amplitudes of temperature cycles were 8 �C (A and G), 6 �C (B and H), 4 �C (C and I), 2 �C (D and J), 1 �C (E and K), or 0 �C (F andL). R values show the degree of rhythmicity. At the average temperature of 20 �C, the values of /E (triangle) were 10.8, 10.0, 9.5, 7.7, 6.1, and 3.5 h in A, B, C, D, E, and F,respectively. At the average temperature of 25 �C, the values of /E were 9.6, 9.6, 8.1, 8.1, 7.0, and 16.6 h in G, H, I, J, K, and L, respectively. The differences in /E were examinedstatistically by the Steel–Dwass test for nonparametric multiple comparison in A–E and in G–K, separately (Hochberg and Tamhane, 1987). The /E values with the same letter(a–d in A–E; e, f in G–K) are not significantly different (P > 0.05).

Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258 1253

pupae had experienced WC 12:12 with high (8 �C) or low (1 �C)amplitude under DD; this was compared to eclosion of pupaecontinuously exposed to temperature cycles and pupae not experi-encing temperature cycles (Fig. 6). When flies were exposed to

high- and low-amplitude temperature change, and not exposedto any temperature change during the experimental period, circa-dian rhythmicity was shown after transfer to a constant tempera-ture; the values of s were 24.4, 23.7, and 24.5 h, respectively

Fig. 3. Various square-wave temperature cycles (average temperature, 20 �C) and distribution of adult eclosion from pupae of Sarcophaga crassipalpis under continuousdarkness. Thermoperiod was 12 h warm phase and 12 h cool phase (WC 12:12, A–F) or 6 h warm phase and 18 h cool phase (WC 6:18, G–L). The temperature differencebetween the 2 phases was 8 �C (A and G), 6 �C (B and H), 4 �C (C and I), 2 �C (D and J), 1 �C (E and K), or 0 �C (F and L). R values show the degree of rhythmicity. Under WC 12:12,the values of /E (triangle) were 12.6, 12.5, 12.5, 12.7, and 12.0 h in A, B, C, D, and E, respectively. Under WC 6:18, the values of /E were 12.5, 11.5, 12.4, 12.7, and 12.7 h in G, H,I, J, and K, respectively. The differences in /E were examined statistically by the Steel–Dwass test for nonparametric multiple comparison in A–E and in G–K, separately(Hochberg and Tamhane, 1987). The /E values with the same letter (a, b in A–E; c–e in G–K) are not significantly different (P > 0.05).

1254 Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258

(Fig. 6B, D and E). The eclosion records showed that under square-wave temperature cycles, /E was reset close to, or just after, thetime at which temperature increased in the thermoperiod, irre-spective of the amplitude of the temperature exposure – althoughWC 12:12, with a 1 �C amplitude, was insufficient to reset the

eclosion phase in all individuals. The advance of eclosion timingthat was dependent upon the exposed temperature amplitudewas also not observed in the free-running rhythm. In the exposureto high-amplitude change, the /E value in the pooled record aftertransfer to 20 �C was significantly different from that under WC

Fig. 4. Effects of pseudo-sinusoidal temperature cycles on the circadian eclosion rhythm of Sarcophaga crassipalpis under continuous darkness. Pupae were exposed totemperature cycles with the temperature amplitude of 8 �C continuously (A); to 20 �C after transfer from temperature cycles with the temperature amplitude of 8 �C (B);to temperature cycles with the temperature amplitude of 1 �C continuously (D); to 20 �C after transfer from temperature cycles with the temperature amplitude of 1 �C (E); orto 20 �C continuously (C and F). Each left panel shows temperature and the record of adult eclosions for a certain 7-day period of the experiment. Each right panel shows thedata for the period indicated by a double-headed arrow in the left panel, as the pooled record of daily eclosion. R values show the degree of rhythmicity. The values of /E

(triangle) in right panels were 10.6, 10.9, 1.4, 7.2, 6.7, and 20.1 h in A, B, C, D, E, and F, respectively.

Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258 1255

12:12 (Wilcoxon rank sum test; Z = 4.77, P < 0.001; Fig. 6A and B),in contrast to that under pseudo-sinusoidal temperature cycles(Fig. 4A and B). This may be attributed to a burst of eclosion

directly invoked by an abrupt temperature increase. This direct re-sponse is induced to some extent by even low-amplitude temper-ature change (Fig. 6C and D).

Fig. 5. Effects of shifting the phase of the pseudo-sinusoidal temperature cycle by 12 h on the circadian eclosion rhythm of Sarcophaga crassipalpis under continuous darkness.Pupae were exposed to temperature cycles with the temperature amplitude of 1 �C continuously (A); or to 20 �C after transfer from temperature cycles with the temperatureamplitude of 1 �C (B). The phase of the temperature cycle in A was shifted by 12 h from that in Fig. 4D. The phase of the temperature cycle in B was shifted by 12 h from that inFig. 4E. Each left panel shows temperature and the record of adult eclosions for a certain 7-day period of the experiment. Each right panel shows the data for the periodindicated by a double-headed arrow in the left panel, as the pooled record of daily eclosion. R values show the degree of rhythmicity. The values of /E (triangle) in right panelswere 18.9 and 17.8 h in A and B, respectively.

1256 Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258

4. Discussion

4.1. Entrainment to temperature cycles during the pupal stage

Although the circadian rhythm of eclosion has been reported inmany insects (Saunders, 2002; Myers, 2003), there are few reportsassociated with the location where eclosion occurs. In the labora-tory and when exposed to temperature cycles that simulate soil-temperature fluctuations, D. antiqua ecloses at the same time asin the outdoor soil (Tanaka and Watari, 2003). The present resultsdemonstrated that the eclosion pattern of S. crassipalpis in the soilwas consistent with that observed in response to pseudo-sinusoi-dal temperature cycles in the laboratory. This indicates the impor-tance of temperature cycles for the eclosion timing of S.crassipalpis.

Flies are generally thought to eclose around dawn (Pittendrigh,1954; Roberts et al., 1983; Denlinger and Zdárek, 1994; Saunders,2002). Experiments under light–dark cycles in the laboratory indi-cate that S. crassipalpis ecloses around dawn (Yocum et al., 1994).The present studies of S. crassipalpis eclosion in the soil, however,showed that this fly ecloses from early morning to mid-afternoon,with a peak in the late morning rather than around dawn. How-ever, similar experiments performed in August 2000 showed thatD. antiqua eclosed from mid-night to noon, with a peak just afterdawn. Most eclosions at both 5- and 20-cm depths were completedbefore the temperature increase, in contrast to S. crassipalpis(Tanaka and Watari, 2003). These results indicate that the eclosiontime of flies in the field varies depending on the species. For exam-ple, the tsetse fly, Glossina morsitans, emerges as an adult from thesoil in mid-afternoon. This eclosion rhythm acutely responds to thethermoperiod but not the photoperiod, and the peak occurs duringthe late temperature increase (Zdárek and Denlinger, 1995).

The eclosion clock of S. crassipalpis is highly sensitive to light–dark cycles during the wandering larval stages and free-runs witha period close to 24 h in DD (Joplin and Moore, 1999). Joplin andMoore (1999) suggested that these properties of the circadianclock allow S. crassipalpis to determine precisely the eclosion phasewith respect to environmental conditions despite a pupal stage

that is spent underground, as suggested by other researchers(Winfree, 1974; Saunders, 1976, 1979; Roberts et al., 1983; Smith,1985). In the present study, when pupae of S. crassipalpis weretransferred to DD at constant temperature within 2–3 h after chill-ing – i.e., in control experiments without daily temperature cycles– the eclosion patterns were not consistent in terms of the rhyth-micity and days of adult emergence (Figs. 2F,L, 3F,L, 4C,F, 6E). Wecannot sufficiently explain what factors caused these variable pat-terns, because there are various possibilities, such as photoperiodicentrainment of the larval clock, resetting of the pupal clock byroom light for a few hours and a temperature step-up from7.5 �C before the experiments, and the difference of responsivenessto external stimuli among pupae. However, we found that temper-ature cycles applied to the pupal stage did produce eclosionentrainment. Temperature cycles not only initiated high rhythmic-ity in eclosion (e.g., Fig. 2E and F) but also reset the circadian clock(e.g., Fig. 4E and F), so that the eclosion peak occurred at the opti-mal time, even under a 1 �C temperature amplitude. We thereforeconsider that in insects that pupate underground, the eclosion timeunder natural conditions should be identified based on the respon-siveness to thermoperiods rather than to photoperiods. This islikely important for pupae that overwinter for a long time, espe-cially in areas such as Sapporo, where the winter soil temperaturedoes not fluctuate, because of snow cover, but begins to fluctuateagain in the spring, after diapause (Watari, 2005).

Daily changes in the soil temperature are varied depending onthe weather and do not show clear fluctuations when the solarradiation is low. Under outdoor conditions, the soil temperatureon August 16, 2007, continued to decline only gradually duringthe day, but the eclosion peak occurred in the late morning(Fig. 1A and B, left panels). This suggests that the circadian clockwas entrained by soil temperature fluctuations that were experi-enced before August 16, with eclosion at the appropriate timeestablished by the internal periodicity, as occurs in the laboratoryafter the transfer from temperature cycles to constant temperatureunder DD. Therefore, a free-running period close to 24 h likelyhelps pupae emerge as adults at the optimal time when the circa-dian rhythm persists without temperature cycles, rather than

Fig. 6. Effects of square-wave cycles of 12 h warm phase and 12 h cool phase on the circadian eclosion rhythm of Sarcophaga crassipalpis under continuous darkness. Pupaewere exposed to temperature cycles with the temperature difference of 8 �C continuously (A); to 20 �C after transfer from temperature cycles with the temperature differenceof 8 �C (B); to temperature cycles with the temperature difference of 1 �C continuously (C); to 20 �C after transfer from temperature cycles with the temperature difference of1 �C (D); or to 20 �C continuously (E). Each left panel shows temperature and the record of adult eclosions for a certain 7-day period of the experiment. Each right panel showsthe data for the period indicated by a double-headed arrow in the left panel, as the pooled record of daily eclosion. R values show the degree of rhythmicity. The values of /E

(triangle) in right panels were 12.5, 13.5, 14.1, 15.8, and 0.2 h in A, B, C, D, and E, respectively.

Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258 1257

photoperiodic cycles. The sensitivity of larvae to photoperiods maybe meaningful for the eclosion rhythm only when the pupal periodis short and there is no reliable temperature signal during the per-iod, such as in locations deep underground or during long periodswith cloudy days.

4.2. Effects of the amplitude of temperature cycles on the eclosiontiming

In the present study, the eclosion phase of S. crassipalpis pro-gressively advanced as the amplitude of the pseudo-sinusoidaltemperature cycle decreased and coincided with the lowest tem-perature when the amplitude was 1 �C. As reported in D. antiqua

(Tanaka and Watari, 2003), this response allows pupae located atany depth in the soil to eclose at the appropriate time of day, de-spite the depth-dependent phase delay of the temperature change.Insects other than flies that pupate underground may also have aneclosion timing that is linked to the amplitude of temperature cy-cles (Tanaka and Watari, 2003). A transfer from pseudo-sinusoidaltemperature cycles to constant temperature revealed that the eclo-sion time is set in response to the amplitude based on the phase ofa circadian clock. This tempts us to investigate, at the molecular le-vel, the responsiveness of the circadian clock of flies to the ampli-tude of temperature change.

The eclosion peak of D. antiqua occurs 3–4 h earlier in WC 12:12with a temperature difference of 1 �C than with a temperature dif-

1258 Y. Miyazaki et al. / Journal of Insect Physiology 57 (2011) 1249–1258

ference of 4 �C (Tanaka and Watari, 2003; Watari and Tanaka,2010). However, the present results did not show an alteration ofeclosion timing in response to the amplitude of the square-wavetemperature cycle in S. crassipalpis, even after transfer to constanttemperature. Although it is unclear why such a difference was ob-served between S. crassipalpis and D. antiqua, it may be attributedto the optimal eclosion time in the field. Delia antiqua eclosesaround dawn, and the peak occurs during the decline of the sinu-soidal temperature cycle and during the cool phase of thesquare-wave temperature cycle (Watari, 2002b; Tanaka andWatari, 2003; Watari and Tanaka, 2010). However, near-surfacepupae of S. crassipalpis emerge as adults in the late morning whenthe soil temperature rapidly begins to increase (see Fig. 1A and D).The rapid temperature increase may be the reliable time signal oflate morning for the circadian clock, and therefore, resets the eclo-sion rhythm to the particular phase. The rapid temperature in-crease of 2 �C or more directly invoked the extremely narroweclosion peak (Figs. 3 and 6). This high sensitivity indicates theimportance of the temperature environment for the eclosion tim-ing of S. crassipalpis.

There was a difference in the eclosion time between the pseudo-sinusoidal cycle and the square-wave cycle with the 1 �C amplitude.Under the pseudo-sinusoidal cycle, the eclosion peak occurred atthe lowest temperature point, whereas under the square-wave cy-cle, the peak occurred around the temperature increase (Figs. 2E,3E, 4D, and 6C). This difference indicates our limited ability to sim-ulate natural conditions based on a square-wave environmental cy-cle. Recent experiments on the locomotor activity of the fruit fly,Drosophila melanogaster, also demonstrated different results be-tween pseudo-sinusoidal and the square-wave temperature cycles(Yoshii et al., 2009). These results suggest the significance of apply-ing sinusoidal temperature cycles to discuss temperature entrain-ment of circadian rhythms under natural conditions.

In summary, the present study of the circadian clock underlyingeclosion of S. crassipalpis revealed the following 3 points. First, theclock is entrained to temperature cycles in the pupal stage withhigh sensitivity under both natural and laboratory conditions. Sec-ond, the eclosion phase of a clock is progressively advanced as theamplitude of the sinusoidal temperature cycle decreases, as shownin D. antiqua. Third, the abrupt temperature increase reset the eclo-sion phase of the circadian clock to the increasing time, regardlessof the amplitude of temperature cycles, unlike D. antiqua. Theseproperties of the circadian clock of S. crassipalpis would enhancethe likelihood of emerging as adults from the soil at the optimaltime of the day (late morning).

Acknowledgements

We thank Dr. Toru Shimizu for providing the computer programfor eclosion recording. We also thank Drs. Kiyomitu Ito, JunichiKaneko, and Kazuhiko Konishi for allowing us to use an experimen-tal field of NARCH. This study was supported by a Grant-in-Aid forScientific Research (C) (19570075) from the Japan Society for thePromotion of Science and, in part, by a special research Grant fromMiyagi Gakuin Women’s University (2008).

References

Denlinger, D.L., 1972. Induction and termination of pupal diapause in Sarcophaga(Diptera: Sarcophagidae). Biological Bulletin 142, 11–24.

Denlinger, D.L., 1981. The physiology of pupal diapause in flesh flies. In: Bhaskaran,G., Friedman, S., Rodriguez, J.G. (Eds.), Current Topics in Insect Endocrinologyand Nutrition. Plenum, New York, pp. 131–160.

Denlinger, D.L., Zdárek, J., 1994. Metamorphosis behavior of flies. Annual Review ofEntomology 39, 243–266.

Hirashima, Y., 1989. A Check List of Japanese Insects II. Entomological Laboratory,Faculty of Agriculture. Kyushu University and Japan Wild Life Research Center,Fukuoka.

Hochberg, Y., Tamhane, A.C., 1987. Multiple Comparison Procedures. Wiley, NewYork.

Joplin, K.H., Moore, D., 1999. Effects of environmental factors on circadianactivity in the flesh fly, Sarcophaga crassipalpis. Physiological Entomology 24,64–71.

Lankinen, P., Lumme, J., 1982. An improved apparatus for recording the eclosionrhythm in Drosophila. Drosophila Information Service 58, 161–163.

Myers, E.M., 2003. The circadian control of eclosion. Chronobiology International20, 775–794.

Pape, T., Dahlem, G., de Mello Patiu, C.A. Giroux, M., 2010. The world of flesh flies(Diptera: Sarcophagidae). http://www.zmuc.dk/entoweb/sarcoweb/sarcweb/sarc_web.htm (accessed 22.1.2011).

Pittendrigh, C.S., 1954. On temperature independence in the clock systemcontrolling emergence time in Drosophila. Proceedings of the NationalAcademy of Sciences of the United States of America 40, 1018–1029.

Roberts, R.M., Northover, J.M., Lewis, R.D., 1983. Circadian clock control of theeclosion rhythm of the brown blowfly, Calliphora stygia (Diptera: Calliphoridae).New Zealand Entomologist 4, 424–431.

Saunders, D.S., 1976. The circadian eclosion rhythm in Sarcophaga argyrostoma:some comparisons with the photoperiodic clock. Journal of ComparativePhysiology 110, 111–133.

Saunders, D.S., 1979. The circadian eclosion rhythm in Sarcophaga argyrostoma:delineation of the responsive period for entrainment. Physiological Entomology4, 263–274.

Saunders, D.S., 2002. Insect Clocks, third ed. Elsevier, Amsterdam.Smith, P.H., 1985. Responsiveness to light of the circadian clock controlling

eclosion in the blowfly, Lucilia cuprina. Physiological Entomology 10, 323–336.

Tanaka, K., Watari, Y., 2003. Adult eclosion timing of the onion fly, Delia antiqua, inresponse to daily cycles of temperature at different soil depths.Naturwissenschaften 90, 76–79.

Truman, J.W., 1972. Involvement of a photoreversible process in the circadian clockcontrolling silkmoth eclosion. Zeitschrift für Vergleichende Physiologie 76, 32–40.

Watari, Y., 2002a. Comparison of the circadian eclosion rhythm between non-diapause and diapause pupae in the onion fly, Delia antiqua. Journal of InsectPhysiology 48, 83–89.

Watari, Y., 2002b. Comparison of the circadian eclosion rhythm between non-diapause and diapause pupae in the onion fly, Delia antiqua: the effect ofthermoperiod. Journal of Insect Physiology 48, 881–886.

Watari, Y., 2005. Comparison of the circadian eclosion rhythm between non-diapause and diapause pupae in the onion fly, Delia antiqua: the change ofrhythmicity. Journal of Insect Physiology 51, 11–16.

Watari, Y., Tanaka, K., 2010. Interacting effect of thermoperiod and photoperiod onthe eclosion rhythm in the onion fly, Delia antiqua supports the two-oscillatormodel. Journal of Insect Physiology 56, 1192–1197.

Winfree, A.T., 1970. Integrated view of resetting a circadian clock. Journal ofTheoretical Biology 28, 327–374.

Winfree, A.T., 1974. Suppressing Drosophila circadian rhythm with dim light.Science 183, 970–972.

Yocum, G.D., Zdárek, J., Joplin, K.H., Lee, R.E., Smith, D.C., Manter, K.D., Denlinger,D.L., 1994. Alteration of the eclosion rhythm and eclosion behavior in the fleshfly, Sarcophaga crassipalpis, by low and high temperature stress. Journal of InsectPhysiology 40, 13–21.

Yoshii, T., Vanin, S., Costa, R., Helfrich-Förster, C., 2009. Synergic entrainment ofDrosophila’s circadian clock by light and temperature. Journal of BiologicalRhythms 24, 452–464.

Zdárek, J., Denlinger, D.L., 1995. Changes in temperature, not photoperiod, controlthe pattern of adult eclosion in the tsetse, Glossina morsitans. PhysiologicalEntomology 20, 362–366.