Circadian Clock in InsectPhotoperiodism

Abstract. Night-interrluption experi-mnents with the wasp Nasonia vitripennismnaintained in 48- and 72-hour light:dark cycles reveal two and three peaksof light sensitivity (diapause inhibi-tion), respectively. The first peak is19 houirs after lights-on in all regimestested; the later peaks occlur at 24-hourintervals thereafter, providing evidencefor the circadian nature of the photo-periodic clock in this insect.

Some of the strongest evidence forthe circadian nature of the seasonalphotoperiodic "clock" comes fromnight-interruption experiments (1) inwhich organisms are maintained in48- or 72-hour cycles that contain ashort light component (8 to 10 hours)and a long dark period systematicallyinterrupted by a supplementary lightpulse. In plants (2) and birds (3) suchexperiments often reveal maximums ofphotoperiodic effect at roughly 24-hourintervals in the extended night. In theBiloxi variety of soybean maintainedin a light: dark cycle (LD) of 8 : 64,for example, flowering was inhibited atZeitgeber time (Zt) 16, Zt 40, and Zt64 (4, 5). These results are usually in-terpreted as showing that a light-sensi-tive phase repeats itself in the extendednight with circadian frequency.

Similar experiments have been per-formed in at least two insect species-with different results. For instance, inthe codling moth Carpocapsa pomoniella(6), maintained in an interrupted cycleof LD 8: 40, maximum inhibition of

Fig. 1. Effect of light breaks (night in-terruption) on the inhibition of diapauseproduction by females of Nasoniia vitripen-ni.v in 48-hour light: dark cycles. The toppanel shows 2-hour light breaks in anLD 12: 36 cycle, the lower panel shows1-hour light breaks in LD 14: 34 (dashedline) and 2-hour light breaks in LD10: 38 (solid line). The horizontal dottedline in the top panel shows the response toan uninterrupted cycle of LD 12 : 36.The first peak of inhibition is at Zt 19 inall regimens, the second peak 24 hourslater. Each point represents the mean ageat the "switch" from developing to dia-pause larvae for about 40 females.

Fig. 2. Effect of 2-hour light breaks in anLD 12: 60 cycle. Three maximums ofphotoperiodic effect are produced, thefirst at Zt 19 and the later peaks at 24-hour intervals. Each point represents themean age at the "switch" for about 40females.1 MAY 1970

Zeitgeber time, (hours)601

on J

uly

8, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

on J

uly

8, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

on J

uly

8, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fro

m

PhilippeComentrio do textoInterrupo noturna

PhilippeComentrio do textomantidos

PhilippeComentrio do textorevelam

PhilippeComentrio do textodepois

PhilippeComentrio do textoluzes acesas

PhilippeComentrio do textoprovado

PhilippeComentrio do textoMais tarde

PhilippeComentrio do textoem

PhilippeComentrio do textodepois disso

PhilippeComentrio do textofornecendo

PhilippeComentrio do textoalguns

PhilippeComentrio do textomais forte

PhilippeComentrio do textochegar

PhilippeComentrio do textomantidos

PhilippeComentrio do textoeste

PhilippeComentrio do textosistematicamente

PhilippeComentrio do textointerrompeu

PhilippeComentrio do textotais

PhilippeComentrio do textofrequentemente

PhilippeComentrio do textoaproximadamente

PhilippeComentrio do textoestendido

PhilippeComentrio do texto uma cidade localizada no litoral do estado do Mississippi.

PhilippeComentrio do textosoja

PhilippeComentrio do textose

PhilippeComentrio do textosido

PhilippeComentrio do textolagartas

larval diapause was found at Zt 16and Zt 40. On the other hand, in theaphid Megoura viciae (7), at LD8: 64, there was only one point oflong-day effect at Zt 16 but no repeti-tion 24 or 48 hours later.

I now report the results of similarexperiments with the parasitic waspNasonia vitripennis, a long-day speciesin which the photoperiod acting uponthe mother controls the onset of larvaldiapause (8). These experiments dem-onstrate a well-defined circadian rhyth-micity in response to the light pulses.Newly emerged females of N. vitri-

pennis were enclosed in small (5 by 1.2cm) glass tubes with a male and threeto five 3-day-old pupae of their fleshflyhost Sarcophaga barbata. Groups of

40 of such tubes were placedKilner jars over a saturatedof NaNO3 to maintain thehumidity within the jars at -percent; the Kilner jars weplaced on different light: dark18C (9). Groups of femaexposed to LD 10: 14, LD 12LD 14: 10 for four cycles atransferred to the appropriatc72-hour cycle. This procediadopted so that the photcrhythm, if present, would be eto the 24-hour light: dark cycibeing "released" into the lonwmens with "weaker" signals. Tsitized pupae of S. barbata wereby fresh pupae only when tiwere on, that is, every 48 or 7

I.1 gAg

Zeitgeber time (hours)Fig. 3. Schematic drawing of the effect of light breaks (night interruption) in Ipennis, showing the relation between the peaks of diapause inhibition obtainecand 72-hour cycles with the two peaks (A and B) obtained in each night of acycle (11). (A) One-hour light breaks in LD 12: 12 producing two peaks, c15 to 16 hours after lights-on, and a second (B) 15 to 16 hours before li(B-E) Effect of light breaks in LD 10: 38, LD 12: 36, LD 12: 60, and LDThe first peak in all of these regimens is 19 hours after lights-on and is consihbe equivalent to peak A in a 24-hour cycle. Later peaks occur at circadian ithereafter. L, long-day effects; 5, short-day effects.

602

in glass They were then kept at 25C for 10solution days and opened to ascertain whetherrelative the parasites within were diapause

about 70 or developing larvae.Dre then In a 24-hour periodic environment,cycles at females of N. vitripennis respond toles were the effects of photoperiod by "switch->: 12, or ing" to the production of diapauseand then larvae after a number of light: dark48- or cycles; short-day effects result in a

ure was rapid switch (after 8 to 11 days), andDperiodic long-day effects result in a delayedmntrained switch (after 20 days or more). Atle before 180C, short-day conditions produceger regi- about 65 to 75 percent of diapause inhe para- the progeny and long-day conditionsreplaced less than 5 percent (8). In the presenthe lights experiments results were calculated as72 hours. the number of calendar days to the

switch so that they are comparable toL earlier results for this species (8).

In the first experiment groups of40 females of N. vitripennis were ex-posed to LD 12: 12 for 4 days andthen transferred to 48-hour cycles each

S with a 12-hour main light componentand a 2-hour supplementary light pulseplaced at different positions in thedark (Fig. 1). Maximum sensitivity tolight (producing long-day effects) wasobserved when the supplementary lightperiods fell at Zt 19 and at Zt 43.

' A maximum short-day effect was ob-tained when the pulses fell betweenZt 22 and 28. In other words, the twopeaks of diapause inhibition were 24

........... hours apart, and the greatest short-dayeffect was observed when the pulse

...............

. was placed in a position where thelights would have come on had the in-sects been maintained in a 24-hourregimen (that is, at the beginning of the"subjective day"). Light pulses fallingin the second half of the subjectiveday (Zt 30 to 36) produced an effectindistinguishable from the uninter-rupted control (LD 12 : 36).

In a similar experiment with a light:dark cycle of LD 12 : 60, three peaksof diapause inhibition were observed(Fig. 2). The first was at Zt 19, thesecond at Zt 43, and the third was afteranother 24-hour period at Zt 67. Lightpulses falling between these peaks-during each "subjective day" (Zt 23 to36 and Zt 48 to 60)-produced short-day effects.

N. vitri- In both of these experiments thei in 48- peaks of diapause inhibition appeared24-hour at identical positions in the night andne (A) occurred at 24-hour intervals. The ex-ights-off. periment using a 72-hour cycle is par-dered to ticularly good evidence for the involve-intervals ment of the circadian system in photo-

periodism and is the first to show clear-SCIENCE, VOL. 168

*cAd

I I,e.

PhilippeComentrio do textoafdio

PhilippeComentrio do textoencontrado

PhilippeComentrio do textomo

PhilippeComentrio do textoh

PhilippeComentrio do textofoi

PhilippeComentrio do textoSomente

PhilippeComentrio do textomais tarde

PhilippeComentrio do textorelatrio

PhilippeComentrio do textocomeo

PhilippeComentrio do textoestes

PhilippeComentrio do textorecentemente

PhilippeComentrio do textovidro

PhilippeComentrio do textoseus

PhilippeComentrio do textomosca varejeira

PhilippeComentrio do textocolocado

PhilippeComentrio do textomanter

PhilippeComentrio do textovidro

PhilippeComentrio do textodentro

ly three peaks of light sensitivity in aninsect system. In Megoura viciae (7)there was no evidence of a circadianrhythm, and in Carpocapsa pomonellamaintained in an LD 8: 64 cycle therewas evidence of a complex timing sys-tem that involved two conceptually dif-ferent clocks: an hour-glass timer as-sociated with "dusk" and a 24-hourrhythm associated with "dawn" (10).The 72-hour experiment also removesthe objection that each peak representsthe result of a direct interaction be-tween the pulse and the main light com-ponent. In other words, the middlepeak is too far removed from eitherthe preceding or following light pe-riods to be caused by a direct inter-action; the nearest effective light signalsare those at the other peaks, both 24hours away.

Since it was possible that the ap-parent 24-hour rhythm was generatedin some way by the choice of light:dark cycles, which both included a 12-hour main component, two furtherexperijnents were carried out with 48-hour cycles that contained a 10-hourand a 14-hour main light component,respectively (Fig. 1). In these experi-ments the females were exposed to LD10: 14 and LD 14: 10 for four cyclesbefore being released into the longerregimens. Both of these experiments re-vealed maximums of photoperiodic ef-fect at Zt 19 and, in the LD 10: 38 ex-periment, a second maximum at Zt 43.This showed that the positions of thepeaks were unaffected by the durationof the main light component. It is notknown why the amplitude of the firstpeak is greater in the experiment withthe 14-hour main light component.

In a 24-hour light: dark cycle twopeaks of diapause inhibition are ob-served in N. vitripennis (11) and otherinsects (12) when the night is scannedby light interruptions. In these insects thefirst peak (A) is always a certain num-ber of hours after lights-on ("dawn")and the second peak (B) always thesame number of hours before lights-off (dusk). These observations formthe basis of the coincidence model forthe photoperiodic clock (13). Thismodel is essentially a rhythm of sensi-tivity to light which is itself phase-set bythe light: dark cycle, and has the sameproperties as overt circadian rhythmssuch as pupal eclosion in Drosophila oroviposition in the moth Pectinophoragossypiella (13, 14). The crux of thismodel is the dual action of light: (i) asan entraining agent and (ii) as an in-ducer when the light coincides with the1 MAY 1970

photoperiodically inducible phase. Ac-cording to Pittendrigh (14), pulses oflight placed at both A and B act as en-training agents, but only one (at eitherA or B) coincides with the induciblephase.

In the present experiments the firstmaximum of photoperiodic effect hadthe same relationship to the beginningof the main light component in each ofthe regimens tested. For this reason itmust be equivalent to a peak A in a24-hour cycle (Fig. 3). The later peaksmust therefore represent this phase ofthe rhythm repeating itself in the nightwith circadian frequency. In N. vitri-pennis maintained in 24-hour night in-terruption experiments, peak A occurs15 to 16 hours after lights-on (11).The fact that it is 19 hours after lights-on in 48- and 72-hour cycles must in-dicate that the photoperiodic rhythmassumes a different phase relationshipwith a 48-hour or 72-hour Zeitgeber.The last peak in the night is not con-sidered to be equivalent to peak B be-cause it is 15 hours before lights-offin the LD 10: 38 cycle and 17 hoursbefore lights-off in the LD 12: 36 andLD 12 : 60 cycles.

In none of the experiments was thereany indication of two peaks of diapauseinhibition in each "subjective night."This was so even in the LD 14: 34 ex-periment in which the first 10 hours ofthe night were scanned by 1-hourpulses, and the two peaks, if present,would be expected to be at their mostdivergent (Fig. 1). It appears, there-fore, that peak B or its equivalent doesnot appear in 48- or 72-hour cycles. Al-though these experiments are conse-quently difficult to interpret in termsof the coincidence model, they do pro-

Terrestrial locomotion and orienta-tion of fishes have been investigated intidewater Fundulus (1) and in Bathy-gobius soporator (2). Bathygobius sop-orator use local landmarks for ori-

vide strong evidence for the circadiannature of the seasonal photoperiodicclock.

D. S. SAUNDERSDepartmentment of Zoology, Universityof Edinburgh, Edinburgh, Scotland

References and Notes

1. In night-interruption experiments, a shortlight break (usually 1 to 2 hours, or aslittle as a few minutes) is introduced intothe dark component of the cycle and re-peated as a daily signal. Different experi-mental groups receive light regimes with theinterruption in a different position in thenight.

2. H. Claes and A. Lang, Z. Naturforsch. 2b,56 (1947); R. Bunsow, Z. Bot. 41, 257 (1953);G. Melchers, Z. Naturforsch. l1b, 544 (1956);E. Bunning, Cold Spring Harbor Symp.Quant. Biol. 25, 249 (1960).

3. W. M. Hamner, in Circadian Clocks, J.Aschoff, Ed. (North-Holland, Amsterdam,1965), p. 379.

4. The Zeitgeber is the forcing oscillation (thatis, the light: dark cycle) which entrains abiological rhythm. Zeitgeber time (Zt) ismeasured in hours after the beginning ofthe main light component.

5. K. C. Hamner, Cold Spring Harbor Symnp.Quant. Biol. 25, 269 (1960).

6. D. M. Peterson and W. M. Hamner, J. InsectPhysiol. 14, 519 (1968).

7. A. D. Lees, Nature 210, 986 (1966).8. D. S. Saunders, J. Insect Physiol. 12, 569

(1966).9. These experiments were conducted in Gallen-

kamp cooled incubators fitted with Philips8W striplights controlled by Londex timeswitches.

10. W. M. Hamner, J. Insect Physiol. 15, 1499(1969).

11. D. S. Saunders, ibid. 14, 433 (1968); Symp.Soc. Exp. Biol. 23, 301 (1969).

12. P. L. Adkisson, Amer. Natur. 98, 357 (1964);Science 154, 234 (1966); R. J. Barker, C. F.Cohen, A. Mayer, ibid. 145, 1195 (1964);N. I. Goryshin and V. P. Tyshchenko, inPhotoperiodic Adaptations in Insects andAcari, A. S. Danilevskii, Ed. (LeningradUniv. Press, Leningrad, 1968), p. 192.

13. C. S. Pittendrigh and D. H. Minis, Amer.Natur. 98, 261 (1964); D. H. Minis, inCircadian Clocks, J. Aschoff, Ed. (North-Holland, Amsterdam, 1965), p. 333.

14. C. S. Pittendrigh, Z. Pflanzenphysiol. 54, 275(1966).

15. I thank Mrs. M. H. Downie and Miss H.Graham for technical assistance and theScience Research Council for financial sup-port.

25 November 1969; revised 3 February 1970 N

entation and do not rely on sun cues;however, the cues used for orientationby Fundulus were not understood. Themosquito fish, Gambusia affinis (3),and several species of amphibians (4,

603

Terrestrial and Aquatic Orientation in theStarhead Topminnow, Fundulus Notti

Abstract. Starhead topminnows from various shores of a small woodland pondwere displaced to unfamiliar surroundings, and their orientation was tested inaquatic and terrestrial arenas. These fish used a sun compass to move in a direc-tion which, at the location of their capture, would have returned them to theland-water interface. The fish accomplished directional terrestrial locomotion byusing the position of the sun to align its body for each jump. On heavily overcastdays many fish were unable to orient their bodies in a consistent direction from,jump to jump; this inability to orient resulted in random rather than linear move-ment. There was considerable individual variation in terrestrial locomotor ability.

PhilippeComentrio do textomeio