Thermoregulation in Sympetrum frequens (Selys), with notes on other Sympetrum species (Anisoptera: Libellulidae)

18 Pages·1998·6 MB·English

by  Ishizawa, N

Odonatologica27(3): 317-334 September1, 1998
ThermoregulationinSympetrumfrequens (Selys),
with notesonother Sympetrumspecies
(Anisoptera: Libellulidae)
N.Ishizawa
Yamaguchi 1644-15, Tokorozawa,Saitama 359-1145, Japan
Received August 11, 1997/Revised andAcceptedJanuary21, 1998
Thermoregulationin S.frequens wasinvestigatedthroughoutits adult life, from
emergencetoreproduction. Datawereanalyzedby the least squaresmethod. Body
temperature was highlycorrelated to ambient temperature in the sun in the teneral
stage andas thelife stage advanced the correlationofbody temperature to ambient
temperature waslowered.Sexual differences wereconfirmed onthermoregulation;in
the male the correlation decreased, whereas,in the female it decreased less. Males
have lower body temperatures athighambient temperatures than the females, and
whenambienttemperatureswerelow atthereproductiveperiodinautumn, theymain-
tained high body temperatures. The sexual difference ofTbranged2.5-3.4°C. This
seemtobedue tothesexualdifference inbody size.Females have smallerthoraxand
are likelytobeinfluencedby ambient temperature. Because ofthe small thorax,fe-
males seemto be tolerant ofhighambient temperature. Highbody temperature in
ovipositingmales werenotcaused by metabolic heatproductionatthe tandem flight,
butbythe elevation ofitduringcopulation.Atlow ambient temperatures maleswere
seenwarmingup at the last stage ofcopulation.Bodytemperature in non-contact
ovipositingfemales approximatedtothat ofovipositingmales in tandemand the du-
ration oftheformerwas shorterthanthatofthelatter.5.frequens,inspiteofsmallsize
and disadvantageforwarm-up, is aperiodicendotherm.
INTRODUCTION
Thermoregulation in dragonflies had been remarked in general by CORBET
(1962). MAY(1976a)was thefirsttoreportquantitative dataonbody temperature
and thermoregulatory behaviour.With regards to thermoregulation in thegenus
Sympetrum, general remarks onseveral species (ISHIZAWA, 1994, 1996;VOGT
& HEINRICH, 1983)or on specific species atthereproductive period (SINGER,
1987;TAGUCHI etal., 1995)havebeen made.But, no reports havefollowedon
thermoregulation throughouttheadultlifeofaspecificspecies.Sympetrumfrequens
318 N.Ishizawa
Selys is oneofthemost commonandfamiliardragonfliesin Japan.Thisdragonfly
is mediuminsize (abdominallength: around28mm,mass: 240mginthemaleand
280 mg in the female, thus not so large as aeshnids) and belongs to ‘perchers’
definedby CORBET (1962). This dragonfly, emerging inlowlands inearly sum-
mer, migrates tohighlands and summerstherethroughout thehotseason.Whenit
matures in autumn,itdescendsfromthehighlands andbegins to reproduce inthe
lowlands.Someofthe dragonflies, however, stay indeciduousforests or atsmall
parks in lowlandsand summer there (MIYAKAWA, 1989, 1990, 1992, 1994;
ISHIZAWA, 1994a, 1995a;UEDA, 1988,1993a, 1993b;ARA1, 1995;ASAHINA,
1990, 1991, 1992). Such a dragonfly that emerges inearly summer andrepro-
ducesin autumnUEDA(1990) definedas ‘autumnalspecies’. InthisreportIwill
showhowS. frequens thermoregulates inresponsetotheseenvironmentalchanges,
especially emphasizing itsthermoregulation atthereproductive period.
METHODS
STUDYSITES.- Fielddatawere collected;inlowlands: inteneralsatricepaddiesatAsaka(20m
a.s.l.), in matured adults atthe reproductiveperiod, atrice paddies atTokorozawa (80 m a.s.l.),
Yokoze (260ma.s.l.)andMotokaji(80ma.s.l.),Hannoata largerain puddlethathad been madeby
a typhoonandinimmature adultsatthe marginofthe deciduousforests inthe SayamaHills (150 m
a.s.l.) atIruma and the Okumusashi Hills (240m a.s.l.)atNishi-AganoatHanno, SaitamaPref.; in
highlands:in mature adults atMt Mitsumine(1150 ma.s.l.) in SaitamaPref.,MtNyugasa(1955 m
a.s.l.), Nyugasamarsh (1650m a.s.l.) andthe Yunomaru Heights (1780m a.s.l.) in Nagano Pref.
duringtheperiodofJuly 1987 toNovember 1996. These sitesare locatedonthe latitude fromaround
35°50'N to36°20'N.
FIELDTECHNIQUES.- Ageof.S',frequens wasestimated asfollows: emerged adults(tenerals,
varyinginadultage fromafewhourstoperhaps2-3days)weredistinguishedbytheir relativelyweak
flutteringflight, glossy wings and their softerexoskeleton, and in migratingteneralsthe agewas
estimatedby adding4 daysformigrationsofar toTokorozawa(20km per dayfromrice paddies in
Chiba) to2-3 days ofrestingat emergence sites. Adultsat Nishi-Aganowere regardedpopulations
which had migratedatthe sametime when theoneatTokorozawa hadbeenseen.Inthepopulationat
theYunomaruHeights,itwas regardedabout twoweeks old,emergedinlate Julyatthefootofthe
heights. The footis around700 m abovesealevel, andin such ahighaltitudeS.frequens emerges
late,aboutin late July.
Body temperatures (Tb) were measured to 0.1°C with a Hoskins F-V-K-002 thermocouple
(chromel-alumel,diameter,0.05mm,the samewithotherprobes)setin ahypodermic needle,which
was insulated with paper covering, connected toaJohn-Fluke 55 double channeled digitalthermo-
meter.The measurements weremadewithinabout 10secondsafternettingthe dragonflies,byinsert-
ing the probethroughthemesh ofthe net tothe center ofthe thorax, 1 mm up the mesothoracic
spiracle,because body temperatures weredifferent between mesothorax and metathorax.
Both mesothoracic and metathoracic temperatures were measuredasfollows: thelocation ofthe
probewasnotvariedandthe samespecimenwasprobedtwice, firstinthe centerofthemetathorax
andnextofthemesothorax. The formerwashigherthanthelatterby 0.6°C inthemaleand 1,2°C in
the female (p<0.01,respectively). Ambient temperaturesatthe heightof 1 m in the sun (Ta)were
measuredwithanOmronHC-100 thermo-clocksoonafter Tbweremeasured.The thermistorprobeis
3mm indiameterandtheresponses ofit toTa arerather slow, soTawereassumed tobe those when
Tbweremeasured.
Thermoregulationin Sympetrumfrequens 319
Bodytemperaturesofthe dragonfliesadopting obeliskpostures weremeasuredontheindividuals
soonafteradoptingtheposture and those thatcontinued the postureforoneminute.
Bodytemperaturesofpairsduringcopulationweremeasuredafterthelapse ofoneminutefromthe
start of copulationandthose ofovipositingpairs weremeasured after oneminute’s,three minutes’
and five minutes’ lapsefromthebeginningofoviposition,respectively.
Inthesepairs measurementsweremade onfemales priortomales.Asbodytemperaturesofpairs
less than I min.after start ofovipositionrangedwidely,these datawereomittedbecause Icould not
easilydistinguishthosepairscomingtoovipositionsitestoovipositafter copulationfromthose soon
after arrival there fromtheir roostingsites. Duration ofcopulationandovipositionweremeasuredat
rice paddiesatTokorozawa,Iruma,HannoandYokoze.Ovipositingpairsandfemales wererecorded
with a SONY CCV 90 8 mm videocamera. Ovipositionrhythm was determined by calculating
frequenciesand durations ofovipositiononaTV display.
INDOOREXPERIMENTS.- Minimum flighttemperatureatwhichlevel flightcanbe sustained
was determined by tossingcooled S.frequensinthe airby thesamemethodofMAY(1976).
Bodytemperatures ofdragonflieswarming up endothermicallywere investigatedasfollows: a
thermocouplewas implantedin the center ofthe thorax ofspecimenswhich were notrestrained. I
used the video cameratorecord continuouslythe output ofthethermocouplethermometer.
Thermoregulationby haemolymphcirculation were investigatedas follows: specimens werere-
strainedonawoodenplatewith their wingsputtogetherby aclothespinwhichwashungdownfrom
anarmofasupport,with oneprobeimplantedinthe abdomen andanother inthe centerofthe thorax
throughapinholetothe depthof2mm;the thorax washeated I cmaway fromthe side with aheater
(National BG-112-P,2.88 W;itsinternal diameter:7.9 mm)thatwasconnected to 3VAC/DC con-
verter. The abdominal temperature and Tbofthe specimenswere measuredat30 seconds intervals.
These experiments weremadeindoors atthe temperature of19-20°C.
Rapidrising ofTbin femalesejectingeggs wereobservedatanexperimentasfollows: specimens
werefixedattheprothoraxwithinsect-pinsonastyrofoamplate,with aprobeimplantedinthecenter
ofthethorax;wingsand thoraxwerecoveredwithaslipofKentpaper, upperside ofwhich waspasted
with aluminum foil; specimenswerelighted20cm away fromabove witha75 Whalogenlampand
Tb weremeasuredat30seconds intervals.
RESULTSAND DISCUSSION
MINIMUMFLIGHTTEMPERATURE
AfemaleS. frequens madealevel flight atTbof 15°C.Someofthespecimens
couldfly below15°C. Dr. M.L.May(pers. comm., 1993)estimatedtheminimum
flight temperature(MET)ofS. vicinumat 15.4°C. VOGT & HEINRICH (1983),
however,reportedthatMETofthesamespecies is 12.5°C,lowerthanMay’s esti-
mation. In thefield,TbofamaleS.parvulum thatmadealevelflightwas 15.9°C
atTa 14.5°C. These suggest MET ofthe species in genus Sympetrum is near at
15°C.
WARMING-UPANDVOLUNTARY ACTIVETEMPERATURE
Dragonflies cantakelevelflightatMET, butthey needhigherTbfortheirvolun-
tary activities in the daytime.S. frequens sometimeswarm upwith their wings
shivering atlowTa. In my experiments, they continuedto do ituntil Tb rose up
320 N. Ishizawa
near to 25°C (Fig. 1). At
Amaike pond (2060ma.s.l.)
atMtYatsugatake, Iobserved
S. frequens thatwere kept in
thecoolshade(Ta; 14°C)and
were put on the shaded side
ofaboulder, a halfofwhich
was shaded by me. All of
themput intheshadewarmed
upforawhileandflewaway.
Ontheotherhand,S.frequens
Fig. 1. Thoracic temperatureduringwing-whirringina 6 S.
frequens.Ta= 19°C. put on the sunny side never
warmed up and they stayed
therestillforawhileandflewaway.Thesurfacetemperatureonthesunny sidewas
27°C.MAY(1976b)reported thatheatlossincreasesmorerapidlywith decreasing
body size, andthat loss is sorapid thatmetabolismcannot fully compensate.As
warm-upcostsmuchenergy, itmay be disadvantageous forS.frequens thathasa
rathersmall thorax (around 140 mg). 5. frequens seems, however, to do it for
maintaining thenecessary Tb (voluntary active temperature) for activities in the
daytime.Then, Tbatwhichitstops warm-upis presumed tobethelowerlimitof
voluntary activerange.On theotherhand,theupperlimitoftherangeisTbwhere
the dragonflybegins tothermoregulate notto rise so far inspiteofrising Ta. InS.
frequens, itmaybeestimatednear36-37°C (shown inFig. 2). These suggestthat
the voluntaryactive rangeinS.frequens may beestimated 25-37°C.
TENERALS
S.frequens emergeatrice
paddies in lowlandson lulls
in the rainy season in late
June. Tenerals soon after
emergence were sighted
perching low in the grass at
irrigated fallow paddies at
Asaka (ISHIZAWA, 1994b,
1995a). When I drove some
ofthem,they flewaroundfor
a whileandagain landedon
thegrasses atthe samearea.
TAKETO (1992) also re-
ported on this behaviour. At Fig. 2. Relation ofTbtoTa in teneral S.frequenssoon after
non-irrigated fallowpaddies emergence and in migratingtenerals restingin the lowlands.
Thermoregulationin Sympetrumfrequens 321
ambienttemperaturesatthesameheight were higher,andnoneofS.frequens was
seen. They seemedtostay thereforafew days(TAKETO, 1992).The meanbody
temperatureofmaleswas 35.5±1.7°C,andthatoffemaleswas 33.5±1.8°C (n=15).
MaleTb was higher than Ta by 8.1°C,andfemale Tb was 6.4°C higher,and as
MARDEN etal.(1996) noted, theseTb’swere relatively lowerthanthose ofma-
ture adults(except maturesin thelowlands inmid-summer).The regression coef-
ficient offemale tenerals soon after emergence was the largest (2.08, r=0.77,
p<0.001) amongthestagesinS.frequens inlowlands(Tab. I).Thatformaleteneral
S.frequens was not availablebecauseofthescarcity ofdata(n=4),butthescatter-
ing of maledatainFig. 2 suggests thatmalecorrelationcoefficientmaybeabout
thesameas thatoftheteneralfemale.MAY(1976a) reportedthatteneralPachydiplax
hadbeen foundin deeply shadedareasandthermoregulated poorly andtendedto
havelower Tbthanadults andthiswas dueto theincompletely hardenedcuticle.
Table 1
Statisticscalculated fromleast squares linearregression ofTbon Tain each stageoffree-rangingS.
frequens;- [al=slope; r=correlation coefficient;p=probability;S.D.= standarddeviation]
Average ExpectedTbat:
Stage Sex N al r P Tb S.D. 20°C 25°C 30°C
Soon after emergence 6 4 - - - 35.5 1.7 - - _
in the lowlands 9 15 2.08 0.77 <0.001 33.5 1.8 - 25.0 39.6
During migration S 11 1.34 0.72 <0.1 36,4 1.2 - 28.2 35.0
inthelowlands 9 10 1.32 0.85 <0.005 36.8 0.9 - 28.6 35.2
In thedaytime 6 8 -0.17 0.10 - 33.0 1.8 - 33.9 33.1
in thelowlands 9 9 1.69 0.84 <0,005 33.8 1.6 - 25.0 33.4
Soon afteradopting 6 7 0.19 0.20 <0.8 38.9 0.7 - 37.2 38.1
obelisk posture 9 8 -0.00 0.00 - 39.3 1.0 - 39.3 39,3
1 min.afteradopting S 7 2.58 0.82 <0.05 38.7 l.l - - 30.4
obeliskposture 9 5 0.66 0.83 <0.1 38.3 0.8 - - 35,9
Early morning 6 12 2.88 0.90 <0.001 26.5 3.7 27.3 - -
inthehighlands 9 10 1.87 0.83 <0.005 25,0 1,9 26.5 - -
In the daytime <J 55 1.08 0.59 <0.001 32.1 3.6 27.7 33.1 38.5
in thehighlands 9 55 1.08 0.58 <0.001 31.8 3.9 27.5 32.9 38.3
Perchingatrepro- 6 21 0,11 0.16 <0.5 34.7 1.7 34.6 35.2 35.7
ductiveperiod 9 6 0.35 0.59 <0,5 33.8 0.9 33.6 35.4 37.1
1 min.after the 6 20 0.43 0.49 <0.05 34.4 1.7 34.1 36.2 38.4
start ofcopulation 9 20 0.84 0.53 <0.02 31.0 2.8 30.3 34.1 38.0
Tandem oviposition c? 21 0.70 0.84 <0.001 36.3 1.4 35.4 38.9 42.3
1 min. after 9 21 0.70 0.59 <0.01 33.8 2.0 32.9 36.4 39.9
3min. after <J 22 0.57 0.61 <0.005 36.0 1.5 35.8 38.7 41.5
9 22 0.87 0.65 <0.001 32.9 2.1 32.6 36.9 41.2
5min. after S 21 0.36 0.56 <0.01 36.3 1.4 36.1 37.9 39.6
9 21 0.85 0.77 <0.001 32.9 2.5 32.5 36.7 40.9
Non-contact
oviposition 9 10 0.54 0.92 <0.001 35.8 2.7 34.4 37.1 39.8
322 N.Ishizawa
UEDA (1988, 1993a, 1993b) also noted the thin cuticle ofteneral S. frequens,
whichisrelatively thinnerthan that ofLestes sponsa,a species thatdoes notmi-
grateto highlands and undergoes prereproductive diapause in summer. The pro-
pensity ofteneralS.frequens tostay atirrigatedareasandmaintainrelativelylower
Tbis probablyrelatedtothethinandincompletely hardenedcuticleandtheconse-
quentdanger ofexcessive water loss.
MIGRATINGS.FREQUENS
S.frequens were foundresting atdeciduousforest inthe Sayama Hills during
theirmigration to highlandson July 2, 1994.They seemedaboutoneweekold. It
was very sultryandTaoftenexceeded33°C.ThemeanTbofmaleswas36.4±1.2°C
(n=l 1) andoffemaleswas 36.8±0.9°C(n=10), higher thanTaby 5.3°Cand5.6°C
inthemalesandfemales,respectively. Theregression coefficientsoftheequations
onbothsexes inmigrating dragonflies wererather smallerthanthoseoftenerals
(Tab.I, Fig. 2).
Inmigrating5.frequens,Tbwas not so high aswouldbepredicted inteneralsat
similarly high Ta,andwas apparently less variablethan in newly emerged speci-
mens. MostofTbwas controlledbelow37°C atTa31°C. Probably with thelapse
oftimethecuticleofteneralS.frequens was hardenedenough for migration and
so theinsectscouldtoleratehighTa.Mostofmigratory swarmswereobservedon
fineorhotdays(MATSUMURA, 1984;HASEGAWA,I991;ARAI, I992;ISHIZA-
WA, 1995c). HighTaseems torelease themigration.
ADULTSINLOWLANDINTHE
HOTDAYTIME
Most teneral S. frequens
migratetohighlands,but some
seemto stay in lowlandsand
summer therein thehot sea-
son.MIYAKAWA(1989,1991,
1992, 1994) was the first to
havemadeprecise reports on
S.frequens summering inthe
lowlandoftheSayama Hills.
These S. frequens, most of
whichwere female,were seen
perchinginthecanopiesoftrees
in thedeciduousforests, and
Fig. 3. Relation ofTbto Ta in immature S.frequensin the theywere inactive duringthe
lowlands in mid-summer. hot daytime.Theirbody tern-
Thermoregulationin Sympetrumfrequens 323
peraturewas maintainedneartheambienttemperatureinthesunflecks(ISHIZAWA,
1994a).
At Nishi-Agano, Hanno,inlowlands,S.frequens, seemingly morethantwomonth
old, were perching atthemarginofadeciduousforest.ThemeanTbof maleswas
33.0±1.8°C (n=8),thatoffemaleswas 33.8±1.6°C (n=9), higherthanTaby 2.4°C
inthe maleand3.6°C in the female,respectively. Inthemale, especially, there-
gression coefficient was less than zero andthe correlationcoefficientwas 0.10
(Tab. I, Fig. 3). It suggested thatthe relation ofTb to Tawas not significantly
correlated,andasTaincreased, theirTbwas ratherlowered. Ontheotherhand,in
thefemalethecorrelationwashighly significant andTbwasalittlehigherthanthat
in the male. Both correlationandregression coefficients in the femalewere far
largerthanthoseinthemale.They didnot seemtoregulateTb,butprobably thisis
due to the scarce data. According to the data in the Sayama Hills(ISHIZAWA,
1994a) correlationcoefficientin adults was smaller:0.45, and regression coeffi-
cient was 0.26. According to MAY (1976a), dragonflies in tropical zone donot
thermoregulate well. In lowlandsTais high enough forS.frequens tobe active,
andtheyonly needtotake carenot tooverheatthemselves.So, they stay attheedge
offorestsorattheopenspaces thereforthemtotakerefugeeasily fromthehighTa
atany time.
ADULTSADOPTINGOBELISKPOSTURES
Atthe sameplace wherethedata abovewere collected, the Tb ofindividuals
adopting obelisk postures
were measured during late
July and early August 1996.
TheseS.frequens also seem-
edmorethanonemonthold.
The mean Tb ofindividuals
soonafteradoption ofobelisk
postures was 38.9±0.7°C
(n=7), 39.3±1.0°C (n=8), in
themaleandinthefemale,re-
spectively. The correlation
coefficientssoon after adop-
tionofobeliskpostures were
strikingly small (Tab. I, Fig.
4,). Butthe meanTbofindi-
viduals that continued the
postures forone minutewas
38.7±1.1°C (n=7), 38.3± Fig.4. Relation ofTb to Ta in S.frequensadopting obelisk
0.8°C (n=5)in the maleand postures.
324 N.Ishizawa
thefemale,respectively. InthemaletheeffectoftheposturetodecreaseTbseemed
small,andinthefemaleitseemedalittlelarger (the differencewas 1.0°C)than in
themale(0.4°C).
CORBET (1962)describedthat perchers tendtopoint theabdomentowardsthe
sun athighertemperatures.Thispostureis saidto minimizeexposure ofthebody
surfaceto thedirectrays ofthesun. MAY(1976a)reportedthatthestrictassocia-
tionoftheobelisk withhigh Tb demonstratedathermoregulatory functionofthe
posture, andwhichwereeffectivein preventing or slowing furtherrise inTb.Ac-
cording tohisdata,inPachydiplax obeliskpostureswere adopted atTbhigherthan
36°C.InS. frequens adopting thispostureinthefield,themeanTbwashigherthan
36°C,andthis coincideswithMAY’Sdata.Tb atwhichS.frequens begins toadopt
obeliskposture variesaccording tothedegreeofmaturity(ISHIZAWA, 1996)and
also varieswithsexes. Teneralsinhighlandsoftenadopted theposture atTblower
than those in lowlands, and malestendto adopt the posture at lowerTbthanfe-
males.
IMMATURE ADULTSIN HIGHLANDS
S.frequens attheYunomaruHeights seemedmorethantwoweeksold, andmost
ofthemspenttimeperching on thegrassesinthemeadowinthehighland. Early in
themorningthey were seenperching onthegrassesor onthetwigsoflarches atthe
marginofthemeadowwiththeirbodiesperpendicular to thesunlight.
As shown in Fig. 5 & 6, Tb of5. frequens in highlands seemed to be highly
correlatedwith Ta, and particularly early in themorning between6:00and 7:00
a.m. the body temperaturesof S.frequens seemed to rise faster, andregression
coefficients were far larger. This may be due to theprobability that S. frequens
were eager toattain quickly
highTbforthemtobeactive.
No differencesofcorrelation
andregression coefficientsin
the daytime were observed
between both sexes (Tab. I).
Their Tbin the daytime was
maintainedattherangefrom
25 to 38°C atTafrom 22 to
26°C; the mean Tb: 32.1±
3.6°C (n=55) in the male,
31.8±3.9°C(n=55) in the fe-
male, respectively, relatively
high compared with those in
Fig.5. Changesoftemperatures ofimmature S.frequensat
lowlandsathigherTa. As Ta
theYunomaruHeights inthe daytime.Five specimens were
collected ofeach sex, atintervals ofonehour. inhighlands is not so higher
ThermoregulationinSympetrumfrequens 325
than in lowlands, once their
Tbiselevated upto thepref-
erablerange,they seemtobe
able to keep their Tb within
thefreerangeofthevoluntary
activetemperature.
TSUBUKI (1987) studied
the flight activity of S.
frequens inrelationtothe en-
vironment factors at
YunomaruHeights, thesame
area where my observation
was done, and reported the
dragonfly was mostactive in
theairattheambienttempera-
ture, 18-24°C in the shade.
UEDA (1988, 1993a) esti- Fig. 6. Relation ofTb to Ta inimmature S.frequensat the
matedthe upper limitofthe Yunomaru Heightsin mid-summer.
mostsuitableambienttemperaturein theshadeforthe dragonfly at23°C,judging
fromtheverticaldistributionofS.frequens insummer.Astheambienttemperature
intheshadeis usuallyby2-3°ClowerthanTa,Taof25°C,whereS.frequens begins
tothermoregulateTbat36°C,is consistentwith theirconclusions.
MATUREADULTS ATTHEREPRODUCTIVEPERIOD
PERCHINGADULTS.-During thereproductive period inautumn, mostS.frequens
cametooviposit intandemtoharvested rice paddies, but somemalescamethere
about 9:00 a.m.,beforepairs arrived, andperched hereand there atridges be-
tween rice paddies. They occasionally flew over the rice paddies, dipping their
abdomens in the puddles, in the same way as females oviposited. These males
rarely succeededincopulating, becausefewerunmatedfemalesappeared there.
Figure 7 shows therelationofTbofperching andcopulating S.frequens toTa.
The mean body temperatureofperching males (34.7±1.7°C, n=21) was not so
higherthan thatoffemales(33.8±0.9°C, n=6);thesexualdifferenceofTb (0.9°C)
wasinsignificant. Regression andcorrelationcoefficientsweresmallerinbothsexes
perching inreproductive period thanintheprevious stagesexcept inadopting ob-
eliskposture; inthemalethedecrease was larger thaninthefemale(Tab. I). This
suggests thatinreproductive period males maintainhigh Tbindependent fromTa
andthis may beusefulformales tocopulate withfemalesandtooviposit in tan-
dem. According to the report by TAGUCHI & WATANABE (1995), Tb ofS.
darwinianum searching forfemaleswasashigh asthatofovipositing males.Prob-
ably thispropensity ofmalesforhigh Tb is not limitedtomales ofthesespecies
326 N.Ishizawa
and may be general in
Odonata thatoviposit in tan-
dem.
COPULATION.- Thehourof
appearance oftandem pairs
varied with the day. In late
summer they appeared early
in the morning, about 7;00
a.m. Whenambienttempera-
ture was lowin midautumn,
thefirstappearancewasrather
late, about9:30 a.m. andlate
autumn they appeared about
10:00a.m.Ambienttempera-
tures seem to affecttheirap-
pearance. As soon as S.
Fig. 7. RelationofTb toTa in perching and copulating S. frequens cameoverintandem
frequens. totherice paddies from their
roosting sites, theyflewover thepaddies toandfrounskillfully. Thisflight,unlike
the rhythmical oviposition, seemed as ifthe males ofthe pairs were checking
whetherthe oviposition sites were preferable fortheirpartners to lay eggs.This
behaviourcan betermed “trialoviposition”. Consequently, whenpairs were first
observed at oviposition sites I could not easily distinguish the trial oviposition
from thereal oviposition. This trialoviposition continuedfor aboutone minute,
andthenmalesinpairstranslocatedspermintheairandsoonlandedontheground
near the oviposition sites or flew away from thereto copulate. They landed in
depressions indead grasses or onseemingly warmer spots, such as sunlitfences,
logs orridgesbetweenpaddies. Thesepairs mighthaveselectedthespotstoavoid
predators, becausethey were very sensitiveto disturbance,andoftenchanged the
perching spots. Someofthemchanged thespots forwarmer places.
Figure 8 shows the relationofdurationofcopulation to ambient temperature.
Themeandurationwas 882.9±401.9sec atTa24.0±1.4°C(n= 18).Thecorrelation
coefficientofthedurationtoTais only marginally significant(r=0.44,p<0.1), but
this seems tosuggestthat asTariseshigher, thedurationis shortened.According
toMICHIELS(1992), inS. danaecopulation shortenedwith increasing tempera-
ture: early morning copulations lastedlonger than midday copulations. Consider-
ing thatTain themorning is usually lowerthan in midday, it seems naturalthat
copulations lastedlongerintheearlymorning. However,heattributedthistosocial
factors.
Figure7 shows therelationofTbto Taofpairs copulating foroneminute.The
meanTbofthepairswas 34.4±0.4°C (malesn=20)and31.0±0.6°C(femalesn=20)
atTa20.9±1.9°C, in themaleand in thefemale,respectively. The sexual differ-