Life history patterns of Onychogomphus uncatus (Charpentier) (Anisoptera: Gomphidae)

16 Pages·1998·2.9 MB·English

by  Schütte, C

Odonatologica27(1): 71-86 March I, 1998
Lifehistory patterns of
Onychogomphusuncatus(Charpentier)
(Anisoptera: Gomphidae)*
C.Schütte,P. Schridde andF.Suhling
ZoologischesInstitut,Technische Universität Braunschweig,
FasanenstraBe 3, D-38092Braunschweig, Germany,
E-Mail: [email protected]
Received March 5, 1997/Revised and AcceptedJune 10,1997
The results ofa 4-yr study arereported. Data from Surber-samples of3 running
waters in southern France werecompared with results fromfield enclosure experi-
ments.Egg developmentlastedabout onemonth underfieldconditions andtwoweeks
at 25°C in the laboratory. O.uncatus has usually but notstrictly 13larval instars and
a3-yrlife cycle.Eachyearemergence startedin earlyJune. The emergencecurvewas
typical of asummer sp. Several aspects of seasonal regulationare discussed. The
monitoringofpopulationstructure indicated ahighmortality offinal instarlarvae (up
to99%)duringwinter. This mortalitydifferedbetween the samplesites ofoneofthe
runningwaters. Thesedifferences dependonpreyavailability and onlarval density.A
field enclosureexperiment revealed density dependenceoflarval size and growth.
INTRODUCTION
Studieson thelife-cycle ofdragonflies are numerousandthebasic principles of
larvaldevelopment arewellknown(e.g.CORBET, 1962; 1980;NORL1NG, 1984a).
CORBET(1954)dividedthe dragonflies oftemperateregions intwo groups,spring
and summer species, according totheir seasonal regulation. Themain factors in-
fluencing seasonal regulation and larvalgrowth are temperatureandphotoperiod
(CORBET, 1955; 1956; 1958;LUTZ, 1974a; 1974b;INGRAM, 1975; NORL1NG,
1984a). Larval density and the availability and type of foodalso affect growth
(HASSAN, 1976; LAWTON et ah. 1980; PICKUP et ah, 1984; BANKS &
THOMPSON, 1987).
* Dedicated tothe memory ofthe late Peter MILLER, whojoinedusatthe Canal deVergierefora
while.
72 C.Schiitte,P.Schridde &F. Suhling
In thispaperwe want to presentresultsofa fouryearstudy on the life cycle of
Onychogomphus uncatus. In particular wewanttofocus onegg development, du-
ration ofthe life cycle and theeffects of density and food availability on larval
growthandpopulation structure. Lifecycle studieshavebeen carried out for sev-
eral gomphid species. Most of these deal with egg and larval development
(POPOWA, 1923;HAWKING& NEW, 1995),larvalsizefrequency distributionin
onehabitat(HUGGINS&DUBOIS, 1982;FOLSOM &MANUEL. 1983;DUDG-
EON, 1989; HAWKING& NEW, 1996)oremergencepatterns (e.g. TESTARD,
1975; MOLLER, 1993a).As faras weknow thereare fewlong-term investigations
combining theseparameters (KURATA. 1971;AOKI, 1994;MULLER, 1995). In
O. uncatus only dataon seasonalregulationbasedonemergencestudieshavebeen
published (FERRERAS-ROMERO, 1994;FERRERAS-ROMERO & CORBET,
1995; SUHLING, 1995).
Onychogomphus uncatus is a western Mediterranean dragonfly. According to
SUHLING & MÜLLER (1996) its distributionstretches fromnorthwesternItaly
over thesouthern andwestern parts ofFrance andtheIberianpeninsula toNorth
Africa. In central Europe an isolated population exists at the Rhine near
Schaffhausen. O. uncatus is generally foundinshady mountainstreams in atypi-
calcommunity withBoyeria ireneundCordulegaster boltonii(JARRY& VIDAL,
1960;FERRERAS-ROMERO, 1984).In southernFranceitalsolivesinlargerriv-
ers and irrigation canals(SUHLING& MULLER, 1996). The larvae are able to
burrowindifferenttypesofbottomsubstratabutprefergravelly sand(SUHLING,
1996).
STUDYAREA
Thestudy wascarried out'at threerunningwaters in southern France: the Canal deVergiferes(CdV),
the CanalCentre Crau(CCC), andtheGaudre d’Aureille (GdA).The GdA isasmallmountain stream
inthe Alpilles, amountainrange ca 15km NE ofthe city ofArles (43°34'N, 4°50' E).The CdVand
the CCCaresituated in the stony steppeofthe Crau,25kmSE ofArles. The CdV borders the nature
reserve“Peau deMeau” tothe NW.TheCCC is situated 2kmSE ofthe CdV.
The runningwaters differed in annual variation oftemperatureand waterlevel (SUHLING, 1994).
WhereastheCCC usuallydries up numeroustimesforperiods of3to 4daysin March andAprilthe
CdVdoes not.The waterlevel oftheGdAwasusuallybetween 10and 30cm but increased to 200cm
duringfloods in the rainy season.The bed ofthe canals consisted ofa concrete like conglomerate
partlycovered by other sediments asmud,sand, gravelor boulder. The vegetationdiffered between
the running waters. In the CdV theemergent vegetationconsisted ofseveral species ofJuncus and
Scirpus, Menthaaquatica,and Berulaerecta, but in the CCC Polygonumamphibiawasdominant. In
both canals the stream channel was partly overgrown by submerged vegetation,particularly by
Characeae andPotamogetoncoloratus(CdV)andbyGroenlandia densaandMyriophyllumsp. (CCC).
The GdA wastotally shaded and there were noaquaticplants. Whereas at ourmain study site,the
CdV,aswell asatthe CCC the larvaecould be found inhighpopulationdensities,densitywaslowin
theGdA(SCHRIDDE & SUHLING, 1994;SUHLING, 1994).Detailed descriptions ofall study sites
are givenin REHFELDTetal. (1991)andSUHLING,(1994).
Life historyofOnychogomphusuncatus 73
METHODS
REARING OFLARVAE FROMEGGS.- Inorderto study the developmentofO. uncatus inthe
field,eggs wereexposedinthe CdV in cages, Atotal of460eggs wereobtained byholdingaplankton
net in the waterunder anovipositing female. No other commonmethod forobtainingeggs from
dragonflies,like “hand held oviposition”succeeded. The developmentof215 ofthese eggs was ob-
served in the CdV,withtheremaining245 reared in thelaboratory.
Fortheexamination ofegg- and larval developmentin the field,weusedplastic box frames(diam-
eter: 11.6 cm, height:6.7 cm)covered with gauze (0,5mm mesh size). They were attached tothe
stream bottomwith pieces ofwire mesh fixedtoboth sidesofthe box andweighteddown in the water
with flat stones. 245eggs weredistributedevenly intoeach ofthreecages. Monitoringofeggdevelop-
mentwasstartedoneweek afteroviposition. Eggswereexamined usingabinocular microscope at 50
Xmagnificationand measured toanaccuracy of0.02 mm.For larvae biggerthan 0.6 mm headwidth,
a magnificationof20 X wasused and measurements made tothe nearest 0.05 mm. Measurements
weretaken ofheadwidth abovethe eyes andlengthofthe abdomen fromthefirstsegmenttothe end
ofthe anal pyramid. Larvae in the cages were notfed additionally.In several experiments it was
confirmed that prey availability as well ascurrent velocity,temperature, and oxygen contentwere
within the naturalrange (SCHUTTE, 1992).
DETERMINATION OF INSTAR NUMBER AND SIZE. - To determine larval instars weused
headwidth frequency histograms ofthose larvae caught in the Canal de Vergiere. Foreach instara
standard distribution ofthis measurearound amean headwith should beexpected.We separatedthe
instars atthose headwidths wherethe lowestnumber oflarvaewerefound(cf. BENKE, 1970).Forthis
we usedonlylarvae from onesample site (2)because the headwidths differed between the sample
sites (see below).Fortheearly instars weused theresultsearnedfromrearingeggs and young larvae
inthefield atthe samesite. Forthoselarvae caughtinthe fieldweusedthe terms forinstarsaccording
toLutz (1968a);F=ultimate instar,F-l =penultimateinstar,etc.
MONITORINGOFPOPULATIONDEVELOPMENT INTHEFIELD.- InApril, 1990,weestab-
lished three samplesites inthe CdV (samplesites I, 2,4).In April, 1991,two additional sites (3 and
5) wereselected. The samplesites had alengthof25 m each and weredistributed alongastretch of
2.6km.The most downstream wassite 1,followed by 2 and soon.Sites 2 and4which shall be dealt
with in detail (seebelow)were500m apart.Atthese regularsamplesiteswe caughtdragonflylarvae
andotheraquaticmacroinvertebrates usingamodified SURBER-sampler(area:0.325 x0.325m).For
detailed information onthe method seeSCHRIDDE & SUHLING (1994). In 1990 thesampleswere
takenin thefirstweek ofApril,inthe secondweekof Julyandthe last weekofOctober. In 1991 we
startedin April and fromthis date wetook samplesatleast every threemonths until April 1993. All
sampledata aregivenin SUHLING (1994), The number ofsamplestaken oneach date is given in
Figure2.Usingthe numberoflarvae caughtpersamplewecalculated thelarvaldensityper m2.Atthe
twoother study sites,the CCC and the GdA,we did notsampleregularlyand used ahand netinstead
ofaSURBER-sampler.Because ofthis wedidnotcalculate larval densityforthese sites.Onall larvae
wetookin situ measurementsofheadwidth,abdominal length,and in 1990 the lengthoftheleft hind
femur. For measuringsmall larvae weused the method described above. Those larvae which were
largerthan I mm headwidth wemeasured tothe nearest 0.01 mm usingadial calliper. Aftermeasur-
ing, the larvae werereleased attheplacetheyhad been caught.
To obtain information aboutthe availabilityofprey organisms wecounted all macroinvertebrates
and fish caughtwiththe SURBER-samplerandcalculated the accumulated densityperm2.Preyavail-
abilitywascalculatedby dividingthemeandensity ofO.uncatuslarvaeby the meandensityofprey
organisms persamplesite. In feedingexperimentswefound thatlarval O.uncatusfeedonnearly any
prey theycanhandle: macroinvertebrates,planktonic species, and fry(SUHLING, 1994).Onlysome
case-bearing caddis larvae (Trichoptera) werenotconsidered as prey because O. uncatus has a low
ability toextractthem fromtheir cases.
74 C.Schiitte,P.Schridde & F.Suhling
MONITORINGOFEMERGENCE. - Annual emergencewasmonitoredinthe years 1991 to 1993
beginninginthe lastdecade ofMay and continued until noadditional exuviae werefoundduringtwo
consecutive visits.Collections weremade at samplesites2and 4attheCdV. Astandardized areaof
the banksandemergentstructuresinthe streamchannel weresearched forexuviaebetween 16.00and
17.00 CET.Allexuviae found werecollected,counted and sexed. The collection sites wereinspected
dailyuntil the first exuviae were found. Later collections weremade at intervals of4to 5 days.
Because ofthe very lightrainfall and no detectable changes in waterlevel during the emergence
period,wearesurethatthe number ofexuviaethat weredislodgedbeforecollection wasnegligible.
Forthe 1992emergence curvewecollectedonlyatsite2 andusedadditional datafromKLEEMEYER
(1994)who collected exuviae intwo dayintervals atsite 4.
FIELD EXPERIMENTON DENSITY EFFECTS.- To get information aboutdensity effects on
larval growth wecarried outan enclosure experiment in the CdV. For this weused four cages of
wooden frames withgauze(meshsize; 1.3mm).Thesecages haddifferentareas: 1.0,0.25,0.125and
0.0625 nr.Each cagewasstockedwith30F-2 larvae. The resultinglarval densitieswere30, 120,240
and 480 specimens nr2.Tomonitor growth larval headwidth wasmeasured atthebeginningofthe
experiment in October 1992,and in January,Apriland June 1993.
RESULTS
EGGDEVELOPMENTAND HATCHING
Thesizeoftheeggswas 0.54x0.40mm±0.01 mm(SDinbothmeasures, n=25,
measured2h after oviposition). In thelaboratory we observed that at25°C,after
one to several hoursthejelly aroundthemicropylar projection stuck tight toany-
thing ittouched. Inthefieldeyes andlimbsoftheembryos couldfirstbe seen24
daysafteroviposition.Thefirstlarvahatchedafter27 days,thelastoneonthe33rd
day. Out of215 eggs weobtainedonly 50larvae(23.3%), theotheroneswere not
Table I
Larval instars ofOnychogomphusuncatus: mean head width and abdominal lengthper instar and
growth ratios between the instars based on measures fromlarvae reared in field enclosures and of
thosecaughtinthe Canal deVergierein 1992. - [Forcalculatinggrowthratio theheadwidth wasused]
Instar (N) Headwidth (mm)±SD Abdominal length(mm) ±SD Growth ratio
I (50) 0.39 ±0.01 0.79 ±0,04
II (69) 0.48 + 0.01 1.00±0.08 1.22
III (36) 0.60 ±0.02 1.37 ±0.12 1.25
IV (50) 0.72 ±0.04 1.68±0.12 1.20
V (71) 0.87 ±0.06 1.97±0.21 1.21
VI (164) 1.10±0.07 2.48±0.36 1.25
VII (348) 1.35±0.09 3.06 ±0.36 1.23
VIII (465) 1.71 ±0.11 3.92±0.42 1.26
IX (415) 2.17±0.14 5.04 ±0.50 1.27
X (303) 2.79 ±0.19 6.41 ±0.64 1.29
XI (230) 3.53 ±0.17 8.12 ±0.64 1.27
XII (149) 4.36 ±0.16 10.15 ±0.66 1.23
XIII (91) 5.46 ±0.12 12.29 ±0.75 1.26
Lifehistory ofOnychogomphusimcatus 75
found again. Under laboratory conditions, 243 larvae hatched from 245 eggs
(99.2%).Onthe 12thdayeyes andlimbscouldbe seeninall larvae.One day later
thefirsthatchwas recorded.On the 15thdayafteroviposition99.9%ofthelarvae
hadhatched.Thelasthatchedon the 16thday. Hatching itselfwas observed three
times.The larvalefttheegg through alongitudinal slit. Whilethelastabdominal
segmentswere stillcaught intheslit, thefirstmoult happened. Theresulting skins
werevery thin andafterbeing cast couldnot berecognized as exuviae.Therewas
nofree prolarva.
LARVAL INSTARS
A complete viewof the instarsofO. uncatus intheCdV was obtainedby com-
bining the results ofsampling and rearing experiments (Tab. I). As a rule there
were 13instars. Immediately afterhatching themean headwidth(HW) was 0.40.
Headwidthincreasedstepwise ateachmoultso thatalllarvaecouldbeassigned to
a certain instar. IntheCdV78%of the larvaehad reachedthethird instar40 days
afterhatching, theotherswere already in thefourthinstar(HW: 0.72± 0.03mm,
n=7). Due to small numbers the fourth instar was the last that couldbe distin-
guished by rearing of larvae,butlarvae inthis instarwere also foundby meansof
sampling. They hadnearly thesamemeansize as thereared ones(0.73mm±0.04
mm, n=43).
Meanlarvalgrowthratiosbetweentheinstarsranged from 1.20(instar4) to 1.29
(instar 10). Theoverall growth ratio was 1.25. Measures taken before and after
moultsthatcouldbeobservedrevealedthattherearealso moultswithin thefixed
instar limitsand that instars can beomitted(Tab. II). Soindividualgrowth ratios
ranged atleastbetween 1.41and 1.07.
STRUCTURE OFTHELARVAL POPULATIONINTHECdV
Fortheanalysis of population development ofO. uncatus iiithe CdVwe useda
total of 5669 larvae from
TableII
SURBER samples from all
Examples for atypical growth in single larvae ofOnycho-
sites.Thuswewantedtoavoid
gomphusuncatus, indicating that instars wereomitted (A)
errors caused by local effects ormoult within the typicallimits (B, seeTab. I).Given are
onlarvaldevelopment. During the headwidth before and immediately afterthe moult and
the investigation most larval the growthratio
instarswerepresent atalmost
Headwidth [mm] Growth ratio
all sites. Those instars com- before moult after moult
binedincategory<F-6inFig-
ure 1 were found throughout A 1.35 -> 1.90 1.41
the year and were especially B 1.40 -> 1.50 1.07
B 1.65 1.85 1.12
abundant between September
76 C. Schiitte,P.Schridde& F, Suhling
Fig. I. InstarfrequencydiagramofOnychogomphusuncatusin theCanal deVergiereoveraperiodof
threeyears (densitynr2perinstarandsampledate).Theyoung instarlarvae weregrouped(<instarF-
-6). The lightly shadedbars indicate the yearclass which emerged in 1993.
andearly June.However, second instarlarvaewere only foundin smallnumbers.
First instarswere notcaught atall.Finalinstarlarvaewereless abundantduringthe
emergence period from Juneto August thanatany othertimeoftheyear.
Except forthesummermonthsJunetoAugust wefoundthree moreorlessclearly
definedpeaks inthesizeclass distributionofO. uncatus. Duringthewintermonths
wefoundmainly small larvae(< F-6) andinstars F-3and F. This distributionwas
foundbetween October and April during the whole investigation with no major
shiftofthepeaks. Frorh May on therewas a shift incommunity composition. In
Junethenumberoffinalinstarlarvaedecreasedandtherewere noneleftin July.As
aresultthedistributioncurve
shows only two peaks at that
time.
EMERGENCE
The emergence curve was
calculatedfromthe accumu-
latedexuviaeofsites2and4.
In 1990, whithoutmonitoring
Fig, Z Three year patterns of annual emergence in the wholeemergenceperiod,
Onychogomphusuncatusbasedoncounts ofexuviae fromtwo we collecteda totalof 663
stretches ofthe Canal de Vergierein 1991, 1992, and 1993.
exuviae. In the following
Day 1 is the first dayon which exuviae were found (1991):
3rd June; - (1992):5th June;- (1993):8th June. From this yearsthenumberswere 1770
date exuviaewerecollected in intervals offour days. (1991), 1725(1992), and739
Life history ofOnychogomphusuncatus 77
(1993), respectively. In these
years emergence lasted be-
tween 64and72 days(Fig. 2).
Inall yearsemergencestarted
in early June and ended be-
tween August 7th and20th. It
took between 25 and33 days
before 50%ofthe insects had
emerged (EM ). In all years
S(I
thenumberofmaleswas lower
Fig. 3. Total larval density of Onychogomphusuncatus at
than that ofthe females. The samplesite 2(upper line)and 4(lowerline)inthe Canal de
percentageofmales was 43.9 VergierefromApril 1991,toApril 1993,measured in inter-
in 1991,46.5in 1992and42.6 vals ofthree months.
in 1993,respectively.
LOCAL DIFFERENCES IN POPULATIONSTRUCTUREAND DENSITY
In theCdV therewere considerabledifferencesincommunity composition and
larval density amongthesample sites, particularly betweensites2 und4.At site2
total larval density was generally higher than it was at site 4at any time.At the
latter itneverexceeded 160larvae per m2 (Fig. 3). In contrast, theproportion of
large larvae was higher atsite 4. Here it variedbetween20 and45%ofthetotal
population whereas atsite 2 itreached a similaramount only in January (20%).
Density ofpenultimateandfinalinstarlarvaetogetherwas significantly differentat
thetwo sites inSeptember/OctoberandAprilbutmoreor less thesamein January
(Fig. 4). In January 1992,theirdensity was even higherat2 thanat4.Thenumbers
ofemerged individualswerealsosignificantly higher atsite4 duringtheinvestiga-
tion(Tab. III).
LARVALDENSITY ANDFOODAVAILABILITY
Table III Thehigh density of O. uncatus atsite 2 co-
Annual emergence of Onychogomphus incidedwithalowabundanceof potentialprey
uncatusattwo sample sites at the Canal organisms. Thisresultedinameanprey avail-
deVergiere between 1990and 1993 ability ofless thanonepreyperlarvathrough-
Year No. ofexuviae atsample site out theinvestigation period.Atsite4prey avail-
2 4 ability was much higher, particularly in July
1991when itwas almost 40timeshigherthan
1990 100 563 atsite2.In 1992thedifferencewas lowerbut
1991 350 1418 therewas still6 timesas muchpreyperlarva.
1992 59 1714
1993 86 653 Regarding allsites (cf. Methods)we found a
correlationbetweentheproportion ofbig lar-
78 C.Schutte, P.Schridde &F.Suhling
vae (last two instars) per site
and both total larval density
and prey availability. It de-
creased whentotaldensity in-
creased(Spearmanrank corre-
lation, r = - 0.825) and in-
creased with increasing prey
availability (r =0.891).
LARVALSIZE
Betweenthesitestherewere
also differences in thesize of
final instar larvae. This was
true for thethreerunning wa-
ters as well as for the sites
withintheCdV(Fig.5).Atsite
Fig, 4.(A)Abundance ofultimateandpenultimateinstarlar-
vaeofOnychogomphus uncatusat the sample sites 2 and 4 2 meanheadwidthwas lowest
in the Canal deVergiereatdifferenttimesoftheyearin 1991, andsignificantly higheratsite
1992,and 1993. The number ofspecimenper nf (+ SE), 4 with a mean difference of
calculated fromSurber samplesisgiven.U-test;NS:notsig- 0.08 mm (t-test, P < 0.01). In
nificant;*P<0.05.(B)PreyavailabilityperlarvaofO.uncatus
based on total density (see Fig 3) at the two sample sites. the GdA and at CdV 1
The datawerecalculated by dividingmacroinvertebrate den- headwidthwas higher by al-
sityby the densityofO.uncatus persamplesite and month. most0.2mm (t-test, P<0.01).
DENSITYEFFECTSINFIELD EXPERIMENTS
Theeffectsofdensity on larvalsize (seeabove) were confirmedexperimentally.
Incageswithdifferentlarval den-
sities exposed in the CdV mean
headwidthincreasedsignificantly
from July to October 1993 (Fig.
6).Almostallsurviving larvaehad
moulted at least once and had
reachedtheF-1 andF instar.Inthe
cagewiththehighest density only
threelarvae hadreachedthe final
instar whereas in the other cages
their number was at least three Fig. 5.Effect oflarval densityon the size oflast instar
times as high. During the winter larvae.The meanheadwidlhs (±SD) oflarvae caughtin
1991 and 1992 at three sample sites in the Canal de
monthsonly larvaeinthelowden-
Vergiere, onein the Canal Centre Crau and onein the
sity cageshowedaslight increase Gaudre d’Aureille aregiven.
Life historyofOnychogomphusuncatus 79
Fig. 6. Influence oflarval densityonthe growthofOnychogomphusuncatus infield enclosures. The
developmentofthe averageheadwidths (±SD) ofthe larvae (N=30percage)in eachoffour different
sizedcages from29July 1992, to9April 1993 aregiven.
inheadwidth.However, density hadnosignificant influenceonthesizeofthefinal
instar. Though individualswere bigger (HW 5.49 ± 0.09 mm SD) at the lowest
density thanwere thoseattheotherdensities(5.38±0.1 mm,5.40±0.16mm,and
5.40±0.05mm)thedifferencewas not significant (ANOVA, P>0.05). Mortality
decreased with decreasing initial density inthecage. BetweenOctober 1992and
January 199325.7% (-7)of thelarvae in thehigh density cagediedand so did
13.8%(=4), 12.0%(= 3),and7.4% (=2),respectively intheothercages(decreas-
ing density).
DISCUSSION
EGGDEVELOPMENT.- The durationofegg development ofgomphids has been
studiedinanumberofspecies (sec reviewinSUHLING& MULLER, 1996).The
minimumdurationpreviously recorded was in Asiagomphus pryeri (Sel.),which
neededninedays fromoviposition tohatch ofthelarvae (AOK1, 1994);the maxi-
mumdurationwas56 days inGomphuspulchellus Sel.(ROBERT, 1959).Therate
ofegg development increases with the ambient temperature (e.g. AOKI, 1994;
HAWKING& NEW, 1995). InO. uncatus thedurationisbetween 13and 16 days
at25°Cinthe laboratory. Inordertoestimatethedurationofthewholelifecycle it
is usefultostudy the eggdevelopment inthenaturalhabitat, as ithasbeen donein
Sympetrum danae(Sulz.) and Cordulegasterboltoniiimmaculifrons (VanclerL.)
(WARINGER, 1983;SCHUTTE, 1997).IntheCanaldeVergière O. uncatusneeds
27to33 days -twiceas long asat 25°C. Becauseoviposition usually beginsinmid
June(seebelow)the first instarlarvaecan not befoundbefore themiddleofJuly.
Thelatestoviposition was observed onSeptember20thwhenmeanwatertempera-
tures intheCdV varybetween 18°Cand 14°C.Atthesetemperaturesegg develop-
80 C.Schiitte,P.Schridde &F.Suhling
meritshouldbe muchsloweror shouldstop,as foundinGomphus f.flavipes (Charp.),
Ophiogomphus cecilia(Fourcroy),andsomeAustraliangomphids (POPOWA, 1923;
MUNCHBERG, 1932;HAWKING & NEW. 1995).
NUMBEROFLARVALINSTARSANDDURATIONOFDEVELOPMENT. - O. uncatususu-
ally has 13 instars. Forotherspecies ithasbeen shown thatthenumberofinstars
can vary. STERNBERG (1990)puts therangeof 11 to 17instarsinAeshnajuncea
(L.) downtoendogenous influenceswhichpossibly have developed as anadapta-
tionto differenthabitats. In O. uncatus there is arangefrom 12to 14instars. The
meaninterinstargrowthratioof O. uncatus is approximatly thatfoundas median
for awiderange ofhemimetabolousinsectsby COLE(1980) andthereforecon-
firms Dyar’s Rule (DYAR, 1890).Butindividualgrowthratiovaries overamuch
widerrange(seeTab. 2)becauseofinter instarmoultsoromittedinstars.
Normally, the lifecycle ofO. uncatus lasts threeyears intheCdV.This canbe
deducedfrom thefact thatinevery yearthere are always three generations atthe
sametime,eggsandimagines included.Formostoftheyearonly larvaearepresent.
Thepresenceofdifferentgenerations orage-groupsis thenreflectedinthethree-
peaked frequency distributionofthe instars. We can, however, not infer the age
fromthedistributionas single larvaemaytaketwo orfouryearsfortheirdevelop-
ment(see below). From mid-Juneonwardswhenthefirstadultsreturn totheCdV
afterthematurationperiod andstarttooviposit wecanfindfourage-groups:imago,
eggs, and twoage-groupsoflarvae. In our study this implies that mostofthose
individualsthatemerged in 1993probably hatchedin summer 1990(seeFig. 1).
Thusthelifecycle is similarindurationwiththatofmostotherEuropean gomphids
(SUHLING & MÜLLER, 1996).
SEASONAL REGULATIONANDEMERGENCEPATTERNS.- CORBET (1954)divided
thedragonflies oftemperateregions intospring and summerspecies according to
theirseasonal regulation. In spring species, such as Anax imperator Leach, the
majorityofthelarvaespend thelastwinterbeforeemergenceas final instarlarvae.
As a consequencethereisa high synchronisation inemergence.MostoftheEuro-
peanGomphus-speciesalsobelongtothiscategory(SUHLING & MÜLLER, 1996).
On theotherhand, thereis ingeneral only lowsynchronisation insummerspecies
as describedbyCORBET& CORBET(1958)forAeshnacyanea(Mull.). O. uncatus
in CdV has to be put into this category because ofthe distributionoflarvae in
January andtheemergenceobservedover several years.
InO. uncatus growth occurs mainly betweenApril andOctoberwhen themean
dailymaximumwater temperaturesexceed 15°C.During thisperiod larvaemoult
between 4 and 7 times according to theirage, as found in other gomphids by
MULLER(1995).Inthewinterperiod development is greatlyreduced.Twoinstar
XlarvaewhichhadbeenmarkedinOctober 1992wererecaptured inApril 1993in
the same instar (SUHLING, 1994). However, there are some individualswhich
also grow in winter. Thus there seems to be a facultative diapause as in Anax
imperator(CORBET, 1957).