Sperm precedence in experimental interspecific multiple matings of hybridizing North American tiger,swallowtail butterfly species (Lepidoptera : Papilionidae)

2006·6.5 MB·

ou
Volume 60, Number 2 v^SN" 'V/^/l/V 65
'
JournaloftheLepidopterists'Society
60(2),2006,65-78
'igfiARlfc*
SPERM PRECEDENCE IN EXPERIMENTALlNTCRSPECIFIC MULTIPLE MATINGS OF
HYBRIDIZING NORTH AMERICAN TIGER SWALLOWTAIL BUTTERFLY SPECIES (LEPIDOPTERA:
PAPILIONIDAE)
Aram D. Stump°
DepartmentofEntomology, MichiganStateUniversity, EastLansing,MI48824USA
AND
Mark Scriber
J.
DepartmentofEntomology, MichiganStateUniversity,EastLansing, MI48824USA,,email:[email protected]
ABSTRACT. PapiliocanadensisandP.glaucushavemultiply-matingfemalesandmales.Laboratoryhand-pairingsallowedustoby-passnat-
uralmatingpreferencesofpolvandrousfemalesandconductsequential(twice-mated)labpairings;oneheterospecificandoneconspecificmale
for each virgin female ofP. glaucus or P. canadensis. Using electrophoresis and species-specific (diagnostic) allozymes, we were able to
determinewhetherrandomfertilization(amixofheterospecificsperm)orspermprecedence(eitherfirstorlastmale)existedforthesetwice-
matedfemales. Sincethenumberofusefuldoublepairingsforlabandfieldfemaleanalyseswaslow,wecanonlyconcludethatofdoublemated
females,mostarefertilizedbvmalesoftheirfirstmating,butthatamixofspermandlastmalespermprecedencedooccuratlowerfrequen-
cies.
Overall,theselabresultssuggestthatifitexists,spermselectionbyfemalesof P. glaucusandP. canadensiswouldnotnecessarilypresidea
clearmechanismofreproductiveisolation. Infact,eggsfrominterspecificprimaryhybridlabcrosses(73familiesofPgxPcand 17familiesof
PcxPghybrids;allsinglepairingF crosses)areasfertileandviableasthosefromparentaltypes.Theisolatingmechanismsthatmaintaineda
narrow hybrid zone from 1980-19l97 have not been sustained during the recent 1998-2003 period where extensive interspecific hvbrid
introgressionandX-chromosome recombination has occurred. Itisclearthateventhough lab-rearedmaleswere regularlyfedasolutionof
honeywater, aminoacids, andelectrolytes,theywerelesssuccessful atproducinglarvaloffspringthatwerefield-captured(wild) maleswhen
matedtovirginfemales.Thisdifferenceinreproductivesuccessmandatescautioninextrapolationfromlabtofieldstudies.Howeverlab-reared
males can show first male orlast male sperm precedence. It is also possible for a female to have a mixture offertilizations from males of
differentspecies. We are therefore hesitantto interpret these preliminaryresults with relativelysmall sample sizes in the context ofsexual
selectiontheory. Theseresultsdocomplicatetheinterpretationofhybridzonedynamics.
Additionalkeywords: Conspecificsperm, Papilioglaucus,P. canadensis, polyandry,hybridzonespermatophore..
INTRODUCTION for the entire life ofthe insect, in female organs called
Females ofmanyanimal species mate multiple times spermathecae (Cook and Wedell 1996; Gage 1998).
Some species with different types of sperm storage
within their fertile lifetime (Smith 1984; Birkhead and
M0ller 1998). There is evidence that these multiple organs (Pitnick et al. 1999) can separatelv store sperm
matings may extend female life-span (Wagner et al. from different males, as in the yellow dung flv (Ward
1998; Hellriegel and Bernasconi2000). This creates die
2001), increase egg production (Wiklund et al. 1993;
Boggs 1997a, 1997b, Karlsson 1998, Eadyet al. 2000), possibility that sperm from two (or more) different
andincrease offspringviability(Lederhouse and Scriber males areincompetition foreggfertilizations, extending
1987; Tregenza and Wedell 2002). It may also increase sexual selection beyond copulation. Packets of sperm
(spermatophores) are transferred to polvandrous
offspring weight/size (Sakaluk et al. 2002) or offspring
reproductive success (Bernasconi and Keller 2001; Pai females of the North American tiger swallowtail
and Yan 2002). However, where males are ofdifferent butterfly species group, raising questions of sperm
precedence and sperm competition from different
species, egg production can be drastically reduced
mates.
(Shapiro 2000) orpost-zygotic hybrid offspring survival
may be poor (Orr and Turelli 2001; Marshall et al. Females ofPapilio glaucus L. and P. canadensis R &
2002). Mixturesofspecies diagnostictraits from asingle J, often mate four to five times during their relativelv
mother could result from fertilization by both short lifetime ofthree to six weeks (Lederhouse et al.
heterospecific and conspecific males in a hybrid zone. 1989; Lederhouse 1995; Scriberetal. 1998a) andmales
Clarification of sperm precedence for naturally can mate at least four times (Lederhouse 1995). We
hybridizingspecies is needed. have shown that fertilitydeclines in P. glaucus females
Sperm are stored for use in fertilization, sometimes after seven to eight days and that a second mating may
restore it (Lederhouse and Scriber 1987; see also
CurrentAddress: Departmentof. BiologicalSciences Watanabe 1988). This may be the result of sperm
UniversityofNotreDame, NotreDame,IN 46556
depletion orotherfactors.
66 Journalofthe Lepidopterists' Society
As in many Lepidoptera, these two species ofPapilio heterospecifics as suspected in grasshoppers (Hewittet
usually package sperm in a spermatophore, which is al. 1989; Bella et al. 1992), crickets (Gregory and
energetically costly (Mann 1984; Watanabe and Hirota Howard 1994; Howard et al. 1998), Drosophila (Price
1999). Sperm are often dependent upon the female's 1998), ladybird beetles (Katakura 1986b), and flour
muscles to move them out ofthe spermatophore and beetles (Wade et al. 1993). If so, strong conspecific
transfer them to the spermatheca (Tschudi-Rein and sperm precedence could help explain why the hybrid
Benz 1990). Whilethepatternofpaternityin butterflies zone has remainedbasicallygeographicallydistinct and
andmothsisvariable, itis believedthatsperm fromone ecologicallystable forthe pasttwo centuries (Scriberet
male takes precedence. However, it mightbe an earlier al. 1982, 1996) despite the capacityforstrongflight and
mate instead ofthe most recent one that sires most or long range dispersal of both species (Scriber et al.
all of the offspring (Clarke and Sheppard 1962; Flint 1998b).
and Kressing 1968; Pair et al. 1977; Bissoondath and We employedspecies-specificdiagnosticallozymesto
Wiklund 1997; Cooketal. 1997;WedellandCook 1998; identify the paternity of larvae produced after two
LaMunyon 1999). The importance of which male's sequential hand pairings of individual virgin females
sperm sires offspring maybe ofeven greater ecological withheterospecificandconspecific (orviceversa) males
significance when males are from different species, as of P. glaucus and P. canadensis. We also compared
can occurinhybridzones ofparapatric species. males of the more distantly related P eunjmedon, P.
Multiple-mating females and males of tiger rutulus, P. multicaudatus, and P. troilus (Hagen and
swallowtail butterflies may encounter heterospecific Scriber 1991). The western species (P. eunjmedon, P.
mates as well as their conspecifics in the vicinity of rutulus, P. multicaudatus) also hybridize to various
hybrid zones such as the one across the Great Lakes degrees ( Sperling 1990, 2003; Scriber et al. 1995;
region (generally at latitudes of 40-43°N) from Layberry et al. 1998; Guppy and Shepard 2001).
Minnesota through Wisconsin, Michigan, New York Different benefits ofmultiple matings (accessorygland
State, and New England (Scriber 1996; Scriber et al. substances promoting increased egg production,
2003). We have documented such parapatric range ejaculates with nutritional benefits; Arnqvist and
overlap andcases ofnatural hybridization ofthe species Nilsson 2000) mayresultfrom repeatedmatingwiththe
in the field (Luebke et al. 1988; Scriber 1990; Deering same male versus different males (where indirect
and Scriber 1998; Scriber et al. 1998b; Stump 2000; genetic benefits may also occur; Jennions and Petrie
Ording 2001; Donovan and Scriber 2003; Hereau and 2000; ZehandZeh2001). Theimportanceofcomparing
Scriber 2003). Mating preferences of both P. glaucus laboratory versus natural polyandrous matings (where
andP. canadensis males inthe fieldarebothverystrong females choose their mates andthus influence bodi die
for tethered P glaucus females in 2-choice tests using direct and indirect benefits; see Tregenza et al. 2003)
size-matched, tetheredvirginfemales,whichrepresents has recentlybeenemphasizedforcrickets (Sakaluketal.
a strong and unexpected asymmetry (Deering 1998; 2002). Therefore, in additionto labpairingswithvirgin
Deering and Scriber 2002). Females were capable of females, wild allopatric females of canadensis and
free flight on the thread tethers andwere able to reject glaucus were captured and subsequently remated to
some male mating attempts. However, few rejections heterospecific males after first collecting her field-
were observed for these lab-reared virgin females. No fertilized (presumably conspecific) eggs for 1-3 days.
evidence ofsex pheromones has been detectedin field Directional posteopulatory sexual selection (involving
tests with crypticallycaged virgin females compared to both sperm competition, where sperm from different
empty control cages behind dead display females males compete to fertilize eggs, and cryptic female
(Deering 1998). choice,where females bias sperm usetowardparticular
Interspecific hybridization for P. glaucus and P. males)hasbeenrevealedbyusingartificialinsemination
canadensis has been documented to increase in guppies to bypass the complications with natural
extensively within and north of the historical hybrid mating preferences and mating order of polyandrous
zone during the recentwarming climatic trends during females (Evans et al. 2003). We have functionally
1998-2003 (Scriber 2002a; Scriber and Ording 2005). accomplished the same thing with laboratory hand-
The potential of post-mating (but prezygotic) pairingtechniques forPapilio.
reproductive isolation hadnot been examinedforthese All female butterflies of the P. glaucus group are
butterflies (e.g. differential selection of sperm by presumably polyandrous and store sperm in
females). Inparticular, wewonderedwhethermultiple- spermathecae We hopedto determinewhether sperm
.
mating females of P. glaucus or P. canadensis may from eitherfirstorlastmales haveprecedence,whether
cryptically select sperm of conspecifics over die female may exert cryptic sexual selection of
.
Volume 60, Number 2 67
conspecificsperm forfertilizinghereggs, orifamixture Female butterflies were fed a 20% honey solution
ofspermfrombothspecies (withpotentialcompetition) and maleswere feda20% honeysolutionsupplemented
wasinvolved. Inadditiontobeingthe firstsuch studyin with amino acids and salts following Lederhouse et al.
these species, the implications ofthe results for hybrid (1990). Lab-reared males were not paired until at least
zone dynamics and the interpretation ofthe genetics of two days following adult eclosion. For comparison,
offspring from wild-caught females could also be various other single male pairings among these species
significant. From asingle motherofhybridzone origins, were conducted (1996-1999). Comparisons ofegg and
amixture ofoddsegregatingtraits in offspring mightbe larval offspring production from field-captured females
explained bv multiple matings with males ofdifferent for various species were also included from years
species (Clarke and Sheppard 1962). 1983-2002. A separate group of primary (F hybrid
L)
crosses between P. canadensis and P. glaucus from
Methods
1995-2002 were compared to the reproductive fitness
Pairings. All doublepairingswere conductedduring ofwild females ofboth species (Table 8) to provide a
the summers of 1996, 1997, and 1998 (Tables 1-4). comparison offecundityandviability.
Both wild-caught and lab-reared male and female Allozyme Electrophoresis. Electrophoresis was
butterflieswereusedforpairings. Females andmalesof carried out on thin-layer cellulose acetate plates (Titan
P. glaucus (G) and P. canadensis (C) and males of P. III, Helena Laboratories, Beaumont, TX), using
rutulus (R), P. eurymedon (E), P. multicaudatus (M) methods of Hagen and Scriber (1991). Small larvae
were used. Lab-reared, newly eclosed, virgin females were homogenized whole in bufferwhile the head and
were hand-paired to males and allowed to oviposit in thoraxes oflarger larvae were homogenized in buffer.
multichoice plastic arenas lined with host plant foliage The distal halfofthe abdomen ofadult males and die
as previously described (Scriber 1993). These were proximal halfofthe abdomen ofadult femaleswas used
remated (aftertwotosixdays andhavinglaidat least20 since the distal part of the female abdomen included
eggs), again by hand pairing, to a male of a species male spermatophores. The enzyme 6-phosphate
differentfrom thatofthe first male. Femaleswere then dehydrogenase (Pgd)wasusedtodetermine differences
allowed to oviposit again. Additionally, wild caught between the species of the P glaucus species group
females were allowed to oviposit in plastic boxes, then since all 5 species are fixed for different diagnostic
rematedafterone to five days byhand-pairingto amale alleles (Hagen and Scriber 1991). Other enzymes with
of a different species, and allowed to oviposit again. diagnostic differences between species were not used
Only females that were actively laying eggs were because those alleles were too faintlystained in larvae.
remated. The duration of pairings was recorded, as Staining ofPgdwas fainter for larvae than adults, butit
transferofspermatophores takes at least 30 minutes. was clearandinterpretable. In other Lepidoptera. fixed
For the P. glaucus species group, larvae from eggs allozyme differences have also been used to identify
laidboth before and afterrematingswere collectedand egg, larval, pupal, andadultspecies (Woodsetal. 2001).
rearedonWild BlackCherry (Prunusserotina Ehrh.), a Paternity- of offspring produced after remating of our
common host of both P glaucus and P canadensis Paf)ilio was established by determining alleles of Pgd
Larvae ofP. troilus were reared on Sassafras albidum allozymes oflarvae produced after remating. In several
Nutt. orLinderabenzoin (L.) Blume (Lauraceae). After cases, the males were lost (or accidentally not frozen
reaching approximately the third instar, larvae were soon enough) and could not be evaluated. Precedence
frozen at -80°C. Adults were frozen before or for each female was expressed as P2, the proportion of
immediately after death. Females were later dissected larvae produced after the remating that were sired bv
to determine how many spermatophores were present the female's secondmate.
We
at death. include females that after dissection only
had a single spermatophore (Tables 3 & 4), becausewe Results
wanted to determine which ofthe two males passed it. Lab and wild pairings. Out of 82 interspecific
However, we deleted these pairings in our summaryof double pairings for lab-reared females, only 32
"last" (P2 = 1) or"first" (P2 = 0) sperm precedence. We produced larvae both before and after being remated
are also aware that sperm transfer without leaving (Table 1). The others either laid no eggs after being
recognizable spermatophores can rarely occur in remated, laidno hatching;esss (all beingeidierinfertile
Lepidoptera (Cordero 1999). Analyses of egg fertility or fertile and non-viable) after remating, orhad laid no
and viability ofa set ofsingle lab-paired females from hatching eggs before being remated. Ofdie 32 females
1996-1999 with no spermatophores (n=45) were producing larvae both before and after being remated,
comparedto thosewith one present (n=274). five had been mated to males that shared allozvmes.
68 Journalofthe Lepidopterists' Society
Some interspecific introgression (or ancestral low-level possible ornot.
polymorphisms, or residuals from earlier north/south Durations of second pairings were recorded. No
Pleistocene hybrid zone movement) is found atthe Pgd secondpairings lasting less than 30 minutes resulted in
locus (Hagenand Scriber 1991; Scriber 1996; Scribere£ sperm replacement. This is notsurprisingbasedontime
al. 1998b) making the determination of paternity for needed for ejaculate transfer in other Lepidoptera
offspring impossible in these cases, leaving 27 broods (Watanabe and Sato 1993). However, most second
where P2was determined. Ofthese females, 6had only pairings lasted for longer than 30 minutes and most
a single spermatophore recovered, and 3 others females were foundto be carrying two spermatophores
produced fewer than 5 larvae, leaving only 18 for P2 (or more forwild caught females).
analyses. Since the sample sizes were so small, we did Singly-mated females illustrate the need for a
notconductt-tests. minimum of 30 minutes of copulation duration for
In addition to the sequential pairings ofvarious lab- successful fertilization (Table 5). A surprisingly high
reared virgin females (Table 1), wild females were also number of females laid eggs even when no
set up in oviposition arenas and later remated to spermatophore was present (0.62, n = 45) compared to
heterospecific males (Table 2). Out of27 wild females, those with a spermatophore (0.75, n=274, Chi-square
20 produced larvae both before and after being 3.32, n.s.). However, only a very small portion of
remated. All of the remated wild females produced matings with no spermatophores detected produced
larvae before being mated. Only one female that had larvae (0.08, n=24) compared to those with a
produced larvae after being remated had a P2 that was spermatophore (0.57, n=189, Chi-square 19.87, p<
not possible to be determined electrophoretically, 0.001).
leaving 19 broods where P2 was determined (Table 2). Even with regular feedings of males with a honey
Of these 19 females, 5 had only one (or no) water/electrolyte "elixir" (as in Lederhouseetal. 1990),
spermatophores and another 7 had fewer than 5 larvae, it is evident that lab-reared males are less successful at
leaving 7 for P2 allozyme analyses. A shared host plant producing broods with larvae than are wild field-
(Prunusserotina) was usedto minimize anydifferential collected males (Table 6). An index of reproductive
mortality of larvae before the third instar samples for success also shows that primary F interspecific hybrids
l
allozyme analysis, but this delay from egg hatch to the between P. glaucus and P. canadensis parents (Pc x Pg
third instar, when allozyme samples were made, andPg x Pc) have offspringsuccess thatis intermediate
nonetheless mayhaveintroducedanunknowndegreeof comparedto the parentaltypes (Table 6).
selection bias due to differential mortality among Average hatchabilities of eggs (# larvae/total, eggs)
genotypes. Larvaehadallozymescorrespondingtotheir from field-captured females ofthese 5 species (ranging
parents in allofthe caseswhere allozymes ofparents (a from 0.43to 0.70; Table 7) were comparable to ourlab-
once-mated female and her mate) were compared to rearedvirgin femalesintheseexperiments (0.25to0.72;
larval offspring allozymes. We used Pgd, which is X- Table 6). The proportions of females that oviposited
linked in P. glaucus and P. canadensis. Male ranged from 33% to 84% among the 7 Papilio species.
homozygotes of Pgd could not be distinguished from Thepercentageproducinglarvae rangedfrom only16%
female hemizygotes in larvae since the sex was not to 52% (Table 7). In a separate study of singly-mated
ascertained. The sample numbers are not high enough females that did produce larvae (from 1995-2002), it
to be able to estimate accurately the actual population was clear that the heterospecific hybrids between P.
allelefrequencies,butthereisnoreasontosuspectnon- glaucus and P. canadensis were at no obvious selective
Mendelian inheritance. disadvantagecomparedtotheparentaltypes;with mean
Of the 18 double-paired interspecific lab crosses egg viability of68.0% (Pc x Pg) and 62.3% (Pg x Pc)
examined, 11 showed first male sperm precedence (P2 comparedwith58.9% forP canadensis and65.3% forP
= 0), 4 showed mixedparentage (P2 = 0.20 - 0.93), and glaucus (Table 8).
3 showed last male sperm precedence (P2 =1). Only
Discussion
one of these three last male fertilizations was
heterospecific in paternity (Table 3). Of the wild Sperm precedence. Second male precedence has
females remated to different species in the lab, 3 been considered to be the rule in Lepidoptera
showed mixed parentage (from both males; P2 = 0.07- (Drummond 1984), but our results showed 3 of 25
0.14) and the other 4 all showed first male sperm successful double interspecific matings complete
precedence (P2 = 0; Table 4). The number of useful second male sperm precedence (P2 = 1) in these tiger
double male pairings was insufficient to evaluate swallowtail butterflies (Tables 3 & 4). Two additional
whether female selection of conspecific sperm was twice-mated females showed a large proportion of
Volume 60, Number 2 69
Table 1.Typesandnumberofdouble-pairedPapiliofemalesandreproductivesuccessfollowingremating. P2iscalculatedastheproportion
oflarvaeproducedafterasecondmatingthatweresiredbythefemale'ssecondmate.
producinglarvae
producinglarvae beforeandafter
Double-pairingtype double-paired layingeggs producinglarvae beforeand rematingwhereP2
(femalex male 1 xmale2) females afterremating afterremating afterremating couldbedetermined
(CxCxG) 18 17 8 6 5
(CxGxC) 13 12 6 5 5
(GxGxC) 17 13 9 6 6
(GxCxG) 20 18 13 8 5
(CxCxE) 7 5 2 2 2
(CxExC) 4 4 2 2 2
(GxGxM)
1 1 1 1 1
(GxMxG)
1 1 1 1
(GxRxG)
] 1 1 1 1
Total 82 72 43 32 27
72/82 =87.8 43/72=59.7 32/4 44.4 27/35!=84.4%
C=canadensis,G=glaucus,M =multicaudatus R = rutulus,E =eurymedon
,
Table2.Typesandnumberofrematedwild-caughtPapilio femalesandreproductivesuccessfollowingremating.
Producing Producinglarvae
Number Layingeggs Producing larvae before beforeandafter
remated after larvaeafter andafter rematingwhereP,
Double-pairingtype(femalewildx male) wild-caught remating remating remating couldbedetermined
P. canadensiswildxP.glaucus(Cwildx G) 8 7 6 6 6
(GwildxC) 2 2 2 2 2
(GwildxE) 7 6 6 6 5
(GwildxM) 4 3 3 3 3
(GwildxR) 6 3 3 3 3
Total 27 21 20 20 19
21/27= 77.8 20/21 =95.: 20/20= 100 19/20=95%
E, M,R= eurymedon,multicaudatus,&rutulus,respectively
mixed sire offspring (P2 = 0.82 and 0.93) and 5 others (Simmons and Siva-Jothy 2001). Last male sperm
showed less extensive mixing (P2 = 0.07-0.36). This precedence occurred (at leastpartially) in only 10 of25
pattern appears to be in contrastto the last mate sperm double matings for which we could analyze diagnostic
precedence believed to be typical ofinsects in general allozymes. However, offspring sired bv the last male
(Smith 1984) and especiallyLepidoptera. Allbut oneof mate occurred in both P. canadensis and P. glaucus
the 20 Lepidoptera species studied have mean P2 females even when diis male was heterospecifie
values greaterthan 0.47 and most are greaterthan 0.60 (canadensis, glaucus, andeurymedon; Tables 3 & 4). In
(seeTable2.3in Simmons2001); although itis reported addition to these inconclusive results, the last male
that distributions are bimodal with or 1 values for sperm precedence patterns in other insects is also
most species. Lepidopterahavingbimodal distributions known to break down when a third male is involved
create intermediate mean values ofP2, and that is one (Zeh and Zeh 1994) as is the case widi many older P.
reason why P2 analyses are not very useful in glaucus and P. canadensis females (Lederhouse and
understanding mechanisms of sperm competition Scriber 1987; Lederhouse et al. 19S9; Lederhouse
70 Journalofthe Lepidopterists' Society
Table3.Lastmalespermprecedence,P2,theproportionofoffspringthatweresiredbythesecondmale,fordouble-pairedlab-rearedfemales.
Alsoindicatedarethenumberoflarvaeproducedafterrematmgthathadpaternitydeterminedbyelectrophoresis(N),originofthemaleused
forremating,thedaysbetweenpairings,thedurationoftherematmg,andthenumberofspermatophorespresentinthefemaleatdeath.
Type
(femalexmale Daysbetween Durationof Spermatophores
xmale2) Maleorigin pairings remating(minutes) present P2 N
CxCxG lab 3 65 2 1 15
lab 3 100 2 27
lab 2 26 1 61
lab 3 60 2 0.2 5
C x GxC lab 3 59 2 7
lab 3 35 1 22
lab 2 >43 2 12
wild 4 57 2 1 21
Gx GxC lab 5 93 1 16
lab 2 >36 2 21
lab 2 >38 2 23
lab 3 62 2 0.36 11
lab 3 63 2 26
lab 3 106 1 14
GxCx G lab 3 >30 2 0.82 11
lab 4 85 2 72
wild 2 87 2 26
wild 4 >30 2 23
CxCxE
wild 4 >41 2 21
wild 2 73 2 0.93 14
CxExC
wild 2 >85 1 21
wild 2 >91 2 1 5
GxRxG
wild 1 99 2 19
1995). rapid after 6-10 days for P. glaucus and after 4-7 days
The relativelyhigh success offirst matings compared for P. canadensis. Females in the field may normally
to the second matings in our studies is due to unknown rejectsecondmales foralongerperiodoftheirlifethan
factors. One possibility is simply that we did not wait was the case for our experimentally forced laboratory
longenough afterthe first pairingfor natural decline in handpairings. It is also possible that the second male's
fertilityorspermvolume before the second matingwas sperm may have had insufficient time to reach the
made. For example, Lederhouse and Scriber (1987) spermathecaoffemales in 2-3 days. Another area that
have shown that fertilityofthese species declines over deserves additional experimentation would be the
the firstweek after mating, but that the decline is most assessment of whether males prevented from their
Volume 60, Number 2 71
Table4. SpermprecedenceP2forrematedwild-caughtfemales.Alsoindicatedarethenumberoflarvaeproducedafterrematingthathad
paternitydeterminedbeelectrophoresis(N),originofthemaleusedforremating,thedaysbetweencollectionofdiefemaleandremating,
thedurationoftheremating,anddienumberofspermatophorespresentinthefemaleatdeath.
Durationof
Daysuntil remating Spermatophores
Type(femalewildxmale) Maleorigin remating (minutes) present P2 N
CwildxG wild 3 43 2 T
wild 4 33 14
wild 3 60 3 0.14 7
wild 3 72 3 5
GwildxC lab 2 42 1 20
lab 2 49 2 20
GwildxE wild 1 >40 4 0.07 14
wild 1 54 2 0.12 8
wild 4 64 1 22
wild 4 >75 2 21
GwildxR wild 1 27 1 11
natural choice of female mates (in hand-paired cases) passive conspecific sperm selection (Birkhead2000). It
would withhold or reduce the amount of sperm has been suggested that genetic benefits from mate
transferred in laboratory pairings compared to natural choice may be context-specific (Qvarnstrom 2001). In
matings. Fewstudies exist ofsperm utilization patterns any case, it is feasible that sperm competition could
in nature (Cobs 1977; Allen et al. 1994; LaMunyon override anycrypticfemale selection shoulditbe found
1994). However, recent research shows that male tooccur.
ejaculate expenditures are dynamic and that males The failure ofmanysecond matings to sire offspring
assess the mating status and relative fecundity of in these interspecific pairing sequences is likelv due to
females and subsequently modulate the quantity or some ofthe sameproblemsencounteredinfirstpairings
quality ofejaculate they pass to females (Wedell et al. (Stump 2000). For example, as with first pairings, it
2002). This also exists as a possible explanation ofour appears that second pairings lasting at least 30 minutes
poorsecond matingsuccess. are required for sperm replacement (Table 3). We
The roles of apyrene (lacking a nucleus) and foundinPapilio singlematingstudies (Table6)diatonly
eupyrene (with nucleus) sperm in competition (Cook about 25-45% ofhand pairings gave rise to some larvae
and Wedell 1996, 1999; Watanabe et al. 2000) are (24.0%canadensis xcanadensis, n=75;32.4%glaucus x
completelyunknown forthese two Papilio species, and glaucus, n=33; 42.1% ofglaucus x canadensis hvbrids.
knowing these might help clarify results such as ours. n=38; and44.0% ofcanadensis xglaucushvbrids, n=50;
The degree ofpolygamy and polyandry will vary with Stump2000). Previousstudiesshowedthat in 19S6onlv
age, extent ofprotandry, the size ofspermatophores, as 19% of 75 pairings produced some larvae (14 fertile
well as species characteristics, all of which makes females averaging27larvae from 73 eggs) Afterfeeding
understanding sperm precedence a complex venture males an elixirsolution ofamino acids, electrolrtes. and
with slowed progress (Kempanaers et al. 2000; Pitnick sugar our results improved to 90% of 20 females
and Brown 2000;Wiklund2003). producing some larvae (18 fertile females averaging44
The females that continued to produce fertile eggs larvae from 105 e^Deos; Lederhouseetal. 1990).
andlarvae siredbyan earliermale (70% ofthose shown We do not know t'he explanation for seeminglv low
in Tables 3 and 4) also did so even when this original percentages of mating success, but it is reasonable to
mate was a heterospecific male (followed in mating assume that similar problems may explain at least a
sequence bya conspecific male). This further suggests portion ofthe lowsecondmating"replacement"success
that these Papilio females maynotbe exertingactive or (70% offemales continued to use earlier sperm). Wild
72 Journalofthe Lepidopterists' Society
TABLE5. Successofsinglematepairingsasindicatedbythosewithviablelarvae,groupedbymatingduration.Thepairings1wereofvarious
types (conspecific,heterospecific, backcross, and F,), involvedvarious Papilio species (P. canadensis, P. glaucus, P. eurymedon, P. rutulus, P.
multicaudatus,andP. troilus), andwerebetween lab-rearedvirginfemalesandeitherlab-rearedorwild-caughtmales. Samplesizesforeach
duration/categoryareindicatedinparentheses. Notallfemalesweredissectedandexaminedforaspermatophore.
Proportionofpairingswitha Proportionofallpairings Proportionofpairingswithlarvae
Duration(minutes) spermatophore2 producingeggs (ofthosewitheggs)
<30 0.0 0/10 61.1 11/18 0.0 0/11
30-90 94.6 210/222 75.3 223/296 59.2 132/223
>90 76.7 23/30 78.7 37/47 56.8 21/37
'all 1996-1999laboratoryhand-pairingsofvariousPapiliopopulationsorspecieswereincluded.
2Some femaleswerenotexaminedforspermatophores
females of the 7 species also had low proportions Sakaluk et al. 2002). Our experiments involved only
successfully ovipositing andproducing larvae (Table 7). combinations that used male mates from two different
Rather than sperm competition, our experimental species. No double conspecific matings were made in
procedures (secondmatings afteronly2-6days) maybe this study. With hand-pairings, females were unable to
apartial explanation for our generallylow second male make natural mate choices seen for females with free
fertilization success. Waiting 7-8 days might better choice either in the lab or in the field (which can also
reflect the natural rematingbehavior andphysiological, vary; Sadek2001).
needsoffemales. Rematingintervalandspermatophore Males ofvarious insect species mayattempt to guard
sizecan alsoaffectthe P2values andnumberofmatings their mates from competing males (Birkhead and
for Lepidoptera (Drummond 1984; Lederhouse et al. Parker 1997) or to secrete mating plugs to block
1989; Simmons and Siva-Jothy2001). competitors access to the female's reproductive tract
First or last male sperm precedence may occur and (Thornhill and Alcock 1983; Orr 1995). In addition,
be operational in different ways for multiple matings males mayattempttogainanadvantageoverpotentially
and be limited to only conspecific males (or only competing sperm by a number of mechanisms. First,
heterospecific males). In other species, viability of there maybe removal ofprevious male sperm (Lefevre
offspring has been shown to increase with additional and Jonsson 1962; Waage 1979; Ono et al. 1989; Orr
conspecific male mates by avoiding the costs of 1995; Price et al. 2001). Secondly, hindering or
inbreeding (Tregenza and Wedell 2002) or by gaining rendering rival sperm non-functional with seminal
other genetic benefits (Yasai 1998; Newcomer et al. secretions may occur (Katakura 1986a; Katakura and
1999; Jennions and Petrie 2000; Zeh and Zeh 2001; Sobu 1986; Harshman and Prout 1994; Clarke et al.
TABLE6. Indexofmatingsuccessforsinglematepairs(1996-1999). Foreachpairingtype,diisiscalculatedastheproportionofpairingspro-
ducinghatchinglarvaemultipliedbytheaveragehatchabilityofbroodsproducinglarvae.Femalelistedfirstineachpairingtype.
PairingType Ofpairingslastingatleast
(Femalex Male) 30minutes,proportion Averageegghatchabilityof
origin (n) producinglarvae pairingsproducinglarvae Inde ofmatingsuccess
CxCwild (17) 0.45 0.49 ± 0.06 0.22
CxClab (7) 0.11 0.40 ± 0.12 0.04
GxGwild (18) 0.78 0.72± 0.13 0.43
Gx Glab (18) 0.42 0.63± 0.07 0.26
Cx Gwild (22) 0.52 0.61 ±0.06 0.32
CxGlab (10) 0.33 0.47± 0.10 0.16
GxCwild (37) 0.59 0.66±0.04 0.39
GxClab (4) 0.20 0.28±0.09 0.06
TxGwild (6) 0.67 0.25± 0.03 0.17
(CxG)xCwild (17) 0.52 0.37±0.06 0.19
(CxG)xCarebackcrossbroodsusinghybridfemalesandwildcanadensismales.
T=P. troilus
Volume 60, Number 2 73
Table7. Pooledtotallaboratorylifetimeeggandlarvalproductionofwildfield-collectedPapiliofemalesofthetigerswallowtail(glaucus)
speciesgroup.
Female Families Population traits
Species Total (n) withEggs with Larvae Totaleggs totallarvae hatchability(1/e)
canadensis 730 353(48%) 230(32%) 13205 6311 .478
glaucus 959 524(55%) 369(38%) 31806 17341 .545
rutulus 65 43(66%) 33(51%) 1479 795 .538
euryinedon 33 11(33%) 5(16%) 297 172 .579
multicaudatus 25 21 (84%) 13(52%) 599 259 .431
alexiares 31 19(61%) 12(38%) 919 475 .517
troilus 88 55(63%) 35(39%) 4417 3083 .698
P. canadensisandP. glaucus (andP. alexiares)dataaretrom 1983-1986only(othersallinclude 1991-2002individualsaswellas 1983-1986).
1995; Rice 1996; Price et al. 1999. 2000). Thirdly, conflict within a species can lead to the rapid evolution
diluting or delaying competing sperm with sperm of of reproductive isolation (Tregenza 2003; Martin &
highervolumes maybeused(Sugawara 1979; Dickinson Hosken 2003). The role such sexual selection mayhave
1986; Gage and Baker 1991; Karlsson 1995; Gage and in the reproductive isolation ofhybridizingspecies such
Barnard 1996; Simmonsand Siva-Jothy 1998). Fourthly; as these Papilio is currently unknown but may be
transferring sperm orspermatophores ofhigher quality extensive. Our study is only preliminary, and was
or effectiveness may give an advantage (Bertran et al. conducted to assess the physiological possibilities of
1996; Smedley and Eisner 1996; Vahed 1998; differential spermusebymultiplymatedfemalesandto
LaMunvon and Ward 1999; Gilchrist and Partridge see ifconsistent patterns ofsperm precedence occurs.
2000; Stjernholm and Karlsson 2000; Price et al. 2001; In these lab studies, we were unable to test sexual
Bergstrom and Wiklund 2002). Finally, there may be selection theory.
production ofhigh volumes ofnon-nucleated apyrene
Conclusions
sperm, which may prevent fertilization by eupyrene
sperm of former males (Silberglied et al. 1984; We have shown that, under our sequential double
Watanabe et al 2000). However, since some females mating regime, females of P. glaucus and P. canadensis
may potentially manipulate sperm storage and use generally produce larvae fertilized by the first rather
(Birkhead et al. 1993; Sakaluk and Eggert 1996; Siva- thansecondmale. Sperm mixing(with eggs fertilizedbv
Jothy and Hooper 1996; Otronen 1997; Wilson et al. both species) andlast male sperm precedence do occur
1997; Neubaum andWolfner 1999; Olsson et al. 1999) at lower frequencies than first male sperm precedence.
to give certain types of sperm a potential competitive However, we had insufficient numbers of useful two-
advantage (e.g. to conspecific males overheterospecific species matings ofour lab andwild females to evaluate
males) (Robinsonet al. 1994; Albuquerque et al. 1996; any potential conspecific cryptic sexual selection ofdie
Eberhardt 1996; Howard 1999), the particular sperm types by females. Since it is so difficult to
mechanisms determiningwhich sperm wins is difficult distinguish selection from sperm competition, it
to discern. It may relate to the nature ofseminal fluids remains controversial (Simmons and Siva-Jothy 2001)
(Price 1997; Simmons 2001) or of the deposition whether females really have some influence or
patterns in all female storage components (Ward 1993). selection-capability in cases with conspecific sperm
Also, both male and female processes may interact to precedence (Howard 1999). However, the poorly
produce sperm displacementortoaffecttheoutcomeof understood role of sperm from different species in
sperm competition (Clark et al. 1999; Simmons et al reproductive isolation in Papilio or other hybrid zones
1999; Eberhard 2000; Ward 2000; Simmons 2001). for polyandrous females may be significant (Arnold
However, males attempting to increase their own 1997; Endler 1998; Turelli et al. 2001; Howardet al.
fertilization success can result in a decrease offemale 2002; Via2002;Andolfattoetal. 2003). Itwasevidentin
fitness, producing sexual conflict (Stockley 1997; these studies that lab-reared males (even when fed an
Simmons 2001; Chapmanetal. 2003; Pizzari and Snook amino acid, salt, sugar elixir solution; as in Lederhouse
2003). At a local population level, such male-female et al. 1990) were still inferior to field-collected "wild"
74 Journalofthe Lepidopterists' Society
TableS. EggviabilityforwildPapilioglaucus,wildP.canadensis,theirreciprocalhybrids(femaleslabreared;maleswild).Datarepresentthe
mean( ±se)ofal.1individuE1femaleis(1995-2002).
(#offemales) Totallarvae Totaleggs EggViability
P.glaucus (246) 77.3(5.1) 110.5(5.7) 65.3(1.6)%
P canadensis (305) 31.7(2.3) 53.5(2.9) 58.9(1.7)%
PrimaryHybrids
PgxPc (73) 108.5(11.1) 167.1 (15.1) 62.3(3.6)%
PcxPg (17) 55.2(13.0) 77.1 (16.6) 68.0(6.2)%
Backcrosses(1997-2002);
PgPcx Pc (3) 118.0(37.3) 148.3(42.9) 78.5(3.3)%
PcPgxPg (6) 18.6(9.9) 33.2(10.8) 44.9(12.2)%
PcPgxPc (20) 33.2(7.0) 80.1 (10.9) 39.9 (5.6)%
PgxPgPc (1) 278 288 96.5%
Pg xPcPg (1) 17 143 12.0%
F,Hybrids
PcPgxPcPg (2) 24.5 80.1 16.9'
PgPcxPgPc (1) 12 63 19.1 '
PapilioglaucusfrompopulationsinvariousstatesofdieUSA(FL,GA,OH,MD,VA, MO, MI)
P. canadensisfromvariouspopulationsinCanadaandtheUSA(AK,WI, MI,NY,VT)
T-testsbetweenP.glaucusvs.P. canadensiswereallsignificant(p <0.01)fortotallarvae,total,eggs,andeggviability. Forhybrids(PgxPcvs.
PcxPg)viabilitieswerenotsignificantlydifferent,buttotaleggsandtotallarvaewere(p< 0.01). Mean eggviabilityofBackcrosstypePcPgx
Pc (39.9%, n=20families) wassignificantlylowerthanbothparentaltypesandbothprimaryhybrids(allp< 0.001).Anothertwobackcross
types(n=6;andn=20families)werenotsignificandydifferentdranparentalsandprimaryhybridsforeggviability(genotypeswithonly1or2
familieswerenotcompared).
males with respect to reproductive potential in hand- Acknowledgements
pairings forthesePapilio. Differentiallarvalfoodquality Thisresearchwassupportedinpartbythe MichiganAgricultural
and natural puddling behavior ofadult males (Pivnick Experiment Station (Project MICL01644), NSF grants
and McNeil 1987; Boggs 1997a, 1997b) can affect male E(DnEdBo9w2m2e0n1t22ofatnhdeDDEeBp9a9r8t1m6e0n8t)oafnEdntthoemoRlaoygya.ndWeBertnhiacnekHJuetsssiocna
physiologicalvirilityandmatingsuccess and maylargely Deering, Mark Deering, Jennifer Donovan, Cheryl Frankfater,
account for the higher reproductive success of wild Keegan Keefover, Rodrigo Mercader, Jenny Muehlhaus, Michelle
males than lab-reared males in ourstudy. Oberlin,GabeOrding,andSaraSandersforassistanceindiefieldand
lab. HarryPavulaanandDavidWrightwereverygenerousinsending
Finally the extent ofhybrid vigor as seen in our F, P. appalachiensis individuals from West Virginia and Pennsylvania
hand-pairedlabcrosses (seealso Donovan2001; Scriber mountains for morphological and eleetrophoretic comparisons. We
etal. 2003) maynot occurin the natural field situation areparticularlygratefulto reviewer#1, whoprovidedextensiveand
particularlyinsightful andhelpfulcomments.
where hybrid males (Davies 1997) or natural enemies
might differentially affect the fitness of hybrids (e.g. Literature Cited
Mallet and Barton 1989) or backcrosses (Hagen and Albuquerque,G S.,C.A.Tauber6c M.J.Tauber. 1996. Post-mat-
Scriber 1995; Dasmahaptra et al. 2002; Presgraves ingreproductiveisolationbetweenChnjsopaquadripunctataand
2002). Recent examples of likely recombinant hybrid Chrysopa slonnae: mechanisms and geographic variation.
Evolution50:1598-1606.
speciation in insects (Schwarz et al. 2005; Scriber and Allen,G.R.,D. Kazmer&R.FLuck. 1994.Post-copulatorymale
J.
Ording 2005; see also Rieseberget al. 2001) dependon behavior, sperm precedence and multiple mating in a solitary
introgressive hybridization. However, the role ofsperm Andoplafraatsittoo,idPw.aJs.pM..ASncirmi.beBreh&av.B.48C:h6a3r5l-e6s4w4o.rth.2003. Noasso-
precedence in the recently apparent ineffectiveness of ciation between mitochondrial DNA haplotypes and a female-
prezygotic barriers between P. glaucus and P. limited mimicry phenotype in Papilio glaucus. Evolution 57:
canadensis (Deering and Scriber 2002; Scriber 2002a; 305-316.
ARNOLD, M. L. 1997. Natural Hybridization and Evolution. Oxford
Donovan and Scriber2003; Scriberetal. 2006) was not Univ. NY.
resolvedbythis study. Bella,J. L... R. K.Butlin,C.Ferris&G. M. Hewitt. 1992.Asym-
metricalhomogenyandunequalsexratiofromreciprocalmating-
order crosses between Chorthippus parallelus subspecies.