The Auk 112(2):473-482, 1995
MONOGAMY
IN LEACH'S
DNA-FINGERPRINTING
STORM-PETREL:
EVIDENCE
ROBERTA. MAUCK, THOMAS A. WAITE, 1 AND PATRICIA G. PARKER
Departmentof Zoology,OhioStateUniversity,Columbus,
Ohio43210, USA
ABSTRACT.--We
usedmultilocusminisatelliteDNA fingerprintingto estimatethe frequency
of extrapairfertilizationsin a populationof Leach'sStorm-Petrel(Oceanodroma
leucorhoa)
on
Kent Island, New Brunswick, Canada. Leach's Storm-Petrel is a member of Procellariiformes,
an orderof long-livedpelagicbirdscharacterized
bylong-termpairbonds,
single-eggclutches,
and extendedperiodsof parentalcare.We found no evidenceof extrapairfertilizationsin
48 families(42 full familiesand 6 partial familiesconsistingof the putativefatherand the
singleoffspring).Thus,our resultsindicatethat the breedingsystem(geneticmonogamy)
matchesthe matingsystem(socialmonogamy)in our studypopulation,a conditionthat no
longercan be assumedin sociallymonogamousbird species.Geneticmonogamyin Leach's
Storm-Petrels
maybemaintainedby last-sperm
precedence
andfrequentcopulationby mates
duringthe female'sfertile period.Suchtacticsemployedby a malemay yield a high probability of fertilizing the singleegg laid by his mate.Received
15 June1994,accepted
31 August
1994.
IN MOSTAVIANspecies,a male and a female
form a pair bond and cooperateto produceoffspring (Lack 1968).Trivers (1972), however, argued that sociallymonogamousmales should
be expectedto pursuea strategythat maximizes
the trade-offbetween parental effort and mating effort outsidethe pair bond. As the benefit
of male parental care decreases,the costof parental neglect lessensand the net benefit to
males pursuing extrapair fertilizations (EPFs)
increases.Femalesshould be expected to seek
EPFs,which may favor the acquisitionof better
genesfor their offspring, increasegenetic diversity amongoffspring,increasethe probability of fertilization, or yield other materialbenefitsthroughcourtshipfeedingor increasedparental care (reviewed by Westneatet al. 1990,
Birkhead and Moller 1992). Thus, there are potential fitness gains associatedwith extrapair
mating activity in any monogamousspecies.
Extrapair copulations(EPCs) have been observed in many socially monogamousspecies
(e.g. Black-cappedChickadees[Parusatricapil-
tem.Malesand femalesof manyapparentlymonogamousspecieshave been shown to engage
in extrapair reproductiveactivities(e.g. House
Sparrows[Passer
domesitcus],
Burke and Bruford
1987;Indigo Buntings [Passerina
cyanea],Westneat1990;White-crownedSparrows[Zonotrichia
leucophrys],
Shermanand Morton 1988;Eastern
Bluebirds[Sialiasialis],Gowatyand Karlin 1984,
Gowaty and Bridges1991).
Clearly, equal genetic contributions to offspring tended by breeding partners (genetic
monogamy)can no longer be safelyassumedin
sociallymonogamous
bird species.ReportedEPF
ratesin suchspeciesrangefrom 0%(e.g.Northern Fulmar, Hunter et al. 1992; Black Vulture
[Coragyps
atratus],Decker et al. 1992) to greater
than 30% (e.g. Indigo Bunting, Westneat1990;
Purple Martin [Prognesubis],Morton et al. 1990;
Tree Swallow [Tachycineta
bicolor],Lifjeld and
Robertson1992). Following the recent discoveries of extrapair parentage among the offspringof sociallymonogamous
birds,it remains
of interest to identify those speciesin which
lus], Smith 1988; Northern Fulmars [Fulmarus apparent reproductive successaccurately reglacialis],Hatch 1987), although until recently flectsrealized reproductivesuccessand to clarEPF detectionwas difficult. Genetic techniques ify the factorspromotinggeneticmonogamy.
Speciesin the order Procellariiformesfulfill
suchas DNA fingerprinting now permit accurate assessment
of parentagewithin a socialsys- the conventional,hypothesizedcriteria for social monogamy(Wittenbergerand Tilson 1980,
Silver et al. 1985, Westneat et al. 1990, Mock
•Present address:School of Forestry, Michigan
Technological
University,Houghton,Michigan49931,
USA.
473
and Fujioka 1990); they are long-lived birds
characterizedby long-term pair bonds,singleegg clutches,and extendedperiodsof biparen-
474
MAUC•C,
WAITE,ANDPAR•CER
[Auk,Vol. 112
tal care during which they forage over great M NaCI, 6 mM EDTA, 0.2%sodiumazide), a nonlytic
preservative.After cells had settled,the clear superly assumedthat breedingadultprocellariiforms natant was removed and replaced with fresh PBS.
are the geneticparentsof the young they raise. Sampleswere inverted several times to resuspend
cellsand,subsequently,
were storedat 4øCor ambient
distances. Researchers, therefore, have routine-
We usedDNA fingerprinting to assessthe octemperaturefor about three months.Samplescolcurrenceof EPFsin a breedingcolonyof Leach's lectedduring the 1992seasonwere immediatelysusStorm-Petrel ( Oceanodromaleucorhoa).
METHODS
Studyspecies
andsite.--Leach'sStorm-Petrelsof both
sexesusuallydelay breeding until four or five years
of age, then breed yearly for up to 30 years (C. E.
Huntington unpubl. data). The female lays a single
egg each year and the male and female shareincubationdutiesduring a 40- to 44-dayincubationperiod
(Gross 1935, Wilbur 1969). Adults alternate incuba-
tion bouts, during which the adult at the nest fasts
for up to sevendays(œ= 3.1 days;R.A.M. and C.E.H.
unpubl. data), losing up to 7.5% of its initial body
massdaily while its partner foragesat sea (Ricklefs
et al. 1986).Incubationis coordinatedby the pair such
that the egg is rarely left unattended.
The nestlingisbroodedfor about5 days,afterwhich
it remainsalone in the burrow for 55 to 65 days.The
nestling is fed during brief nocturnal visits by its
parentsreturning from feeding areasmany kilometers out to sea. Leach's Storm-Petrels
are surface feed-
ers, preying on euphausidsand other zooplankton,
which are concentratedin areasof upwelling and
alonglocal current clines(Haney 1985,Brown 1988,
Pittman and Ballance 1990). Such oceanic feeding
zones are unpredictable in space and time (Brown
1980, 1988,Duffy 1989).Thus, adult storm-petrelsreturn to the nest sporadically,though typically a parent returnsevery secondor third night until the chick
fledges(Ricklefset al. 1985).Parentalcare in Leach's
Storm-Petrellasts100to 110daysfrom egg laying to
fledging.
We sampledfamilies from the breeding colonyof
about2,000pairsof Leach'sStorm-Petrelsat the Bowdoin CollegeBiologicalStationon Kent Island,New
Brunswick,Canada(66ø45'W,44ø35'N).We captured
each adult storm-petrelin its burrow in late June or
early July (during incubation),or in late August or
early September(during the provisioning period).
Presencein the burrow during either of theseperiods
pendedin 1 ml of a lysisbuffer (100mM Tris, pH 8.0,
100mM EDTA, 10 mM NaC1,0.5%SDS;Longmireet
al. 1988),which requiredno further handling.
DNA was extracted from 144 blood samples representing44 completefamilies (matesplus their single chick) and 6 partial families. Extractionbegan
with addition of proteinaseK (200 •g) and SDS (to
0.8%)to each of the preservedblood samples,which
were then incubatedat 55øCovernight.Subsequently,
four extractionswere performedon the 1991samples:
one with phenol; two with 25:24:1phenol:chloroform: isoamylalcohol;and one with 24:1chloroform:
isoamylalcohol.Five extractionswere performed on
the 1992samples:two with phenol; two with 25:24:1
phenol: chloroform: isoamyl alcohol; and one with
24:1chloroform:isoamylalcohol.Following the last
extraction, the aqueous phase was dialyzed extensively againstTNE2(10 mM Tris, pH 7.4,10 mM NaC1,
2 mM EDTA). Concentrationsand purlties of extracted DNA were assessed
by spectrophotometry.
Optical
densitieswere determinedat wavelengthsof 260 nm
(for nucleicacids)and 280nm (for proteins).Estimates
of DNA concentration and assessments
of DNA purity were then corroboratedby running 1.5 •g of
undigestedDNA from eachindividualthrougha 0.8%
agarosegel at 80 V for about2 h.
Gels were constructedof arbitrary assortmentsof
families, with the three membersof completefamilies
(Fig. 1), or two membersof partial families situated
in adjacentlanes. DNA (5 •g) from each individual
was digestedwith 5 x excessrestriction endonuclease
(HaeIII) at 37øCfor about 3 h. Resultingfragments
were separatedthrough a 0.8%agarosegel (22 cm) at
20 V for 64 to 65 h (until all fragmentssmallerthan
1,600basepairshad been run off the gel). Fragments
were then transferredto nylon by Southern blot in
10x SSC buffer and were fixed to the membrane by
UV crosslinking.Jeffreys'multilocus minisatellite
probe33.15 (Jeffreyset al. 1985b,c) was radiolabelled
by primer extensionwith [32P]dCTP.Hybridizations
were run overnight at 62øCin 1.5x SSC, 0.1% SDS,
qualified an individual as a putative parent. We col-
5 x Denhardt's solution, and 6% w/v dextran sulfate.
lectedbloodfrom chicksduring the four weeksprior
to fledging.Wesampled15full families(female,male,
and offspring)and 6 partial families (male and offspring)during the 1991breedingseason,and 29 full
families during the 1992 season.
Hybridized filters were put through four washesof
DNA methods.--Two50-•1blood sampleswere taken
from eachbird by punctureof the brachialvein. During the 1991field season,eachsamplewas immediately suspendedin 1 ml of phosphate-bufferedsaline
(PBS;3 mM KC1, 8 mM Na2HPO4, 2 mM KH2PO4, 0.14
at least 30 min each at 62øC in 1.5x SSC, 0.1% SDS.
Filters were then exposedto x-ray film at -20øC for
at least 50 and up to 212 h (usually with an intensifying screen).A secondhybridization,usingJeffreys'
multilocus minisatellite probe 33.6 (Jeffreys et al.
1985b, c), was run under the same conditions and a
secondset of autoradiographs
produced(Fig. 1).
Band-matching
analysis.--Werecorded the number
of bands in a chick's lane that were not attributable
April 1995]
Monogamy
in Leach's
Storm-Petrel
475
Family Number
I
M
C
2
F M
C
3
F
M
F
4
C
C
P
P
12-
Fig. 1. Multilocus minisatellite DNA fingerprints of 12 Leach'sStorm-Petrelsfrom four completefamilies
(both parentsand single chick) using Jeffreys'probe 33.15 and enzyme Hae IlL Lettersindicate:(M) putative
male parent, (C) chick;(F) putativefemale parent;(P) putativeparent of unknown sex.
to either of the two adults under consideration.
Since
all bands(or nearly all; seebelow) in a chick'sfingerprint should be accountedfor by the combined
fingerprintsof its parents,we comparedthe banding
pattern in eachchick'sfingerprint with that of two
sortsof dyadsof adults.The firstof theseband-match-
ing comparisons
involvedrecordingnumbersof bands
that were unattributableto a chick'sputativeparents.
The distribution
of such unattributable
bands across
the sample(Fig. 2) allows calculationof the rate of
mutation (i.e. rate of appearanceof bands unattributableto parentaldyads;Jeffreyset al. 1988,Raben-
476
MAUC•C,
WAITE,
ANDPARKER
old et al. 1990,Westneat1990).We applied the Poisson
40
distributionfunctionto the frequencydistributionof
35
the number
of bands that could not be attributed
to
either of the putativeparents(Wilkinson 1989).Thus,
we evaluatedthe probabilitiesassociated
with various
[Auk,Vol. 112
30
25
numbers of unattributable bands (Weatneat 1990) to
arrive at a criterion
number
of unattributable
bands
for excluding a putative parent.
The secondband-matching comparisoninvolved
recording the number of bands in a chick's fingerprint that couldnot be accountedfor by the combined
fingerprints of one putative parent and one outside
adult (i.e. a putative nonparent from outside family
triad). Where the sexesof the membersof matedpairs
were known, we selected an "outside adult" that was
of the oppositesexfrom the putativeparent.To insure
independenceof the data, we randomly choseone
putativeparent from eachfamily for thesecomparisons.To minimize error in scoringdue to the distance
between the lanes under comparison(Piper and Parker Rabenold1992),we alwayschosean outsideadult
whose
lane was within
five lanes of the focal off-
spring's lane, with the restriction that no lane was
used more than once as an outside
bution
adult.
of bands that were unattributable
The distrito either
of
the two adults (Fig. 2) was then fitted to a normaldistribution function, which permitted the evaluation
of the probability that an outside adult could be misidentified as a parent.
To evaluate
the likelihood
that two or more novel
bandswill arisein an offspring,we first assessed
the
fit of the distribution of bands unattributableto putative parent dyads to a Poissondistribution (twotailedKolmogorov-Smirnov
two-sampletest,P > 0.30).
Assuming mutations occurred randomly acrossindividual offspring and loci, the Poissonprobability
was0.013 that the combinedfingerprintsof the two
geneticparentswould fail to accountfor two or more
bandsin a chick'sfingerprint, where the mean occurrenceof novel bands per chick, taken as the expectedproportion,was0.17 (7 novelbands/42chicks).
Similarly, to evaluatethe probability that at leasttwo
bandsin a chick'sfingerprintwouldbe unattributable
to the combinedfingerprintsof one putative parent
and one outside adult, we first assessedthe fit of the
distribution of bands unattributableto putative parent-outsideadult dyadsto a normal distribution (normal probability plot, r2 = 0.99, P < 0.001). The lower
95% confidence limit of this distribution was 1.3 un-
attributablebands;therefore, only rarely would the
combinedfingerprintsof one putativeparent and one
outside adult account for all or all but one of the bands
in a chick'sfingerprint.
1
2
3
4
5
6
7
8
Number of Unattributable Bands
Fig. 2. Frequencydistributionof numberof bands
in fingerprintof eachof 42 Leach'sStorm-Petrelchicks
unattributableto combinedDNA fingerprintsof putative parents (solid bars), and combined DNA fin-
gerprintsof one putativeparentand an outsideadult
(shaded bars).
freys' 33.6). We consideredan outcomeof one or no
unattributablebands in a chick'sfingerprint, summing acrossprobes,to be an adequatecriterion for
the assignmentof parentage.
Band-sharing
analysis.--Wecalculatedthe proportion of bandsshared(X,) by eachpair of individuals
within eachfamily as
X• = 2S/(2S + A + B),
(1)
where S is the numberof bandsof indistinguishable
mobility and similar intensity in the two lanesunder
comparison,A is the numberof bandsunique to one
member of the dyad, and B is the number of bands
unique to the other memberof the dyad (Wetton et
al. 1987).The proportionsof bandssharedby dyads
of unrelatedindividuals,termed"background"
band
sharing,can be used as a basisfor comparingthe
similarityof bandingpatternsof individualswith variousdegreesof relatedness
(Georges
et al. 1988,Lynch
1990,ParkerRabenoldet al. 1991,Piper and Parker
Rabenold1992).We took the distributionof the proportionof bandssharedby matesto be representative
of the backgroundband sharing in the population
sampled(seeDiscussion).While all putative parentoffspringband-sharingproportionswere calculated,
onerandomlychosendyadfrom eachfamilywasused
to calculatethe distribution to insure independence
of the data. To corroboratethe assignmentof parentagebasedon the band-matchinganalysis,we then
compared the distribution of between-mate band-
sharingvalueswith the distributionof putativeparent-offspringband-sharingvalues.We performeda
For any casein which one unattributable band was
paired t-test to determine whether the between-mate
found in the band-matchinganalysis(either within
putativefamily, or within triad composedof putative
parent and an outside adult), we analyzed band
matchingin the correspondingautoradiographproducedby hybridizationwith the secondprobe (Jef-
band-sharingvalues and the putative parent-offspringband-sharingvalues(averagedover both parent-offspringdyads)representdistinctpopulations.
Probability
of assignment
errors.--Usingthe results
from boththe band-matchingand band-sharinganal-
April 1995]
Monogamy
in Leach's
Storm-Petrel
TAnrs 1. Calculations of probabilities of errors in
assignmentsof parentage based on DNA band-
sharingin familiesof Leach'sStorm-Petrelsusing
HaeIII and Jeffrey'smultilocusprobe 33.15.
Parameter
and definition
Estimate
f Mean number bandsscoredper lane.
21.4
œ Mean proportion of bandssharedbetween
mates.
0.58
q Allele frequency, where x = 2q - q2.
0.35
m Expected number of maternally derived
bands,whereto =f(1 + q- q2)/(2- q).
Expected number of paternally derived
bands= f - m.
s Expectedproportionof bandssharedby siblings, where s = (4 + 5q - 6q2 + q•)/
4(2 - q) (Jeffreyset al. 1985a).
Pu Probability of misassigningunrelated bird
as either parent, where Pu = x'.
PR Probabilityof misassigningfirst-orderrelative of parent as parent, where PR= s'.
16.0
e
5.4
477
Inclusionof the 6 offspring-fatherpartial families
with the 42 completefamiliesin our analysesof monogamyin Leach'Storm-Petrelrequiresthe assumption of no broodparasitism,a reasonableassumption
given that storm-petrelslay a single egg. Any egg
laidin a burrowpriorto incubationpresumably
would
berejectedby theresidentfemale.Duringincubation,
vulnerabilityto broodparasitism
wouldbe restricted
to periodsof egg neglect,requiringeither removal
of the original egg, or placing the new egg next to
the original.Consequently,
we wouldexpectto find
burrowscontaining two or more eggssimultaneously,
a situation documented only six times in 35 years
(> 11,000burrow-years;C. E. Huntington pers.comm.)
of study on Kent Island. Broodparasitismthus has
no appreciableeffecton the matingsystem,and the
inclusionof the sixfather-offspringdyadsin our anal0.052 ysesseemsjustified.
0.77
0.24
RESULTS
Band matching.--The distributions of the
yses, we computed probabilities of assigning the
wrong individualsasparents.The meanband-sharing
scorefor dyads of unrelated individuals (œin Table
1) wasusedto derive,the meanallelefrequencyacross
the family of loci screenedby the probe,which allows
calculation of the probability of misidentifying individuals as parents (Table 1; Jeffreyset al. 1985a,
Georgeset al. 1988). We calculatedthe probability
that the fingerprint of an unrelated male could account for all of the exclusivelypaternal bands (Pu in
Table 1). The calculatedprobability (0.052) of misidentifyingan unrelatedmaleasthe fatheris likewise
the probabilityof misidentifyingan unrelatedfemale
as the mother. We also calculatedthe probability of
misassigningas a parent somecloserelative of the
actual parent, such as a brother of the actual father.
Thus, we calculatedthe probability that the fingerprint of an uncle could account for all of the exclusively paternalbandsin the fingerprint of his niece
or nephew (0.24). This probability should be multiplied by the incalculableprobability that the uncle
would be caught in his brother's burrow (see Discussion).
Finally, we evaluatedthe probability of finding no
evidenceof extrapairparentagein 42 completefamilies, aswell asthe probabilityof finding no evidence
of extrapairpaternity in those42 families plus 6 partial families. Specifically,for a range of possibleincidencesof extrapairparentagein the population,we
calculatedthe probability of excluding none of the
nest attendentsas geneticparent (E) as
E = 1 - QN,
(2)
where Q is the proportion of chickswhose putative
parentsare the actual parents and N is the number
of families.
numbersof unattributablebandsshown in Figure 2 includethosefor dyadsof putativeparents
and for dyads composedof a putative parent
paired with an outsideadult. For all 42 complete
families, the number of bands in the chick's
profile that could not be accountedfor by the
combined fingerprints of putative parent-outsideadult dyads(median= 4) wasgreaterthan
the number
of bands that could not be account-
ed for by the combined fingerprints of the
chick's putative parents (median = 0; binomial
P < 0.001)
Among the 42 chicksfrom completefamilies,
putative parentsaccountedfor every band in
35 chicks'
lanes and all but one band
in the
remaining 7 chicks' lanes. To corroborateour
assignmentof parentage in these seven cases,
we analyzed band matching in those families
usingthe 33.6 hybridization. No unattributable
bandswere found in any of the seven cases.
We found
three cases in which
the combined
fingerprintsof one putativeparentand oneoutside adult accounted for all but one of the bands
in a chick'sfingerprint(Fig.2). In all threecases,
more than one unattributable
band was found
(3, 3, and 2) when the 33.6 hybridization was
analyzed. None of these three casesinvolved
any of the sevenchickswhoseputative parents
accounted
for all but one band.
With 7 of the 42 chicks exhibiting one unattributableband,the averagemutationratewas
calculated to be 0.17 novel bands/chick. Since,
on average,21.4fragmentswere scoredfor each
478
MAUCK,WAITE,ANDPARKER
0.4-
0 .4' 0• m•5m
0•5m• )0•5' 0'7m
0•5' • [0•5) • m
0.•5
' lm.O
Band-sharing
Score
N= 42
ß N=48
•
0,0
0.00 0.;2 0.;)4 0.'06 0.•)8
Fig.3. Proportionof bandsin DNA fingerprintof
each of 42 Leach's Storm-Petrel
[Auk, Vol. 112
chicks shared with
thosein one (randomly chosen)putative parent'sfingerprints (shaded),and proportion of bands shared
between mates(diagonal).
Hypothetical
Inciclence
of EPFsin Population
Fig. 4. Probabilityof finding no evidenceof extrapairparentage(i.e. zero exclusions)
over rangeof
hypothetical
levelsof extrapairparentage
(i.e. 1 - Q;
seeMethods)in Leach's
Storm-Petrel
population,giv-
chick (f in Table 1), the per-bandrate of occurrence of novel bands was 0.17/21.4 or 0.008 mu-
en samplessizesof 42 (completefamilies) and 48 (42
completefamiliesplus 6 partial familieswhere DNA
tationsto new-lengthallelesper locusper mei-
fingerprintsof chickanditsputativefatherwerepro-
otic event. This rate is similar
to those at mini-
duced).
satellite loci in other species(see Decker et al.
1992 and references therein).
Bandsharing.--Figure3 shows the distributions of band-sharingscoresfor dyads of firstorder relatives (proportion of an offspring's
bands that also occurred in each parent's fingerprint; • = 0.78 + SD of 0.066) and dyadsof
presumablyunrelatedindividuals (the 42 mated pairs;• = 0.58 + 0.059).Evenwith the 58.1%
level of backgroundband sharing (• in Table
bined fingerprints of that chick's putative parents (Fig. 2). In four of the seven families, both
putative parents had parent-offspring band-
sharing values exceedingthe 99% upper confidencelevel (0.72) for the proportionof bands
sharedbetween presumablyunrelated individuals (range 0.74-0.86). In each of the three re-
mainingfamilies,one putativeparent-offspring
dyad exceededthe 99% upper confidencelimit
and the remaining putative parent-offspring
ficientlydistinct(Fig. 3) sothat the probability dyadsexceededthe 80%upper confidencelimit
of misassigninga nonrelative asa putative par- for the proportion of bandssharedbetween preent is low (P = 0.052; Table 1). In addition, the sumablyunrelated individuals (83, 92, and 98%;
band-sharing score for every parent-offspring correspondingto the 2, 5, and 15%lower condyad was greater than for the corresponding fidence limits for related individuals). Therebetween-parent dyad, and the mean difference fore,while the band-sharingdatafor thesethree
between thesescoressuggeststhat thesedyads individuals are equivocal, there is no compelwere drawn from distinctstatisticalpopulations ling reason to exclude them as parents.
Finally, we used the resultsof the band-shar(paired t-test; t = 27.18, P < 0.001, one-tailed).
Five of the 84 putative parent-offspring dyads ing analysisto evaluatewhether extrapairpahad band-sharingscoresbelow the upper limit ternity might have occurred in any of the six
of the 95% confidence interval (0.678) of the partial familiesfrom which blood sampleswere
distribution of between-mate band-sharing collected. The proportion of bands shared bescores. Those scores involved
an adult whose
tween the putative father and the chick (• =
fingerprint together with that of its mate ac- 0.80, range 0.77-0.85) exceededthe upper 99%
counted for either all (three cases)or all but one confidencelimit for presumablyunrelated in(two cases)of the bands in the chick's finger- dividuals (0.72) in all six cases.Thus, we found
print.
no evidence of extrapair paternity in six partial
We usedthe resultsof the band-sharinganal- families.
ysisto corroborateour assignmentof parentage
Probability
ofassignment
errors.--Figure4 shows
1), first-order relatives and nonrelatives are suf-
in those cases in which
one band in the chick's
fingerprint could not be attributed to the com-
the probabilityfunctionsfor excludingnone of
the putative fathers given sample sizes of 42
April 1995]
Monogamy
in Leach's
Storm-Petrel
and 48. With a sample size of 48, if the true
level of EPFs in the population were 5%, the
probability that we would have excludednone
of the fathers is less than 0.085. As the hypothetical level of EPFsin the population increas-
es, the probability of finding no evidence of
EPFsquickly approacheszero given our sample
sizes.
DISCUSSION
Our results indicate
that Leach's Storm-Pet-
rels in our study population were strictly monogamous in that breeding partners were the
geneticparentsof the young they provisioned.
This conclusiondependson the low probability
of misassigning nonrelatives as putative parents (P = 0.052; Table 1). If full siblings, or
parentsand their adult offspring,tended to nest
in closeproximity, this conclusionwould be
suspectsincethe probability of misassigninga
first-orderrelative was considerablyhigher (P
479
(Wilbur 1969, Mauck pers. obs.),the latter perhaps reducing its probability of survival to
fledging (Ricklefs and Schew 1994), or to maturity (cf. Perrins et al. 1973). Parental care apparently is unshareable,meaningthat one procellariiform parent cannot divide its care between two offspringsuccessfully.
No caseshave
been reported of a Leach'sStorm-Petrelparent
simultaneously raising two broods and, in
brood-enlargement experiments, both the
Leach'sStorm-Petrel(Huntington 1963)and the
closely related Fork-tailed Petrel (O. furcata;
Boersma et al. 1980) failed to raise more than
one chick.
Wittenberger and Tilson (1980) defined the
conditionsthat would seemto favor monogamy
over other mating systems. In view of the
mounting evidence for EPCsand EPFsin birds,
Wittenberger and Tilson's model has been expanded to incorporate the idea that the conditions that promote social monogamyalso promote genetic monogamy (e.g. Westneat et al.
= 0.24; Table 1). On Kent Island, however, natal
1990,Decker et al. 1992).In particular, Birkhead
philopatry to the islandasa whole is rare (< 1% and Moller (1992) suggestedthat when reof more than 10,000bandedchicks;Huntington sourcesbecome so unpredictable in spaceand
and Mauck in prep.). Thus, we are confident time that two parents are required to raise offthat our resultsare not confoundedby the pres- springsuccessfully,the pursuit of EPCsis greatence of breeding first-order relatives. If EPFs ly devalued. This argument assumesthat puroccurredat all in our study population, they suit of EPCs carriesa prohibitive opportunity
cost(i.e. allocating time to extrapair mating efapparentlydid so only rarely (Fig. 4).
In the population of Leach'sStorm-Petrelswe fort at expense of parental effort would critistudied,the breeding system(geneticmonog- cally reduceprobability of intrapair reproducamy) matched the mating system (social mo- tive success).
nogamy). Wittenberger and Tilson (1980) preThis assumptionmay not be valid for Leach's
dicted social monogamy when male parental Storm-Petrel.Procellariiformchicksexperience
careis "both non-shareableand indispensable," long and irregular periods of fasting between
and when males are more successful with one
parental visits (Ricklefs et al. 1985, Warham
mate than with two. Long commutesto ephem- 1990). Thus, it is not clear how spending relaeral food supplies,making food difficult to ob- tively small amounts of time pursuing EPCs
tain and deliver, are conditions under which
would affectchickgrowth and survivorship.Albiparental care may be obligatory(Wittenber- though we have provided evidence that EPFs
ger and Tilson 1980,Westneatet al. 1990).These are rareor absentin our studypopulation,nothconditions, therefore, are associated with moing is known of the occurrence of EPCs in
nogamous mating and long-term pair bonds Leach's Storm-Petrel. EPCs have been observed
(Mock and Fujioka 1990).
in another procellariiform, the Northern FulThe ecology'ofsmallprocellariiformsfits this mar (Fulmarusglacialis;Hatch 1987, Hunter et
description. Biparental care may be indispen- al. 1992).Hunter et al. (1992) reportedthat 7.2%
sable, meaning that a single parent probably of all observed copulations in one fulmar colcannotsuccessfullyraisea chick. Incubation by ony were EPCs,although they found no evione parent alone would result in so much ne- denceof EPFsin 85 families. A simple extension
glect of the egg that hatching probably would of Wittenberger and Tilson's model (1980) to
not occur (Gross 1935, Wilbur 1969, Boerstoa include genetic monogamy seemsinadequate
and Wheelwright 1979). A chick that loses a to explainthe fulmar dataor, by extension,the
parent either dies or gains massvery slowly storm-petreldata.
480
MAUCK,
WAITE,
ANDPARKER
A more parsimonious explanation for our
failure to detectany EPFsis that last-spermprecedenceand frequent copulationby matesduring the female'sfertile period may be sufficient
to insure a high probability of fertilizing the
single egg laid by the female. This mechanistic
explanationis promptedby recentfindingsfrom
work on other avian species.In Zebra Finches
(Taeniopygia
guttata),the last male to copulate
with
a female
fathered
84% of that
female's
chicks, even when that male copulatedonly
once(Birkheadet al. 1988).Thus,merely being
the lastmaleto copulatebeforefertilizationmay
greatly increasea male's probability of paternity. In domestic fowl, Martin et al. (1974)
[Auk,Vol. 112
paper,the strategicimplicationsof smallclutch
size merit careful
theoretical
attention
in future
work.
ACKNOWLEDGMENTS
This work wassupportedby the College of Biological Sciences,the Ohio StateUniversity, and in facilitiessupportedby a NationalScienceFoundationgrant
to P.G.P.(BSR9111848).
T.A.W wassupportedin part
by an NSERC of Canada International Post-doctoral
Fellowship.R.A.Mwassupportedin partby theFrank
M. ChapmanFund of the American Museum of Natural Historyanda CooperOrnithologicalSocietyPaul
A. Stewart Award. We thank Nancy Yorinks, Katherine Gill, Claudine Cordray, and Tigerin Peare for
technicalassistance.
We thank Chuck Huntington for
his adviceand support.This is publicationnumber
118 from the Bowdoin College BiologicalStationat
showedthat the percentof young (in a multipie-egg clutch) sired by individual males was
positivelycorrelatedwith the quantity of sperm Kent Island.
delivered. In the caseof the single-eggclutch,
this relationshipcouldbe interpretedto suggest
LITERATURE CITED
that a male'scertaintyof paternityincreaseswith
the quantityof spermhe delivers.Frequentcop- BIRKHEAD,
T. R., AND A. P. MOLLER.1992. Sperm
ulations during the entire fertile period and
competition in birds: Evolutionary causesand
consequences.Academic Press,London.
near the onset of ovulation might greatly increasethe probability of intrapair fertilization.
In their study, Hunter et al. (1992) showed that
BIRKHEAD,T. R., J. E. PELLET,AND F. M. HUNTER. 1988.
Extra-paircopulationand spermcompetitionin
the Zebra
Finch.
Nature.
334:60-62.
male Northern Fulmars closelyguarded their
D. D., ANDN. T. WHEELWRIGHT.
1979. Egg
matesduring the fertile period and immediate- BOERSMA,
neglect in Procellariiformes:Reproductiveadly followed any EPCs they appearedto detect
aptationsof the Fork-tailedStorm-Petrel(Oceanwith multiple copulations of their own.
A male Leach'sStorm-Petrelemployingthese
tacticsthroughoutthe prelaying period might
have a high probabilityof fertilizing his mate's
single egg. Such tacticsmay be sufficiently effective to severelyrestrictthe chanceof an EPF
in any given storm-petrelfamily and make unlikely the detectionof a single EPF in 48 families. We cannotsay whether geneticmonogamy in Leach'sStorm-Petrelreflectsthe rarity
(or absence)of extrapair copulations,the effectiveness of male sperm competition tactics,or
both. We can say, however, that for Leach's
Storm-Petrelsin this study population there
seems to be no distinction between apparent
and realized reproductive success.
Our result adds to an emerging pattern of
strict(genetic)monogamyin specieswith small
clutches(e.g. Leach'sStorm-Petrel,this study;
Northern Fulmar, Hunter et al. 1992; Black Vul-
tures,Decker et al. 1992).Among sociallymonogamousbirds,all reportedoccurrences
of EPFs
have involved specieswith clutchsizesgreater
than two. Although beyond the scopeof this
odromafurcata).Condor 81:157-165.
BOERSMA,D. D., N. T. WHEELWRIGHT,M. K. NERINI,
AND E. $. WHEELWRIGHT.
1980. The breeding
biology of the Fork-tailed Storm-Petrel (Oceanodromafurcata).Auk 97:268-282.
BROWN, R. G. B. 1980. Seabirds as marine animals.
Pages
1-31inBehavior
ofmarineanir•als
(J.Burger, B. L. Olla, and H. E. Winn, Eds.). Plenum
Press, New York.
BROWN,
R. G.B. 1988. The influenceof oceanographic anomalieson the distributionof storm-petrels
(Hydrobatidae) in Nova ScotJanWaters. Colon.
Waterbirds
11:1-8.
BURKE,
T., ANDM. W. BRUFORD.1987. DNA fingerprinting in birds. Nature 327:149-152.
DECKER,
M.D., P. G. PARKER,
D. J. MINCHELLA,AND K.
N. RAnENOLD.1992. Monogamy in Black Vultures: Genetic evidence from DNA fingerprint-
ing.Behav.
Ec01.
4:29-35.
DUFFY,D.C. 1989. Seabird foraging aggregations:
A comparisonof two southern upwellings. Colon. Waterbirds
12:164-175.
GEORGES,
M., A.-S. LEQUARRE,
M. CASTELLI,R. HANSET,
ANDG. VASSART.1988. DNA fingerprinting in
domesticanimals using four different minisatellite probes.Cytogenet.Cell Genet. 47:!27-131.
April 1995]
Monogamy
in Leach's
Storm-Petrel
481
GOWATY,
P. A., ANDA. A. KARLIN.1984. Multiple
portionsof offspring.J. Reprod.Fert. 39:251-258.
maternity and paternity in single broodsof ap- MOCK,D. W., ANDM. FUIIOKA.1990. Monogamyand
parently monogamousEasternBluebirds,Sialia
long-term pair bonding in vertebrates.Trends
sialis. Behav. Ecol. Sociobiol.
Ecol. & Evol. 5:39-43.
15:91-95.
GOWATY,P. A., AND W. BRIDGES. 1991. Nestbox avail-
abilityaffectsextrapairfertilizationsandconspecific nest parasitismin EasternBluebirds,Sialia
sialis. Anita.
Petrel (Oceanodroma
leucorhoaleucorhoa)on the
outer seaislandsof the Bay of Fundy. Auk 52:
currencesin relationto upwelling off the coast
of SE U.S. Wilson
Bull. 97:543-547.
HATCH,S. A. 1987. Copulationand mate guarding
in the Northern
Fulmar. Auk 104:450-461.
Hu•rrER, F. M., T. BURKE,AND S. E. WATTS. 1992. Fre-
quent copulationasa methodof paternityassurance in the Northern
Fulmar.
Anita.
Behav. 44:
149-156.
HUNTINGTON,
C. E. 1963. Populationdynamicsof
Leach'sPetrel,Oceanodroma
leucorhoa.
Pages701705 in ProceedingsXIII International OrnithologicalCongress(C. G. Sibley,Ed.).Ithaca,New
York, 1962. American Ornithologists' Union,
1985a.Positiveidentification
of an immigration
test-caseusingDNA fingerprints.Nature 317:818-
printing to determine relatednessin a tropical
wren. Mol. Ecol. 1:69-78.
PITTMAN,R. L., ANDL. T. BALLANCE1990. Daytime
feeding by Leach'sStorm-Petrelon a midwater
fish in the Eastern Tropical Pacific. Condor 92:
524-527.
RABENOLD,P. P., K. N. RABENOLD,W. H. PIPER,J.
HAYDOCK,ANDS. W. ZACK. 1990. Shared pater-
nity revealedby geneticanalysisin cooperatively
breeding tropical wrens. Nature 348:538-540.
RABENOLD,P. PARKER,K. N. RABENOLD,W. H. PIPER,
M.D. DECKER,
AND J. HAYDOCK. 1991. Using
DNA fingerprintingto assess
kinshipandgenetic
structurein avian populations.Pages611-620 in
IV InternationalCongressof Systematicand EvolutionaryBiology(E.C. Dudley,Ed.).Dioscorides
Press,Portland,Oregon.
RICKLEFS,
R. E., C. H. DAY, C. E. HUNTINGTON,AND J.
819.
JEFFREYS,
A. J., V. WILSON,AND S. L. THEIN. 1985b.
Hypervariable'minisatellite'regionsin human
Nature
314:67-73.
J•EREYs, A. J., V. WILSON, AND S. L. THEIN. 1985c.
Individual-specific
'fingerprints'of humanDNA.
316:76-79.
JEFFREYS,
A. J., N.J. ROYLE,V. WILSON,AND Z. WONG.
1988. Spontaneous
mutationratesto newlength
allelesat tandem-repetitivehypervariableloci in
DNA.
Nature
332:278-281.
LACK,D. 1968. Ecological
adaptations
for breeding
in birds. Chapmanand Hall, London.
LIFJELD,J. T., AND R. J. ROBERTSON.1992. Female
control of extra-pair fertilization in Tree Swallows. Behav. Ecol. Sociobiol.
31:89-96.
LONGMIRE,J. L., A. K. LEWIS,N. C. BROWN,J. M.
BUCKINGHAM,
L. M. CLARK,M.D. JONES,L. J.
MEINCKE,J. M. MEYNE,R. L. RATLIFF,F. A. R•Y,
g. P. WAGNER,AND g. K. MOYZIS. 1988. Isolation
andmolecularcharacterization
of a highlypolymorphiccentromeric
tandemrepeatin thefamily
Falconidae. Genomics 2:14-24.
LYNCH,M. 1990. The similarityindex and DNA fingerprinting. Mol. Biol. Evol. 7:478-484.
MARTIN,P. A., T. J. REIMERS,
J. R. LODGE,AND P. J.
DZIUK.
Ibis 115:535-548.
PIPER, W. H., AND P. PARKERRABENOLD. 1992. Use
The unity of evolutionarybiology.Proceedings
Washington,D.C.
JEFFREYS,
A. J., J. F. Y. BROOKFIELD,
AND R. SEMEONOFF.
human
Survival of Manx ShearwatersPuffinuspuffinus.
of fragment-sharing
estimates
from DNA finger-
382-389.
HANEY,J. C. 1985. Band-rumpedStorm-Petreloc-
Nature
trapair fertilizations and the evolution of colonial
breeding in Purple Martins. Auk 107:275-283.
PERRINS,C. M., M.P. HARRIS,AND C. K. BRITTON. 1973.
Behav. 41:661-675.
GROSS,
W. A.O. 1935.Thelife historycycleof Leach's
DNA.
MORTON, E. S., L. FORMAN,AND M. BRAUN. 1990. Ex-
1974.
The effect of ratios and numbers
of spermatozoa
mixed from two maleson pro-
B. WILLIAMS.1985. Variability in feeding rate
and meal
size of Leach's
Storm-Petrel
at Kent
Island, New Brunswick. J. Anita. Ecol. 54:883898.
RICKLEFS,
R. E., D. D. ROBY,ANDJ. B. WILLIAMS. 1986.
Daily energyexpenditureby adultLeach'sStormPetrelsduring the nesting cycle.Physiol.Zool.
59:649-660.
RICKLEFS,
R. E., ANDW. A. SCHEW.1994. Foraging
stochasticity
and lipid accumulation
by nestling
petrels. Funct. Ecol. 8:159-170.
SHERMAN, P. W., AND M. L. MORTON. 1988.
Extra-
pair fertilizations in Mountain White-crowned
Sparrows.Behav. Ecol. Sociobiol. 22:413-420.
SILVER,R., H. ANDREWS,ANt) G. F. BALL. 1985. Pa-
rental carein an ecologicalperspective:A quantitative analysisof avian subfamilies.Am. Zool.
25:823-840.
SMrrH,S.M. 1988. Extra-paircopulationsin the Blackcapped Chickadee: The role of the female. Behaviour
TRIVERS,R.
107:15-23.
1972. Parental investment and sexual
selection.Pages136-179in Sexualselectionand
the descentof man, 1871-1971(B.Campbell,Ed.).
Aldine, Chicago.
WARHAM,J. 1990. The petrels.AcademicPress,London.
482
M^UCK,W.qIT•,aNDP•,I•K•
WESTNEAT,D. F., P. W. SHE•M•,
•D
M. L. MORTON.
1990. The ecology and evolution of extrapair
copulationsin birds. Curt. Ornithol. 7:331-369.
WEStNEAt,D. F. 1990. Genetic parentage in the Indigo Bunting: A study using DNA fingerprinting. Behav.Ecol.Sociobiol.27:67-76.
WESTON,.J. H., R. E. C^R•rER,D. T. PARraN, •ND D.
WAITERS. 1987. Demographicstudy of a wild
House Sparrowpopulation by DNA fingerprinting. Nature 327:147-149.
[Auk, VoL 112
WILBUR,
H.M. 1969. The breedingbiologyof Leach's
Storm-Petrel, Oceanodromaleucorhoa.Auk 86:433442.
WILKn•$oN,L. 1989. SYSTAT: The systemfor statistics. SYSTAT, Inc., Evanston, Illinois.
WI•rENBERGE•, J. F., •a•D R. L. TInSoN. 1980. The
evolution of monogamy: Hypotheses and evidence. Annu. Rev. Ecol. Syst. 11:197-232.