UMI TO

NOTE TO USERS

Page not included in the original manuscript are unavailable from the author or university. The

manuscript was microfilmed as received.

18

This reproduction is the best copy available.

UMI

1.1

NationalLbrary ofCaoada Acquisitions and BibliographicServices 395W~Slr_

=,ONK1AOl'U

BibliotnequenationaJe du Canada Acquisitionset servicesbibliographiques

=C:~0N4

The author bas granted a non- exclusive licence allowing the National Library of Canada to reproduce. loan. distribute or seU copies of this thesisinmicroform, paper or electronic formats.

The author retains ownership of the copyrightinthis thesis. Neither the thesis nor substantial extracts fromit maybeprinted or otherwise reproduced without the author's permission..

L'auteur a accorde une licence non exclusive pennettant

a

la Bibliotheque nationale du Canada de reproduire, preter, distribuee au vendre des copies de cettc theseSOllS la forme de microfiche/film. de reproduction sur papier au sur format electronique.

L'auteur conserve la propriete du droit d'auteur qui protege cette these.

Nila

these

nides extraits substantiels de celle..cjne doivent etre imprimes au autrement reproduits sans son autorisatioD.

0-612-55126-1

Canada

EGG PRODUCTION IN THE TIUCK-BlLLED MURRE (Uria lamvia) AND RAZORBILL (Alca 'ordo); A LIFE-ffiSTORY PERSPECTIVE

by 1.Marl<_

A thesis submitted to the Schoolof GraduateStudies inpartial.fuIfilmentof the requiremcols forthedegree of

Doctorate ofPhilosophyinScicoee

June2000

StJobn's

ABSTRAcr

Akeytenetoflife--histmy theoryisthatcosts ofrepcoduction lead tophysiologicaland evolutiOIWy trade-offsamong fitness componeots. Althoughavian egg productionwuakeytopicinthe development of life-history theory, the significaoce of egg production cosuwithinthe life histories ofbirds remains poorly understood. Two aspects ofegg productionthathave received considerable attention are variation amoag femalesintheircapacityto reoest foDowingdutch loss, and variation in the size ofeggtheylay.Iexaminedtheseupcct5ofegg productionin two pelagic seabirds,Thick- biDed Murre:s Uria Iomvia andRazorbillsAlca torda.[tis geoeraIly thoughtthategg production costs arerdativdylowin pelagic seabirdsthatlaysingle-eggclutches., butag&and date-specific patternsinre:nesting capacityandegg size suggest:thatsigDificaDt constraints on egg production operateinthese birds.

The proportion ofThick-billed MWTCS that renests roUowing egg loss declineswiththe date ofloss.However, early-laying females that had their eggs removed continued to renest UDtilla1e in thelayingperiod.. FlCSt lllldreplaceme:n1eggs wen:similarinelUde colllpOSition. but replacemCllt eggs bad low proteinCODtcn1.Despitethis,replacement eggs~nolesslikdytohatchthanwere firsteggs. andc:bi.cksfromreplacement eggs wereDOless likely tosurvivetoDCSldeparture.and to recruitment age(4-5 years). I conclude that variationinthe egg-production capacity offcmalesthat lose their eggsearlyand late (probably age/experience effects)drivesthe seasonal declinesin renestingrates. Capable femaJes will relay uotillatein the layina periodbecause.forthese bi.rds, the poteotial fitness payofffrom a replacemcateggissimilartothat from afirstegg. It remains to be detc:nninedwbethCl" thereare survival costsassociated with the production of replaccmcnt eggsfoe femaleThick:-billed Murres.

Thick:-billed Murre eggs vary considerablyinsize, and eggsizeaffcetll offspring performance:

chicksfrom large eggs have theirwingfeathers grow more quickly than do those from small eggs.

Onehypothesisoften iDvoked to explaintheexistenceof'YUiationinegg sizein thefAce ofexpected directionalselectionforlarge eggs isthattheoptimalegg size varies withenviroomc:atalconditions;

thishypothesis predictsthatbmditsofhatcbiDgfrom a large eggwillbemagnifiedwhen

fcediDs

conditionsare unfavOU1'B.ble.I tested thisbycomparingbetweencoloniesthatexperience favourable (Coats Island) and unfavourable (Digges Island)conclitiOtlS,usinganexperimental. egg-switcbing protocoL Cootrary to prediction. theeffect of egg size00wing feather growthwasDOgreaterat Digges Islandthan at Coats bland.

This effectofegg size on wing-feathlC1" growth has not been detectedinothlC1"birds,suggesting thatitmightreflect adaptations to the unique "intcnnediate" developmental strategy employed by murresandRazorbillJ.Totestthishypot.besis, Iexamined the effect ofegg

me

onpost-batcbiog developmeotintheRazorbillwin&:thesameegg-switchingprotocolResults clearlysupportedthe hypotbesU:asinThiclt-billedMmres, Iarge-eggRazoIbillsaperieooedenhanc:edearlywing-feather growth.There are anumbez- of adaptationsinthe dcvdopmeDt of intermediate alcidsthatmight explainwhythisegg-size effectis readily detectable in these birds. k therewasnoevidence for a trade-offbetweeo egg size and provisioningineither Thick-billed Murres or Razorbill5,theexistence ofconsiderable female-specificvariationinegg size rema.ins unexplained.

iii

ACKNUWLEDGEMENTS

During my timeinNewfoundland,Ihave come into contact with some truly remarkable people and made some reallygood friends. Their WIWaVering suppan and frieodsbip bas meant more to me than they could possibly know.Thanksin partiwlarto everyoneinthe Biopsychology PCOgralI1lDC, and toallwho tolerated my incompctcnce during Sunday afternoon soccer games, Fridaygamesof pool at Bitter's,etc. For various reasons,Iowe special thanks toGailFraser,Martha Hiscock,Keith Lewis.Ian Stenhouse (opinionated, but JtOt coofrontational),aDd Carolyn Walsh.

Workatthevarious field siteswas always eujoyable because oftbe peoplewhoWeRlthere.

I am especially grateful to Rachel Bryant for the energy, enthusiasm, patience,andsense of wonder she brought to everyday life, and to ourresearch,onGC4.Thanlcsalso toShauna Baillie, Monica Kidd, Alejandra Nude:z de I. Mora,"i'umoPoldmaa, TonyRoberts, SbcnyIynnRowe,and Brian Veitch for their invaluable COntributiOWl at the Gannet Islands. Brian's stories of growing upin NewfoundJand willbe some of my favourite memories ormy time here ("So, Mary...got something tosay to ma?"). I am also deeply indebted 10 Charlie Cornwallis, David and John Geale, Kentin Kober. RichLemer", Jo5iah Nakoo!a.k,UliSleiner, andKerry Woo for all oftheir beJpand companionshipatCoatslsIandand DiggesIsland.Debbie Martin deserves specialthanks forcanying out the research presentedinChapter 2 atCoatsIslandin 1997.

For outstanding logistical supportin Labrador, manythankstoRandyCahill.Blair Gillis, TrishaKinsey, HatryMartinand Janet RussdI.. ForalpportatCoats,thanks tothe staffoftbe Nunavut Researchlostinrte in Iqaluit, theCanadianWildlife Service, and the Polar Continental Sbelf Project ofNatural~urcesCanada. Funding to carry out this research was providedbyMemorial University of Newfoundland, the Northen Scientific TrainiDg Program,the AtlanticCoopentive Wddlife Ecology Research Network, the Canadian Wddlife Service, and the Natund Sci.cocesand Engineering Research Council of Canada (through grants to Anne Storeyand Tony Gaston).

iv

Tha.nksalso toDrs.SeanBrosnanandGeneHerzbergoftbe Department ofBiochemistryat Memorial University fortheir expertise, and to Donna Hunt and Craig Slcinner for doing the protein andaminoacid analysesontheThick~illedMurreeggs.

Ialwaysreceivedoutstanding support from my thesis committee -thankstoJohnCbardine for help with statistics in particular, andthanks to Bill Mootevecchi forunfailingenthus.iasmand for sharing his knowledge of marinebirds.Finally, I couldn't begin to express my gratitude to mythesis supervisor's, Anne Storey and TonyGaston.The support IrcccM:dfromthem throughoccasionally diffiOJlttimeswuaboveand beyondthe call, andIam verygratefulInaddition, theywere always available to make invaluable contributionsto this research.

TABLE OF CONTENTS

_.

Acknowledgments . Table of Contents . List of Tables . List ofFiguces .

ChapterI -Introductionand Overview 1.1 Literature Cited 1.2 Co-authorship Statement

. ... ii

• (y

. ... vi

.. ,w ... 1 .7 .10 Chapter

n -

SeasonalDeclinesin ReplacementEgg-Laying!inaLong-Lived,ArcticSeabird:

CosuofLateBreedin.gorVariationinFemaieQuality? . . II

2.1 Abst:raet. ..•. .11

2.2 Introduction ..

2.3 Study Area and Methods 2.3.1 Data analysis . 2.4 Results .

2.4.1 Timingofbreeding .

2.4.2 Date ofegg Loss and probability ofreplaccment laying .

..12 .14

. 16

.17 .17 17 2.4.3 Probability of relaying in relation to egg size and incubation period .... 18 2.4.4 Relaying interval and replacement egg size .

2.4.5 Reproductivesucoessofexperimeotalpain.

2.5 Discussion.

.22 ..••.. ..•.••.. .25 .25 2.5.1 Seasonal declinesinreplacementlayings: effects ofetate or quality? . .2S 2.5.2 Eggsize, incubation period, and probability ofreLaying . . 30

vi

2.5.3 RelayingiDterval andreplacementegg size ... 31 2.6 LiteratureCitcd ...•..•... 34 Chapter

m -

Fitness Consequences ofRenestiog for Thick-billed Murres:

SurvivalofOffspring to ReauitmeDt Age ..

3.1 AbstnJct.

Introduction ..

3.3 Study Areaand Methocb . 3.4 Results .

3.5 Discussion.

4\

. 41

. 42

. 42

. 46

. 48

4.2 lotroduction.

3.6 Literature Cited

ChapterIV -CompositionofFint aDdReplaoememEggs:

ConstraintsandStntegies for Rmesting lbick:-billed Murres 4.1 Abstract.

4.3 Methods.

... 52

.56 .56 ..57 . .... S9 4.3. 1 Fieldmethods .

4.3.2 AnalysisofClUde composition.

4.3.3 Analysis of biochemical composition.

4.4 RelUlts

4.4.1 Crude composition . 4.4.2 Biochemical composition . 4.5 Discussion.

. S9

. 60

..60 ... 61

. 61

. .. 64 ... 67

4.5.1 CompositionofThick-billed Murre eggs . . 67

4.5.2 Constraints and strategies for renesting Thick-billed Murres . . . 73

4.6 literatueeCited. . 75 Chap~V - Food Supply and the Consequences of Egg Size for lbick-billod Murres 80

5.1 Abstract. .80

5.2 Introduction ..

5.3 Study Area and Methods 5.3.1 Dataanalysis.

SA Resulu.

5.4.1 Experimentalprotocol andbctween~lonydifferences .

. 81

. . . 83 . ... 85 . ... 85

. 85

5.4.2 The relationships between Initial and Foster volumes and growth .... 86

5.5 DiscussioQ. . 89

5.5.1 Egg size and post-batcbing development: intcr-c:olony compariJon ...91

5.5.2 Consequcoccsofcggsizeforadults. . .. 92

5.6 Literature Cited .. . .... 95

Chapter VI - The Effect of Egg Size on Post·Hatching Development in the Razorbill:

AnExperimentalStudy. . 99

6.2 Introduction.

6.3 Study AreaandMethods 6.4 Results ..

6.4.1 Inter-year differences and experimental protocol .

6.1 Abstra.ct ...•••. .99

. 100

... 101 .102 .102 6.4.2 Egg-sizeeffectsonbatc:blingmassandsize . . 105 6.4.3 The rel.ationshipbetweenchick gro'Wthand FosterVohune 105 6.4.4 The rdationshipbetweenchickgrowth and Initial Volume 110 6.5 Discussion.

viii

110

... 110 . ... 111 .113 .... 113 . ... IIS . ...•... 119 . ... 128 6.S.1 Eggsizeand post.hatchiDg developmeminlUzorbills

6.S.2 Evolutionary implications . 6.5.3 Does eggsizereflectfemalequality? . 6.S.4 Cooc:tusions

6.6 Literature Cited . Chapter

vn .

Summary .

7.1 Litenmre Cited .

ix

LIST OF TABLES

Table 2.1: Results of a forward stepwise logistic regression analysisClWDining variationinthe proportion ofcontrol Thick-billed Murre pairsthatrelaid following oatural egg loss(n -200). The constant was eDtered into the model.

Variables available to the model were date of egg lou, colony-year, and the date • COlOny-year interaction tena Selection of variables WlU bued onthe changeinthelikelihood-ratio with the variable deleted from themodel.Variables enteredthe model atP<0.05,andwell!removed. from the model atP>0.1 19 Table 2.2: Results ofaforwacdstepwiselogistic regressionanalysisc:xamiDingvariation

in the proportion of control(n - 79) andexperimental(n "" 83) Thick-billed Murre pain that relaid followingnatural egglossoreggremoval during the period of overlapbetweenthetwogroups. The constantwasentered into the model Variablesavailable to the modelweredateofegg lou(grouped bythree-day intervals), colony-year,treatment(controland experimental), andalltw().way and theone three-way interaction terms. Selection of variableswasbued on the change in the likelihood-ratio with the variable deleted from themodel.Variables entered the model atP<0.05, and were removed. from the model atP> 0.1 . . 21 Table 2.3: Regressions relatingthe relaying interval to: (I) the date of removal offirst

eggs, and (2) thenumberof daysthatthefirsteggs were incubated prior to removal (covariatesinthe ANCOVAs), forexperimentalThick-billed. Murresat the GannetlsIand5 (n -24) and Coatsldand (n -33) (grouping variablesinthe

ANCOVAs)in 1997. . 23

Table 3.1: Tuning ofnest departurein1994 and 1995for fint-egg(n'" 58 and 57, respectively) and replacement-egg(n -26and 18) Thick-billed Murre chicks, and

the proportions of these birds resigbted at the colonyin1999 ...••. .47

65 ... 62

..70 Table 4.1: Crudecomposition (mean!: SO) of24 first and 24 replacement

Thick.-billcd Murre eggs !aidby the same femalesatCoats Island Table 4.2:Analyses of covariance (ANCOVAI) rdating log componeD1 mass

tologeggmass for 24firstand24 replacement Thick-billed Murre eggs laid by the same femalesatCoats Island .

Table4.3: Amino acid content (mean!: SO,in J.UDg.1protein) ofsevenfirstand

seven replacement Thick-billed Murre eggs laid by the same females . . . 68 Table4.4: Amino acid content(g) ofseveofirstand seven replacement

Thick:-billedMurre eggs laidbythe same females. . 69

Table4.5:Proteinandlipidconteo1(g)ofsevenfirstandseven.replaoemeat Thick-billed Murre eggs laidbythe same females .

Table 5.1: Measures (mean±SO)ofThick-billed Murre egg.size and chick.growthat Digges and Coats lsIand.! (Coats data from Hipfoer and Gaston 1999a) . . . 87 Table5.2:Multiple regressions celating measures ofchick growth tol:oitiaI and Foster

volumes. For clarity, significant results areinbold

Table 6.1: Comparisons (mean±SO) of egg size andgrowthofthe Razorbill cbicks inthe twoyean.None afthebetween-yeardifferencesweresignificaDt (allr:5o:n. all P2:;0.48).Coefficientsofvariation(CV) are reported with the two years combined .

Table 6.2: RelationshipsbetweenFostcI" Volume aDd growthofthe RazorbillChien Datafrom1996 and 1997werepooled .

••

.104

.107

LIST OF FIGURES

Figure 2.1: PCI"CCltofcontrol (dotted line) and experimeotal (dasbed tine) Thick:-billed Murre pairs thatreplacedlosteggs in relation to the date ofegg loss or removal (grouped by3 day intervals, in days from the median laying date) in thethree colony-years. Numbers besidedata points are sample sizes . . .. ; ... 20 Figure 2.2: Relaying intervals ofThick-billed Mwres in relation to the number ofdays

that firsteggswereinwbated prior toexperimentalremoval Datafromthe GannetIslandsaDd Coats Island in 1997 were pooled. . ... 24 Figure 2.3:Hatchingsuccess(lIS), 6edgiug success (FS) and reproductive success(as)

of control (open bars) aDd experimental (batcbed bars) Thick-billed Murrepain at the Gannet Islands in 1997. Controlpairs were those that laid by the date of the firstexperimental egg removal. All measures of success were similarbetweenCODttOu andexperimentalpairs.Numben aboveban are sample sizes . . . .. 26 Figure 3.1: Percentage of birds resightedinrelation to their date of departure forconttol

chicks, andthe proportion ofexperimental chicks resighted.. Sample sizes above

bon . .49

Figure 4.1.Relativemasses of whole eggs and egg components for firstandreplacement Thick·billed Murre eggs.Thedashed tines indicate I: I relationships . . .... 63 Figure 4.2. Yolk to albumen ratiosin firstand. replacement Thick-billed Murreeggs.

The dashed lineiodicatesa 1:1 relationship. 66

Figure 5.1: Scatter plot sbowing the re.lationship between 14 daywinglength andeggsize for the fostered Thick-billed chicks atDiggesIsland (open triangles) andCoats IslaDd (filledcircles).Theslope of the Iioe at Digges Island(0.14:t 0.04 (S8) mm cm..J) wassimilarto that at Coats Island (0.15:t 0.04 (88) rom cm..J; ANCOVA,F",fI'J -0.02,

xii

P -0.89). Coats Islaod data from Hipfuer and Gaston (l999a) . . 90 Figure 6.1: Scancrplots showing the relatiooship between egg size and (a) hatchling

mass.

and (b) hatcbliogtarsus lengtb. for Razorbillschicks. Oatapoints from 1996 ace

opencircles(n-IS),tbosefroml997arefilledciccles(n-23) 106

Figure 6.2: Scatterplots showing the relationship between Foste£ Volume and (a) 14 day ma.ss.and (b) 14 daywingleogth, forRazorbill chicks. Datapoinlsfrom 1996 are open circles (n - 30), those from 1997 arc filled circles (n - 26) . . .... 108 Figure 6.3: Wmg length (mean±SO)inrelation to age for Razorbill chicks that batched

from the largest ODI>third of

ew

(open circles), and thesmallestone-third ofeggs (bothn"18). The bottom figure shows the differenceinmeanwinglengthsbetween

the two groups . . 109

Chapter 1: Introduction and Overview

Research into avian egg productionwas at the forefront ofthe development oflife--history theory, largdyas a consequence ofDavid Lack's early studies ofcluteh size (Lack 1947). Lack: suggested thategg productionwas DOt a particularly dcmaoding process for birds, at leastincomparison to chick-rearing; this conclusionwas based lugdy on theappacemcase with which many species of birds replaced lost eggs and clutches. Focmany yearsafterwards,researchintothe evolution ofavian life-histories concentrated heavily on the reproductive costsassociated.with chick-rearing, while largely ignoringthose associatedwithegg production (Stearns 1992). HoweYa", anumberofrecent studies have shownthatfemales thatproducemoce than theusuallll.lmberofeggs subsequentlyfcecl their chicks less efficiently (Heaney and Monaghan 1995, Monaghan etal 1998). Moreover, eggs produced above and beyond the normal number may be oflow quality, as indicated by low hatching success (Beintemaet al 1991), or bigber-than-expected mortality of the nestlings that hatch from them(Monaghan etal 1995). AJ a consequence ofoew experimental studies, the view that has c:meJBedinrecent yean: isthatavian egg production is much more demanding than Lack: believed, andthat thereproductive costsassociated with this stage ofthe breeding cycle have been important inshaping avian life histories (Monaghan and Nager 1991).

This thesisexam.ineJtwo aspects ofavianeggproductionwithinalife--bistory framework: the detem1inao1sIDd o(mscquc:nces ofreplaccmelltegg-layingamong Thic.k:-billed MurresUrialomvia (Chapters 2 -4). and theconscquences ofeggsize for Tbick:-billed Murres (Chapter S)andRazorbills Aleatorda(Chapter 6). Thick-billed Murres and Razorbills aceboreal andArctic-nestingmarine birds ofthe family Alcidaewhich, like most seabirds, employhigh-survival,low-fecundity life-history

strategies:bothspecies laysingle-eggdutcbes, and bothexperienceaDDUa1adultsurvival. ratesof about00-/0(Hudson 1985). SpeQc:sthat employ this sort oflife.bistorystmegyareexpectedto investcautiously in reproductioninanysingle breeding

season.

becausetheirmaximumreproductive outputinanyone year represents only a very small proportion of their potentiallifetimeoutput,if they survive. While their generallife-history strategy is typical ofseabirds, murres andRazorbilb are uniquein employingwhatSealy( 1973)ooioedthe"'intermediate"devel.opmeDta1 strategy: their chicks remain atthe oest for onlyabout20days,before departing to sea with their malepareolE <30% of adultmassandcoveredinatransitional,mesoptileplumage. to completetheirgrowthatsea.The greatly reducedperiod ofdevelopment at the colony is thought to reflect the extremedemandsof raising young on prey capturedfarfrom the colony, and that is unpredictableinspace and time(Lack 1968). Constraints onchickprovisioniog are especially severe for murres andRazorbills.because they have extremely high wingAoadiDgthatlimits the size of preythatadults can carry bK.k to their offspring at the nest site, and makestrips back andfunhfrom feeding areas to the nest energetically expensive(pennycuick 1987). M in other marine birds. and perhaps especially for these two species, the single-egg clutch is generally thought to reflect constraints operating at thechick-rearingstage oftbe breeding cycle, ratherthanattheegg-production stage.Infact,direct measurements of the energetic COn1e:D.t of the yolk:r~to the female's standard metabolic rate indicatethat,in comparison to other groups ofbirds, yolk: formation may not be a particularly demanding process for murres and most other seabirds (Astbeimer and Grau1990).H~,as noted byMonaghanand Nager (1997), theenergy conteot oftbe fiDaJ. product does not take account oCtile additional energy expended by the femaleinfu~to acquire the nutrients to put into theegg.For many pelagic.

marine birds., these additional requirements are likely to be considerable.Inaddition,bothThick-

billed MUlIl:S (Hipfucr" etaL1997) andRuorbills(Uoyd1979)show apn:la1edtreDdsinegg size that would not be expectedifegg-production costs were insignificant.

In Cbaptef"Z.Iexamine the underlyingcausesof theseasonaJ.declinein theproportion of Thick-billed Murrepairs thatreplace losteggJ.This is widelythoughttooccurbecausefemaleshave little togain. andmuchto lose., byrenestinglateintheseason::due toa seasoQIJ.reductioninfood supply, birds that breed latejeopa.rdizetheirown~whiletheirlate-batchedoffspring; have little chanceofsurviving. However. birds that lose their eggs lateinthe season typicallyace younger. less experiencedfemales. birdswhoseegg-produc:tioo.capacity (Le. theirabilityoc'"wil.lingness' toiIIvest.

heavilyin producingeggs) is low (Hipfuer etaL (997). Therefore, analternativehypothesis holds that it is the low egg-production capacity offemales that lose their eggs late,ratherthanthefactthat theydo lose theireggslate. that makes them unlikely to relay. I use anexperimentalapproach that controls for variatioo in the quality of binb that breedearlyand late in the season. to distinguish between these alternatives.

InChapter 3. I compare survival betWeen colony departure and cecruitment age (4-5 yr of age) for Thick-billed Murres that hatchedfrom firstand replacement eggsinexperimental studies earriedoutin 1994 and 1995(Hipfuer 1997). Empirical studics (e.g.. Birkheadaod Nettlcs.hip 1981), and theoretical models (Birkhead and Nettleship 1982, Ydenberg 1989), of breeding by Aretic seabirds generally assumethatyoung that depart from the colony lateinthe season are unlikely to swvive. AsinChapter I, the~approach wasdesigned to control for variationinthe qualityofpairsthatbreed early and lateinthe season.

10 Cbapter4. Ieamine characteristics offirst and replacement eggs laid by the same early- laying C'bigb quality") female Thick-billed Murre!. A number of studies on other species of birds

have compared the characteristics of first and replacementeggs'and clutches, and attempted to evaluate whetherany diffeunccs between them. reflected evolutionary strategies that females employ to mitigate the costs of reuestmg, or were consequences of physiological constraints (e.g., Arnold 1993, Nilsson 2000).

Among Thick-billed Murres, there is a remarkably consistent 5-6% average reductioninegg sizebetween first and replacement eggs. Birlchead and Nettleship (1982) suggested that female murres reduce thesize ofreplacement eggs in order to advance laying ahead ofthe seasonal reduction infood supply.Insupport ofthis,CommonMurresUriaaaigewere found to form replacement eggs more quickly thanfusteggs(Birkheadmd delNevo 1987), and thisappearedto bettue forThick·

billed Murres as well (Hipfner et aI. 1997). However. the alternative hypothesis, that thesma1lsize of replacement reflects constraints onegg production, seems to have been rarely considered.I evaluate the two hypotheses by comparing the size, and crude and biochemical composition, offirst and replacement eggs. 1ms chapter also presents data ontheaminoacid composition ofthe protein inalbumen and yolk in Thick-billed Murre eggs. information that is,available for very few species of wild birds.

Chapters S and 6turntoinvestigations on a secondaspect ofavian egg production, variation inegg size.InChapter S. I test a widely-discussed, but rarely investigated, hypothesis regarding the fitness consequences ofegg size for birds: namdy, that the advantaBe(s) ofhatching from a large egg is magnified when feeding conditions are unfavourable.Inmany taxa, offSpring that batch from large eggsenjoy advantages over those from small eggs (Azevedo etal 1997). However, there are surprisingly

few

data to support this notioninbirds, even though the relationship between eggsize and offspring fitness probably has been investigated frequentlyinbirds (Wtlliams 1994). Several

reseaccben bave suggestedthatthelack ofevidence for an effect ofeggsize on offspring fitness may be due to thefacttha1cx;perimeDtal studiesthatexamined this relationshipWCI'Ccarried out in situationsinwhich breedingbirdsexperieoccd favourable feeding conditions (e.g., Reid and Boenma 1990, Smithetat1995). Mon:ovcr,ithas been suggestedthatvariation infeedingcoDditionsmigb1 explainwhy considerable variation ineggsize,an inheritedbittbougbtto affect fitDess,persists withinbirdpopulations (e.g.,AnkDeyand Bissett 1976). lnthis view, a range ofegg·sizegeootypes can coexist within a population because the the optimaleggsize variesacross years, due to variation in feeding conditions. Consequently, in favountblc years, the small (ornon-existent.) fitness advantage thatoffspring gainbyhatching from a larger egg is moce than offsetby the added fitness cost incurred bythe adult female for producing that larger egg. rn unfavourable years, a large egg offers sufficient.

advantage that it outweighs the added costs of producing it. This hypothesis has rarely been tested, because few avian studies haveexaminedthe consequences of egg size across a range of feeding conditions.Icompare the consequences of egg size for Thick.-billedMurresata colonywhere feeding conditions during chick..rearingareunfavourable toresuluofa previous sbJdy (Bipfuer and Gaston 1999a) carriedoutatacolony where fccdiDg conditionswere more favourable.Tbick-billed Murres offer the advantage ofbeingODeofvery few species for whichthereiscx;perimeDtal evidence thateggsize affects offspringfit:Desa(Hipfuer and Gaston 1999&).

FinaUy, inQlapter6, I employanexperimentalapproac.bto test the bypotbesis that the positive rdationship betweco egg size and post·hatcbing wing.feathergrowthobserved previously in Thick.-billedMurres reflects adaptations of"intermediate" alcids (Hipfuer and Guton. 1999a). To do this, I carried out an egg-switching experimentinRazorbills. For theintermediatealcids, rapid growth of primary coverts (the longest feathers on the wings of nestling murres and Razorbills)

during thebriefnestling period appears to be critical to enable thechick:to make the transition from life at the nest site tolifeat sea quickly and safely (Birkhead 1977, Hipfucr and Gaston 1999b). This isbecause these species typically breed high on cliffs, and at nest departure, the chick glides and flutters from the nest to the sea, .accompanied by the male parent. Mortality rates are veryhighon young thatfailto reach the sea directly (e.g., Gilchrist and Gaston 1997). There appear to be a numberof adaptations inegg characteristicsand patterns inthepost-hatching development ofthe intermediate alcids that reflect the urgency ofgrowing long covert feathers. These adaptations are discussed in detail.

1.1 Literature Cited

A.nkney,C. D.•andAR.Bissett.. 1976.Anexplanation ofegg weight 'YViationin tbelessersnow goose. Journal ofWddlife Management 40:729-734.

Arnold,.T.W.1993. FactorsaffeaingreoestinginAmerican Coots. Coodor95: 273-281.

A.stehimcr", LB., and CA. Grau. 1990. A comparison ofyolk growth ratcsinseabirdeggs.Ibis 132:

380-394.

Azevedo.R.B.R..V.French,aDdL.Partridge. 1997. Life-history consequences of eggsizein Drosophila melanogaster. American Naturalist 150:250-282.

Beintema, A.J., T. Baarspul, and J.P. Krijger. 1997.CalciumdeficiencyinBlack: Terns Chlidonias niger nesting on acid bogs. Ibis 139: 396-397.

Birkhead.,T.R.1977. Adaptive significance oftbe oestling period of GuillemotsUria aalge. Ibis 199:544-549.

Birkhead, T.R., and D.N. Nettleship. 1981. Reproductive biology of Thick:-billed MurresUria lomvia: an inter-colony comparison.Auk:98: 258-269.

Birkhead, T.R., and AJ. ddNevo. 1987.Egg focmation and tbe pre-layingperiodof tile Common GuillemotUriaaaige.louma1of Zoology. London 211: 83-88.

Birkhead, T.R.... and D.N. Nettlcsbip. 1982. Adaptive significance ofeggsize and layingdatein Thick-billedMurresUrialomvia. Ecalogy 63: 300-306.

Gilchrist,KG.• and AI. Gaston. 1997. Factors affecting the success ofcolony departurebyTbick- billed MurreclUcks.Condor 99: 345-352.

Heaney, N., and P. Monaghan. 1995. A within-clutch trade-off betweenegg production and rearinginbirds. Proceedings oftbe Royal Society ofLondoQ (8) 261 :361-365.

Hipfiler,IM 1997.Theeffeeuofparentalqualityandtimingof breed.ingonthegrowthofnestling Tbick::-billed Murres. Condor 99: 353·360.

Hipfuer,I. M.,andA.J.Gaston.1999a.The relationship between egg size and post-hatching developmentinthe Thick-billed Murre. Ecology 80: 1289-1297.

Hipfner, I. M, andA.I.Gaston.1999b. TllDing of nest departureintheThick-billc:d Murre and Razotbill: tests ofYdenberg's model Ecology 80:587-596.

Hipfuer,I.M, A.I.Gaston, and L.N. de Forest. 1997. The role offemaleage in: determining egg size and laying date ofThick-billed Murres.Journalof Avian Biology 28: 271-278.

Hudsoo, PJ.Population panmctcn for the AtlanticAlaw.Pages 233-261inD.N. Nettleship and T.R. Birkhead,editors.TheAtlanIicAlcidae. Academic Press,London.

Lack,.D. 1947.Thesignificanceofclutcb size. Ibis 89: 302·352.

Lack,D.1968.Ecological adaptations for breedinginbirds. Methuen, London.

Uoyd, C.S. 1979. Factors affecting breedingofRazorbillsAlca rorda onSk:o~olm.Ibis 121: 165- 176.

Monaghan. P.,M.Bolton, aod D.C. Houston. 1995. Egg productionconstraintsandthe evolution of avian clutch size. Proceedings of the Royal Society ofLondon(8) 259: 189-191.

Monaghan, P., and R.O. Nager. 1997. Why don't birds lay more eggs? Trendsin Ecology and Evolution 12: 270-274.

Monaghan, P., R. G. Nager, and D. C. Houston. 1998.Thepriceofeggs: incr"easedU1vestmeDtin eggproduction reducesthe offspring-rearing capacity of parents. Proceedings of theRoyal Society of London(8) 265:1731-1135.

Nilsson, I-A. 2000. Time-dependent reproductivedeci5ionsintheblue tit. Oikos 88: 351-361.

Pennycuick. CJ. 1987. Flights ofauks (Alcidae) and other nortbem. seabirds comparedwithsoutbcm ProceIlariiform:omitbodolite observatioos.lowna1 ofExperimentalBiology 128: 335-348.

Reid,W.V., and PD. Boersma. L99O. Parmtalqualityand selection oneggsizeintbeMagdlanic Penguin. EvofutiOll44: 1780-1786.

Sealy, S.G. 1973. Adaptivesignificanceofpost-barcbing developmentpatternsandgrowthmesin theAlcidae. 0miJ ScaDdmavica 4: 113-121.

Smith,H.G., T. Olssen, and K.. 1. Wettc:nnaIk. 1995. Adaptivesignificanceofegg.mein the EuropeanStarling:experimentaltests.Ecology 76: 1-7.

Steams, S. C. 1992. The evolution oftife histories. Oxford University Press, Oxford.

Williams, T. D. 1994. Intraspecific variation ineggsizeand eggcompositioninbirds~effects on offspring fitness. Biological Reviews 68: 35-59.

Ydenberg,

R.e.

1989. Growtb-mortalitytrade-offs and the evolution ofjuveni.le life-histories in the Alcidae. Ecology 70; 1494-1506.

1.2 Co-A....onhip Stat_at

My contributions to the research describedinthisthesis were: (1) I proposed and developed the research qucsttons,andwasprimarilyresporwble for the design oftbe experiments used to address these questions; (2) Iwasprimarily responsibleforcarrying outallofthe fidd work desa:ibed, except for the fiddwod:carriedoutat Coats hlandin1997(secQlapter"2).The analyses of the biochemical composition ofThick-billedMurre eggs (see Chapter 4) were ClUried out by, or under tbcdirection o( Drs. J. Brosoanand G. Herzberg, DcpartmentofBiochemi:suy. Memorial University of Newfoundland;(3) I analysedalloftbe data; and (4) I wroteallof thecbaptenimaIllLSClipts.

Chapter 2 oflhis thesishualreadybeen published as:

Hipfuer,J.M, AI. Gaston, D.L. Martin, and I.L. Jones. 1999. SeasoDBI declinesinreplacementegg.

layingsina long-lived Arctic seabird: costs of late breeding or variationinfemalequality?

Joumal of Animal Ecology 68: 988-998.

Chapter 6 of this thesishu been accepted for publication as:

Hipfuer. I.M 2000. The effect of eggsize on post-hatching developmentintheRuorbill:an experimental study. Journal of Avian Biology 3 L

10

Chapter"2: Seasonal DeclinesinReplacement Egg-Layingsina Long-L.ived.

Arctic Seabird:CostsofLateBreeding or VariationinFemale Quality?

2.1. Abstract

Inmany species ofbints. the proportion offe:males that relays after losing theirfirstclutchvaries.

R.eplacemeatclutclles can make importaDt cootributiODS to aniDdividuaJ.',lifetimereproductive sucoeu,sothedecisioo00whetherorDOttordaymayhave significant conscqueoces for fitDCSS.

Manystudies reportthatfemalebirdsof-high quality" (e.g., older and more~enccd)weremore likelyto relay foUowing clutch loss, and that the proportion of females that relaid declinedasthe brccdi.ng season progressed. High quality females tend to nest early. and may be less likc1y to lose theirclutch.~a result, it is unclear whether few females relay late in the season because there are prolnDitivecostsassociated with latebreedill&or because those females that lay andlosetheir eggs lateare of low quality. and therefore incapable of relaying. 1 ccamined the roles ofdate and quality in causingseasonaldeclines in replac:emeot layings in the Tbick-billedMurre UriaIom'l'ia,a 10118- Jived,Arctic marine bird that lays a ooe-egg clutch. At two low-Arctic colonies in eastern Canada in1996 and 1997. the single eggwasexperimentally removed from samples of early-layingpairs (presumably those ofhigh quality)at3 day intervalsbegi.nningon the median laying dateaDd ending 12 days later.Inthegeneralpopulation, the proportion ofpainthatrelaid foUowingnatunJ.eggloss dedioed with the date ofloss.Incootn.st, the proportion ofexperimeDlal pain that relaid remained.

bighirrespcctiveofthedate their eggwasremoved. These results support the quality hypothesis. but not the date hypothesis. 1 conclude that seasonal declines in replacement layiogsoccurredprimarily because increasing proportiODS of low quality females lost eggs as the seasonadvanced. The

11

experimental pairs also suffered no oven reductionin reproductive success as aR:SU1tofthe delay, atleast up to the timethattheir chicksdepartedtosea. This suggeststhattimingofbreedingbad little effect.ontheimmediate success of morecapablepain..These resultsmayhave importan1 implicationsabout the seasonality oflow-Arctic marine cnviromneats, and thelifehistories of birds that inhabit them.

1.1IDtrodUctiOD

According to life.historytbcory, individuals ofitcroparous species.will attemptto balance the bcnc613 of investingincurrent reproductionagainstthe costs that reduce their ability to investinfuture reproductioninordcrtomuimizetheir lifetime reproductive success (Stearns 1992). One life-history decision that many birdsfaceduring a breeding seasonis whcthcrto lay a replacement clutchifthey losetheirfirst. Replacement clutches potentially make important contributions to an individual's lifetime success, especiallyinspecies or populations prone to clutch lou(Martinctal. 1989,~eegan andCntwford 1993). However, many studies report thatonlya low proportion offcmale~dsthat losttheirfirstclutch relaid (e.g., Duncan 1987, Connelly etaL1993), suggestingthatthe costs can be prohlbitive for maoy individuals.In sevenlstudies, femalesthatrdaidfollowingclutch losswece olderandmoreexperienced(Wooller 1980, BoekeI.heide and Ainley 1989,WbeelwrightandScbuJ.tz 1994), or in betta"coodition (HegyiandSuviri1998), than thosethatdid not relay.Other studies reportthat femalesthatreoested laid more eggs (Rooneem and Robertson 1997) or larger eggs (Sandercock: and Pederson 1994)intheirfirstclutches, and that they laid theirfirstclutches earlier in the season (e.g., Kelly and van Home 1997). TheseohSCIVations suggest that "high quality"

females might be more likely to relay.Inaddition,the proportion ofindividuals that relays invariably

12

declines with the date ofchrtcb.loss (Feare 1976. Parsons 1976. Arnold 1993, Barbaetal1995).

Thismightreflectthe highcost·to-bcnditratio associatedwithlatebreeding (SiIYerin 1980).

Consequeody.ooe ofthe mostimportantevolutiooarya.ctva.ut.gesofearlylaying

roc

birdsmight be the incceased potentia.l to renestinthe eventthatthe firstclutchis lost (Hannon etaI.1988).

lbick:-billedMurres arecolonial, clilf-nesting, marine birdsinwhich annual adult survival is high(c. 90";'. Gaston etal 1994) and annual reproductive output is low (they lay ao~ggclutch).

Species with theselife-bistorycharacte:ristics, wbichincludemanymarine birds, areexpectedto invest cautiouslyinreproduction in any single yearinorder to avoid jeopardizing their future reproductive potemial Moreover, lbick:-billed Murres breedinthe Arctic where the penalties forlatebrecdins arethought to besevere (e.g..reduced reproductivesuccess.Birkhead and Net:tlesbip 1981). Itis DOtsurprisingthentba1few lbick-billed Murresreplacelosteggs(typically 20..30'% oftbosethatLose anegg may), andthatthe proportionthatreplaces declines abtuptlywith the date ofegg 1055 (Gaston and Nett1esbip 1981. Birkhead and Nettleship 1987a, deForest and Gaston 1996). Howl:\lu. many eggs that are lost, and particularly those lost late in the season, belong to young, inexperienced.

females because these females lay later (Hipfuer etal 1997). and are more likel.y to lose their eggs than are older. more experienced females (de ForestandGaston 1996).~a result, it is unclear whethertheseasonaldeclines in the proportion ofThick:-billcd Murresthatlayreplacemeot eggs are causedbydateeffectsperse. or by late breeding and lateeggloss oflow quality females.

Inthis paper. Iexaminetheroles ofdate and female qualityincausing seasonal declines in replacemcnr.layings by Thick:-billed MWTeS. I removed the singleeggfrom earty"ayiog pain (those that appear to be most capable,. de Forest and Gaston 1996) at successively later dateI that spanned the interval duringwhichdeclining proportions of control painwerereplacing lost eggs.Ifthe

13

declines arecausedbydate effects and are iodepeodem offemale quality (the daU hypo/hesis), tbe:n the experimentalpairs shoukl fonow the population-wide ded.iocs. Convenely,ifthe declines are quality effects and indepeodmt ofdate (thequalityhypothuis),tbe:n the proportion ofexperimcmal females thatrelayssbouId remainhighirrespective of the date theiregg is removed. I also compare the reproductive success of the experimental pairs to that of early-laying. unmaoipulated pairs to investigate whether replacement eggs are leu likely to produce fledged chicks than are first eggs for Thick-billed Murre pairsthatlay their first eggs earlyinthe season. Finally, 1 examine causes of variationinthe number of days that elapse between removal of the first egg and thelayingof a replacementegg(hereafterthe relaying interval), and causes of variationinthe size of replacement eggs compared tofirsteggslaid bythe same female.

2.3 Stvdy Ana and Methods

The studywas conducted at the Tbick-billcd Murre colonies at the Gannet Islands, Labrador,Canada.

(S3°S6'N,S6"32'W)in1996 and 1997, and at Coats Is.land. Nunuavut Territory. Canada (62°S7'N, 82"OO'W)in1997. Both colonies are situatedinthe low-Arctic oceanographic zone (Nettlcship and Evan5 1985).

1monitored the breeding chronology and breeding success ofThick-billed Murrcs usingthe Type ImC'tbods ofBirkhead and Nettlesbip (1980). Expcri.meata1pairswere selected from among those pairs that laid by the estimated median layingdateforthe whole colony. At both coloniesin 1997,medianlaying dates wereestimated fromplots observedinprevious yeanthatheld known numbers ofbrceding pairs. Because 1996was ourfirstyearlitthe GannetIslands,the mcdian.laying date had to be estimated based on site occupancy on plots observed during the pre- and early-laying

14

periods.These methodsprovedreliable: theestimatedmediaD. laying dates were within 1dayofthe truemedian (estimatedfromlaying dates offirsteggsfur- allstudy pairs) in

an

samples.Ineachof thethree colony-yean ofstudy,the siDgle eggwasremoved (underpermitfromtheCaDadianWildlife Service) from five to eightc:xpc:rimemalpainon each oftive occasions, beginning onthe median laying date, and continuing at 3 day intervals upto the median+12 days. At theGannet lslands,a total of30eggs wasremoved in 1996,and25 in 1997; 33 eggs were removed atCoatsIsland in 1997.Control samples inthese threecolony-yean were 35, 42, and 123 pairs thatlosttbeirfirsteggs naturally. The length 8Ddmaximumbreadth of each removedegg wasmeasured

<:t

0.1 mm) using calipers, anda volumeindexOeogth •breadth')thatbasa stronglinear relationship with the fresh mass ofThick-billed Murreeggs(1' -0.952, Birlc:bead and Nc:ttIeship 1984)wu used u a measure ofeggsize. At Coats Island, where many Thick-billed Murrebreedingsiteswereaccessible, most of the experimental replacement.eggs were measured. Tomio.imiz.e disturbance, replacement eggs were notmeasuredat the Gannet Islands, where most sites were less accessible.

Wherever possible (see below), the following informationwasrecorded for both control and experimentalpairs:(i)laying date offirstegg;(n)date oClo" (orexperimentalremoval) offirstegg (Iattemptedto excludeinfertileeggs -those that were incubatedtofulltermbutWledto hatch -by includingonlyeggsthatdisappeared <31 daysafterlaying,theminio:aunincubation period forThick- billr.dMurreeggsatboth colonies; HipfueretaLinpress);(ail) laying date of replacementegg;(tv) hatching date; (v) date ofdisappearanceofchick.Cbick.swereIISSWIledtohave SUJ'Vivedtodepart from the colony iftbeydisappeared~15 days afterbatcl1ing(the minimum ageat which Thick-billed Murrechicks are known to depart the nest oftheir own volition, Gaston and Nettlesbip 1981); it was assumed that they had died prior to departureifthey disappeared at<15 days of age.

IS

Qbsentationsat theGannetlsI.andsin1996 cootiauedwrtilallcontroleggs,andall apcrimanal replacementeggs.bad batcbed. However, fieldaewslefttheislaodsbeforeall experimentalchicksbaddeparted sothatfledgiDg success data were uaavailable forthatyear.In addition, the layingdates offewc:xperimeat&l6nt eggswereknown in 1996 (althoughallwere laid by theestimated media.a laying date).In1997, observatioas at the Gannet Islands spanned theentire breeding season, but observatioas at Coats Island ceased prior to the hatching ofallexperimeuta.L egg'.

2.3.1DaIaAnalysis

Mostanalyseswere conducted usiDg SigmaSw (venion.2.0) statisticaI software,wtUchmrtomatically examines theoonnaIityandcqualit.yofvariaDce assumptions of parametricstatistical tests. Non- parametric tests wereusedwhere theseassumptions were notmet. Toanalyzevariationin the proportion of pairs that relaidin relation to date of egg loss (and other factors, as appropriate), logistic regression analyses were conducted using SPSS (version 7.5)statistical software. I started with logistic regressionmodels that included oaly the constant and thenuseda forward stepwise variable selection pl'OCeClw'e withall maineffectsandallintenction terms available. Vviable selection wasbasedon the <:bangeinthe Iikdibood-ratiothatoccutTedwith the variabledeletedfrom the model Signi6canc.e levels for entry into the model were set at P<O.OS,and at P>0.1 foe removaIfrom the model (the default values in SPSS).

To examine aatunJ. variationinthe incidenceofreplaccmcntegg-layinginrelation tothe date of egg loss, the logistic regressions were run first onallcontrol pairs,withdate ofloss (relative to the median laying dateinthat colony·yeac) entered to the exact day. Because eggs were removed

16

at3 dayinterValsover a set timeperiod.relaying probabilitywas compared between control and c:xpcrimco1aIpairs with the data: (i) grouped.by3day intervals relative to themedianlaying datein each colony-year, and (0) limited to the period of overlap betweconatutalegglossesand experimentaleggremova15 (fromthemedianlaying date - 1 day, to the median+13days. see Ftg..

I). Allmeans ace reported:!: 1 SO, IIDdaUreported p·values are two-tailed.

2.4 Results 2.4.1 TJ1J'Iing

01

breeding

TUDingofbreediog by Thid:::-billedMurres was similar at the Gannet Islands in the two years,with medianlayingdatesoffirsteggsfalliDgon 26 June (1996)and28 June (1997).Thiswas similacto thetiming recordedinprevious yeanat this colony(Birldle&dand Nettleship 1987b).Themedian laying dateatCoats Island fell on 21 June in 1997, whichwas relatively early in comparison to previousyeanat this colony (Gaston andHipfuer1998).

2.4.2 Date ofegg loss and the probability o/replacement laying

Two experimentalpairs stole eggs fromDCigbbouringThick-billed Murres after theirowneggwas removedattheGannet I.slandsin 1996, as did one paicin 1997.Inaddition, two experimcznalpairs stoleeggsfrom Common MurreUriaaalgeocighboursin1996. These five pain were n:moved from allanalyses. Two experimcznal pairs stoleeggs at Coats ls1and, but these eggswere returnedby us to their original sites immediatdy.

Inthe logistic regression analysis, therewas a negative relationship between the relative date ofegg loss and the proportion ofcontrol pairs that relaid after they losttheirfirstegg (Table 2. I, Fig.

17

NOTE TO USERS

Page not included in the original manuscript are unavailable from the author or university. The

manuscript was microfilmed as received.

18

This reproduction is the best copy available.

UMI

Table2.1: Results of a forward stepwise logistic regression analysisexaminingvariationinthe proportion ofcontrol Thick-billed Murre pairsthatrelaid following naturaleggloss (n - 200).The consta.otwasentered into the model Variables available tothe moddweredate ofeggloss.colooy- year, andthedate • colony-year interaction tcrIn. Sdection ofvariableswasbased on the change in the Iikdihood-ratio with the 'fViable deleted from themodel Variables entered the modda1P<

0.05, and wereremo~from the modd atP>0.1.

Variable

Constant Date ofeggLoss

Change in Iikelihood-ratio elf

103.92

19 p

<0.001

<0.001 R

-0.43

6 4 5 Gannet Is.

100

•...

~ / -

...

1996

i

⁷⁵

3 4· ...~/ "-

.···3·.

^'\.5

50

'.'

a.

6

25 '.4

13 0

100

.~ / - - - .

^{5 5 5 4 GarYlet}1997^Is.

~

^{75 5}

^{/'. ...}

^{6 .}"~

If

⁵⁰

25

.1 _1~

0

6 6 CoatsI.

100

'-~/--"'" ^~

¹⁹⁹⁷

~

^{75 5} o'"S"', ~/

Q) 50

. '.

a.

11

-;"'-9-"

25 0 .78

<-,

^{-1-1 2-4 5-7 6-10 11-13 >14}

Date ofeggloss or removal

Ficure2.1: Percent ofcontrol (dotted line) and experimental (dashed line) Thick-billed Murrepairs thatreplaced losteggsinrelation to thedate ofeggloss or removal (grouped by 3 dayintervals,in daysfrom the median layingdate)inthe three colony-years. Sample sizes beside points.

20

Table2.2:Results of a forward stepwise logistic regressionanalysiseumi.ningvariationinthe proportionofcontrol (n - 79) andexptrimeotal(n - 83) Tbicl::-billedMurrepairsthatrelaid foUowiDgoatunJegg Jossoreggremoval duriog the periodofovutapbetweenthe twogroups.The constant was emceedintothe model Variables availa.ble to the model were date ofeggloss (grouped bythree-day interVals), colony-year, treatment (control and experimental), andalltwo-way and the one. three-wa.y interaction tenDS. Selection ofvariablesWi15based on thechangeinthe likelibood-ratio withthe variable deleted from themodel. Variables entered the model atP< O.OS, and were removed fromthcmodel atP> 0.1.

Variable

Date of loss • treatment 41.73

21

<if ^p

<0.001

<0.001 R

-0,40

threefor15 days. Inwbation periods ofexpc:rimmtal6.rst:eggsthatwerereplaoed rangedfrom0 to 16 days.Duringtheperiodofoverlap between naturalegglossesandexperimental removals at the GannetIslandsin1997(theonly sample in which I bad complete data), controlpainhadiocubatcd theirfirsteggsfor less time beforetheylost them. than bad the experimentalpairspriortohaving their eggsremoved (coDttols:median" 2 days, " .. 21; experimentals: median - 8 days, n - 24;Mann- Whitney test,p ..0.01).

2.4.4Relayingintervalandnp/oc%menteggstu

[ used multiple regression toeominehowrelayingiDtervalswere affectedby:(i) the date ofegg removal, and(ii)the length oftimethatfirsteggs were incubated, using data fromthe two coloniC!

in1997.Tbemultiplereg:ressionwashigblysignificant(F},~-12.52,P<0.OOl).H~,neither of the independcm variables appeared tobe important in this analysis(bothP ;:: 0.42), probably becausetheywerestronglyWeI'COrre1ated (ru -0.96,P<O.OOI)(seeZar 1996,p.419).Insimple regressions,the relaying interval increased with both the date ofegg removal and the numberofdays that thefirstegghadbeenincubated (Table 2.3). Theregression lines didDOtdiff~significantlyin eitbc:r elcvatioa orslope betweencolonies (Table 2..3), and Fig. 2.2 shows therdatiombip between relaying interval andchntion ofincubation with the two samples combined.

Experitnentalreplacement eggs that were measured at CoatsIslandaveraged 6.1±2.8%

small~thanfirsteggslaid bythe same female (first eggs: mean" 211.3±15.4cm"; replaccment eggs: mean- 198.3±IS.3 em),both" -27; paired~-11.89,P <0.(01). The sizesoffint aDd replacement eggs laidbythe samefemale were strongly correlated(Tn"0.93,P<0.001).The percent differenceinvolume index between6rst and cc:placemcnt eggs showed little relationshipwith

22

Table 2.3:Regressions rdatiDg the relayingiotcrval to: (1) the date ofremovaLoflirsteggs.and (2) the number ofdays thatthe first eggs wereiDaaba1edprior toranoval (covariatesintbcANCOVM), for expe:rimcntal 'Thiele-billed Murresatthe Gannet Islands(n -=24) aDd Coats Is1and(n -33) (grouping variablesinthe ANCOVAs)in1997.

ANCOVA

GanDetIslands Coats Island Colony 1otcnction

Variables F p F P F P

Rcmovaldate 0.220.18 Incubation period 0.25 0.17

0.63 0.26 0.57 0.22

23

31.70 <0.1)01 3.26 0:08 0.92 0.34 30.06 <0.001 1.69 0.20 0.36 0.55

20

18

§:

~

¹⁶

~ '" "

¹⁴

:;

'"

'"

12

10 0

4 •

y

=

13.2+0.21x

4 B 12 16

Numberofdays firsteggincubated

Flpre2.Z:Relaying irtterva1sofThick-biDedMwTesin relationtothe llIJIDb«ofdaysthatfirsteggs wereincubatedpriortoexperimemaIremoval.Data from theGannetIslands andCoatsIsland in 1997werepooled.

24

the date ofeggremoval(r=-0.01, FU' -0.41,P-=0.S3),tbe number ofdays thatthe6ntcgghad beenincubated(r ""0.02,F~-0.47,P-O.SO),

or

the relaying interval(r -

om,

F_{1,2S -}0.24,P -0.68).

2.4.5&protbIcmcSllccusofe:xperimenlaJpairs

Atthe Gannet IsJands in 1996, thehatchingsuccess ofexperimeatal replacement eggs(87%,if.,23) was similar to that offirsteggs laid by unmanipuiatcd pairsthat laid by the dateofthefirstegg rcmoval(93%,n=93;rl-0.S2,P-0.47,Vates' corrcctionapplicd). Hatcbingsuc:ccss<r,-1.47, P - 0.23, Yates' correction applied). Hedging success (Ftsbcr'sexacttest,P - 0.71), and reproductive success (X¹1 -0.57,P - 0.45, Vates' correction applied), ofcxperimcntal painwere similarto those of control early painatthe Gannet Islands in 1997 (Fig. 2.3).

2,3 Discussion

2.5.1 Seasonal declines In replacement /ayings: effects ofdate or quality?

Asinprevious studies (Gaston and Ncttleship 1981, Birkhead and Ncttleship 19B7&,de Forest and Gaston 1996),the proportionofThick-billed Murn:s that laid rcpJacement eggs dcdinedwiththe date ofegg lossinthe gcnenJ.population.However, the proportion of experimeatal femalesthatrelaid remainedhighirrespective oftbe date their egg was removed. Consequently,theseresults support the quality hypothesis, but do not support thedatehypothesis. I cooctudethatscasonaI.declinesin replacement layings oc:o.uredinthegenen.I population primarily becauseinaea.siogproportions of iow-quality females lost eggs u the season advanced. A previous study found thatyoung, inexperienced female Thick·billcd Murres laid laterthanolder, experienced females (Hipfil.eret al.

100

45 18

80 23

23

~

^{60 73} 73

c7l ~

⁴⁰

20

HS FS RS

F'.pret.3:HatclJiDgsuccess (HS), 8edging success (FS) and reproductiveSUCCCSI(RS) ofcomrol (openban)andexperimental(hatched ban) Thick-billed Murre pairs at tbeGannetIslandsin 1997.

Controlpairswerethosethatlaidbythe date ofthefirstexperimental CS8 removal. All measures of succesa were similarbetween controlsandexperimental

pairs.

Numbers above bars are samplesizes.

26

1997). YOUDg birds alsowerelesslikelytolaya replacemeutegg.and those that did hadlong relaying inteIvals (de Forest aDd Gaston 1996). These observationsiodicatethata femaleThck- billed. Murre'scapacity(Le.,herabilityor -wiJliDgness") to produce areplacementeggioa'ease3 with herage and/orexperience,as does hercapacityto produce atint egg(Bipfi:ler etaI.. 1997). I suggest that the female quality effeeu documentedinthis study largely reflect effects ofage and/or experience.

Many recent experimental studies, most involving temperate-nesting species, haveexamined the extent towhichdate effects and parental quality effects caused avian reproductive success tovary within seasons (e.g., BrinkbofetaI..1993, Norris 1994, Wiggins etaL 1994, SiikimIki 1998). Date effects wen: more important than parental effectsinmost studies. For egmple, experi.meotally_

delayed,earty.layingGreatTitsPanu maicxwere less likely thanunmanipulatedearly-layingfemales tolaybothreplacemeot(BarbaetaL1995)and second (VcdJulstetaL 1995. VerbovenaDd Verlmlst 1996) clutches. Formanybirds, there is a highly predictable declineinfood availability (e.g., insects) lateinthe breeding season that is thought to compromise the survival oflate-hatchedyoungand late- breeding aduiu (perrins 1970, Daan et aL 1988). Late-breeding adulu (Gustaffson et ai. 1994), and late nestlings (Sorci et al. 1997), may be more susceptible to disease, andlate8edglings sometimes fare poorlyincompetitionwiththoxthat8edgeearly(NHssoo 1989. Spear and Nur 1994).In addition, adulu that delay their post-ouptialmouhmightproduce a basic plumage of poor quality.

which can inaeasc thermoregulatory costs, and reduce ovuwinter swvivaI. (N....lsson and Svensson 1996).Consequently. thea often is a high cost-to-bcnefit I1ltio associated with late breeding that couldstrongly influence avian reproductive behaviour.A$with seasonal decli.nesinclutch size (e.g., Wmlder and Allen 1996) and the proportion of pairs that attempt second broodsinthe same season

27

(Verboven and Verlwlst 1996), seasonal declines in the proportion of pairsthat replace lost clutches mightbe evolutionary responsesto the reduced fitness of late adults and late younginmany species ofbirds (Silverin 1980).

Thebreediogseasooisthoughttobestrictlytime-<:onstrainedin.Accticcuviroamcuts.Despite this,e::Jq)erimentaJThic!t-billedMurresthatlaid their firsteggs cady in the season c:ontioued to lay replacementeggsu.ntiIlateinthelayiDgperiod. This suggests that the evolutionary beDefits obtaiocd byrelaying andraisingyoung even verylate in the season exceed the costs for the more capable members ofThick--billedMLlrTepopulations. Life...history theory predictsthatspecies withhigh survival and lowfecundity(like most marine birds, including1IlUl1'eS)willinvest cautiously incu.rreDt reproduction because their lifetime reproductive successis influenced morebysurvival, and therefore the number of breeding episodes, thanbyannual fecundity (Clunon-Brock: 1988). Congruent with this theory, studies on other marine birds oilen have found that adults racedwithexperimental trade- offs ravoured their own well-beingover-thatof their offspring (Sztberet aI. 1993,Mauckand Grubb 1995, Wemham and Bryant 1998; but sce Johnseo et al. 1994,lacobsen et al 1995).Ifthis istrue.

thenthe most important implicatioo

or

this SbJdy isthatrelaying,breeding late, and departing late from their low·Arctic coloniesinflictrelativelylittle additional cost on capableadultlbick-billed Murres. Rowevet'. thisremainsto be tested directly.

The reproductive success of the experimeotalIy-ddayed Thick-billed Murrepairswassimilar tothatof umnaoipulated early.laying pairsat the Gannet lslands,atleastup to the time ofdeparture oftheirchicks to sea. de Forest and Gaston (1996) obtainedthe same resultintwo previow years at Coats Island.Incontrast, experimental studies on otherbirdspecies often findthat replacement eggs foUow natural., population·wide seasonal declines in success (Feare 1976, Barba etat1995,

28

VetboveoandVlSSCr1998; butsce Panoos 1975, Ha1chwdJ.1991). Moreover, Thick-billedMwre chicksraisedbyearly-l.ayiog.expcrimentally-ddayed pain:grewas wdl and were as heavy at nest departureas chicks raised by unmanipulated early pairsinthree previous yean atCoatsIsland (de FOf"eSt llDdGaston 1996, Hipfuer 1997). Consequently,itisunliJcetythatthere15asystematic seasonaldecliDe in the availability offood that affects the viability ofThick:-billed Murre chicksraised by capable parents, at least at the low-Arctic coloniesinthis study (chick: diets at the twostudy colonies consist mainly of fish, bothpelagic,schooling species and benthic species; Gaston 1985, Bryant etal 1999). lisa, this differs from the situation formany otherbinis. However, I have no information00food availability after departurefromthese coloaies, nor doIknowwhether deputure date affects post-departure survival of young lbick-billed Mw:res.Insome years, Common Murre chicksthatdepaned lateinthe seasonfroma tempe:ratecolonywereless likd.yto be resightedinthe future (Harris etal1992;but see Hedgren 1981), but late chicks might have hadlowsurvivalbecause they were tended at sea by less capable parents, rather than because theydeparted late.

Murres also might obtain indirect benefits by laying replacement eggs late

m

the seasonifthis acts as asignalof the female's.quality to her mate. Many studies have demonstra1ed thebenefitsof re-matiDgwitha partner in a subsequent year, andthatbreeding failureisfoDowed by divorce more ofteothanexpectedbychance(Rowley1983). Therefore, a femalethatrelays lateinthe laying periodmight benefitthrough aninaeasedlikdiboodofn::matingwithhercurreD1partnerand belding on to her nest site for breeding inafuture year. Ainley etaI.(1990)suggested that this might partly explain why some female Cassin's Auldetsptychoramphus a/eu.ticuslaysecondeggsinODeseason aftertheirfirstchick fledges.

29

2.5.2Eggsize. incubationpttriod. andprohohiIityofreiaymg

1bcre was little celationshipbetwec:o.firstegg size and the probability of relaying among the experimental Thick-billed Murrepairs.Incontrast,WillowPtarmiganLagopus/agopfuthatlaid large eggs were more likely to relay following clutch loss, suggesting thatboth fiLcton reflected a female's capacity to produce eggs (Sandercock and Pederson 1994). By removing the single eggs frompairsthatlaid early, I excludedmostyoung,inexperiencedfemales, birdsthatteDdto lay small eggs late in the season (Hipfuec etat1997).However,eggsize varies coDSidenbly even among older Thick-billedMwres,whichtend to lay early. aDdegg size remains rdativelyfixedbetweenyears for older females (Hipfner etat1997). Therefore,. even the small eggs removedinthis study probably were laid by capable females. .I(as I have suggested, age llIldIor experience largely determines relaying propensityinThick-billed Murres., thenitis DOt surprisingthat relaying propensity was not reflected infirsteggsizeamong early-laid eggs.

Thelength of time that ticsteggs were incubated,likethe date ofegg removal withwhichit was higbly correlated, also appearM to have little overt effect on relaying probability among the experimental.pairs. More importantly, rela)i.ng declined withdate among control pairs,butnot experimental pairs, even though experimental pairs bad incubatedtheirfirsteggs for almost a week longer 00 average prior to loss. Young Thick-billed Murre pairs typically lay late and losetheiregg quickly (de Forest and Gastoo 1996), andthistendencywas displayedbycontrolpairs.MW'phy and Schauer (1994) also foundthaJ:theduration of incubation bad littleeffect00 relaying probabilityin a non-experimeotal study ofCommon Murres.

AtCoats Island, Thick-billed MllITCIusually gainmass ovee the course ofincubation (Croll et aI.1991, Gaston and Hipfuer 1998), suggesting thattheyhave little difficulty meeting the energetic

)0

dema.ods ofinc::ubation 11thatcoLony. Therefore.,itseems unlikelythata longer incubation period wouldreduce a female's ability to summon theenergyrequired to produoe areplaceme:megg.

However, body massmi.gb1notreflect the availability ofthe morespecializedresources required for egg production, which in some species of birds havetobedrawnfrom endogenous reserves laid down before egg production begins (Houston et al.1995).Itis poSSIblethatonly high quality females will have sufficient reserves to produce both afirstand a replacement egg. There was, however, a cleareffect ofincubation duration on relaying intervals among the experimental females. indicating that physiological changesthataffect eggproduction take placeevenearlyin incubation(see below).

Such changes likely cause a gradual. reductioninthe female'segg-production capacity, andthis probably contributed totheverylow l1ltes ofrelaying amoogbirdsthatlost theireggsveryl.ate.. To investigate the influeocc ofincubationduration on probability ofrepl.a.cemcotlayinginisolation from othCf", potentially confounding factors (e.g., female quality) will requirefurtherstudy.

2.5.3 Relaying in!erval and replacement egg size

Relaying intervals increasedwithboth the date ofegg removal and thenumberofdays that first eggs were incubated prior to experimental removal. [suggestthatthefirstofthesepositive relationships was caused bytbe second. as WooUer(1980} suggested for Black-legged KittiwakesRLuatridactyla..

Neither proximateDOrultimatefactorsseemtoexplain whyrdayingintervals should inaeaseas a function of removaldate for Thick-billed Murres. Food probably becomes more available through the laying period as sea ice retreatsinthesummer-(Gaston and Hipfuer 1998). Therefore. it seems unlikelythatrelaying intervals would increase due to a decreaseinfood availability, although simple food availability might not reflect the availability of resourcesrequiredfor egg production.

31

Morcc:JVer, the departure ages of'lbick:-billed Murre cbicks often decline withha1chingdate (Hipfuer and Gaston 1999a). and late Common Murre dUcksweremore likely to betaken bypredaton prior to depanuceinsome studies (Hatchwell 1991, Spear 1993). TheR:fore.itseems unlikely that delaying the laying of a replacementeggwould confer an evolutionary advantage on murres.

Where researchers have collectedbreedingTbick-billed Murres at their colonies, the largest follicle presentinthe ovary increasedinsize up to approximatdy the median laying date,then deaeased (Swartz 1966, Gaston and Nettlesbip 1981). Thismight reflect an increaseinprimacy follicle size up to laying, followed by atresiaofaccessory follicJ.es (CommcxlMurres collected with fully-sbellcd eggsinthe oviduct had other follicles deveJoping as well, Mmphy and Schauer 1994).

Therefore, a murrethatloses her egg soon after laying may have a head start in the form of a larger secoDd folliclewithwhich to produce a replacemcnt egg.IncoDjunction with this,aDd perhaps more importantly, her bonnooal balance might. be more favounble for egg produetiOD becauae shehas incubated for less time, e.g., perilaps she would have higher levels offollicie-stinnLIating hormone, and lowerlevels of prolactin (WLDgfieId and Famrr 1979, Zadwomy et aL 1989).

Asobserved at other Thick-billed Murre colomes (Birkhead andN~eship1982), and previously at Coats Island (de Forest 1993, Hipfuer et a1.. 1997), replacemeDt eggs averaged 6%

smaller thanfirst.eggs,andthe sizes of first and replacementeggswere strongly correlated. None of thefactorsthatI examined influenced the relative changein sizebetween first andreplacemeol eggs. This suggests that replacemcmeggsize,like£cst egg size (Hipfilcr et al 1997),is rdativdy fixed. Birkhead and Nettleship (1982) suggestedthatfemale anures might laysmaUrepl..acement eggs inorder to advancetheir laying date. A previous study of Common Murres confirmedthat the smaller replacement eggs were formed more quicklythanfirst eggs (Birkhead and del Neva 1987),

32

and indirectevideoce suggested thatthiswas true inThic.k:-billcdMurres as wdl(Hipfueretal 1997). Howeva-.the wing fcatben ofTbick-billed MUJTecbic.k::sthatbatchfromlargeeggsgrow morequicklythan dothose on cbicks from small eggs (perhaps becawelarge-eggchicks hatchwith larger yolk: reserves, Birkhead and Nettlesbip 1984). and rapidgrowthofwing-feathers may enable the chick to go to sea quicldy(HipfuCl" and Gaston 1999 b). Con.sequc:ntly, there might be little time- saving advantage forThick:..()illedMurresinproducing small replacementeggs..Itstillisunclear why murreslaysmall replacement eggs, although timc-savingremainsa possibility.

Tosw:nmariz.e.these resultsshow that Thick-billedMurres that layearlyinthe season (probably older, more experieoced females) have considerable capacityto produce replacement eggs.

Layinga replacement egg might coofer direct and/or indirect fitness benefits on murres _this warrants further investigation (see Chapter 3). The capacity of bigh.qua.lity female Thick-billed Murres to produce replacemc::nl:eggs until very Late in the layingperiodcontrasts with observations00some tempe:nte-oestiogbirdspecies,and mightthereforebave imponant implications about the seasonality oflow-A.rctic marine environments.

33

2.6 LitenhU'e Cited

Ainley,D.G., R.I. Boekelheidc, S.HMorre~andC.S. Strong. 1990 Cassin's:Auldet.Seabirds of the Fanillone Islands:: Ecology, dynamies: and structure ofan upwelling-system community (eels.

D.G. Ainley and R.I. Boekdbeidc), pp. 306-338. Stanford UniversityPress:,Palo Alto.

Arnold,T.W.1993.FaaonaffectingrenestiDgin AmericanCoots:.Coodor95: 273-281.

Barba, E.,lA Gil-Delgado, andI.S. Monros. 1995. Thecosts ofbei.Dg late: consequences:ofde1ayiDg Great TitParvs major first clutches. Journal ofAnima1Ecology 64: 642-651.

Birkhead, T.R.,andAI. del Neva. 1987.Eggformation and the prc-Iaying period ofthe Common GuillemotUriaaalge.Jouma1of Zoology (London) 211: 83-88.

Birkhead, T.R., and D.N. Nettleship. 1980. Census methods formurresUriaspecies - a unified approach. Canadian Wildlife Service Occuional Papen.. No. 43.

Birkhead, T.R.,andD. N. Nettiesbip. 1981. ReproductivebiologyoITbic.k:-billedMurresUrialomvia;

aninter-eolonycomparison. Auk 98: 258-269.

Birkhead, T.R., andD. N. Nettlesbip. 1982. Adaptive significance ofegg size and laying date inThick- billed MurresUria lomll;a.Ecology 63: 300-306.

Birkbead, T.R., and D.N. Nettleship. 1984.Eggsize,compo:sition, and off'springqu.alityinsome Alcidae (Aves: Cbaradriiformes). Journal ofZoology (Lond.) 202: 177-194.

Birkhead., TK, and D.N. Nettleship. 1987a. Ecological reiatioDShipsbetweenCommon MurresUria aaJge and Thick-billed Mw:res Uria lomviaatthe GannetIslands:,. Labrador. ll. Breeding success and nestsite characteristics.CanadianJournalof Zoology 65: 1622-1629.

Birt.head, T.R., and D.N. Nettle:ship. 1987b. Ecological reJationsbipsbetween CommonMurresUria aalgeand Thick-billedMurresUria Jomviaat the Gannet Islands, Labrador. I. Morphometries:

34

andtimingofbreeding.Canadian Journalof Zoology65: 1621-1629.

Boekelbeide,RJ.,and D.G. Ainley. 1989.Age,resource availability, andbuedingeffortinBrandt's Cormorant. Auk106: 389401.

Brinkhof; M.W.G., A 1. Cave, F.1.Hagc.amiS. Verlwlst. 1993. Tuning ofreproduction and fiedging successinthe coot F",licaatnr. evidence

roc

a causal relationship. Journal ofAn.imal Ecology 62: 577-587.

Bryant,R., 1.L.Joncs, and 1. M. Hipfuer. 1999. Responses to. changes in prey-availability-by Common Murres and Thick-billed Murres at the Gannet Islands, Labrador. Can.adian·Jownal Zoology 77: 1278-1287.

Clutton-Brock, T.K (ed.). 1988.lU:productiveSUcce5.S.Ucivcnity ofChicagoPress,Chicago.

Conn.d1Y. J.W.,R.A.Y1SCbcr, A.D.Apa,K..P.Reese,and W.L. Wakkinen. 1993. R.eoestingbySage grousein Southeastern Idaho. Condor 95: 1041-1043.

Croll, D.A, AJ. Gaston, and D.G. Noble. 1991. Adaptive lossofmass in Thick-billed Murres. Condor 91: 496-502.

Daan, S., C.Dijkstra, aDd JM. TIDberge:n. 1988. Familyplanning intheKestrel: the ultimate control ofcovariation of laying date and clutchsize..Behaviour 114: 83-116

de Forest,L.N.1993. Theeffectsof age,timingof breeding, and breeding sitecbaracterist:i.cson the reproductive success of the Thick:-billed Murre. MScthesis.,UniversityofOttawa.

de Forest, L.N.• andAJ.Gaston. 1996.The effect of ageontimingofbreedi:ng and reproductive successintheThi<:k-billed Murre. Ecology 77: 1501-1511.

Duncan,. D.C.1987. Nesting ofNorthe:m PintailsinAlberta: laying date, clutch size, and renestiDg..

CanadillDJownalofZoology 65: 234-246.

35