UMI TO

NOTE TO USERS

Pagels) missing in number only; text follows.

Microfilmed as received.

18

This reproduction is the best copy available.

UMI

Chick diet and nestling condition among Atlantic Puffins at three Nonhwest Atlantic colonies

by

©Janet Russell

A thesis submitted to the School of Graduar:e Studiesinpartial fulfillment of the requirements for the degree of Master of Science

Biopsychology Program Memorial University of Newfoundland

S1. John's, NewfOWldland, Canada 1999

Abstnlct

Seabirds forage in a variable environment. Theoretical investigations into seabirdforagingeook)gyand practicaleffortsto utilizeseabirdsas sampling agents of marine dynamics havebeenhampered by an inactequate base of natural histOlYdata.Priortothisstudy, the nestling diets ofAtJanticPuffinsaJong the coast of insular Newfoundland inthenorthwest Atlantic were assumed tobe homogeneous. Successful breedingbyAtlantic Puffins intheregion had been linked to the availability of mature capelin wtlich was thought tobean essential componentofpuffin nestling diet due tothepaucity ofSlitab6ealternative prey.

This studyreportsdata onthediets and condition of nestlings collected at Baccalieu, Funk and Small Islands between 1992 and 1995. BiN-loads offresh preywerecollectedfrom aduttAtlanticPuffins provisioning nestlings. BiN-load size and the contribution to the diets made by mature capelin were compared between areas and years in conjunctionwithnestling COndition, measured as body mass at8given wing-length. Proximate organic composition analyses were conducted onthelesser knownpreyitems and energy densities estimated.

These datawereused to test the assumptions that 1) the dietsofpuffin nestlings along thecoastof insular Newfoundland are homogeneous and 2)that there are nopreyof comparable quality to mature capelin availabie to breeding puffins in the region. Thedatadonotsuppport either assumption. Dietsof nestling puffins along the northeast coast of insular Newfoundland are not

homogeneous. Rather than uniformly dominated by mature capetill, the dietsof nestling puffins on Funk and Small Islandsweresometimes composed principally of either posdarvat o-group sandlence or larval 1.group capelin. The condition of nestlingsfedonpost~larvalo-group sandlance was the bestobservedin this study. exceeding that of nestlingsfeddiets dominated bymattrecapelin. 1.group capelinhada wet energy density similar to male capelin while postlarval O-group sandlancehadawetenergy densitywhichexceeded that of mature cape6in during a comparable timeofyear.

Iconclude that thedietsofAtlantic Puffinnestlings inthenorthwest Atlantic are more variable than previously considered, and that there are occasions whenpreyof at least comperab'e energetic density to thatofmature capetin are available to adult puffinsprovisioningyoung in this region.

Support forthisstudy eame in manyformsfrommany sources.Thereare alotofthingsthatI heve mysupervisor.BillMonteveochi,to thankfor.

Paramount amongthemhasbeen themoral support and open-mindedness that are Bill'suniQuegifttohisenvirons andthete anttimeswhenitmade all the difference in the wottd. My committee member Ram Myers began thetaskof teachingmehowtolookatdataandprovidedme wtth much logisticalandrT1Of1N support. Thegoodnature of both of thesepeoplesparedme much grief and I would like to thank them for their tolerance of all my deviations.

Iwas more than fortunate in the field assistants that Bill provided. Sian Frenchin1992andJillCaseyin1993werebothhetpful colteaguesandtheir company SO good I stiN enjoyit..Ulrike and Ecke Zuschlag and Nick Monteveoc:tli alsoc:oHecteddataforthisstudy. Pierre RyanandCaroline Walsh\¥lM'8~pful on FunkIsland. Dave MethvenletmetagaJongonsome of hisbeachseining tripsandstealsome fish.Along withDon 5. .andJohn Greenhealsohelped identify prey specimens.HieBarrowman, courtesy of Ram, showed mehowto useS-Plus and8rlSW'8f'8damillion stupidquestions8$ifthey 'N8f8not. David Schneidermadehelpful comments on my thesis proposal earty on.

Thecanadian CoastGuard maderoom for us on more than one heticopter and provided use of the lighthouse on Baccalieu where the lighthouse keepers were mosthospitable.I especially wish to thank Linus Walsh and Tony

Broader.s.VVhentherewasno room on a c::hoppertheHyde's ofRedheedCove gotus safely onandoffofBaccalieu.Wallace WeHondid thesame forSmall I.nd CWldtheEaston',ofCannanvittemadethesteamto Funk a

P'eaute.

BernieceSt Croix, Kim French,BrendaNoftaU Md Maureen Howardofthe PsychologyOepertmentwerealwayshelpfulwhenI neededitDonnaButter, Bill's right hand. did morethingsformethanI probably even realize.

FinancialsupportwasprovidedbyaMemorialUni¥efsityofNewfoundland Fellowship and Canadian WildlifeServiceStudent Grant to mysetf andby NSERC funding to W.A. Montevecchiwhichpaidfor fiekt assistants. Supportof various forms also came from the Department of Fisheries and Oceans both directty to W.AMontevec::chi and throughR.A.Myers.

The most important support came from Merrill Francis and myparents.

TABLE OF CONTENTS

Abstract . Acknow6edgments ..

Table of Contents ..

List of Tables . List of Figures . List of Appendices ..

Chapter1.Introduction ...

1.1 GeneralIntroduction...

1.2 Study sites . 1.3Study objectives...

. ii

. iv

. vi

... ix .. xiii

. xviii

.11 Chapter 2. Prey Composition of Atlantic Puffin nestling diet ... . 19

2. 1 Introduction.... . 19

2.2 Methods.. . 22

2.3 Results.. . 25

2.3.1 Size of prey types... . . 25

2.3.2 Contribution made by mature capelin to puffin nestling diet... ...26 2.3.3 Contribution madebyprey otherthan mature capelin to

nesUing diet . 28

2.3."Size,mass andenergycontent of biM·loads deliveredto puffin nestlings ... ... 33 2.3.... 1 Numberofindividualpreyitems in bill·loads

cottected from Atlantic Puffin nestlings .. ..33 2.3.4.2Mass (9)ofbill-loads collected from Atlantic Puffins

provisioning nestlings.. . 34

2.3.4.3Estimatedenergycontent(kJ)of biH·1oadscollected from Atlantic Puffin nestlings... ..35 2.3.4.4Theretatiooship between bil140adsiZe,mass and

2.5Disc:usskHl ..

energy content .. .... 36

...38 Chapter 3. Proximate composition of prey delivered to Atlantic Puffin nestlings

3.1Introduction .... 82

3.2 Methods .. 83

3.3Resutts .. 86

3.4Discussion .. 88

3.4.1 Water content ... 88

3.4.2 Lipid content .. 90

3.4.3 Protein content .. 91

3.4.4 Ash content .. 92

3.4.5 Energy densities .... 93

vii

. 109

. 112

112 ... 105

4.3.1Intra-c:otony variation in nestlingbodymass....

4.3.2Inter-annual and inter-colony variation in nestling

bodymass... . 113

4.4 Discussion.. 116

Chapter 4. Atlantic Puffinnestling condition 4.1 Introduction ..

4.2Methods ...

4.3Results ..

Chapter5.0Summary . Appendices . Literature Cited .

... 145

. 152

. 179

viii

Table2.1. Mean mass and standard deviation(g)peremlengthdass of majorpreytypesdeliveredtoAtlantiCPuffin nestlings within the study

area between1992and1995. 53

Table 2.2. Criteria fortheassignment of mass(g)to preyitemsrecorded without measurements of either mass or length and forwhichno other

mass data are availablefromthis study. 54

Table 2.3. Criteria for the assignment of energetic value (kJ/gwetmass) to prey itemsfor thecatculation of percentcontributionmadeby different prey to the energy contentofnestling diet... 55

Table2.4. The median and range of mass and total lengths of the mainprey types collected from Atlantic Puffins provisioning nestlings on Bacealieu. Funk and Smalll*nds between1992and1995combined

...57

Table2.5 Proportional representation of mature capelln in bill-loads collected from Atlantic Puffins provisioningnestlingson Baccalieu, Funk and Small

Islands between1992and1995 . 58

Table 2.6 Numberofindividualpreyitemsinbill-loeds c:dlected from Atlantic Puffins provisioning nestlingsonBaccalieu, Funk and Small Islands

between1992and1995 . . 67

Table 2.7 Mass (9) of bill-loads collected from Atiantic Puffins provisioning nestlings on Baccalieu, Funk and Smailisiands between1992and

1~... . ~

Tabte 2.8 Estimated energy content(kJ)of bill-loads collected from Atlantic Puffins provisaoning nestlingsonBaccalieu, Funk and Small Islands

between1992and1995 .. 69

Table3.1. Proximate composition of dietary items of nestling Atlantic

Puffins. 100

Table3.2. Energy densityestimatesforprey itemsfedto nestling Atlantic

Puffins. ...101

Table3.3. Proximate composition and energy density data forbeachseine

caught fishin1994. . 102

Table 3.4. A sampling of conversionfactorsinuse for estimationofenergy

density from proximate composition data 103

Tab6e3.5 Recent energydenSityvaluesforaduttcapelininsummerreported

from thenorthwestAtlantic... 104

Table 4.1. Results of an anatysis of covariance comparingtherelationship between Atlantic Puffin nesUing mass and wing-length sampled on

Baccalieu Island in1992and1993.... . 125

Table 4.2. Results of an analysisofcovariance comparing the relationship betweenAliantic Puffin nestling mass and wing4ength at the foursub- colonies sampled on Baccalieu Island in1992and1993. 126

Table 4.3. Results of an analysisofcovariance comparingtherelationship between Atlantic Puffin nestling mass and wing-ktngth on Baocalieu.

funk and Small Islands in1992 - 1995. . 127

Table 4.4. Results of pre-P'anned comparisons from within an analysis of covariance (reported in TatMe 4.2) comparing the relationship between Atlantic Puffin nestling mass and wing-length on Baccalieu, Funk, and

Small Islands in1992·1995. . 128

Table 4.5. Results of the re-running of an analysis of covariance(reported in Tab'e 4.2) comparing the relationship between AUantic Puffin nestlingmassandwi~minusthetwolargestpositive andthe twolargest negative outliers. ... 129

TatHe 4.6. Results of pre-planned comparisons from the re-runningofan analysisofcovariance (reported in Table 4.4) comparing the relationship betweenAtlanticPuffin nestling mass and wing-length on Bacealieu, Funk and Small Islands in 1992·1995 minus thetwolargest positive and thetwolargest negative outliers. ... 130

ListofFigu....

Figure 1.1. Mapofthenorthwest Atlantic showingthelocationofseabird breeding islands sampled for this study (Baccalieu. Funk andSma"

Islands) or mentioned inthetext (Great Island) .. 13

Figure 1.2 Topographtcmapand aerial photographofBaccaHeuIstand showing the location ofthefour Atlantte Puffin sub-<:olonies sampled during 1992 and 1993 indudingWoodyCove, theonty sutH:06ony sampled in 1994 and corresponding to the top right arrow . 15

Figure 1.3 Mapofthenortheast coastofk\sular Newfoundland showing the location of seabird breeding islands sampledtorthis study (Baccalieu.

FunkandSmall Islands) and their proximity toshore. 17

Figure 2.1 Percent OCCUrTenCe of adult capelin in sampled Atlantic Puffin nestling diets on Bac:caHeu, Funk and Small Islands during 1992 - 1995.

Grey shading

=

adultcapelin,blackshading

=

^{other.. . 59}

Figure2.2 Percent frequencyofadutt capelininsampMd Atlantic Puffin nestling dietsonBaccalieu, Funk and Smallislanc:ts during1992·1995.

Grey shading=adutt capetin,~shading=other.. 61

Figure 2.3 Percent massofadult capelin in sampfed Atlanbc Puffin nestling dietsonBaccalieu,Funk and Small Islands during1992·1995.Grey shading=adult capelin,blackshading=other... 63

Figure2.4 Percentofestimated dietaryenergycontent represented by adutt capelln in sampled Atlantic Puffin nestling diets on Baccalieu, Funk and Smallls!anc:ts during1992 - 1995.Grey shading=adult capetin.black

shading=other... 65

Figure2.5 Boxplotsshowingthesize(numberofpreyitemsperbill.Joed) of bill·loads delivered by &dutt Atlantic Puffinsto nestlings on Baccalieu, Funk and Small Islands during1992·1995. Boxptotbrackets indicate the range of values,blackboxes the inter-quarti6e range, whitelines through black boxesthemedians and dashes the outliers (defined as points that are further away from the median than1.5times theintef".

quartilerange).The grey shading indicatesthe 95%confidencelimit

around the median... .. 70

Figure 2.6 Boxptots showing themass(g)ofbilf.4oads delfver'edby adult Atlantic Puffins to nestlingson 8accalieu, Funk and Small Islands during 1992 - 1995.80xpklt braCkets indicate therangeof values, b&ackboxes the inter-quartile range, white lines through black boxesthemedians and dasheStheoutliers (defined as pointsthat arefurtherawayfromthe median than 1.5 times the inter-quartile range). The grey shading indicates the 95 %confidencelimit aroundthemedian. . .. . 72

Figure 2.7 Boxplol:s showing the estimated energy content (kJ) of bill-loads deliveredbyadult Atlantic Puffinsto nestlings on Bacc8lieu, Funk and Small Islands during 1992· 1995. Boxplot brackets indiC!te therangeof values,~boxesthe inter-quartife range,white lines throughblack boxes the medians and dashestheoutlier5 (defined as points that are further away from the median than 1.5 times the inter-quartile range). The grey shading indtcates the 95 % confidence limit aroundthemedian ... 74

Figure 2.8 Bill-load mass (g) versus the number of prey items per bill·load for all locations and ye8r5 sampled duringweek5 ofthenesttingperiod on Baccalieu, Funk and Small Islands 1992·1995 76

Figure 2.9 BiM-1oad mass (g) versus estimatedbill-1oad energycontent (kJ)for all locations and years sampled during week 5 of the nestling periOd on Baccalieu, Funk and Small Islands 1992 - 1995... . 78

Figure 2.10 BiII40ad size (number of prey items per bill-load) versus estimated bill·1oadenergycontent (kJ)forall toc:ations and years sampkJd during week 5 of the nesttingperiodon Baccalieu, Funk and Small I$lands 1992

-1995... 80

Figure 4.1 Total numberofmeasurementsofnesUing body mass (g) and winst- length (em)collectedat~ieu,Funk and Small Islands during

1992 - 1995 By date... . 121

Figure 4.2 Number of independent measurementsofnestlingbOdymass (g) andwing-length (em)colk!ctedat Baccaltau, FunkandSmall Islands during 1992·1995 by date and used as the input data for an analysis of

covariance... ... 123

Figure 4.3 Relationship between Atlantic Puffin nestling mass and wing-tength between the years of 1992·Ht94 on Baccalieu Island. . 131

Figure 4.4 Relationshipbet'NeenAt1antic Puffin nestlingmassand wing-length betweentheyears of1992-1995on Funk Island. . ..133

Figure4.5 Rekltionship between Atlantic Puffin nestlingmass and wing-ktngth between the years of1994-1995on$malll.nd. . 135

Figure4.6 Relationshipbetween Atlantic Puffin nesUing mass and wing-length

on Baccalieu and Funk Islands in1992. . 137

Figure4.7 RNtionSh~between AUsntic Puffin nestling mass and wing-tength

onBaccalieuandFunk Islands in1993. .. 139

Figure4.8 Relationshipbetween Atlantic Puffin nestling mass and wing-length on Baccalieu, Funk and $mallislands in1994. ... 141

Figure4.9 RelationshipbetweenAtlanticPuffin nesUingmass and wing-iength

on Funkand $mallistands in1995. . 143

xvii

ListofAppendices

Appendix2.1 Proportionalrepresentation of 1-groupca~inin bill-loads collected from Atlantic Puffinsprovisioningnestlings on Baccalteu, Funk and Small Islands between1992 and 1995 .. . .. ... ... ... .... ... ... '52

Appendix2.2 Proportionalrepresentation of adult sandlance in bill-108ds collected from Atlantic Puffins provisioning nestlingson8aocalieu, Funk andSmall Islandsbet'Neen 1992 and1995 .. . 153

Appendix 2.3 Proportional representation of laNai G-group sandlance in bill-toads collectedfrom Atlantic Puffins provisioning nestltngs on Baccalieu, Funk and Small Islands between1992 and 1995 154

Appendix 2.4 Proportional representation ofpostl8N81Q..group sandlance in bill-loads collectedfrom Atlantic Puffins provisioning nestlingson Baccalieu, Funk and Small Islands between1992 and 1995 155

Appendix2.5 Proportional representation ofStichaetdsin bill-loads collected from Atlantic Puffins provisioning nestlingsonBaccalieu, Funk and Small Islands between 1992 and 1995 .. . 156

_ _ _ oICottidaein_lXlIKted fromAtlantic Puffinsprovisioningnest1ings on Becc:alieu, Funk and Sma"

IsIMdsbetween1992and1995 .. 157

Appendix2.7 Proportional representation ofjuvenileWhite Hake in bill·

loadsCXlIIeCted fromAttanticPuffins provisjoningnestlingsonBac::caMeu, Funk and Small Islands between 1992 and 1995 .... '58

Appendix 2.8 Proportional _ of().groupHetTinginbill..Jo8ds c:oIIectedfrom Attentic Puffinsprovisioningnestlings onBaocaIieu, Funk and Small Islands between 1992 and 1995.. . 159

Appendix2.9 Proportionalrepresentation ofAgonidae inbil140ads collected fromAtlanticPuffinsprovisioningnestlings onBacc:aIieu.Funk

andSmaIIlstands between 1992and1995 160

Appendix 2.10 Proportional representation ofUparis$p.in bill-loads collected from Atlantic Puffins provisioning nesttings on Baocalieu, Funk andSmal1lSiands between 1992and1995 .. _{. 161}

Appendix2. 11 Proportional representation of squid inbiH-loadscollected from Atlantic Puffins provisioning nestlingsonBaccalieu, Funk and Small

Islands between 1992 and 1995 .. 162

Appendix 2.12 Proportional representationofcrustaceans in bitl-toads COI6ected from Atlantic Puffins provisioning nes1tings on Baccalieu, Funk

and Small Islands between 1992 and 1995 163

Appendix4.1. Scatterpk)tsofAtlantic Puffin nestling bodymassand wing-length measurements for all locations and years sampled 164

Appendix4.2. Scatter plotsofthe truncated datasets on AtlantiC Puffins nestling body mass and wing..Jength used as input for analysis of

covariance. .66

Appendix 4.3. The residuals from an anatysis of covariance on AtlantiC Puffin nestling body mass and wing-length plotted against the quantiles

of a standard normal distribution. 168

Appendix 4.4. The residuals from an analysisofcovariance on At1antic Puffin nestling bOdymass andwing-length plotted against the log of nestling wing length(theindependentvariable fromthesame anatysis).

'70

Appendix 4.5. The residuals from an analysis of covariance on Atlantic Puffin nestling bOdymassand wing-length pkrtted against the predicted

values fromthesame analysis. . 172

Appendix4.6 Summary of measurements of puffin nestJing wing-length (em) and bOdy mass (g) includinglongitudinald8ta .. . 174

Appendix4.7 Summary of measurementsofpuffin nestling wing-length (em) andbOdymass (g) used in the analysisofcovariance reported in

Chapter 4 .. .. .. . 177

1.0 Introduction 1.1 General Introduction

Studies comparing the dfetsofc:onspedfic seabird species at dif'fefent colonies andatthe same colony between years are useful in assessing the envlrotlmental variation encountered by a speOes in time andspace(e.g. Hams and Hislop 1978; Schneider and Hunt 1982;Barrettetat 1987; Anker-Nilssen and lorentsen, 1990; Baird 1990; Hatch and sanger 1992; Bertram and Kaiser 1993; Montevec:chi 1993; Monteveechi and Myers 1995; Springer et al. 1996) If one assumes that conspecific seabirds at differentcolonieSinthesame oceanographic region have the same basicfoodrequirements for survival and rearing offspring it 'NOU1dfollOw,all other things being equal, thatthese conspecific seabirds would employ roughly equivaient foraging strategies. If this assumption is valid, differences in conspecific seabird dietsbet'Neenlocations and at different times wouldbedue to variation in local menus.

The two general sources of variation in seabird nestling diet are parental foraging behavior and foraging conditions. In addition, thesetwosources of variation interact. There is evidence ofchangesin seabird nestting diets in association with changes in the availability of commonly taken prey specieswhich demonstrates the plasticity of seabird foraging behavior (e.g. Hislop and Harris 1985,Barrett et. al. 1987, Martin 1989, Anker-Nilssen 1992, Montevecchi and Myers 1996). seabirds arenotsorigid in their behavior that in theabsenceOf usual prey they refrain from Pl"ovisioning their young. Rather, they Pl"ovision them

diffefentty,oMthvaryingdegreesofsuccess (e.g. Anker-NiItsseo 1987,Martin 1989).

ThedynamicintenIclion betweenforagingconditions andseabird foraging behavior complic:ates the interpretation of seabird nestling diets with regard tothe foragingconditionstheyrepresent(seeHuntet81. 1991 fot" an exc:eUent discussionofseabirds as sampling agents). Because oftheflexible nature of foragngbehaviol'"the potentialrangeofsuchbehavior within aseabird speciesis unlikely to change01"1the temporal scaJeoMthwtltchweare interested, i.e.00the scale of decades, unless thereisanewselection pressure exerted within that time frame such8Sa profound change in environmental variability.Thepotential range of foraging beh8vtor under given conditions is roughly illustrated at any given time bythevariation innestlingdietsamong c:onspecific individuals present within a colony.

Theinherently flexiblebutoftenpredicUII:llenature ofputl'inforaging behavior illustrated by long-term studies inthenortheastAtlantic (Anker·Nilssen 1987, Martin 1989) is consistentoMthassuming thatthedrivingforce behind variation in nestling diet among years at a given colony or amongooIoniesin a given yearisnotdiffetencesinthepotentialforagingbeh8vior of the parentsbut ratherdiffefencesintheforagingenvironment,the same variabfeenvironmentin whichit'MJU1dhave been lIdvantageous to have evolved aplastic fonIging phenotype. Changesinnestling diet variationcantherefore reasonablybe assumed to reflect a combinationofchangesin for8ging conditions and the

interactionbetweenfcngingconditionsandperentaIforagingbehavior.ttis unlikely thattheyre1'Iec:tchangesinparentalforaging behavioralone asthisisnot expec::ted to d1ange fromyear to year. Ditfefenoesinnesttingdietamong conspecific seabirdsnestingat differentc:ofoniesor atthe same cok>nyin different years therefore to some degree measureditfefencesin foraging conditions. However. differences in foraging conditions arenotalways manifest as differences in nestling diet. For exampte, in a yearof verytowsandlance abundancenearShetland,SCotlandmurrenestling diet was still dominated by sandlanc:ebecauseadultswereableto increesetheirforagingenergy expenditure to (X)lTlpensate forthe lowerabundanceofpreferredprey(Monaghan etal. 1994, 1996). Asa consequence, nestlingdietdidnotretIectLargechanges inpreyabundance.

The ability of adult seabirds provisioning chicks to adjust foragingeffort and thus maintain stability in nestling diet composition complicates the use of nestlingdietas an indicatorofchanges inpreyavailability. When changes in nestlingdietcompositiondooccur it is thereforelikely,butnot necessarily the case, thatsuchchangeshavebeenprecededbyeffortstomaintain historical nestling diet compositionby inc:r-easingparental foraging effort (e.g. increasing foraging dtstsnc:e). Theabil1ty to increase the frequencywithwtMchchicks are provisK>ned is another way inwhichseabirds provisioning young can compensate forchanges in foraging conditions (e.g.UttIeyetal.1994but see also Burger and PiaU 1990). Increased feeding frequency can compensate for a Cleerease inthe

dietaryvalueof individualmealSbutdoesnotmaskch8nges inthespecies compositionofnestlingdiet

Feedingconditionsthatatfedthecomposition ofseebit'dnestlingdietare a compositeofthepresence, abundance,relative.boodanceandbehav;or of variouspreytypes1)withinthe bird'sforagingrangeofthe ooIony,2)withinthe bird's foraging depth range and 3) dUring a timeperiod whichover1ap$withthe bird's nestlingperiod.

Puffin COlOniessituatedoffthe northeastcoastofinsularNewfoundl8nd (Fig 1.1) are aU under the influence of the inshore branch ofthelabrador Current which is a mixture oflowsalinityHudsonBay water andthearctic waters of the BaffinIsland Current which then

now

thelength of labrador andthe east coast of insular Newfoundland (Drinkwater1996).InlerannU81 variabilityinforaging conditionsoff insular Newfoundlandis Iargefya consequenceofinterannual variation inthe temperature and salinity oftheoceanandits effect on somatic growth ofprey,phenOlogyofpreyreproductionand behavior,distribution r d preyavailabilitywfthintherMwntforaging dtstance anddepthfor the seabird speciesunderconsideration (e.g. Nakashima 1994,Carsceddenel. al. 1991).

The main sourceofintercolony variability in foraging conditionsina given year aredif'rerencesinkx:aIbiophysicalooncfitions,someofwhichareconstllnl such as bottom topography and some of which varyfromyear to yearSUChas water temperatureand S8linity.

large scale pe4agic surveysconductedoff the noftheast coast of insular

Newfoundlandand thecoastofsouthernlabrador since 1994 dividetheregion into four broad scale zoogeographic domains: 1} the shelf waters of the Northeast NewfoundlandShelfand the SouthernlabradorShelf;2)theNorthern Grand Bank: 3)theSouthem Grand Bank: and 4) the inshore bays alOng the northeast coast ofNewfouncland(fromAnderson andDalley1997a). Puffin colonies offthe coast of insular Newfoundland differ in their proximity to the different zoogeographic domains identifiedabove.Starting from the south,theWItiess Bay seabird colonies (e.g. Gun and Great Islands where most data on Atlantic Puffins in thenorthwestAtlantic havebeengathered) arelocatedinShOre ofthe deep Avalon Channel and directly adjacent to the extensive and shallow Grand Banks (e.g. Nettleship 1991, Roc:tway and Monlevecchi 1996). Baccalieu Island is loCated inshore inllMativ~shOal free deep water off the headland separating Trinity and Conception Bays and near the boundarybetweentheGrand Banks to the south and the Northeast Newfoundland Shelf to the north. Small Island, in the Wadham Island archipelago is located inshore in shoal waters, just south of Notre Dame Bay with thedeepNortheast Newfoundland~fdirecily offshore.

Funk Island is in shoal waters surrounded by deeper water and is direcilyoffshore from Small Island and within the NortheastNewfoundlandShelf area. The above descriptions are based on the broad-scale delineation of the Newfoundland region into subareasadoptedbythePelagicJuvenile Fish Survey conducted by the Canadian Department of Fisheries and Oceans (DFO) ineachof the years since 199-4(Andersonand Dalley 1997a).

Cac*in in thenorthwestAdanticfor themostpartmigrateinshore during thesummer months tospawnon or near beaches (Templeman 1948, Ganscaddenetat1989). WtWeinshore,spawningcapelinprovideanaggregated source of lipidrich food withinforagingrangeofall puffinsbreedingoffthe northeastcoastofinsul8rNewiot.n:Iand.Becausetheyswiminaggregations, spawning schools of capetin, once encountered,providethe opportunity for seabirds to forage exdusively on one species andmaturitYstageofpreyand this has generallybeenObSerVedtobe thecase for Atlantic Puffins proviSioning young (Brownand NettJeship1984, Piatt 1987, Creelman and Storey 1991). The arrival of capelin inshore tospawnvaries both annually (e.g. roughly 1 month later in the early1990$than duringthe 19805) and geographically, i.e.

progressivefy later as one moves north alongtheIlOrtheast coast of insular Newfoundland (Templeman 1948, Shackelletal. 1994, NakashimaandWinters 1996,Carscaddeneta'- 1997, Therriaultetat 1996). The puffin nestlingperiod is sufficienttylong50 thatdirect0Yef1apwith the availabilityofspawningc:apelin whilenotgauranteed for the full durationofthe nestling provisioning periodis boundtoovertapwithsomeportionofit (HarrisandBirkhead 1985)."Nhich portionovenaps andfor howlongisofrelevance to puffin nestling diet

The diets of Atlantic Puffin (Fratere:ula afCtica) nestlings havebeen compared amongcolonies and years throughout much of the northeast Atlantic (e.g. Corkhill 1973, Harris and Hislop 1978,Ashcroft 1979, Barrettetat.1987, Martin 1989. Anker-NilSsen and Lorentsen 1990) but not in the northwest Attantic

(i.e. limrtedtoNetUesh~1972, 1991, CreeIIlwll991, RedwayandMontevecx:hi 1996). In a reviewofthe importanceofmature capelin toNewfoundland

sea_,

en_. _ _

I'984l00ndud0dlhat_---foo'

thesucx:essfulftedgingofAttantic Puffinchicks and that therewereno suitable alternate prey available insoutheastNewfoundlandwaters.NetUeship(1972) assumed further that Atlantic Puffin nestling diets at various colonies along the northeast coastof insular Newfoundland(namelyFunk, Small and GreattsIands) were essentiallythesame. Absence of mature capetin in puffinnestingdietsat Great Island has beenassociatedwith breedingfailure(Nettieship1991).

ProlongedperiodsofbreedingfaikJresattributed toscarcityofenergyrichprey have been reportedinnorthernEurope (Anker·Nituen 1992,Baird1990)and concern has been expressedthatthe breedingsuoc:essofseabirdsin Newfoundland is dir8Ctty dependenton theavailability inshore ofspawning capelin (Brown and NettleShip 1984, Nettleship 1991).

Although puffins spedaliZe in feeding on smallpetagicfish, prey harvests are often variable (Cor1dlilI1973. HarTis 19&4,Ba«ett.et81. 1987, Martin 1989).

However,themenu offered puffin nestlings by provisioning adutts atthe major NewfoundlandCX)Ion~appeersmuch lessvaried than inthenortheastAtlantic.

In all periods sampled, capelin dominated the diet in the northwest Atlantic (Brown and Nettleship 19&4,Bradstreetanc:lBrown1985, Piatt 1987, Creelman and Storey 1991, Rodway and Montevecchi 1996).The apparent consistency of Atlantic Puffin nestlingdietsin the northwest Atlantic in contrast to the diversity

experienoed inthenortheaStAd8nticmaybean artifact of the shorttimeseries of dataavailabletorthe northwest and/or the smallgeographicarea overwhichthe datawet8cdIected.Allpublisheddatafor puffinnestlingdietoff insular Newfoundlandhas beencollectedateitherGufIOf"Great Island,bothlocated in doseproximitytooneanotheroffthesoutheastcoastinWidessBay (F9J1"81.1;

Piatt1987,Creelman andStorey 1991,NetUeship1991, Rodwayand Montevecchi1996).

AtJantic Puffinsbreedin several locations stong theeastcoast of the islandof insular Newfoundland (cairnsetat 1989) and offthe southernlabrador coast. Major breeding sites are indicated in Figure 1.1. Breedingadultsanive at their respectivecokJnies in earty April at Baccalieu and Great Islands (Harris and Bif1thead 1985)and departinAugustlSeptember. Breedingphenologyisvari8ble andprogressivelylater as one movesnorthwardfromGreat Island.Oneeggis laid in a ground burroworless frequentlyinarodtcrevice and incubated for roughly 40-45days. Aftera~periodof6-7days,hatchednestlingsare able to thermoreg\nate and the continuous presence oftheparents is no k>nger required. Both parentsprovisiontheyoungwithfood(CreelmanandStorey 1991,Corkhin 1973)whichiscarried crosswiseinthe bill anddeliveredto lhe d'licks in afreshstate. Nesttings arefedbytheparentsuntilthey ftedge.

Feedingoccurs onthe order of 2to8 timesperday during daylight, usuallyfor 38-41 days,withextremes to more than 70 days (Harris and Birkhead, 1985).

1.2 StudysillM;

Data on the diet.-1dconditionofAdanticPuffin nesttingsVlIl8feooIIected fromBaccalieuIsland (48°09'N.s2'"48W) M11992, 1993and1994, Funk Island ("9⁰"S'N. 53011W) in 1992, 1993, 19904and1995. and SmallIslandCW8dham Islands Figure1.1; ..9⁰35'N,53° <46W)in19904and1995(see Figure1.1for location ofstudyOCJk)nies).

Baccalieu Islandmeasures aproximatety 1 x6 kmand is located approximately"km off shore fromthe northerntip oftheAvalOn Peninsula on the northeastcoast ofNewfoundland (seeFigure1.2and1.3). Thepuffinpopulation includesabout45,000pairsand is expanding (Monteveechi1996). Theisland is surrounded by diffsandsteepslopesofgrass and/or talus. Baccalieo is the largest seabird island in thel1Ol1hwestAtlantic and thereisan abundance of suitablenestinghabitatforpuffins. Colony expan$ionoutsidethemore densely occupied areasisevidentfrom newty excavated burrows.Asmall fox popu&ation has deterred guns fromnesting directlyon Baccalieu Island thoughthey donest on nearby Puffin Island (Sk'epkovych andMonIeYecchi1989).BevenspecieSof seabirds also breed on Baccalteu, including leach's Stotm-Petr8ts (OcellnodfOtrlllIeuoorhoe). murrespp.(Un.

88'Pe,

U. Iomvia), Black-legged Kittiwakes(Rissatridttetyla)and RazorbiIls (Ab tonia; Monteveochi and Tuck 1987). PuffinburrowsonBaa::8lieufsIandare usuallyonsteepseawardslopeS.

Tunnels often extendfordistances longer than 1 m. withbou4dersand scree incorporated into the subStratum creating nal'T'OW and winding tunnel paths. The

slopegradientaHowsv«ydense boo'oMng astunnebdOnotlimitthesurface area avaitable for entrances asisthecasein5eveIhabitat (i.e. Funk and Small Islands).

FunkIslandmeast.K8Sapproximatety800x400mandissituated50km offthenortheast coast of insular Newfoundland (Montevecchi and Tuck 1987;

Figure 1.3). The island supportsasmall Atlantic Puffin population of about 2000 pairs (Montevecchi unpubl. data) restricted to the one central meadow composed of gravel andthedecomposedremains oftheGreat Auks(Pinguinusimpenni$) thatoncenested andwereslaughteredthere (Montevecchiand Tuck 1987, Montevecchi and Kirk 1997).The puffinnesting . .issurrounded bybarerock onwhichmany thousands of murres nestandgannets.several peirs ofHerring Gulls(Lallls argent.tus) andGreatB*::k-b8cked Gulls(LBtlIs marinus)nest among the puffins. Bothgullspeciesstealpreybeing delivered to puffin chicks, and some GreatBlack·backedGullskilladult puffins (RussellalldMontevecchi 1996). The puffinpopulationonFunk '$!and is limited bythesmall surface area of the islandintowhich burTOWSmaybedug, themajority oftheislandbeing exposedgranite,andtheIeYeInature of the puffin habitatwhich limits the potentialdenSityofburrowsas tuMels are not deepanderosion is aprc::lbkHn. In the shallowsoiloftheis&8ncI,nestc:hambersarefrequentlyless then 1mfromthe tunnel entrance.

Smalt Island measures approximat8ty 520 x 360 m and is oneofseven islands in the Wadham Islandsarchi~gosituatedatthe entrance to Hamilton

Sound 00thenortheastcoastofinsular Newfoundland(FtgUre 1.3).AnAtlantic Puffin poputationof some 20,000breedingpairs occupy theftatmeadowyhabitat that00V<lf$roughly20 ...

"'the

island ( R _and ...1996).The island perimeter is the mostdensetyburrowed area, althoughthereis extensive suitable habitat and expansion intotheisland's center is ongoing.Ason Funk Island there are nesting Great BIack·baeked Gulls. Their presence is a recent development, and the source of retatively heavy predation 00 adult puffins (RussellandMontevec:ctli1996).Puffins onSmall Islandburrow inlevelground composed of gravel andpeetinvarying mixtures~frompurepeatto almost pure gravelwithsomeboutders. Ason Funk Island, nest chamberS are frequenttylessthan1mfromtheburrowentrance.

1.3Studyolljoc:ti.-

The present study willdescribeand compare thedietsof Atlantic Puffin nestlings from three colonies: Baccatieu (1992,1993 and 1994), Funk (1992.1995 inclusive) and Small (1994 - 1995) ISlands inthenorthwest Atlantic offthe northeast coast of insular Newtoundiand to test the assumptions that 1. thecompositionofpuffin nestling dietsoffinsu.Newfoundlandis

homogeneousand

2. that there are no preyofcomparablequatityto mature C8ptHin availableto breeding puffins in theregion

Diet quality willbemeasureddirectlythrough proximate composition analyses of prey items andindirectlythrough measurements of nestling body condition.

ThesedatawtIIconstituteanimportantextensionintimeandspaceofwhatis • relativefylimitedsetof naturalhistorydataonthe nesUingdtetsofAtlanticPuffins in thenorthwestAtlantic.

Figure 1. 1. Mepof thenorthwestManticshowingthe kxation ofmajor Ad8ntic Puffinbreedingsites atong the east coastofinsular Newfoundland and off theSOlJIhemlabradorcoast

North At/antic

13

Figure 1.2 Topographic mapandaerialphotographofBecxaiieu Is&end showingthelocatin of the four Atlantic Puffin sub-coloniessam~during 1992and1993inductingWoodyCove, theonfysub-<xl6onysamP'edin 1994 andcorrespondingtothelopright8fTOW.

35

34

I

33·

L

65

32

~

I

31I-

365000 66 67

15

Figure 1.3 MapofthenortheaStcoastof insularNewfoundland showingthe k>cationof seabirdbreedingislands sampled for this study(Baccalieu.

Funk and Smallls!ands) and their proximitytoshore.

North Atlantic

-+1

I

* 2

Trinity Bay

1 I

~Sland

~~

I . * ³

r Ip~ _{I G'rp,,~} "'\

I

(.---c

~/)

~ ... ... So,,::;.'a

I

~/

Study Sites

I ,,/J\)

*

1.

Bacca~eu

^Island

~ f-..,~ \

I

2.Funk Island /

~

;.$'-<?-(

I 3.S~III"."" ~/

#. )

17

2.0Compoaition01Atlantic Puffinneetling diets 2.1 Introduction

Thereismuc:tltheoretical and practicalinterest to how seabirds sample their environment and towhatextent an understanding of this maybeused to understand variation in seabird moftaIity rates (Harris and Bailey1992), reproductivesocces.s(Anker·Nilssenand lorentsen 1990)and indicatechanges inthebehaviour,distribution, recruitment,and/orabsoluteorrelativeabundance of preyspecies(Cairns1987,1992, Bertram andKaiser1993, Montevecchi and Myers 1995).Dataonseabirddietsalso contributeto the estimationofnatura4 mortality experienced by prey speQes (e.g. Anker-Nilssen 1992, Barrett et at 1990, Hatchand5anger 1992, RedwayandMonteveeehi1996). Changesin seabirddietsmay reftect shffts or converselythestability ofoceenographic regimes and marine foodwebs (Springeret.at 1984,MonteveochiandMyers 1996). Studies using dietaryinformation fromseabirdsdependon anadequate foundation of natural history data. Serious consideration of the general foraging ecology of Atlantic Puffins, for instance, relies on an adequatedescriptionof their prey consumption.

Forc::oIontaity breedingbirdswhich provision dependentland-basedchicks, nestling diet maybe observed without theneedtosaaificeor inordinate4y distU'b animals (Rodwey andMontevec:chi1996). The constraint on breedingbirdsof having to obtain prey from withinforagingrangeofthecolOny fortuitously limits the potential sources ofpreywhichthe researcher must conSider and facilitates

comparisons of prey harvestedwithpreyknowntobepresent inthearea at the time if such additional information is available.

P~species within the foraging range of adutt Atlantic Puffins provisioning nestlings vary in abundance, availability to predators, susceptibility to capture, digestibility and energetic and nutrient content (Bradstreet and Brown 1985).At any given time there is inter- and intra-specific variation in size, maturity, reproductive status, abundance, availability and dietary value of prey (Clarke and Prince 1980, Montevecchi and Piatt 1984, Hislop et al. 1991). Therefore, in addition to taxonomic identification, infOrmation on prey size, maturity and reproductive status isneededfor a proper consideration of differences between prey items.

Data onthediets of At1antic Puffin nestlings inthenortheast Atlantic have been reported more frequently than in the northwest Atlantic (order of100 breeding seasons compared with 10), over a broader geographical area (> 10 degrees latitude comparedwith<1 degree latitude) and from more varied foraging contexts (forage fish assemblages dominatedbyfish species other than capelin i.e. sandlance and herring). The diets of puffin nesttings in the northeast Atlantic have varied considerably both within (inter and intra-annually) and between colonies(e.g. Harris and Hislop 1978. Martin 1989). \Nhile puffin nestling diets are often dominated by a singlepreytype (i.e. capelinatthe northern tip of Norway and post-ta...a1D-group herringfurthersouthinNorway, Barrett et at 1987; O-group and 1-group sandlance in the Shetland Islands, Martin

1989),23 prey types at thespecieslevelalone havebeenreported from a single colony (Anker-Nilssen 1987). Replacement of the usual dominant prey type by another at a given site (in aSSOCiation with both failed and successful breeding) has been reported (Anker-Nilssen 1992). In contrast, the diet of puffin nestlings in the northwest Atlantic hasbeen reported as dOminated uniformly both in space and time by reproductively mature capelin (Brown and NettJeship 1984. Piatt 1987, Creelman and Storey 1991, Redway and Montevecchi 1996) with the exception of one year when the replacement of capelin by juvenile gadids was associated with breeding failure. In 1981, a year of apparent low capetin availability in Witless Bay, Newfoundland, puffins on Great Island fed their chicks 68% immature gadids, 16 % sandlance and only 10 % capetin versus 78 - 100 % capelin in the six other years previously sampled (NettleShip 1991). The apparent lack of diversity in the diet of Atlantic Puffin nestlings in the northwest Atlantic as compared with the northeast is likely an artifact of the Short timeseries and small geographical extent of sampling effort in the northwest as compared with the northeast Atlantic. Nonetheless, a comparison of the effect of kleptoparasitismby gulls on puffin breeding success at different colonies assumed that the diets of nestling puffins on Funk, Small and Great Islands didnotdiffer (Nettleship 1972).

The obteetive of this study was to test the assumption of homogeneity of the diets of nestling Atlantic Puffins along the northeast coast of Newfoundland by documenting the diversity of prey fed to nestlings at Baccalieu, Funk and Small Islands between the years 1992 and 1995.

2.2Methods

Adult puffins provisioning nestlingswereintercepted and the dropped bill·

loads collected for direct measurement. Mist-nets and fine meshed giUnets deployed vertically onpoles'Nef'eused to intercept birds in flight andwerealso used to cover burrow entrances inducing some landing birds to drop bill-loads at burrow entrances.

Bill-loads and individual prey items (unless desiccated) were weighed to the nearest 0.1 9 with 10 9 Pesola scales (1995 only) or to the nearest 0.5 9 with 100 9 scales. In the absence of 'Neigh scales of adequateprecision,themass of individual larval fish from biU·!oads containing numerous items of a single prey typeand length class was estimatedbydividing the bill-load mass by the number of fish in the bill-load. Crustaceans and the heads and tails of partial specimens were counted. Most preywereidentified to species level in the field and otherwise retainedforfurtheridentification on shore (Scott andSCott1988,J.

Green, D. Methven, D. Steele and G.H. Winters, pers.comms.). Total length of the larger whole fish specimens (snout or lowerjawtip to tip of longest tail fin smoothed back) was measured tothenearest mm using a stopped metal ruler.

Often only the range of total length wasrecordedforthe numerous larval fish from a given bill-load. Squid mantle length was measured to the nearest mm.

larval and juvenile fish are distinguished fromeachother on the basis of settling behaviour and/or metamorphosis to adult pigmentation and body form Larval fish are transparent, have not metamorphosed to the adultfofmand are

found in the uwer water column(Kamler 1992).Juvenile fish may have assumed both adult pigmentation and form and/or eXchanged pelagic lifefora benthic one.

The term larvalis usedheretorefertofishwithtransparentbodieswhile the tenn juvenile refers to fish which havenotreached adultsizes butwhosebodiesare no longer transparent Theyesr class distinction common in fishery science whereby young of the year are designated as O-group and young of the previous spawning year as 1--group is usedwheneverthereis reasonable confidence in prey age. Fish possessing the complete form, size and cdoration of adults but showing noobvioussignof&ext.IiUmaturity are refemed to as immature. The tenn post-larvalisalso used to refer to any or aH fish beyondthetransparent larval bodyphasewithoutconsideration ofage.

The larger capelin (Mallotus vlIlosus) and sandlanc::e (Ammodytesspp.) were surgically opened for examination of reproductive status. Fish containing egg masses are referred to as gravid while those with residual eggs (i.e. one to several eggs in an otherHise empty body cavity) are referred to as spent.

Reprodudivety mature male capelin are easily dentifiedbythe presence of spawningridges(Temp6eman 1948). Reproductive maturity was inferred fromthe presence ofeggsor spawning ridges in the case of capelin, and from the presence of eggs or milt inthecase of sandlance.

The estimation of percent occurrence (percent of bill-toads inwhfcha prey type is present), percent frequency (numerical frequency of a prey among all those sampled), percent mass (percent of total sampled prey mass) and percent

estimated energy(percentoftotalsampledestimatedenergy)ofvariousprey requiredtheassignment of masstothoseitemswithoutone. Mass of un-weighed specimens was estimated following R0dw8y and Montevecchi (1996) by calculating the mean mass percmlengthdan ofcomparate prey (Table 2.1).

The mass ofthOSepreywhichlacked a Hmgth measurementwasestimated by calculatingthemean mass of comparabte prey forthatsampling location. Small and/or rare itemsforwhich no massdatawereavailablefromttUsstudywere assignedaminimal mass aocon:ling to criteria outlinedinTable 2.2.

Theenergy valueofpreyitemswasestimated by multiplying the measured or estimated mass (see above) of a prey item by an energy density value (kJ/g) obtained from this study (see Chapter 3 and Tab&es 3.1 .3.'-) or extrapolation from published values(seeTable2.3fa(details). Values from this study were estimated from proximate compositiondata (dry mass energy density ( {dry energy] )=(%protein·20.0kJlg dry mass) +(%lipid·38.0kJ/g dry mass) . Ricklefs and Schew1994;'Netmassenergydensity ([ 'Netenergy) =«100·% water)· (dry energy)/100).Energy density values reported in Table2.3from Percyand Fife1981arederived from proximate compositiondatausing the mid- range values of proximate composition values reported byPercyandFifeas input to the equationsprovided above.

Thetimingofsamplingvaried between rstandsandyears limiting the inter- annual and inter-cok>ny comparisonswhichcouldreasonablybemade. Data were aggregated using a seven day'Neekasthe unit of aggregation. To facilitate

theorganisationand presentationatthedata nine standard weekscoveringthe nestling period were defined by ca~rdete beginning on July 13.

Comparisonsbetweenyears and coloniesweremade whenthereweresufficient data forweeks ofthesameordinalrank. Aminimumsamplesizeof 5 was arbitrarilyse'ectedand weekswhichdfdnotmeetthiscriteriawereeliminated from consideration.As puffin nestling diet is not homogeneous throughout the breeding season (Harris and HisJop 1978, Rodway and Montevecchi 1996) data are presented graphicallysuchthatthe temporal distribution of sampling and gaps in samplingare evident. To avoid confusion the x and y axes are consistent for those figures likely tobecompared. To facilitate consideration of differences among colonies andyears in bill-load size, mass and energy content, the95% confidence intervals around the median are shown on Figures 2.5 - 2.7.

2.3 Results 2.3.1 Sizeof preytypes

Overall thesize ofpreyitems fedtonestlings rangedfrom 1.5·21.1 em in length and 0.1 • 28.0 9 in mass (Table 2.4). Theheaviestprey itemcollected was amate c:apetin~thek)ngestwas an ovid sandlance. 1-groopcapelin and G-group sandlanceweregenerally similar toeachotherinsize(range 3.0 • 9.9and3.5 • 9.1em)withthenotab'eexception of 1995whenG-group sandlance were post-larval in form (i.e. metamorphosed to adult form, size range 7.2 • 12.9 em) and much targer thaninotheryears. In 1995 post-larval o.group sandlance

had a median mass of2.3gcomparedwith • median massof0.5gwhen inthe larvalform.

2.3.2 ContributionrnHe bymatu...capell"topuffin needingdlet8 The sample sizes relevant to nestling diet composition and the contribution made by mature capefinbypercent occorrenoe,frequency.mass and estimated dietary energy content are summarised in Tab'e 2.5. A graphK::a1 summary using barplotstoillustrate the proportionate contributionmade bymaturecapelloto the nestling dietofpuffinsatBaccalieu. Funk and SmallIslandsbetween 1992·1995 and described be'ow isshowninFigures2.1 -2.4inclusive.

Percent occurrence

OnBaccalieuISlandadult capelinwereusuallypresentin the majority of bill·loads (Figure 2.1). In 1993.theonlyyearwith fullseasoncoverage.the proportion of bill·loads containing adult capelin increased overthefirst fewweeks from an initiallowof45%to remain consistentty high for the Ianer half ofthe season at between89and 100 %. Thesame pattern was apparent in1994 although samplingooverage was not as complete. In contrast. on FunkIsland the presexe of adult capelin in bill·loads was infrequent in all of the4years sampled with a maximum occurrence of30% inone'N88kof 1993 comparedwith a range of0 •11%otherwise. Adult capelinwereonly Slightly more common in bill-loads on Small Islandwhere theyweremore oftenpresentin1994 (range: 17

·37%)than1995 (range: .. ·16 %).

-

^The

^...

^numerical

^-

^{abund8nceofaduttcapelioinsampednestlingdietwas}

greatest onBaccalieuIslandwhereit often exceeded 70 % compared with Funk and Small Islandswherepercent frequency never exceeded 5 % and was usually closer to zero (Figure 2.2). Numerical abundance on Baccalieu Island was variablewitha pattern of lower abundance duringtheearty season compared with the lale season for thetwoyears with extended sampling. Adult capelin accounted for a smalterproportionofdietary items in late 1992thaninlate 1993 and 1994.

Pereentma. .

The contributionofadult capelin by masswasgreaterthanby numerical frequencyand lessvariable (Figure 2.3). Aswithpercentfrequency,the proportionate contributionofadult capetin by masswasgreatest on Baccalieu Island compared with Funk and Small Islands. Thesame pattern observedwith percent frequency oflowerlevels in ear1y season increasing to higher levels later was evident in1993on Baccalieu Island but the differences in magnitude were not as pronounced as withpercentfrequency. Thecontribution by mass of mature capelin onSmaJllslandwas greater in1994than in1995and during 1994 was also greater on Small than on Funk lsand (58 vs 6 %respectivetyduring a comparable samplingperiod),but much less than on Baccalieu Island (33 vs 90

%respectively during a comparable samplingperiod).

Percentntimn.d dietaryenergy

The contribution to estimated nestling dietary energy by adult capelin doseIyresernbtedtheproportionaterepresentation by mass described above, both in magnitude and overall pattern (Figure2.4).

2.3.3 Contributionrn8deby preyotherthinm8tUNeapelin tonestlingdillt.

A detailed summary oftheprey types presentinthedietsofAtlantic Puffin nestlings by percent occurrence, frequency, mass and energy isshownin Appendices 2.1 - 2.12. In all,aminimum of 10 fish and5invertebrate species were present in the puffin nestling diet. While mature capelin dominated the diet by percent occurrence, mass and energy consistently on Baccalieu Island and by percent mass and energy on Small l$and inweek5 of 1994,thedtet on Funk Island ineachofthe years 1992 • 1994 was domrnated by1~roupcapelinby percent frequency, mass. occurrence and energy. In 1995, post-tarval (metamorphOSed)()...groupsandLance dOminated by percent frequency, mass, occurrence and energy onbothSmall and Funk Islands. Abriefdescriptionof results organisedbyprey type follows.

1-group C8pelin

1-group~inwerepresent inthenesttingdietof allcok)nie$samP'ed in all years, butnotin all weeks (Appendix 2.1). OnBaccalieu in 1993 when the entire season was sampled the presenceof1.group capelinwasgreatest and most important energetically early in the season. Among the three years (1992·

199-' inclusive) sampled on Baccalieu, 1-group capelinweremost common in bill-

loads and importantenergeticallyduMg1994. Amongthe threecolonies sampled 1-group capelinweremostc::onYT"IOnon Funk I.nd where they were present inmostbilJ..loadsin mostyears with thenotal:*!exception of 1995.They were also important energetically except in 1995. On Small Island theywere more common andmadea greater contribution todietaryenergycontent in19904 than 1995 a"hough never reachinglevels of oc:currenceCKenergetic contribution observed on FunkIsland.

Mature sandlanee (Ammodyfes.p.)

MaturesandlanceweteonlyobseNedon Baccalieu Islandwherethey were infrequently fed to nestlings (Appendix 2.2). When present, however. their energetic contributionwasusually intherangeof 19 -44'Yo.They'M8f"eobserved frombotheartyseason (1993) and late season (1992).

~roup sandla~(Ammodytessp.)

Because oftheappearance oftwodistinctiy differentsize dasses of 0- group sandlance with different morphology,the twosize dassesarepresented separately and referred to as either larvalorpost·larval.

LarvalO-groupsandlancewerenotalwaySpresent,beinggeneralty scarce and unimportant energeticallyon8ac:calieu Island(Appendix 2.3). Theywere common onFunkIslandin all

ve-s

except1994,atthough energetic contribution never exceeded 10 % andwhMe1992 and 1993weresimilarlynear this maximum, their contribution in 1994 and 1995wasnegligible. On Small Island they were often present in biU·loads in 1994 but not in 1995.

Post-4arval Q-group sandillnce dominated nestlingdiet ineveryway,i.e.

percentoc:x:urrence,frequency,massandenergy,Ol"lbothFunk and Small Islands in 1995(Appendix2.4). On Small Islandtheywerealsopresentin 1994 but less frequent and important energetically. They were notably absent from BaccaJieuIsland and in every year sampled on FunkI~ndexcept 1995.

Stichaeids

AU Stichaeids present inthenestlingdietslNerelarvalandwerenot identified tospecies(Appendix 2.5). Thetwomost likelyspeciesto whichthey belong areArcticShannyStichaeus punctatusandRadiated ShannyUlvaria subbifurcata (J. Green pers. com.).

Stichaeidswereobserved at all colonies and in all years although not all

~ssampled. Theyweresomewhat commonin theeal1yseason on Baccalieu and generany absent there later on excepttor1992. Energetic contribution on Baccaheu never exceeded 3 %.Stichaeidsweremore common in bil140ads on both Funk and Small Islands where energetic contribution ranged from a high of 7 to < 1 %. On Funk Island theyweremore common and contributed more energetically in 1992 and 1994 than in the othertwoyears. On Snlallisland they were more common and contributed more energetically in 1994 than 1995.

Cottids

CottidsW'ef'8presentinthenestlingdietsateachofthethreec:oblies although not in all 'Ne8J(s sampled, particularly on Baocalieu Island where they were absent or scarce in August-Septembersam~andmost common in lale

July (Appendix2.6). Theyweremore common onBac::caIieuin1994 than 1993 and rather than the usual <1%contributionto dietary energytheCXlf'ltribution in early1994reached11%. CottidsVlI'8realways presentin periodssampled on Funk Island and usuallyonSmall Island. Energetic contribution at both colonies was usually negligible withthe exceptionof1992onFunk Island.

Herring

Herring weregenerally absentfrom nestliog dietswith the notable exception of Small Island in 1994when~roupHerringinweeks 5 and 6 respectively contributed 13 and7% by frequency, 13and 12'Yobymass, occurred in 32 and50%of bill·loads and contributed10%of the estimated dietary energy delivered to nestlings (Appendix2.7).

White Hake

White Hakewererecordedat eachoftheook:lniesbutnotinallyears (AppendiX2.8). Theyweregenerally absent or very scarce and made negligible energetic contributions. White Hake occurred most commonly in bill-foads on Funk Island in1993.They were recordedfrom both the early and late season on Ba<xalieu l548nd.

Agonidae

Agonidaeweteeitt1er-abSent or scarce in sampled nestlingdiet(Appendix 2.9). They occurredmostoommonly (22 %) intheearly seasonof 1994 on Baccalieu Island. OnFunkIslandtheyweremost commonin 1992 (11'Yo)and

wereabsent therein 1995. They'N'8l'enotrecordedfromSmall Island andatno lime did their energetic contribution anywhe«t exceed1%.

li".ris

.p.

Liparissp.werescarce everfNhefewiththeexception of1992onFunk Islandwhentheyoccurredin23% ofbill40ads(Appendix 2.10). Theywere notably absent from FunkIslandin1994and1995and wereneverobservedon Small Island. They appeared more often in the early season than at any other limeon8accaIteu. Nowheredidtheyever contribute morethan 1%ofdietary ene<gy.

Squid

Squid rarelyoccurred onBaccalieu and Small Islands, atthough in the late seasons of1992and1993onBaccalieu their energetic contribution reached 5·9

%(Appendix2.11). On Funktsland theyweremostcommoninbil~1oadsin1992 (23%)contributing18% of dietaryenergyat that time. In1994their energy contribution was also notnegligibleat 11 % whileinthe othef"years sampledon Funk Island they were less important, particular1y in 1995.

Crustaceans

Althoughcrustaceanswere obsetved at eachof the colonies ineachofthe years sampled except for 1995,theywere often absent(Appendix2_12). On Baccalieu crustaceansW'eI'emost common in the earty season. On Funk Island they were less common in 1994 than in the previous 2 years. They were equally

as scarce onSmaII"ndin1994as on FunkIslandand at no time orP'acedid they contribute>1 % ofthedietary energy.

Miscellaneous

The baIanoe of collected nestlingcfietwasmadeupof rarely occurringprey items which induded unidentified juvenile gadids, Nereis sp., larval Pleuronectidae and another unidentified larval fish.

2.3.4 Size, man .nd energy content ofbllMCMlds delivered to puffin nestlings

The enumeratiorl of bill-load contents and measurement of bilMoad mass werenotcollected from every sample (see methods) and thus the sample size for bitl·1oadsizetsoften greaterthanthe samplesizeforbilJ.loadmass. Asummary of the sample sizes, masses, numbers of prey items and estimated energy content ot bill·1oads intercepted from adult puffins provisioning nestlings is shown in Tables 2.6 - 2.8. A graphical summary of the same data using boxplots to display the range and distribution of data values and showingthe 95 % confidenCe intervals aroundthemedian isshown inFigures 2.5·2.7.

2.3.4.1 Number of individualprey ...in bill-loMls collec1lldfromAtlantic Puffin nntlings

The number ofpreyitems inbil.toedsooIIectedrromAtlanticPuffin nestlings was Quite variable (range: 1 - 50; see Table 2.6 and Figure 2.5 for details). BiII·loadsweresmallest on Baccalieu Island and generally much larger on Funk and Small Islands. On Baccalieu during the 2 seasons with both early

and late sampling coverage (1993 and 1994) bill-bad$weredramatically larger at the beginning of the nestlingperiodthan there after. Aswell, bitl-loeds in early 1994oontained more preyitemsthan inearty 1993. On Funk IsJand bill-6oads -were largest and most variablein1992and 1993,lessso in1994and contained the fewest prey withtheleast variation in 1995. While the 95 % confidence intervals around median bill-load size on Funk Island during 1992*1994 O1I'ef1apped witheach other, those for1995 didnot overtap with theothers. On Small Island in 1994biM-1oac:tscontainedmore preyduringweek 6than in 1995 ando,vere morevariab6e inweek 5than in 1995.

During sampling periods when data are available from more than one site, there are some dear diffec'ences, for exampAe, during 'Neeks4and5of 1993 bill*

IoedsonFunk Istandweredramaticalylargerandmorevariable thanthoseon Baccalieu lsiand. During week 6 of 1994 bill-loads on Small ISlandwerealso larger and more variabJe than on Baccalieu Island. Aweekbefore,when comparison is possible between Small and Funk's1ands.bill-k)ads on Funk Island are larger than on Small Island. A yearlatet'",in1995, duringthesame time periodbill-load size was notablyreduced andtheFunk and Small Islandbil~s contain similar numbers of prey.

2.3.4.2 Mne (g)ofbill-loads collectedfromAtlllntie Puffins provisioning nestlings

The magnitude of variation in bill·1oad masswasgenerally greater and mol"e consistent than that of bill-load size {see Table 2.7 and Figure 2.6 for

details). Overa. bilJ.Jo8d mass had a range of 0.5 • 33.5g.OnBac:calteuIsland during 1993 and 1994, there wasareciprocal relationship between bill-load mass and size. Earlyseasonbil40adsweighedlessthan duringthebalanceof the season in1993. Bill·108dmasswaslessyariab5e'inthelatter part of1992than in thesameperiodduring1993and1994.

Funk Island bill-loads were generally somewhat heavier in1995but the95

%confidence limits around the median bin40ad massamongthefour years sampled on Funk Island all overtapped. The smallest median bilf-klad mass on Funk Island was7.6 9in1994while the largest median was11.0 9during1995.

Bill-loads were heavier on Small Island in1995than in 1994.

8et'Neencolonyditrerences in bill-lo8d mass aremuch lesspronounced than was thecasefor bilJ-ioadsiZe. In 1993 during a comparable timeperiod (weelo:;s 4 and 5) Funk Island bil1-bads'N8f8generally lighter andlessvariable than on Baccalieu Island but95%confidence limits around the medians still had some overlap. In 1994 during comparabM! weeks the median bill-load mass was roughly similarbetweencolonies. In1995during a comparable'Neekmedian bill- load mass on Small Island tendedtoexteed that of Funk Island.

2.3.4.3 Estimated energy content(kJ)ofbill-lc.dscollected from At..ntic Puffin nestlings

Estimated energycontent of bin-loads coIleded from Atlantic Puffin nestlings ranged from a lOw of 0.5 kJ to a high of 189 kJ. Estimated energy content of bill-loads followed much the samepatternas that of bill-load mass (see

"

Table 2.8andFigure 2.7fordetails). Thesameintra-~pattem of lower values earty in the nestling season comparedwithlater.observedforbin-load mass on Bacx:alieu Islandin1993, isaboapparent forenergy content. Bill..toads on Baccalieu Istand from comparable weeks betweenyearscontained the least energy in1994,as did those on Funk andSma"lsiands. Bill-4oadson both Funk andSmanIslands had highest energy contents in 1995. On Funk Island, 1995 bill40ad energy contents were also more variablethan in previousyeaf$.

Median values of bill-lOad energy contentrangefrom a low of20.7kJon Funk Island in 1994 to a high of &4.7kJon Small Island in 1995. Bill-loads on Funk Island in 1993 contained less energy than those on Baccalieu Island during the same time period. Bill-bads on Funk Island in 1994 contatnedlessenergy thanthoseon Small Island,whilethoseon Small Islandtendedto containless than those on Baccalieu Island, althoughthedifferencebetweenSmall and Baccalieu Islands(medians: 33.04 vs39.0kJ)wasnotasgreatasthedifference between Small and Funk lsiands (medians:33.7vs20.7kJ).

2.3... The relationship between bill-kMld size. mae• •nd.""rgycontent The relationships between bill-load mass and size. bill·load mass and energy content and bin-loadsize andenergycontent are illustrated for a comparable samplingperiod (week5of nestling season) for Baocalieu Island 1993, Funk Island 1992 - 1995 and Small 1s18nd 1994 - 1995 in Fjgures 2.8 • 2.10.

Thereis~variation In the rMitionshipbetweenbill-load mass (g) and bill-loadsize (number of prey items per bill·load) among colonies and years duringacomparable timeperiod(e.g.weeK5 of thenestlingseasonin1993, Figure2.8). OnBaocalieuin1993 theheaviestbill-toadswere atso thelargest.

However,theextremelysmall variation inbill-loadsizesuggests somethingelse is responsible for increasingbill~loadmasses, i.e. size of individual fish. On Funk Island,the relationshipbetweenbill-badmassandsize isdit'refenttot"each ofthe four years sampled ranging fromnon-existent in1994through l0oseiypositive with alotof variation in1992,to a IOW'siope tightpositive association in1995and a steep slope positive association in1993with variation intermediate between the extremesof1992 and 1995.On Small Island,therelationship between bill-load mass andsizealsodiffersdistinc:ttybet'Neen1~and1995. In1~,thereare twogroups of points, one aloose duster resembling Funk Island in1992and the othera flat string of points resembling Baccalieu in1993. Thesetwopanems representbill-lOads containing multipleand variable prey items in the first instanceand biU40ads CDl"ltaining varyingnumbers of asingledominantprey (adult capelin) in thesecond. Thepattern on Small Island in 1995 is different again and resembles Funk Island in 1995. Both instances depict bill-loads containing variable numbers of asingle dominant large preytype(post-larval 0- groupsandlance).

The retationship betweenbill-lo8c:lmass and estimatedenergy contentis strongly positive and generally tight with minimal variation (Figure 2.9). Funk

"

Island in 1992wasnotabfymore 't8riab'ethanthe otheryearsandcolonies examinedduringthesamesamplingperiod.

The number of prey items in abin..Joadhada varyingeffectonenergy content (Figure 2.10). On Funk Island the relationShip was similarly positive and variable in1992and1993,ambiguous in1994and steeply positive in1995.

Small Island in1995resembledFunkIslandin thesameyear. Baccalieu Island hada distinct pattern duetothelackof variation inbill-4oad sizeandSmallIstind 1994containedtwopatterns includingonesimilar to Baccalieu Island and the other resembting Funk Island in1992and1993.

2.5Discus.ion

Sampling ofbilt-4oadsfromadultsprovisioning nestltngs relied on retrieval of a dropped samplefromthegl'OUnd.Direct examination ofretrievedpreyitems is possible but recovery of100% of the bin-load isnotassured (Rodway and Montevecc:hi1996). Thesmaflertransparentpreyitemsare morelikely tobe over-looked among the ground cover than largerorfleshier prey. Consequently.

numbers of larval fish and crustaceans are underestimated with theSe methods (ROdway and Monteveechi 1996). These data therefore represent minimal estimates ofthe pen::entoccurrence and numerical frequency ofthesmallerprey items and especially the transparent larval forms_ Even so,thenumber of theSe preytypespresent in neslting dietswasfar greater thanexpected. The estimation of contribution by mass of these prey is not strongly biased by the tendency to underestimate numerical abundance as the mass contrtbutionper

unit fish isvery small. Therefiol'e. wtI~ reported pen::ent frequency and occurrence for small prey shookSbetreatedas minimal estimatesthepercent contributionbymass and energyofthesesame preyaremorerobust to the sampling bias identified above.

Considef'8tionof diffet'ences among CQk)nies and yearsW8Sbasedon comparingweeksofthesameordinal rank.i.e.comparisons were among similar calendarperiods. Nestlingage ofAtlanticPuffin nestlingsmay affect mealsize.

i.e. the youngest chicksmaybeunable to ingestthelargerprey consumed when older andmaybe fedsmalterpreybytheprovisioningadult (Bradstreet and Brown 1985). Comparing diets of nestlings among colonies for the same calendar period is potentiaUy confoundedbydif'rerenc:esamongcok)nies in breeding phenok>gy. If thequestionbeingaddf"essedislimited toconsiderationof the differences inforagingconciitions among colonies at a particular time then this is notaproblemunless confoundedbythetendency of adults tofeedthesmalleSt chicks smallerprey.The timing of puffin breedingonFunk and Small Islands was roughlyawoeek later thanonBaccalieu Island during the years of this study (Russell and Montevecchi unpubl. data). The puffin breedingseason onFunk and SmallIslandin theperiodssampledbythisstudy wasfurtheradvanced than thatwhich wouldpresent a concernforthe interpretation of nestling diet comparisonwith Baccalieu Island. Therefore.thedifferencein puffin breeding phenology among the study ook>nies is insufficient to explain the observed differences innestlingdiet and I infer that the abundance in neslling diet on Funk

Island during 1992 - 1994ofsmallpreyitemsisareflection offoragingconditions and notconfounded bynestlingage.

The composition of Atlantic Puffin nestting diets at Bac:calteu, Funk and Small Islands (1992 - 1995) were not homogeneous. Mature capetin dominated nestling diets in allyears onBaccalieo and SmaU Islandin 199>4 butotherwise were~as theprimarydiet constituent by larval 1-group capelin (Funk Island: 1992-19904)or post4arvalG-group sandlance (Funkand Small Islands 1995) Contributions to estimated nestling dietary energy of 10 % or morewere occasionallymadebya numberofpreytypes otherthanmatureca~n. In addition to the principal exceptions mentioned above, the following made lesser but noteworthy contributions to puffin nestling diet in the northwest Atlantic during 1992 - 1995: 1) adultsandlanceon BaccalieuIsland, 2) O-group larval sandlance onFunk and Small Islands, 3) cottidaeonBac:calieu and Funk Islands. 4) Q-group Herring on BaccaIieu and Sman Islands and 5) squidon Funk Island.

Invertebrateswere uncommon prey items but much more frequent than recorded in the northeast Atlantic. In 27 colony years in Great Britain, Harris and Hislop (1978) observed only 1load containing squid andone crustacean. In contrast.

although infrequent, crustac8answererecorded from all locations andyearsin this study and squidmadeup 11 % bymassof the nestling dietonFunk Island in 1992.

Thesmale miscellaneouspreywhichare most common in the early seasons at Baccalieu Island and in each year sampled on Funk Island are