The 116(3):690-701, 1999

BREEDING BIOLOGY OF CRESTED AUKLETS AT BULDIR AND KASATOCHI ISLANDS,

GAIL FRASER?6 IAN L. JONES,2 JEFFREY C. WILLIAMS,3 FIONA M. HUNTER,4 LISA SCHARF,3 AND G. VERNON BYRD 5 •BiopsychologyProgramme, Memorial University of Newfoundland,St. Johnõ,Newfoundland A1B 3X9, Canada; 2Departmentof Biology,Memorial University of Newfoundland,St. Johnõ,Newfoundland A1B 3X9, Canada; 3AlaskaMaritime National Wildlife Refuge, Aleutian Island Unit, P.O.Box 5251, Adak, Alaska 99546, USA; 4Departmentof Zoology,University of Cambridge,Cambridge CB2 3E J, UnitedKingdom; and 5AlaskaMaritime National Wildlife Refuge, 2355 KachemakBay Drive, Homer, Alaska 99603, USA

ABSTRACT.--Wequantified breeding parametersof Crested Auklets (Aethiacristatella) at Buldir and Kasatochiislands in the , Alaska, in 1996 and 1997. Crested Auk- lets incubatedtheir eggsfor about36 daysand chicksweighed about 35 g within the first threedays of hatching(14% of adultmass; Buldir, n = 58). Growthrates averaged about 9.9 g per day duringthe linearphase (Buldir, n = 58; Kasatochi,n = 17),and chicks fledged at an averagemass of 248 g (95% of adult mass;Buldir, n = 63) and a wing lengthof 123 mm (88%of adultwing length; Buldir, n = 37) at 34 daysafter hatching. We found no differences in intraislandand intrayear chick growth for Buldir and Kasatochi. Productivity (the product of hatchingsuccess and fledgingsuccess) averaged more than 60% for the two yearsat Ka- satochiand for eightyears (1990 to 1997)at Buldir.Intercolony comparisons of productivity parametersrevealed differences in hatchingdate, age of chicksat fledging,and hatching and fledgingsuccess. Adult massdiffered significantly between the sexes(267 g for males,253 g for females)and amongyears. At Buldir, we observedno effect of variouslevels of inves- tigatordisturbance on hatchingand fledgingsuccess or on otherbreeding parameters. We foundno negativerelationships between hatching date and fledgingage, hatching date and fledgingmass, or fledgingmass and fledgingage, contrary to the predictionsof Ydenberg's (1989)model of intraspecificvariation in timing of fledgingof alcid chicks. chicks,like thoseof otherdiurnally active species of auklets,grow relatively fast and depart at a youngerage compared with chicksof two nocturnalspecies of auklets.Received 24 April 1998, accepted17 November1998.

CRESTEDAUKLETS ( cristatella)are co- (e.g.B•dard 1969,Hipfner andByrd 1993,Jones lonialseabirds that lay a clutchsize of oneand 1993b), and the Whiskered Auklets is the least havesemiprecocial chicks that are provisioned well known (Byrd and Williams1994). Detailed by both parents(Jones 1993b, Gaston and Jones knowledgeof the biologyof theseauklets is re- 1998). This speciesis part of a remarkable quired for an understandingof their life-his- adaptiveradiation of five small planktivorous tory variation,adaptive radiation, and ecolog- auklets(Cassin's [Ptychoramphus aleuticus], Par- ical relationships. akeet [Aethiapsittacula], Least [A. pusilia],and Alcids show considerable inter- and intra- Whiskered[A. pygmaea]auklets) that are en- specificvariation in the age and massof their demicto the North PacificOcean and Bering chicksat the time of departurefrom nesting andOkhotsk seas. These auklets range in mean colonies.For example, the tiny precocialchicks bodymass from the LeastAuklet at 85 g to the of Ancient Murrelets (Synthliboramphusantiq- ParakeetAuklet at 289 g. Adult and chickdiets uus)depart two days after hatching,whereas and chickdevelopment patterns (Gaston 1985, the semiprecocialchicks of RhinocerosAuklets Gastonand Jones1998) alsovary amongspe- (Cerorhincamonocerata) depart close to adult cies.The biologyof Cassin'sand Leastauklets havebeen relatively well studied(see Manuwal size at 38 to 58 daysold (Gaston1985, Gaston and Thoresen1993, Jones1993c), whereas Par- and Jones1998). Ydenberg(1989) and Yden- akeet and Crested auklets are less well known berget al. (1995)provided the first comprehen- sive modelthat explainedvariation in life-his- tory traits in alcidsand offeredtestable predic- E-mail: [email protected] tions. Two main assumptionsof the model

69O July1999] CrestedAuklet Breeding Biology 691

FiG. 1. Mapof theNorth Pacific, including St. Lawrence Island and our study sites in theAleutians.

were thatjuvenile mortality is lowerat the nest chickgrowth, productivity, and breeding chro- and growthrate is higherat sea.Here, we test nology in Crested Auklets, including eight two of the model'spredictions about intraspe- yearsof data from a singlecolony in the Aleu- cificvariation in life-historytraits using data tians. from CrestedAuklets: (1) fast-growingchicks The objectivesof our studywere to (1) quan- departthe nest younger and heavier (i.e. a neg- tify chickgrowth; (2) examineeight years of ativerelationship exists between fledging age variation in reproductive performance and and fledging mass), and (2) late-hatching breedingchronology from one colonyin the chicksdepart the nest younger and lighter. Al- western Aleutians (1990 to 1997) and make in- though somedata from Cassin'sand Rhinoc- teryearand intercolonycomparisons with an- eros aukletsare consistentwith thesepredic- othersite located in the centralAleutians (1996 tions(Harfenist 1995, Harfenist and Ydenberg to 1997);(3) evaluatethe effectsof investigator 1996,Morbey and Ydenberg 1997), further test- disturbanceon breedingsuccess; and (4) test ing of the model is needed (Gastonand Jones two key predictionsof Ydenberg's(1989) model 1998, Hipfner and Gaston 1999).Crested Auk- to explaintiming of chickdeparture for a semi- let chicksare semiprecocial,depart the nestat precocialspecies of auklet. closeto adult size,fit all of the assumptionsof themodel (Ydenberg et al. 1995),and thus pro- STUDY AREA AND METHODS vide an opportunity to examine the model's predictionsfor the first time within the Study areas.--We studied auklets on Buldir Aethia. (52ø21'N, 175ø56'E) and Kasatochi (52ø11'N, Much of the information on Crested Auklet 175ø30'W) islands in the Aleutian chain of Alaska breedingparameters comes from northerncol- (Fig. 1). Buldir, locatedin the westernpart of the onies on St. Lawrence Island (B6dard 1969, chain,contains one of the largestand mostdiverse Piatt et al. 1990;see Fig. 1) wherethe nest concentrationsof seabirdsin the Aleutians (Sowlset later than their Aleutian counterparts.The al. 1978,Byrd and Day 1986).Our studyarea on Bul- Aleutian Islandscomprise a substantialpor- dir waslocated at "Main Talus,"a colonywith ap- tion of the CrestedAuklet's breeding range. proximately250,000 Crested and Leastauklets (ratio of 2:1 Crested to Least;Knudtson and Byrd 1982, Nevertheless,Knudtson and Byrd's(1982) and Byrd et al. 1983). Kasatochiis locatedin the central Hipfner and Byrd's(1993) reports on laying Aleutians about 480 km east of Buldir. Our study and hatchingdates, productivity,and nest- areathere was in a colonyon a northeast-facingtalus creviceattributes provide the only published slopewith a minimumof 35,000Least and Crested data from the Aleutians.Here, we presentthe auklets (2:1 Least to Crested; Scharf et al. 1996). first comparisonsacross years and coloniesof Chick growth, productivity,and breedingchro- 692 FRASERET AL. [Auk, Vol. 116 nology were recordedon both islands,whereas in- data from Buldir to a logisticmodel (Ricklefs1967). vestigatordisturbance and effectsof crevicelocation The model'sproducts (asymptotic mass [a] and the were evaluatedonly on Buldir.We selectedcrevices growth constant[k]) were used to calculatethe in- that were configuredso that parentsor chicksin stantaneousgrowth rate at thepoint of inflection(i.e. nestscould not easily hide from view. Becauseour k(a)/4; Hussell1972), which is consideredto be the samplesof creviceswere taken from large areasand maximum growth rate (Hussell 1972, Sealy 1973, were of differentcrevice types (except for very deep Gaston 1985, Piatt et al. 1990). crevices),we believe that the creviceswe monitored Adult mass.--Toquantify variationin adult mass were representativeof the entire coloniesat Buldir within and betweenbreeding seasons and islands, and Kasatochi. For all crevices, chicks were consid- and to compareadult mass with chickmass at fledg- ered fledged(i.e. left the nest)if they were 26 days ing, we caughtbirds at a singlestudy plot centrally or olderupon disappearance(i.e. the crevicefailed if located on Main Talus, , and at a similar chicksdisappeared at lessthan 26 daysof age). plot on KasatochiIsland. We determinedthe sex of Incubationand chick growth.--In 1997, we followed these birds using bill shape (Jones1993a) and eight pairs from laying until hatchingto determine weighedthem to the nearest1 g with a springscale. the duration of incubation on Buldir. We selected Measurementsof adults were taken between May previouslyused crevicesthat were unoccupiedin and August from 1990 to 1997 on Buldir and from mid-Mayand checked them daily until anincubating 1996 to 1997 on Kasatochi. was present.To minimize disturbanceearly in Productivityand breeding chronology.--Crested Auk- incubation,we checkedoccupied crevices only once let creviceswere checked approximately every seven a week for the first 29 days. After 29 days, we daysthroughout each breeding season between 1990 checkedcrevices daily to obtain the exact date of and 1996 (1996 to 1997 at Kasatochi).In 1997, we hatching. checkedcrevices every four daysbetween the onset In 1996and 1997,we studiedchick growth on Bul- and terminationof hatchingand fledgingperiods to dir andKasatochi. Because we hadlarger sample siz- obtainmore precise estimates of the timing of these es and wanted more precisedata from Buldir, our events.To minimize bias resulting from nestfailures methodologyvaried slightly betweenthe two colo- early in the breeding season(i.e. overestimating nies. To determinehatching dates on Buldir, nests hatchingsuccess), only crevicesfound prior to 15 were checkedonce a week until mid-June(again, to Junewere used to estimateproductivity. We used the minimizedisturbance early duringincubation), then midpoint between visits to estimatehatching and every two daysuntil hatching.Chick age was esti- fledging dates and the even-numberedJulian date matedto within oneday based on appearance(Jones whenan evennumber of daysoccurred between vis- 1993b).Chicks were then measuredevery three days its. until they departedfrom their crevice;in mostcases, Effect of crevice location and investigatordistur- fledging dateswere known to the nearestday. On bance.--Weevaluated differences in habitatquality Kasatochi,we determinedhatching dates by check- and compared various parametersof productivity, ing nestsat four-dayintervals prior to hatchingand nesting chronology,and chick conditionbetween estimatedchick age to within two days. We mea- siteslocated in differenttypes of habitaton Main Ta- suredchicks only during the linear growth period on lus, Buldir. We subdivided the area into three sec- Kasatochi;therefore, chicks were handled every four tions (1) the low area (27 x 45 m), which was char- days (from ages6 to 24 days), and fledging dates acterizedby large boulders(>1 m3) that provided were estimatedto within two to four days.We mea- deeperand larger crevicesapparently preferred by suredchick mass to the nearest1 g, flattenedwing CrestedAuklets (Knudtsonand Byrd 1982,B•dard (Buldir) and wing chord(Kasatochi) to the nearest1 1969);(2) the middle section(42 x 45 m), which con- ram, and tarsuslength to the nearest0.1 mm. tained a mixture of large and smallboulders; and (3) We used regressionresiduals from mass versus the high section(67 x 35 m), which was composed age and wing versusage to estimatethe linearpor- primarilyof smallboulders with morevegetation. tion of the growthcurve from composite data. We as- We evaluatedthe disturbancecaused by our mon- sumedlinear growth when the residualswere ran- itoring activities by comparing successof nests vis- domly distributedaround zero. The slopesof simple ited at differentfrequencies during the incubation linear modelsfor massand wing length from each and chick-growthperiods. In 1996, we visually chickprovided comparative statistics. The sample checkedcrevices during incubation(1) oncea week units were individual chicks with at least three mea- (i.e. low disturbance),(2) every two days (medium surementscollected during the linear growth period. disturbance),or (3) twice a day for 15 days before We used P < 0.05 as a thresholdfor concludingsta- hatching(high disturbance).During 1997,we had tistical significance. two levelsof disturbanceduring the incubationpe- To compareour growth datawith thosefrom stud- riod (1) weekly checks(low disturbance)and (2) ev- ies of CrestedAuklets on St. LawrenceIsland (Sealy ery two days (medium disturbance).During the 1968,Piatt et al. 1990),we fitted individual growth chick-rearingperiods of both years, we evaluated July1999] CrestedAuklet Breeding Biology 693 three levels of disturbance(1) crevicesthat we vi- cant relationshipbetween mass and age of suallychecked approximately every seven days (low fledging(day 25 only) in either 1996or 1997 disturbance),(2) crevicesin which we captureda (1996, P = 0.60; 1997, P = 0.10). chickevery three days (mediumdisturbance), and On average,Crested Auklet chicksat Kasa- (3) crevicesin whichwe captureda chickevery three days and in which we capturedat least one adult tochigained approximately 9 g per day during oncewithin six days after hatching(high distur- the lineargrowth phase and remainedin their bance; 1996, n - 16 nests; 1997, n = 36 nests). To as- crevicefor 35 days (range26 to 40; Table1). We sessthe effectsof handling chicks,we compared found no significantdifferences between years mass,wing, and tarsus measurements of chickshan- for the linear phaseof growth in mass(P = dled regularlythroughout the rearingperiod with 0.20)or wing length(P = 0.08),nor anysignif- chicksthat we checkedvisually every three days, but icant differencesbetween islands or years for capturedand measuredonly oncebetween ages 28 the linearphase of growthin mass(island, P = and 30 days(i.e. control chicks). 0.30;year, P = 0.20;island x year, P = 0.06) or fledgingage (island,P = 0.10;year, P = 0.90; RESULTS island x year, P = 0.80). We were unable to sta- tisticallycompare linear growth of thewing be- Incubationperiod and chickgrowth.--In 1997, causeof differencesin methodologyfor wing the durationof incubationaveraged 35.9 + SD measurements(Buldir, œ = 4.5 -+ 0.36 mm per of 4.8 days(n = 8, range 29 to 44 days)on Bul- day;Kasatochi, œ = 3.1 _+0.61 mm per day). dir. On average,Crested Auklet chicks on Bul- Adultmass.--On average, male CrestedAuk- dir weighedabout 35 g within three days of hatching(13.4% of adultmass) and gainedap- lets (œ= 267 + 19 g, range 211 to 314 g, n = 347)were 14 g heavierthan females (œ = 253 q- proximately10 g per day during the linear phaseof growth.They remainedin their crev- 1.0 g, range210 to 322 g, n = 345;t = 10.1,df icesfor about35 days(range 26 to 41 days)and = 1, P < 0.0001; Fig. 3). Mass varied signifi- fledged at a massof 248 g (95% of adult value; cantly amongyears on Buldir (1990to 1997;F Tables1 and 2, Figs.2A, B) and a wing length = 6.2,df = 6 and565, P < 0.0001),being high- of 123mm (88%of adult value;Fig. 2C). Chick estin 1993and 1997,which corresponded with massdeclined prior to fledgingby 5% in 1996 two yearsof highbreeding productivity (Table and by 3% in 1997. 4). We foundno evidencethat massdiffered be- At Buldir, we found no significantdifferenc- tween islands in 1996 and 1997 (P = 0.20). es betweenyears in hatchingmass (P = 0.07), Productivityand breeding chronology.--Over an growth rate (P = 0.60), maximum mass (P = eight-yearperiod at Buldir, hatchingsuccess 0.20), fledging mass (P = 0.70), and fledging averaged81%, fledgingsuccess 76%, and pro- age (P = 0.10). Hatchingdate (t = 2.4, df = 1, ductivity67% (Table4). Forthe two studyyears P = 0.005, n = 91) and fledging date (t = 2.3, on Kasatochi(1996 and 1997),Crested Auklets df = 1, P = 0.02, n = 67) differed significantly averaged85% hatchingsuccess, 76% fledging betweenyears (Table 1). success,and 65% for productivity (Table5). Wefound no significantrelationships in 1996 The annualbreeding chronology for Crested or 1997between hatching dates and the follow- Auklets on Buldir was somewhat flexible dur- ing: fledgingmass, fledging age, fledging wing ing the eight yearswe monitoredthe colony length,and linear growthrate for mass(Table (Table4). The median hatchingdate was 26 3). When we controlledfor linear growth in June,but hatchingdates varied by nine days 1997,we foundthat fast growing chicks fledged amongyears; typically, the first eggs hatched in at a heaviermass, but not at an earlierage (i.e. mid-June,and the last eggsin mid-July.The a positiverelationship between linear growth median fledging date was 29 July. Median and fledgingmass exists, but no significantre- fledgingdates differed as muchas sevendays lationshipbetween hatching date and fledging betweenbreeding seasons, but generallythe mass,or fledging age;Table 3). We alsofound firstchicks fledged on about 19 July, and the last a significantnegative relationship (r 2 = 0.32,P chicksfledged on around 12 August.On aver- = 0.02,n = 17) betweenwing lengthand age age,the aukletson Kasatochi appeared to beon of fledgingin 1997(i.e. day 25, which is oneday a slightlylater schedulethan on Buldir (Table beforethe first fledgingage; see Hipfner and 5). The first chickshatched on 26 June,the last Gaston 1999). However, we found no signifi- chickson 15July, and the median hatching date 694 FRASERET AL. [Auk, Vol. 116 July1999] CrestedAuklet Breeding Biology 695

TABLE2. CrestedAuklet chick growth data (logistic model) from Buldir and St. Lawrenceislands.

Asymp- Fledging totic Adult age Fledging mass(g) mass(g)• (days) mass(g) kb k(a)/4c Island Source 254.4 286 34 228 (80)a 0.197 12.5 St. Lawrence Sealy(1968, 1973) 269.0 260-283 ------12.8 St. Lawrence Piatt et al. (1990) -- 267 ------11.1 St. Lawrence Searing (1977) 260.8 262 34 244 (93) 0.184 12.0 Buldir This study (1996) 258.5 262 35 246 (94) 0.172 11.1 Buldir This study (1997) • Massof incubatingadults (seeFig. 2). bMean instantaneousgrowth rate calculatedfrom individual chicksfit to model. ßMaximum instantaneous growth rate, where a - asymptoticmass (Hussell 1972, Sealy 1973). • Fledgingmass at % of adult massin parentheses.

was30 June.The firstchicks fledged on 22 July, the last onesby 12 August,and the median fledgingdate was 4 August. Interislandcomparisons of breedingchronology and productivity.--Incomparing reproductive 3OO A variablesfor fourcolony-year combinations, we 250- foundsignificant differences in hatchingdate (F = 68.9, df = 1 and 229, P < 0.0001;Tables 4 200- and 5) and hatchingand fledgingsuccess in 150- 1997(hatching, X2 = 3.86,P = 0.05;fledging, X 2 = 4.0,P = 0.05).We alsofound a significantin- 100- teractionbetween year and islandfor fledging

50- 1996 age (year,F = 0.05, df = 1 and 171, P = 0.80; island,F = 3.0, df = 1 and 171,P = 0.08;year o 300 x island, F = 16.9, df = 1 and 171, P < 0.0001), and pairwise comparisonsrevealed significant 250 - differencesin fledging age betweenislands

200 - (1996, F = 12.9, df = 1 and 69, P = 0.0006;But- dir, • = 35.2 days;Kasatochi, • = 31.7 days; 150- 1997, F = 3.8, df = 1, 102, P = 0.05; Buldir, œ =

lOO - 32.6days, Kasatochi, œ = 34.0 days).We found no differencesin hatchingand fledging success 50 in 1996or in fledgingdate in eitheryear (Ps > 1997 0.05). o 14o Effectof crevicelocation on productivity.--The

120 locationof a nestingcrevice on Main Talushad some influence on when a chick hatched and on lOO the likelihoodthat it wouldsurvive to fledging 8o age.Chicks in crevicesin thelow or middlear- eas on Main Talus hatchedearlier (F = 6.2, df 6o = 2 and115, P = 0.002;Scheffe's post-hoc, high 40 vs. low,P = 0.005;high vs. middle,P = 0.04; 20 1997 middle vs. low, P = 0.90; low, œ= 24 June;mid- dle,œ = 24 June;high, œ = 27 June)and had a o 4• so greaterchance of fledging(both years com- Chick Age (days) bined, X2 = 16.9, P = 0.0002;high = 53.3%, F•c. 2. Growth of Crested Auklet chicks on Buldir middle = 92.6%,low = 83.7%)than chicksto- Island. Changesin body massin (A) 1996 and (B) catedhigh on the talus.We foundno location 1997;(C) wing lengthin 1997.Values are œ+ SE. effects(all Ps > 0.05)in eitheryear for hatching 696 FRASERET AL. [Auk, Vol. 116

TABLE3. Data from CrestedAuklets with respectto Ydenberg'spredictions for timing of fiedgingin alcid chicks.Dependent variables centered in bold.

Hatchingdate; Linear growth Hatchingdate linear growth of mass• of mass Year n P r 2 n P r 2 n P r 2 Fledging mass 1996 25 0.8 0.003 25 0.8, 0.3 0.05 23 0.3 0.04 1997 35 0.2 0.04 35 0.5, 0.02 0.14 35 0.01 0.18 Fledging age 1996 25 0.8 0.01 25 0.6, 0.4 0.04 26 0.4 0.03 1997 35 0.9 0.001 35 0.9, 0.7 0.01 35 0.6 0.01 Fledging wing length 1996 ...... 1997 35 0.4 0.02 35 0.3, 0.5 0.04 35 0.6 0.01 Linear growth of body mass 1996 24 0.7 0.01 ...... 1997 32 0.1 0.09 ......

Hatchingdate and linear growth as independent variables in a multipleregression. mass,hatching wing length,hatching success, fledgingsuccess varied with disturbancelevel or for the slopeof the linear growth for mass in 1997,although in a directionwe would not andfledging mass, fledging wing length,fledg- havepredicted: nests with higherlevels of dis- ing age, or fledging date. turbanceappeared to have higher fledging suc- Investigatordis turbance.--Crested Auklets tol- cess(X 2 = 7.4, P = 0.03; low = 82%, medium = eratedlarge amounts of disturbanceduring the 55.6%,high = 92.3%). incubationperiod; hatching success did not dif- CrestedAuklet chicks that were handled reg- fer significantlywith low,medium, or highlev- ularly were the samesize betweendays 28 to els of disturbance(1996, P = 0.40; 1997, P = 30 as chicksthat we handledonly once.Chicks 0.20). Creviceswhere we either captured an handled only once (n = 10) were similar in adult (onceafter hatching),and/or handleda mass(P = 0.10),wing length(P = 0.20),tarsus chickregularly did not experiencelower fledg- length(P = 0.70),and fledgingdate (P = 0.40) ing successin 1996 (P = 0.40). In contrast, comparedwith chickshandled every three days(ages 28 to 30 daysin both cases).

laying hatching fledging DISCUSSION 3OO 29 9 41 82 25 51 38 45 12 14 20 16 27 64 24 61 52 46 10 23 .• 290 Incubationand chick growth.--Similar to other 280 studies(Sealy 1984, Platt et al. 1990),we found IE 270 the duration of the incubationperiod to be -o• 260 highlyvariably, although the mean duration on o Buldir was comparableto that on St. Lawrence -a 250 Island.We believethat the wide rangein in- • 240 cubationperiods is related to egg neglect,as • 230 Sealy (1984) reported for Crested Auklets, 22• Least Auklets, and Ancient Murrelets. Crested May June July Aug Auklet eggscan endureperiods of coolingand Date still hatch;an egg that we found unattended FIG. 3. Variationin massof adult male (opencir- severaltimes hatched after 44 days.The ability cle) and female (closed circle) Crested Auklets at Bul- of eggsto endurecooling likely is importantfor dir and KasatochiIslands, Alaska (values are • _+ a speciesthat foragesfar from the colonyon 95% confidencelimit; samplesizes indicated for each ephemeralpatches of food. week). We observedno intracolonydifferences in July1999] CrestedAuklet Breeding Biology 697

o o"• 698 FRASERET AL. [Auk, Vol. 116

TABLE5. Reproductiveperformance and variationin adult massof CrestedAuklets at KasatochiIsland, 1996 and 1997.

Parameter 1996 1997 No. eggs found 43 76 No. of eggshatched 37 64 No. chicksfledged 32 42 Medianhatching date 30 June 1 July Rangehatching date 26 Juneto 17 July 27 Juneto 14 July Median fledgingdate 31 July 8 August Rangefledging date 22 Julyto 8 August 27 Julyto 12 August Hatchingsuccess • 0.86 0.84 Fledging successb 0.86 0.66 Productivityc 0.74 0.55 Male massa 269.1 -+ 17.6 (29) 256.2 _+18.6 (19) Female massd 259.1 _+19.5 (30) 248.8 ñ 15.1 (31) • Proportionof eggsthat hatched. bProportion of chicksthat survivedto fledgingage (26 daysbefore disappearing or 22 daysat the time of the last visit). ßHatching success x fiedgingsuccess. d• + SD,with n in parentheses. growthparameters (Table 1). AlthoughBuldir and chickgrowth. The model also predicteda CrestedAuklet chicks fledged at a similarage positive relationshipbetween fledging mass and grew at ratescomparable to St. Lawrence and chickgrowth. Studieson two semipreco- Island chicks,fledging mass differed between cial species,Cassin's Auklet (Morbeyand Yden- the two sites(Sealy 1968, 1973;see Table 2). We berg1997) and RhinocerosAuklet (Ydenberg et would predictthat fledglingson St. Lawrence al. 1995, Harfenist 1995), supportedthe mod- are heavier than those on Buldir because adult el's predictions.However, we found no signif- mass,on average,was greaterin the northern icantnegative relationships between these var- part of their range (Sealy1968, Jones1993b). iables for Crested Auklets at Buldir Island in ei- However,St. LawrenceIsland chicks weighed ther of our two study years. When we con- lessat fledging(228 g) and lostmore mass prior trolled for linear growth, we observed a to fledging(11% decline) than chicks on Buldir positive relationshipbetween linear growth (Sealy1968, 1973). The Ydenberg model (Yden- and fledgingmass only in 1997,which implied berg et al. 1995)for intraspecificvariation pre- that faster-growingchicks fledged heavier, but dictsthat in colonieswith fastergrowth, nest- not earlier (Table3). The lack of a seasonalde- lingswill fledgeheavier and younger; however, clinein fledgingmass with ageat Buldircould Buldir chicksfledged heavier but not at a youn- haveresulted from the lack of a strongseason- gerage. There are three possible reasons for the ality in food abundancearound this near-oce- observeddifferences in fledgingmass between anic island. chicks from St. Lawrence Island and those from Hipfner and Gaston(1999) suggestedthat the Aleutians: (1) methods differed between wing lengthis a betterpredictor of the timing the two studies (i.e. chicks that were handled of chick departure than body mass and pre- every day fledgedlighter than thosehandled dicted a negativerelationship between wing lessfrequently), (2) environmentaldifferences lengthat minimum fledgingage minus one day between the two areas resulted in the Aleutian and fledgingage. We observed this relationship chicksbeing fed moreoften or on higher-qual- for Crested Auklet chicks on Buldir. Crested ity food, or (3) differencesresulted from inter- Aukletshave a highersurvival rate if they are year variation related to differencesin prey ableto fly to searather than walk, becausefliers availability. are more adept at avoiding predators and at Ydenberg'smodel (Ydenberg 1989, Ydenberg gettingpast the surf(G. Fraser pers. obs., Jones et al. 1995),concerning the transitionfrom nest 1993b).Therefore, wing lengthmight be more siteto seaby alcidchicks, predicted three neg- crucialthan body massin determiningthe tim- ative relationships:(1) betweenhatching date ing of departureof CrestedAuklet chicks. and fledgingmass, (2) betweenhatching date Among the five planktivorousauklets (Cas- and fledgingage, and (3) betweenfledging age sin's,Parakeet, Least, Whiskered, and Crested) July1999] CrestedAuklet Breeding Biology 699

60 ductivitycorresponded with laterhatching and fledging dates (Table 3). Although hatching '•50- dates did not vary much acrossyears, lower productivitycoincided with a later fledging 40 dates on Kasatochi (Table 4). Knudtson and ?c. 3.5 tCaAu CrAu PaAu Byrd (1982) reported productivity from Main 30 5.5 Talus on Buldir from 1976 as 51% (proportion of eggsresulting in fledgedchicks), which is ß 20 31% lower than the averageproductivity dur- 10 ing the 1990s.Lower productivity in 1976also 1'00 1'50 2'00 2'50 300 correspondedwith a very late peak hatching Adultmass (g) date (July8). During our study,intercolony and interyear differencesin productivity were dra- FiG. 4. Relationshipbetween adult mass,fledg- matic. On Buldir, productivity increased by ing age (bars indicate range of values recorded in 27% between 1996 and 1997, whereas on Ka- published studies),and maximum growth rate (be- low acronyms,in g per day) of auklet chicksin the satochi,productivity dropped by 26%.The tim- Aethiini (see Gaston and Jones1998 for references). ing of breedingand the productivityof an auk- LeAu = , WhAu = , let colonywere almost certainly related to local CaAu = Cassin's Auklet, CrAu = Crested Auklet, prey availability. Data on prey types, prior to PaAu = . and during chickrearing, between islands and yearswould allow us to furtherexamine these relationships.Another factor that could influ- thatcoexist in theBering Sea and adjacent parts of the North Pacific,the major dichotomyin encetiming and successof breedingare local weather conditions such as sea state, sea tem- chickprovisioning occurs between species with nocturnal versus diurnal colony attendance perature,wind speed,and rainfall during the winter and summer months. (Fig. 4). The nocturnalCassin's and Whiskered auklets,which are constrainedto feedingtheir Other productivitydata for CrestedAuklets are scarce. From St. Lawrence Island, Platt et al. chicksno morethan once or twiceper day dur- ing the hoursof darkness,have relatively slow- (1990) reported productivity for 1987 at 48%. growingchicks that fledgeat 40 daysor older, Althoughthis valuefell within the rangeob- whereasthe diurnal species(particularly Least served at Buldir (1990), it is 19% lower than the Auklets) have relatively fast-growingchicks average productivity at Buldir during the that fledgebetween 29 and36 daysafter hatch- 1990s.One major differencebetween our study sites and the colonies on St. Lawrence Island is ing. A nocturnal lifestyle for Cassin'sand Whiskeredauklets, while presumably reducing a lack of mammalian predators.We would pre- adult mortality from predation,prolongs the dict that productivity would be lower at colo- nestlingperiod of their chicks(Sealy 1973). Di- nies that have mammalian predators.Sealy urnal activity at coloniesis the ancestralchar- (1968) and Piatt et al. (1990) documented that acterstate for planktivorousalcids (Gaston and predationon aukletchicks by voles(Clethrion- Jones1998). The evolutionof nocturnalcolony omysrutilus and Microtusoeconomus ) and arctic activity by thesetwo speciesmay havebeen fa- foxes(Alopex lagopus) was an importantfactor vored becauseit reduced competitionfor nest that depressedproductivity on St.Lawrence Is- sites, permitted colonization of areas with land. large numbersof avian predators,and/or en- Effectof crevicelocation and investigator distur- hanceddiurnal foraging opportunities. bance.--Knudtsonand Byrd (1982)showed that Productivityand breeding chronology.--Crested auklets selected nest sites based on crevice vol- Auklet productivityand breedingchronology ume and substratetype. The differencesin appear to be somewhatflexible. Productivity hatching dates and fledging successwe ob- on Buldir steadily increasedfor the first four served were somewhat correlated with rock years (1990 to 1993) and then fluctuatedin al- size on Main Talus. Crested Auklets that se- ternatingyears (1994 to 1997).A relationship lectedcrevices in largerboulder areas had ear- betweenproductivity and breedingchronolo- lier hatchingdates and higherfledging success gy may exist;in two of threeyears, lower pro- than ones that nested in areas of smaller-sized 700 FRASERETAL. [Auk,Vol. 116 boulders. These differences could also be relat- graphicCommittee for Researchand Exploration, ed to othermicrosite features, such as prox- andthe Behavior Society provided funding. imity to the seaand exposureto predators.For Y. Morbey,M. Hipfner, B. Montevecchi,A. Storey, instance,lower areasmight be preferredbe- and two anonymousreviewers provided many use- causethey are closerto the sea,or becausear- ful commentson the manuscript. eas high on the talus slopemay experience higherrates of predationby Glaucous-winged LITERATURE CITED (Larusglaucescens). B•2DARO,J. 1969. The nestingof the Crested,Least, Our finding that investigatordisturbance and Parakeet auklets on St. Lawrence Island, hadno significant influence on fledging success Alaska. Condor 71:386-398. of Crested Auklets was similar to that for Par- BYRD, G. V., AND R. H. DAY. 1986. The Avifauna of akeetAuklets (Hipfner and Byrd 1993),but it Buldir Island, Aleutian Islands, Alaska. Arctic differedfrom Piatt et al.'s (1990)study at St. 39:109-118. Lawrence,where hatching and fledging success BYRO, G. V., R. H. DAY, AND E. P. KNUOTSON.1983. of LeastAuklets decreased in highlydisturbed Patternsof colonyattendance and censusingof plots. Again, mammalianpredators may ex- auklets at Buldir Island, Alaska. Condor 85:274- plainthe difference in ourresults; higher levels 280. of human disturbance on St. Lawrence Island BYRO, G. V., AND J. C. WILLIAMS. 1994. Whiskered mayhave caused auklet chicks to be morevul- Auklet (Aethiapygmaea). In The birds of North nerableto mammalianpredation. America, no. 76 (A. Poole and E Gill, Eds.). Academyof NaturalSciences, Philadelphia, and Regularhandling of chicksoften can retard AmericanOrnithologists' Union, Washington, chickgrowth and causechicks to fledgepre- D.C. maturely(e.g. Harris and Wanless1984, Lyngs GISTON,A. J.1985. Development of the youngin the 1994). However, we found no differencesin tar- Atlantic Alcidae.Pages 319-354 in The Atlantic suslength, wing length, or bodymass between Alcidae (T. Birkheadand D. Nettleship,Eds.). regularlyhandled chicks and chicksmeasured Academic Press, London. onlyonce. This suggests that measuring chicks GASTON,A. J., AND I. L. JONES.1998. The . Ox- onceevery three days is an appropriateproto- ford UniversityPress, Oxford. col, althougha longerinterval between mea- HARFENIST,A. 1995.Effects of growth-ratevariation surementsduring the linear growth phase on fledging of RhinocerosAuklets (Cerorhinca monocerata).Auk 112:60-66. would suffice (Harfenist 1995). HARFENIST, g., AND g.C. YDENBERG. 1996. Parental Conclusions.--Thepaucity of data on breed- provisioningand predationrisk in Rhinoceros ing biologyof WhiskeredAuklets presents the Auklets(Cerorhinca monocerata). Effects on nest- biggest gap in knowledgewithin the Aethia ling growthand fledging. Behavioral Ecology 6: group; oncethis gap is bridged,a more in- 82-86. depthcomparison of aukletlife-history strate- HARRIS, M.P., AND S. WANLESS.1984. The effects of gies will be possible.Furthermore, an under- disturbanceon survival, age and weight of standingof how productivity,breeding chro- youngGuillemots aalge. 7:42-46. nology, and growth parameterscorrespond HIPFNER,J. M., AND G. V. BYRD.1993. Breeding bi- with interyearvariation in preytypes, and ide- ologyof the ParakeetAuklet compared to other crevice-nestingspecies at Buldir Island, Alaska. ally,how the availability of preytypes changes Colonial Waterbirds 16:128-138. throughoutthe breeding season, is essentialfor HIPFNER,J. M., ANDA. J. GASTON.1999. Timing of a comprehensiveexamination of auklet ecolo- nestdeparture in theThick-billed Murre andRa- gy, includinghow largechanges in the marine zorbill: Testsof Ydenberg'smodel. Ecology 80: ecosystemmay influencebreeding parameters. 587-596. HUSSELL,D. J.T. 1972.Factors affecting clutch size in ACKNOWLEDGMENTS arctic passerines.Ecological Monographs 42: 317-362. We thank CaptainKevin Bell and the crew of the JONES,I. L. 1993a. Sexualdifferences in bill shape R/V Tiglaxfor their superblogistical support. L. and external measurements of Crested Auklets Cowen,J. Fischer, L. Meehan,M. Ortwerth,N. Rojek, (Aethiacristatella). Wilson Bulletin 105:525-529. P.Ryan, G. Thompson,and H. Walshcontributed in- JONES,I. L. 1993b.Crested Auklet (Aethiacristatella). valuable field assistance. The Natural Sciences and In The birdsof ,no. 70 (A. Poole EngineeringResearch Council, the National Geo- and E Gill, Eds.).Academy of NaturalSciences, July1999] CrestedAuklet Breeding Biology 701

Philadelphia,and American Ornithologists' and Wildlife ServiceReport, AMNWR 96/11, Union,Washington, D.C. Adak, Alaska. JONES,I. L. 1993c.Least Auklet (Aethia pusilia). In The SEALY,S. G. 1968. A comparativestudy of breeding birds of North America,no. 69 (A. Pooleand E ecology and timing in plankton-feedingalcids Gill, Eds.).Academy of Natural Sciences,Phila- (Cyclorrhynchusand Aethiaspp.) on St. Lawrence delphia,and AmericanOrnithologists' Union, Island,Alaska. M.S. thesis,University of British Washington,D.C. Columbia, Vancouver. KNUDTSON,E. P., ANDG. V. BYRD.1982. Breeding bi- SEALY,S. G. 1973. Adaptive significanceof post- ology of Crested,Least and Whiskeredauklets hatchingdevelopmental patterns and growth on Buldir Island, Alaska. Condor 84:197-202. rates in the Alcidae. Ornis Scandinavica 4:113- LYNGS,P. 1994.The effects of disturbanceon growth 121. rate and survivalof youngRazorbills Alca torda. SEALY,S. G. 1984.Interruptions extend incubation by Seabird 16:46-49. Ancient Murrelets, Crested Auklets and Least MANUWAL, D. g. AND g. C. THORESEN.1993. Cassin's Auklets. Murrelet 65:53-56. Auklet (Ptychoramphusaleuticus). In The birds of SEARING,C. E 1977.Some aspects of the ecologyof North America, no. 70 (A. Poole and E Gill, cliff-nesting at Kongkok Bay, St. Eds.). Academy of Natural Sciences,Philadel- LawrenceIsland, Alaska, during 1976. Pages phia, and American Ornithologists'Union, 263-412 in Environmental assessment of the Washington,D.C. Alaskan Continental Shelf. Annual Report. MORBEy,Y. M., ANDR. C. YDENBERG.1997. Intra-spe- BLM, NOAA, OSEAP, Boulder, Colorado. cificvariability in nestlinggrowth and fledging SOWLS,A. L., S. A. HATCH, AND C. J. LENSINK.1978. behaviorof Cassin'sAuklets Ptychoramphus aleu- Catalog of Alaskan seabird colonies.United ticusat TriangleIsland, British Columbia. Con- StatesFish and Wildlife Service,FWS/OBS-78/ dor 99:361-371. 78. PLATT,J. E, B. D. ROBERTS,W. W. LIDSTER,J. L. WELLS, YDENBERG,g. C. 1989.Growth-mortality trade-offs AND S. g. HATCH. 1990. Effects of human dis- and the evolutionof juvenilelife historiesin the turbanceon breeding Leastand Crestedauklets Alcidae.Ecology 70:1494-1506. at St. Lawrence Island, Alaska. Auk 107:342-350. YDENBERG,R. C., C. W. CLARK, AND A. HARFENIST. RICKLEES,R. E. 1967.A graphicalmethod of fitting 1995.Intraspecific fledging mass variation in the equationsto growthcurves. Ecology 48:978-983. Alcidae,with specialreference to the seasonal SCHARF, L., J. C. WILLIAMS, AND G. L. THOMPSON. fledging massdecline. American Naturalist 145: 412-433. 1996.Biological monitoring in the centralAleu- tian Islands, Alaska in 1996. United States Associate Editor: E. Greene