The Auk 109(3):576-584, 1992

FACTORS AFFECTING SURVIVAL OF ADULT LEAST AUKLETS (AETHIA PUSILLA) AT ST. PAUL ISLAND, ALASKA

IAN L. JONES• Departmentof Biology,Queen's University, Kingston, Ontario K7L 3N6, Canada

ABSTRACr.--Survivalof LeastAuklets (Aethiapusilia) was investigatedat St. Paul Island, Pribilof Islands,Alaska, during three breedingseasons (1987-1989). Based on resightingsof color-marked birds, an overall annual adult survival rate of about 0.75 was observed. Because of their fidelity to nesting and display sites,as well as their insensitivity to capture early in the breedingseason, this estimatewas unlikely to be substantiallybiased by emigrationfrom the study plot or disturbancecaused by handling. An index of adult body condition was highly variable during the 1988breeding season,but was weakly correlatedwith subsequent survival. However, the body-conditionindex was significantlycorrelated to likelihood of breedingin the sameand following years.Survival was not relatedto adults'variable plumage, although lightnessof plumagecorrelated with age and dominanceand functionsfor status signalling in this species.Survival was significantlylowered in individualswith damageto their webbed feet. Subadults(two-year-olds) showed low site fidelity comparedto adults. Age of first breeding was at three years. During the breeding season,known causesof mortality included predation by foxes,gulls, and humans.The low annual survival of Least Auklets relative to other alcids, together with large year-to-yearvariation in reproductive performance,suggests there is greater potential for populationfluctuations in this species. Survival estimationmay be a partial solutionto the difficulty of monitoring alcid populations directly. Received22 April 1992,accepted 12 February1992.

ANNUALsurvival rate is a critical population b), but none has attempted to estimatesurvival parameterfor theoreticalinvestigations of avi- parameters.Furthermore, there is no accepted an life histories, as well as conservation of wild- method of estimating the size of Least Auklet life populations.Survival rate estimateswill be breeding populationsor monitoring population required for a full understandingof life-history trends (Platt et al. 1990a, Jones 1992). strategiesof various seabird species.Further- Here I present information on survival of more, survival ratesmay be usedto assessvul- LeastAuklets obtainedas part of a detailedstudy nerability of seabirdpopulations to human and of behavioralecology of the speciesin the Prib- other sourcesof mortality, as well as for long- ilof Islands, Alaska. My objectiveswere: (1) to term monitoring of population status.Howev- estimate the annual survival rate of Least Auk- er, there are few estimates of survival rates for lets at St. Paul Island, Alaska; (2) to identify Pacific alcids (Croxall and Gaston 1988). factors that influence adult survival; and (3) to The Least Auklet (Aethiapusilla) is the small- examinethe feasibilityof long-term monitoring est member of the family Alcidae and one of of auklet survival rates for conservation and the most abundant North Pacific seabirds (SowIs management purposes. et al. 1978). The speciesis monogamouswith a clutch size of one, and the chick is fed by both METHODS parentsat the nestingcrevice until nearly adult size. Least Auklets have both a high rate of Field work was conductedat a colony of more than chickdevelopment and fledging mass (Roby and 10,000 Least Auklets near Tolstoi Point, St. Paul Is- land, Pribilof Islands,Alaska (57ø08'N, 170ø17'W)dur- Brink 1986a). There have been a number of ing May through August of 1987, 1988, and 1989. studiesof LeastAuklet breeding biology (B•- Auklets nested along a strip of habitat about 10 m dard 1969,Knudtson and Byrd 1982,Byrd et al. wide, consistingof sparselyvegetated boulders and 1983, Roby and Brink 1986a, Platt et al. 1990a, talusjust abovesea level alongseveral kilometers of shoreline. All banding and most observationswere made at one densely occupied 10 m x 15 m study • Presentaddress: Department of Zoology, Univer- plot on this talus. sityof Cambridge,Cambridge CB2 3EJ, United King- With help from field assistants,I capturedand color dom. marked 263 Least Auklets (234 adults, 29 subadults)

576 July1992] Survivalof Least Auklets 577 in 1987, 369 (306 adults, 63 subadults) in 1988, and An index of body condition was obtained for many 145 (all adults) in 1989 using mist nets and noose birds using measuresof body masscorrected for vari- carpets(totals include recaptures).To minimize dis- ation in body size.Mass measurements were obtained turbance, birds were banded, measured and released during prelaying and incubationperiods when birds' as quickly as possibleafter capture;banding was re- masswas relatively constant(breeding LeastAuklets strictedto every fourth day during the prelaying pe- lose massafter hatching; Jonesunpubl. data). Mass riod. Subadults(two yearsold) were distinguishedby was measured at capture and by repeated weighings their worn forehead and flight feathersand spotted of markedbirds usingfour electronicbalances (Ohaus throats(B•dard and Sealy 1984).Each auklet was giv- D 1000LA, accurateto _+1 g) placedon the study plot. en a numbered U.S. Fish and Wildlife Service stain- Variation in both massand tarsus length within in- less-steelband and a unique combination of three dividuals was less than variation between individuals Darvik plasticcolor bands. Upon capture,auklets were becauserepeatability of mass(r• = 0.50 _+SE of 0.05, weighedto the nearest1 g usinga Pesolaspring scale, F•09,413= 5.61, P < 0.0001) and tarsus(r• = 0.73 _+0.05, and tarsuslength was measuredto the nearest0.1 mm F74,75= 6.27, P < 0.0001) were significant and greater using calipers. than zero (seeZar 1984:323-325).The repeatabilityof Damage to foot webbing, consistingof holes or tarsuslength, basedon measuresof 75 individuals in tears(always fully healedand appearingto be a result both 1988 and 1989,suggests that this measurecould of pastinjuries or disease),was recordedfor eachbird provide a reliable indication of body size. I regressed captured in 1987 and 1988. The condition of each mean masson tarsuslength and used the residuals birds' webbing was scoredas: (0) websbetween toes (the difference between actual mean massof each bird completely intact; (1) small holes or a minor (<1 cm) and predictedmean massof a bird of its tarsuslength) tear in webbing of one foot; or (2) major tearsor large as the condition index (for further details, see An- holes (>1 cm) in webbing of both feet. Later, I ex- dersson1989, Jonesand Montgomerie 1992). amined the relation between web damage and sub- Survival was evaluatedby resighting color-marked sequentsurvival. Ventral plumage coloration varies birds at the colony in yearsfollowing banding. Daily greatly in LeastAuklets, so each marked auklet was watchesduring prelaying, laying, incubation,chick classifiedinto one of four plumage categories(white, rearing and chick departure periods (mid-May to ear- lightly flecked,intermediate, or heavily flecked;see ly August) were conductedto identify marked birds Jones 1990). I checked for a relationship between that bred, or were in regular attendanceon the study plumage and survival. plot in each year of the study (plot residents).Birds Since sexualdimorphism is slight in LeastAuklets that did not appear in one season,but had been in (B•dardand Sealy1984, Jones and Montgomerie1992), regular attendanceor breeding on the study plot in capturedbirds could not be sexedwith certaintyfrom the previous year, were classifiedas having disap- external measurements.However, I identified a large peared. These individuals must either have died or sample of males by watching for vocal advertising moved from the study plot to another part of the displays (restricted to males; Jones 1992). Female colony. To see if disappearancesresulted from birds membersof mated pairs were identified by associa- moving off the study plot, I looked for banded birds tion with males. Mated pairs were identified by re- during daily walks along most of the length of the peated associationand courtship behavior prior to Tolstoi Point colony throughoutthe study. To make laying (Jonesand Montgomerie 1991). To monitor further checks for marked birds, I searched for color- attendance of marked birds and observe their behav- markedLeast Auklets throughoutcolonies at Tolstoi ior, 4-h watches covering the time of peak auklet Point, Village Cove (1 km distant) and Zapadni Point activity (1100-1500 Alaska StandardTime) were con- (4 km distant)at leastthree timeseach breeding sea- ducted daily from mid-May to early August. Peak son.Additional checksof the colony 100 m on either auklet activity occurredlater in the day at St. Paul side of the study plot were made from the blind. comparedto either St. GeorgeIsland (Roby and Brink Further evaluation of breeding-site and display-ter- 1986a)or St. Lawrence Island (Piatt et al. 1990a).Birds ritory fidelity was made by comparing the exact lo- observeddelivering food to chicks on at least two cationsof nesting crevicesand display areasof color- occasions were classified as breeders; those that were marked birds among years from plot photographs. never seen delivering food were classifiedas non- Wear of both plastic and stainless-steelbands was breeders. Use of the term breeder implies successat observed, but this led to few band losses. Of the total least to hatching of the chick. Reproductiveperfor- of 480 individuals color marked, 14 lost one of their manceof marked aukletswas quantified by monitor- three color bands during the study period, but all ing chick provisioningduring daily 7-h watchesat were still identifiable, and 9 of thesewere recaptured the study plot throughout the chick-provisioningpe- and the missingband replaced.Band wear of birds riod, timed to coincide with mostfood deliveries (Jones originally marked in 1987 probably led to significant and Montgomerie1992). To minimize disturbance,all lossesof colorbands by 1990(Art Sowls,pers. comm.), observations were made from a blind set near the but resightingsfrom that year were not included in study plot. (or available for) this analysisof survival. 578 IAIqL. Jo•qEs [Auk, Vol. 109

TABLE 1. Survival of Least Auklets at Tolstoi Point, St. Paul Island. a

Breeding season 1987 1988 1989

Bandedresident population 208 233 209 Breedingadults 172 125 155 Nonbreeding adults 36 I08 54 Adults surviving to following season 152 (73%) 177 (75%) -- Breeding adults surviving to following season 128 (74%) 104 (83%) -- Nonbreeding adults surviving to following season 22 (61%) 70 (64%) -- Surviving breedersthat bred again in following year 83 84 -- Annual survival rate given in parentheses.

I used unpaired two-tailed t-tests to assessdiffer- Site fidelity of both breeders and nonbreeders ences in body condition between auklets that sur- was high. For example, among 83 birds that vived and disappeared.Log-likelihood ratio tests(see bred in 1987 and returned to breed again in Zar 1984:52)were usedto comparesurvival frequency 1988, none moved to a nesting crevice more data between the sexes, between breeders and non- than a meter away. Most appeared to use the breeders,and between capturedand uncapturedauk- same crevice entrance from year to year. No lets in 1988 and 1989. To further assessinvestigator disturbanceeffects of capture and handling, repro- resident birds moved to a nesting site outside ductive performance measuresof captured and un- the boundariesof the study plot between years, capturedauklets were comparedwith Mann-Whitney although the plot was small and densely oc- U-tests.Throughout this paper,two-year ranges (e.g. cupied by marked breeding birds; this was true 1987-1988) refer to overwinter survival, while ref- even when erosion by winter-storm waves al- erencesto a singleyear (e.g. 1988)refer to a particular tered the nesting habitat. Similarly, nonbreed- breeding season. ing adult residentswere faithful to a small area occupied year after year. Each year a number RESULTS of adults, presumablyprospectors, were not re- sighted on the plot after capture and color Estimationof survivalrate.--Minimum annual marking: 12 (5.1%of birds captured)in 1987;73 survival of adult Least Auklets was close to 75% (23.8% of adults captured) in 1988; 6 (4.1% of in each year, the first estimatefor this species birds captured) in 1989. This proportion dif- (Table 1). At St. Paul, fewer nonbreeding auk- fered significantlyamong years (X 2 = 39.6, df lets survivedthan breedingbirds in both years, = 2, P < 0.0001),possibly an artifact of the extra but thesedifferences were not significant(1987- captureeffort in 1988. Intensive effortsin 1988 1988, G = 0.45, df = 1, P = 0.6; 1988-1989, G = using mist nets were likely to have captured a 1.52, df = 1, P = 0.2; Table 1). Similarly, there higher proportion of adult prospectors.Non- was no evidence of a relationship between sur- resident birds were excluded from further anal- vival rate and sex. Over 1987-1988, survival of ysisbecause either mortality or emigrationcould a sample of known-sex males was 86% (56/65) explain their disappearance;therefore, my sur- and female survival was 90% (43/48; G = 0.30, vival estimates include only resident adult df = 1, P = 0.8). Over 1988-1989, male survival breedersand nonbreeders.Subadults (two-year- was 86% (43/50) and female survival 95% (38/ old nonbreeders)showed little site fidelity, and 40; G = 2.14, df = 1, P = 0.2). mostwere not seenagain on the studyplot after The proportion of adultsclassified as breeders banding. For example,4 of 29 subadultsbanded differed significantly among years. The 1988 in 1987 becameplot residentsand only 4 of 63 breedingseason was poor for reproduction,with subadultsbanded in 1988 were later seen reg- a smaller proportion of adults attempting to ularly. During the three yearsof this study,five breed; 54% (125/233) of marked adults bred in birds banded as subadults were later observed 1988, compared to 83% (172/208) in 1987 and on parts of the colony as far as !km from the 74% (155/209) in 1989 (overall X 2 = 8.77, df = study plot. Three of the total of four marked 2, P = 0.01). subadultsresident on the plot in 1988 returned Careful checks at Tolstoi Point revealed that in 1989, and two of these three-year-oldsbred, of the adults resident on the plot in one year suggestingthat first breeding in LeastAuklets none were found elsewherein subsequentyears. occursat three years of age. The third three- July1992] Survivalof Least Auklets 579

TABLE2. Numberand percentof adult LeastAuklets (1987-1989) of differentplumage categories that survived to subsequentseason.

Plumagecategory ß Years Fate White Lightly flecked Intermediate Heavily flecked 1987-1988 Survived 8 (67%) 30 (91%) 68 (71%) 26 (65%) Disappeared 4 (33%) 3 (9%) 28 (29%) 14 (35%) 1988-1989 Survived 16 (80%) 33 (81%) 79 (78%) 26 (65%) Disappeared 4 (20%) 8 (19%) 22 (22%) 14 (35%) SeeJones (1990) for further explanationof how plumage categorieswere scored. year-old that returned remained as a nonbreed- and handling birds had no detectableeffect on er. In 1987,several two-year-olds were observed timing of breeding, chick provisioning rate or at the study plot with throat pouchesfull of reproductive success (for further details on food, but none obtained mates or bred. Nearly measuresof reproductive effort and success,see all subadultsarrived at the colony for the first Jonesand Montgomerie 1992). For example, the time well after the peak of laying (Jones1992). mean hatch date did not differ between cap- Becauseof their low site fidelity, it was not tured and uncaptured auklets in 1988 (Mann- possible to estimate subadult survival quanti- Whitney U, Z = 1.3, n = 129, P = 0.2) or 1989 tatively. However, anecdotal evidence suggests (Z = 0.4, n = 141, P = 0.7). In addition, capture that immature and subadult auklets may have had no detectableeffect on chick provisioning a survival rate as high or higher than adults. rate in 1988 (Mann-Whitney U, Z = 0.4, n = The number of breeders on the plot in 1987 was 128, P = 0.7) or 1989 (Z = 0.9, n = 141, P = 0.4), about 170 pairs (Jones 1992), which produced nor on reproductive successlater the same sea- an estimated130 fledglings. Two years later in son as measured by chick feeding interval 1989,up to 125two-year-olds attended the study (Mann-Whitney U: 1988, Z = 1.9, n = 129, P = plot simultaneously (50% of auklets present; 0.1; 1989, Z = 1.4, n = 144, P = 0.3). Jones 1992). This proportion of subadults at- Factorsaffecting survivaL--In 1988, the mean tending seemedto hold for all parts of the col- body-conditionindex of LeastAuklet survivors ony. The presence of these remarkably large was somewhat greater than that of birds that numbers of two-year-old subadultson the col- failed to return the following year,but this dif- ony can only be explained by a high survival ference was not significant (survivors' body rate between fledging and two years of age. condition = 1.86 + SE of 0.41 g; disappeared To evaluate possibledisturbance effects,I in- birds' condition = 0.63 + 0.70 g; t = 1.48, df = vestigatedthe relationshipbetween capture and 159, P = 0.07; Fig. 1A). However, among birds handling of birds and their likelihood of breed- that did return, mean body condition was sig- ing later the sameseason and returning to the nificantly greater in 1988 among auklets that study plot the following season. In 1988, 50 bred in the following year (breeder body con- birds that were color marked in the previous dition = 2.30 + 0.51 g; nonbreederbody con- year bred on the study plot, but were not cap- dition = 0.83 + 0.48 g; t = 2.10, df = 159, P = tured (i.e. not disturbed), and another 173 0.02; Fig. lB). breederswere captured in the prelaying period The variable ventral plumage color of Least (disturbed). Capture and handling had no de- Auklets is correlated with reproductive perfor- tectable effect on survival or on the likelihood manceand dominancestatus, and lightens with of returning the following season:among un- age (Jones1990). However, there was little ev- disturbed birds 43 of 50 (86%) returned the next idence that plumage variation correlatedwith year; among disturbed (i.e. captured) birds 138 survival (1987-1988, X 2 = 7.13, df = 3, P = 0.07; of 173 (79.8%) returned (G = 1.53, df = 1, P = 1988-1989, X 2 = 0.87, df = 3, P = 0.83; Table 2). 0.2). Auklets captured in 1988 were not signif- There may be a weak trend towards darker- icantly less likely to breed in 1989 than the plumaged(i.e. heavily flecked plumage) birds undisturbed sample:among disturbedbirds 92 having lower survival, but this was not statis- of 138 (66.7%) bred; among undisturbed birds tically significant. 33 of 43 (76.7%) bred (G = 1.63, df = 1, P = In both years,fewer LeastAuklets with dam- 0.3). Similarly, disturbancecaused by capturing aged webs survived than did birds with un- 580 IANL. JONES [Auk,Vol. 109

N = 37 B N = 79 N = 82 A N• 124 I I 20 2O o

o 15 o lO' 10

5

O' 0 1-jT!I......

-5

-10 SurvivorDisappeared -10 NonbreederBreeder Status in 1989 Status in 1989 Fig. 1. Boxplots showing:(A) body conditionduring 1988breeding season of birds that survivedto 1989 breeding seasonand of those that disappeared;(B) body condition during 1988breeding seasonof surviving birds that subsequentlybred during 1989breeding seasoncompared to thosethat returned as nonbreeders. Horizontal lines indicate 5th percentlie, 25th percentile, median, 75th percentlie and 95th percentties;open circles represent outliers. damagedwebs (1987-1988, X 2 = 6.35, df = 1, P son was not correlated with web damage (un- = 0.01; 1988-1989, X 2 = 9.51, df = 1, P < 0.01; damagedcondition = -0.05 g, n = 249;slightly Table 3). This effectwas remarkablyconsistent damagedcondition = 0.21g, n = 14;extensively amongyears. Birds with severelydamaged webs damagedcondition = -0.24 g, n = 6; ANOVA, had lower survival than birds with minor dam- F = 0.022,226,P = 0.98), but the sample size of age, but the sample of birds with severe web web-damagedauklets of known condition was damage was small and the difference among small;thus, this testhas very low power. thesegroups was not significant(Table 3). Web Arctic foxes (Alopex lagopus)were abundant damage occurredat a low rate (4 of 65 [6.2%] at St. Paul Island at the time of this study and adults examinedin 1987had additional damage were a predator of adults, eggs,and chicksat in 1988), but there were few clues as to how the study plot. Foxesmay be the single largest auklets incur web damage.Some auklet chicks sourceof auklet mortality at St. Paul Island and had similar damageto their feet, so injuriesmay nearbySt. George Island (Roby and Brink 1986b). occurin nestingcrevices. Web damagemay also Predation by Glaucous-winged Gulls (Larus result from bumblefoot, a condition that can glaucescens)occurred at St. Paul, but this species result from the septicconditions in the nesting rarely breeds in the Pribilofs and few individ- crevice (D. D. Roby, pers. comm.). The mean ualsspecialized in capturingauklets. Least Auk- body-conditionindex during the breedingsea- let remains were found in stomachs of halibut

TABLE3. Number and percentof LeastAuklets (1987-1989) with differentdegrees of web damagethat survived to subsequentyear.

Condition of webbing between toes Years Fate Undamaged Slight damage Extensivedamage 1987-1988 Survived 110 (80%) 11 (58%) 2 (40%) Disappeared 27 (20%) 8 (42%) 3 (60%) 1988-1989 Survived 115 (80%) 7 (58%) 1 (25%) Disappeared 28 (20%) 5 (42%) 3 (75%) July1992] SurvivalofLeast Auklets 581

(Hippoglossusstenolepis) at a fish-processingplant nificant impact on breeding successand prob- at nearby St. George Island (Vernon Byrd, pers. ably other behaviorof aukletsand other cavity. comm.). Hunting of Least Auklets by Aleuts nesting seabirds(Roby and Brink 1986a, Platt occursat several breeding sites close to roads et al. 1990b, Gaston et al. 1988). However, this at St. Paul Island, but no auklets were taken at was not a factor in the present study because my study plot, so this sourceof mortality prob- nestingcrevices were left entirely undisturbed. ably did not affect my survival estimates.Dur- Ancient Murrelets (Synthliboramphusantiquus) are ing the 1988 breeding season,many birds re- very sensitiveto disturbance,probably even to turned to the colony with their plumage fouled captureand handling away from their burrow by oily waste dischargedfrom fish-processing (Gaston 1990). This did not appear to be the vesselsanchored off St. Paul Island. This may case with Least Auklets. Band wear occurred but have causedsome mortality, but no disappear- resulted in few band losses and, thus, could not ance of a marked auklet was attributable to this have biasedsurvival estimatesduring the short fouling. duration of this study. However, alternative color-band material will have to be used for

DISCUSSION long-termstudies of Alaskanauklets because of the highly abrasivevolcanic talus at many nest- My survival-rate estimatesfor Least Auklets ing colonies. were derived from resightingsof color-marked If my estimate of survival rate is represen- birds, so they must be evaluatedin light of var- tative of Least Auklet populationsgenerally, ious biases that are known to influence esti- the speciesmay have lower adult survival than mates of survival rates (Hudson 1985). Marked all other alcids except Ancient Murrelets. Gas- aukletswere presumedto havedied if they were ton (1990) calculated that annual survival of resident on the study plot in one year and did breeding adult Ancient Murrelets averaged not return the following year. This method about 0.77. Ancient Murrelet body size (mean would underestimatenatural survival rate if any massabout 200 g) is more than double that of birds emigratedfrom the plot, either naturally Least Auklet, so their apparently similar sur- or due to investigator disturbance (Hudson vival rate is noteworthy. Mean annual survival 1985).This is a confoundingfactor in all meth- for breeding LeastAuklets at St. Paul was sim- odsof survivorshipestimation (Clobert and Le- ilar at 0.79. Inclusion of all adult Least Auklets breton 1991). Becauseof the large population in the analysisyielded a mean survival rate of of auklets at St. Paul, I cannot eliminate the 0.75. One year of Least Auklet survival data possibility that a small proportion of banded from Buldir Island in the Aleutian Islands birds that disappearedwere actually alive and (52ø21'N, 175ø56'E)indicated a similar survival attending some remote part of the colony, or rate of 77% (unpubl. data). Additional estimates remained at sea through the breeding season. of LeastAuklet survival from other partsof their However, becausecolor-marked breeding and range suchas St. LawrenceIsland, where there nonbreedingadults were highly faithful to their are largercolonies with differentpredators, will nesting and display sites,and becauseno plot be required before any generalizations about residents catergorized as having disappeared their demography can be made. Furthermore, were found elsewhereduring extensivechecks it will be useful to investigate the degree of of the colony,I believe emigrationwas unlikely interyear fluctuationin adult survival in a lon- to have significantly biased the survival esti- ger-term study. With a population of about mate. Roby and Brink (1986a)recorded similar- 25,000 (pers. observ.), the Least Auklet colony ly high nest-sitefidelity of LeastAuklets at St. at St. Paul Island is small relative to that at GeorgeIsland. The low subadultsite fidelity in nearby St. George,and to other Alaskan colo- Least Auklets follows a trend among other al- nies. Only further work on survival at other cids (i.e. where site fidelity is extremely high colonieswill reveal whether the St. Paul pop- among adults, but low among young birds; ulation of 1987-1989is typical. Hudson 1985). Survival data have been collected for two oth- Disturbance associated with capture and er auklet species.The Cassin'sAuklet (Ptycho- banding of the study population appeared to ramphusaleuticus) had adult survival ratesof 0.83 have a negligible effecton LeastAuklet behav- in California (Speich and Manuwal 1974) and ior and on rates of return and survival. Direct 0.86 in British Columbia (Gaston,unpubl. data). disturbanceof nesting crevicesdoes have a sig- The survival rate of Crested Auklets (Aethia cris- 582 I^NL. JONES [Auk,Vol. 109 tatella) was 0.86 at Buldir Island, Alaska (one may weakly correlatewith subsequentsurvival year of monitoring; unpubl. data). Large alcids becauseconsiderable seasonal changes in mass suchas murres (Uria spp.),Razorbill (Alca torda) and, consequently,body conditionoccur by the and puffins (Fraterculaspp.), like most other sea- end of the breedingseason, and thesechanges birds, have survival rates of 0.87 to 0.96 (Hudson vary accordingto breedingstatus (unpubl. data). 1985, Croxall and Gaston 1988, Hatchwell and I alsofound little indication that plumage-color Birkhead 1991). Given their single-eggclutch, differences,which relate to age and differences the low survival rate of Least Auklets relative in social status(Jones 1990), were related to sur- to most other alcids must be compensatedfor vival. This finding would be surprisingif plum- by relatively high reproductivesuccess, an ear- age statussignals were usedin competitionfor lier age of first breeding, and/or high recruit- food. However, statussignalling by plumageis ment. Breeding-successestimates for LeastAuk- presumablyused only on the colony,probably lets have varied between50 and 72%(Roby and in competition for nest sites and mates (Jones Brink 1986a, Platt et al. 1990b), similar to other 1990). alcids. However, Least Auklets, with some in- Reduced survival of Least Auklets with small dividualsbreeding at three yearsof age, breed injuries to their feet indicates the degree to early compared to most other alcids (Hudson which defectsmay determine individuals' fit- 1985). Furthermore, there is evidence that re- ness. This effect could have been due to reduced cruitment rate may be unusually high as well, foraging successof individuals with damaged becausesurvival from fledging to two yearsof feet.Alcids use their feet asrudders when pur- age seemsto be high. Overall, these data sug- suing prey underwater, so web defects could gest that with low adult survival and a clutch reduce underwater maneuverability and ulti- of but one egg, Least Auklet populations may mately foraging success.If this is true, the lack be vulnerable to additional adult mortality or of a relationship between web and body con- breeding failure. dition in May and Junesuggests that when food The low survival rate of breedingLeast Auk- is abundant the effect of web damage is not lets may result partly from their small body size important (some auklets with damagedwebs and diurnal colonyattendance, which maymake were able to breed successfully).Mortality due them vulnerableto predators.Foxes were a ma- to web damagemay be more likely to occurin jor sourceof auklet mortality at St. Paul. This winter, when conditions are harsher. Web de- contrastswith some Aleutian Island breeding fectsmay have a particularlysignificant effect sites, where foxes are absent, and St. Lawrence on fitnessin LeastAuklets because of their high Island, where few foxes are present (Roby and buoyancyand poor diving ability comparedto Brink 1986a). Overall, human predation takes other alcids (Haney 1991). An alternative ex- place at such a low level in the Pribilofs that planation for the effect of web damageis that its impact on the overall auklet population is it indicatespast disease, and is presentin weak- probably negligible. Marine pollution repre- enedindividuals that are morelikely to die. The sentsa significantthreat to auklet populations, penalty correlated with web defects also raises particularly in the Pribilofs where large num- the question of whether bands could have a bers of fish-processingvessels operate close to similar effect, by increasingunderwater drag. shore.Causes of mortality outsidethe breeding This couldnot be assessedin this studybecause season remain unknown. it was impossible to monitor survival of un- My index of LeastAuklet body condition dur- banded individuals. ing one breeding seasondid not significantly Becauseof many statisticaland logisticaldif- correlate with survival to the following year, ficultiesinherent in attemptsto directly mon- although it was significantly correlated with itor auk populations using counts of birds on likelihood of breedingthe following year. This the surface of colonies (Gaston et al. 1988, Platt finding is somewhatsurprising, because in gen- et al. 1990a, Jones 1992), measurement of adult eral a relationshipbetween body condition and survival may be a useful additional method of survival would be expected. However, body monitoring auklet populations. This may be condition in autumn and winter, when plank- particularly true for LeastAuklets, becausesur- ton populationsare low and weather conditions facecounts in this speciesprovide only a weak are harsh,is likely to be mostcrucial to survival. indication of population changes(Jones 1992). Measuresof body condition in May and June Monitoring of adult survival in color-marked July 1992] Survivalof LeastAuklets 583

populations of auklets with diurnal colony at- Bf•DARD,J. 1969. The nesting of Crested,Least and tendance (e.g. Least Auklets, Crested Auklets, Parakeet auklets on St. Lawrence Island, Alaska. and ParakeetAuklets [Cyclorhynchuspsittacula]) Can. J. Zool. 47:1025-1050. could alert us to trouble before our relatively Bf•DARD,J., AND S. G. SEALY. 1984. Moults and feather weak censustechniques detect a population de- generations in the Least, Crested and Parakeet cline. An initial effort to establish a marked auklets. J. Zool. (Lond.) 202:461-488. BYRD, G. V., R. H. DAY, AND E. P. KNUDTSON. 1983. population of several hundred birds would be Patternsof colony attendanceand certsusingof required in the first year of such a study. Af- auklets at Buldir Island, Alaska. Condor 85:274- terwards, researcherswould need only return 280. for a few weeks each summer to resight sur- CLOSERT,J., AND J.-D. LEBRETON.1991. Estimation of vivors and band new birds to maintain the demographic parameters in bird populations. markedpopulation. This would permit year-to- Pages 75-104 in Bird population studies: Rele- year monitoring of adult survival, a crucialpop- vance to conservation and management (C. M. ulation parameter.Information on reproductive Perrins, J.-D. Lebreton, and G. J. M. Hirons, Eds.). performance, such as proportion of adults Oxford Univ. Press, Oxford. CROXALL,J.P., AND A. J. GASTON. 1988. Patterns of breeding and even fledging success,could be reproduction in high-latitude Northern- and obtained by observation of marked birds car- Southern-Hemisphereseabirds. Pages 1176-1194 rying food to chicks (Jonesand Montgomerie in Acta XIX Congressus Internationalis Orni- 1992). thologici (H. Ouellet, Ed.). Ottawa, Ontario, 1986. Recruitment, an additional component of a Univ. Ottawa Press, Ottawa. population-dynamicsmodel, would remain un- GASTON,A. J. 1990. Population parameters of the measured.This could be a problem if differing Ancient Murrelet. Condor 92:998-1011. populationtrends among coloniesrelate to re- GASTON,A. J., I. L. JONES,AND D. G. NOSLE. 1988. cruitment, suchas may be the casefor Common Monitoring Ancient Murrelet breeding popula- tions. Colonial Waterbirds I 1:58-66. Murres (Uria aalge; Harris and Wanless 1988, Hatchwell and Birkhead 1991). Least Auklet HANm,, J. C. 1991. Influence of pycocline topogra- phy and water-column structure on marine dis- chicksare difficult to capture in their usually tributions of alcids (Aves: Alcidae) in Anadyr inaccessiblecrevices, so recruitmentis unlikely Strait, Northern Bering Sea, Alaska. Mar. Biol. to be easy to measurein this species.Nonethe- 110:419-435. less, annual estimates of survival and repro- HARRIS,M.P., ANDS. WANLESS.1988. The breeding ductive performance may provide valuable biology of Guillemots Uria aalgeon the Isle of measuresfor population monitoring. May over a six year period. Ibis 130:172-192. HATCHWELL,B. J., AND T. R. BIRKHEAD.1991. Popu- ACKNOWLEDGMENTS lation dynamicsof CommonGuillemots Uria aalge on Skomer Island, Wales. Ornis Scand. 22:55-59. This study is dedicated to the causefor environ- HUDSON,P.J. 1985. Population parameters for the mental preservation of the Pribilof Islands. Thanks Atlantic Alcidae. Pages 233-254 in The Atlantic to Anne Harfenist, Simon Gawn and Robert A. Sund- Alcidae (D. N. Nettleship and T. R. Birkhead, stromfor help with field work, and to Vernon Byrd, Eds.). Academic Press, London. Art SowIsand SteveZimmerman for logisticaid pro- JONES,I.L. 1990. Plumage variability functions for vided by the U.S. Fish and Wildlife Service and Na- statussignalling in Least Auklets. Anim. Behav. tional Marine Fisheries Service. I appreciated the 39:967-975. commentsof Tony Gaston,Ben Hatchwell, John Platt, JONES,I.L. 1992. ColonyattendanceofLeastAuklets Dan Roby, J. C. Haney and Jim Rodgerson an earlier Aethiapusilla at St. Paul Island, Alaska: Implica- version of this manuscript.Financial assistancewas tions for populationmonitoring. Condor 94:93- provided by a grant from the National Geographic 100. SocietyCommittee for Researchand Exploration,the JONES,I. L., ANDR. D. MONTGOMERIE.1991. Mating Natural Sciencesand Engineering ResearchCouncil, patternsof LeastAuklets Aethiapusilla in relation and the Frank M. Chapman Fund of the American to plumage and ornamentation. Behav. Ecol. 2: Museum of Natural History. 249-257. JONES,I. L., AND R. D. MONTGOMERIE. 1992. Least

Ln'ER•RE CITED Auklet ornaments:Do they function as quality indicators? Behav. Ecol. Sociobiol. 30:43-52. ANDERSSON,S. 1989. Sexual selection and cues for KNUDTSON,E. P., AND G. V. BYRD. 1982. Breeding female choice in leks of Jackson'sWidowbird Eu- biology of Crested, Least and Whiskered auklets plectesjacksoni. Behav. Ecol. Sociobiol. 25:403-410. at Buldir Island, Alaska. Condor 84:197-202. 584 IANL. JONES [Auk,Vol. 109

PLier,J., B. D. ROBERTS,AND S. HATCH. 1990a. Colony ROBY, D. D., AND K. L. BRINK. 1986b. Decline of attendanceand populationmonitoring of Least breeding Least Auklets on St. George Island, and Crested auklets on St. Lawrence Island, Alas- Alaska. J. Field OrnithoL 57:57-59. ka. Condor 92:97-106. SOWLS,A. L., S. A. HATCH, AND C. J. LENSINK. 1978. PL•vr, J., B. D. ROBr.RVS, W. Lmstrm, J. WELts, AND S. Catalog of Alaskan seabird colonies. U.S. Dep. HATCH. 1990b. Effects of human disturbance on Interior, Fish Wildl. Serv. FWS/OBS-78/78. breeding Least and Crested auklets at St. Law- SPEICH,S., AND D. A. MANUwAL. 1974. Gular pouch rence Island, Alaska. Auk 107:342-350. development and population structure in Cas- ROB¾,D. D., AND K. L. BRINK. 1986a. Breedingbi- sin's Auklet. Auk 91:291-306. ology of Least Auklets on the Pribilof Islands, ZAR,J.H. 1984. Biostatisticalanalysis, 2nd ed. Pren- Alaska. Condor 88:336-346. tice-Hall, EnglewoodCliffs, New Jersey.