An Unusual Sequence of Flight-Feather Molt in Common Murres and Its Evolutionary Implications

Home , Common murre, Synthliboramphus, Uria

The Auk 115(3):653-669, 1998

AN UNUSUAL SEQUENCE OF FLIGHT-FEATHER MOLT IN COMMON MURRES AND ITS EVOLUTIONARY IMPLICATIONS

CHRISTOPHERW. THOMPSON,TM MONIQUE L. WILSON,2 EDWARDE MELVIN,2 AND D. JOHN PIERCE3 •WashingtonDepartment of Fishand Wildlife,16018 Mill CreekBoulevard, Mill Creek,Washington 98012, USA, and BurkeMuseum, Box 353010, Universityof Washington,Seattle, Washington 98195, USA; 2WashingtonSea Grant, 3716 BrooklynAvenue NE, Seattle,Washington 98105, USA;and 3Washington Department of Fishand Wildlife,600 CapitolWay North, Olympia,Washington 98501, USA

ABSTRACT.--CommonMurres (Uria aalge)exhibit an unusualmolt sequence.Primary molt beginsat a focusbetween P4 and P7 and progressesin two rapid concurrentwaves, proximally to Pl, and distallyto P10.The only other birds known to havea similarmolt sequence arecaracaras and falcons (Falconidae), parrots (Psittaciformes), and Pied Kingfishers (Ceryle rudis).Great Auks (Pinguinus impennis) also appear to havefollowed the sameprimary-molt sequence.Phylogenies for the Alcidaeindicate that Great Auks, CommonMurres, Thick- billed Murres (Uria lornvia),Razorbills (Alca torda), and Dovekies(Alle alle) share a common ancestorand are moreclosely related to oneanother than to otheralcids. This suggests that the unusualsequence of primarymolt in CommonMurres is a shared-derivedcharacter that occursin the otherfour speciesin their cladebut hasbeen overlooked. Adult maleCommon Murreshave significantly shorter secondaries and longerprimaries, on average,than do adult females,resulting in a slightlyhigher aspect ratio in males.Secondary molt begins whenprimary molt is morethan one-third completed. Secondaries are replaced rapidly but sequentially(not synchronouslyor simultaneously);molt appearsto proceedfrom two foci, proximallyfrom S1to S4,and both proximally and distally from S8, but moredata are need- ed to clarify this point. Rectrixmolt beginswhen primary molt is two-thirdscompleted. Rectrixloss and replacement occur rapidly, possibly synchronously, and in no apparentor- der.Adults molt abouttwo weekslater than nonbreeding subadults. We found no differences in the timing of molt betweenthe sexesin adultsor subadults.Duration of flight-feather molt can vary from lessthan 25 daysto more than 80 days,possibly reflecting interyear variationin prey abundance.Received 12 February1997, accepted 9 December 1997.

RELATIVE TO MOST OTHER TOPICS in ornithol- Murres. The phenologyof breedingin Com- ogy,virtually all aspectsof the moltingprocess mon Murres differs by two monthsor more are poorly documentedand understood(Pyle among geographicareas. Thus, flight-feather et al. 1987, Jenniand Winkler 1994). This is es- molt scoresin postbreedingadults might be pecially true in seabirdsbecause most species, useful for identifying breedinglocation (e.g. includingall speciesthat become flightless dur- Oregonvs. Washington);this is importantfor ing molt, undergomolt at seaduring the non- determiningthe demographiceffect of mortal- breeding season (Palmer 1962, Glutz von ity causedby gill nets or other anthropogenic Blotzheimand Bauer1982, Cramp 1977, 1983, activitieson differentbreeding populations of Warham 1996).Among alcids,molt has been Common Murres. better studied in Common Murres (Uria aalge) than in manyother species (Verwey 1922, 1924, MATERIALS AND METHODS Salomonsen 1944, Stresemann and Stresemann 1966,Birkhead and Taylor1977), but many as- In 1993,commercial gill net fisheriesoccurred for summer sockeyesalmon (Onchorhynchusnerka) in pectsof their molt remainpoorly known (con- northernPuget Sound and for fall chumsalmon (O. tra Harris and Wanless1990). keta)in Hood Canaland centralPuget Sound, Wash- As an extension of studies on seabird entan- ington(Pierce et al. 1994).To evaluatevarious mod- glementin gill netsin PugetSound (Thompson ifiedgill net designs,test fisheries were done in 1993 et al. 1998), we studied molt in Common by WashingtonSea Grant, and in 1993and 1996by WashingtonSea Grant and WashingtonDepartment of Fish and Wildlife. Entangledseabirds were col- E-mail:thompcwt@dfw. wa.gov lected:(1) from commercialfisheries in 1993from 1

653 654 THOMPSONETAL. [Auk,Vol. 115

August to 5 Septemberand 5 Octoberto 23 Novem- eachfeather. For eachbird, the lengthsof the pri- ber; and (2) from test fisheriesin 1993from 6 Julyto maries, secondaries, and rectrices were summed to 1 August, and 2 Septemberto 5 October (Boessow give cumulativelengths of all primaries,secondar- 1996),and in 1996from 28 Julyto 29 August (Melvin ies, and rectrices,respectively. For moltingbirds, all et al. 1997). growing featherswere measuredin the samefash- Data collectedincluded bill length, body mass, ion.Old feathersreceived a scoreof 0; emptyfollicles plumagecoloration, status of flight-feathermolt, size were assigneda value of 1.0 mm to indicatethat the and physicaldescription of bursaof Fabricius,gonad old feather was lost. As above, all values were size,and presenceor absenceof an incubationpatch; summedto give a cumulativemeasure of growth. in females,the largestfollicle in the ovarywas mea- Primary-moltscore for eachmolting bird was cal- sured, and oviduct condition was noted. Bursa con- culatedby dividing the cumulativeprimary growth dition was categorizedinto one of three categories: of a moltingbird by the mean total primary length (1) large and fleshy;(2) thin-walled;or (3) membra- of a nonmoltingbird of samesex and multiplying by nous,or no bursa.Sex was determinedby gonadal 100, i.e. primary-molt scorewas calculatedas a per- inspection.Birds were classifiedas hatching-year centageof total regrowth,the minimum molt score (born during the current calendaryear), subadult being0.01 and the maximum being 99.99. Secondary- (born at leastone calendaryear previously,but not and rectrix-moltscores were calculatedin analogous yet reproductive),or adult (physiologicallycapable fashion. of reproduction)using a combinationof dataon bur- Statisticsand estimatingduration of flight-feather saof Fabricius(Broughton 1994), body-plumage col- molt.--Statisticswere conductedusing SYSTAT5.0 oration,presence or absenceof flight-feathermolt, for windows (Wilkinson 1992). Remex and rectrix culmenlength (Baker 1993, C. Thompsonunpubl. measurements were correlated within individual data), and reproductivecondition, including pres- birds.Therefore, the relative shapes of theprimaries, ence or absenceof an incubationpatch. Molt and secondaries, and rectrices of males and females were plumageterminology follow Humphreyand Parkes analyzedusing two-way repeated-measures ANOVA (1959).Birds with fleshybursas, no indicationof re- whereprimary, secondary, or rectrixlength was one productiveorgan maturity (e.g. tiny, relatively un- factor and sex was the other. differentiatedgonads and associatedreproductive Mostmethods for estimatingthe durationof flight- structuressuch as oviductsand vasdeferens), in ju- feathermolt regressdate on molt score(Pimm 1976) venalor firstbasic plumage (or firstprebasic molt be- or vice versa (Ginn and Melville 1983). Thesemeth- tween theseplumages), lacking flight-feather molt, ods assumethat flight-feathermolt scoreincreases andwith a culmenlength less than 43 mm werecon- linearly from onsetto completionof molt. This as- sideredto be hatching-yearbirds. Birds in alternate sumptionis based on the observationthat flight plumage, definitive basic plumage,or with flight- feathersgrow at a fairly constantrate regardlessof feathermolt and that had a fleshlyor thin-walled theirtotal length when fully grown(but see Ashmole bursa,culmen length greater than 43 mm, and an im- 1962).How accuratelythese methods estimate molt mature reproductive system were consideredto be durationdepends on variousfactors but is especially subadults.Birds in alternateplumage, definitive ba- influencedby the methodsused to scoremolt. The sicplumage, or with flight-feathermolt and a mem- mostcommonly used method for scoringmolt (New- branousbursa or no bursa (or rarely a thin-walled ton 1966, Ginn and Melville 1983) assignsequal bursa),culmen length greater than 43 mm, anda ma- weightto all flight feathers.However, because short ture reproductivesystem (e.g. large ovarian follicles feathersbegin and finish growth in lesstime than and hypertrophiedoviduct) were consideredto be longerfeathers, and aregiven as much weight as lon- adults. In 1993, 12 adult males, 6 adult females, 6 gerfeathers, this scoring system causes flight-feather subadult males, and no subadult females were ex- molt scoreto increasenonlinearly over the courseof amined. In 1996, 42 adult males, 29 adult females, 14 molt in speciessuch as alcidsand shorebirds(Ash- subadult males, and 19 subadult females were ex- mole 1962, Summers1980) that have remigesthat amined. differ considerablyin length.The methodswe used Molt-score calculation.--Primaries and secondaries to scoreflight-feather molt (describedabove) yield are numberedfrom the innermost(P1) to outermost molt scoresthat increaselinearly over time and, (P10; not including vestigial Pll), and outermost thereforeare ideal for accuratelyestimating duration (S1) to innermost(usually S16), respectively.Rectri- of flight-feathermolt using linear regressionmeth- cesare numberedfrom the centralpair (R1) to the ods. Our data alsoare especiallywell suitedfor anal- outermostpair (R6). For nonmolting adult males (n ysis by a method developedby Underhill and Zuc- = 20) and females(n = 15), the lengthof eachof their chini (1988) and Underhill et al. (1990) that over- 10 functionalprimaries, 16 secondaries,and 12 rec- comesmany of the potentialbiases and limitations trices(Nitzsch 1840, C. Thompsonunpubl. data)was of Pimm's (1976), Newton's(1966), and other linear measured with a ruler to the nearest 0.5 mm from the regressionmethods, and yields more accurateesti- point of insertion in the skin to the terminal end of matesof molt duration as a result. Unfortunately, July1998] Molt in CommonMurres 655

Underhill and his colleaguesprovided no mecha- nism (e.g. softwareprogram) to enableother re- searchersto employtheir methodwithout their as- •,130- ---e--Males (a) sistance. As a result, we used Pimm's method to es- timatethe timing,rate, and duration of flight-feather molt. ,120 FollowingPimm (1976),collection date was used as the dependentvariable and was regressedon flight-feather(primary, secondary, or rectrix)molt 110 score.Using collection date as the dependentrather /x• = Females thanthe independent variable yields estimates of the •100 timing, duration, and rate of molt for individual birds,which were among the goals of thispaper. Re- n=15 versingthe axesyields estimates of the timing, du- 80 ration, and rate of molt for populationsof birds (Pimm 1976).In regressionanalyses, comparisons of 12345678910 time of onsetand rateof flight-feathermolt between agesand sexeswere madeusing two-samplet-tests Primary Number with separatevariances of slope and y-intercept data; otherwise,Mann-Whitney-U tests were used 80 when molt-scoredata were not significantlycorrelated with collection date. E•,75- RESULTS • 70

Wing length,tail length,and body mass.--Adult cø'65 males have significantly longer primaries, (b) shortersecondaries, and longerrectrices than 60 --e-- Males do adult females(primaries: F = 10.87,df = 1 • 55 and 9, P < 0.001; secondaries: F = 7.95, df = 1 and 9, P = 0.005; rectrices: F = 11.41, df = 1 õ 50 n=20 and 9, P = 0.001;Figs. 1A-C). Femaleprimar- ies, secondaries,and rectricesaveraged 1,095.7 45 _+ SE of 10.0 mm, 1,031.6 _+ 8.7 mm, and 289.6 2 4 6 8 10 12 14 16 -+ 3.0 mm, respectively.Male primaries, secondaries,and rectricesaveraged 1,113.0 +_ 3.5 SecondaryNumber mm, 1,012 +- 7.1 mm, and 301.4 +- 4.4 mm, respectively. Both subadult and adult males are signifi- --e-- Males (C) cantly heavierthan females(subadults: males •' 54 - n=20 = 1,050 _+15 g, females= 992 -+ 14 g, t = 2.76, & 52- df = 40, P = 0.009; adults: males = 1,096 -+ 7 g, females= 990 +--7 g, t = 3.02, df = 198, P < 0.001). -- Females '= 46- n=15 F•G. 1. (A) Primary length of adult Common n- 44- Murres in relation to primary number. Primaries numberedfrom proximal(P1) to distal(P10). (B) Sec- 42 ondarylength of adult CommonMurres in relation to secondarynumber. Secondariesnumbered from 1 2 3 4 5 6 distal (S1) to proximal(S16). (C) Rectrixlength of Rectrix Number adult Common Murres in relation to rectrix number. Rectricesnumbered from central pair (R1) to outermost pair (R6). Valuesare œ_+ SE. 656 THOMPSONET AL. [Auk, Vol. 115

A k"3

2 B

c FIG.2. (A) Moltingwing of adultCommon Murre missing P4 and P5; P3 and P6 indicated by numbers onthe figure. (B) Molting wing of adult Common Murre missing P3 to P9; P2 and P10 indicated by numbers onthe figure. (C) Molting wing of adult Common Murre missing P2 to P10; P1 indicated on the figure.

Sequenceof flight-featherloss and replace- 1), P1 droppingafter P2. The firstprimaries to ment.--Subadult and adult Common Murres be lost alsoare the first to be replaced;this is replacetheir flightfeathers once a yearin late reflectedby the greaterlength of the middle summer or fall. Common Murres exhibit an primariesrelative to moreproximal and distal unusual sequenceof primary replacement primariesduring the first one to twoweeks of amongbirds. Contrary to all publishedliter- primary molt (Figs.3A-B, 4). ature,primaries are not droppedsynchro- Secondarymolt is delayeduntil afterall old nously.Instead, molt is initiatedat a singlefo- primariesare lost. On average,secondary molt cusin the middle of the primariesbetween P4 beginsand finishes when primaries are 27 and and P7 and progressesin two concurrent,rap- 99%grown, respectively (Fig. 5A). Of thesam- id wavesproximally to P1 and distallyto P10 ple of moltingbirds examined,we found35 (Fig.2A-C, Table1). P1 andP2 arenot lost un- birds that had lost somebut not all primaries til after P10has been dropped (Fig. 2C, Table (Table1, Fig. 2A-C), but only two specimens July 1998] Molt in CommonMurres 657

TABLE1. Locationof missingprimaries in CommonMurres undergoing definitive prebasic molt.

No. Locationof missingprimaries missing primaries na P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 I I 0 0 0 I 0 0 0 0 0 0 2 I 0 0 0 I I 0 0 0 0 0 3 2 0 0 0 0 2 2 2 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 5 4 0 0 2 2 2 4 4 2 2 2 6 3 0 0 0 1 3 3 3 3 3 2 7 3 0 0 3 3 3 3 3 3 3 0 8 6 0 0 6 6 6 6 6 6 6 6 9 15 0 15 15 15 15 15 15 15 15 15 10 104 104 104 104 104 104 104 104 104 104 104

Number of birds replacinga given numberof primaries. that had lost somebut not all secondaries(Fig. Timing and durationof,flight-feather molt.- 3C). Basedon this, we presumethat secondar- Regressionsof collection date on primary- iesare lost more rapidly than the primaries, the molt score for adult versus subadult males in innermostones being the lastto drop (Fig.3C). 1996indicate that, on average,subadult males Growthof new secondariesappears to beginat began and finishedprimary molt on 2 and 26 two foci (S1 and probably about S8), and to August, respectively,versus 15 August and 9 progressin two concurrentwaves: proximally Septemberfor adult males (t = 4.29, df = 53, from S1 to S4,and both proximallyand distally P < 0.001;Fig. 6A-B). Giventhat subadultand from S8. Like the primaries,this appearsto be adult malesboth begin and finishflight-feath- illustratedby the greater length of S1 and S8 er about two weeks offset from one another, it comparedwith their adjacentsecondaries (Fig. is not surprising that they did not differ in 3B).This growthpattern suggests that second- their rate of primary molt (t = 0.02, df = 52, P ariesmay be lost in the samesequence as well. > 0.9; Fig. 6A-B); i.e. total averageduration (œ These conclusions are tentative, however, and --- 95% CI) of molt for adult and subadult merit further study. males was 24.9 - 22.6 days and 23.9 +--22.6 Like the remiges,old rectricesare shedrap- days,respectively. idly; however,we could not discern any se- The sameanalyses conducted on data from quenceof rectrixloss or regrowth.Regression males collected in 1993 indicate that, in con- of rectrix-moltscore on primary-moltscore in- trast to 1996, subadultand adult malesbegan dicates that, on average,rectrix molt begins primary molt at approximatelythe sametime when primaries are 43% grown, and is 88% (11 and 13 August, respectively;t = 1.64, df = completedwhen primary molt is finished(Fig. 15, P > 0.1; Fig. 6A-B). However,like malesin 5B). Similarly,regression of rectrix-moltscore 1996, the rate of primary molt did not differ on secondary-moltscore indicates that, on av- between subadults and adults (t = 0.02, df = erage,rectrix molt begins when secondaries are 14, P > 0.9; Fig. 6A-B); i.e. total averagedu- 20% grown, and is 81% completedwhen sec- ration of molt for adult and subadult males ondary molt is finished.Rectrix molt is com- was 81.1 + 8.6 days and 75.4 + 8.1 days, re- pleted 3 to 10 daysafter completionof second- spectively.It is noteworthy,however, that pri- ary molt, and 2 to 7 days after completionof mary molt in both subadultand adult males primary molt, dependingon the rate of feather required more than three times as long to growth during a given molt; this can be de- completein 1993 than in 1996; perhapsbe- duced from: (1) the rate of primary growth as causeof small sample sizes, however, these indicatedby the regressionof dateon primary- differenceswere not statistically significant molt score(Fig. 6A-B; discussedbelow), and (adults: t = 1.29, df = 49, P > 0.2; subadults: (2) the rateof rectrixmolt in relationto primary t = 1.30, df = 16, P > 0.2). and secondarymolt asindicated by regressions Regressionanalyses for adult and subadult of rectrix-moltscore on primary- and second- femalesindicate no significantrelationship be- ary-molt score(Fig. 5B-C). tween primary-molt score and date in 1996 658 THOMPSONETAL. [Auk,Vol. 115

1 1

FIc. 3. (A) Molting wing of adult CommonMurre growing P2 to P10;P5 and P6 indicatedby numbers on the figure.(B) Molting wing of adult CommonMurre. P1, S1and S8 indicatedby numberson the figure. Note the greaterlengths of the middle primariesrelative to more proximaland distal primaries,and the decreasinglength of the secondariesfrom S1 to ,%4,and from S8 both distallyand proximally.(C) Molting wing of adult CommonMurre growingP1 to P10;S1 to Sll are missing,and S12to S16are old.

(adult females: t = -0.65, P = 0.522; subadult between sexes (within age classes)in 1996. females:t = 0.35, P = 0.732;Fig. 6C). However, Samplesizes of adult (n = 6) and subadult (n primary-molt scoresof subadult females(œ = = 0) females in 1993 were too small for mean- 26.3 + 4.8) weresignificantly higher than those ingful statisticalanalyses. of adultfemales (œ = 2.5 + 1.3)indicating that, From the discussion above, it can be de- like males, subadult females molt earlier than duced that secondaryand rectrix molt are adult females(Mann-Whitney U-test, U = 85, completed,on average,in about 72 and 65%, df = 1, P < 0.001).Median primary-moltscores respectively,of the time requiredfor primary did not differ between subadult males and fe- molt. Becausethe duration of primary molt males (U = 103, df = 1, P = 0.274) or adult varies widely amongyears (Fig. 6A-B), the malesand females(U = 593, df = 1, P = 0.851), duration of secondaryand rectrix molt varies suggestingthat timing of molt did not differ as well. July 1998] Molt in CommonMurres 659

5O n = 25 .e100 Y = -37.9 +1.40x 4O O2•80 R2=0'958 •O.30

• 20 õ 20 o_ 10 co0 • , •(a) 20 40 60 80 100

1 2 3 4 5 6 7 8 9 10 Prima• Molt S•re

LongestGrowing Primary 100 - FIe. 4. Frequencyof longestgrowing primary in Y = -67.5+ 1.55X 25 adultCommon Murres during early stagesof de- finitiveprebasic molt. Because distal primaries grow oø 80 R2= 0.881 oj•'/• significantlyfaster than proximal primaries (C. Thompsonunpubl. data), data were tabulated only n= 19 for specimensin whichgrowing primaries did not exceed30 mm in length. =r_40 DISCUSSION P<0'001 o•ø CommonMurres are distributedthroughout o temperate to arctic marine waters of the North- ,,, , ernHemisphere (Cramp 1985). As a result,they 0 20 40 60 80 100 exhibitextensive geographic variation in mor- phology,as is reflectedby theseven or sosub- PrimaryMolt Score speciesthat are commonlyrecognized (Storer 1952).Therefore, the results of ourstudy apply to the subspeciescalifornica and possiblyinor- 80__] Y=-20.8+ 1.02X• nata(AOU 1957)that we studied,but theypo- tentiallymay not applyto othersubspecies. • R2= 0-828 / Wingand tail shape.--Adult males have slight- • 60__]n=15 // ly butsignificantly longer and narrower wings, and thus a higheraspect ratio, than adult fe-

• 2o

FIG.5. (A) Regressionof secondary-moltscore on primary-molt scorefor all subadult and adult male o and female Common Murres collected in 1993 and 1996with moltingprimaries and secondaries. (B) Re- 0 20 40 60 80 100 gressionof rectrix-moltscore on primary-moltscore for all subadult and adult male and female Common SecondaryMolt Score Murrescollected in 1993and 1996 with moltingpri- mariesand rectrices.(C) Regressionof rectrix-molt 1993and 1996with moltingsecondaries and rectri- score on secondary-moltscore for all subadult and ces.Regression lines are shown _+95% confidencein- adult male and female Common Murres collected in tervals. 660 THOMPSONET AL. [Auk, Vol. 115

males.To our knowledge,such a differenceis Oct 15 _•o= 1993, n=12 not known within a singlepopulation of any Y=223 + 0.811X speciesthat is relativelymonomorphic in body _ R2=0.981 size, but it has been described between sexesin Sept 15 raptors that are sexually size dimorphic P<0.00•,• (Mueller et al. 1981), and in migratory versus eL JJ•" [] = 1996,n=42 nonmigratorypopulations of various passerines(e.g. Berthold and Querner1982, Chandler Aug 15 • Y=227+ 0.249X and Mulvihill 1988).The only potentialadap- tive significancethat we are awareof that may explainthe differencein wing shapebetween Jul 15 I I I I I I I I I male and female murres relates to differences 0 20 40 60 80 100 in body mass.Among different populations of CommonMurres, no consistentdimorphism in wing lengthor body massexists between males O= 1993, n=6 and females.Males are slightlyheavier and/or Oct 15 - longer-wingedin somepopulations and slight- ly smaller and/or shorter-wingedin others (Cramp 1985,Harris and Wanless1988). How- .

independentlyin eachof theselineages. In ad- thatthe rectrices complete regrowth two to sev- dition,given the differentnatural histories and en daysafter primary molt is complete.In con- morphologies(e.g. wing shapes) of these trast,Birkhead and Taylor (1977) estimated that groups,the adaptive significanceof this molt rectrixmolt wasnot finisheduntil 30 daysafter strategyis unclear. completionof primary molt. Previous authors have stated that Common Calendartiming.--Flight-feather molt in both Murres shed their primaries nearly simulta- 1993 and 1996 began in early to mid-August. neously, e.g. "synchronously"(Stresemann However, because this molt took more than and Stresemann1966), "simultaneously" (Tav- threetimes as longto completein 1993than in erner 1929, Witherby et al. 1941, Salomonsen 1996,molt finishedin late Augustto mid-Sep- 1944, Tuck 1960,Croll 1990), "all shed at one temberin 1996 versuslate Octoberto early No- time" (Storer 1952), "almost simultaneously" vemberin 1993.Elsewhere, flight-feather molt (Ginn and Melville 1983), "almost all at once" hasbeen reported to occurin late Julyor early (Dement'evand Gladkov 1951)," within oneto Augustthrough mid- to late September(For- three days" (Cramp 1985), and "within a few bush 1925, Dement'ev and Gladkov 1951, days of each other," based on captivebirds Cramp 1985,Hope Jonesand Rees1985, Harris (Birkheadand Taylor1977). and Wanless1990), July to early or lateNovem- Secondarymolt begins and finisheswhen ber (Verwey1924, Witherby et al. 1941, Salo- primaries are about38 and 99% grown, respec- monsen 1944), September and October (De- tively (seeBirkhead and Taylor1977). New sec- ment'ev and Gladkov 1951, Kozlova 1957), and ondariesappear to be lost and regrownbegin- "beginning sometimesin Augustbut often not ning at two foci (S1 and probably S8) and pro- until September,"its ending date unspecified gressingin two concurrentwaves: proximally (Bent1919). Some of the variationamong these from S1to S4,and proximallyand distallyfrom datesreflects the fact that northernpopulations S8. As with primary molt, all previoussources molt later, on average,than more southerly statethat secondariesare lostand regrown"si- populations(Cramp 1985), and that young sub- multaneously"(e.g. Birkhead and Taylor1977). adults as well as older adult nonbreeders and Regardingthe timing of secondarymolt rela- failed breedersbegin and finish molt earlier tive to primary molt, previous sources vary than successfulbreeders (Verwey 1924, Birk- from stating that secondariesare dropped at head and Taylor 1977, Swennen1977, Hope the sametime, or nearly so, as the primaries Jonesand Rees1985, Harris and Wanless1990). (Salomonsen 1944, Dement'ev and Gladkov Age and sex differences.--Asfound in many 1951,Storer 1952, Tuck 1960)to when the pri- otherseabirds (e.g. Palmer 1962, Cramp 1977, mariesare "almosthalf grown"(Birkhead and 1983, 1985, Warham 1996), we found that sub- Taylor 1977), or after the primaries have adult Common Murres molt earlier than adults. dropped (Witherby et al. 1941). Swennen(1977) reported the sameresults re- Rectricesare lost very rapidly and in no dis- garding a captive population of Common cernableorder, as foundby Birkheadand Tay- Murres. Because subadults do not breed or care lor (1977) and Storer (1952). On average,the for young, they are able to molt earlier than rectricesbegin to regrow when the primaries adults,especially adults that breedsuccessful- and secondariesare 43 and 20% grown, re- ly. spectively,but at the completionof primary We detectedno differencein timing of molt and secondarymolt, they are only 88 and 81% between sexes within age classes.Primary- grown. Previousauthors have stated that rec- molt score was significantlycorrelated with tricesare dropped"simultaneously" with the Julian date in both adult and subadult males primaries(Tuck 1960),"shortly after primaries but not in adult or subadult females (Fig. 6A havebegun to regrow" (Storer1952), "simul- and 6B vs. 6C). The reason for this is unclear taneously"with the secondaries(Witherby et Anecdotalevidence suggests that males molt al. 1941), when the primaries are "about half eitherearlier or laterthan females in otherpop- grown" (Birkheadand Taylor1977), when the ulations (Hope Jonesand Rees 1985, Taskeret primariesare "lessthan 120mm" (a little more al. 1987).Common Murre chicksfledge at about than half grown;Salomonsen 1944), or "later" 21 to 26 daysof agebefore their primarieshave than the primaries(Kozlova 1957). We found begunto grow.As a result,they are accompa- 662 THOMPSONET AL. [Auk, Vol. 115 nied at seaand providedfood by their fathers Other less-quantitativeestimates of flight- until they are capableof flight at about100 days feathermolt durationinclude "about 60 days" of age, and possiblelonger (Birkhead1984, (Ginn and Melville 1983) and "less than 70 1993; Hope Jonesand Rees 1985). For a few days" (Harris and Wanless1990). weeksafter their younghave left the colony,fe- Evolutionaryimplications.--The unusual se- malescontinue to return to thebreeding colony quenceof replacementof primariesin Common (Varoujean et al. 1979, Wanless and Harris Murres raises the intriguing question:Why 1986).Depending on the rate of molt in a given doestheir primarymolt begin at a centralfocus year,Common Murres are thoughtto be flight- in the middle of the primariesand progressin less for 45 to 60 days during molt (Glutz von two concurrent waves? Because murres lose all Blotzheim and Bauer 1982). As a result, rather of their primarieswithin a few days, it is dif- than delayingmolt until they are finishedcar- ficult to imaginethat the unusualsequence by ing for their young,one might expectsuccess- which they lose them is more advantageous ful breedingmales to molt earlier than their than the moretypical "synchronous"or rapid matesbecause they may be constrainedto do- primary replacementof otherlarge alcidsand ing little or no flyingwhile accompanyingtheir diving seabirds(e.g. May 1930, Palmer 1962, flightlessyoung at sea. Alternatively,because Stresemann and Stresemann 1966, Watson females are emancipatedfrom parental duties 1968, Bellrose 1980, Piersma 1988, Warham after their younggo to sea,they maybe ableto 1996). The most likely a priori explanationfor begin molting as soonas they leavethe breed- the primary-moltsequence in CommonMurres ing colony,whereas the continuedparental du- is that it reflectstheir phylogenetichistory, i.e. ties of males may precludethem from begin- that an ancestor exhibited the same molt se- ning molt until their youngare more self suf- quencefor reasonsunknown. ficient. CommonMurres are most closelyrelated to Molt duration.--Durationof flight-feather Thick-billed Murres (Uria lomvia), Razorbills molt appearsto be remarkablyvariable. In both (Alca torda), Dovekies (Alle alle), and Great subadultand adult malesin 1996,primary molt Auks(Pinguinus impennis; Strauch 1985, Friesen required 24 to 25 days, whereasin 1993 it re- et al. 1993, Moum et al. 1994; contra Chandler quired 75 to 81 days,on average.Adult murres 1990). Like the literature on Common Murres, entangledin PugetSound in the fall are com- that on Thick-billed Murres, Razorbills, and prised of postbreedersfrom both Oregonand Dovekiesstates that they molt their flight feath- Washington.Postbreeding murres from Ore- ers synchronously(Taverner 1929, Salomonsen gon molt aboutfour to six weeksearlier than 1944, Dement'ev and Gladkov 1951, Strese- those from Washington.Thus, between-year mann and Stresemann 1966, Bradstreet 1982, differences in molt duration in adults could be B6dard1985, Cramp 1985,Hope Jonesand Rees due to different proportionsof postbreeding 1985, Harris and Wanless1990). However, the Oregonversus Washington murres. However, oneand only known moltingspecimen of Great becausenonbreeding subadult males showed Auk (Grieve1885, Meldgaard 1988) appears to nearly identical between-year differences in indicatethat primary molt beganin the middle molt duration,this possibilityis unlikely. of the primaries as in CommonMurres. This On average,secondary and rectrix molt re- specimenhas "on the right wing, a pin feath- quire 72 and 65% as much time as primary er..near the middle of the seriesof primaries; molt. Basedon quantitativeanalyses similar to and on the left wing, one primary [P4 or P5] ours,Birkhead and Taylor(1977) estimated the appearsnew whereasthe rest of the primaries duration of primary, secondary,and rectrix are faded and badly worn" (Storer1960). Finn molt to be 62.7 ---16.3 days,25.0 -+ 8.4 days,and Salomonsen made similar comments about this 61.0 --- 16.8 days, respectively,giving a total specimento the Stresemanns(Stresemann and flight-feathermolt duration of about 93 days, Stresemann 1966). This indicates that Great on average.Similarly, Glutz von Blotzheimand Auks molted their primaries slowly. In addi- Bauer (1982) stated the duration of primary, tion, we speculatethat they molted their pri- secondaryand rectrix molt to be 42 to 90 days, mariessequentially, presumably in the sameor 25 days, and 35 to 86 days, respectively,al- similar order as Common Murres (Salomonsen thoughthe source(s)of thesedata is unclear. 1945, Storer 1960, Stresemann and Stresemann July1998] Molt in CommonMurres 663

1966).In addition,in the absenceof specimens drop and regrow their primariesmore rapidly from whichthis couldbe deducedempirically, thanby droppingand regrowingthem in a sin- simple physical principlessuggest that this gle wave,as hasbeen argued for otherspecies must have been the case. in whichthe long durationof molt may inter- The wing sizeand shapeof volantalcids are fere with other energeticallydemanding peri- a compromisebetween selectionpressures for odsin the annualcycle (e.g. Langston and Roh- flying in the air and swimming underwater. wer 1996). Therefore,the wing is not optimallydesigned These conclusions raise two additional inter- for either function.As a result,compared with estingquestions. First, do Thick-billedMurres, non-diving seabirdsof comparablebody sizes Razorbills, and Dovekiesmolt their primaries (Warham 1977), alcids have relatively high synchronouslyas the literature suggests?We wing loading,ranging from lessthan 1.5 g per speculatethat they do not. The molt sequence cm2 of wing area for most small alcidsup to in CommonMurres was overlookedby previ- about 2 g per cm2 for large alcids, including ousresearchers for many obviousreasons: (1) CommonMurres (Livezey 1988).Because Great during molt, individual birds loseall of their Auks were flightless,their wings were more primariesvery rapidly,probably within a few optimallyadapted for flight underwater.Thus, days;thus, depending on the synchrony of molt like penguins,their wing area was small rela- within a population,a large number of speci- tive to their body size, resultingin very high mens must be examined within a narrow win- wing loading,estimated to havebeen about 22 dow of time to detect the sequenceof lossof g per cm2 of wing area,more than 10 timesthe primaries;(2) like other large alcids,they molt maximumwing-loading exhibited by Common at sea where it is more difficult to collect birds Murres (Livezey 1988). Similarly,the ratio of than on a breedingcolony; (3) molting speci- their wing area to cross-sectionalbody area mensare more difficult to prepareand less aes- also was very high. If Great Auks lost all of thetically attractiveas study skins than are their primariesrapidly like largevolant alcids, nonmoltingbirds; thus, historicallyand even they would not have had sufficientwing area currently,specimen preparators have preferred remaining to propel themselvesfast enough not to preparemolting specimens;(4) for the underwaterto obtainfood and, perhaps,to es- two precedingreasons, relatively few molting capepredators. Therefore, they must have re- specimensare in museums;and (5) because placedtheir primariesslowly and sequentially. nearly all study skinsare preparedas "round" This raisesanother questionthat we will not skinswith both wings folded,examining rem- address,but that is worth pondering:Why did igesfor molt is difficultor impossiblewithout Great Auks molt their flight feathersin two damagingspecimens. Thus, we suggestthat slow and concurrent sequentialwaves pro- Thick-billed Murres, Razorbills, and Dovekies gressingproximally and distallyfrom a focus molt in the samesequence as CommonMurres in the middle of the primaries,rather than in a and Great Auks. Given the tens of thousands of single wave or possiblytwo waves from two alcidskilled in commercialfisheries each year foci? in both the North Atlantic and North Pacific Returningto our original question:Why do during their moltingseasons, obtaining speci- CommonMurres replace their primariesin two mensto answerthis questionshould be rela- concurrentwaves from a single focus in the tively easy(King 1984,Ogi 1984,Piatt and Net- middle of the primaries?Unfortunately, there is tleship1987, DeGange et al. 1993).Alternative- no obviousanswer to this question.Uria ap- ly, observationsof captive alcids in zoos and pearsto have evolvedbefore Pinguinus. Thus, aquaria also could help answerthis question the presumedslow molt of Great Auks proba- (Douma and Carlson 1994, Gunther 1994). bly is a derivedcharacter, whereas the two con- The secondquestion is why do Dovekieslose currentwaves of molt likely may be a shared- theirprimaries rapidly? Other than the Whisk- derived character of the clade to which Alle, ered Auklet (Aethiapygmaea) and LeastAuklet Uria, Alca,and Pinguinusbelong, and possibly (A. pusilia),Dovekies are the smallestin mass of otheralcids as well (Braune1987). The only of the 23 extant alcids (Bent 1919,Johnson 1944, selectiveadvantage that we canthink of is that Cramp 1985).All othersmall- to medium-size this molt strategymay allow thesealcids to alcids (Aethia, Ptychoramphus,Cyclorrhynchus, 664 THOMPSONET AL. [Auk, Vol. 115 and Brachyramphus)replace their primariesse- large alcidshave muchhigher wing loading quentiallyand more slowly than larger alcids than small alcids. (Payne1965; B•dard and Sealy1984; Emslie et In contrast,for a large alcid to maintain the al. 1990;Byrd andWilliams 1993;Manuwal and sameunderwater flight ability as a smallalcid, Thoresen1993; Carter and Stein1995; Konyu- it would need to maintain the same ratio of khov and Kitaysky1995; Jones 1993a, b; Nelson wing-surfacearea to body cross-sectionalarea. 1997),whereas larger alcids(Synthliboramphus, Because wing-surface area and body cross-sec- Cerorhinca,Cepphus, Alca, and Fratercula)re- tional areaboth scalewith the squareof linear placetheir primaries much more rapidly (Har- bodymeasurements, this ratio is maintainedby ris and Yule 1977, Ewins 1988, 1993, Drost and allometryalone. Thus, as body mass increases, Lewis 1995, Gastonand Dechesne1996). Like selectionfavors relatively larger wings for ae- our discussionof wing loading above,simple rial versusunderwater flight-such that the dif- physicalprinciples and allometricconsidera- ferencebetween the wing sizesthat are optimal tionsexplain why primariesshould be molted for each"flight" modealso increases. In short, slowly in small alcidsand rapidly in large al- this meansthat the wings of small alcidsare cids.The explanationfor thisgeneral difference nearlyoptimal in sizefor aerialflight when ful- between small and large alcids ostensiblylies ly extendedand also for underwaterflight in understanding what determines optimal when partially folded,whereas the wings of wing size for aerial versusunderwater flight. large alcidsare smallerthan optimalfor aerial Optimal wing area for aerial flight is dictated flight and largerthan optimalfor underwater by wing loading. In contrast,for underwater flight. The supposedconsequence of this for flight,optimal wing areais dictatedby theratio molt in small alcidsis that if they lose all of of the cross-sectionalarea of a bird'sbody di- theirprimaries rapidly, their wing-surface area vided by the surfacearea of its partly folded will bereduced so far belowwhat is optimalfor underwaterflight that their ability to fly un- wings (the partly foldedwings during under- derwater is seriously compromised.In addi- water flight increasestructural support of the tion, as in large alcids,rapid lossof primaries wing). If largealcids have the same proportions wouldmake aerial flight impossible for mostof as small alcids,then as linear body measures the moltingperiod until the new primariesare (bodydiameter and wing length)increase from about80% grown (Birkheadand Taylor1977). small to large species,body volume (mass) As a result,most small alcids replace their pri- shouldincrease with the cube,whereas body maries(and other flight feathers) slowly, there- cross-sectionalarea and wing area shouldin- by maintainingefficient underwater flight abil- creasewith the square.In addition,wing area ity and also retaining their power of aerial shouldincrease with body massto the 0.667 flight throughoutmolt. In contrast,large alcids power, resulting in increasingwing loading havesuch high wing loadingthat evena slow, with increasingbody mass(Storer 1960, Gree- sequentialflight-feather molt would increase newalt 1962, 1975).Thus, for a large alcid to their wing loadingsufficiently to precludeae- maintain the same aerial flight ability as a rial flight during molt (Meunier1951). In ad- small alcid, it would need to maintain the same dition, rapid lossof all of their primariesre- wing loading (or increasewing power, i.e. in- ducestheir wing-surfacearea to a size that is creasewingbeat frequency).To do so, wing- probablyclose to the optimal size for under- surface area would have to increase with the waterflight. Thus, all largealcids lose their pri- cube,rather than the square,of body mass,i.e. maries very rapidly. wing area would have to increasewith body The questionremains, why do Dovekieslose massto the 1.0 power, rather than to the 0.667 their primariesrapidly? As mentionedabove, power.In fact,however, empirical data indicate they havethe third lightestbody massof any that wing area of extant alcidsscales to 0.588 alcid.With thepossible exception of LeastAuk- (Warham1977) to 0.632_+ 0.003 (Livezey 1988); lets (Livezey 1988),they also have the lowest both empiricallyderived exponents are statis- wing loading of any alcid (0.54 to 0.94 g/cm2; ticallysignificantly less than the power of 0.667 Magnan 1922, Poole 1938, Livezey 1988), yet predicted by the law of isometry described many other relativelysmall alcidsthat have above (Warham 1977, Livezey 1988). Thus, considerablyhigher wing loading(1.02 to 1.32 July1998] Moltin CommonMurres 665 g/cm2) molt their remiges sequentially and rel- dant in 1996 (J. Parrish unpubl. data). Male ativelyslowly. In light of the physicalprinci- Common Murres molted three times faster in ples discussedabove, the Dovekie'ssynchro- 1996than in 1993.Such dramatic interyear vari- nous molt appears to be a curious paradox. ationin molt rate is virtually unknownin birds. However,we suggesta possibleexplanation. In addition, subadult males molted two weeks The argumentsdiscussed above are based earlier than adult males in 1996, but not in entirelyon theory;more importantly, perhaps, 1993.Although speculative, we suggestthat they implicitly assumethat selectionfavors these differences in molt reflect differences in maintainingaerial flight ability throughout prey availabilitybetween years. flight-feathermolt in all alcidspecies that are sufficientlysmall to be physicallycapable of ACKNOWLEDGMENTS doing so.We questionthis assumptionand ar- guethat selectionmay favorflight-feather molt This paper is dedicatedto CWT'swife, Krystyna Simm,for 15 yearsof unwaveringlove and support. to be rapid andnearly synchronous rather than We are gratefulto ChrisWood and The BurkeMu- slow and sequentialin Dovekies because,un- seumZoology Division for logisticalsupport while like anyother small alcid, in the fall adultsand conducting necropsiesand processingspecimens. young of all populationsof Dovekiesundergo JonFisher and the WesternFoundation of Vertebrate long-distancemigration to moresoutherly win- Zoologykindly providedus with manyobscure ref- tering areas(Bent 1919,Dement'ev and Glad- erences that would have been much more difficult to kov 1951, Salomonsen 1950, Kozlova 1957, B•- acquire elsewhere.Doug Siegel-Causey,Janet Hin- dard 1985, Cramp 1985). As a consequence, shaw(Josselyn Van Tyne Library, University of Mich- Dovekiesmust molt before,during, or after mi- igan),The EdwardGrey Institute of FieldOrnithol- ogy (Oxford University),Canada Institute for Sci- gration.Like mostother birds, Dovekies do not entific and TechnicalInformation (National Research undergoflight-feather molt during migration Council of Canada), National TranslationsCenter or on their wintering grounds.Instead, they (Chicago), National TechnicalInformation Service molt on their breedinggrounds, presumably (Springfield,Virginia), and the CanadianWildlife becausethe costof doingso is lessthan that of Service(Ottawa) provided many usefultranslations molting at other times, possiblybecause of of foreignliterature. Matthias Leu kindly translated greater food availability and/or lower preda- many Germanreferences. Bob Mulvihil! provided tion risk on the breedinggrounds. As a result, referencesregarding wing shapein birds. Dennis however,they are constrained to completetheir Paulsonallowed us to examinespecimens in the Slat- er Museumof Natural History, Universityof Puget flight-feathermolt within a relativelyshort pe- Sound.We are especiallygrateful to Harry Carter, riod of time. As a consequence,a slow sequen- Dennis Paulson, and Ken Warheit, whose keen in- tial molt of flight feathersis not possiblebe- sight and vast knowledgeabout seabird biology in- causeit would require too much time. For ex- spired many valuable and productive discussions ample,Least Auklets, the ecologicalequivalent aboutevolution of life historiesand molt strategies. of Dovekiesin the PacificOcean (Jones1993b), Jeff Marks, Ken Parkes,Robert Storer,Eric Taylor, requireat leastthree months to completetheir and an anonymousreviewer provided extensive, de- flight-feathermolt (B•dard and Sealy 1984). tailed,and very helpful comments on the finaldraft of this paper. Similarly, nonbreedingsubadult Cassin'sAuk- lets (Ptychoramphusaleuticus) require an aver- LITERATURE CITED age of 146 days to finish flight-feathermolt (Emslie et al. 1990). As a result, Dovekies must AMERICAN ORNITHOLOGISTS' UNION. 1957. Check- molt theirflight feathers nearly simultaneously list of North American birds, 5th ed. American like large alcids in order to complete their Ornithologists'Union, Baltimore. flight-feathermolt before they begin fall migra- ASHMOLE,N. P. 1962.The BlackNoddy Anoustenui- tion. rostris on Ascension Island. Part 1. General bi- Correlateswith preyabundance.--E1 Nifio con- ology.Ibis 103b:235-273. BAKER,K. 1993. Identification guide to European ditionsprevailed along the entireWest Coast of non-passerines.British Trust for Ornithology North Americain 1993,resulting in low food Guide No. 24. British Trust for Ornithology, availabilityand nearly completereproductive Thetford,United Kingdom. failure by murres(R. Lowe and J. Parrishun- BgDARD,J. 1985.Evolution and characteristicsof the publ. data).In contrast,food supply was abun- AtlanticAlcidae. Pages 1-51 in The AtlanticA1- 666 THOMPSONET AL. [Auk, Vol. 115

cidae:The evolution,distribution and biologyof the Middle East and North Africa. Volume I. Os- the auksinhabiting the AtlanticOcean and ad- trich to ducks.Oxford UniversityPress, Oxford. jacent water areas (D. N. Nettleship and T. R. CIdAMP,S. 1983.Handbook of the birds of Europe, Birkhead,Eds.). AcademicPress, New York. the Middle East and North Africa. Vol. III, Wad- BI•DARD,J., AND S. G. SEALY.1984. Moults and feather ers to gulls. Oxford University Press,Oxford. generationsin the Least, Crestedand Parakeet CIdAMP,S. 1985.Handbook of the birds of Europe, auklets.Journal of Zoology (London) 202:461- the Middle East and North Africa. Volume IV. 488. Ternsto woodpeckers.Oxford University Press, BELLROSE,E C. 1980. Ducks, geese and swans of Oxford. North America,3rd ed. StackpoleBooks, Har- CI•OLL,D. A. 1990.Physical and biologicaldetermi- risburg, Pennsylvania. nants of the abundance,distribution, and diet of BENT, A. C. 1919. Life histories of North American the CommonMurre in MontereyBay, California. diving birds. Order Pygopodes. United States Pages139-148 in Auks at sea (S.G. Sealy,Ed.). National Museum Bulletin No. 107. Studiesin Avian BiologyNo.14. BERTHOLD,l•., AND U. QUERNER. 1982. Genetic basis DEGANGE, A. R., R. H. DAY, J. A. TAKEKAWA,AND V. of moult, winglength,and body weight in a mi- M. MENDENHALL.1993. Losses of seabirdsin gill gratory bird species,Sylvia atricapilla. Experien- net fisheriesin the North Pacific.Pages 204-211 tia 38:801-802. in The status,ecology and conservationof ma- BIRKHEAD, t. R. 1984. Distribution of the bridled rine birds of the North Pacific(K. Vermeer,K. T. form of the CommonGuillemot Uria aalgein the Briggs, K. H. Morgan, and D. Siegel-Causey, North Atlantic. Journalof Zoology (London) Eds.).Special Publication of the CanadianWild- 202:165-176. life Service, Ottawa, Ontario. BIRKHEAD, t. R. 1993. Great Auk islands. A field bi- DEMENT'EV, G. P., AND N. A. GLADKOV. 1951. Birds ologistin the arctic.T. and A.D. Poyser,London. of the SovietUnion. Vol. 2. Israel Programfor BIRKHEAD, t. R., AND A.M. TAYLOR. 1977. Moult of Scientific Translations, Jerusalem. the GuillemotUria aalge.Ibis 119:80-85. DOUMA,B., AND M. CARLSON.1994. Breeding alcids BOLSSOW,S. 1996.1994 Coho test fishery.Final Re- at the SeattleAquarium. International Zoo Year- port, WashingtonDepartment of Fishand Wild- book 33:142-147. life, Olympia. DROST, C. A., AND D. B. LEWIS. 1995. Xantus' Murre- BRADSTREET,M. S. W. 1982.Pelagic feeding ecology let (Synthliborarnphushypoleucus). In The birdsof of Dovekies, Alle alle, in Lancaster Sound and North America, no. 164 (A. Poole and E Gill, WesternBaffin Bay. Arctic 35:126-140. Eds.). Academy of Natural Sciences,Philadel- BRAUNE,B. M. 1987. Comparisonof total mercury phia, and American Ornithologists' Union, levelsin relationto diet andmolt for ninespecies Washington,D.C. of marine birds. Archives of Environmental EMSLIE, S. D., R. P. HENDERSON, AND D. G. AINLEY. Contaminationand Toxicology16:217-224. 1990.Annual variationof primary-moltwith age BROUGHTONJ. M. 1994. Size of the bursa of Fabricius and sex in Cassin's Auklet. Auk 107:689-695. in relationto gonadsize and agein Laysanand EWINS,P. J. 1988.The timing of moult in BlackGuil- Black-footed albatrosses. Condor 96:203-207. lemotsCepphus grylle in Shetland.Ringing and BYRD, G. V., AND J. C. WILLIAMS. 1993. Whiskered Migration 9:5-10. Auklet (Aethiapygrnaea). In The birds of North EWiNS,P. J. 1993. Pigeon Guillemot (Cepphuscolurn- America,no.76 (A. Pooleand E Gill, Eds.).Acad- ba). In The birds of North America, no. 49 (A. emy of Natural Sciences,Philadelphia, and Pooleand E Gill, Eds.).Academy of Natural Sci- American Ornithologists'Union, Washington, ences,Philadelphia, and AmericanOrnitholo- D.C. gists'Union, Washington,D.C CARTER,H. R., ANDJ. L. STEIN.1995. Molts and plum- FOIlBUSH,E. H. 1925. Birds of Massachusettsand oth- agesin the annualcycle of theMarbled Murrelet. er New England states. Part I. Water birds, Pages99-109 in Ecologyand conservationof the marsh birds and shore birds. Commonwealth of Marbled Murrelet (C. J.Ralph, G. L. Hunt, Jr.,M. Massachusetts, Boston. G. Rafael,and J.E Piatt, Eds.).Pacific Southwest FORSHAW,J. M., AND W. t. COOPER.1989. Parrots of ResearchStation General Technical Report No. the world, 3rd ed. Lansdowne Editions, Mel- 152. bourne, Australia. CHANDLER, C. R., AND R. S. MULVIHILL. 1988. The FRIESEN,V. L., W. g. MONTEVECCHI, AND W. S. DA- useof wing shapeindices: An evaluation.Ornis VIDSON.1993. Cytochromeb nucleotide se- Scandinavica 19:212-216. quencevariation among the Atlantic Alcidae. CHANDLER,R. M. 1990.Phylogenetic analysis of the Hereditas 119:245-252. alcids.Ph.D. dissertation,University of Kansas, GISTON, A. J., AND S. B.C. DECHESNE.1996. Rhinoc- Lawrence. eros Auklet (Cerorhincarnonocerata). In The birds CIdAMP,S. 1977. Handbookof the birds of Europe, of North America, no. 212 (A. Pooleand F. Gill, July1998] Moltin CommonMurres 667

Eds.). Academy of Natural Sciences,Philadel- KONYUKHOV,N. B., AND A. S. KITAYSKY.1995. The phia, and American Ornithologists'Union, Asianrace of the MarbledMurrelet. Pages 23-29 Washington,D.C. in Ecologyand Conservationof the Marbled GINN, H. B., AND D. S. MELVILLE. 1983. Moult in Murrelet (C. J.Ralph, G. L. Hunt Jr.,M. G. Rafael, birds. British Trust for OrnithologyGuide No. and J. E Piatt, Eds.). PacificSouthwest Research 19, BritishTrust for Ornithology,Tring, United StationGeneral Technical Report No. 152. Kingdom. KOZLOVA,E. V. 1957. Fauna of the USSR. Birds. Vol. GLUTZ YON BLOTZHEIM, U. N., AND K. M. BAUER. 2, Part 3. Charadriiformes, suborder Alcae. Is- 1982. Handbuchder V/SgelMitteleuropas. Aka- rael Programfor ScientificTranslations, Jerusa- demischeVerlagsgesellschaft, Weisbaden, Ger- lem. many. LANGSTON,N. E., AND S. ROHWER.1996. Molt/breed- GREENEWALT,C. H. 1962.Dimensional relationships ing tradeoffsin albatrosses:Implications for un- for flying animals. SmithsonianMiscellaneous derstandinglifehistory variables. Oikos 76:498- Collections 144(2):1-46. 510. GREENEWALT,C. H. 1975.The flight of birds. Trans- LIVEZEY,B.C. 1988.Morphometrics of flightlessness actionsof the American PhilosophicalSociety in the Alcidae. Auk 105:681-698. 65:1-67. MAGNAN, M. A. 1922. Les caracteristiques des GRIEVE,S. 1885.The GreatAuk or Garefowl.(Alca im- oiseaux suivant le mode de vol. Annales des Sci- pennis,Linn.). Its history,archeology, and re- encesNaturelies (Series10) 5:125-334. mains. Thomas C. Jack,London. MANUWAL, D. A., AND A. C. THORESEN.1993. Cas- GUNTHER, M. R. 1994. Alcids in North American sin's Auklet (Ptychoramphusaleuticus). In The zoos and aquaria. International Zoo Yearbook birds of North America,no. 50 (A. Pooleand E 33:136-141. Gill, Eds.).Academy of Natural Sciences,Phila- HARRIS, M. E, AND S. WANLESS.1988. Measurements delphia, and AmericanOrnithologists' Union, and seasonalchanges in weight of Guillemots Washington,D.C. Uria aalgeat a breedingcolony. Ringing and Mi- MAY, J. B. 1930. Simultaneousloss of primaries in gration 9:32-36. prenuptialmolt of loon.Auk 47:412-414. HARRIS, M.P., AND S. WANLESS. 1990. Moult and au- MELDGAARD,M. 1988. The Great Auk, Pinguinusim- tumn colony attendanceof auks. British Birds pennis (L.) in Greenland. Historical Biology 1: 83:55-66. 145-178. HARRIS,M.P., AND R. E YULE. 1977. The moult of the MELVIN, E. E, L. CONQUEST,AND J. K. PARRISH.1997. Puffin, Fratercula arctica. Ibis 119:535-541. Seabirdbycatch reduction: New toolsfor Puget HOPEJONES, P., AND E. I. S. REES.1985. Appearance Sound drift gillnet salmon fisheries.1996 sock- and behavior of immature Guillemots and Ra- eye and 1995chum salmonfisheries. Final Re- zorbills at sea. British Birds 79:370-377. port, WashingtonSea Grant Program,Seattle. HUMPHREY,P.S., AND K. C. PARKES.1959. An ap- MEUNIER, K. 1951. Korrelation und Umkonstruktion proachto the studyof moltsand plumages. Auk in den Gr/Sssenbeziehungenzwischen Vogelflu- 76:1-31. gel und Vogelk/Srper.Biologia Generalis 19:403- JENNI,L, AND R. WINKLER.1994. Moult and ageing 443. of European passerines.Academic Press, New MOUM, T., S. JOHANSEN,K. E. ERIKSTAD,AND J. E York. PIATT. 1994. Phylogenyand evolution of the JOHNSON,R. A. 1944.Weight records for someAtlan- auks(subfamily Alcinae) based on mitochondri- tic Alcidae. Wilson Bulletin 56:161-168. al DNA sequences.Proceedings of the National JONES,I. L. 1993a. Crested Auklet (Aethiacristatella). Academyof Sciences(USA) 91:7912-7916. In The birds of North America,no. 70 (A. Poole MUELLER, H. C., D. D. BERGER,AND G. ALLEZ. 1981. and F. Gill, Eds.). Academyof Natural Sciences, Age and sex differencesin wing loading and Philadelphia, and American Ornithologists' other aerodynamic characteristicsof Sharp- Union, Washington,D.C. shinned Hawks. Wilson Bulletin 93:491-499. JONES,I. L. 1993b.Least Auklet (Aethiapusilia). In The NELSON,S. K. 1997. Marbled Murrelet (Brachyram- birds of North America, no. 69 (A. Poole and E phusmarmoratus). In The birdsof North America, Gill, Eds.).Academy of Natural Sciences,Phila- no.276(A. Pooleand E Gill, Eds.).Academy of delphia, PA, and American Ornithologists' Natural Sciences,Philadelphia, and American Union,Washington, D.C. Ornithologists'Union, Washington, D.C. KING, W. B. 1984. Incidentalmortality of seabirdsin NEWTON,I. 1966.The moult of the BullfinchPyrrhula gillnets in the North Pacific.Pages 709-715 in pyrrhula.Ibis 108:41-67. Status and conservation of the world's seabirds NITZSCH,C. L. 1840.System der Pterylographie.1867 (J.P.Croxall, P. G. H. Evans,and R. W. Schreiber, EnglishTranslation by P. L. Sclater.Ray Society, Eds.). International Council for Bird Preserva- London. tion Technical Publication No. 2. OcI, H. 1984.Seabird mortality incidental to theJap- 668 THOMPSONET AL. [Auk, Vol. 115

anesesalmon gill-net fishery.Pages 717-721 in OrnithologicalCongress (G. Bergmann,K.O. Status and conservation of the world's seabirds Donner, and L. von Haartman, Eds.). Helsinki, (J.P.Croxall, P. G. H. Evans,and R. W. Schreiber, 1958.Tilgmannin Kirjapaino, Helsinki. Eds.). International Council for Bird Preserva- STRAUCHJ. G., JR.1985. The phylogenyof the Alci- tion Technical Publication No. 2. dae. Auk 102:520-539. PALMER,R. S. 1962. Handbook of North American STRESEMANN, E., AND V. STRESEMANN. 1966. Die birds, vol. 1. Loonsthrough Flamingos. Yale Mauserder V6gel.Journal fiir Ornithologie107 University Press,New Haven, Connecticut. Sonderheft. PAYNE,R. g. 1965.The molt of breedingCassin Auk- STRESEMANN, V., AND E. STRESEMANN. 1960. Die lets. Condor 67:220-228. Handschwingenmauserder Tagraubv6gel. Jour- PIATT, J. E, AND D. N. NETTLESHIP.1987. Incidental nal fiir Ornithologie101:373-403. catchof marinebirds and mammalsin fishing SUMMERS,R. W. 1980.On the rateof changein moult nets off Newfoundland, Canada. Marine Pollu- scoresin waders.Wader Study Group Bulletin tion Bulletin 18:344-349. 28:24. PIERCE,D. J., W. P. RITCHIE,AND R. KREUZIGER.1994. SwENNEN,C. 1977.Laboratory research on sea-birds. Preliminary findings of seabird interactions Netherlands Institute for Sea Research, Texel, with the non-treatysalmon gill net fishery:Pu- The Netherlands. get Soundand Hood CanalWashington. Final TASKER, M. L., A. WEBB, A. J. HALL, M. W. PIEN- Report, WashingtonDepartment of Fish and KOWSKI,AND D. g. LANGSLOw. 1987. Seabirds in Wildlife, Olympia. theNorth Sea. Nature Conservancy Council, Pe- PIERSMA,t. 1988.The annualmoult cycleof Great terborough,United Kingdom. Crested Grebes. Ardea 76:82-95. TAVERNER,P. A. 1929. The summer molt of the Ra- PIMM, S. 1976. Estimation of the duration of bird zor-billed Auk (Alca torda).Auk 46:223-224. molt. Condor 78:550. THOMPSON, C. W., M. L. WILSON, D. J. PIERCE,AND POOLE,E. L. 1938.Weights and wing areasin North D. DEGHETTO.1998. Population characteristics American birds. Auk 55:511-517. of Common Murres and Rhinoceros Auklets en- PYLE, P., S. N. G. HOWELL, R. P. YUNICK, AND D. E tangled in gillnet fisheriesin Puget Sound, DESANTE.1987. Identificationguide to North Washington1993-1994. Northwestern Natural- Americanpasserines. Slate Creek Press, Bolinas, ist 79: in press. California. TUCK,L. 1960.The murres;their distribution,pop- RUSSELL,S. M., R. O. RUSSELL,AND S. WETHINGTON. ulationsand biology:A study of the genusUria. 1994.Lucifer Hummingbirds banded in south- Canadian Wildlife Service, Ottawa, Ontario. eastern Arizona. North American Bird Bander UNDERHILL,L. G., AND W. ZUCCHINI. 1988. A model 19:96-98. for avianprimary moult. Ibis 130:358-372. SALOMONSEN,E 1944. The Atlantic Alcidae: The sea- UNDERHILL, L. G., W. ZUCCHINI, AND R. W. SUMMERS. sonaland geographicvariation of the auksin- 1990. A model for avian primary moult-data habitingthe Atlantic Ocean and the adjacent wa- types based on migration strategiesand an ex- ters.G6teborgs Kungliga Vetenskaps-och Vitter- ample using the RedshankTringa totanus. Ibis hets-Samh•illesHandlingar, Sj•itte F61jden, Serie 132:118-123. B, 3(5):1-138. VAROUJEAN,D. H., S. D. SANDERS,M. R. GRAYBILL, SALOMONSEN,F. 1945. Gejrfuglen et Hundredaars AND L. SPEAR.1979. Aspectsof Common Murre minde. Aarbog for UniversitetZoologiske Mu- breedingbiology. Pacific Seabird Group Bulletin seum 1944-1945:99-100. 6(2):28. SALOMONSEN,E 1950. The birdsof Greenland.Ejnar VERWEY,J. 1922.The molt of Uria troilleuria (L.) and Munksgaard,Copenhagen, Sweden. Alca torda(L.). Ardea 11:99-116. SCOTT,P. E. 1994. Lucifer Hummingbird (Calothorax VERWEY,J. 1924.Results of a studyof the moult of lucifer).In The birds of North America,no.134 the Guillemot,Razorbill and native speciesof (A. Pooleand F. Gill, Eds.).Academy of Natural divers. Tijdschriftder NederlandscheDierkun- Sciences,Philadelphia, and American Ornithol- dige VereenigingLeiden 2(19):vii-viii. ogists'Union, Washington, D.C. WAGNER,H. O. 1955.The molt of hummingbirds. STILES,E G. 1995.Intraspecific and intergenericvari- Auk 72:286-291. ation in molt patternsof some tropical hum- WANLESS,S., AND M.P. HARRIS.1986. Time spent at mingbirds.Auk 112:118-132. the colonyby maleand femaleGuillemots Uria STORER,g. W. 1952.A comparisonof variation,be- aalgeand RazorbillsAlca torda.Bird Study 33: haviorand evolutionin the seabird generaUria 168-176. and Cepphus.University of California Publica- WARHAM,J. 1977.Wing loadings,wing shapes,and tions in Zoology 52:121-222. flightcapabilities of Procellariiformes.New Zea- STORER,R. W. 1960.Evolution in the diving birds. land Journalof Zoology4:73-83. Pages694-707 in ProceedingsXII International WARHAM,J. 1996. The behaviour, population biology July1998] Moltin CommonMurres 669

and physiologyof the petrels.Academic Press, WITHERBY,H. E, E C. R. JOURDAIN,N. E TICEHURST, New York. AND B. W. TUCKER. 1941. The handbook of Brit- WATSON,G. E. 1968.Synchronous wing andtail molt ish birds,vol. 5. Ternsto game-birds.H. E and in diving petrels.Condor 70:182-183. G. Witherby,London. WILKINSON, L. 1992. SYSTAT for windows: Statistics, version 5 ed. SYSTAT Inc., Evanston, Illinois. AssociateEditor: J. S. Marks