Population-Genetic Structure of a Philopatric, Colonially Nesting Seabird, the Short-Tailed Shearwater (Puffinus Tenuirostris)

The Auk 111(1):70-79, 1994

POPULATION-GENETIC STRUCTURE OF A PHILOPATRIC, COLONIALLY NESTING SEABIRD, THE SHORT-TAILED SHEARWATER (PUFFINUS TENUIROSTRIS)

JEREMYJ. AUSTIN, ROBERTW. G. WHITE, AND JENNIFERR. OVENDEN Departmentof Zoology,University of Tasmania,GPO Box252 C, Hobart, Tasmania 7001, Australia

ABSTl•CT.--Short-tailedShearwaters (Puffinus tenuirostris) are a numerous,colonially nesting seabird that is strongly philopatric. We applied restriction-enzymeanalysis of mitochondrial DNA (mtDNA) to 335 individuals from 11 colonies across southeastern Australia to assesspopulation-genetic structure and the amount of genetic variability in this species. Eleven 6/5.33-baseand four 4-baserestriction enzymes revealed 25 and 48 mtDNA haplotypes in two overlapping surveysof 215 individuals from seven coloniesand 231 individuals from eight colonies,respectively. A low mean sequencediversity among individuals (0.247%)and lack of spatialstructuring of mtDNA haplotypessuggest a lack of population-geneticstructure and a reduced ancestralpopulation size during glaciation, followed by a population and range expansion.Intracolony mtDNA diversitiesin three recently establishedcolonies and in one colonythat hasexperienced a recentbottleneck were comparableto mtDNA diversities within larger and older colonies.This suggeststhat, despite strictphilopatry in thosecolonies, colony founding and recovery from population reduction occursvia immigration of a large number of individuals. Received27 April 1993,accepted 11 November1993.

DIRECT(genetic) and indirect (observational) is highly polymorphic among individuals (Av- estimates of the type of genetic-population ise et al. 1987). Maternal inheritance and hap- structure within a species often can produce loid transmission result in an effective popu- conflicting results (Moum et al. 1991, Avise et lation size for the mitochondrial genome about al. 1992, Birt-Friesen et al. 1992). Evidence from one-quarter of that for nuclear genes. Mito- band returns suggeststhat many bird species chondrial genes, therefore, are more sensitive exhibit strong natal philopatry to a particular to the effects of random genetic drift, founder site or region. Based on this indirect evidence, events and bottlenecks than are nuclear genes Quinn and White (1987) suggestedthat genetic (Birky et al. 1983). Direct estimatesof the type differentiation among populationsof philopat- of genetic structure of populations, based on ric speciesmay be substantialas a consequence the geographic distribution of extant mtDNA of limited gene flow. However, indirect meth- haplotypesthat retain a phylogeneticrecord of ods are limited in terms of the number of col- population connectednessand historical de- onies (space) and the number of years (time) mographic changes, therefore, reflect past as from which data can be collected. These meth- well as contemporary gene flow. ods, therefore, can often fail to detect rare, but The Short-tailedShearwater (Puffinus tenuiro- evolutionarily important, events such as large- stris),or muttonbird, is a colonially nesting sea- scaledemographic changes and episodicor con- bird breeding on islands and headlands along tinuous, low-level dispersal that can affect the the southern coastline of Australia. Population genetic structuringof populations(Slatkin 1987, size is estimated to be 23 million breeding birds Avise et al. 1992). with numbers concentrated around Tasmania Analysisof mitochondrial DNA (mtDNA) has and the islands of BassStrait (Skira et al. 1985). proved to be an informative and useful method Although Short-tailed Shearwaters are trans- for examining population-geneticstructure in equatorialmigrants, long-term banding studies a variety of organisms,including birds (Ball et suggeststrong philopatry, with only limited na- al. 1988, Ovenden et al. 1991, Avise et al. 1992, tal dispersalto adjacentcolonies (Wooller et al. Birt-Friesen et al. 1992). Evolution of mtDNA 1990). For example, of 23 banded adults found is more rapid than for nuclear DNA (Brown et in a survey of 4,573 birds breeding on Little al. 1979)and the nucleotidesequence of mt DNA Green Island in BassStrait during 1988, 17 had 70 January1994] Geneticsof Short-tailedShearwaters 71

been banded on Little Green Island as adults or nestlings, 3 as nestlings on Great Dog Island, and 3 as surface adults on Fisher Island (Skira pers. comm.). The latter two colonies are less than 5 km from Little Green Island. In addition, approximately73,000 individuals were banded at various colonies in Victoria, New South Wales (NSW), South Australia and elsewhere in Tas- mania over 18 years, but none have been re- covered in a number of large surveyson Little Green Island or in annual surveys on Fisher Island (Serventy and Curry 1984). Populations of the Short-tailed Shearwater have experienced considerable flux within the last 15,000 years. Most current breeding sites would have been unsuitablebreeding habitat during the last ice age, when the sea level was at least 100 m lower than it is today (Milliman and Emery 1968); thus, these sites must have -45os been colonized within the last several thousand years. Prior to European settlement (ca. 1800), 140 øE 145øE 150øE Aborigines in Tasmania harvested eggs and chicksduring the breedingseason for food(Ryan Fig. 1. Locationsof breeding colonies of Short- tailed Shearwaters(approximate size in parentheses): 1981). Subsequently, shearwatershave formed (1) WhalebonePoint (3,000 burrows; Naarding 1980); the basisof a muttonbird industry with up to (2) Cape Direction (20,000burrows; Naarding 1980); one million chicks taken each year for their (3) CapeDeslacs (20,000 burrows; Naarding 1980);(4) meat, feathers and oil (Skira et al. 1985). In the Trial Harbour (15,000 burrows;Naarding 1980);(5) last 100 years many new colonies have been Little GreenIsland (150,000 burrows; Naarding 1980); established both within and outside the histor- (6) GreatDog Island (375,000pairs; Skira et al. 1985); ical range of the species,representing both a (6) Port Fairy (52,100 burrows; Harris and Norman major range, and possiblepopulation, expan- 1981); (8) Cape Woolamai (356,000 burrows; Harris sion (Davies 1959, Harris and Bode 1981). In- and Bode1981); (9) DoughboyIsland (2,000burrows; dividual colonies have become extinct, fluctu- Harris and Norman 1981);(10) Gabo Island (30,000- 40,000pairs; Reilly 1977);(11) Montague Island (15,000 ated in size, and suffered severe bottleneck pairs;Lane 1979).Dots on map inset indicatepositions events (Gilham 1962, Harris and Bode 1981). of coloniesat extremesof breedingrange. We applied restriction-enzyme analysis of mtDNA to examine the population-genetic natal origins of capturedbirds. This could pro- structure of Short-tailed Shearwaters,repre- vide valuableinformation on the relationship senting1 ! distinctnesting colonies, throughout betweenbirds during migrationand foraging, southeasternAustralia. We included one colony on wintering grounds,and during the breeding that had been subjectedto a recentpopulation season.In addition, the natal origins of found- bottleneck (Gabo Island) and several recently ing femalesof the recentlyestablished colonies establishedcolonies (Cape Direction, Cape Des- could be determined. However, although lacsand MontagueIsland). The remainingsev- breeding coloniesmay be genetically isolated en colonieswere all known or presumedto have through contemporaryphiloparry, patterns of predatedEuropean settlement. Philopatry may mtDNA differentiationmay be complicatedby be acting to maintain geneticdifferentiation be- the unstabledemographic history of this spe- tween breeding coloniesin this species,as has cies. been shown in colonially nesting Fairy Prions (Pachyptilaturtur) occupying a similar breeding range (Ovenden et al. 1991). If individual col- METHODS onies represent distinct intraspecific assem- We collected 335 shearwater chicks during the blages of mtDNA haplotypes, colony-specific southernsummer breeding seasonsfrom 1989to 1992 mtDNA markers could be used to identify the from 11 coloniesin southeasternAustralia (Fig. 1), 72 AUSTII•,W•ITE, ^I•I• OVEI•I•EI• [Auk,Vol. 111

T^BLE1. Numbers of 6/5.33-base restriction enzyme haplotypesscored from each of seven Short-tailed Shearwatercolonies. Each haplotype composed of morphdesignations for restrictionenzymes Aft II, ApaLI, Ava I, BanI, Bcl I, Bgl II, BstXI, EcoRV, Hae II, Hin dIII and Pvu II, respectively.

Locality (n) i 3 2 4 5 6 11 Whale- Cape Cape Trial Little Great Mon- bone Deslacs Direc- Harbour Green Dog tague Point tion Island Island Island Haplotype (29) (29) (30) (30) (34) (39) (22) Total a AAAAAAAAAAA 14 19 17 17 23 25 13 128 b AAAAAAAABAA 6 5 6 5 4 7 4 37 c AAAAAAAACAA 2 1 2 4 1 2 i 13 d AABAAAAABAA 1 0 2 0 2 2 2 9 e AAAAAABAAAA 1 0 2 0 0 0 0 3 f AAAAABAABAA 0 0 0 i 0 1 0 2 g AAAACAAAAAA 0 1 0 0 I 0 0 2 h AAAAAAAADAA 2 0 0 0 0 0 0 2 i AABAAABAEAA i 0 0 0 0 0 0 1 j AAAADAAABAA i 0 0 0 0 0 0 1 k AAAAAADAAAA 1 0 0 0 0 0 0 1 I AAAAAAAAFAA 0 1 0 0 0 0 0 1 m AAAAAABABAA 0 1 0 0 0 0 0 1 n AAAAAACAAAA 0 i 0 0 0 0 0 1 o AABABAABBAA 0 0 i 0 0 0 0 1 p AAABAAAAAAA 0 0 0 1 0 0 0 1 q ABAAAAAAAAA 0 0 0 1 0 0 0 1 r BABAAAAABAA 0 0 0 i 0 0 0 1 s AAAAABAACAA 0 0 0 0 1 0 0 1 t AABABAAABAA 0 0 0 0 1 0 0 1 u BAAAAAAAAAA 0 0 0 0 i 0 0 1 v AABAAACABAA 0 0 0 0 0 1 0 1 w AAAAABAAAAA 0 0 0 0 0 1 0 1 x AAADAAAAAAA 0 0 0 0 0 0 1 1 y AABCAAAABAA 0 0 0 0 0 0 1 1

comprising: 6 colonies in Tasmania at Whalebone colonies and based on the results of the 6/5.33-base Point, Bruny Island (43ø28'S,147ø14'E, n = 29), Cape enzyme survey, the mtDNA of 231 individuals from Deslacs(42ø59'S, 147ø33'E, n = 30), Cape Direction WhalebonePoint, Trial Harbour, Great Dog Island, (43ø06'S, 147ø25'E, n = 30), Trial Harbour (41ø55'S, Port Fairy, Cape Woolamai, Doughboy Island, Gabo 145ø09'E,n = 30), Little Green Island (40ø13'S,148ø15'E, Island and Montague Island were analyzedwith four n = 35) and Great Dog Island (40ø15'S,148ø14'E, n = 4-baserestriction enzymes (Hha I, Hinf I, Msp I, Taq 39); 4 colonies in Victoria at Port Fairy (38ø24'S, I). All of the birds from Whalebone Point, Trial Har- 142ø15'E, n = 30), Cape Woolamai, Phillip Island bour,Great Dog Islandand MontagueIsland analyzed (38ø34'S,145ø22'E, n = 30), Doughboy Island (38ø46'S, with 4-baserestriction enzymes were included in the 146ø17'E, n = 30) and Gabo Island (37ø34'S, 149ø55'E, 6/5.33-baseenzyme survey. Sampleswere digested n = 30); and Montague Island (36ø15'S,150ø13'E, n = with up to a 10-fold excessof enzyme, but otherwise 22) in NSW. Liver tissue was dissected from dead according to the supplier's (New England Biolabs, birds and stored in liquid nitrogen. USA) instructions.Restriction fragments produced by We extracted mtDNA from thawed liver samples 6/5.33-base and 4-baserestriction enzymeswere ra- by a modification (Brasher et al. 1992) of the method dioactively labelled with 35S-deoxycytosinetriphos- of Chapmanand Powers(unpubl. report). The mtDNA phate (Ovenden et al. 1989) and separatedby molec- of 215 individuals collected from Whalebone Point, ular weight through 1.4%agarose gels (Ovenden et Cape Deslacs, Cape Direction, Trial Harbour, Little al. 1989) and 5% polyacrylamidegels (Smolenskiet Green Island, Great Dog Island and Montague Island al. 1993), respectively. was digested with eight 6-base restriction enzymes The 6/5.33-baseand 4-baserestriction-enzyme data (Aft II, Apa LI, Bcl I, Bgl II, Bst XI, EcoRV, Hin dIII, were analyzedseparately. Unique fragmentpatterns Pvu II) and three 5.33-baserestriction enzymes (Ava identified with each enzyme were assignedan up- I, Ban I, Hae II). For sample sizes see Table I. Subse- percaseletter, the mostcommon fragment pattern, or quently, with the availability of samples from more morph, for each enzyme being designated"A." For January1994] Geneticsof Short-tailedShearwaters 73 the 6/5.33-baserestriction-fragment profiles, the loss For four enzymes an approximately 200-base- or gain of a restriction site was determined by the pair insert in two individuals comparedto all loss or gain of appropriately sized fragments. It was other mtDNA genomeswas observed.These in- difficultto determine4-base restriction-site gains and/ dividuals were omitted from subsequentanal- or lossesbetween individuals due to the large number yses. The size (œ + SE) of the shearwater ge- of sites.Subsequent analyses of the 4-basedata, therefore, were basedon restriction-fragmentpresence and nome was estimatedto be 20,768 ñ 62 basepairs absence. Each individual in the 6/5.33-base and 4-base from the sum of all fragments for a single in- enzyme surveys was assigned a composite 11- or 4- dividual representing each enzyme morph on letter code or haplotype, respectively,corresponding all gels. to the fragment patterns produced from that individ- Each restriction enzyme yielded one to six ual for each restriction enzyme. For the 6/5.33-base morphs ([1] Hin dIII and Pvu II; [2] Aft II, Apa enzyme data, nucleotide sequence divergence was LI, Ava I, Bgl II and EcoRV; [4] Ban I, Bcl I and calculatedbetween pairs of mitochondrial genomes, Bst XI; [6] Hae II), which in combination iden- with standard errors, using the maximum-likelihood tified 25 haplotypes,differing by 1 to 10 restric- method of Nei and Tajima (1983). An estimateof se- tion-site gains or losses.There were 40 mono- quence divergence between individuals was calculated from the presenceor absenceof 4-baserestric- morphic restriction sites (present in all tion-fragment data using equation 20 in Nei and Li haplotypes)and 23 polymorphic sites (occur- (1979). Net nucleotide sequence divergences and ring in somebut not all haplotypes).The three standard errors between pairs of populations were most common haplotypes (a, b and c) occurred estimated by the method of Nei and Jin (1989). A in all seven coloniesand collectively accounted Student's t-test, with Bonferroni correction for the for 84% of all individuals surveyed (Table 1). large number of pairwise comparisons(Rice 1989), The remaining 22 haplotypeswere represented was used to determine whether the net divergence by only one to nine individuals and, with one estimateswere significantly different from zero. exception(haplotype d), were shared between A chi-squaretest was used to compareobserved and two colonies(haplotypes e, f and g) or exclusive expected haplotype frequencies to detect possible to single colonies(haplotypes h-y). Parsimony population subdivisionamong colonies.The significanceof the chi-squarevalue wastested using a Mon- analysis of the substitutional relationships te-Carlomethod (Roff and Bentzen1989), as expected among the 25 haplotypes produced 675 most- class sizes were often less than five. An estimate of parsimoniousnetworks of length 31 steps. A geneticvariability due to geographicsubdivision was 50%majority-rule consensusnetwork, of length gainedusing G•T analysis (Takahata and Palumbi1985). 34 steps,revealed two groupscentered on com- We usedequations 17 and 19 of Takahataand Palumbi mon haplotypesa and d, separatedby five mu- (1985) to estimate the intrademe (I) and interdeme (J) rational steps(Fig. 2). Two of thesechanges in- identity probability using restriction-site(6/5.33-base volved a single Bgl II restriction site (i.e. a enzymedata) or restriction-fragment(4-base enzyme reversal) and a third (Ava I site 2) exhibited data) presence/absencedata from each population. The significanceof the GsTvalue was evaluatedby considerablehomoplasy in the group centered bootstrapping(Palumbi and Wilson 1990),where the on haplotype d. Most haplotypes within each true G•, value is comparedto 1,000 random Gsrvalues group were separatedby only two or three re- obtainedfrom randomrearrangements of the raw data. striction-site changes. If the true G•T value was greater than 95% of the The mean mtDNA sequencediversity among randomG• values,we concludedthat populationsub- the 213 shearwatergenomes was 0.247 ñ 0.100%, divisionwas present.Wagner-parsimony analysis of with a range of 0 to 1.850%.No variation in the the 6/5.33-baserestriction-enzyme site data was per- intrapopulational sequencediversities was de- formed using the tree-bisection-reconnectionbranch- tected between the seven shearwater colonies swappingalgorithm of version 3.1 of the computer (Table2). The net interpopulationalnucleotide program PAUP (supplied by D. L. Swofford, Illinois Natural History Survey, Champaign). sequencedivergence between birds from pairs of colonieswas zero (n = 17) or not significantly different from zero (P) 0.2; n = 4) in all 21 RESULTS possiblepairwise comparisons. Haplotype fre- quenciesfor individuals sampledfrom the sev- 6/5.33-baseenzyme survey.--The 11 restriction en colonieswere alsonot significantlydifferent enzymesidentified 63 restrictionsites among (X 2 = 41.24, P = 0.49 +_ 0.032). The amount of 215shearwater mitochondrial genomes, with 44 genetic variation among colonies due to geo- to 52 restrictionsites assayed per mtDNA clone. graphic subdivision(Gsr) was 0.284. This value 74 AUSTIN,WHITE, AND OVENDEN [Auk,Vol. 111

Bs2 • "

c Bg 3 __ Ba5

Ha4 Bg 3

Ha 3

Ha7 Av 2

Bc 3

Bg 3

2 Ha5 Av 2

Av 2 Ha 6 b

Fig. 2. Unrooted consensusnetwork describingrelationships among 25 Short-tailed ShearwatermtDNA haplotypes.Values in boxesalongside branches represent percentage of times that group abovethat branch (assuminghaplotype d representsbottom of network) occurredin 675 parsimonious31-step trees. Haplotype designationsas in Table 1. Haplotype H1 representsa hypotheticalhaplotype that was not observed.Labelled barsand arrows on lines joining haplotypesindicate direction of lossof restrictionsites. Each restriction site is denotedby first two lettersof enzymeand site numberfor that enzyme (e.g. Bg 3 = BglI site 3). Open and solid barsrepresent site changesthat were autapomorphicwith respectto haplotypesand clades,respectively. Shadedbars represent convergent synapomorphic site changes.Size of eachcircle is proportionalto frequency of that haplotype. doesnot demonstratesignificant population dif- nomes. Almost twice as many fragments (83- ferentiation asit is greater than only 46% of the 86) were identified per mtDNA clone, com- 1,000 bootstrapped G•r values (range 0.264 to pared to the 6/5.33-base enzyme survey. 0.323). We identified 48 compositehaplotypes, con- 4-base enzyme survey.--The four restriction- sisting of 6 to 15 morphs per restriction enzyme enzymes produced 117 restriction fragments ([6] Hinf I; [8] TaqI; [11] Msp I; [15] Hha I). Three from the 231 shearwater mitochondrial ge- haplotypes(1, 2 and 3) were commonto all eight January1994] Geneticsof Shorttailed Shearwaters 75

TABLE2. Percent intrapopulational mtDNA nucleotide sequencediversity (27+ SE) for 11 Short-tailed Shear- water colonies.

Locality 6/5.33-base enzyme 4-baseenzyme 1 Whale Point 0.3286 + 0.1396 0.2038 + 0.1104 4 Trial Harbour 0.2710 _+ 0.1342 0.2304 _+ 0.0980 2 Cape Direction 0.2565 _+0.1312 -- 3 Cape Deslacs 0.2142 + 0.1132 -- 6 Great Dog Island 0.2205 _+0.1215 0.2639 +_0.1004 5 Little Green Island 0.2142 _+ 0.1026 -- 7 Port Fairy -- 0.2455 + 0.1000 8 Cape Woolamai -- 0.2573 _+0.0943 9 Doughboy Island -- 0.2952 + 0.1078 10 Gabo Island -- 0.2145 _+ 0.0948 11 Montague Island 0.2581 -+ 0.1300 0.2462 _+0.1081 colonies sampled and represented 39% of all mtDNA haplotypes, with the most common individuals (Table 3). Three haplotypes(4, 5 and haplotypes present in colonies throughout the 6) were found in sevencolonies. The remaining surveyedrange. The phylogenetic network (Fig. 42 haplotypes were representedby eight or 2) revealedtwo closelyrelated groups of mtDNA fewer individuals, and occurred in fewer than haplotypes that were geographically wide- fiveof the eightcolonies. Thirty haplotypes were spread. Together, these aspectsof shearwater exclusiveto a single colony. mtDNA variability demonstratea lack of de- The mean mtDNA sequencediversity among tectable phylogenetic population structure in the 231 shearwatergenomes was 0.247 _+0.060%, the Short-tailed Shearwater. with a range of 0 to 0.890%.Mean intrapopu- This result is, in part, unexpected for a num- lational diversities were similar both among ber of reasons. First, the colonies included in colonies and with estimates derived from the this study encompassthe majority (in terms of 6/5.33-base enzyme survey (Table 2). Pairwise numbersof birds) of the species'breeding range; comparisonsof all eight colonies revealed 27 individual colonieswere separatedby distances net interpopulational mtDNA divergence esti- of up to 860 km. Second,no long-distancenatal mates that were either zero (n = 13) or not sig- or breeding dispersalhas been recorded in this nificantly different from zero (n = 14). The one speciesdespite large-scaleand long-term band- remaining comparisonbetween the Doughboy ing and recovery surveys. Over a 40-year pe- Island and Trial Harbour coloniesyielded a sig- riod, only 23 successfulnatal dispersalshave nificantly large net interpopulationalmtDNA been recorded in surveysof tens of thousands divergence of 0.0158 _+ 0.0048% (0.002 > P > of burrows;each dispersal event involved a dis- 0.001). A chi-squaretest of homogeneityof tance of less than 5 km between natal and breed- morphsfor all haplotypesoccurring more than ing colony (Skira pers. comm., Serventy and once was nonsignificant (X • 130.65, P - 0.195 Curry 1984, Wooller et al. 1990). Avise et al. + 0.025).The GsTvaluefor the eightpopulations (1992) have discussedthe problems raised by was 0.1922, being greater than only 42% of the such contrast between band-return studies and 1,000 bootstrappedGs•- values (range 0.172 to the geographicdistributions of mtDNA. They 0.209). suggested that "an understanding of population structure requires the integration of both DISCUSSION evolutionary (genetic) and contemporary (direct observational) perspectives." Third, strict The phylogeographicstructure of a species philopatry has been demonstratedby restric- comprisestwo, partly interdependent compo- tion-enzyme analysisof mtDNA in at least one nents: (1) the magnitude of mtDNA sequence colony of the Fairy Prion, a colonially nesting diversity; and (2) the degree of spatial struc- seabird occupying a breeding range similar to turing of mtDNA haplotypes (Avise 1989). that of the Short-tailed Shearwater (Ovenden Short-tailed Shearwaters across southeastern et al. 1991). Australia exhibited a low estimated mean se- The significantlylarge net interpopulational quence diversity among pairs of mtDNA ge- sequencedivergence between the Doughboy homes and no geographic structuring among Island and Trial Harbour coloniessuggests lim- 76 AUSTIN,WHITE, AND OVENDEN [Auk, Vol. 111

TABLE3. Numbers of 4-baserestriction enzyme haplotypesscored from eachof eight Short-tailedShearwater colonies. Each haplotype composedof morph designationsfor restriction enzymes Hha I, Msp I, Hinf I and Taq I, respectively.

Locality (n) 1 4 6 7 8 9 10 11 Whale- Trial Great Port Fairy Cape Dough- Gabo Montague bone Harbour Dog Woolamai Boy Island Island Point Island Island Haplotype (29) (30) (30) (30) (30) (30) (30) (22) Total 1 AAAA 11 7 3 5 10 2 5 3 46 2 ABAA 3 3 5 4 1 3 5 2 26 3 CAAA 4 3 2 2 1 2 3 1 18 4 BAAA 0 4 4 3 4 2 6 5 28 5 BBAA 1 2 2 5 2 1 3 0 16 6 DAAA 3 2 1 2 2 0 1 2 13 7 CBAA 3 0 1 0 0 2 0 2 8 8 CDAA 1 1 0 0 2 1 0 1 6 9 DBAA 0 2 0 0 1 2 1 0 6 10 EAAA 1 0 0 2 0 2 0 0 5 11 FAAA 0 0 3 0 0 1 0 1 5 12 CAAB 0 1 2 1 0 0 1 0 5 13 FBAA 0 0 0 1 0 3 1 0 5 14 CCAA 0 0 0 1 0 1 1 2 5 15 EBAA 0 1 1 0 0 0 1 0 3 16 AAAC 0 0 1 0 1 0 0 0 2 17 EAAB 0 0 1 0 0 1 0 0 2 18 CFAA 0 0 0 0 1 1 0 0 2 19 IAAA 1 0 0 0 0 0 0 0 1 20 BGAA 1 0 0 0 0 0 0 0 1 21 AADA 0 1 0 0 0 0 0 0 1 22 ABCA 0 1 0 0 0 0 0 0 1 23 GAAA 0 1 0 0 0 0 0 0 1 24 BAAE 0 1 0 0 0 0 0 0 1 25 BCAA 0 0 1 0 0 0 0 0 1 26 HDAD 0 0 1 0 0 0 0 0 1 27 JBBA 0 0 1 0 0 0 0 0 1 28 BBAF 0 0 1 0 0 0 0 0 1 29 BIAA 0 0 0 1 0 0 0 0 1 30 KAAA 0 0 0 1 0 0 0 0 1 31 BEAA 0 0 0 1 0 0 0 0 1 32 CAAC 0 0 0 1 0 0 0 0 1 33 CCAG 0 0 0 0 1 0 0 0 1 34 CAAH 0 0 0 0 1 0 0 0 1 35 DDAA 0 0 0 0 1 0 0 0 1 36 LAAC 0 0 0 0 1 0 0 0 1 37 AHAA 0 0 0 0 1 0 0 0 1 38 AJAA 0 0 0 0 0 1 0 0 1 39 AEAA 0 0 0 0 0 1 0 0 1 40 DKAA 0 0 0 0 0 1 0 0 1 41 ECEA 0 0 0 0 0 1 0 0 1 42 EAFA 0 0 0 0 0 1 0 0 1 43 MCBA 0 0 0 0 0 1 0 0 1 44 DAAB 0 0 0 0 0 0 1 0 1 45 AAAB 0 0 0 0 0 0 1 0 1 46 NAAA 0 0 0 0 0 0 0 1 1 47 ACAA 0 0 0 0 0 0 0 1 1 48 OAAA 0 0 0 0 0 0 0 1 1

ited gene flow between them. This result is dif- apparently unique charactersin terms of its size, ficult to understand since: (1) the Doughboy age or history compared to any other colony; Island colony is situated near the center of the and (3) neither colony appearsto be genetically species'range; (2) neither colony displaysany isolated from the three more geographically January1994] Geneticsof Short-tailedShearwaters 77 distant coloniesat Montague Island, Gabo Is- The Gabo Island colony suffered a severe pop- land and Whalebone Point. ulation bottleneck in 1959 (Gillham 1962) from The limited geographicstructuring among which it hasrapidly recoveredto a size of 30,000 mtDNA haplotypessuggests contemporary gene to 40,000 pairs (Reilly 1977). The genetic data flow and/or recent evolutionary connections support this hypothesis.The intrapopulational among populations (Slatkin and Maddison sequence diversities (Table 2) in the recently 1989). Indeed, the geographically widespread established colonies at Cape Direction, Cape occurrenceof both the common and rare hap- Deslacsand Montague Island, and in the Gabo lotypes, observed in more than one individual, Island colony following the bottleneck event indicates an extensive and recent intercolony are similar to the genetic diversity for the pop- exchange. ulation overall. This suggestslarge founder Birky et al. (1983) and Slatkin (1987) have populations and population recovery via a sub- shown that an exchangeof asfew asone or two stantial, and probably opportunistic, immigra- females per generation would be sufficient to tion of birds (Slatkin and Maddison 1989). Birt- counteract genetic drift •n the mitochondrial Friesen et al. (1992) have suggesteda similar genome. Assuming genetic equilibrium within scenario for philopatric, colonially nesting populations,the effectivegene flow, or number Thick-billed Murres (Uria lomvia). In contrast, a of females exchanged per generation (N•,m), recently formed Fairy Prion colony was found among colonies can be estimated as to be devoid of mtDNA restriction-site variation, suggestinga very small founder popula- N,.m = (1/Gs,•. 1)/2 (1) tion and a lack of large-scalegene flow (Oven- (Takahata and Palumbi 1985). The GsTvalues den et al. 1991). calculated from the 6/5.33-base and 4-base en- Regardlessof current levels of gene flow, the zyme data give Nem estimates of 1.4 and 2.1, limited mtDNA differentiation among colonies respectively.This suggestsa low level of gene may have occurredthrough recent evolutionary flow, which would be sufficient to maintain the interconnections. Most current breeding colo- observedgenetic homogeneityof mtDNA hap- niesmust have originated in the last 10,000years lotypes. However, due to recent fluctuations in and may have been derived from a restricted overall population and individual colony size breeding population during the Pleistocene and colony number (see below), it is unlikely glaciation. that Short-tailed Shearwater populations have The mean nucleotide sequence diversity reachedgenetic equilibrium. It is not possible, among pairs of shearwatermtDNA genomesis therefore, to distinguish between contempo- consistentwith the common pattern of low in- rary gene flow or historical associationsof col- traspecificmtDNA variability found in birds onies using this method (Slatkin and Maddison (Ball et al. 1988, Moum et al. 1991, Birt-Friesen 1989). Banding studies suggeststrong philo- et al. 1992). A low intraspecific mtDNA se- patry, although a small number of natal dis- quence diversity suggestsa reduced long-term persals to geographically very close colonies effective population size (N•.) relative to the cur- have been recorded (Wooller et al. 1990). How- rent breeding population size, asa consequence ever, no long-distancedispersal has been re- of a restricted number of female ancestors that ported. Basedon this evidenceit is unlikely that have contributed to the current gene pool (Av- a low level of intercolony movementby females ise et al. 1988). The last glacial period played a is responsiblefor the observedhomogeneity of major role in influencing mtDNA diversity in mtDNA haplotypes. avian species through population bottlenecks Gene flow among colonies,however, may in- and founder events during subsequentpopu- volve rare, large-scale movements of individ- lation expansion from a restricted number of uals (Slatkin 1985, Avise et al. 1992 ). Circum- breeding refugia (Tegelstr•m et al. 1990,Moum stantial evidence of this type of gene flow is et al. 1991, Ovenden et al. 1991, Birt-Friesen et available, given the low intrinsic population al. 1992). growth through reproduction (Wooller et al. The low mtDNA sequencediversity in shear- 1988). The Montague Island colony and others waters indicates similar fluctuations in popu- on the NSW coast were founded in the last 35 lation size, probably involving one or more years(Lane 1979)and have grown rapidly. Sim- population bottlenecksduring glaciationand a ilarly, coloniesat Cape Direction and CapeDes- recent expansion to current population levels. lacsappeared early this century (Bryden 1966). Shearwaters may have experienced a very re- 78 AUSTIN,WHITE, AND OVENDEN [Auk,Vol. 111 cent population expansion given that a number on mitochondrial DNA lineages and inbreeding of new colonies have been formed both inside theory for neutral mutations. Mol. Biol. Evol. and outside the historical breeding range (Bry- 5:331-344. den 1966, Harris and Bode 1981). BALL, R. M., S. FREEMAN,F. C. JAMES,E. BERMINGHAM, The pattern of population-genetic structure AND J. C. AVISE. 1988. Phylogeographic popu- in the Short-tailed Shearwater indicates that lation structure of Red-winged Blackbirds as- sessedby mitochondrial DNA. Proc. Natl. Acad. shearwater colonies have had extensive and Sci. U.S.A. 85:1558-1562. evolutionarily recent genetic contact.These re- BIRKY, W. C., T. MARUYAMA, AND P. FUERST. 1983. An suits do not refute observational data indicating approachto population and evolutionary genetic strong contemporary philopatry, nor do they theory for genes in mitochondria and chloro- imply present-day panmixia. Instead the mt- plasts,and some results. Genetics 103:513-527. DNA data suggest a scenario involving a pop- BIRT-FRIESEN,V. L., W. A. MONTEVECCHI, A. J. GASTON, ulation bottleneck during the last glaciation, AND W. S. DAVIDSON. 1992. Genetic structure of and subsequent population and range expan- Thick-billed Murre (Uria lornvia)populations ex- sion that has continued until present times. In amined using direct sequenceanalysis of ampli- fied DNA. Evolution 46:267-272. combination with observational data, the ge- BRASHER,D. J., J. R. OVENDEN,AND R. W. G. WHITE. netic data also suggestgene flow occurring via 1992. Mitochondrial DNA variation and phy- rare, large-scaledispersal events. In contrastto logenetic relationships of Jasusspp. (Decapoda: the sympatric species,Fairy Prion, where very Palinuridae). J. Zool. (Lond.) 227:1-16. small founding populations are implicated, es- BROWN, W. M., M. GEORGE,JR., AND A. C. WILSON. tablishment of Short-tailed Shearwater colonies 1979. Rapid evolution of animal mitochondrial and recovery from population bottlenecksap- DNA. Proc. Natl. Acad. Sci. U.S.A. 76:1967-1971. pear to involve an immigration of large num- BRYDEN, W. 1966. TasmanJan Museum and Art Gal- bers of individuals. lery, Historical Notes. Pap. Proc. R. Soc.Tasman. 100:21-26. DAVIES,S. J. J. F. 1959. A note on the shearwaters ACKNOWLEDGMENTS breeding on the Tollgates Islands, NSW. Emu 59: 287-288. We thank the following for logistical support: the GILLHAM, g. E. 1962. Granite islands of south-east Department of Parks, Wildlife and Heritage, Tasma- Victoria as a seabird habitat. Proc. R. Soc. Vic. 75: nia; Department of Conservation and Environment, 45-63. Victoria; and the NSW National Parks and Wildlife HARRIS,M.P., AND K. G. BODE. 1981. Populationsof Service. Matt Edmunds,Irynej Skira, Tim Lamb, Greg Little Penguins, Short-tailed Shearwaters and Austin, Peter Fullagar and Mike Crowley assistedwith other seabirds on Phillip Island, Victoria, 1978. collection of material. Eldredge Birmingham, Vicki Emu 81:20-28. Birt-Friesen and Bill Montevecchi provided helpful HARRIS, M.P., AND F. I. NORMAN. 1981. Distribution commentsthat improved the manuscript.The project and status of coastal colonies of seabirds in Vic- was funded by PasmincoMetals-EZ and a grant from toria. Mem. Natl. Mus. Vic. 42:89-106. the Victorian Institute of Marine Science. LANE,S.G. 1979. Summary of breeding seabirdson New South Wales coastal islands. Corella 3:7-10.

LITERATURE CITED MILLIMAN, J. D., AND K. O. EMERY. 1968. Sea levels during the past 35,000 years. Science162:1121- AVISE,J. C. 1989. Gene trees and organismal histo- 1123. ries: A phylogenetic approachto population bi- MOUM, T., K. E. ERIKSTAD, AND E. BJORKLID. 1991. ology. Evolution 43:1192-1208. Restriction fragment analysis of mitochondrial AVISE, J. C., R. T. ALISAUSKAS,W. S. NELSON,AND C. DNA in Common Murres, Uria aalge,from four DAVIDSONANKNEY. 1992. Matriarchal popula- Norwegianseabird colonies. Can. J. Zool. 69:1577- tion genetic structure in an avian specieswith 1584. female natal philopatry. Evolution 46:1084-1096. NAARDING,J. A. 1980. Study of the Short-tailed AvIsE, J. C., J. ARNOLD, R. M. BALL, E. BERMINGHAM, ShearwaterPuffinus tenuirostris in Tasmania.Rep. T. LAMB, J. E. NEIGEL, C. A. REEB,AND N. C. Tasman. Natl. Park Wildl. Serv., Hobart. SAUNDERS.1987. Intraspecific phylogeography: NEI, g., AND L. JIN. 1989. Variance of the average The mitochondrial DNA bridge between popu- numbers of nucleotide substitutions within and lation geneticsand systematics.Annu. Rev. Ecol. between populations.Mol. Biol. Evol. 6:290-300. Syst. 18:489-522. NEI, M., AND W-H, LI. 1979. Mathematical model for AVIsE,J. C., R. M. BALL,AND J. ARNOLD. 1988. Current studying genetic variation in terms of restriction versus historical population sizes in vertebrate endonucleases. Proc. Natl. Acad. Sci. U.S.A. 76: specieswith high gene flow: A comparisonbased 5269-5273. January1994] Geneticsof Short-tailedShearwaters 79

NF2, M., AND F. TAJIMA. 1983. Maximum likelihood S•RA, I. J., J. E. WAPSTRA,G. N. TOWNEY, AND J. A. estimation of the number of nucleotide substi- NAARDING. 1985. Conservation status of the tutions from restriction sites data. Genetics 105: Short-tailed Shearwater Puffinustenuirostris in 207-217. Tasmania, Australia. Biol. Conserv. 37:225-235. OVENDEN, J. R., A. J. SMOLENSKI,AND R. W. G. WHITE. SLATKIN,M. 1985. Gene flow in natural populations. 1989. Mitochondrial DNA restriction site vari- Annu. Rev. Ecol. Syst. 16:393-430. ation in Tasmanianpopulations of orangeroughy SL^TK•N,M. 1987. Gene flow and the geographic (Hoplostethusatlanticus), a deep-watermarine te- structureof natural populations.Science 236:787- leost. Aust. J. Mar. Freshwater Res. 40:1-9. 792. OVENDEN, J. R., A. WUST-SAUCY, R. BYWATER, N. SLATKIN, M., AND W. P. MADDISON. 1989. A cladistic BROTHERS,AND g. W. G. WHITE. 1991. Genetic measureof gene flow inferred from the phylog- evidence for philopatry in a colonially nesting enies of alleles. Genetics 123:603-613. seabird,the Fairy Prion (Pachyptilafurfur). Auk SMOLENSKI,A. J., J. R. OVENDEN, AND R. W. G. WHITE. 108:688-694. 1993. Evidenceof stock separationin Southern PALUMBI, S. R., AND A. C. WILSON. 1990. Mitochon- Hemisphere orange roughy (Hoplostethusatlan- drial DNA diversity in seaurchins. Evolution 44: ticus,Trachichthyidae) from restriction enzyme 403-415. analysis of mitochondrial DNA. Mar. Biol. 116: QUINN, T. W., AND B. N. WHITE. 1987. Analysis of 219-230. DNA sequencevariation. Pages163-198 in Avian TAKAHATA, N., AND S. R. PALUMBI. 1985. Extranu- genetics (F. Cooke and P. A. Buckley, Eds.). Ac- clear differentiation and gene flow in the finite ademic Press, London. island model. Genetics 109:441-457. REILLY, P. N. 1977. Seabird Islands no. 45, Gabo Is- TEGELSTR6M,H., H. P. GELTER,AND M. JAAROLA.1990. land, Victoria. Corella 1:51-53. Variation in Pied Flycatcher(Ficedula hypoleuca) RICE,W.R. 1989. Analyzing tablesof statisticaltests. mitochondrial DNA. Auk 107:730-736. Evolution 43:223-225. WOOLLER,R. D., J. S. BRADLEY,D. L. SERVENTY,AND I. ROFF, D. A., AND P. BENTZEN. 1989. The statistical J. SKIPA. 1988. Factorscontributing to repro- analysisof mitochondrial DNA polymorphisms: ductivesuccess in Short-tailedShearwaters (Puf- X2 and the problem of small samples.Mol. Biol. finustenuirostris). Pages 848-856 in Acta XIX Con- Evol. 6:539-545. gressusInternationalis Ornithologici (H. Ouellet, RY^N, L. 1981. The Aboriginal Tasmanians.Univ. Ed.). Ottawa, Ontario, 1986. National Museum of Queensland Press, St. Lucia. Natural Science, Ottawa. SERVENTY,D. L., AND P. J. CURRY. 1984. Observations WOOLLER,R. D., J. S. BRADLEY,I. J. SKIRA,AND D. L. on colony size, breeding success,recruitment and SERVENTY.1990. Reproductive successof Short- inter-colony dispersalin a Tasmaniancolony of tailed ShearwatersPuffinus tenuirostris in relation Short-tailedShearwaters Puffinus tenuirostris over to their age and breeding experience.J. Anim. a 30-yearperiod. Emu 84:71-79. Ecol. 59:161-170.