Subspecific Status and Population Genetic Structure of Least Terns (<I

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

USGS Staff -- Published Research US Geological Survey

2010

Subspecific Status and opulationP Genetic Structure of Least Terns (Sternula antillarum) Inferred by Mitochondrial DNA Control- Region Sequences and Microsatellite DNA

Hope M. Draheim Michigan State University

Mark P. Miller U.S. Geological Survey, [email protected]

Patricia Baird Simon Fraser University

Susan M. Haig U.S. Geological Survey, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/usgsstaffpub

Draheim, Hope M.; Miller, Mark P.; Baird, Patricia; and Haig, Susan M., "Subspecific Status and Population Genetic Structure of Least Terns (Sternula antillarum) Inferred by Mitochondrial DNA Control-Region Sequences and Microsatellite DNA" (2010). USGS Staff -- Published Research. 667. https://digitalcommons.unl.edu/usgsstaffpub/667

This Article is brought to you for free and open access by the US Geological Survey at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in USGS Staff -- Published Research by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. The Auk 127(4):807 819, 2010 The American Ornithologists’ Union, 2010. Printed in USA.

SUBSPECIFIC STATUS AND POPULATION GENETIC STRUCTURE OF LEAST TERNS (STERNULA ANTILLARUM) INFERRED BY MITOCHONDRIAL DNA CONTROL-REGION SEQUENCES AND MICROSATELLITE DNA

HOPE M. DRAHEIM,1,3 MARK P. M ILLER,1,4 PATRICIA BAIRD,2 AND SUSAN M. HAIG1

1U.S. Geological Survey Forest and Rangeland Ecosystem Science Center, 3200 SW Jeﬀerson Way, Corvallis, Oregon 97331, USA; and 2Centre for Wildlife Ecology, Simon Fraser University, Burnaby, British Columbia V5A 1S6, Canada

Abstract.—The taxonomic identity of endangered populations of the Least Tern Sternula( antillarum) has long been debated. Their current conservation status provides even more impetus to examine the taxonomic distinctness of these groups. We used rapidly evolving mitochondrial DNA control-region sequences ( base pairs; n ) and microsatellite DNA data ( loci; n ) to examine genetic structure within and among three subspecies that occur within the United States: California Least Tern (S. a. browni), Interior Least Tern (S. a. athalassos), and Eastern Least Tern (S. a. antillarum). Although significant genetic structure was observed among breeding populations from across the species’ range, our data indicated little evidence of genetic structure within traditional subspecific groups. Isolation-by-distance analyses, however, identified subtle patterns that may reflect sex-specific differences in dispersal behavior. Our analyses likewise demonstrated little population subdivision among subspecific groups, which raises questions regarding the taxonomic status of traditionally defined subspecies. Our findings can therefore be used to consider a reevaluation of Least Tern subspecies by the American Ornithologists’ Union’s Committee on Taxonomy and Nomenclature. We further emphasize the need for studies of range-wide breeding-site fidelity and natal philopatry to better understand interpopulation movements of individuals throughout the annual cycle. Received  November , accepted  March .

Key words: Least Tern, microsatellites, mitochondrial DNA, population structure, Sternula antillarum, subspecies.

Estatus Subespecíﬁco y Estructura Genética Poblacional de Sternula antillarum Inferidos Mediante Secuencias de la Región Control del ADN Mitocondrial y ADN Microsatelital

Resumen.—Por mucho tiempo se ha debatido la identidad taxonómica de las poblaciones en peligro de Sternula antillarum. Su estatus de conservación actual genera un incentivo aún mayor para examinar la distinción taxonómica de estos grupos. Usamos secuencias de ADN mitocondrial de la región control de rápida evolución ( pares de bases; n ) y datos de ADN microsatelital ( loci; n ) para examinar la estructura genética dentro y entre tres subespecies que se encuentran en Estados Unidos: S. a. browni, S. a. athalassos y S. a. antillarum. A pesar de que se observó estructura genética entre poblaciones reproductivas dentro del área de distribución de la especie, nuestros datos indicaron poca evidencia de estructura genética entre grupos subespecíficos tradicionales. Los análisis de aislamiento por distancia revelaron patrones que podrían reflejar diferencias sexuales en el comportamiento de dispersión. Nuestros análisis también mostraron poca subdivisión poblacional entre grupos subespecíficos, lo que pone en duda el estatus taxonómico de las subespecies definidas tradicionalmente. Nuestros resultados pueden ser usados para considerar una revaluación de las subespecies de S. antillarum por el comité de taxonomía y nomenclatura de la American Ornithologists’ Union. Además, enfatizamos la necesidad de estudios sobre la fidelidad de sitio reproductivo y filopatría natal en toda el área de distribución de la especie para entender mejor los movimientos de individuos entre poblaciones a lo largo de todo el ciclo anual.

The subspecies concept has been extensively applied within and Brown , Phillimore and Owens , Winker and Haig avian taxa since Linnaeus first introduced intraspecific classifica- ). Definitions have varied, from inclusion of geographically tions in  (American Ornithologists’ Union [AOU] ). Indeed, distinct natural populations that are not sufficiently different to ornithologists have spent considerable time and effort refining the be considered separate species (Mayr ) to more quantitative concept and debating its utility (Mayr , Amadon , Wilson definitions such as the “% rule,” which states that a population

3Present address: Department of Zoology, 203 Natural Science Building, Michigan State University, East Lansing, Michigan 48824, USA. 4Address correspondence to this author. E-mail: [email protected]

The Auk, Vol. , Number , pages  . ISSN -, electronic ISSN -.  by The American Ornithologists’ Union. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions website, http://www.ucpressjournals. com/reprintInfo.asp. DOI: ./auk.. — 807 — 808 — DRAHEIM ET AL. — AUK, VOL. 127

FIG. 1. Map showing the breeding areas of Least Terns included in our study. Breeding areas are listed in Table 1. Breeding distributions of currently recognized subspecies are highlighted. Sampling locations corresponding to the California, Interior, and Eastern subspecies are indicted by circles, triangles, and squares, respectively. can be described as a separate subspecies only if % of its indi- Regional Working Group and Southeast Regional Working Group viduals differ from a previously described subspecies (Amadon [see Acknowledgments]). Although actual numbers are unknown, ). it is thought that Least Terns were historically abundant through- Today, the debate over taxonomic definitions has widened out their range. During the th century, however, the species ex- with the passage of conservation legislation that mandates or al- perienced large population declines as a result of anthropogenic lows protection of groups below the species level (e.g., subspecies, pressures (USFWS , ; Burger ; Kruse et al. ). As evolutionarily significant units, distinct population segments, and a result, the California subspecies is listed as endangered under more; Haig et al. , Haig and D’Elia ). Thus, there can be the U.S. Endangered Species Act (USFWS ). The Interior sub- legal ramifications, depending on how these units are defined. species was not listed as a subspecies because of taxonomic uncer- These issues come to the forefront with endangered species such tainty at the time of listing; however, the USFWS designated “the as the Least Tern (Sternula antillarum), various populations of populations of Least Terns occurring in the interior of the United which are listed under the U.S. Endangered Species Act (U.S. Fish States” as endangered (USFWS ). The Eastern subspecies is and Wildlife Service [USFWS] and National Marine Fisheries Ser- state-listed as threatened or endangered in most states in which it vice ). occurs (USFWS , ). At least five Least Tern subspecies have been described on The need to clarify appropriate taxonomic units for Least the basis of morphological characteristics (Thompson et al. ). Terns led to two studies that revealed little genetic differentiation Three U.S. subspecies are recognized by the AOU that correspond among traditional Least Tern subspecies. Using  polymorphic to the eastern United States (S. a. antillarum; Lesson ), inte- allozyme loci, Thompson et al. () found no genetic differen- rior United States (S. a. athalassos; Burleigh and Lowery ), tiation between the Interior (n ) and Eastern subspecies (n and California (S. a. browni; Mearns ) (AOU ; Fig. ). The ). Subsequently, Whittier et al. () used single-strand con- taxonomic status of the two subspecies described from Mexico, formation polymorphism analyses to examine variation in the mi- S. a. mexicana (van Rossem and Hachisuka ) and S. a. staebleri tochondrial DNA (mtDNA) cytochrome-b region and two nuclear (Brodkorb ), is uncertain (García and Ceballos , Patten introns among the U.S. subspecies (Eastern, n ; Interior, n and Erickson ). ; California, n ). All genetic markers revealed low variabil- Recent population counts estimate the breeding population ity (three mtDNA haplotypes; three alleles at one nuclear intron, of Least Terns in the United States to be ~, birds (California, one allele at the second). The variable intron indicated some ge- ,; Interior, ,; Eastern, ,) (Marschalek , Lott netic differentiation between the California and Interior breed-

, and data from the Mid-Atlantic/New England/Maritimes ing areas (FST .), but the pattern was not corroborated by the OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 809

TABLE 1. Sample sizes and genetic diversity parameters a for mtDNA control region (840 bp) and seven microsatellite loci in Least Terns (Sternula antillarum). Signiﬁcant values for R2 and FS statistics (P 0.05) are followed by asterisks.

MtDNA Microsatellites

Number of

Subspecies Breeding area County and state n haplotypes (h)(P) FS R2 nA HE HO

California (S. a. browni) 20 7 0.76 0.0021 −1.584 0.118 50 3.57 (3.57) 0.464 0.474 NCA Alameda, California 10 6 0.89 0.0029 −1.363 0.174 26 3.29 (2.77) 0.457 0.472 SCA San Diego, California 10 4 0.53 0.0010 −1.345* 0.166 24 3.29 (2.80) 0.472 0.477 Interior (S. a. athalassos) 85 28 0.95 0.0046 −12.811* 0.080 185 5.43 (4.38) 0.500 0.495 NDMOR McLean, North Dakota 8 6 0.89 0.0026 −2.444* 0.141* 20 3.86 (3.25) 0.508 0.442 SDMOR Yankton, South Dakota 9 7 0.94 0.0054 −1.453 0.178 30 3.57 (3.08) 0.508 0.538 KSKSR Pottawatomie, Kansas 10 7 0.91 0.0048 −1.244 0.138 18 3.29 (3.02) 0.516 0.549 MOMSR New Madrid, Missouri 9 8 0.97 0.0034 −4.550* 0.111* 14 3.71 (3.26) 0.512 0.510 OKCR Woods, Oklahoma 10 9 0.98 0.0058 −3.696* 0.139 14 3.57 (3.23) 0.498 0.422 OKAR Tulsa, Oklahoma 9 7 0.94 0.0054 −1.453 0.175 35 4.14 (3.2) 0.497 0.473 OKRR McCurtain, Oklahoma 10 9 0.98 0.0067 −3.220* 0.145 18 3.71 (3.10) 0.458 0.521 TXINT Dallas, Texas 10 6 0.89 0.0028 −1.459 0.153 17 3.43 (2.97) 0.474 0.487 MSMSR Bolivar, Mississippi 10 7 0.87 0.0028 −2.815* 0.098* 19 3.86 (3.18) 0.506 0.506 Eastern (S. a. antillarum) 83 44 0.96 0.0057 −25.590* 0.058 182 6.14 (4.94) 0.494 0.480 ME Knox, Maine 7 4 0.81 0.0048 −1.247 0.294 21 3.71 (2.82) 0.453 0.478 MA Barnstable, Massachusetts 11 8 0.93 0.0054 −1.724 0.167 61 4.43 (3.00) 0.477 0.457 NJ Cape May, New Jersey 10 8 0.96 0.0065 −1.760 0.158 12 3.57 (3.24) 0.500 0.500 VA Accomack, Virginia 9 9 1.00 0.0048 −5.661* 0.144 10 3.29 (3.07) 0.467 0.500 GA Glenn, Georgia 8 8 1.00 0.0069 −3.497* 0.124 8 3.57 (3.57) 0.564 0.554 USVI St. Croix, Virginia 10 7 0.91 0.0060 −0.716 0.145 24 4.00 (3.32) 0.532 0.537 FLGC Bay, Florida 8 8 1.00 0.0061 −3.381* 0.160 15 3.29 (2.93) 0.464 0.411 MSGC Harrison, Mississippi 10 8 0.93 0.0052 −2.377 0.150 16 3.57 (3.10) 0.503 0.527 TXGC Brazoria, Texas 10 8 0.96 0.0058 −2.063 0.112* 15 3.71 (3.32) 0.503 0.440 a Number of individuals sampled (n), haplotype diversity (h), number of haplotypes, nucleotide diversity (P), mean number of alleles per locus (A; rareﬁed estimate accounting for different sample sizes provided in parentheses), expected heterozygosity (HE), and observed heterozygosity (HO).

mtDNA data (FST ). Sample sizes were small in both studies, but extractions as previously described (Haig et al. ). Initially, a each concluded that traditional subspecific distinctions were un- ~,-base-pair (bp) segment containing the NADH dehydrogenase resolved. Thus, we conducted rigorous sampling and applied rap- subunit  gene (ND) and control region of the mtDNA genome was idly evolving loci (mtDNA control region and microsatellites) to amplified and sequenced in  specimens by long polymerase chain more definitively assess genetic variability and population genetic reaction (PCR; GeneAmp XL PCR Kit; Roche Molecular Systems, structure in Least Terns across their North American range. Our Branchburg, New Jersey) using conserved mtDNA primers L primary objectives were to () characterize range-wide breeding- (`-TGGTCTTGTAARCCAAARANYGAAG-; Desjardins and site genetic structure and diversity patterns and () provide a com- Morais ) and H (`-CATCTTCAGTGCCATGCTTT-`; prehensive evaluation of Least Tern subspecies designations. Tarr ). Sequences were aligned with known NADH dehydrogenase subunit  gene (ND) and control-region sequences of a va- METHODS riety of tern and gull (i.e., Charadriiformes) species from GenBank to confirm that the sequence was mitochondrial and not a nuclear Sampling.—We obtained  Least Tern samples from several tis- homolog. Likewise, the transition:transversion ratio of the se- sue sources: blood from live specimens, salvaged carcasses, and quence was :, which suggests a strong transition bias as expected embryos from collected eggs. To ensure that tissues were repre- in mtDNA (Wakeley ). Our alignment was used to design new sentative of local breeding areas, sampling was limited to breed- internal primers LETE  L (`-ATACGCTCACATGCACCT-`) ing adults and young-of-year fledglings collected at the breeding and LETE  H (`-ACTGTCGTTGACGTATAACAA-`) that area before individuals moved to migration staging areas. Eight to amplified  bp of the Least Tern mtDNA control region. Prim-  samples were collected from each of  breeding areas through- ers annealed ~ bp downstream from the ` end of control region out the Least Tern’s breeding range (Fig.  and Table ). Coastal domain I and ~ bp upstream of the AC repeat at the ` end of breeding areas were defined as groups of individual samples col- domain III. For subsequent PCR reactions, a total reaction volume lected within a breeding colony or collected from multiple adja- of  ML was used with the following concentrations:  mM Tris- cent colonies. Breeding areas along interior rivers were defined HCl at pH .,  mM KCl, .% gelatin, . mM MgCl,  Mm as a group of individual samples collected within  river miles. of each dNTP, . Mm of each primer, – ng of template, and Additionally, breeding areas that occurred within the described . U AmpliTaq Gold Polymerase (Perkin Elmer, Waltham, Mas- geographic ranges of the traditional subspecies were grouped ac- sachusetts). The following parameters were used for amplifica- cordingly (Table ). tions:  min denaturation at nC, followed by  cycles of  s at DNA extraction, marker isolation, and amplification.—DNA nC, annealing at nC for  s, and elongation at nC for  min. was obtained from samples using standard phenol–chloroform A final -min period of elongation at nC followed the last cycle. 810 — DRAHEIM ET AL. — AUK, VOL. 127

The PCR amplicons were visualized on % agarose gels and subse- used to detect recent population bottlenecks within each breed- quently cleaned and concentrated by centrifugation dialysis using area (Cornuet and Luikart ). Analyses were run under ing Microcon , MW cutoff filters (Amicon Bioseparations, the two-phase model (TPM) assuming a TPM variance of  and Bedford, Massachusetts). Complete bidirectional sequences were with % of mutations corresponding to a pure stepwise muta- obtained using primers LETE  L, LETE  H, LETE  H (`- tional model. Given the number of loci examined and sample sizes CATAACTTGATTAATCCTTTCAAC-`), and LETE  L (`- within breeding areas, we note that these analyses may have lim- CTCGAATACCTCAATGAGAC-`). Sequences were generated ited power for our data set (Cornuet and Luikart ). However, using ABI Prism Big Dye Terminator Cycle Sequencing chemistry given the lack of power, we may possibly expect to hold more confi- on an ABI  DNA sequencer located in the Central Services dence in any significant result that is identified, especially if other Laboratory at Oregon State University. Sequences were aligned analyses can possibly corroborate inferences made in these analy- using BIOEDIT, version .. (Hall ), and archived in GenBank ses. Therefore, we used BOTTLENECK to perform an indepen- (accession nos. EU–EU). In total,  specimens were dent test that screened for skewed allele-frequency distributions used in mtDNA analyses (Table ). in each breeding area, which also provides heuristic evidence of We used seven variable microsatellite loci in our analyses: the effects of prior bottleneck events (Luikart et al. ). Hbau (PCR annealing temperature TA nC; J. R. Gust et al. We quantified and tested for genetic structure and differ- unpubl. data); K, K (TA nC; Tirard et al. ); RBG entiation among breeding areas using the maximum-likelihood (TA nC); RBG, RBG (TA nC; Given et al. ); and estimator of D described in Jost (). Most conventional F- SDAAT  (TA nC; Szczys et al. ). DNA was amplified statistics variants have upper bounds that are constrained by un- using a PCR profile with the following steps: initial denaturation derlying levels of genetic diversity (Hedrick ). These issues may for  min at nC, followed by  cycles of  s at nC,  s at the make comparisons across marker types or study systems problem- specified annealing temperature,  s at nC, then an additional atic, because higher-diversity loci will generate lower overall FST -min extension step at nC. Ten-microliter reactions were pre- values—even in cases where populations are completely differen- pared using – ng of DNA in  mM Tris-HCl;  mM KCl; tiated. D does not suffer from these issues (Jost ) and consis-

. mM MgCl; . mM of each dNTP;  Mm of each primer; and tently represents differentiation of samples as values that fall along . units of Taq polymerase (Promega, Madison, Wisconsin). Am- the continuous interval from zero to unity. Thus, global and pair- plified products were sized on an ABI  automated DNA se- wise estimates of D among all  breeding areas were obtained quencer at Oregon State University’s Central Services Laboratory. separately for microsatellite (Dmic) and mtDNA data (Dmit). Com- Genotype analysis was performed using the software applications parable global differentiation measures were likewise obtained for GENESCAN ANALYSIS, version ., and GENOTYPER, version breeding areas within each traditional subspecies. For the micro-

. (Applied Biosystems, Carlsbad, California). A total of  indi- satellite data, multilocus estimates of Dmic were constructed using viduals were used for microsatellite analyses (Table ). an approximation to the harmonic mean of locus-specific values, Range-wide genetic structure and diversity.—We used ARLE- calculated as QUIN, version . (Excoffier et al. ), to calculate haplotype 23 diversity (h), the probability that two randomly chosen individuals DDDmic 11/[( / A )S D ( 1 / A ) ] have different haplotypes; and nucleotide diversity (P), the average  pairwise nucleotide differences for control-region haplotypes at where DA and S D are the arithmetic mean and variance, respec- each breeding area of the traditional subspecies. Fu’s FS (Fu ) tively, of the locus-specific D values (A. Chao pers. comm.). P val- and the R statistic of Ramos-Onsins and Rozas () were also ues for single-locus and multilocus D values were obtained through used to identify the signal of historical population expansions. FS a randomization procedure based on , randomization repli- and R have been identified as having the greatest power for iden- cates. In each replicate, individuals (and their respective genotypes tifying these patterns (Ramos-Onsins and Rozas ). ARLE- or mtDNA haplotypes) were randomly allocated to breeding areas

QUIN was used for FS calculations, whereas DNASP, version . while keeping the sample sizes of breeding areas coincident with (Rozas et al. ), was used to calculate R. In both cases, the sig- the original data. P values were ultimately obtained as the propor- niﬁcance of observed values was inferred through the use of , tion of randomized data sets producing values of D as large as or coalescent-based simulations. larger than the original D values. Comparable global estimates of D The program GDA, version . (Lewis and Zaykin ), was were also obtained for each traditional subspecies. All calculations used to quantify microsatellite genetic diversity in each breeding for D were performed using a short computer program written by area and within each traditional subspecies grouping using mean M.P.M. Mantel tests (Mantel ) based on , randomization number of alleles (A), observed heterozygosity (HO), and expected replicates were used to quantify correlations between pairwise heterozygosity (HE) for each locus and over all loci. The program Dmic and Dmit values of breeding areas using the program NT-SYS, HP-RARE (Kalinowski ) was used to obtain rareﬁed esti- version . (Exeter, Setauket, New York). mates of allelic diversity within these units to better account for Mantel tests were also used to identify isolation-by-distance sample-size variation. GDA was also used to identify deviations patterns by assessing the correlation between D values and the from Hardy-Weinberg proportions and to test for linkage disequi- logarithm of geographic distances between breeding areas. These librium between pairs of loci within each breeding area. In Hardy- analyses were performed separately for the mtDNA and micro- Weinberg tests, P values over loci were combined and evaluated satellite data and were likewise also performed () across the full using the Z-transform test (Whitlock ) to obtain a composite range of Least Terns in the United States and () separately within result for each breeding area. The program BOTTLENECK was the Eastern and Interior subspecies groups. Mantel tests were not OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 811

possible within the California subspecies group because of the small number of breeding sites (n ) examined. Genetic diﬀerences among traditional subspecies.—We used

three approaches to examine differentiation among the three (continued) 1 1 traditionally defined Least Tern subspecies. First, relationships 3 2 1 among Least Tern control-region haplotypes were inferred by estimating a statistical haplotype network with % parsimoni- 1 1 2 ous connections using the program TCS, version . (Clement 1 et al. ). If traditional subspecies are valid, we expected to observe strong associations between haplotype lineages and sets of 31 birds collected within the ranges of the three traditional subspecies groups. Second, we used STRUCTURE, version . (Pritchard et al. ), in conjunction with our microsatellite data to infer S. a. antillarum the number of Least Tern genetic clusters (K). We performed  independent runs for each value of K – using * iterations after a burn-in period of * steps. Analyses were performed us- 1 1 ; locations of breeding areas are given in Table 1). of breeding; locations Table areas in are given ing the correlated-allele-frequencies model and admixture model implemented in the program. The most likely number of clusters was determined by identifying values of K that produced the highest average log likelihood values. If genetic structure was largely 3 1 1 1 1 2 2 congruent with traditional subspecies definitions, then we ex-

pected to observe the highest likelihood values for the K  case Sternula antillarum 1 1 and likewise expected to see the majority of individuals from each 1 of the three subspecies assigned to cohesive genetic clusters. Fi- nally, global and pairwise estimates of Jost’s D were calculated 3 4 as described above using traditionally defined subspecies as the 1 operational unit of interest. The significance of these values was determined using , randomization replicates. As with our 2 1 1 other analyses, we expected our results to reflect levels of diver- 1 gence that were consistent with low gene flow among groups. 1 1 1 RESULTS 1

Genetic structure and diversity.—Sixty-seven haplotypes were observed among the  sampled individuals,  of which were 2 shared among traditional subspecies (Table ). Thirty-six haplotypes were observed only once, reﬂecting a high underlying level S. a. athalassos 1 1 1 1 1 1 of mtDNA diversity. Control-region sequences ( bp) were 1 characterized by  polymorphic sites, and no insertions or dele- tions were present. Observed nucleotide composition (A, .%; C, .%; T, .%; G, .%) was similar to that of other char- 1 1 1 2 1 1 2 1 3 adriiform species (Wenink et al. , Buehler and Baker , Funk et al. ). Within-breeding-area haplotype diversity was high (mean o SD . o .) and ranged from . (SCA) to . 1 1 2 2 (FLGC, GA, and VA) (Table ). 1 Tests for population expansions revealed signiﬁcant (at the

A . level) negative values of FS in  of the  breeding areas 1 1 1 1 11 1 examined, whereas significant R statistics were observed at four 1 breeding areas (Table ). In three of the four cases involving significant R values, the corresponding values of FS were also significant. Across regions, the signal of population expansions was 1 1 1 identified within the Interior and Eastern groups, but only from tests using FS as an indicator (Table ). 11 3 2 3 1 7 2

The total number of microsatellite alleles per locus ranged S. a. browni from  (locus K) to  (locus K). Genetic diversity was similar among breeding areas (Table ), with the mean number of alleles per locus ranging from . (NCA, SCA, VA, and FLGC) to 2. and frequencymtDNA control-region Distribution haplotypes ( breeding areas of 67 among of Least all Terns

. (MA). Rareﬁed estimates of allelic richness were also similar ABLE H18 H19 H15 H16 H17 H7 H8 H14 H11 H9 H10 H5 H6 H13 H12 H20 H3 H4 H1 H2 T Haplotype SCA NCA NDMOR SDMOR KSKSR MOMSR OKAR OKCR OKRR TXINT MSMSR ME MA NJ VA GA USVI FLGC MSGC TXGC 812 — DRAHEIM ET AL. — AUK, VOL. 127 1 1 1 1 1 1 1 1 1 11 1 1 1 1 1 1 1 1 1 1 S. a. antillarum 11 1 1 1 1 1 1 2 11 1 1 11 31 12 1 11 1 12 1 1 1 1 2 1 12 1 11 1 2 111 1 2 1 1 S. a. athalassos 1 1 1 S. a. browni 2. Continued. ABLE H21 H24 H25 H26 H27 H28 H23 H22 H46 Haplotype SCA NCA NDMOR SDMOR KSKSR MOMSR OKAR OKCR OKRRH29 TXINTH30 MSMSR ME MA NJ VAH35 GA USVI FLGC MSGC TXGC H47 H48 H49 H50 H51 H31 H32 H33 H34 H52 H53 H54 H55 H56 H57 H58 H59 H60 H36 H42 H43 H44 H45 H61 H62 H63 H64 H65 H66 H67 Total 10 10 8 9 10 9 9 10 10 10 10 7 11 10 9 8 10 8 10 10 H37 H38 H39 H40 H41 T OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 813

TABLE 3. Indicators of genetic differentiation (D) and associated P values for different hierarchical levels and subsets of the genetic data set for Least Terns. The number of signiﬁcant locus-speciﬁc values of D (out of 7 total) observed for the microsatellite data is also provided.

Microsatellite data Mitochondrial data Number of signiﬁcant loci at A 0.05 level DPDP

All breeding areas 4 0.038 0.001 0.835 0.001 Interior breeding areas 0 0.026 0.090 0.730 0.038 Eastern breeding area 1 0.027 0.342 0.767 0.101 California breeding areas 0 0.008 0.581 0.472 0.063 All three subspecies 6 0.023 0.001 0.831 0.001 California vs. Interior 5 0.020 0.001 0.873 0.001 California vs. Eastern 5 0.018 0.001 0.859 0.001 Interior vs. Eastern 4 0.021 0.001 0.620 0.001

a Number of signiﬁcant loci at A 0.05 level.

among breeding areas (range: .–.). Average observed and ex- P .). The Eastern subspecies demonstrated similar trends pected heterozygosity within breeding areas ranged from . (Dmic ., P .; Dmit ., P .). Within the Inte- and . to . and ., respectively (Table ). According to rior subspecies, the microsatellite data also revealed no significant our combined analyses over loci, no significant deviations from structure (Dmic ., P .). However, Dmit for the Interior Hardy-Weinberg genotypic proportions were observed within any subspecies, though numerically similar to that observed among breeding area (P .). Fifteen significant tests for linkage dis- Eastern breeding areas, was significant at the A . level (Dmit equilibrium were observed at the A . level among the  ., P .). tests performed (.%;  locus combinations per breeding area r Mantel tests designed to identify correlations of genetic and  breeding areas  total tests), a result that could have been geographic distances between pairs of breeding sites indicated observed by chance alone. These significant tests were also evenly that significant spatial genetic structure exists within Least Terns distributed among locus pairs and populations, which further (Table ). Across all  breeding areas investigated, significant indicated the absence of linkage disequilibrium. The program correlations between genetic and geographic distances were ob- BOTTLENECK detected a significant excess of heterozygosity served for microsatellite data and mtDNA sequence data (Table ). within the SDMOR and KSKSR breeding areas (P . and However, different results were obtained when we analyzed sub- P ., respectively). However, these analysis results may be sets of breeding areas that encompassed the Interior and Eastern a chance outcome from multiple tests, given that samples from groups. In this case, microsatellite analyses identified isolation- both locations (and all other breeding areas) also demonstrated by-distance patterns within regions (Table ). However, mtDNA normal L-shaped allele frequency distributions typical of non- analyses identified no significant spatial structure (Table ). bottlenecked populations (Luikart et al ). Furthermore, Analyses of traditional subspecies.—Although  control- P values from both of these tests were nonsignificant after se- region haplotypes were restricted to single traditional subspecies quential Bonferroni corrections. (Table ), the mtDNA haplotype network revealed no definitive Our analyses of population structure identified multiple associations between haplotype lineages, geography, or tradi- trends. Numerical values of Dmit were generally an order of mag- tional subspecies definitions (Fig. ). Some phylogenetic clustering nitude larger (or more) than comparable values of Dmic (Table  of haplotypes was observed within the California traditional sub- and Appendix), which reflects the overall high observed haplo- species (Fig. ), but these haplotypes were also generally shared type diversity and low degree of allele sharing between groups with the other two traditional subspecies groups. (Table ). Nonetheless, similar general patterns were observed be- According to STRUCTURE analyses of the microsatellite tween the microsatellite and mtDNA sequence data sets. For ex- data, the highest average log-likelihood value (−,.) was ample, the global matrices of Dmic and Dmit calculated between all pairwise combinations of breeding areas showed loose, but sig- TABLE nificant, correlations with one another (r ., P .; Ap- 4. Results of Mantel tests that assessed the significance of correlations between pairwise values of D and the logarithm of geographic dis- pendix). Likewise, with one exception, the significance of Dmit tances between breeding areas of Least Terns. and Dmic calculated for different partitions and hierarchical levels in the data set were similar (Table ). Specifically, both data Microsatellite Mitochondrial sets indicated significant genetic structure among the  breeding areas investigated (Dmit ., P .; Dmic ., P rP rP .). However, analyses based solely on each traditional subspecies provided slightly different results. No significant differentia- All breeding areas 0.385 0.001 0.394 0.002 tion was observed between the two breeding areas representing Interior breeding areas only 0.517 0.011 −0.190 0.845 Eastern breeding area only 0.368 0.049 −0.007 0.471 the California subspecies (Dmic ., P .; Dmit ., 814 — DRAHEIM ET AL. — AUK, VOL. 127

FIG. 2. The statistical 95% parsimony network generated by the program TCS, based on mtDNA control-region haplotypes of Least Terns. Circle sizes are proportional to the number of individuals that share the haplotype (frequencies of each haplotype are given in Table 2). Shading refers to the pro- portion of samples that came from a traditional subspecies designation: California Least Tern haplotypes are shown in white, Interior Least Tern haplotypes in gray, and Eastern Least Tern haplotypes in black. Dashes represent inferred haplotypes. observed for K  (Fig. ), which suggests that there was no strong comparisons between subspecies groups (Table ). However, the clustering of individuals into traditionally defined subspecies magnitude of differentiation observed at this level was largely groups. We further note that for the K  and K  cases, assign- comparable to that observed among breeding areas within each ment probabilities of individuals to clusters were generally on the traditional subspecies group (Table ). Differences in P values as- order of . (for K ) or . (for K ) and also showed no associated with Dmit and Dmic at these two hierarchical levels may sociations with traditional subspecies groups. By contrast, how- reflect the larger sample sizes (and associated power of tests) for ever, global indicators of subspecies differentiation were highly subspecies-level comparisons in relation to comparisons of breed- significant (Dmit ., P .; Dmic ., P .) when ing areas within traditional subspecies groups. traditional subspecies groups were used as the operational unit of interest (Table ). Comparable patterns were identified in pairwise DISCUSSION

Least Tern genetic diversity and structure.—An important concern with regard to many endangered species is the loss of genetic diversity that results from population declines. Superficially, our analyses of Least Tern mtDNA and microsatellite data did not suggest that genetic diversity was low within the species. For ex- ample, our analyses revealed the presence of a large complement of mtDNA haplotypes (Table ). Furthermore, though direct comparisons among studies can prove difficult when different loci are examined, measures of genetic diversity in Least Terns appear to exceed those observed in other tern species. Our microsatellite analyses (Table ) revealed average numbers of alleles per locus, observed heterozygosities, and expected heterozygosities that were on par with or exceeded those found in colonies of Common Terns (S. hirundo; Sruoga et al. ). Also, the nucleotide diver- FIG. 3. Results of analyses of 417 Least Terns using the program STRUC- sity and total number of mtDNA control-region haplotypes in our TURE. Analyses suggested that a single cluster (K 1) was most likely, analyses exceeded those previously noted in Sooty Terns (Ony- given data from the seven microsatellite loci examined. choprion fuscatus Sterna fuscata; Peck and Congdon ). OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 815

Our more formal analyses of microsatellite data provided no microsatellite data were examined (Table ). The microsatellites conclusive evidence of pervasive past bottleneck events. Instead, revealed slight but significant correlations between geographic the mtDNA data more frequently identified the signal of popula- distances and Dmic within traditional subspecies; however, no tion expansions (Table ). Of the  separate bottleneck analyses comparable patterns existed for the mtDNA data. Because of its performed within breeding areas, significant results were obtained smaller effective population size, analyses of the mitochondrial in only two cases from South Dakota and Kansas (SDMOR and genome may produce different patterns than those observed with KSKSR). These results are difficult to discern from random expec- nuclear markers solely because of differences in the effects of ge- tations and, moreover, contradict the allele frequency distributions netic drift. However, the different patterns may also indicate differ- at these sites that provided no evidence for prior bottlenecks. We ences in dispersal tendencies between males and females (Chappel note, however, that SDMOR and KSKSR are geographically proxi- et al. , Miller et al. ). Because of their maternal inheri- mate to one another (Fig. ). Thus, if this pattern is not coinciden- tance, genetic structure observed in mtDNA reflects female behav- tal, our results may actually indicate that population bottlenecks ior patterns. By contrast, because they are biparentally inherited, have occurred in a small part of the Least Tern’s geographic range. genetic structure at microsatellite loci reflects the joint behavior Recent population surveys have indicated moderate increases in of both sexes. Thus, the contrasting isolation-by-distance patterns population abundance since the mid-s and mid-s (for observed in our analyses of microsatellites and mtDNA may indi- SDMOR and KSKSR, respectively; Lott ), which may reflect cate that male dispersal is limited compared with that of females. population increases following such bottleneck events. Additional Distinguishing between these scenarios is difficult given our cur- detailed investigations within this region may be required to more rent data. Quantitative estimates of differentiation within each conclusively establish this pattern. Because bottlenecks are gener- traditional subspecies were higher for mtDNA than for microsatel- ally detectable for only a few generations following the population- lites, but significance levels were not appreciably different between reduction event (Cornuet and Luikart ), such analyses should marker types (Table  and Appendix). Sample-size limitations be performed in the near future if they are deemed to be worth associated with mtDNA analyses prevented us from determin- pursuing for conservation and management purposes. ing whether this was an artifact associated with limited sampling Our analyses indentified significant genetic structure among from a highly diverse, nondifferentiated group of populations, or breeding areas. However, this pattern was primarily observed when whether the limited sampling provided insufficient power to detect we examined global differentiation by treating all  breeding true population differentiation. We note, however, that the mini- areas as operational units of interest (Table ). Significant isolation- mal structure observed with mtDNA may, in this case, highlight by-distance patterns were likewise noted across all breeding female dispersal. Although sample sizes were also limited, previ- areas examined (Table ). As a general rule, the strength of genetic ous genetic analyses of Least Terns have arrived at similar conclu- structure will increase as the degree of natal and breeding-site fi- sions (Whittier et al. ). Because breeding pairs form on the delity increases within a species (i.e., minimizing gene-flow rates). nesting grounds (Thompson et al. ), the isolation-by-distance Empirical field observations, however, indicate that nesting-site pattern observed only at microsatellite loci may therefore point to fidelity is variable in Least Terns. Using estimates based on band- increased natal nesting-site fidelity of males in relation to females ing and resight methods, natal philopatry ranges from % to %, and corroborate previously documented patterns among the Lari- whereas breeding-site fidelity ranges from % to % (Atwood dae in general (Greenwood and Harvey ). and Massey , Massey and Fancher , Boyd , Renken Least Tern subspecies.—Our analyses of mtDNA and micro- and Smith ). These widely differing results among studies may satellite data did not provide conclusive support for the three tra- depend on behavioral differences attributable to landscape type ditional subspecies of Least Terns and reiterated results from two (i.e., coastal vs. interior rivers; Renken and Smith ) as well as previous studies of the species (Thompson et al. , Whittier et the extent of banding and resight efforts. Traditionally, disper- al. ). Although the number of haplotypes restricted to tradi- sal studies have focused on smaller scales (i.e., between colonies tional subspecies was high (), it is important to recognize that  within traditional subspecies; Boyd , Johnson and Castrale of the  total detected haplotypes were observed only once and , Lingle ). However, records exist of one juvenile banded are therefore uninformative with respect to determining the de- at its natal site on the Gulf Coast of Texas (in the range of the East- gree of allele sharing between or among groups. Nonetheless, we ern traditional subspecies) that was later found nesting in Kansas note that –% of individuals within each traditional subspecies (Boyd and Thompson ). Long-distance movements of individ- shared haplotypes with individuals originating from another sub- uals to new nesting locations (on the order of –, km) have species (Table ). Consequently, mtDNA control-region sequences also been reported (Renken and Smith ). Collectively, these do not appear to provide support for traditional Least Tern subspe- types of findings may explain why little evidence of genetic struc- cies designations (Amadon , Patten and Unitt ). ture was observed within each traditional subspecies (Table ). In- Our STRUCTURE analyses also indicated little support for deed, of the six analyses performed within a traditional subspecies, the existence of different subspecies, in that our results illustrated only one identified a significant pattern (the analysis of mtDNA that the K  solution was most likely for the Least Tern microsat- data within the Interior breeding areas; Table ). We note, how- ellite data set (Fig. ). Superficially, this pattern conflicts with re- ever, that this specific finding may actually be spurious, because sults of analyses based on Dmic, which detected low (but significant) only a single pairwise contrast involving mtDNA from Interior differentiation among traditionally defined subspecies units (Table breeding areas was significant at the A . level (Appendix). ). This discrepancy may be attributable to STRUCTURE’s inabil- Within traditional subspecies, our analyses of isolation-by- ity to detect weak genetic structure or isolation-by-distance pat- distance patterns produced different results when mtDNA and terns (see sections . and . of STRUCTURE’s documentation; 816 — DRAHEIM ET AL. — AUK, VOL. 127

Latch et al. , Schwartz and McKelvey ). In our analyses, our findings emphasize the need for range-wide information on both of these factors may be relevant. In all comparisons between breeding-site fidelity and natal philopatry as well as an under- or among subspecies, values of Dmic among subspecies were ex- standing of population-specific movements throughout the an- tremely low and on par with values observed within traditional nual cycle in order to best plan for their future success. subspecies units (Table ). The significance of values associated with subspecies comparisons is most likely due to the larger sample ACKNOWLEDGMENTS sizes of operational units at this level than in analyses of breeding areas within each traditional subspecies. Furthermore, our Mantel We thank C. Lott, R. Hurt, R. Boyd, M. McCollough, K. Stubbs, tests suggested that the overarching structural pattern may reflect K. Kughen, T. Pover, J. Castrale, J. Boylan, R. Vanderlee, C. Kruse, isolation-by-distance patterns of breeding areas across the Least P. Glass, H. Nass, R. Renken, K. Jones, S. Dinsmore, G. Pavelka, R. Tern’s range (Table ). Thus, the observed significant differences Boettcher, B. Winn, and P. Kelly for assistance and coordination among traditional subspecies may merely reflect the relatively large of sample collection. We also thank T. Mullins, R. Bellinger, M. geographic distances between breeding areas found within the dif- Schwartz, C. Funk, A. Liston, and K. Dugger for insightful discus- ferent traditionally defined subspecies (Fig. ). sion and helpful comments on the analysis and manuscript. For In addition to molecular data, morphological, behavioral, information on the Mid-Atlantic/New England/Maritimes Regional and geographic range information can also be used to determine Working Group, see http://www.waterbirdconservation.org/manem. whether a subspecies is “diagnosably distinct” (Mayr and Ashlock html. For information on the Southeast Regional Working Group, , Winker and Haig ). However, previous studies that ex- see http://www.waterbirdconservation.org/southeast_us.html. Sup- amined factors such as vocalizations, behavior, and morphological port for this project was generously provided by the U.S. Geological characteristics in Least Terns found little support for differences Survey Forest and Rangeland Ecosystem Science Center and the between traditional subspecies and concluded that any distinc- U.S. Fish and Wildlife Service. Any use of trade, product, or firm tions were arbitrary or clinal (Burleigh and Lowery , Massey names in this publication is for descriptive purposes only and does , Thompson et al. ). One morphological study based on not imply endorsement by the U.S. Government. spectrophotometric analysis of feathers nominally provided vali- dation for the three traditional subspecies (Johnson et al. ). LITERATURE CITED If we assume that plumage differences are genetically associated, then spectrophotometric analysis may be used as genetic sup- Amadon, D. . The seventy-five percent rule for subspecies. Con- port of traditional Least Tern subspecies. However, Whittier et al. dor :–. () suggested that plumage differences may be related to other American Ornithologists’ Union. . Check-List of North factors because the eumelanin that forms gray hues in Least Tern American Birds, th ed. American Ornithologists’ Union, Wash- feathers can be influenced by environment or food sources (Welty ington, D.C. and Baptista ). Atwood, J. L., and B. W. Massey. . Site fidelity of Least Terns Conservation implications.—Vignieri et al. () argued that in California. Condor :–. no single approach should be used as a “taxonomic litmus test” for Boyd, R. L. . Site tenacity, philopatry, longevity, and population taxa of concern. However, we would predict that “subspecies” that trends of Least Terns in Kansas and northwestern Oklahoma. represent unique evolutionary entities should demonstrate con- Pages – in Proceedings, Missouri River and Its Tributaries: gruent evidence of evolutionary distinctiveness. On the basis of Piping Plover and Least Tern Symposium (K. F. Higgins and our analyses, we cannot conclusively validate the traditional sub- M. R. Brashier, Eds.). South Dakota State University, Brookings. species designations within Least Terns using our neutral mtDNA Boyd, R. L., and B. C. Thompson. . Evidence for reproductive control-region or microsatellite data. Our findings can be used to mixing of Least Tern populations. Journal of Field Ornithology consider a reevaluation of Least Tern subspecies by the AOU Com- :–. mittee on Taxonomy and Nomenclature. California, Interior, and Brodkorb, P. . New birds from southern Mexico. Auk :– Eastern Least Terns appear to exhibit high genetic connectivity . among groups. However, genetic connectivity and demographic Buehler, D. M., and A. J. Baker.  Characterization of the Red connectivity are not necessarily synonymous, because only a few Knot (Calidris canutus) mitochondrial control region. Genome migrants in each generation are needed to genetically homogenize :–. disparate breeding populations, whereas the same level of move- Burger, J. . Colony stability in Least Terns. Condor :–. ment cannot maintain demographically stable populations or per- Burleigh, T. D., and G. H. Lowery. . An inland race of Sterna mit recolonization of an extinct population (Wright , ; albifrons. Occasional Papers of the Museum of Zoology, Louisi- Mills and Allendorf ). ana State University :–. Molecular tools have a demonstrated ability to identify evo- Chappell, D. E., R. A. Van Den Bussche, J. Krizan, and B. lutionarily divergent lineages. However, most studies, includ- Patterson. . Contrasting levels of genetic differentiation ing ours, sample only a small part of the genome. Thus, neutral among populations of wolves (Gulo gulo) from northern Canada mtDNA control-region and microsatellite loci are not likely to revealed by nuclear and mitochondrial loci. Conservation Genet- reflect adaptive variation that may be relevant in different envi- ics :–. ronments or for different life histories (McKay and Latta ). Clement, M., D. Posada, and K. A. Crandall. . TCS: A Although California, Interior, and Eastern Least Terns may con- computer program to estimate gene genealogies. Molecular Ecol- tinue to function as demographically independent populations, ogy :–. OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 817

Cornuet, J. M., and G. Luikart. . Description and power Latch, E., G. Dharmarajan, J. C. Glaubitz, and O. E. Rhodes, analysis of two tests for detecting recent population bottlenecks Jr. . Relative performance of Bayesian clustering software from allele frequency data. Genetics :–. for inferring population substructure and individual assignment Desjardins, P., and R. Morais. . Sequence and gene organiza- at low levels of population differentiation. Conservation Genetics tion of the chicken mitochondrial genome. A novel gene order in :–. higher vertebrates. Journal of Molecular Biology :–. Lesson, R. P. . Histoire Naturelle des Colibris, Suive d’un Sup- Excoffier, L., G. Laval, and S. Schneider. . ARLEQUIN plement a L’histoire Naturelles des Oiseaux-mouches. Bertrand, (version .): An integrated software package for population genet- Paris. ics data analysis. Evolutionary Bioinformatics Online :–. Lewis, P. O., and D. Zaykin. . Genetic Data Analysis: Com- Fu, Y. X. . Statistical tests of neutrality of mutations against pop- puter program for the analysis of allelic data. Version .. Free ulation growth, hitchhiking and background selection. Genetics program distributed by the authors. [Online.] Available at http:// :– hydrodictyon.eeb.uconn.edu/people/plewis/software.php. Funk, C. W., T. D. Mullins, and S. M. Haig. . Conservation Lingle, G. R. . Site fidelity and movements of Least Terns and genetics of Snowy Plovers (Charadrius alexandrinus) in the West- Piping Plovers along the Platte River, Nebraska. Pages – ern Hemisphere: Population genetic structure and delineation of in Proceedings, Missouri River and Its Tributaries: Piping Plo- subspecies. Conservation Genetics :–. ver and Least Tern Symposium (K. F. Higgins and M. R. Brashier, García, A., and G. Ceballos. . Reproduction and breeding Eds.). South Dakota State University, Brookings. success of California Least Terns in Jalisco, Mexico. Condor : Lott, C. A. . Distribution and abundance of the interior popu- –. lation of the Least Tern (Sterna antillarum), : A review of the Given, A. D., J. A. Mills, and A. J. Baker. . Isolation of poly- first complete range-wide survey in the context of historic and morphic microsatellite loci from the Red-billed Gull (Larus ongoing monitoring efforts. Dredging Operations and Environ- novaehollandiae scopulinus) and amplification in related species. mental Research Program Technical Notes Collection, U.S. Army Molecular Ecology Notes :–. Corps Engineer Research and Development Center, Vicksburg, Greenwood, P. J., and P. H. Harvey. . The natal and breed- Mississippi. ing dispersal of birds. Annual Review of Ecology and Systematics Luikart, G., F. W. Allendorf, J.-M. Cornuet, and W. B. Sher- :–. win. . Distortion of allele frequency distributions provides Haig, S. M., E. A. Beever, S. M. Chambers, H. M. Draheim, B. D. a test for recent population bottlenecks. Journal of Heredity : Dugger, S. Dunham, E. Elliott-Smith, J. B. Fontaine, –. D. C. Kesler, B. J. Knaus, and others. . Taxonomic Mantel, N. . The detection of disease clustering and a general- considerations in listing subspecies under the U.S. Endangered ized regression approach. Cancer Research :–. Species Act. Conservation Biology :–. Marschalek, D. A. . California Least Tern breeding sur- Haig, S. M., and J. D’Elia. . Avian subspecies and the U.S. vey,  season. Species Conservation and Recovery Program Endangered Species Act. Pages – in Avian Subspecies Report, -. California Department of Fish and Game, Habi- (K. Winker and S. M. Haig, Eds.). Ornithological Monographs, tat Conservation and Planning Branch, Sacramento, California. no. . Massey, B. W. . Vocal differences between American Least Haig, S. M., E. D. Forsman, and T. D. Mullins. . Subspecies Terns and the European Little Tern. Auk :–. relationships and genetic structure in the Spotted Owl. Conser- Massey, B. W., and J. M. Fancher. . Renesting by California vation Genetics :–. Least Terns. Journal of Field Ornithology :–. Hall, T. A. . BioEdit: A user-friendly biological sequence Mayr, E. . Systematics and the Origin of Species. Columbia alignment editor and analysis program for Windows //NT. University Press, New York. Nucleic Acids Symposium Series :–. Mayr, E., and P. D. Ashlock. . Principles of Systematic Zool- Hedrick, P. W. . A standardized genetic differentiation mea- ogy, nd ed. McGraw-Hill, New York. sure. Evolution :–. McKay, J. K., and R. G. Latta. . Adaptive population diver- Johnson, N. K., J. V. Remsen, Jr., and C. Cicero. . Refined gence: Markers, QTL and traits. Trends in Ecology and Evolution colorimetry validates endangered subspecies of the Least Tern. :– Condor :–. Mearns, E. . Description of a new subspecies of the American Johnson, R. R., and J. S. Castrale. . Management of breeding Least Tern. Proceedings of the Biological Society of Washington Interior Least Terns in Indiana. Proceedings of the Indiana Acad- :. emy Science :–. Miller, M. P., S. M. Haig, and R. S. Wagner. . Conflicting

Jost, L. . GST and its relatives do not measure differentiation. patterns of genetic structure produced by nuclear and mitochon- Molecular Ecology :–. drial markers in the Oregon slender salamander (Batrachoseps Kalinowski, S. T. . HP-Rare: A computer program for per- wrighti): Implications for conservation efforts and species man- forming rarefaction on measures of allelic diversity. Molecular agement. Conservation Genetics :–. Ecology Notes :–. Mills, L. S., and F. W. Allendorf. . The one-migrant- Kruse, C. D., K. F. Higgins, and B. A. Vander Lee. . Influ- per-generation rule in conservation and management. Conserva- ence of predation on Piping Plover, Charadrius melodus, and tion Biology :–. Least Tern, Sterna antillarum, productivity along the Missouri Patten, M. A., and R. A. Erickson. . Subspecies of the Least River in South Dakota. Canadian Field-Naturalist :–. Tern in Mexico. Condor :–. 818 — DRAHEIM ET AL. — AUK, VOL. 127

Patten, M. A., and P. Unitt. . Diagnosability versus mean U.S. Fish and Wildlife Service. . California Least Tern differences of Sage Sparrow subspecies. Auk :–. recovery plan. U.S. Fish and Wildlife Service, Portland, Oregon. Peck, D. R., and B. C. Congdon. . Reconciling historical pro- U.S. Fish and Wildlife Service. . Endangered and threat- cesses and population structure in the Sooty Tern Sterna fuscata. ened wildlife and plants; interior population of Least Tern deter- Journal of Avian Biology :–. mined to be endangered. Federal Register :–. Phillimore, A. B., and I. P. F. Owens. . Are subspecies use- U.S. Fish and Wildlife Service. . Recovery plan for the ful in evolutionary and conservation biology? Proceedings of the interior population of the Least Tern (Sterna antillarum). U.S. Royal Society of London, Series B :–. Fish and Wildlife Service, Twin Cities, Minnesota. Pritchard, J. K., M. Stephens, and P. Donnelly. . Infer- U.S. Fish and Wildlife Service and National Marine Fish- ence of population structure using multilocus genotype data. eries Service. . Policy regarding the recognition of distinct Genetics :–. vertebrate population segments under the Endangered Species Ramos-Onsins, S. E., and J. Rozas. . Statistical properties of Act. Federal Register :–. new neutrality tests against population growth. Molecular Biol- van Rossem, A. J., and The Marquess Hachisuka. . A ogy and Evolution :–. further report on birds from Sonora, Mexico, with description of Renken, R. B., and J. W. Smith. . Interior Least Tern site fidel- two new races. Transactions of the San Diego Society of Natural ity and dispersal. Colonial Waterbirds :–. History :–. Rozas, J., J. C. Sánchez-DelBarrio, X. Messeguer, and R. Vignieri, S. N., E. M. Hallerman, B. J. Bergstrom, D. J. Hafner, Rozas. . DnaSP, DNA polymorphism analyses by the coales- A. P. Martin, P. Devers, P. Grobler, and N. Hitt. . Mis- cent and other methods. Bioinformatics :–. taken view of taxonomic validity undermines conservation of an Schwartz, M. K., and K. S. McKelvey. . Why sampling evolutionarily distinctive mouse: A response to Ramey et al. . scheme matters: The effect of sampling scheme on landscape Animal Conservation :–. genetic results. Conservation Genetics :–. Wakeley, J. . The excess of transitions among nucleotide sub- Sruoga, A., D. Butkauskas, P. Prakas, and A. Paulauskas. stitutions: New methods of estimating transition bias underscore . Evaluation of the genetic structure of the breeding Com- its importance. Trends in Ecology and Evolution :–. mon Tern (Sterna hirundo) population by means of microsatellite Welty, J. C., and L. F. Baptista. . The Life of Birds, th ed. markers. Biologija :–. Saunders College Publishing, Philadelphia. Szczys, P., C. R. Hughes, and R. V. Kesseli. . Novel mic- Wenink, P. W., A. J. Baker, and M. G. J. Tilanus. . Mito- rosatellite markers used to determine the population genetic chondrial control-region sequences in two shorebird species, the structure of the endangered Roseate Tern, Sterna dougallii, Turnstone and the Dunlin, and their utility in population genetic in Northwest Atlantic and Western Australia. Conservation studies. Molecular Biology and Evolution :–. Genetics :–. Whitlock, M. C. . Combining probability from independent Tarr, C. L. . Amplification and determination of mitochondrial tests: The weighted Z-method is superior to Fisher’s approach. control-region sequences in oscine passerines. Molecular Ecol- Journal of Evolutionary Biology :–. ogy :–. Whittier, J. B., D. M. Leslie, Jr., and R. A. Van Den Bussche. Thompson, B. C., J. A. Jackson, J. Burger, L. A. Hill, E. M. . Genetic variation among subspecies of Least Tern (Sterna Kirsch, and J. L. Atwood. . Least Tern (Sterna antil- antillarum): Implications for conservation. Waterbirds :–. larum). In The Birds of North America, no.  (A. Poole and Wilson, E. O., and W. L. Brown, Jr. . The subspecies concept F. Gill, Eds.). Academy of Natural Sciences, Philadelphia, and and its taxonomic application. Systematic Zoology :–. American Ornithologists’ Union, Washington, D.C. Winker, K., and S. M. Haig, Eds. . Avian subspecies. Ornitho- Thompson, B. C., M. E. Schmidt, S. W. Calhoun, D. C. Morizot, logical Monographs, no. . and R. D. Slack. . Subspecific status of Least Tern popu- Wright, S. . Evolution in Mendelian populations. Genetics lations in Texas: North American implications. Wilson Bulletin :–. :–. Wright, S. . Breeding structure of populations in relation to Tirard, C., F. Helfenstein, and E. Danchin. . Polymorphic speciation. American Naturalist :–. microsatellites in the Black-legged Kittiwake Rissa tridactyla. Molecular Ecology Notes :–. Associate Editor: J. P. Dumbacher OCTOBER 2010 — LEAST TERN GENETIC ANALYSES — 819 1.000 0.941 1.000 0.653 0.778 0.825 0.811 0.833 0.647 0.947 1.000 0.014 0.467 0.727 0.802 0.582 1.000 1.000 0.906 0.933 1.000 (below reflect diagonal) mi- mic 0.947 1.000 0.842 0.615 0.833 0.765 0.032 0.038 1.000 1.000 1.000 1.000 D 0.05 level (provided as for level a basis only 0.05 0.038 0.590 0.824 0.875 1.000 Eastern A 0.929 0.923 0.772 0.751 0.824 0.834 0.326 0.360 1.000 1.000 1.000 1.000 0.047 n Table 1). Values of Values 1). n Table 0.018 0.037 0.037 0.032 0.024 0.018 0.022 1.000 1.000 1.000 1.000 1.000 0.022 1.000 1.000 1.000 0.011 0.694 0.849 0.919 0.838 0.758 0.713 0.633 0.774 0.796 0.023 0.019 0.020 0.765 0.847 0.882 0.754 0.823 1.000 0.831 0.949 1.000 1.000 1.000 1.000 0.017 0.012 0.007 0.823 0.811 0.024 0.607 0.028 0.019 0.010 0.014 0.024 0.918 0.875 0.649 0.811 0.381 0.027 0.031 0.021 6 0.02 0.030 0.016 0.012 0.010 0.013 0.031 0.022 0.045 0.021 0.015 0.008 0.009 0.028 0.015 0.037 0.025 0.059 0.054 0.034 0.043 0.039 1.000 1.000 1.000 0.933 0.804 0.789 0.650 0.039 0.027 0.025 0.750 0.762 0.813 0.882 0.665 0.803 1.000 0.904 0.898 0.933 0.898 0.714 0.615 0.040 0.033 0.049 0.023 0.039 0.039 0.566 0.739 0.583 0.801 0.813 0.641 0.793 0.919 0.900 0.787 0.930 0.893 0.474 0.597 0.047 1.000 0.054 Interior 35 0.018 0.020 0.022 0.018 0.013 0.020 0.017 0.018 0.859 0.868 0.875 0.750 0.040 0.054 0.041 0.046 0.935 0.667 0.683 0.762 0.615 0.692 0.708 0.592 0.873 0.926 0.818 0.708 0.869 0.903 0.515 0.704 0.032 0.035 0.031 1.000 0.048 0.037 0.017 0.022 0.020 0.017 0.018 0.016 0.009 0.015 0.006 0.027 0.027 0.022 0.854 0.561 0.687 0.631 0.779 0.854 0.761 0.658 0.592 0.637 0.026 0.045 (above diagonal) reflect the mitochondrial control-region data. Values in bold reflect significance at the at reflect bold in significance reflect control-region data. Values the mitochondrial (above diagonal) mit D 0.017 0.019 0.031 0.013 0.031 0.020 0.010 0.013 0.011 0.010 0.014 0.445 0.474 0.728 0.936 0.926 0.909 0.805 0.935 0.032 0.051 0.035 0.052 0.037 0.026 0.024 0.047 0.898 0.771 1.000 0.966 0.969 0.750 1.000 1.000 1.000 0.931 0.946 1.000 1.000 1.000 1.000 0.690 0.971 0.939 0.048 0.019 0.014 0.029 0.011 0.023 0.018 0.015 0.015 0.016 0.013 0.018 0.472 0.642 0.877 0.025 0.022 California NCA SCA NDMOR SDMOR KSKR MOMSR OKRR OKCR OKAR TXINT MSMSR ME MA NJ VA GA USVI FLGC MSGC TXGC 0.037 0.041 0.031 0.028 0.057 0.037 0.025 0.026 0.032 0.037 0.043 0.029 0.020 0.032 0.041 0.035 0.083 0.055 0.099 0.053 0.083 0.078 0.081 0.067 0.051 0.021 0.028 0.029 0.029 0.029 0.041 0.024 0.022 0.054 0.042 0.034 0.030 0.034 0.065 0.041 0.049 0.043 0.039 0.044 0.050 . (locations breeding of breeding areas areas i among 20 of Least are differentiation of genetic given Terns indicators Pairwise PPENDIX illustrating general trends). general illustrating TXGC 0.016 0.019 0.030 0.015 0.028 0.0 A SCA 0.008 SDMOR NDMOR NCA crosatellite data, whereas of values crosatellite KSKSR MOMSR OKRR MSGC 0.012 0.015 OKCR 0.025 0.030 0.027 0.017 0.026 0.015 0.015 0.762 0.938 0.882 0.866 0.869 1.000 1.000 1.000 1.000 0.796 0.857 0.846 FLGC USVI OKAR TXINT MSMSR ME MA NJ GA 0.036 0.035 0.053 VA