Liberty UniversityDigitalCommons@Liberty

University

Faculty Publications and Presentations Department of Biology and Chemistry

2005

Reproductive Success Across the Black-cappedChickadee (Poecile atricapillus) and CarolinaChickadee (P. carolinensis) Hybrid Zone in OhioC. L. Bronson

Thomas C. Grubb, Jr.

Gene D. SattlerLiberty University, edsattle@liberty.edu

Michael J. Braun

Follow this and additional works at: http://digitalcommons.liberty.edu/bio_chem_fac_pubs

This Article is brought to you for free and open access by the Department of Biology and Chemistry at DigitalCommons@Liberty University. It hasbeen accepted for inclusion in Faculty Publications and Presentations by an authorized administrator of DigitalCommons@Liberty University. Formore information, please contact scholarlycommunication@liberty.edu.

Recommended CitationBronson, C. L.; Grubb, Jr., Thomas C.; Sattler, Gene D.; and Braun, Michael J., "Reproductive Success Across the Black-cappedChickadee (Poecile atricapillus) and Carolina Chickadee (P. carolinensis) Hybrid Zone in Ohio" (2005). Faculty Publications andPresentations. Paper 35.http://digitalcommons.liberty.edu/bio_chem_fac_pubs/35

^LThe Aiik 122(3):759-772, 2005© The Americiin Ornithologists' Union, 2005.Printed in USA.

REPRODUCTIVE SUCCESS ACROSS THE BLACK-CAPPEDCHICKADEE {POECILE ATRICAPILLUS) AND CAROLINACHICKADEE (P. CAROLINENSJS) HYBRID ZONE IN OHIO

C. L. BRONSON,' ' THOMAS C . GRUBB, JR. / GENE D. SATTLER,'AND MICHAEL J.

'Dqmrtiticnt of Evolution. Ecohi^y, and Organlsmal Biology, The Ohio Staff Uiiivcrsifii, 318 West Tllli AveiUKColumims, Ohio 43210. USA:

'Di-piut)Jiciit of Biolo;^y mid Cliemistri/, Libert]/ Uniz'crsity, 1971 University Boiilcvnrii, i\/mhbuy^,Vir^iniiJ 24502, USA; mid

iciit ofVcrldinUe Zoolo;^y, Nntioiuil Miisciofi ofNntiirnl History, Siiiillisoniiin Institution,4210 Silver tliil Roud. Siiiliand, Mmyhmd 20746, USA

CT. —Black-capped Chickadees (/'(ifc;7fi7fr;L"(7/)/7/i/s) and Carolina Chickadees(P. cawliueusis) hybridize in an east-west band from New Jersey to Kansas. Withinthe past century, the Ohio portion of this hybrid /one and the Carolina Chickadeerange to the south have been moving northward, whereas the Black-cappedChickadee range has retracted. In Ohio, we characterized the genetic compositionof the hybrid zone using five diagnostic molecular loci. Although there was no evi-dence of assortative mating in the center of the hybrid zone, we found a relative pau-city of genetically intermediate breeding females as compared with breeding males.That suggests viability selection against female hybrids, in line with Haldane's rule.On the basis of reproductive variables (number of nestlings, reproductive success),we found a decrease in productivity of breeding pairs in the hybrid zone that issignificantly and positively related to their probability of producing homozygousoffspring at each autosomal or sex-linked locus. We also found that the decrease inproductivity was significantly and positively related to the genetic composition ofthe male of the pair {i.e. pure male chickadees more productive). These data stronglysuggest that hybrids are at a selective disadvantage. Because the zone of reducedreproductive success was considerably narrower than the zone of introgression, ourresults demonstrate that genetic introgression is occurring in the face of substantialselection against hybrids. Received 16 April 2004, accepted W jmniary 2005.

Key words: Black-capped Chickadee, Carolina Chickadee, genetic indices, hybridzone, Poecile iitricapillus, Poecile carolinensis, reproductive success.

Exito Reproductivo a traves de la Zona de Hibridacion de Poecile ntricapillusy P. carotiiicnsis en Ohio

RESUMEN. —Las especies Poecile atricapillus y P. caroliucnsis hibridan en una franjaorientada de este a oeste desde New jersey hasta Kansas. Durante el ultimo siglo,la seccion de Ohio de esta zona de hibridacion y el rango de P. carolinensis al surde esta se han desplazado hacia el norte, mientras que ei rango de P. atricapillus seha contraido. En este estudio, caracterizamos la composicion genetica de la zonade hibridacion en Ohio usando cinco loci moleculares diagnosticos. Aunque nocxistio evidencia de apareamiento asociativo en el centro de la zona de hibridacion.

'Present address: Department of Internal Medicine, Division of Immunology, The Ohio State University, 473West Twelfth Avenue, Columbus, Ohio 43210, USA. E-mail; bronson.3("osu.odti

759

760 BRON-ION Lr AL. [Auk, Vol. 122

encontramos una relativa carencia de hembras reproducHvas geneticamenteintermedias en comparacion con los machos reproductivos. Esto sugiere la existenciade seleccion por viabilidad en contra de las hetnbras hibridas, lo que concuerda conla regla de Haldane. Con base en variables reproductivas (nt'imero de pichones,exito reproductivo), encontramos una disminucion en la productividad de lasparejas en la zona de hibridacion que esta significativa y po.sitivamente rclacionadacon SLi probabilidad de producir crias homocigoticas en cada locus autosomico oligado a! sexo. Tambien encontramos que la disminucion en la productividad estuvosignificativa y positivamertte relacionada con la composicion genetica del machode la pareja (i.e. los machos puros fueron mas prodiictivos). Estos datos sugierenfuertemente que los hibridos se encuentran en desventaja seloctiva. Debido a quela zona de exito reproductivo reducido fue considerablemente mas estrecha quela zona de introgresion, nuestros resultados demuestran que a pesar de que existeseleccion en contra de los hibridos, esta sucediendo introgresion genetica.

ZONE ij-iNAMics is a fertile area forresearch on natural selection and speciation,because of the exchange of genes betweendistinct groups (Harrison 1990, 1993; Hewitt1988). Within birds (see review in Grant andGrant 1992), the hybridization of many NorthAmerican species has been studied (for reviewof Great Plains hybrid zones, see Rising 1983).For example, Black-Capped Chickadees (Poecileatricapillus) and Carolina Chickadees {P. caroli-ncnsis) are known to hybridize in many areasalong their common border (e.g. Kansas:Rising 1968; Missouri: Braun and Robbins 1986,Sawaya 1990; Illinois: Brewer 1963; Ohio: Grubbet al. 1994; Virginia: Johnston 1971, Sattler 1996,Sattler and Braun 2000; West Virginia: Sattler1996, Sattler and Braun 2000; Pennsylvania:Ward and Ward 1974, Cornell 2001). Becausethese species may not be sister taxa (Gill et al.1989, 1993; but see Sattler and Braun 2000 fordiscussion), only limited hybridization mightbe expected.

In North America, except for a peninsulardistribution in the Appalachian MoLintains,the Black-capped Chickadee distribution abutsthe north edge of the Carolina Chickadee di.s-tribution (Mostrom ct al. 2002). In the southernAppalachians, Black-capped Chickadees areoften found at high elevations, and CarolinaChickadees at lower elevations. In the early1880s, Carolina Chickadees were described aspermanent residents only within the southernportion of Ohio (Wheaton 1882). By the late1930s, the hybrid /one was probably locatedacrt)ss the middle of the state, approximately inthe location of the east-west U.S. Interstate 70(Trautman 1940). Currently, the zone is located

-100 km farther north, approximately along U.S.Highway 30 (Grubb et al. 1994, Peterjohn 2001).In other words, the Black-capped Chickadeedistrihtition has been receding northward.

The chickadee hybrid zone is quite narrow,with genetic cline widths on the order of 20to 30 km (Sattler 1996, Sattler and Braun 2000).Given the likely age of contact and the dispersalcapabilities ot chickadees (Weise and Meyer1979), the narrow cline widths suggest thatsome sort of selection may oppose introgressionacross their hybrid zone {Barton and Gale 1993).In Illinois, Brewer (1963) found that hatchingsuccess was lower in the chickadee hybrid zonethan for either parental species, but he hadcomplete data on only four hybrid zone nests.He attributed the reduction to infertility andretarded development of eggs.

We wished to study the relationship betweenhybridization and reproductive success in greaterdetail. The objectives here were to employ geneticmarkers to map one segment of the hybrid zonein north-central Ohio and to examine the rela-tionship between genetic composition of theparents and reproductive success.

Field nicthods. —The area of the hybrid zonestudied within Ashland County, Ohio (4O''5O'N,82"15'W) was bounded by County Road 700 onthe north. State Route 95 on the south. StateRoute 89 on the west, and County Road 175 onthe east (Fig. 1). The study area was 23 km fromnorth to south and 6 km from west to east. Thelandscape was about equally divided amongpasture, row crops, and woodlands. To limit the

July 2005] clive Success in tli/hriii Zone 761

overlap of points in Figures 2-4, the 21 sampling In November of 1993 and 1994, we placedlocations were condensed to 10 pooled s^imples remote-controlled Feeder traps (Pierce andbased on similar latitudes. Frt>m north to south, Grubb 1979) filled with sunflower seeds withinthe groupings were a-b, c, d, e, g-f, h, i-m, n-p, privately owned woodlands af all study sitesq-r, and s-u (Fig. 1). within the zone. From December through

5-

CR700

10-

km

15-

25-

FIG. 1. Tbe study transect (light gray area) in Ashland County, Ohio, indicating sampling points(letters) and major roadways.

762 B K O N S O N hi ' AL. [Auk, Vol. 122

February, we trapped or mist-netted chickadeesvisiting each feeder. In late February, we placedartificial nesting snags (Grubb and Bronson1995) in the woodlands and monitored themthrough the chickadee breeding season (to theend of June).

At the time of capture, we banded each birdwith a federal aluminum band and a coloredleg streamer (Sullivan 1984) tor individualidentification from a distance. We weighedeacb bird to the nearest 0.1 g using a springbalance. Unflattened wing chord and tail lengthwere measured to tbe nearest 0.5 mm, and tar-sus length (from the hent "elbow" to tbe hent"wrist") to tbe nearest O.l mm. Sex was initiallydetermined tbrougb bebavioral observationsubsequent to capture (i.e. males dominant tofemales) and relative size of tbe members ofa pair (Desrochers 1990, Smith 1991). Sex waslater verified for many individuals on the basisof vocalizations (e.g. singing males, beggingfemales) and morpbology (e.g. male cloacalprotuberance, female brood patch). Finally, sexwas determined through genetic techniques(see sex-linked marker below).

Maleailnr nietlwtis. — Methods for collectingblood, extracting DNA, genetic analysis, andparental analysis are detailed in Bronson etal. (2003). Tbe genetic markers employed arediagnostic restriction-fragment-length poly-morphisms (RFLP) of three types (Sawaya1990, Sattler 1996, Sattler and Braun 2000). Theenzyme/prohe combinations of Eco Rl/ski, Bg!II/RP104, and Ava II/RP7 detect autosomal loci.Tbe combination of Pst I/C7 detects a sex-linkedlocus on the Z chromosome, and the combina-tion of P^t l/mtDNA was used to genotypethe maternally inherited mitochondrial DNA(mtDNA).

Data aimh/fiis.—On tbe basis of the statisti-cal models of Boecklen and Howard (1997),only four or five markers might be adequateto coarsely categorize individuals in a bybridzone. However, tbe models assume tbat no I' xbackcross or backcross x backcross matingoccurs witbin tbe zone. We were not willing tomake that assimiption tor tbis cbickadee byhridzone, so we followed Boecklen and Howard's(1997) suggestion and created a genetic indexbased on "tbe percentage of loci tbat are cbarac-teristic of a pure species."

We combined tbe five molecular genotypes{Eco Rlhki, Bgl II/RP104, Ami 11/RP7, P^l i/C7,

and Pst l/mtDNA) into a genetic index (Cl)for each individual, calculated as tbe numberof Carolina Chickadee alleles divided by tbetotal number of alleles examined (Bronson et al.2003). Tbere are two alleles for eacb autosomalmarker and one for tbe mtDNA haplotype. TbeZ-linked marker has two alleles in males andone in females (females are tbe heterogameticsex in birds). Thus, Gl was hased on up to eightmarker alleles for females and nine markeralleles for males. For some of the nonparametriccorrelations, Gl was transformed

Gr = |GI-0.5|

to adjust for the potential underlying parabolicdistribution of Gl. Transformed Gl' ranges from0.5 for either pure Carolina Chickadee or pureBlack-capped Cbickadee to 0 for maximal inter-mediate birds.

For each set of parents, a compatibility index(Cl) was calculated on tbe basis of tbe expectedproportion of homozygous offspring tbey couldproduce at eacb autosomal or sex-linked locus,averaged across loci (Bronson et al. 2003):

wbere ^ are the autosomal loci (fi, = Eco Rl/ski,c-, = B<il il/RP104, t>, - Ava 1I/RP7; (1 = I if tbe par-ents are identical bomozygotes, fi = 0.5 if at leastone is beterozygous, and 5 = 0 if they are oppositehomozygotes) and t is the sex-linked locus Pst 1/C7 (t = 0.5 if the parents are identical homozy-gotes, f = 0.25 if the male is beterozygous, and t =0 if tbey are opposite bomozygotes). To allow forequal weighting of sex-linked loci in tbe Cl, onlythe expected proportions of male offspring areconsidered for those loci (females cannot be het-erozygous), lt seems important to at least equal-ize the contribution of the sex-linked markers toCl because of the disproportionate involvementof sex chromosomes in reproductive isolation(Coyne and Orr 1989). Compatibility indexranges from 0 to 1, with higher values indicatingmore complementary genetic backgrounds.

MINITAB, version 13.1 (Minitab, StateCollege, Pennsylvania) was used for generatingnonparametric correlations based on the ranks.Spearman's rbo (p). To reduce tbe cbance ofmaking Type 1 errors, we employed tbe sequen-tial Bonferroni tecbnique (Hochbcrg 1988) tocorrect for the number of similar tests.

July 2005] Rfprodiictivc Siiccf^^ iu Hybru 763

We expected /; \iriori tbat several variables(all reproductive measures and Cl of a hreedingpair) would bave reduced values in tbe middleof the hybrid zone. Consequently, for figuresinvolving the relevant comparisons, a second-order polynomial trend line was included(instead of a linear trend line), but both typesof lines are shown, if only to facilitate visualassessment ot the pattern.

RLSULT.S

The frequencies of alleles at marker loci con-sistently changed in a clinal fasbion across tbebyhrid zone, from a low proportion of CarolinaCbickadee alleles in tbe nortb to a bigh propor-tion in tbe soutb, witb the midpoint in allelefrequency between 10 and 15 km in all five cases(Fig. 2). Thougb coincident in position, somevariation among markers in dine width wasapparent. For example, tbe frequency oftbe Pst 1/C7 marker allele changed from 0 to 1 between7.4 and Id.2 km, whereas Eco RMski only variedfrom about 0.1 to 0.8 over the entire transect.

Across tbe byhrid zone, we obtained recordsof reproductive output for 29 pairs of chicka-dees. Tbere was no significant correlationbetween genetic indices of tbe male versusfemale of each pair (Spearman's p = 0.310, P =O.IOI, /; = 29), wbich suggests that mating wasnonassortative. Genetic confirmation of parent-age for two nests could not be obtained, becauseof a laboratory accident that caused the loss ofthe DNA for tbe nestlings. We analyzed DNAfingerprints for the remaining 27 nests, findingno evidence of extrapair fertilizations. Of tbe100 offspring tested, 14 had one or more unat-tributable bands. None of those 14 individualscould be excluded as tbe offspring of the puta-tive parents on the basis of band-sbaring scores(Wetton et al. 1987).

For all 29 pairs, tbere was no loss of offspringbetween batching and fledging. Therefore, forthose 29 pairs of chickadees, number of nestlingsequaled number of tledgUngs, fledging success(ratio of fledglings to nestlings) was lOO'fi., andreproductive success (ratio of tledglings to eggs)equaled batching success (ratio of nestlings toeggs) (Table 1). Figure 3 places various reproduc-tive measures within the study landscape (witha second-order polynomial trend line includedfor easier visualization). No trend was appar-ent in clutch size across the zone (Fig. 3A), but

botb number of nestlings or fledglings (I-'ig. 3B)and reproductive success (Fig. 3C) displayedapparent troughs near the midway point of thezone. The effect on reproductive output in thosetroughs appears to be substantial; at least halftbe eggs failed to batch in 14 of 22 nests between5 and 15 km, whereas 0 of 7 nests outside tbatzone had batcbing success <0.7. However, tbosetroughs were also narrow; all nests witb repro-ductive success <0.5 wore found in tbe regionfrom 7.4 to 13.0 km.

For breeding individuals, Gl of eacb sexhad a positive and significant relationshipwith location in the study landscape (femaleGl: Spearman's p = 0.553, /' = 0.002, n = 29, Fig.4A; male Gl; Spearman's p = 0.769, P < 0.001,II = 29, Fig. 4B). There was a paucity of breedingfemales of intermediate Gl in comparison withbreeding males (e.g. no females vs. 12 males inthe Gl range from 0.3 to 0.6; Fig. 4A-B). Tbe Clof breeding pairs was lowest in the middle ofthe transect (Fig. 4C). That trough coincidedwith tbe trough in productivity (Fig. 3C).

Transformed Gl of tbe female ranged from0.12 to 0.50 (Fig. 5A-B) and was not signifi-cantly related to any reproductive measure(Table 2). Tbe lack of females (only one) witha transformed Gl <0.25 could bave weakenedthe correlation. The transformed Gl of tbe male,which ranged from 0.05 to 0.50 (Fig. 5C-D), wasnot significantly related to clutch size (Table2), but was positively and significantly relatedto number of nestlings or fledglings (Table 2;Fig. 5C) and to reproductive success (ratio offledglings to egg; Table 2; Fig. 5D).

The Cl between tbe male and female of abreeding pair ranged from 0.25 to 1.0 (Fig. 6).Although Cl was not significantly related toclutch size (Table 3), it was positively and signifi-cantly related to number of nestlings or fledglings(Table 3; Fig. 6A) and to reproductive success(ratio of fledglings to egg; Table 3; Fig. 6B).

DISCUSSION

St'iection and hybrid zone malnti'nance.—Ourresults demonstrate tbat tbere is a narrow regionof reduced reproductive success at the center oftbe chickadee hybrid zone in Obio. We moni-tored a 23-km transect of the /one witbin wbicbwe observed substantially reduced reproduc-tive output of populations in tbe center of thetransect as compared with tbose at eacb end. On

764 BRONSON ET AL. fAuk, Vol. 122

O O B '

•5 0g 0 4 H'£ 03 'oQ . 0 2 '

North-south distance (km) North-south distance (km)

u"SU

O{

coa.

06

0.5

0.4

03

0.2

North-south distance (km)

u"5c0•c0a

U6

05

0.*0.3

0.2

North-south distance (km)

North-south distance (km)

FIG. 2. Distribution of Carolina Cbickadee alleles al eacb of five loci across tbe study site in nortb-ccntral Ohio. Kilometer 0 is Asbland County Road 700, tbe nortbern boundtiry of the study area.(A) EcoR Vski, II = 143; (B) Ava II/RI'7, u - 121; (C) Bgl 1I/RP104, n - 142; (D) Pst I/C7, ii = 60; (H) Pst 1/mtDNA, n 142. Only adults, not nestlings, were used for tbe analysis to limit nonindcpendence ofdata points because of relatedness. All available adults (breeders and nonbreeders) were included.Varying sample sizes are attributable to (1) difficulty in scoring a few nonbrceding individuals,especially for Ava TI/RP7; and (2) a need to know tbe sex of eacb individual to determine tbe num-ber of alieles to be considered for Pst I/C7. Consequently, only individuals observed breeding wereincluded for tbat locus to insure tbe correct attribution of sex.

tbe basis ot geograpbic distributions of allelefrequencies for five diagnostic genetic markers,tbe transect we monitored spanned tbe core oftbe bybrid zone. Brewer (1963) also providedanecdotal evidence of reduced reproductivesuccess in the cbickadee bybrid zone. Altbougbcomparable reproductive success data (tledg-lings per successful clutch) are not availablefor either Black-capped nr Carolina chickadeesin areas immediately adjacent to tbe study

area or for Carolina Chickadees in general, tbeextremes of our transect bad a similar numberof Black-capped Chickadee fledglings to tbesouthern peninsula of Micbigan (5.5 vs. 6.6;Nickell 1956). Therefore, we are confident tbatthe observed reduction in productivity is lim-ited to tbe bybrid zone and is not a widespreadcbickadee phent)menon. Reduced reproductivesuccess indicates tbat some form of selection isoperative in the bybrid zone.

July 2005] Rcjirodiictivc Success in Hybrid Zone 765

TABLE 1. Genetic and reproductive data for each pair of chickadees observed. All nestlings fledged.The location column gives the letter designation of each site (see Fig. 1) and its distance inkilometers from the northern end of the study transect.

Location

Km

1.11.13.47.07.47.47.49.89.89.89.8

11.512,212.212.212.212.212.212.213.013.013.013.013.013.016.216.222.422.4

Site

aacdeecff

g8h

iiiik11nnnoo

Pqrtu

Genetic index

Female

0.1250.2500.1250.0000.7500.0000.1250.2500.2500.0000.0000.1251.0000.1250.8750.1250.7500.6250.8750.7500.0000.0000.7500.2500.7501.0001.0001.0001.000

Male

0.1110.222

o.m0.1110.3330.4440.7780.0000.2220.3330.4440.3330.3330.5560.5560.7780.2220.3330.3330.444O.f)670.6670.5560.6671.0000.8891.0001.0001.000

Compatibilityindex

1.0000.7500.8750.8750.6250.5000.2500.6250.8750.6250.6250.6250.3750.3750.6250.5000.5000.3750.3750.5000.3750.3750.6250.3750.7500.8751.0001.0001.000

Reproductive variables

Clutch

47887779576888878777686467977

Nestlings

45863261422153254113173457976

What is the nature of selection in tho chicka-dee hybrid zone? Reduced reproductive successwas linked to genetic intermodiacy of males andto the genetic compatibility of a breeding pair.Those links suggest that intrinsic genetic incom-patibihties are responsible for the reducedreproductive success. However, althoughreproductive measures were related to geneticcharacteristics of the breeding pairs, both repro-ductive measures and genetic characteristicswere also related to geographic position withinthe zone (Fig. 3). Thus, parental genotypesmay not have been an exclusive cause for thereduced productivity in the middle of oursample transect. For example, environmentalattributes (e.g. food availability, temperature.

precipitation) also could have been involved.Although we did not detect any gradients orother inconsistencies in environmental char-acteristics across our sample transect, suchexogenous factors (Harrison 1990, Arnold 1997)could have existed and been causal. Because ofits observational nature, our study cannot dif-ferentiate between intrinsic or extrinsic factorsin the reduced reproductive success.

The best method for separating those causesis to perform a manipulative experiment(Moore and Price 1993). Therefore, as aresult of the observations reported here, werelocated chickadees of bolh parental speciesatid hybrids into isolated island woodiotswithin the hybrid zone and again observed

766 BRONSON ET AL. |Auk, Vol. 122

B

uauuin0)

tiv

3• D0•.aQ:

03

OT

OS

05

04

03

0 2

10 15

North-south distance (km)

North-south distance (km)

North-south distance (km)

Fic. 3. Relationship of reproductive variableswith location in the hybrid zone {with thesecond-order polynomial trend line for easiervisualization). Kilometer 0 is Ashland CountyRoad 700, the northern boundary of the studyarea. The four sizes of circles from smallest tolargest indicate sample sizes of 1, 2, 3, and 4,respectively. (A) Clutch size, (B) number ofnestlings or fledglings, and (C) reproductivesuccess (ratio of fledglings to eggs).

10-

09-

£ 0 7 -

O 06-

1.0 •

0 . 9 -

o•D 0 5 -

•0 15

North-south distance (km)

B

10 Ib 30 2b

North-south distance (km)

"5 0 9

•5O8

S"'

= 04

+3030.0;

0 D 1 •

" 0 0

\ ^ • • • •

• • •

• •

c0 5 10 15 20 25

North-south distance (km)

FIG. 4. Relationship of genetic and compat-ibility indices of breeding individuals withlocation in the hybrid zone. Kilometer 0 isAshland County Road 700, the northern bound-ary of the study area. The four sizes of circlesfrom smallest to largest indicate sample sizes of1, 2, 3, and 4, respectively. (A) Female geneticindex; (B) male genetic index; and (C) compat-ibility index of breeding pair, including thesecond-order polynomial trend line for easiervisualization. The CI is calculated as the pro-portion of Carolina Chickadee alleles presentin an individual {0 = Black-capped Chickadee;1 = Carolina Chickadee). The Cl was calculatedusing the average of the proportion of homozy-gous offspring a breeding pair could produce ateach of the loci (0 = least compatible; 1 = mostcompatible).

Julv 20051 Rcpwductizv Si(ft"fss i)i Hybrid Zone 767

Transformed genetic index of female

B

Transformed genetic index of female

Transformed genetic index of male

0 9 '

0 7 •

OB

OA

0 3

02 '

Cl

0 0

D m •

•

Transformed genetic index of male

FIG. 5. Relationship of reproductive variables with transformed genetic indices, and lineartrend lines for easier visual comparison. The four sizes of circles from smallest to largest indicatesample sizes of 1, 2, 3, and 4, respectively- (A) Number of nestlings or fledglings versus femaleGI', Spearman's p = 0.305; (B) reproductive success (ratio of fledglings to egg) versus female GI',Spearman's p = 0.229; (C) number of nestlings or fledglings versus male GI', Spearman's p = 0.579; (D)reproductive success (ratio of fledglings to egg) versus maleGI', Spearman's p =0.540. TheGr(Gr =IGl -0-^1) is calculated from the proportion of homozygous allele pairings present in an individual(0.5 = Black-capped or Carolina Chickadee; 0 = maximal intermediate birds).

T.\B].r 2. Spearman's rho (p) ft>r chickadee reproduction on the transformed genetic indices (GI' =IGI - 0.51) of the female and male of a pair. Sample size for all correlations is 29, Significantrelationships based on the sequential Bonferroni technique (/i ,. = /' x [4 - Rank]; Hochberg 1988}are shown in bold.

Sex

Female

Male

Reproductive variable

Clutch sizeNestlings or fledglingsReproductive successClutch sizeNestlings or tledglingsReproductive success

P

0,0760.3050.2290.1650.5790.540

P0-6960.1070.2310.3910.0010.002

Rank

312312

Padi

0.6960-3210.4620.3910.0030.004

reproductive success (Bronson et al. 2003).The results of that experiment indicated thatendogenous factors are primarily responsiblefor selection in the Ohio section of thechickadee hybrid zone. When pure pairs weremoved into the hybrid zone, they still had

greater reproductive success than hybrid pairssimilarly moved within the zone. Thus, withthe environment held relatively constant, theobserved reproductive decline of hybrids inthe hybrid zone must be mainly attributable tointrinsic genetic factors.

768 BRONSON F,T AL. |Aiik, Vol. 122

Compatibility index of breeding pair

Selection in fhe zone is probably balancedby inward dispersal of naive parentals to cre-ate a stable, narrow hybrid zone (Barton andHewitt 1989). A variety of evidence suggeststhat this hybrid zone is temporally stable (e.g.Tanner 1952, Rising 1968, Robbins et al. 1986,Grubb et al. 1994, Sattler and Braun 2000; butsee Merritt 1981). The habitat in which thesechickadees meet lacks obvious barriers and hasbeen relatively unchanged since the last glacialmaximum, so it is plausible that the hybrid zoneexisted long before it was detected. There is noevidence of assortative mating in our data or inprevious studies (Robbins et al. 1986). Also, allgenetic clines in this and previous studies of thehybrid zone (Sawaya 1990, Sattler 1996, Sattlerand Braun 2000) were coincident in position,and dine widths were generally narrow withrespect to the dispersal capabilities of chicka-dees (Weise and Meyer 1979). All these factssuggest that a narrow hybrid zone has existedbetween these chickadees for a relatively longperiod, and that a balance of selection and dis-persal maintains the zone.

Both the demonstration of intrinsic selec-tion against hybrids and the recent northwardmovement of the zone are indications fhat thisportion of fhe chickadee hybrid zone functionsas a "tension zone" (Key 1968, Barton andHewitt 1985). In such cases, the location of thezone reflects a balance between dispersal fromparental populations and selection againstindividuals of mixed ancestry, regardless ofenvironment. The zone then moves until itreaches a location limiting either popula-tion density or dispersal (Harrison 1993).Tension zones are believed to be one of themost common kinds of hybrid zones in nature(Barton and Hewitt 1989).

Caveats.—The CI used here and bv Bronsonet al. (2003) is one of a suite of possible compat-ibility indices. Our CI considers each available

TABLE 3. Spearman's rho (p) for chickadee reproduction on the cumpaHbilityindex of the breeding pair. Sample size for all correlations is 29. Significantrelationships based on the sequential Bonferroni technique ( i j =p [4-RankJ;Hochberg 1988) are shown in bold.

1.0 1

0.9

IgOBd>u ato10 06

0)> 05

3 04

a« 0 2IT

0 1 •

00B

• • i

*

41

Compatibility index of breeding pair

FIG. 6. Relationship of reproductive variableswith CI of breeding pair, and linear trendlines for easier visual comparison. Small,medium, and large circles indicate samplesizes of 1, 2, and 3, respectively. (A) Numberof nestlings or fledglings. Spearman's p = 0.425;(B) reproductive success (ratio of fledglingsto egg). Spearman's p = 0.450. The CI wascalculated using the average of the proportionof homozygous offspring a breeding pair couldproduce at each of the loci (0 = least compatible;1 = most compatible).

Reproductive variableClutch sizeNestlings or fledglingsReproductive success

Rank

-0.0270.4250.450

0.8880.0210.014

0.8880.0420.042

July 20031 Reproductive Si<cce^!i in Hybrid Zone 769

locus independently and additively. A subset ofalfernative CIs would include epistatic interac-tions across loci and he nonadditive. Withinthis alternative subset, a pair would be catego-rized as pure only if all loci were homozygousfor alleles representative of the same species(i.e. had only Carolina Chickadee alleles). Bycontrast, our CI will miscategorize some hybridpairings as pure. For example, hybrid pairsthat are homozygous and matching at all locibut do not have all alleles of one species willbe miscategorized as pure. As a result, ourconclusions based on this CI are conservative.We accepted that limitation primarily becauseit is conservative and, with nearly equalimportance, because of the realization that ourlimited number of markers and families wouldlikely cause some hybrid pairings to be miscat-egorized regardless.

Our observation of no extrapair fertilizationsis remarkable and deserves some attention.Extrapair fertilizations have been documentedin many songbirds, including Black-cappedChickadees in continuous habitat (Otter ct al.1998), and we expected to see some extrapairoffspring. Although we do not have a definitiveanswer for our lack of exfrapair offspring, wedoubt that it is a consequence of our parentalanalysis method. We used multilocus finger-printing (Jefferys'probes 33.15 and 33.6;Jefferyset al. 1985a, b) conducting the standard pater-nity and maternity analyses (Parker and Burley1998). Although alternative methods nowexist (e.g. microsatellite markers) that mightbe considered more powerful, the multilocusfingerprinting method has been successful inother avian populations with similar popula-tion band-sharing levels and failed to be effec-tive only in highly inbred populations. Anotherpossible explanation is that fragmentation ofthe breeding habitat in our study area limitedopportunities for chickadees to seek extrapairfertilizations. However, although our studyarea was fragmented, chickadees are known tocross habitat gaps daily in areas of greater frag-mentation (Grubb and Doherfy 1999).

A more likely cause for the lack of extra-pair offspring is thaf we may have primar-ily sampled dominant breeding pairs. In fhewinter flocks, these species form wifhin-sexdominance hierarchies (Hartzler 1970). Femaleswill desert their mate (in the winter) or obtainextrapair fertilizations (in the breeding season)

in favor of a male that ranks higher than theirmate (Smith 1991, Otter and Ratcliffe 1996,Otter ft al. 1998). Consequently, few extrapairoffspring are observed in fhe dominant pair ofa winter flock. Additionally, the dominant pairgenerally obtains a large breeding territory fhatencompasses most of the winter flock's territory(Smith 1991). We initially captured the birdsin winter at temporary feeding stations andthen focused our artificial nest placements inthe vicinity of those feeders to obtain breed-ing data on the birds banded at those stations.As a result, we likely sampled the dominantbreeding pairs while the subordinate pairs wereforced to breed in unsampled areas. Althoughthis sampling bias may result in detection offew exfrapair fertilizations, we do not believeit would unduly bias our data with respect toreproductive success. Our genetic analyses ofthe zone indicate quite a hybrid swarm, andno pattern was observed in the relationshipbetween the genetic indices of the observedbreeding pairs (i.e. dominance status was notcreating assortative mating based on genetics).We also do not believe our results regardingextrapair fertilizations should be generalized toall chickadee populations.

Variation ID dine widths. —A\\ genetic clinesexamined here were generally narrow, on theorder of 20-30 km or less, which is consistentwith other studies of fhe chickadee hyhrid zonein Missouri and Appalachia (Sawaya 1990,Sattler 1996, Sattler and Braun 2000). However,there was some interesting evidence of varia-tion in cline width. The dine for the autosoma!marker Eco Rl/f^ki cline was the broadest of thediagnostic markers in Ohio, as also observed inAppalachian transects (Sattler 1996, Sattler andBraun 2000). The cline for the sex-linked markerP^l I/C7 was the narrowest, again as observedin Appalachia by Sattler (1996). The variationin cline width suggests that selection againsthybrids is stronger af some loci than at others.Limited introgression of sex-linked markers hasbeen observed in other hybrid zones (e.g. Dodet al. 1993, Ferris et al. 1993) and is consistentwith fhe expectation that selection will tend tobe stronger on sex-linked loci (e.g. Charlesworthetal. 1987).

Haldimc's RK/C —Haldane's rule holds thatin hybridization events, reduced viability orfertility should be more evident in the het-erogametic sex —the female in birds (Haldane

770 BRONSON ET AL. [Auk. Vol. 122

1922, Orr 1997, Turelli 1998). An analysis of fhenestling sex ratio in the chickadee hybrid zonein Pennsylvania yielded a lack of significantsupport for Haldane's rule (Cornell 2001). Ourdata relating to fertility also show no supportfor fhe rule; the observed reduction in produc-tivity was related to male genetic composition,not female. It should be noted, however, thatthe small number of genetic markers employedcould have contributed to the difference in therelationships of female GI and male CI wifhreproductive measures. Male GI had a largernumber of possible values because of the inclu-sion of a sex-linked marker.

Haldane's rule may not directly apply tocomparisons of parenfal genetics with repro-ductive measures in all cases. Reproductivevariables such as hatching success should becompared to the sex of the offspring themselves.Unfortunately, we do not know the sex of nesf-Hngs or unhatched eggs. If viability is affected,Haldane's rule would predict an increased ratioof females to males in unhatched eggs with acompensatory decrease in the ratio for viableoffspring.

Other data from the present study indicatethaf there may be a viability effect on females.For example, a distinct gap between 0.3 and0.6 exists in the distribution of adulf breeding-female GI (Fig. 4A). By confrasf, 12 males hadGIs between 0.3 and 0.6 (Fig. 4B). Apparently,highly heterozygous females had been removedfrom our population samples before reproduc-tion, the sampling point of this study. Therefore,Haldane's rule might well have been manifestedby reduced viability of females prior to repro-ductive age.

Conclusion.— Our results indicate that thewidth of the zone based on reduced reproduc-tive success (-6 km) is less than half that basedon the genetic indices (>15 km). Furthermore,this relationship of reproductive and geneticindices of hybridization is likely conservative,because we know that two components of GI{Eco R\hki and B;^l II/RP1(14) show introgressionof Black-capped Chickadee alleles in southernOhio (Sattler 1996). The pattern of reproductivesuccess across a hybrid zone provides informa-tion about fhe firsf stage of hybridization and itseffect, the actual barrier to the passage of geneticmaterial between two taxa. Possibly, the widthof the zone of decreased reproductive successcould indicate the strength of thaf barrier. By

comparison, the patterns of introgression inthe genetic markers across a zone also provideinformation about the strength of the barrier,because they represent the long-term effect ofhybridization. When hybridization is effective(i.e. backcrosses are presenf), a hybrid zone isless analogous fo a wall and more like a semi-permeable membrane that allows alleles to passthrough at various rates, depending on allele-specific selection factors (Barton 1983).

ACKNOWLEDGMENTS

We thank many woodland owners forallowing us to work on their property; S.Humrichouser for his untiring help with fieldwork; N. Arguedas and C. Huddleston forassistance with lab work; and R. Curry, S.Gaunt, F. Gill, D. Nelson, P. Parker, the OhioState Universify (OSU) Behavioral EcologyGroup, the Smithsonian Laboratory ofMolecular Systematics lunch group, and threeanonymous reviewers for discussion and com-ments. This study was funded by the AmericanOrnithologists' Union (Betty and Herbert CarnesAward) and American Museum of NaturalHistory (Frank M. Chapman Memorial Fund) foC.L.B., and was performed under OSU ILACUCprotocol 93A0206, federal banding permif 20653,and Ohio banding and collecting permit 509.

LITERATURE Cirpn

ARNOLD, M . L. 1997. Natural Hybridization andEvolution. Oxford University Press, NewYork.

BARION, N . H . 1983. Multilocusciines. Evolution37:454-471.

BARTON, N . H. , AND K. S. GALE. 1993. Genetic

analysis of hybrid zones. Pages 13-45 inHybrid Zones and fhe Evolutionary Process(R. G. Harrison, Ed.). Oxford UniversityPress, New York.

BARTON, N . H. , ANDG. M . Hhwnx 1985. Analysisof hybrid zones. Annual Review of Ecologyand Systematics 16:113-148.

BARTON, N . H. , AND G. M . HKWITI. 1989.

Adaptation, speciafion and hybrid zones.Nature 341:497-503.

BOFCKLEN, W. J., AND D. J. HoWARD. 1997.Genetic analysis of hybrid zones: Numbersof markers and power of resolution. Ecology78:2611-2616.

July Reproductive Suceess in Hybrid Zone 771

BRAUN, M. J., ANDM. B. ROHHINS. 1986. Extensiveprotein similarity of the hybridizing chicka-dees Parus atriciipiUus and P. curoliiieiisis.Auk 103:667-675.

BRFVVER, R. 1963. Fcological and reproductiverelationships of Black-capped and Carolinachickadees. Auk 80:9^7.

BRONSON, C. L., T. C. GRUESB, JR., AND M. J.BRAUN. 2003. A test of the endogenous andexogenous selection hypotheses for themaintenance of a narrow avian hybrid zone.Evolufion 57:630-637.

CHARLESWORTH, B., J. A. COYNE, AND N. H.BARTON. 1987. The relative rates of evolu-tion of sex chromosomes and autosomes.American Naturalist 130:113-146.

CORNELL, K. L. 2001. Hatching success and nest-ling sex ratio in Black-capped and Carolinachickadees: Do hybridizing chickadees fol-low Haldano's Rule? M.Sc. thesis, VillanovaUniversity, Villanova, Pennsylvania.

COYNE, J. A., AND H. A. ORR. 1989. Patterns of spe-ciation in Drosophiia. Evolufion 43:362-381.

DESROCUERS, A. 1990. Sex determination ofBlack-capped Chickadees with a discrimi-nant analysis. Journal of Field Ornifhology61:79-84.'

DOD, B., L. S. JERMIIN, P. BOURSOT, V. H.CHAPMAN, ]. T. NTELSEN, AND F. BONHOMME.1993. Counterselection on sex chromosomesin the Mtts miiscuhis European hybrid zone.journal of Evolutionary Biok)gy 6:529-546.

FERRIS, C , J. M. RUBIO, L. SERRANO, J. GOSALVEZ,AND G. M. HEWHT. 1993. One way intro-gression of a subspecific sex-chromosomemarker in a hybrid zone. Heredity 71:119-129.

Gui, F. B., D. H. FUNK, AND B. SILVERIN. 1989.Protein relationships among titmice (Pnrus).Wilson Bulletin 101:182-197.

GILL, F. B., A. M. MOSTROM, AN[) A. L. MACK.1993. Speciatinn in North American chicka-dees: I, Patterns of mtDNA genetic diver-gence. Evolufion 47:195-212.

GRANT, P. R., ANDB.R. GRANT. 1992. Hybridizationof bird species. Science 256:193-197.

GRUBB, T. C , JR., AND C. L. BRONSON. 1995.Artificial snags as nesting sites for chicka-dees. Condor 97:1067-1070.

GRUBB, T C , JR., AND P. P. DOHERTY, JR. 1999. Onhome-range gap-crossing. Auk 116:618-628.

GRUBB, T C , JR., R. A. MAUCK, AND S. L. EARNST.1994. On no-chickadee zones in Midwestern

North America: Evidence from the OhioBreeding Bird Atlas and fhe North AmericanBreeding Bird Survey. Auk 111:191-197.

HAI.DANE, J. B. S. 1922. Sex ratio and unisexualsterility in hybrid animals. Journal ofGenetics 12:101-109.

HARRISON, R. G. 1990. Hybrid zones: Windowson evolutionary process. Oxford Surveys inEvolutionary Biology 7:69-128.

HARRISON, R. G., ED. 1993. Hybrid Zones andthe Evolutionary Process. Oxford UniversityPress, New York.

HAKIZLER, J. E. 1970. Winter dominance rela-tionship in Black-capped Chickadees.Wilson Bulletin 82:427-434.

HEWITI; G. M. 1988. Hybrid zones-Natural lab-oratories for evolutionary studies. Trends inEcology and Evolution 3:158-167.

HOCHBERG, Y. 1988. A sharper Bonferroni pro-cedure for multiple tests of significance.Biometrika 75:800-802.

JEIIREYS, A. J., V. WILSON, AND S. L. THFIN.1985a. Hypervariable minisatellite regionsin human DNA. Nature 316:67-73.

JEFFREYS, A. J., V. WILSON, AND S. L.THEIN. 1985b.Individual-specific fingerprints of humanDNA. Nature 316:76-79.

JoiiNsioN, D. W. 1971. Ecological aspects ofhybridizing chickadees (Panis) in Virginia.American Midland Naturalist 85:124-134.

KEY, K. H. L. 1968. The concept of sfasipafricspeciation. Systematic Zoology 17:14-22.

MERRLI r, P. G. 1981. Narrowly disjunct allopatrybetween Black-capped and Carolina chicka-dees in northern Indiana. Wilson Bulletin93:54-66.

MOORE, W. S., AND J. T. PRICE. 1993. Nature ofselection in the Northern Flicker hybridzone and its implications for speciafion the-ory. Pages 196-225 in Hybrid Zones and theEvolutionary Process (R. G. Harrison, Ed.).Oxford University Press, New York.

MOSTROM, A. M., R. L. CURRY, AND B. LOHR. 2002.Carolina Chickadee. In The Birds of NorthAmerica, no. 636 (A. Poole and F. Gill, Eds.).Birds of North America, Philadelphia.

NICKELL, W. P 1956, Nesting of the Black-capped Chickadee in southern peninsula ofMichigan. Jack-Pine Warbler 34:127-138.

ORR, H. A. 1997. Flaldane's rule. Annual Reviewof Ecology and Systematics 28:195-218.

OrrtR, K., AND L. RATCLIFFE. 1996. Female initi-ated divorce in a monogamous songbird:

772 BRONSON ET AL. [Aak, Vnl. 122

Abandoning mates for males of higherqualify. Proceedings of the Royal Society ofLondon, Series B 263:351-355.

OTTER, K., L. RATCLIFFE, D. MICHAUD, ANDP. T. BoAc. 1998. Do female Black-cappedChickadees prefer high-ranking males asextra-pair partners? Behavioral Ecology andSociobiology 43:25-36.

PARKER, P. G., AND N . T. BURLEY, EDS. 1998. Avianreproductive tactics: Female and male per-spectives. Ornithological Monographs, no.49.

PETERJOHN, B. G. 2001. The Birds of Ohio.Wooster Book Company, Wooster, Ohio.

PIERCE, V., AND T. C. GRUBB, JR. 1981. Laboratorystudies of foraging in four bird species ofdeciduous woodland. Auk 98:307-320.

RISING, J. D. 1968. A multivariafe assessment ofinterbreeding between the chickadees Panisatricapillus and P. carolinensis. SystematicZoology 17:160-169.

RISING, J. D. 1983. The Great Plains hybrid zones.Pages 131-157 )H Current Ornithology, vol. 1(R. F. Johnston, Ed.). Plenum Press, NewYork.

ROBBINS, M. B., M. J. BRAUN, AND E. A. TOBEY.1986. Morphological and vocal variationacross a contact zone between the chicka-dees Parus atricapillus and P. carolinensis.Auk 103:655-666.

S.\TrLER, E. D. 1996. The dynamics of vocal,morphological, and molecular interactionbetween hybridizing Black-capped andCarolina chickadees. Ph.D. dissertation.University of Maryland, College Park.

SAITLER, G. D., AND M. J. BRAUN. 2000.Morphometric variation as an indicator ofgenetic interactions betw^een Black-cappedand Carolina chickadees af a contact zone

in the Appalachian Mountains. Auk 117:427-444.

SAWAYA, P L. 1990. A detailed analysis of thegenetic interaction at a hybrid zone betweenthe chickadees Parus atricapillus and P.carolinensis as revealed by nuclear and mito-chondrial DNA restriction fragment lengfhvariation. Ph.D. dissertation. University ofCincinnati, Cincinnafi, Ohio.

SMITH, S. M. 1991. The Black-capped Chickadee:Behavioral Ecology and Natural History.Comstock Publishing, Ithaca, New York.

SULLIVAN, K. A. 1984. The advantages of socialforaging in Downy Woodpeckers. AnimalBehaviour 32:16-22.

TANNER, J. T. 1952. Black-capped and Carolinachickadees in the southern AppalachainMountains. Auk 69:407^24.

TRAUIMAN, M. B. 1940. The Birds of BuckeyeLake, Ohio. Universify of Michigan Press,Ann Arbor.

TURELLI, M. 1998. Evolutionary genetics—Thecauses of Haldane's rule. Science 282:889-891.

WARD, R., AND D. A. WARD. 1974. Songs incontiguous populations of Black-cappedand Carolina chickadees in Pennsylvania.Wilson Bulletin 86:344-356.

WFISE, C. M., AND J. R. MEYFR. 1979. Juveniledispersal and development of site-fidelity inthe Black-capped Chickadee. Auk 96:40-55.

WETION, J. H., R. E. CARTIR, D. T. PARKIN, AND

D. WALTERS. 1987. Demographic study of awild House Sparrow population by DNAfingerprinting. Nature 327:147-149.

WHEATON, J. M. 1882. Report on the birds ofOhio. Ohio Geological Survey Bulletin 4:187-628.

Associate Editor: M. T. Murphy