[228]
The Condor 105:228?238
q The Cooper Ornithological Society 2003
MITOCHONDRIAL DNA PHYLOGEOGRAPHY OF THE BAY
WREN (TROGLODYTIDAE: THRYOTHORUS NIGRICAPILLUS)
COMPLEX
MARIBEL A. GONZA? LEZ1, JESSICA R. EBERHARD1,2, IRBY J. LOVETTE1,3, STORRS L. OLSON4
AND ELDREDGE BERMINGHAM1,5
1Smithsonian Tropical Research Institute, Naos Laboratories, Apartado 2072, Balboa, Republic of Panama
2Department of Biological Sciences and Museum of Natural Science, Louisiana State University,
Baton Rouge, LA 70803
3Cornell Laboratory of Ornithology, 159 Sapsucker Woods Rd, Ithaca, NY 14850
4Division of Birds, MRC 116, National Museum of Natural History, Smithsonian Institution,
Washington DC, 20560
Abstract. The Bay Wren (Thryothorus nigricapillus) is distributed from Costa Rica to
Ecuador and includes seven described subspecies, five of which occur in the Caribbean
lowlands of Panama. The subspecies vary in plumage characters, with particularly striking
differences between Bay Wrens from western Panama (to the north), and eastern Panama
(to the south). We surveyed mitochondrial DNA (mtDNA) sequence variation from a geo-
graphically broad sample of Bay Wrens and compared the phylogeographic structure of
mtDNA diversity with previously described patterns of morphological variation. The
mtDNA-based phylogeographic reconstructions revealed a basal split separating populations
in far eastern Panama and South America from those in central Panama through Costa Rica.
These two clades are concordant with previous morphology-based groupings of T. nigri-
capillus subspecies into the ??castaneus group?? (costaricensis, odicus, castaneus, and redi-
tus) and the ??nigricapillus group?? (schottii, connectens, and nigricapillus). Morphological
intergradation between the two groups takes place in central Panama, but all intergrades
possess the mtDNA haplotype of the castaneus group, suggesting that mitochondrial gene
flow is introgressing from west to east. In spite of the marked body size and plumage
variation present among subspecies of the castaneus group, mtDNA variation within this
group was low. At a deeper phylogenetic level, the mtDNA data support recognition of the
Riverside Wren, T. semibadius, as a full species. This taxon has sometimes been considered
conspecific with T. nigricapillus, but the high mtDNA divergence between these species is
consistent with previous suggestions that the morphological similarity results from conver-
gence in plumage traits.
Key words: Bay Wren, Panama isthmus, phylogeny, phylogeography, plumage, specia-
tion, Thryothorus nigricapillus.
Filogeograf??a del ADN Mitocondrial del Complejo de Thryothorus nigricapillus
Resumen. Thryothorus nigricapillus se distribuye desde Costa Rica hasta Ecuador e
incluye siete subespecies, de las cuales cinco se encuentran en las tierras bajas cariben?as de
Panama?. Las subespecies var??an en plumaje, con diferencias particularmente notables entre
Thryothorus nigricapillus del occidente de Panama? (hacia el norte), y aquellas del oriente
de Panama? (hacia el sur). Examinamos la variacio?n entre secuencias de ADN mitocondrial
(mtADN) de una muestra geogra?ficamente amplia de Thryothorus nigricapillus y compa-
ramos la estructura filogeogra?fica de la diversidad de mtADN con patrones previamente
descritos de variacio?n morfolo?gica. Las reconstrucciones filogeogra?ficas basadas en las se-
cuencias de mtADN revelaron una divisio?n basal entre las poblaciones del este de Panama?
y Sudame?rica, y las poblaciones que se encuentran desde el centro de Panama? hasta Costa
Rica. Estos dos clados corresponden a las agrupaciones previamente definidas con base en
caracteres morfolo?gicos, dividiendo las subespecies de T. nigricapillus en dos grupos: el
??grupo castaneus?? (costaricensis, odicus, castaneus y reditus) y el ??grupo nigricapillus??
(schottii, connectens y nigricapillus). Entre los dos grupos ocurre intergradacio?n morfolo?gica
en Panama? central, pero las formas intermedias tienen haplotipos de mtADN caracter??sticos
del grupos castaneus, sugiriendo que el flujo gene?tico mitocondrial es introgresivo de oeste
Manuscript received 29 July 2002; accepted 2 January 2003.
5 Corresponding author. E-mail: eb@naos.si.edu
BAY WREN PHYLOGEOGRAPHY 229
a este. A pesar de la notable variacio?n en taman?o corporal y plumaje entre las subespecies
del grupo castaneus, la variacio?n de mtADN dentro de este grupo fue baja. A un nivel
filogene?tico ma?s profundo, los datos de mtADN apoyan el reconocimiento de T. semibadius
como especie. Este taxo?n ocasionalmente ha sido considerado coespec??fico con T. nigrica-
pillus, pero la marcada divergencia a nivel de mtADN entre estas especies es consistente
con previas sugerencias de que la semejanza morfolo?gica es resultado de convergencia en
caracteres del plumaje.
INTRODUCTION
The location and dynamic geological history of
the Isthmus of Panama renders it a region of
particular interest for the study of genetic vari-
ation in Neotropical birds. The best-known ef-
fect of the Pliocene formation of the Panama-
nian landbridge was the mixing of distinct North
and South American faunas in a process termed
the Great American Biotic Interchange (Simpson
1980, Marshall 1988, Coates and Obando 1996,
Webb 1997). In addition, the complex geograph-
ical development of the isthmian region has fos-
tered intraspecific phenotypic and genetic vari-
ation among indigenous taxa (Peterson et al.
1992, Seutin et al. 1993, Bermingham and Mar-
tin 1998). One such case is the Bay Wren
(Thryothorus nigricapillus), with five of seven
described subspecies occurring in the Caribbean
lowlands of Panama (Wetmore 1959, Paynter
1960, Wetmore et al. 1984).
The Bay Wren complex is of particular bio-
geographic interest because of the striking plum-
age variation among subspecies, their uncertain
distributional boundaries, and the presence of a
possible zone of secondary contact in east-cen-
tral Panama. Subspecies at the northern (T. n.
costaricensis) and southern (T. n. nigricapillus)
extremes of the range are very different in ap-
pearance from one another, and the seven sub-
species have sometimes been divided into two
groups (AOU 1998).
Here we investigate the magnitude and geo-
graphic structure of mitochondrial DNA
(mtDNA) variation among Panamanian and
Ecuadorian subspecies of the Bay Wren. Genetic
comparisons permit an independent assessment
of population divergence, one that allows strong
inference regarding the phylogenetic basis of
phenotypic variation. If the magnitude of plum-
age differentiation is roughly proportional to the
time populations have been isolated, the phe-
notypic differences that distinguish geographic
populations of Bay Wren would indicate sub-
stantial periods of separation and predict com-
mensurate levels of DNA nucleotide substitution
and phylogenetic separation. Alternatively, the
absence of significant phylogenetic divergence
would suggest that plumage differences among
subspecies of Bay Wren are more probably due
to strong selective pressures or a small number
of alleles with large phenotypic effect (e.g., Om-
land and Lanyon 2000). In either case, our
mtDNA-based analysis of Bay Wrens provides
a phylogeographic context for future compara-
tive studies of intraspecific variation in isthmian
birds.
METHODS
STUDY TAXA
The Bay Wren has a broad distribution in Cen-
tral America and northern South America, with
a range that extends from the Caribbean slope
of Nicaragua south through Costa Rica and Pan-
ama into western Colombia and western Ecua-
dor (Fig. 1). Throughout its range, the Bay Wren
is a generally common resident of forests and
second-growth borders, where it is typically
found along streams and roadsides in the tropi-
cal and lower subtropical habitat zones (Wet-
more et al. 1984, Stiles and Skutch 1989).
Differences in overall plumage types have led
to the somewhat oversimplified recognition of
two groups of subspecies: a northern ??castaneus
group?? (T. n. costaricensis, T. n. odicus, T. n.
castaneus, and T. n. reditus) and a southern ??ni-
gricapillus group?? (T. n. schottii, T. n. connec-
tens, and T. n. nigricapillus; AOU 1998).
Thryothorus n. costaricensis and T. n. castaneus
together were given the specific rank of T. cas-
taneus early in the last century (Ridgway 1904)
but have subsequently been considered conspe-
cific with T. nigricapillus. Plumage and size dif-
ferences are subtler among the races within each
of these two main groups; Wetmore (1959) and
Wetmore et al. (1984) treated this morphological
variation in detail. The T. nigricapillus subspe-
cies distributions presented in Figure 1 are based
on the examination of 281 specimens listed in
the Appendix.
230 MARIBEL A. GONZA? LEZ ET AL.
FIGURE 1. Geographic distribution of subspecies of Thryothorus nigricapillus in Panama. Dots indicate sam-
pling locations; numbers indicate sequenced individuals listed in Table 1. Subspecies ranges are based on the
specimens examined in the Appendix; ????? denotes regions where the boundaries between subspecies are not
known. An X marks the type locality of Thryothorus n. reditus, drawing attention to the nomenclatural problem
referred to in the Appendix. Inset depicts the entire Central and South American range of T. nigricapillus,
including the two subspecies not found in Panama: T. n. connectens of southwest Colombia (not sampled) and
T. n. nigricapillus of western Ecuador. The country of Panama is boxed on the inset map.
To avoid potential confusion over the use of
??nigricapillus?? to describe a subspecies, a sub-
species complex, and a species, we hereafter use
??T. n. nigricapillus?? when referring to the Ecua-
dorian subspecies alone; ??nigricapillus group??
when referring to the subspecies complex com-
posed of T. n. nigricapillus, T. n. connectens,
and T. n. schottii; and ??Bay Wren?? or ??T. ni-
gricapillus?? when referring collectively to all
forms of the species.
In addition to the uncertain relationships
among the seven subspecies of the Bay Wren,
the Riverside Wren (T. semibadius) of south-
western Costa Rica has variously been consid-
ered conspecific with T. nigricapillus (Hellmayr
1934, Paynter 1960), placed in a superspecies
complex with T. nigricapillus (Sibley and Mon-
roe 1990), or given full species status (e.g., Wet-
more 1959, Wetmore et al. 1984, AOU 1998).
To test these hypotheses of relationship, we in-
cluded a sample of T. semibadius in our recon-
structions. Three specimens of T. rutilus (Ru-
fous-breasted Wren) collected in central Panama
provided an unambiguous outgroup for the phy-
logenetic analyses (Table 1).
MITOCHONDRIAL DNA ANALYSIS
Samples were obtained from the tissue collec-
tions of the U.S. National Museum of Natural
History and the Museum of Natural Science at
Louisiana State University, and the Academy of
Natural Sciences, Philadelphia (Table 1). Total
cellular DNA was extracted by phenol-chloro-
form extraction, followed by dialysis. We used
the polymerase chain reaction (PCR) to amplify
a 1040-base pair (bp) fragment of the mitochon-
BAY WREN PHYLOGEOGRAPHY 231
TABLE 1. Geographic locations, museum specimen accession numbers, tissue identification numbers, and
GenBank DNA sequence accession numbers for Thryothorus nigricapillus and outgroup Thryothorus species
analyzed for a study of Bay Wren phylogeography. Museum abbreviations are as follows: ANSP, Academy of
Natural Sciences, Philadelphia, Pennsylvania; LSU, Louisiana State University Museum of Natural Science,
Baton Rouge, Louisiana; STRI, Smithsonian Tropical Research Institute, Balboa, Panama; USNM, U.S. National
Museum of Natural History, New York. Localities are in Panama except where another country is listed.
Sam-
ple Taxon Museum
Specimen
accession
no.
Tissue
catalog
no.
GenBank
accession
no.
Collection location
(Province; locality)
1 T. n. costaricensis USNM 612420 B00494 AY103289 Bocas del Toro; Tierra
Oscura, mainland S
of Isla San Cristo?bal
2 T. n. costaricensis USNM 607004 B00302 AY103287 Bocas del Toro; Isla San
Cristo?bal, Bocatorito
3 T. n. costaricensis USNM 607005 B00305 AY103288 Bocas del Toro; Isla San
Cristobal, Bocatorito
4 T. n. costaricensis USNM 614082 B01986 AY103284 Bocas del Toro; Chiri-
qu?? Grande
5 T. n. costaricensis USNM 605408 B01745 AY103282 Bocas del Toro; Isla
Bastimentos, Old
Point
6 T. n. costaricensis USNM 605409 B01762 AY103283 Bocas del Toro; Isla
Bastimentos, Old
Point
7 T. n. costaricensis USNM 607871 B01103 AY103279 Bocas del Toro; Cayo
Agua
8 T. n. costaricensis USNM 607879 B01249 AY103280 Bocas del Toro; Pen??n-
sula Valiente
9 T. n. costaricensis USNM 607876 B01250 AY103281 Bocas del Toro; Pen??n-
sula Valiente
10 T. n. odicus USNM 613500 B01028 AY103277 Bocas del Toro; Isla Es-
cudo de Veraguas
11 T. n. odicus USNM 613508 B01029 AY103278 Bocas del Toro; Isla Es-
cudo de Veraguas
12 T. n. castaneus LSU 164291 B28552 AY103285 Colo?n; Achiote Road at
Rio Providencia
13 T. n. castaneus LSU 164292 B28559 AY103286 Colo?n; Achiote Road at
Rio Providencia
14 T. n. reditus LSU 163696 B26392 AY103290 Panama?; Serran??a de
San Blas; Chepo
15 T. n. reditus LSU 163697 B26393 AY103291 Panama?; Serran??a de
San Blas; Chepo
16 T. n. schottii LSU 108537 B2269 AY103292 Darie?n; Cana on E
slope Cerro Pirre
17 T. n. schottii LSU 108536 B2272 AY103293 Darie?n; Cana on E
slope Cerro Pirre
18 T. n. nigricapillus ANSP 180437 LSU B12047 AY103294 Ecuador, Pinchicha;
Mindo, 1300 m
19 T. n. nigricapillus ANSP 180436 LSU B12053 AY103295 Ecuador, Pinchicha;
Mindo, 1300 m
20 T. semibadius LSU 138767 B16101 AY103273 Costa Rica, Puntarenas;
Rio Copey, 4 km E
Jaco?
21 T. rutilus LSU 163699 B26902 AY103274 Panama?; Old Gamboa
Road
22 T. rutilus LSU 163700 B26903 AY103275 Panama?; Old Gamboa
Road
23 T. rutilus STRI ? PA-THU66 AY103276 Panama?; Old Gamboa
Road
232 MARIBEL A. GONZA? LEZ ET AL.
drial genome from all individuals. The primer
pair CO2GQL and CO3HMH (Bermingham
2003) was used to amplify a region spanning the
full tRNALys, ATPase 8, and ATPase 6 genes.
PCR reaction components and conditions are de-
tailed in Hunt et al. (2001). Amplification prod-
ucts were cleaned and checked via electropho-
resis on low-melting-point agarose gels. We then
conducted Dyedeoxy terminator cycle sequenc-
ing reactions (Applied Biosystems Division of
Perkin Elmer, Inc., Fullerton, California). We se-
quenced partially overlapping portions of the
light strand of the ATPase region with the prim-
ers CO2GQL, A8PWL, and a segment of the
heavy strand with primer CO3HMH (Berming-
ham 2003). These products were then electro-
phoresed in an Applied Biosystems model 377
automated DNA sequencer.
Genetic data analyses. Sequenced fragments
obtained using the three primers were aligned
and proofread using Sequencher 3.1 (Gene
Codes Corporation 1998). All products used in
the analyses presented here had very clean chro-
matograms with a high signal-to-noise ratio, no
conspicuous double peaks, and no confounding
sequencing artifacts. Phylogenetic analyses were
based on the 842-bp coding sequence of the
complete ATPase 6 and ATPase 8 genes. These
genes have a frame-shift overlap of 10 bp; this
short region of overlap did not vary among our
taxa and was excluded from analyses parame-
terized by codon position. Genetic distances
among haplotypes were estimated using
PAUP*4.0b10 (Swofford 2002); distances re-
ported here are uncorrected percent divergence.
Phylogenetic reconstructions among all haplo-
types were generated via a maximum likelihood
(ML) approach using the Markov chain Monte
Carlo method (Steel 1994) implemented in
MrBayes (Huelsenbeck and Ronquist 2001).
This analysis employed the general time-revers-
ible model (nst 5 6), with site-specific rate var-
iation partitioned by codon. Four chains were
run for 500 000 generations and sampled every
1000 generations. Inspection of the resulting ML
scores suggested that parameter and likelihood
stationarity was reached by 10 000 generations;
we discarded the topologies sampled from the
first 20 000 generations. A majority-rule consen-
sus of the remaining 480 sampled trees was gen-
erated in PAUP* to provide a phylogenetic hy-
pothesis with associated posterior probability
values for internal branches.
For comparison, distance- and parsimony-
based reconstructions were also conducted using
PAUP* and a variety of distance metrics and
character weighting methods. As all highly sup-
ported (100% posterior probability) nodes in the
ML tree were also invariably present in the
neighbor-joining and maximum parsimony re-
constructions, only the results of the ML anal-
ysis are presented in detail here.
Phylogenetic reconstructions of Bay Wrens
identified a clade of very closely related haplo-
types representing the castaneus group, which
included all individuals sampled from central
and western Panama. To examine the relation-
ships of these castaneus group haplotypes in
more detail, we generated a gene genealogy us-
ing the program TCS (Clement et al. 2000). This
network-based approach enhances the interpre-
tation of the relationships between closely allied
haplotypes because it permits extant haplotypes
to be derived from other extant haplotypes.
RESULTS
We obtained the complete sequence (842 nucle-
otides) of the ATPase 8 and ATPase 6 coding
regions from a total of 23 samples of Thryot-
horus, including 17 T. nigricapillus sampled
along a geographic transect spanning western
Bocas del Toro province through the eastern Da-
rie?n lowlands of Panama and two T. nigricapi-
llus from Ecuador (Fig. 1). We also included in
our analyses four samples representing two other
species of Thryothorus (Table 1). All sequences
have been accessioned in GenBank (see Table 1
for accession numbers). A total of 194 nucleo-
tide sites varied among all 23 samples, and a
total of 74 sites varied among the 17 unique hap-
lotypes of T. nigricapillus.
The ATPase haplotypes of T. nigricapillus
differed from those of Thryothorus semibadius
and T. rutilus by 90?120 substitutions (10.7?
14.3%). MtDNA distances between T. nigrica-
pillus and T. semibadius (mean 5 11.4%), two
taxa sometimes considered conspecific (Hell-
mayr 1934, Paynter 1960), were nearly equiva-
lent to the mean distance (13.5%) between T.
nigricapillus and the morphologically more dis-
tinct T. rutilus. Divergence among mtDNA hap-
lotypes within the Bay Wren complex was mod-
est, with a maximum difference of 5.6%.
A deep basal bifurcation separated the 19 Bay
Wren samples into two clades whose geographic
ranges appear to meet in east-central Panama
BAY WREN PHYLOGEOGRAPHY 233
FIGURE 2. Phylogenetic relationships among individuals of Thryothorus based on the mitochondrial ATPase
6 and 8 genes. Individuals are numbered as in Table 1 and Figure 1. The consensus tree pictured represents the
maximum likelihood analysis of mtDNA haplotypes. Numbers above branches indicate all posterior probability
values .90%, and numbers below branches indicate uncorrected genetic distances based on mtDNA sequence
divergence.
FIGURE 3. A genealogical network for the mtDNA
haplotypes observed in the castaneus group, which in-
cludes the four Thryothorus nigricapillus subspecies
from western and central Panama (castaneus, costari-
censis, odicus, and reditus). Shaded circles indicate in-
ferred ancestral haplotypes at network nodes; bars in-
dicate the number of nucleotide substitutions between
mtDNA haplotypes (e.g., individuals 2 and 3 differ at
one nucleotide site, individuals 12 and 14 at two sites).
(Fig. 2). These clades correspond to the casta-
neus and nigricapillus subspecies groups (AOU
1998), and haplotypes from the two groups dif-
fered by 34?47 substitutions (4.0?5.6%). In the
phylogenetic analysis, support for the reciprocal
monophyly of these two haplotype groups was
high (Fig. 2).
The castaneus group was represented by 13
unique haplotypes recovered from the 15 rep-
resentatives of the subspecies costaricensis, odi-
cus, castaneus, and reditus. The maximum di-
vergence among the mtDNA haplotypes of the
castaneus group was 10 nucleotide substitutions
(1.2%). The haplotype network depicting the re-
lationship of these very closely related haplo-
types is shown in Figure 3. Although individuals
sampled at particular locations tended to have
similar haplotypes, there was little evidence of
strong geographic structuring within this clade.
For example, an identical mtDNA haplotype was
recovered from individuals representing the sub-
species costaricensis and reditus sampled at
sites approximately 330 km apart (samples 8, 9,
and 14 in Table 1 and Fig. 1). Thryothorus n.
odicus, the best-differentiated subspecies in our
sample of the castaneus complex, had a mean
genetic distance of only 0.9% from the remain-
der of the subspecies in this group.
The nigricapillus group was represented by
two samples of T. n. schottii from eastern Darie?n
234 MARIBEL A. GONZA? LEZ ET AL.
province and two T. n. nigricapillus from Ec-
uador. Haplotypes from these two widely sepa-
rated localities differed by 24?27 substitutions
(2.9?3.2%), suggesting substantial geographi-
cally structured diversity within the nigricapillus
group. Additional samples, particularly from
Colombian populations of T. n. connectens,
would be required to fully evaluate the pattern
and magnitude of phylogeographic diversity
within the South American representatives of
this complex.
DISCUSSION
PHYLOGEOGRAPHIC VARIATION IN
MAINLAND POPULATIONS OF THE BAY WREN
The most striking of our results was the large
(4.0?5.6%) mtDNA divergence in pairwise
comparisons between haplotypes representing
Bay Wrens from far eastern Panama and South
America (nigricapillus group) and those from
central Panama through Costa Rica (castaneus
group). This is concordant with the previous rec-
ognition of two distinct subspecies groups based
on morphological criteria (AOU 1998). The de-
gree of mtDNA sequence divergence within the
Bay Wren equals or surpasses that observed be-
tween many avian sister species (e.g., Berming-
ham et al. 1992, Klicka and Zink 1997, Lovette
et al. 1998). It is also similar in magnitude to
the genetic differentiation among conspecific
populations of freshwater fish distributed across
the same region (Bermingham and Martin 1998,
Perdices et al. 2002), suggesting that the same
historical factors were responsible for the diver-
gences within the Bay Wren and within isthmian
freshwater fish. Bermingham and Martin (1998)
posited that changes in sea level from the end
of the Miocene through the early Pliocene might
have caused several episodes of regional isola-
tion across the rising isthmus, thus setting the
stage for allopatric differentiation. The magni-
tude of mtDNA divergence between the casta-
neus and nigricapillus groups is consistent with
the expansion of the common ancestor of these
groups across the newly formed Panamanian
isthmus approximately three million years ago
(Coates and Obando 1996).
Although there is concordance between mor-
phology and mtDNA at the level of the two ma-
jor subspecies groups, within each group there
is discordance. Plumage differences between
schottii and nigricapillus are not striking (see
Appendix), yet these two subspecies differ by
roughly 3% mtDNA sequence divergence. With-
in the castaneus group, however, we found a re-
verse pattern in which subspecies showing
marked differences in body size or plumage pat-
tern were often indistinguishable on the basis of
their mtDNA haplotypes. There is extensive var-
iation in plumage traits from west to east in the
subspecies costaricensis, castaneus, and reditus,
with breast color changing from dark chestnut
to white and ventral barring changing from ab-
sent or indistinct to extensive (Wetmore et al.
1984). Yet these three subspecies show only
slight differences in mtDNA sequence (Fig. 2,
3).
Phenotypically, the two subspecies from cen-
tral Panama, T. n. castaneus and T. n. reditus,
are intergrades between T. n. costaricensis and
T. n. schottii. The range occupied by castaneus
and reditus could thus be considered a pheno-
typic hybrid zone in which no pure parental
types occur, much as is seen, although over a
much more limited area, in the northern and
southern forms of the Variable Seedeater (Spo-
rophila americana; Olson 1981). Although it is
most like costaricensis, castaneus shows the in-
fluence of nigricapillus/schottii in the paler
chestnut of the underparts, more extensive white
throat, and greater frequency and extent of black
ventral barring. The populations that Wetmore
treated under the name T. n. reditus (Appendix)
are extremely variable, but have much more
white in the underparts and are always more
heavily barred below than in castaneus. All re-
tain considerable amounts of chestnut coloration
in the underparts, which is not expressed in
schottii. The demarcation between these two
populations of intergrades is quite sharp, with
castaneus being known from Cocle? to the vicin-
ity of the Panama Canal, and reditus occurring
in the headwaters of the R??o Chagres and east-
ward. One specimen from the former Canal
Zone at R??o Frijoles is referable to reditus. The
point of contact between castaneus and reditus
is extremely similar to that between the distinc-
tive northern and southern forms of the Buff-
rumped Warbler (Basileuterus fulvicauda; Wet-
more et al. 1984).
The very abrupt transition from castaneus to
reditus in the narrowest part of the Panamanian
isthmus suggests that the two isolates might
have come into contact at this point. In this sce-
nario, mtDNA and presumably nuclear genes
have flowed eastward toward South America to
BAY WREN PHYLOGEOGRAPHY 235
produce the highly variable intergrades now
called reditus. Morphological characters indicate
that some nigricapillus nuclear genes have ap-
parently introgressed in the opposite direction,
giving rise to what is now recognized as casta-
neus. Nevertheless, the mtDNA evidence sug-
gests that the zone of contact between the cas-
taneus group and the nigricapillus group occurs
not in the canal region but farther to the east on
the Caribbean slope.
Differential introgression of plumage and
mtDNA could result, for example, from female
preference for males with dark breast barring on
a white background. Analogous patterns of
plumage introgression possibly driven by female
choice have been well documented in a Mana-
cus hybrid zone in western Panama (Parsons et
al. 1993, Brumfield et al. 2001).
The partial discordance between patterns of
genetic and morphological variation in the Bay
Wren is intriguing given the absence of major
geographic disjunctions in this species? contin-
uous continental distribution. The vocal behav-
ior of the subspecies castaneus has been studied
in central Panama, and vocal duetting has been
shown to be important in the contexts of sexual
selection and social competition (Levin 1996a,
1996b), both of which could be linked with the
origin and maintenance of the subspecific mor-
phological diversity. Further sampling and be-
havioral studies, particularly in the zone where
reditus and schottii meet, are key to understand-
ing the mechanisms underlying the patterns de-
scribed here.
DIFFERENTIATION IN THE ISLANDS OF
BOCAS DEL TORO
The most pronounced genetic difference within
the castaneus subspecies group, supported by a
single mtDNA synapomorphy, was observed be-
tween the insular odicus and related mainland
forms. The subspecies T. n. odicus, confined to
Isla Escudo de Veraguas, is characterized by its
much larger size (9?16%) and paler chestnut of
the underparts (Wetmore 1959). Two other en-
demic subspecies of birds on Escudo de Vera-
guas are also much larger than their counterparts
on the adjacent mainland: the Rufous-tailed
Hummingbird (Amazilia tzacatl handleyi) and
the Golden-collared Manakin (Manacus vitelli-
nus amitinus). In addition, a larger, heavier bill
distinguishes the Blue-gray Tanager (Thraupis
episcopus caesitia) on the island from its main-
land form. Escudo de Veraguas has been isolated
from the mainland longer than any of the other
islands of Bocas del Toro, about 8900 years, al-
though this separation may simply be the last in
a series of isolating events caused by fluctuating
Pleistocene sea levels (Summers et al. 1997). An
endemic species of sloth (Anderson and Handley
2001) provides an additional measure of the rel-
ative isolation of Isla Escudo de Veraguas,
which appears to be reflected in the mtDNA se-
quence divergence (0.06?1.2%) between odicus
and the remainder of the castaneus group. None-
theless, the genetic divergence of odicus only
slightly exceeds nucleotide diversity in the cas-
taneus group, and the overall similarity of
mtDNA haplotypes is revealed in the genealog-
ical network presented in Figure 3. The maxi-
mum mtDNA divergence between odicus and
related subspecies in the castaneus group is al-
most three times lower than observed between
subspecies in the nigricapillus group.
We detected no morphological or genetic dif-
ferences among individuals of costaricensis
from three islands in Chiriqu?? Lagoon or be-
tween the islands and the adjacent mainland of
Bocas del Toro. These islands were separated
from the mainland in the mid- to late Holocene
from 4700 to 1000 years ago (Anderson and
Handley 2001), so either insufficient time has
passed to accumulate measurable phenotypic
and mtDNA differences among populations, or
dispersal has been sufficient to prevent differ-
entiation. It is interesting to note that with the
exception of House Wrens (Troglodytes aedon)
around the artificial pasturelands of Isla San
Cristobal, the Bay Wren is the only one of a
diverse fauna of eight species of wren in low-
land Bocas del Toro that occurs on the islands.
RELATIONSHIPS OF T. SEMIBADIUS AND T.
NIGRICAPILLUS
The Riverside Wren (Thryothorus semibadius)
of the Pacific lowlands of southwestern Costa
Rica and extreme western Panama was treated
as a subspecies of T. nigricapillus by Hellmayr
(1934) and Paynter (1960). However, Wetmore
(1959) and Slud (1964) argued for specific sep-
aration and Wetmore et al. (1984:94) noted that
T. semibadius differed ??definitely and complete-
ly?? from all of the forms of T. nigricapillus. It
is smaller, has a crown the same color as the
back, and its barring results from feathers with
three black bars, whereas the ventral feathers of
236 MARIBEL A. GONZA? LEZ ET AL.
nigricapillus have two black bars, suggesting
that these taxa may have evolved black barring
independently. Nevertheless, the two continue to
be considered a superspecies (AOU 1998).
The only apparent reason for regarding T. se-
mibadius and T. nigricapillus as close relatives
is their allopatric distribution combined with
barred underparts, although the barred forms of
T. nigricapillus are those farthest removed geo-
graphically from T. semibadius. If geography
and barring are the criteria, then T. semibadius
shares an allopatric distribution with other con-
geners with barred underparts such as the Band-
ed Wren (T. pleurostictus), which occurs from
Mexico to northwestern Costa Rica, also in Pa-
cific lowlands. It would perhaps fit even better
into the disjunct range of the Spot-breasted Wren
(T. maculipectus), with subspecies ranging from
Mexico to Costa Rica and with supposed outli-
ers T. m. paucimaculatus and T. m. sclateri (the
most similar to T. semibadius) in Ecuador and
Peru. These last have also been treated as sub-
species of T. rutilus (e.g., Hellmayr 1934).
Our mtDNA-based analyses identified 11.4%
sequence divergence between semibadius and
members of the Bay Wren clade, and therefore
strongly support the recognition (Wetmore et al.
1984, AOU 1998) of semibadius as a species-
level taxon distinct from T. nigricapillus. Owing
to the fact that we have only sampled three of
the 27 species of Thryothorus (AOU 1998), we
can only speculate based on geographic and phe-
notypic considerations and level of mtDNA di-
vergence that T. semibadius and T. nigricapillus
probably do not have a sister-group or supers-
pecies-level relationship. Clearly there is much
yet to be learned about relationships within this
complex genus of wrens.
ACKNOWLEDGMENTS
This project was initially inspired by conversations in
the field with Donna Dittmann and Steve Cardiff after
collecting several T. n. schottii, and we thank them for
their suggestions and insights. We thank the U.S. Na-
tional Museum of Natural History, the Louisiana State
University Museum of Natural Science Collection of
Genetic Resources, and the Academy of Natural Sci-
ences, Philadelphia, for granting us loans of tissue
samples from their holdings. We are very grateful to
the countries of Panama, Costa Rica, and Ecuador for
approving and supporting the collection of these sam-
ples. We thank J. Hunt for his expert technical assis-
tance, Deborah Siegel and Grethel Grajales for map-
ping the geographical localities of Bay Wrens in Pan-
ama, and Douglas Causey for lending specimens from
the Museum of Comparative Zoology, Harvard Uni-
versity, Cambridge, Massachusetts. The Smithsonian
Institution Molecular Systematics Program and a
Smithsonian Tropical Research Institute short-term fel-
lowship to MAG funded this research.
LITERATURE CITED
AMERICAN ORNITHOLOGISTS? UNION. 1998. Check-list
of North American Birds. 8th ed. American Or-
nithologists? Union, Washington, DC.
ANDERSON, R. P., AND C. O. HANDLEY JR. 2001. A new
species of three-toed sloth (Mammalia: Xenarthra)
from Panama?, with a review of the genus Brady-
pus. Proceedings of the Biological Society of
Washington 114:1?33.
BERMINGHAM, E. [ONLINE]. 2003. Bermingham lab
home page. ,http://nmg.si.edu/bermlab/berm-
lab.htm. (30 January 2003).
BERMINGHAM, E., AND A. P. MARTIN. 1998. Compara-
tive mtDNA phylogeography of Neotropical
freshwater fish: testing shared history to infer the
evolutionary landscape of lower Central America.
Molecular Ecology 7:499?517.
BERMINGHAM, E., S. ROWHER, S. FREEMAN, AND C.
WOOD. 1992. Vicariance biogeography in the
Pleistocene and speciation in North American
wood warblers: a test of Mengel?s model. Pro-
ceedings of the National Academy of Sciences 89:
6624?6628.
BRUMFIELD, R. T., R. W. JERNIGAN, D. B. MCDONALD,
AND M. J. BRAUN. 2001. Evolutionary implica-
tions of divergent clines in an avian (Manacus:
Aves) hybrid zone. Evolution 55:2070?2087.
CLEMENT, M., D. POSADA, AND K. CRANDALL. 2000.
TCS: a computer program to estimate gene ge-
nealogies. Molecular Ecology 9:1657?1660.
COATES, A. G., AND J. A. OBANDO. 1996. The geologic
evolution of the Central American Isthmus, p. 21?
56. In J. B. C. Jackson, A. F. Budd, and A. G.
Coates [EDS.], Evolution and environment in trop-
ical America. University of Chicago Press, Chi-
cago.
GENE CODES CORPORATION. 1998. Sequencher, v. 4.1.
Gene Codes Corporation, Ann Arbor, MI.
GRISCOM, L. 1932. The ornithology of the Caribbean
coast of extreme eastern Panama. Bulletin of the
Museum of Comparative Zoology 72:303?372.
HELLMAYR, C. E. 1934. Catalogue of the birds of the
Americas, part 7. Zoological Series, Vol. 13. Field
Museum of Natural History, Chicago.
HUELSENBECK, J. P., AND F. R. RONQUIST. 2001. MR-
BAYES: Bayesian inference of phylogeny. Bioin-
formatics 17:754?755.
HUNT, J. S., E. BERMINGHAM, AND R. E. RICKLEFS.
2001. Molecular systematics and biogeography of
Antillean thrashers, tremblers, and mockingbirds
(Aves: Mimidae). Auk 118:35?55.
KLICKA, J., AND R. M. ZINK. 1997. The importance of
recent ice ages in speciation: a failed paradigm.
Science 277:1666?1669.
LEVIN, R. N. 1996a. Song behaviour and reproductive
strategies in a duetting wren, Thryothorus nigri-
capillus: I. Removal experiments. Animal Behav-
iour 52:1093?1106.
BAY WREN PHYLOGEOGRAPHY 237
LEVIN, R. N. 1996b. Song behaviour and reproductive
strategies in a duetting wren, Thryothorus nigri-
capillus: II. Playback experiments. Animal Behav-
iour 52:1107?1117.
LOVETTE, I. J., E. BERMINGHAM, G. SEUTIN, AND R. E.
RICKLEFS. 1998. Evolutionary differentiation in
three endemic West Indian warblers. Auk 115:
890?903.
MARSHALL, L. G. 1988. Land mammals and the Great
American Interchange. American Scientist 76:
380?388.
OLSON, S. L. 1981. The nature of the variability in the
Variable Seedeater in Panama (Sporophila ame-
ricana: Emberizinae). Proceedings of the Biolog-
ical Society of Washington 94:380?390.
OMLAND, K. E., AND S. M. LANYON. 2000. Recon-
structing plumage evolution in orioles (Icterus):
repeated convergence and reversal in patterns.
Evolution 54:2119?2133.
PARSONS, T. J., S. L. OLSON, AND M. J. BRAUN. 1993.
Undirectional spread of secondary sexual plumage
traits across an avian hybrid zone. Science 260:
1643?1646.
PAYNTER, R. A. 1960. Family Troglodytidae, p. 379?
440. In E. Mayr and J. C. Greenway Jr. [EDS.],
Check-list of birds of the world. Vol. IX. Museum
of Comparative Zoology, Cambridge, MA.
PAYNTER, R. A., JR. 1993. Ornithological gazetteer of
Ecuador. 2nd ed. Museum of Comparative Zool-
ogy, Cambridge, MA.
PAYNTER, R. A., JR. 1997. Ornithological gazetteer of
Colombia. 2nd ed. Museum of Comparative Zo-
ology, Cambridge, MA.
PERDICES, A., E. BERMINGHAM, A. MONTILLA, AND I.
DOADRIO. 2002. Evolutionary history of the genus
Rhamdia (Teleostei: Pimelodidae) in Central
America. Molecular Phylogenetics and Evolution
25:172?189.
PETERSON, A. T., P. ESCALANTE, AND A. NAVARRO S.
1992. Genetic variation and differentiation in
Mexican population of Common Bush-Tanagers
and Chestnut-Capped Brush-Finches. Condor 94:
244?253.
RIDGWAY, R. 1904. The birds of North and Middle
America, part III. Bulletin of the United States
National Museum No. 50.
SEUTIN, G., J. BRAWN, R. E. RICKLEFS, AND E. BER-
MINGHAM. 1993. Genetic divergence among pop-
ulations of a tropical passerine, the Streaked Sal-
tator (Saltator albicollis). Auk 110:117?126.
SIBLEY, C., AND B. L. MONROE JR. 1990. Distribution
and taxonomy of birds of the world. Yale Univer-
sity Press, New Haven, CT.
SIMPSON, G. G. 1980. Splendid isolation: the curious
history of South American mammals. Yale Uni-
versity Press, New Haven, CT.
SLUD, P. 1964. The birds of Costa Rica. Distribution
and ecology. Bulletin of the American Museum of
Natural History 128:1?430.
STEEL, M. 1994. Recovering a tree from the Markov
model. Applied Mathematics Letters 7:19?23.
STILES, F. G., AND A. F. SKUTCH. 1989. A guide to the
birds of Costa Rica. Comstock Publishing Asso-
ciates, Ithaca, NY.
SUMMERS, K., E. BERMINGHAM, L. WEIGT, S. MC-
CAFFERTY, AND L. DAHLSTROM. 1997. Phenotypic
and genetic divergence in three species of dart-
poison frogs with contrasting parental behavior.
Journal of Heredity 88:8?13.
SWOFFORD, D. L. 2002. PAUP* Phylogenetic analysis
using parsimony (*and other methods) version
4.0b10. Sinauer Associates, Sunderland, MA.
UNITED STATES BOARD ON GEOGRAPHIC NAMES. 1990.
Gazetteer of Panama. Defense Mapping Agency,
Washington, DC.
WEBB, D. 1997. The Great American Faunal Inter-
change, p. 197?222. In A. G. Coates [ED.], Central
America: a natural and cultural history. Yale Uni-
versity Press, New Haven, CT.
WETMORE, A. 1959. The birds of Isla Escudo de Ve-
raguas, Panama?. Smithsonian Miscellaneous Col-
lections 139(2):1?27.
WETMORE, A., R. F. PASQUIER, AND S. L. OLSON. 1984.
The birds of the Republic of Panama. Smithsonian
Institution Press, Washington, DC.
APPENDIX
Bay Wren specimens examined for this study and a
nomenclatural note regarding the name Thryothorus n.
reditus. All specimens are from the National Museum
of Natural History, Smithsonian Institution, Washing-
ton, DC (USNM), except as marked from the Museum
of Comparative Zoology, Harvard University, Cam-
bridge, Massachusetts (MCZ). Place names are stan-
dardized according to the following sources: Costa
Rica (Slud 1964), Panama (USBGN 1990), Colombia
(Paynter 1997), Ecuador (Paynter 1993).
Thryothorus nigricapillus costaricensis (n 5 69)
Nicaragua: R??o Escondido (2); Greytown (2); Los
Sabalos (1). Costa Rica: Alajuela: R??o Frio (3); Car-
tago: Bonilla (1); Pacuare (2, including Paqua [sic]);
Reventazo?n (1); Volca?n Turrialba (1); Limon: Jime?nez
(6); Siquirres, 16 km south (1); San Jose?: Guayabo 5
Guayabal? (1); San Jose? (2). Panama: Bocas del Toro:
R??o Changuena, 727 m (1); Boca del Drago, Chan-
guinola Canal (1); Almirante, Milla 2 (1); R??o Oeste
(7); Valle de Agua (3); Tierra Oscura (2); Isla Basti-
mentos (9); Cayo Nancy (5); Isla San Cristo?bal (4);
Isla Popa (4); Cayo Agua (4); Cayo Patterson (1); Pla-
ya Verde, Pen??nsula Valiente (1); Punta Alegre, Pen-
??nsula Valiente (4).
Thryothorus nigricapillus odicus (n 5 16)
Panama: Bocas del Toro: Isla Escudo de Veraguas
(16, including holotype).
Thryothorus nigricapillus castaneus (n 5 51)
Panama: Cocle?: Boca de Uracillo (6); Cascajal (2);
Cerro La India Dormida (2); Tigre, head of R??o Guabal
(3); El Valle, head of R??o Mata Ahogada (2). Colo?n:
El Chilar, R??o Indio (5); R??o Membrillar, R??o Indio (3);
Chilar, Quebrada Seren?a (1); Chilar, Quebrada Torno
Rompido (3); Colo?n (former Panama Canal Zone): Ga-
tu?n (9); R??o Indio, near Gatu?n (7); R??o Indio, near
mouth (2); Frijoles (1); Lion Hill (2); Marajal (1);
Monte Lirio, R??o Gatu?n (1); Pipeline Road (1).
238 MARIBEL A. GONZA? LEZ ET AL.
Thryothorus nigricapillus reditus (n 5 42)
Panama: Colo?n (former Panama Canal Zone):
Gamboa, R??o Frijoles (1). Colo?n: Cerro Bruja (1); Por-
tobelo (3). Panama?: Estacio?n Hidro El Candelaria (10);
Estacio?n Hidro El Peluca (4); San Miguel, Cerro Azul
(2); Bajo Grande, Cerro Azul (1); Quebrada Joro?n,
Chiman (1); R??o Corotu?, Chiman (3); R??o Maje?, Char-
co del Toro (4); Quebrada Cauchero, Cerro Chucant??
(9); San Blas: Mandinga (3).
There is a nomenclatural problem associated with
the name reditus. Griscom (1932) based this subspe-
cies on a series from easternmost San Blas, Panama,
at Perme? (the type locality, Fig. 1) and Puerto Obald??a.
The subspecies was supposed to differ from schottii in
having the throat white, as in nominate nigricapillus,
with no mention of the characters Wetmore (1959) lat-
er attributed to reditus elsewhere in Panama. The white
throat, as mentioned previously, appears sporadically
elsewhere in schottii. Wetmore (1959:21) considered
birds from Puerto Obald??a and Armila, only a few ki-
lometers to the east of Perme?, to be referable to schottii
and that Perme? was ??barely within the range?? of re-
ditus. We examined nine paratypes of reditus from Per-
me? and Puerto Obald??a and found the amount of white
in the throat to be variable. There was a hint of a more
chestnut wash in the underparts of one specimen from
Perme?, so that it remains possible that some genetic
influence from the reditus form may extend as far east
as Perme?, although this would have to be demonstrated
through DNA analysis. But otherwise reditus would
have to be synonymized with schottii and a new name
given to the birds listed above that Wetmore recog-
nized under reditus.
Thryothorus nigricapillus schottii (n 5 91)
Panama: San Blas: Perme? (5 MCZ), Armila (1),
Puerto Obaldia (1 USNM, 4 MCZ). Darie?n: Tacarcuna
Village (2); Pu?curo (3); R??o Paya mouth (9); Cana (7).
Colombia: Cordoba: R??o Salvaj??n, R??o Esmeralda (5);
Socorre?, R??o Sinu? (4). Bolivar: Volador (1); Regene-
racio?n (3). Antioquia: Alto Bonito (4); La Bodega (1);
El Pescado (6); El Real (3); Villa Arteaga (8); Haci-
enda Bele?n (5). Choco?: Acand?? (3); Nuqu?? (7); R??o
Jurubida? (3); R??o Truando? (2, syntypes). Valle del Cau-
ca: Buenaventura (1); San Jose? (1); Punto Muchimbo,
R??o San Juan (2).
Thryothorus nigricapillus connectens (n 5 5)
Colombia: Cauca: Guap?? (2); Narino: Guayacana
(2); Barbacoas (1).
Thryothorus nigricapillus nigricapillus (n 5 7)
Ecuador: No locality (1). Guayas: Guayaquil (2);
Huerta Negra (4).
In T. n. nigricapillus of Ecuador, the underparts are
white, becoming brownish or buff posteriorly, and
finely barred with black except on the throat and upper
breast. In T. n. schottii of Colombia and eastern Pan-
ama, the barring tends to be heavier and more exten-
sive on the throat. There is considerable variation,
however, and birds with white, unbarred throats may
appear almost anywhere in the range. A female from
El Real, Antioquia, Colombia, is pale, with a white
throat and fine ventral barring, and would be difficult
to separate from a series of T. n. nigricapillus. A male
from the same locality is dark and very heavily barred
nearly to the chin, and a second male is intermediate.
The subspecies T. n. connectens of southern Colombia,
as implied by its name, was supposed to bridge the
differences between nigricapillus and schottii but was
synonymized with schottii by Wetmore (1959).