Why are the phylogenetic relationships among wrens of interest? Many species of wrens exhibit a social system termed cooperative breeding, in which offspring remain with their parents for one or more breeding seasons after reaching sexual maturity, often helping to feed and care for their new siblings. The evolutionary origins and selective pressures maintaining (or constraining) this behavior remain an area of controversy and active research. One way of getting at the factors which might be important is to look at multiple, independently-derived cases of cooperative breeding, to seek common underlying ecological, behavioral, and demographic correlates. At least three genera of wrens (Campylorhynchus, Cinnycerthia, and Cyphorhinus; follow this link for a summary of evidence for cooperative breeding in wrens) show good evidence for the occurrence of cooperative breeding in one or more species, but relationships among these genera were unknown. Another species, the black-capped donacobius (Donacobius atricapilla) of Central and South America, is also a cooperative breeder, and was recently reclassified as wren. Thus, wrens could represent as many as four origins of cooperative breeding (and thus four independent chances to test potential correlates), or as few as one, depending on their interrelationships.

Another interesting behavior exhibited by many wren species is vocal duetting. This occurs when both members of a breeding pair participate in song production, coordinating their songs to a greater or lesser degree (the University of Michigan's Animal Diversity Web has some great examples). Most wren species in the temperate zone do not perform duets, whereas most species in the tropics do (follow this link for a summary of evidence for duetting behavior in wrens). One exception is the Carolina wren (Thryothorus ludovicianus), where males produce a “cheerily cheerily cheerily” song familiar to many in North America, and females often produce an accompanying “chatter” song. The function of duets such as these was first studied experimentally by Rachel Levin (Pomona College), and has subsequently been studied in a variety of species (T. leucotis, Sharon Gill; T. rufalbus, Dan Mennill; T. fasciatoventris, David Logue). Song use by the Carolina wren has been studied by Eugene Morton (now at the Smithsonian Migratory Bird Center), and Sandra Vehrencamp's group is pursuing studies of a non-duetting wren (the banded wren Thryothorus pleurostictus). As with cooperative breeding, the number of origins and losses of duetting behavior in these birds can only be inferred using phylogenetic data on the group, which have been unavailable until recently. In collaboration with Peter Slater's group at the University of St. Andrews, I am studying the evolution of duetting behavior in wrens (and particularly in the genus Thryothorus) using phylogenetic comparative methods.

A final aspect of wrens—their biogeography—makes them a compelling subject of study. Many groups of New World birds are clearly northern or southern in origin, but wrens are difficult to classify in this way. For instance, wren species-level diversity reaches its peak in Central America, and there are endemic genera in both North and South America. Phylogenetic analyses of these birds may help to clarify their continental origins by identifying where the basal divergences occurred.

Presented above is Figure 4 from Barker (2004), summarizing relationships among genera of wrens (Troglodytidae). This unrooted tree is based on partial sequences of the mitochondrial cytochrome b gene, in combination with sequences from the fourth intron of the b-fibrinogen gene (see data). The root of this tree is ambiguous given these data, but additional sequences (the nuclear loci RAG-1, RAG-2, b-fibrinogen intron 7, and d-EF1) indicate that in all likelihood it lies on the branch separating Salpinctes, Microcerculus, Catherpes, and Hylorchilus from other wrens. Importantly, the three genera containing cooperatively-breeding species are not closely related, indicating that this behavior was independently derived in each. Note that Donacobius is not in this tree. Molecular data strongly support exclusion of this species from wrens, placing it with the Old World warblers and allies (see Barker 2004). Unfortunately, reconstructing the history of duetting behavior on this tree is still problematic, given the relatively poor sampling in most groups, especially the large genus Thryothorus. Notably, Thryothorus is itself paraphyletic, indicating that comparisons between species in the genus (e.g., tropical versus temperate) must be interpreted with caution. Note that the alternative rootings of this tree have important implications for the biogeography of the family. For instance, Salpinctes, Catherpes, and Hylorchilus are all endemic to North America (defined as land north of the Panamanian isthmus), while Microcerculus is found on both sides of the isthmus, and Odontorchilus is endemic to South America. The rooting favored by unpublished data in hand yields ambiguity in the continental origin of wrens. The phylogenetic structure in their sister group (gnatcatchers and gnatwrens) may be informative with regard to this question.

Species illustrated:
Salpinctes obsoletus Rock wren.
Catherpes mexicanus Canyon wren.
Odontorchilus cinereus Tooth-billed wren.
Troglodytes troglodytes Winter wren.
Troglodytes aedon House wren.
Cistothorus palustris Marsh wren.
Campylorhynchus brunneicapillus Cactus wren.
Thryothorus ludovicianus Carolina wren.
Thryomanes bewickii Bewick's wren.
Polioptila caerulea Blue-gray gnatcatcher.
Certhia familiaris Brown creeper.
Sitta canadensis Red-breasted nuthatch.
 
 
 

Four genera (Salpinctes, Microcerculus, Catherpes, and Hylorchilus) appear at the base of the wren family tree. Initial results suggested that the rock wren (Salpinctes obsoletus, left) might be the sister group to all other wrens.  However, ongoing study of the group with additional data suggests that this species is part of a monophyletic group containing the other petrophilous (rock-loving) species in these genera, such as the canyon wren (Catherpes mexicanus, right).

 
 
 

The South American genus Odontorchilus (2 species, O. cinereus pictured at right) forms the sister group to all the other "non-petrophilous" wrens.

	The marsh wren group (Cistothorus), and the house/winter wren group (Troglodytes) form a clade along with the little-known timberline wren (Thryorchilus browni) from Central America.  This clade includes such familiar species as the winter wren (Troglodytes troglodytes, left), the house wren (Troglodytes aedon, below left), and the marsh wren (Cistothorus palustris, below right).

Wrens in the genus Campylorhynchus (the cactus wren C. brunneicapillus, above, and allies) form a strongly-supported clade along with the Carolina wren (Thryothorus ludivicianus, upper right), and the Bewick's wren (Thryomanes bewickii, right). Most other Thryothorus (it remains to be seen how many) fall into a separate group, along with the tropical genera Henicorhina, Uropsila, Cyphorhinus, and Cinnycerthia. Nearly all Campylorhynchus species are cooperative breeders (excluding C. brunneicapillus, above), as are several species in the genera Cyphorhinus and Cinnycerthia.

This blue-gray gnatcatcher (Polioptila caerulea) is representative of the sister-group to wrens, which includes all gnatcatchers, along with the Neotropical "gnatwrens" in the genera Microbates and Ramphocaenus. Contrary to Sibley and Ahlquist's Tapestry, which also places the verdin (Auriparus) in this group, additional DNA-DNA hybridization and DNA sequence data confirm its traditional relationship with the penduline tits (Remizidae).

Three traditional families (Troglodytidae, Certhiidae, and Sittidae), along with the gnatcatchers and gnatwrens (see above) form a well-supported clade.  However, aside from the relationship of wrens (Troglodytidae) with the gnatcatchers and allies, relationships among these lineages remain poorly resolved.  The sister group to the wrens and gnatcatchers may be the creepers (Certhiidae), such as the brown creeper (Certhia familiaris) shown above.  Nuthatches, such as this red-breasted (Sitta canadensis, below) may be the sister group to the the other members of this assemblage. The relationships of yet another member of the group, the European wallcreeper Tichodroma—which has been allied to both the Certhiidae and Sittidae in previous classifications—remain obscure even after the collection of over 7000 bases of nuclear DNA sequence.