Skip to main content
Dryad logo

Multilocus phylogeography, population genetics and niche evolution of Australian Brown and Black-tailed Treecreepers (Aves: Climacteris)


Edwards, Scott et al. (2022), Multilocus phylogeography, population genetics and niche evolution of Australian Brown and Black-tailed Treecreepers (Aves: Climacteris), Dryad, Dataset,


The Carpentarian barrier across northeastern Australia is a major biogeographic barrier and a generator of biodiversity within the Australian Monsoonal Tropics. Here we present a continent-wide analysis of mitochondrial (control region) and autosomal (14 anonymous loci) sequence and indel variation and niche modeling of Brown and Black-tailed Treecreepers (Climacteris picumnus and Cmelanurus), a clade with a classic distribution on either side of the Carpentarian barrier. mtDNA control region sequences exhibited reciprocal monophyly and strong differentiation (Fst = 0.91), and reveals a signature of a recent selective sweep in C. picumnus. No loci among 14 anonymous autosomal markers exhibited reciprocal monophyly between species, and a variety of tests support an isolation-with-migration model of divergence, albeit with low levels of gene flow across the Carpentarian barrier and a divergence time between species of ~1.7 – 2.8 MYA, depending on the model and assumptions about generation time. Paleo-ecological niche models show that both range size as measured by available habitat and estimated historical population sizes of both species declined in the last ~600 kyr and that the area of range overlap was never historically large, perhaps decreasing opportunities for extensive gene flow. The relatively long divergence time and low opportunity for gene flow may have facilitated speciation more so than in other co-distributed bird taxa across the Australian Monsoonal Tropics.


Avian tissue was collected in Australia under appropriate permits. Total genomic DNA was extracted from heart, liver or muscle tissue using phenol chloroform methods. To generate nuclear loci, we constructed a plasmid library from total genomic DNA from a single female C. picumnus, shearing, repairing and size-selecting the DNA into 2-3 kb fragments using methods. We used PCR to amplify a 363 base-pair segment of region 1 of the mitochondrial control region (CR1) and amplification conditions for each anonymous locus was optimized by thermal gradient PCR and annealing temperatures for PCR varied from 50 – 55ºC.  Sequences were trimmed, aligned and edited, and putatively heterozygous sites identified, using Aliview. Unphased diploid DNA sequences for each locus were phased using Phase 2.0. Phased haplotypes were converted into Nexus and Phylip formats for subsequent analyses.  Standard population genetic statistics, such as genetic diversity within and between populations and species , Tajima’s D and Fst among populations based on pairwise differences of haplotypes, were calculated using the R package PopGenome v. 2.7.5. Gene trees were constructed using the maximum likelihood algorithm in IQ-TREE v. 1.6.12, using best substitution models for each locus as estimated in ModelFinder. Net nucleotide diversity between populations and species was calculated using PopGenome . Indels were scored as biallelic or multiallelic and Fst for each indel was calculated using the hierarchical method implemented in hierFstat. We analyzed the data via principal components analysis (PCA) in smartpca v. 8000. We used STRUCTURE v. to understand the basic population units within and between species. We used Phrapl to test hypotheses of major classes of phylogeographic models. We used bpp v. 4.3 to estimate the population phylogeny and divergence times. We used migrate-n (version 5.0, Beerli & Felsenstein, 1999; Beerli, 2006) to conduct additional phylogeographic model selection (divergence with gene flow) using the full data set as well as equilibrium models of gene flow within C. picumnus and C. melanurus separately. The between-species analyses using migrate-n utilized the full sequence data from C. picumnus and C. melanurus, treating each species as a panmictic population.  For each species, we used ENMeval to test the best model parameters for species distribution models and downstream analysis with Maxent v. 3.4.1. We used Stairway plots to estimate variation in effective population size through time for C. picumnus and C. melanurus.

Usage Notes

Phrapl (R package)
PopGenome (R package)
Bayesian Phylogeography and Phylogenetics (bpp)
hierfstat (R package)



National Science Foundation, Award: DEB-0500862

National Science Foundation, Award: DBI- 1145999

National Science Foundation, Award: DBI-1564822

National Science Foundation, Award: DBI-2019989