Skip to main content

A test of island biogeographic theory applied to estimates of gene flow in a Fijian bird is largely consistent with neutral expectations

Cite this dataset

Gyllenhaal, Ethan et al. (2020). A test of island biogeographic theory applied to estimates of gene flow in a Fijian bird is largely consistent with neutral expectations [Dataset]. Dryad.


Islands were key to the development of allopatric speciation theory because they are a natural laboratory of repeated barriers to gene flow caused by open water gaps. Despite their proclivity for promoting divergence, little empirical work has quantified the extent of gene flow among island populations. Following classic island biogeographic theory, two metrics of interest are relative island size and distance. Fiji presents an ideal system for studying these dynamics, with four main islands that form two large-small pairs. We sequenced thousands of ultraconserved elements (UCEs) of the Fiji bush-warbler Horornis ruficapilla, a passerine distributed on these four Fijian islands, and performed a demographic analysis to test hypotheses of the effects of island size and distance on rates of gene flow. Our demographic analysis inferred low levels of gene flow from each large island to its small counterpart and little or none in the opposite direction. The difference in the distance between these two island pairs manifested itself in lower levels of gene flow between more distant islands. Both findings are generally concordant with classic island biogeography. The amount of reduction in gene flow based on distance was consistent with predictions from island biogeographic equations, while the reduction from small to large islands was possibly greater than expected. These findings offer a hypothesis and framework to guide future study of inter-island gene flow in archipelagos as the study of island biogeography progresses into the genomic era.


The dataset is based on ultraconserved elements (UCEs) and SNPs called from them. The input files for phylogenetic and population genetic analyses are included here. Scripts used to produce and use these files can be found at

Usage notes

These are descriptions of files in this Dryad upload for the paper:

TableS3_FullFastsimcoalFull.xlsx: Supplemental table with full fastsimcoal output with all Kadavu individuals.
TableS4_FullFastsimcoalNoAdmix.xlsx: Supplemental table with full fastsimcoal output without the putative admixed Kadavu individual.
horo_100_rand.vcf: VCF file used for most pop gen analyses (not sNMF).
horo_100_rand_noSingle.vcf: VCF file used for sNMF analysis.
horornis_partition_finder.cfg: Configuration file used to run PartitionFinder2.
horornis_partitions.txt: Partitions output from PartitionFinder2.
horornis_min95_renamed.phy: Input phylip file used for RAxML analysis.

Scripts used to generate this input and run analyses can be found at:


National Science Foundation, Award: DGE-1650114

National Science Foundation, Award: DEB-1557051

National Science Foundation, Award: DEB-1557053