Additional file 1: Table S1. List of selected health centres and villages for Intensification Plan.

Additional file 1: Table S1. List of selected health centres and villages for Intensification Plan.

Southeast Asia harbours abundant biodiversity, hypothesized to have been generated by Pliocene and Pleistocene climatic and environmental change. Vicariance between the island of Borneo, the remaining Indonesian archipelago and mainland Southeast Asia caused by elevated sea levels during interglacial periods has been proposed to lead to diversification in the littoral zone mosquito Anopheles (Cellia) sundaicus (Rodenwaldt) sensu lato. To test this biogeographical hypothesis, we inferred the population history and assessed gene flow of A. sundaicus s.l. sampled from 18 populations across its pan-Asian species range, using sequences from mitochondrial cytochrome c oxidase subunit 1 (CO1), the internal transcribed spacer 2 (ITS2) and the mannose phosphate isomerase (Mpi) gene. A hypothesis of ecological speciation for A. sundaicus involving divergent adaptation to brackish and freshwater larval habitats was also previously proposed, based on a deficiency of heterozygotes for Mpi allozyme alleles in sympatry. This hypothesis was not supported by Mpi sequence data, which exhibited no fixed differences between brackish and freshwater larval habitats. Mpi and CO1 supported the presence of up to eight genetically distinct population groupings. Counter to the hypothesis of three allopatric species, divergence was often no greater between Borneo, Sumatra/Java and the Southeast Asian mainland than it was between genetic groupings within these landmasses. An isolation-with-migration (IM) model indicates recurrent gene flow between the current major landmasses. Such gene flow would have been possible during glacial periods when the current landmasses merged, presenting opportunities for dispersal along expanding and contracting coastlines. Consequently, Pleistocene climatic variation has proved a homogenizing, rather than diversifying, force for A. sundaicus diversity.

Southeast Asia is one of the world’s richest regions in terms of biodiversity. An understanding of the distribution of diversity and the factors shaping it is lacking, yet essential for identifying conservation priorities for the region’s highly threatened biodiversity. Here we take a large scale comparative approach, combining data from nine forest associated Anopheles mosquito species and using statistical phylogeographic methods to disentangle the effects of environmental history, species specific ecology, and random coalescent effects. Spatially explicit modelling of Pleistocene demographic history supports a common influence of environmental events in shaping the genetic diversity of all species examined, despite differences in species' mtDNA gene trees. Populations were periodically restricted to allopatric northeastern and northwestern refugia, most likely due to Pleistocene forest fragmentation. Subsequent southwards post-glacial recolonisation is supported by a north-south gradient of decreasing genetic diversity. Repeated allopatric fragmentation and recolonisation has led to the formation of deeply divergent geographical lineages within four species and a suture zone where these intraspecific lineages meet along the Thai-Myanmar border. A common environmental influence for this divergence was further indicated by strong support for simultaneous divergence within the same four species, dating to approximately 900 kya. Differences in the geographical structuring of genetic diversity between species are likely the result of varying species’ biology. The findings have important implications for conservation planning; if the refugial regions and suture zone identified here are shared by other forest taxa, the unique and high levels of genetic diversity they house will make these areas conservation priorities.

Tropical forests have undergone repeated fragmentation and expansion during Pleistocene glacial and interglacial periods, respectively. The effects of this repeated forest fragmentation in driving vicariance in tropical taxa have been well studied. However, relatively little is known about how often this process results in allopatric speciation, since it may be inhibited by recurrent gene flow during repeated secondary contact, or to what extent Pleistocene-dated speciation results from ecological specialisation in the face of gene flow. Here, divergence times and gene flow between three closely-related mosquito species of the Anopheles dirus species complex endemic to the forests of Southeast Asia, are inferred using coalescent based Bayesian analysis. An Isolation with Migration model is applied to sequences of two mitochondrial and three nuclear genes, and 11 microsatellites. The divergence of An. scanloni has occurred despite unidirectional nuclear gene flow from this species into An. dirus. The inferred asymmetric gene flow may result from the unique evolutionary adaptation of An. scanloni to limestone karst habitat, and therefore the fitness advantage of this species over An. dirus in regions of sympatry. Mitochondrial introgression has led to the complete replacement of An. dirus haplotypes with those of An. baimaii through a recent (~62 kya) selective sweep. Speciation of An. baimaii and An. dirus is inferred to have involved allopatric divergence throughout much of the Pleistocene. Secondary contact and bidirectional gene flow has occurred only within the last 100,000 years, by which time the process of allopatric speciation seems to have been largely completed.