Deciphering the mechanisms of species formation is one of the most fascinating and challenging areas of evolutionary biology (Darwin and Wallace 1858; Darwin 1859; Mayr and Diamond 2001; Coyne and Orr 2004; Fitzpatrick et al. 2009; Santini et al. 2012). Many studies on biogeographic and ecological factors promoting speciation have helped to establish the separation by physical barriers (“vicariance”) followed by genomic isolation as the null hypothesis that begets new species (Lynch 1989; Barraclough and Vogler 2000; Mayr and Diamond 2001; Johannesson 2010; Santini et al. 2012). However, although restriction of gene flow and allopatric speciation appear to be the most common mechanisms, different processes have also been documented through the years (White 1968; de Aguiar et al. 2009; Johannesson 2010). For instance, sympatric speciation, suggested already by Darwin (1859) before being tested and supported for many metazoan taxa (e.g., Crow et al. 2010), implies that speciation events can occur within the same population through genetic polymorphism. Furthermore, special cases of allopatric speciation are found in parapatric and peripatric models, which invoke processes at the distributional periphery of an ancestral species, where individuals might enter a new habitat most likely facilitated by divergent ecological characteristics. In this case of peripheral speciation, also known as “budding speciation,” either the geographic isolation, ecological factors, or a combination of the two leads to a cessation of gene flow, thus enabling the speciation process (Mayr 1982; Fitzpatrick and Turelli 2006).
Here, we focus on the Australasian archipelago and surrounding areas, a region that shelters a rich yet highly threatened biodiversity; seven biodiversity hotspots are situated in the Indomalayan-Australasian region (Mittermeier et al. 2004; conservation.org). This region, despite its highly complex geological history (e.g., Hall 2011; Metcalfe 2011), represents an ideal laboratory to study lineage diversification and speciation (Wallace 1860; Mayr and Diamond 2001; Condamine et al. 2013). Thousands of islands, many of them scattered across the Equator, varying in size from tiny patches to continental sized landmasses, ranging from young to geologically old and low-lying to high altitudes including snow-capped summits, harbor hyperdiverse biota and exceptional radiations. Particularly across the megadiverse Wallacea and Melanesia, most studies to date have investigated larger scale evolutionary patterns, whereas factors promoting speciation remain scarcely addressed, and the evolutionary processes involved are little known despite increased recent efforts (e.g., Von Rintelen et al. 2004; de Bruyn and Mather 2007; Joseph and Omland 2009; Craft et al. 2010; Deiner et al. 2011; Klaus et al. 2013). A unique feature in the Indomalayan-Australasian archipelago is its long chain of islands often with high mountains, usually surrounded by tropical lowland rain- or dry-forest. “Sky island” ecosystems are isolated patches surrounded by dramatically different lowland ecosystems (Heald 1967), in this case, isolated further from each other by ocean. Recently, such highland ecosystems have been shown to act as evolutionary cradles shaping a flourishing biota in diverse regions of the World (Hall 2005; Smith and Farrrell 2005; Robin et al. 2010; Schultheis et al. 2012). In the Australasian region, sky islands are geologically young (<5 Mya; Cloos et al. 2005), yet highly diverse (e.g., Mittermeier et al. 2004). Numerous studies on Australian and New Zealand mountain ranges have investigated speciation patterns and indicate important vicariant effects from mountain uplift per se, and climate change as further promoter of species diversity (e.g., Trewick et al. 2000; McCulloch et al. 2010; Hawlitschek et al. 2012). The mountains of Indonesia and northern Australasia show especially striking altitudinal gradients, usually being surrounded by tropical lowland rain- or dry-forest. Within this biodiversity-rich assemblage of mountainous ecosystems, one of the largest and most remote highland regions is the central New Guinean cordillera, with vast expanses of tropical montane and subalpine habitat, and extensive areas above 3500-m altitude, including numerous summits above 4500 m. Their role as a diversity pump for the archipelago remains poorly studied and little appreciated, despite their vast geographic extent and extreme structuring (but see Mayr and Diamond 1976).
Here, we study pond-dwelling Rhantus diving beetles, often abundant in tropical montane and subalpine pond habitats across the Australasian region. There are at least 30 endemic species in the region including Oceania, mostly narrow endemics restricted to a single high valley, or few mountain tops. There is one striking exception, however: Rhantus suturalis MacLeay, 1825 (Coleoptera, Dytiscidae, Colymbetini) (Fig. 1), is a widespread ecological supertramp, ranging from the Azores islands to New Zealand, and inhabits diverse lentic habitats in mountainous or subalpine environments, for example, high altitude lakes (Fig. 2) and highland peat swamps (with pH around 4–5). It occurs in temperate lowland swamps, many anthropogenic habitats (freshly dug fish ponds, reservoirs, roadside ditches, cattle troughs), saline desert wetlands in North African deserts, and many more. Along with its closer relatives, it has never been found in tropical lowlands (Balke 1993, 2001; Balke et al. 2009). Rhantus suturalis is often an early colonizer of newly available habitats, hence referred to as a “supertramp species” (Balke et al. 2009). It is an ecological generalist, with high physiological tolerance, for example, in terms of salinity, temperature, and acidity. A comprehensive molecular phylogeny of R. suturalis from across its wide range revealed two major clades – a northern one, from Portugal to Sumatra, and a southern one from adjacent Java eastward to New Caledonia (Fig. 3) (Balke et al. 2009). Both R. suturalis clades contain one or more narrow-endemic species previously described based on marked morphological divergence. This species paraphyly was supported by extensive mitochondrial and nuclear DNA sampling (>7000 bp). A recent origin of R. suturalis, c. 6.0–2.7 Ma ago, was suggested, possibly in New Guinea, followed by an ancestral colonization of the Malay Archipelago and a large part of the Australian region. The rise of a widespread generalist out of a clade of narrow endemics not only refutes the assumption of “specialisation as an evolutionary dead-end” (Mayr 1963; coined by Cope 1896 “the law of the unspecialized”), but offers an opportunity to study the early phases of lineage diversification across a wide species range, which is nevertheless constrained by climate and other ecological factors.
In this study, we use extensive sampling across the Indomalayan-Australasian region to (1) reconstruct phylogenetic relationships within the southern clade of the widespread R. suturalis, (2) investigate phylogeographic patterns using haplotype network inferences, (3) infer the historical demography and timing of divergence of this group in a paleoclimatic framework, in order to test the hypothesis of ongoing peripatric speciation in Australasia, particularly New Guinea, sky islands, and (4) examine whether those mountain chains act as a species pump in a “cradle of evolution” model, or as an ancient biotic pool, in a “museum” model.