Introduction
Theoretical predictions (MacArthur & Wilson, 1967) and empirical data (Crottini et al. , 2012; Economo et al. , 2015) suggest that islands are rarely a source of diversity for continents. This broad pattern is thought to be underpinned by varying combinations of the typically lower diversity of species on islands (Matthews et al. , 2019), tendency for insular taxa to lose adaptations for dispersal or competition (Bellemain & Ricklefs, 2008), and putatively higher rates of extinction in less stable island systems (Whittaker et al. , 2017). Because biotic communities on larger continents are often more species-rich they are also predicted to generate higher numbers of outwards colonists (MacArthur & Wilson, 1967) and act as a strong ecological filter against upstream colonisation (White et al. , 2021). However, running counter to these broad predictions, phylogenetic analyses and ancestral state reconstructions have demonstrated that some continental lineages have insular origins (upstream colonisation or reverse colonisation) (Carine et al. , 2004; Filardi & Moyle, 2005; Bocek & Bocak, 2019), especially in the Caribbean region (Nicholson et al. , 2005; Bellemain & Ricklefs, 2008). Understanding where, when and how islands contribute diversity to continents has the potential to provide broader insights into the factors that shape biotic dispersal and colonisation (Bellemain & Ricklefs, 2008).
The Indo-Australian Archipelago (IAA) spans from the Sundaland continental promontory to New Guinea, and is sometime also referred to as Malesia (as opposed to Melanesia). This region provides a prominent example of putative upstream colonisation with subsequent radiation in the global radiation of Oscine songbirds, which comprises of over 5000 extant species, all hypothesised to be descended from ancestors which occurred in islands of the IAA (Jønsson et al. , 2010; Moyleet al. , 2016). Upstream colonisation of south-east Asia from the IAA has also been recently inferred in two insect lineages (Letschet al. , 2020; Bank, Cumming, & Bradler, 2021). The original geographic source of these upstream colonisations has been variably linked to terranes accreted to the northern edge of the Australian continent, often referred to as ‘the proto-Papuan Archipelago’ (Jønssonet al. , 2010), or the islands of “Wallacea” (Moyle et al. , 2016) (see more discussion of the geology of these regions below). These results emphasise the potential significance of the islands of eastern IAA as a source for diversity on nearby continents. However, the dynamics of insular diversification and island-continent dispersal across this region remains poorly understood.
Pigeons and doves (Columbiformes) are a moderately diverse global radiation of fruit and seed-eating birds (~350 nominal species), that also show a strong signature of insular diversification (Cibois et al. , 2014, 2017) and upstream colonisation (Lapiedraet al. , 2021). The islands of the IAA, and especially New Guinea and surrounds, are a world centre of pigeon ecological and lineage diversity, and were even more diverse pre-human settlement across the Pacific (Steadman, 2006; Steadman & Takano, 2020). Conversely, the oldest known Columbiform fossils are from early Miocene deposits in Australia (Worthy 2012) and New Zealand (Worthy et al. 2009), and on this basis it has been suggested that pigeons may have originated in Australia, or nearby islands (Low, 2014). To better understand the temporal and spatial dynamics of pigeon dispersal and diversification between islands and continents in the IAA, we assembled a supermatrix dated phylogeny for the pigeons, and from this we estimate, 1) the timing and geographic foci of insular pigeon diversification, with particular focus on the potential role of island arc terranes in the IAA in the early radiation of the group, and 2) estimate the relative frequency and direction of dispersal events between major island groups, and nearby continents.