Discussion
We employed a large number of presumably unlinked neutral markers to test for species replacement of marbled newts in Portugal. Northwards hybrid zone movement has been suggested in this system, with the reported T. marmoratus enclave signalling the competitive advance of T. pygmaeus with incomplete replacement (Wielstra,Burke, Butlin, & Arntzen, 2017a). Species displacement might be driven by the desertification of the region, conferring T. pygmaeus a competitive advantage over its sister species. It is knownT. pygmaeus thrives in ephemeral water bodies, and environmental modelling here suggests it is favoured by arid precipitation regimes, whereas T. marmoratus prefers smaller and more permanent breeding sites under more humid conditions (Espregueira Themudo & Arntzen, 2007; Harrison & Rand, 1989). In fact, climatic simulations also suggest the North Atlantic jet stream, associated with wetter conditions, was at its southernmost position during the Last Glacial Maximum, and it experienced a positive latitudinal shift after this period, parallel toT. marmoratus ’ estimated northwards retreat (Beghin et al., 2016). Unidirectional introgression of functionally and physically unlinked neutral markers was expected in the wake of a moving hybrid zone (Wielstra, Burke, Butlin, & Arntzen, 2017a). Yet, we found no evidence for an additional enclave or a genomic footprint consistent with enclave formation. Thus, two main biogeographic scenarios of enclave formation arise in the absence of a footprint: i) the species could have undergone negligible introgression or ii) displacement could have in fact occurred with introgressive hybridisation, but with the signal subsequently lost.
Species replacement in the absence of locally extensive hybridisation during T. pygmaeus’ expansion could be responsible for enclave formation unaccompanied by a genomic footprint. Displacement might have occurred in allopatry, involving an initial recession of T. marmoratus’ range during the Pleistocene glaciations, followed by a northwards postglacial expansion of T. pygmaeus . Under this scenario, the species would not have been in contact across their western ranges until meeting at the current hybrid zone. The T. marmoratus enclave would have locally persisted during the retreat of its main range, alike the fragmenting pattern of contracting species ranges described by Wilson, Thomas, Fox, Roy, and Kunin (2004), withT. pygmaeus later enveloping the pocket. However, it seems unlikely the marbled newts did not meet after the Last Glacial Maximum, given the similarity in suitable habitats predicted by the environmental modelling.
Alternatively, species replacement may have taken place sympatrically, with strong reproductive isolation preventing introgressive hybridisation. Prezygotic and postzygotic barriers could be limiting hybridisation among the species, resulting in reproductive isolation. Prezygotic effects, such as mating preference or genetic incompatibility, have not been reported in Triturus species, whereas postzygotic effects occur between crested and marbled newts, brought by the direction of the interspecific cross (Arntzen, Jehle, Bardakci, Burke, & Wallis, 2009). The northwards advance of the contiguous species ranges could therefore have been driven by species replacement without introgressive hybridisation, with the enclave likely persisting under T. marmoratus’ presumed strong adaptation to local conditions.
The scenario of species replacement with hybrid zone movement is derived from previous findings supporting a moving contact zone between the marbled newts (Espregueira Themudo & Arntzen, 2007; Espregueira Themudo, Nieman, & Arntzen, 2012). Assuming widespread interspecific hybridisation, the erosion of T. marmoratus’ molecular signal would explain the absence of a genomic footprint. The time passed sinceT. marmoratus inhabited the Lisbon Peninsula, environmental effects, and the competitive advance of T. pygmaeus might have led to the erosion of the footprint. Additionally, strong selection against hybrids, as suggested by the strong bimodality of the species’ contact zone with few admixed individuals in Arntzen (2018), could have swiftly erased the footprint. Purifying selection could have operated to maintain the species’ functional integrity, by eliminating deleterious sequences being pulled into the receding taxon (Oleksyk, Smith, & O’Brien, 2010). Under strong purifying selection, existing introgression of neutral variants might not have been detected despite the large number of markers here employed. In fact, the significant instances of heterozygote deficit and admixture linkage disequilibrium are likely due to a combination of incomplete lineage sorting or ancestral polymorphism and relatedness amongst the sampled larvae, resulting from the collection of siblings that can potentially bias the landscape genetic structure of populations (Goldberg & Waits, 2010).
The proposed biogeographic scenarios of species replacement are subject to the unknown past position of the species boundary. While environmental modelling supports T. marmoratus inhabiting the study area, it remains possible that this species did not previously inhabit the Lisbon Peninsula. Moreover, the absence of an additional enclave or disconnected footprints in mountainous areas, such as Serra de Sintra, does not clarify T. marmoratus’ past range nor whether species replacement occurred with or without hybridisation. If in the past, the southernmost range of T. marmoratus included Caldas da Rainha, enclave formation could have occurred via incomplete species replacement. Thus, testing for a genomic footprint north of the enclave could potentially elucidate among the aforementioned biogeographic scenarios. We expect future studies on the area between the T. marmoratus enclave and its main distribution range to unravel a genomic footprint of species replacement with hybridisation, confirming the dynamic nature of the hybrid zone movement between the marbled newts.
Remarkably, the enclave in Caldas da Rainha showed notable levels of introgression, possibly signalling the beginning of its erosion, as predicted by Espregueira Themudo & Arntzen (2007), and further illustrating the dynamics of species replacement. The enclave is expected to eventually disappear under T. pygmaeus’ competitive advance, leading to the loss of T. marmoratus gene variants over time. The expansion of T. pygmaeus is likely influenced by the loss of breeding sites in southern Spain and Portugal, an area with unique biodiversity patterns stricken by a decline in temporary ponds, driven by climate warming, desertification and agricultural intensification (Arntzen et al. , 2004; Thomas, Franco, & Hill, 2006; van de Vliet et al. , 2014). Understanding shifts in species distributions, particularly when driven by climate change and anthropogenic activities, therefore becomes especially relevant in deciphering the dynamics of species replacement (Taylor, Larson, & Harrison, 2015).