The population structure of A. mantiqueira between mountains ranges and implications for conservation
The best number of clusters detected by the clustering analyses for all samples was K = 2. This K value, however, did not indicate any population structure for A. mantiqueira through its distribution. Nonetheless, phylogenetic inference and DAPC analysis revealed some intraspecific structure, implying that a population structure exists that was not detected by cluster analyses using the optimal K = 2. All methods proposed to determine genetic structure a posteriori will face detection limits when differentiation between groups is low, and some factors such as uneven sampling and the number of individuals may affect the optimal K value (Miller et al., 2020). In simulations to test the efficiency of DAPC to assess population genetic structure, it was observed that there is considerable inaccuracy in the face of low levels of differentiation (FST values <0.1) (Miller et al., 2020). This may explain the inability to detect intraspecific population structure in A. mantiqueira since the mean FST within populations is 0.0122.
Clustering methods depict a simplified version of a complicated reality that does not always correspond to existing population genetic models (Jombart, 2012). For this reason, it is important to interpret clustering results with caution and consider additional genetic and demographic factors that may influence population structure. Additionally, incorporating information from multiple clustering methods or conducting further analyses can provide a more comprehensive understanding of genetic diversity within populations. Therefore, other K values were investigated, and the K = 4 value recovered the same clustering pattern in all analyses, regardless of the different assumptions underlying these methods. Furthermore, this clustering pattern was preferred because it is congruent with A. mantiqueiramountain range distribution, implying biological relevance. Indeed, a process of differentiation between the populations from Serra da Bocaina and the inland populations may be ongoing, as evidenced by the slight variations in wing patterns that exist between the two.
The phylogeographic break recovered between Serra da Mantiqueira and Serra do Mar Mountain ranges has been observed in Atlantic Forest orchids, and dispersal is a more plausible explanation than vicariance for the origin of this population structure (Pinheiro et al., 2013). The lowland regions between these mountains were potentially phylogeographic barriers to dispersal for endemic birds as well (Chaves et al., 2015).
Singularly, the sample of A. mantiqueira from Poços de Caldas (MG) stands out from the rest of the Serra da Mantiqueira. The Poços de Caldas plateau is a circular structure of Mesozoic age comprising a suite of alkaline volcanic and plutonic rocks with elevations up to 1500–1600 m above sea level in its borders. Its original vegetation coverage consisted of the Atlantic Forest biome, but this region has suffered high degradation, losing areas of native vegetation to pastures (Grohmann et al., 2007; RadamBrasil, 1983; Schorscher & Shea, 1992). The highlands of Poços de Caldas also have a distinct geology when compared to the rest of the Serra da Mantiqueira complex (Schorscher & Shea, 1992), making them a distinct and representative area of endemism for the region’s anurans (Neves et al., 2018). Because of the strong influence of a different domain (the Cerrado savannas) in this location, phytophisiognomy factors may explain some of the observed genetic discontinuity. Furthermore, this location is one of the most geographically distant from the others in Serra da Mantiqueira and may be under greater isolation by distance (IBD) effect.
The discontinuity between mountain ranges may be partially due to historical processes, such as the founder effect, which involves the colonization of new mountains through dispersal. By this model, new populations would be established by few individuals from a large ancestral population, suffering the effects of genetic drift after establishment and large changes in allele frequencies due to sampling error (Templeton, 2008). Another potential outcome is the reduction in genetic diversity caused by genetic drift following the disruption of gene flow between distinct mountain ranges. In this case, the ancestral population would have been distributed over the entire sampled extent, and later, climatic barriers during interglacial periods may have caused isolation in sub-populations. Natural selection may also have influenced these polymorphisms in different locations, selecting those more adapted to the ecological, climatic, and niche variations of these mountain ranges, promoting faster differentiation among populations (Matsubayashi & Fujiyama, 2016). Our results suggests that gene flow between populations of A. mantiqueira may be limited, leading to potential genetic divergence within the species. Therefore, the same factors that lead to allopatric speciation in A. mantiqueira andA. alalia may also operate at the intrapopulation level inA. mantiqueira .
Climate change is a potential danger to the survival of these species that inhabit mountainous regions. This is particularly true for the Atlantic Forest, a region of high biodiversity that is currently facing habitat fragmentation as a result of urbanisation and human activities, leaving only 12% of its original vegetation intact. (SOS Mata Atlântica, 2019). According to the niche models generated in the present study, the fragments suitable for these species will become scarcer as the mean temperature rises. Habitat fragmentation may reduce dispersal and, consequently, genetic connectivity among these populations (Cunningham & Moritz, 1998; Dayanandan et al., 1999; Gerlach & Musolf, 2000). This major landscape alteration is likely to change gene flow and promote genetic drift in natural populations (Dixo et al., 2009; Vandergast et al., 2006). The loss of genetic connectivity may be detrimental to long-term species persistence (Gilpin & Soulé, 1986; Templeton et al., 1990). For this reason, more phylogeographic, climate modeling, and demographic studies on the fauna of MAF will be crucial for the conservation of the biodiversity of this hotspot biome in scenarios of habitat fragmentation and climate change.