Influence of latitudinal region on FST
Population differentiation was higher in the tropics and subtropics than in temperate regions (Fig. 1e). This result supports the idea that patterns of local diversity, such as the partitioning of genetic diversity among plant populations, cannot be explained in isolation from the geographic and historic processes of each region (Ricklefs, 1987, 2004, 2006). Some factors that may contribute include regional differences in seasonality, macroevolution, and geography, differences which have more generally been hypothesized to contribute to the latitudinal diversity gradient (i.e. increased species richness closer to the equator) (Mittelbach et al., 2007; Rolland, Condamine, Jiguet, & Morlon, 2014; Schemske, Mittelbach, Cornell, Sobel, & Roy, 2009). Below we discuss some of these ideas, including the ‘asynchrony of seasons hypothesis’ (ASH) (Martin et al., 2009), the ‘time/area hypothesis’ (Fine & Ree, 2006), and the ‘niche conservatism hypothesis’ (Kerkhoff, Moriarty, & Weiser, 2014).
One compelling explanation for the regional differences in FST is based on the idea that the tropics can have highly asynchronous rainfall patterns over small spatial scales (Martin et al., 2009). Given that most plants time their flowering to seasons (Crimmins, Crimmins, & Bertelsen, 2011; Gaudinier & Blackman, 2019), and that seasons are largely determined by rainfall in the tropics, small-scale differences in rainfall potentially disrupt gene flow and cause high population differentiation over short distances compared to the temperate zones. This is the aforementioned ASH, and our analyses support the prediction of higher population differentiation in the tropics. We note that the tropics and subtropics did not differ in FST, and that these regions have comparable climatic patterns (Sitnikov, 2009), thus the ASH may extend to subtropical regions.
Higher FST in the tropics/subtropics than in the temperate zones can also be due to the different history of plant lineages in each region. The ‘time/area hypothesis’ (Fine & Ree, 2006) and the ‘niche conservatism hypothesis’ (Kerkhoff et al., 2014) allude to the idea that tropical clades are older and tend to live in the same environments throughout their evolutionary history, while temperate clades diversified more recently after switching to novel environments once cooling began in the Oligocene. Thus, most temperate species likely expanded their populations fairly recently post-glaciation (34 Mya), resulting in lower population differentiation due to recent gene flow maintaining cohesion. In contrast, tropical species may have been in the same place longer and their populations have had more time to isolate due to dispersal limitations and build up genetic differentiation (Kisel & Barraclough, 2010; Smith et al., 2014). Tropics and subtropics share strong floristic affinities (Sarmiento, 1972), which corresponds to the similar FST between them.
Finally, gene flow is likely more restricted in the tropics due to its heterogeneous orogeny and rich fluvial systems. Such geographic differences have also been hypothesized to contribute to the latitudinal diversity gradient (e.g., Smith et al., 2014; Wallace, 1854). This argument becomes particularly compelling in combination with the fact that temperature does not vary as extremely through the year in the tropics. Given this, different subpopulations would be expected to evolve narrower physiological niches that adapt them to particular altitudinal zones, and a similarly sized mountain would impose a greater barrier to dispersal, and thus to gene flow among subpopulations, in tropical than in temperate regions (Ghalambor, 2006; Janzen, 1967).
Thus, overall, our results are in line with hypotheses that suggest greater species diversity in the tropics is due to higher speciation rates rather than lower extinction rates. While the specific mechanisms differ, including those mentioned above and others (see Mittelbach et al., 2007), these hypotheses all posit greater population-level differentiation that then scales up to faster speciation rates in a model of allopatric or parapatric speciation. Direct tests on the influence of population differentiation on speciation rates are necessary in order to establish that population differentiation is a rate-limiting step of the speciation process (Harvey et al., 2019). Such tests are scarce and have only focused on vertebrates, finding a positive association in New World birds (Harvey et al., 2017), and no association in Australian lizards (Singhal et al., 2018). We encourage similar tests in seed plants at a global scale. Nevertheless, ours is the first study that we are aware of to clearly document such a pattern of greater population differentiation in the tropics for seed plants (see Martin & McKay, 2004 for a study in vertebrates).