**3.2. Limitations of current methods to link historical events with biological diversification**

of Malagasy vangids, early dispersal from Africa to Madagascar promoted an increase in species diversification rates by ecological opportunity followed by a decline through time suggesting an

**Figure 1.** A hypothetical example showing inference of dispersal and vicariance and shifts in diversification dynamics across the phylogenetic diversification of a taxa. Nodes labeled with circles denote the position in the tree where it was inferred a dispersal (square) or a vicariance event (circles). Nodes labeled with stars denote where shifts in macroevolutionary regimes occurred. These shifts might involve increase in speciation rates or decrease in extinction rates or shifts in diversification dynamics from a time-dependent process to diversity-dependent cladogenesis (star). Under a dispersal scenario, where a lineage colonizes a new region with ecological opportunity, it is expected that ecologically relevant traits evolved through a pattern of niche divergence (dash). Whether all dispersal events promote speciation through ecological opportunity, it is expected that all events generated the same pattern of niche-trait divergence. By contrast, in a vicariance event, where the emergence of geographical barriers (e.g., a mountain uplift) promotes allopatric speciation, it is expected that ecologically relevant traits evolved little (i.e., phylogenetic niche conservatism) or under a neutral (drift) process in response to the changing environment (cross). Similarly, it is expected that all vicariant events generated the same pattern of niche conservatism or neutral evolution. Many other dispersal or vicariance events can be nested in the phylogeny; therefore, the inference of an event does not preclude that subsequent

Although these studies suggest a historical association between one or several dispersal events and shifts in net species diversification rates, it is still not clear which mechanisms were involved in this link between biogeography and macroevolution. Even more, it is not clear exactly how these past biogeographic events facilitate speciation in some regions and

In contrast to the role of dispersal, vicariance has been associated as the only valid explanation for current distributional patterns and speciation for traditional biogeographers (e.g., Refs. [23, 45, 46]). However, few studies reviewed identified a vicariance event as a trigger of species diversification. Although these associations are congruent with the hypothesis of speciation

ecological limit for this adaptive radiation [57].

speciation events occur in sympatry or allopatry.

18 Pure and Applied Biogeography

not in others or the mode of speciation involved in each event.

Some studies only provide a weak evidence of a consistent link between historical biogeographical events and shifts in species diversification rates. In particular, these few studies did not test whether dispersal or vicariance events effectively generated increases in net species diversification rates (e.g., Refs. [4, 7, 9, 68]. For instance, Chaves et al. [9] suggested that Andean uplift promoted increases in species diversification in the *Adelomyia* hummingbird genus. However, they only provide a correlation between divergence times for *Adelomyia* clades with some phases of the Andean uplift. Here, it is necessary to test whether Andean clades diversified more than other clades or the entire group. Accordingly, it is possible to suggest with higher confidence that dispersal to the Andes promoted posterior cladogenesis likely through allopatric speciation.

The link between historical biogeography and species diversification need to be explicitly tested using a two-step framework. First, it is necessary to infer historical biogeographic events with a higher confidence using powerful statistical methods [34]. These methods are ideal to identify rare dispersal events (founder-event speciation; [34]) and allow us to incorporate explicit geological information (e.g., timing of emergence of island landmasses; see Ref. [48] for an example using Caribbean *Anolis* lizards). Many empirical datasets show that founder-event speciation have left a strong imprint on distributional patterns of many taxa [34]. However, it is necessary to develop more complex models allowing the possibility to incorporate life-history traits and how these traits can affect rates of dispersal between regions (e.g., Ref. [35]). After the historical inference is conducted, it is crucial to identify whether these same nodes where a dispersal or vicariance event was inferred exhibit a notable shift in diversification dynamics [2, 3, 66].

Several methods allow to detect specific shifts in diversification dynamics in phylogenies, particularly increases in diversification rates [36, 37, 43, 53, 54]. These methods can be divided into temporal or topological methods, depending on the phylogenetic dataset at hand [43]. These methods also allow us to identify shifts in macroevolutionary regimes (i.e., shifts in diversification process; e.g., bayesian analysis of macroevolutionary mixtures (BAMM) approach developed by [53]; but see Ref. [42] for a criticism about the statistical power). However, these phylogenetic methods are not integrated fully with methods of historical biogeographic inference. Therefore, the association between dispersal or vicariance events and notables shifts in species diversification or diversification dynamics remains narrative [23]. Therefore, new methodologies are necessary to establish with more confidence whether these biogeographical events promote changes in biological diversification. Even more, it is possible that unmeasured traits (e.g., a "hidden" key innovation) affecting speciation and/or extinction rates evolve in the same nodes of a phylogenetic tree where a dispersal or vicariance event was inferred. In these cases, it might be very hard to distinguish whether diversification was affected by the presence (or absence) of a trait or by the movement to a new region (see Ref. [75]).

More robust methods to link historical biogeography and macroevolution are necessary. However, these new phylogenetic methods only can detect these historical associations, but these are very limited to establish the potential ecological or evolutionary process underlying this association. I consider that an integration of ecological studies (including testable predictions of trait-niche evolution) with a historical biogeographic approach will be useful to discern these mechanisms promoting speciation. The association between a biogeographic event and shifts in macroevolutionary dynamics is not enough to understand how biodiversity was generated in many regions. For instance, it has been found in warblers (*Phylloscopus*) that habitat isolation was the first form of niche differentiation to evolve in sympatry, followed by prey-size selection and feeding strategies [74]. It will be important to evaluate whether lineages diversifying after a dispersal event exhibit more niche differentiation [47] than lineages diversifying after a vicariant event. This can help to establish whether the initial trigger of speciation in a lineage was the colonization of a new region and the subsequent habitat isolation in sympatry [12, 13].
