Extended Data Fig. 3: A worked example for defining assembly processes.
From: The late rise of sky-island vegetation in the European Alps
A worked example for defining assembly processes within Viola. The coloured squares on the tips represent the current elevational and geographic states for each species. The pie charts on the internal nodes display the likelihoods of each state based on ancestral state reconstructions. All species that occur in the sky-islands (denoted by red squares) are considered sky-island species. Endemism status (green and yellow squares) allow us to differentiate between species that have assembled in the sky-islands through colonisation or through single speciation. Blue text on tip labels represent colonist species (for example V. palustris). Colonists are not endemic to the EAS (green squares) and their ancestors occurred at low elevations or in non-EAS regions. Orange text on tip labels represent species that originated via single speciation (for example V. thomasiana). Single-speciation species are endemic to the EAS (yellow squares) and their ancestors occurred at low elevations or in non-EAS regions, but they are not (yet) part of an EASI clade. Red tip labels represent cladogenetic species (for example V. cenisia). Cladogenetic species are endemic to the EAS (yellow squares) and their ancestors occurred at high-elevations and were EAS endemics. The size of a sky-island clade is determined by the node depth to which high-elevation and EAS endemism is reconstructed. In this example, all nodes between the tips and the starred node are reconstructed as high-elevation EAS endemics, and therefore all species below the starred node (with the exceptions of V. declinata and V. dacica) are considered to be part of a single clade that diversified in situ. V. declinata and V. dacica do not occur in the sky-islands, and thus are not considered part of the sky-island clade. The speciation events leading to their formation are not used in our calculation of cladogenetic rate. All other branching events within the clade are, however, used in the calculation of cladogenetic rate.
