Table 1.

Common metrics of phylogenetic structure and their analogues for quantifying the terragenetic structure of landscapes

Phylogenetic metric	Description	Terragenetic equivalent	Interpretation	Calculation
Mean nodal phylogenetic distance	The mean of the number of nodes separating all pairwise combinations of species on a phylogeny (Gregory 2008)	Median nodal terragenetic distance	The number of nodes quantifies how many fragment separation events occurred in the history of a pair of fragments since they separated from their most recent ancestor fragment. Fragments separated by few fragment separation events are closely related	We calculated a pairwise distance matrix for all fragments on the terrageny, and report the median for terragenies because the distribution was skewed
Mean phylogenetic distance	The mean of the branch lengths separating all pairwise combinations of species on a phylogeny (Webb. 2002)	Median terragenetic distance	Branch lengths give an indication of how recently the two fragments separated from their most recent common ancestor fragment, with long branch lengths indicating fragments that have been separated for a long time period. Thus, fragments separated by large terragenetic distances have been isolated from each other for long time periods	A pairwise distance matrix for all fragments in the terrageny was calculated using the ‘cophenetic.phylo’ function in the R package ‘ape’ (Paradis. 2004). We report the median for terragenies because the distribution was skewed
Topological balance	The extent to which nodes on a phylogeny define subgroups of equal sizes (Mooers & Heard 1997). Balanced phylogenies tend to appear more symmetrical	Terragenetic balance	Terragenetic balance quantifies the degree of asymmetry in a terragenetic tree. A symmetrical tree would suggest that all fragments in a landscape are equally likely to separate into the same number of child fragments, whereas an asymmetrical tree would suggest that some fragments were more likely to separate into child fragments than others	Our terragenies had numerous polytomies, or situations where a fragment splits into > 2 child fragments, so we calculated terragenetic balance with the metric I′ (Purvis. 2002), a modified form of Fusco and Cronk's imbalance statistic (Fusco & Cronk 1995). I′ was calculated using the ‘fusco.test’ function in the R package ‘caper’ (Orme. 2012)
Evolutionary distinctiveness	The phylogenetic diversity of a clade split equally among its members (Isaac. 2007). Species in older, less-speciose clades are more evolutionarily distinct, while species in younger, diverse clades are less evolutionarily distinct	Terragenetic distinctiveness	High terragenetic distinctiveness indicates fragments that have been isolated from all other fragments for a long time period. It is greatest in fragments that have few siblings or have been separated from other fragments for long time periods	We used the function ‘ed.calc’ in the R package ‘caper’ (Orme. 2012)
Pagel's lambda	The strength of phylogenetic signal in species traits (Pagel 1999). A value of zero suggests trait evolution is independent of phylogeny, a value of one that traits are evolving according to Brownian motion, and intermediate values suggest an effect of phylogeny that is weaker than the Brownian model	Terragenetic Pagel's lambda	We treat fragment size as a ‘trait’ of a fragment, although other physical (e.g. fractal dimension, edge:area ratio) or biological (e.g. species richness, number of local endemics) features could equally be used. If fragments always separated into children that have equal traits, then we would expect closely related fragments to have similar trait values and terragenetic Pagel's lambda to be close to one	We used the function ‘pgls’ in the R package ‘caper’ (Orme. 2012)