Choice between phylogram and chronogram can have a dramatic impact on the location of phylogenetic diversity hotspots

Aim: Biodiversity is declining rapidly at a global scale, fuelling the need for urgent conservation action. With limited resources available, areas for conservation often need to be prioritized. Phylo-diversity metrics such as phylogenetic diversity (PD) and phylogenetic endemism (PE) combine phylogenies and spatial distribution data and can be used to rank candidate areas for conservation. Phylogenetic trees underlying phylo-diversity metrics are available in two main forms, phylograms and chronograms. We examined the use of these two types of phylogenies in spatial studies and tested whether there is a difference in the resultant hotspot localities when employing either for analysing the same dataset. Location: Australia and New Zealand. Methods: We reanalysed three published datasets and one new dataset of geocoded species presence data and DNA sequence data. In all cases, we inferred first a phylogram and then time-calibrated it to keep tree topology constant. We conducted spatial analyses with equal area grid cells and compared the distribution of PD and PE hotspots (top 5% values) as well as Categorical Analysis of Neo- and Palaeo-Endemism between phylogram and chronogram. Results: We found that both types of trees were frequently used in spatial studies but that there is no explicit discussion in the relevant community about either choice. Differences in the locations of hotspots between phylogram and chronogram ranged from 11% to 91%, with no clear correlation between tree shape and effect size. Main conclusions: We demonstrate for the first time that in some cases the choice between phylogram and chronogram can have a dramatic effect on the resulting hotspot distributions. Branch lengths of phylograms indicate divergence in features (or at least in those used to infer the phylogeny), while those of chronograms indicate time since divergence. The interpretation of the resulting phylo-diversity hotspots should reflect this difference.