Dendrograms and measures of functional diversity: a second instalment

The ecological literature has seen the increasing application of dendrogram-based measures of functional diversity, which capture how functional trait composition varies amongst assemblages (Roscher et al. 2004, Thompson et al. 2005, Ernst et al. 2006, Leps et al. 2006, Barnett and Beisner 2007, Barnett et al. 2007). Podani and Schmera (2006) recently raised a number of criticisms of these approaches, including the FD measure originally proposed by Petchey and Gaston (2002b). Petchey and Gaston (2007) (hereafter P and (2) to clarify the points over which P and five real communities have relatively similar (linear) relationships between species richness and functional diversity (Petchey and Gaston 2002b). In P only the relative differences among local assemblages from the same regional assemblage are pertinent. We consider absolute branch lengths unimportant because they are a function of the regional species pool chosen for the analysis (Petchey and Gaston 2002b). Consequently, finding that the absolute value FD is influenced by the choice of clustering method (Podani and Schmera 2006) seems irrelevant. More important is whether the relative differences between the FD of assemblages within a region are affected greatly by the clustering method used. It is certain that they will differ somewhat between clustering method and perhaps this will provide an important opportunity for distinguishing the relative usefulness of different distance measures and clustering methods. In the same way as Petchey et al. (2004) tested the performance of measures of functional diversity, one could test the performance of different distance measures and clustering methods. It is uncertain, however, that differences between clustering methods are sufficiently large to alter the broad conclusions, for example about the importance of species richness for functional diversity (Petchey and Gaston 2002b). P the FD of these local assemblages would be calculated as the branch length of the regional dendrogram required to join the species (Petchey and Gaston 2006, 2007). Any study of the functional importance of invasive species must include the identity and functional characteristics of the invasive species. As long as these characteristics (traits) are the same as those known for the resident species (see Agreement 4) then a regional dendrogram can be calculated. This regional dendrogram contains the resident and invasive species; from this it is relatively straightforward to find, for example, the addition in functional diversity that might result from an invasion without any violation of set monotonicity. The concept of functional diversity has many potential important applications in ecology, from linking species and ecosystems, to estimating the functional consequences of species extinctions, and informing about the processes by which communities assemble. It is thus vital that functional diversity is measured appropriately, and debate about how this is best achieved is a healthy component of ensuring that this happens and is to be encouraged. Such discussions should concern only scientific issues, a goal that we hope to have achieved here. Finally, we would like to apologise for any misunderstandings or misinterpretations to which we may inadvertently have contributed. OLP is a Royal Society University Research Fellow. KJG holds a Royal Society-Wolfson Research Merit Award.