Optimising linear taxon sequences derived from phylogenetic trees—a reply to Haston & al.

Numerous linear taxon sequences can be generated from large phylogenetic trees raising questions about what criteria should be used to select an optimal sequence for collections of specimens and published taxonomic outputs including field guides. Possible criteria and ways to implement them are discussed in relation to the sequence of Angiosperm families proposed by Haston & al. (2007). We define a particular spatial arrangement in one or more dimensions of the terminal nodes of a tree as a projection of that tree; a projection method is the set of rules by which the projection is achieved. We show that the projection of the Angiosperm phylogeny to a linear sequence, proposed by Haston & al. (2007), is not ideal because it is incompletely optimized for incompletely specified criteria. Tree projection methods that promote stability and incorporation of morphological data used in herbaria and other identification guides are proposed.