Treemap Versus BPA (Again): A Response to Dowling

Abstract

TreeMap is a computer program for analysing host-parasite cospeciation. We respond to Dowling’s (Cladistics, 18: 416-435) recent comparison of TreeMap and Brooks Parsimony Analysis (BPA) by showing that Dowling’s comparison suffers from several mistakes and flaws. We discuss the problems with both BPA and TreeMap, and show that BPA incorrectly counts the true number coevolutionary events more often than TreeMap 1. We also discuss the two main limitations of TreeMap 1 correctly identified by Dowling, namely its inability to handle widespread parasites, and its coarse optimality criterion (the number of cospeciation events). We suggest a simple fix for widespread parasites. The newly released TreeMap 2 uses a more sensitive optimality criterion than TreeMap 1, addressing Dowling’s second concern.