Molecular and osteological heron phylogenies: Sources of incongruence

Abstract

Payne and Risley's (1976) comparison of 33 osteological characters of herons was the first cladistic estimate of heron phylogeny. Among their findings were two major clades: (1) Boat-billed Heron (Cochlearius cochlearius), night-herons, and bitterns; and (2) tiger-herons and day-herons. In contrast, more recent DNA-DNA hybridization comparisons, cladistic analyses of vocalizations, and mtDNA sequence data portray a more asymmetric phylogeny, with day-herons and night-herons forming a clade with bitterns as their sister group, and tiger-herons and the Boat-billed Heron branching basally. To explore the source of the disagreement between these phylogenetic estimates, we reanalyzed the osteological data using modern cladistic methods and compared the results with the DNA-DNA hybridization tree using taxonomic congruence analysis. Character-by-character comparisons between trees and among lineages within trees suggest that similar cranial morphology in the relatively unrelated tiger-herons and day-herons has resulted in the misleading attraction of these two lineages in osteological estimates of phylogeny. Apparent convergence in bill morphology and modifications of orbital structures for nocturnal feeding in night-herons and Boat-billed herons have led to further disagreement between data sets. In part, problems in the osteological data stem from the relatively small character matrix of Payne and Risley (1976), but ultimately they may derive from using highly adaptive characters to reconstruct phylogeny. In this case, the cranial characters are functionally correlated as part of the piscivorous heron Bauplan. As such, they relate to the forces responsible for speciation and divergence in the early history of the group but may not be useful for phylogenetic inference. The discovery of bias in cranial characters underscores the value of taxonomic congruence analysis and the need to explore cases of phylogenetic incongruence.