Morphology versus molecules: the phylogenetic relationships of Sepsidae (Diptera: Cyclorrhapha) based on morphology and DNA sequence data from ten genes

• Published in: Cladistics Vol. 24; no. 6; pp. 902 - 916
• Main Authors: Feng-Yi Su, Kathy, Narayanan Kutty, Sujatha, Meier, Rudolf
• Format: Journal Article
• Language: English
• Published: Oxford, UK Blackwell Publishing Ltd 01.12.2008
• Subjects: Diptera; Orygma; Ropalomeridae; Sepsidae
• ISSN: 0748-3007; 1096-0031
• DOI: 10.1111/j.1096-0031.2008.00222.x

The Sepsidae is, with approximately 300 described species, a relatively small family of cyclorrhaphan flies whose behaviour, morphology, and development have been extensively studied. However, currently the only available tree for Sepsidae is more than 10 years old and was based entirely on morphological characters. Here, we present the results of parsimony and Bayesian analyses based on 75 species, ten genes, and morphology. Parsimony and Bayesian analyses produce largely congruent and well‐supported topologies regardless of whether indels are coded as 5th character states, as missing values, or all sites with indels are removed. The tree confirms the monophyly of Sepsidae and identifies the Ropalomeridae as its sister group. With regard to higher‐level relationships, we identify widespread conflict between the morphological and the DNA sequence data. The proposed hypothesis based on both partitions largely reflects the signal in the molecular data. Particularly surprising is the rejection of two relationship hypotheses with strong morphological support, namely the sister group relationship between Orygma and the remaining Sepsidae and the monophyly of the Sepsis species group. Our partitioned Bremer support (PBS) analyses imply that indel coding has a stronger effect on the relative performance of individual gene partitions than the exclusion of alignment‐ambiguous sequences or the location of a gene on the mitochondrial or nuclear genome. However, these analyses also reveal unexpectedly strong fluctuations in PBS values given that indel treatment has only a minor effect on tree topology and jacknife support. These unexpected fluctuations highlight the need for a comparative study across multiple data sets that investigates the influence of conflict and indel treatment on PBS values. © The Willi Hennig Society 2008.