Percomorph Phylogeny - a Survey of Acanthomorphs and a New Proposal

Abstract

The interrelationships of acanthomorph fishes are reviewed. We recognize seven monophyletic terminal taxa among acanthomorphs: Lampridiformes, Polymixiiformes, Paracanthopterygii, Stephanoberyciformes, Beryciformes, Zeiformes, and a new taxon named Smegmamorpha. The Percomorpha, as currently constituted, are polyphyletic, and the Perciformes are probably paraphyletic. The smegmamorphs comprise five subgroups: Synbranchiformes (Synbranchoidei and Mastacembeloidei), Mugilomorpha (Mugiloidei), Elassomatidae (Elassoma), Gasterosteiformes, and Atherinomorpha. Monophyly of Lampridiformes is justified elsewhere; we have found no new characters to substantiate the monophyly of Polymixiiformes (which is not in doubt) or Paracanthopterygii. Stephanoberyciformes uniquely share a modification of the extrascapular, and Beryciformes a modification of the anterior part of the supraorbital and infraorbital sensory canals, here named Jakubowski's organ. Our Zeiformes excludes the Caproidae, and characters are proposed to justify the monophyly of the group in that restricted sense. The Smegmamorpha are thought to be monophyletic principally because of the configuration of the first vertebra and its intermuscular bone. Within the Smegmamorpha, the Atherinomorpha and Mugilomorpha are shown to be monophyletic elsewhere. Our Gasterosteiformes includes the syngnathoids and the Pegasiformes (Pegasus) and Indostomiformes (Indostomus), two groups which are shown to be immediately related to syngnathoids by modifications of the gill filaments and their skeletal supports. Monophyly of the Gasterosteiformes in this sense is justified by several characters. We are unable to resolve the interrelationships among the five subgroups of Smegmamorpha. The remaining percomorphs are the Perciformes (including Caproidae), Scorpaeniformes, Dactylopteriformes, Pleuronectiformes and Tetraodontiformes; we have found nothing to indicate that Percomorpha in that sense are monophyletic, although our survey does not cover Tetmodontiformes. We believe that Scorpaeniformes and Pleuronectiformes are nested within Perciformes, but again have found nothing to indicate that Perciformes in this expanded sense are monophyletic. We recommend extending the Percomorpha to include the Atherinomorpha (and other smegmamorphs), and argue that this larger group is monophyletic. A scheme of relationships of the seven groups Lampridiformes, Polymixiiformes, Paracanthopterygii, Stephanoberyciformes, Zeiformes, Beryciformes and the expanded Percomorpha is presented and supported by apomorphies. New names for higher acanthomorph taxa are proposed as follows: Euacanthomorpha (Acanthomorpha minus Lampridiformes), Holacanthopterygii (Eucanthomorpha minus Polymixiiformes), and Euacanthopterygii (Acanthopterygii minus Stephanoberyciformes). Monophyly of Beryciformes s.l. (including stephanoberyciforms) is rejected because Beryciformes s.s. share several apomorphies with the expanded Percomorpha, all of which are absent in Stephanoberyciformes. The Zeiformes are the most problematic of the acanthomorph groups; with the characters that we have been able to assess, the zeiforms are placed most parsimoniously as the sister-group of Euacanthopterygii (i.e., between stephanoberyciforms and beryciforms on the cladogram), but we do not propose a name for the taxon so formed. There is a disturbing incidence of homoplasy in the characters that we have investigated in acanthomorphs.