Aspects of the phylogeny, biogeography and taxonomy of galaxioid fishes Jonathan Michael Waters, BSc. (Hons.) Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy, / 2- Oo ( 01 f University of Tasmania (August, 1996) Paragalaxias dissim1/is (Regan); illustrated by David Crook Statements I declare that this thesis contains no material which has been accepted for the award of any other degree or diploma in any tertiary institution and, to the best of my knowledge and belief, this thesis contains no material previously published o:r written by another person, except where due reference is made in the text. This thesis is not to be made available for loan or copying for two years following the date this statement is signed. Following that time the thesis may be made available for loan and limited copying in accordance with the Copyright Act 1968. Signed Summary This study used two distinct methods to infer phylogenetic relationships of members of the Galaxioidea. The first approach involved direct sequencing of mitochondrial DNA to produce a molecular phylogeny. Secondly, a thorough osteological study of the galaxiines was the basis of a cladistic analysis to produce a morphological phylogeny. Phylogenetic analysis of 303 base pairs of mitochondrial cytochrome b _supported the monophyly of Neochanna, Paragalaxias and Galaxiella. This gene also reinforced recognised groups such as Galaxias truttaceus-G. auratus and G. fasciatus-G. argenteus. In a previously unrecognised grouping, Galaxias olidus and G. parvus were united as a sister clade to Paragalaxias. In addition, Nesogalaxias neocaledonicus and G. paucispondylus were included in a clade containing G. brevipinnis-like species. A high level of intraspecific diversity was detected between geographic isolates of the South African G. zebratus and the widespread G. maculatus. The intraspecific divergences are the highest yet reported for fish cytochrome b, suggesting that these taxa may represent species complexes. Phylogenetic analysis of 16S rDNA supported the monophyly of Paragalaxias, Neochanna, Galaxiella, and the G. truttaceus-G. auratus group. Again, G. parvus and G. olidus formed a clade as the sister of Paragalaxias and G. cleaveri was placed as the sister group of the New Zealand Neochanna (a grouping weakly supported in the cytochrome b analysis). A large G. brevipinnis clade including Nesogalaxias was supported, and substantial genetic divergence was detected within G. maculatus and G. zebratus. Despite their different evolutionary properties, separate mitochondrial genes produced largely congruent phylogenetic trees, reflecting their common history. Deep phylogenetic splits within the Galaxiinae generally received low bootstrap support in the molecular analyses. Similarly, there was little resolution of galaxioid relationships. This is probably because nucleotide sites that are free to vary become saturated with changes over time. However, one grouping that was weakly supported in both separate molecular analyses received substantial support from the combined molecular analysis. Specifically, the South American Brachygalaxias bullocki formed a clade with the Australian Galaxiella. While there have been several morphological studies of the Galaxiinae, until now there has been no cladistic synthesis of these data. In this study, a parsimony analysis of 51 characters in 18 galaxiine specfos supported some widely accepted groups: Galaxiella, Paragalaxias, Neochanna, and Galaxias truttaceus-G. auratus. In addition, the clades [G. cleaveri, Neochanna] and [G. zebratus [Brachygalaxias, Galaxiella]] received Summary strong support. Well supported morphological conclusions were corroborated by the molecular and global parsimony analyses, with the exception of the position of G. zebratus. The molecular data did not conclusively resolve the position of G. zebratus, but weakly supported the above morphological placement. Galaxiine taxonomy is revised to better represent galaxiine phylogeny, as determined by both morphological and molecular analyses. Specifically, G. cleaveri is removed from Galaxias and placed in the genus Neochanna. Similarly, the three species currently placed in Galaxiella are allocated to Brachygalaxias. Biogeographical explanations are proposed to account for galaxiine distribution in the light of hypothesised phylogenetic relationships and molecular clock calibrations. The wide distribution of G. maculatus is probably due to oceanic dispersal, but the high divergences within this species indicate that its dispersal powers are more limited than previously_ suggested. The presence of related mudfish on either side of the Tasman Sea is best explained by marine dispersal. Similarly, Nesogalaxias neocaledonicus is a descendant of a G. brevipinnis-like ancestor that probably colonised New Caledonia in the Pliocene. On the other hand, it is likely that Galaxiella and Brachygalaxias represent an ancient Gondwanan radiation. It is hypothesised that G. zebratus is an ancient Gondwanan ancestor of this clade. Acknowledgements I sincerely thank my supervisor, Robert White, for his unparalleled support and guidance throughout the course of this project. Many thanks to the following people who gave up their time to collect samples on my behalf, organised collections or provided samples from their own collections: Jim Cambray (Rhodes University, Albany Museum, South Africa); Paul Skelton (JLB Institute of Ichthyology, South Africa); Dean Impson (Western Cape Nature Conservation, South Africa); Conor Nolan, R. Maddocks, M. Marsh, and J. Smith (Fisheries Department, Falkland Islands); David Crook, Ron Mawbey, Jenny Ovenden, John Purser and Robert White (University of Tasmania); P. Boxall, R. Gasior, Paul Humphries, Brett Mawbey, and Andrew Sanger (Inland Fisheries Commission, Tasmania); Howard Gill (Murdoch University, Western Australia); Klause Busse (Museum Alexander Koenig, Germany); Karl-Hermann Kock (BFA fur Fischerei, Germany); Brendan Hicks (University of Waikato, New Zealand); Charles Mitchell (New Zealand); Tony Eldon (NIWA, New Zealand); Mark Lintermans (Parks and Conservation Service, ACT); Hugo Campos (Universidad Austral de Chile, Chile); Christine Pollabauer and M. Boulet (Service de !'Environment, New Caledonia); Alastair Graham (CSIRO Division of Fisheries, Hobart); Gillian von Bertouch (CCAMLR, Hobart). Some collections of live fishes for this study were made under permits from the Inland Fisheries Commission of Tasmania. Jeremy Austin, Darren Brasher, Chris Burridge, Sue Dobson, Sharee Mccammon, Kellie Robinson, Adam Smolenski and Tony van den Enden are thanked for their assistance and company in the lab. In addition, Brenda Bick, Sherrin Bowden, Alan Dumphy, Kate Hamilton, Richard Holmes, Wayne Kelly, Ron Mawbey, Barry Rumbold, Adam Stephens and Kit Williams are thanked for their constant help. Adam Smolenski, Rene Vaillancourt and Robert White kindly gave me access to their computers for the analysis of molecular data. Nick Elliot, Peter Grewe, Bronwyn Innes and Bob Ward (CSIRO Division of Fisheries) are thanked for their assistance with automated sequencing. Eric Anderson, Jeremy Austin, Doug Begle, Clive Burrett, Jim Cambray, Jean Chazeau, Peter Davies, Howard Gill, Robin Gutell, Bob Hill, Paul Humphries, Jerry Lim, Mike Pole, and Andrew Sanger are thanked for their invaluable information, advice and assistance. Acknowledgements I thank Ephrime Metillo, Robby Gaffney, Jeremy Austin, Darren Brasher, David Donald, Chris Burridge and Kellie Robinson for their friendship. Tim Reid and Kate Hamilton provided novel insights into art-house cinema. Wayne Kelly introduced me to the world of the circuit-class. Captain Ron Mawbey had sufficient foresight to elect me Vice Captain of the Zoology XI; the CSIRO XI provided evidence of the futility of training. Special thanks go to Jerry and Ephrime for their fine Asian cuisine, to Robert for the free lunches, and to Adi for reinforcing my grasp of the English language. David Crook and Tim Reid provided me with ample opportunity to demonstrate my aptitude for fieldwork, both in the river and out at sea. Finally, I thank Margaret, and my family, especially Mum and Dad, for their continued support and encouragement. Table of Contents 1. General Introduction 1.1 Salmoniform intrarelationships 1 1.2 The galaxioid fishes 1 1.3 Fossil record 9 1.4 Galaxioid intrarelationships 11 1.5 Generic placement of the galaxiines 12 1.6 Biogeographical theories 13 1.7 Galaxiine phylogeny 15 1.8 Aims 16 2. Molecular Systematics of the Galaxioidea I 2.1 Introduction 20 2.1.1 Mitochondrial genes 20 2.1.2 Character weighting and homoplasy 21 2.1.3 Sequence alignment and gap weighting 23 2.1.4 Congruence between separate mtDNA data sets 23 2.1.5 Incorrect phylogenetic inferences from molecular data 24 2.1.6 Phylogeny reconstruction 27 2.1 .7 Assessing phylogenetic confidence 29 2.2 Materials and Methods 32 2.2.1 PCR with universal primers 32 2.2.2 Collection and preservation of specimens 33 2.2.3 DNA extractions 33 2.2.4 PCR primers 37 2.2.5 PCR amplification 38 2.2.6 Purification of PCR products 39 2.2.7 DNA sequencing 41 2.2.8 Electrophoresis and autoradiography 43 2.2.9 Sequence alignment and phylogenetic analysis 44 2.3 Results 48 2.3.1 Cytochrome b gene phylogeny 48 2.3.2 16S rRNA gene phylogeny 66 2.3.3 Combined mtDNA sequence phylogeny of the galaxiines 84 2.3.4 Combined mtDNA sequence phylogeny of the galaxioids 89 2.4 Discussion 91 3. Morphological Systematics of the Galaxiinae 3.1 Introduction