bs_bs_banner Biological Journal of the Linnean Society, 2014, 111, 521–540. With 6 figures A multigene molecular assessment of cryptic biodiversity in the iconic freshwater blackfishes (Teleostei: Percichthyidae: Gadopsis) of south-eastern Australia MICHAEL P. HAMMER1,2,3*, PETER J. UNMACK4,5, MARK ADAMS1,2, TARMO A. RAADIK6 and JERALD B. JOHNSON4 1 Evolutionary Biology Unit, South Australian Museum, North Terrace, SA 5000, Australia Australian Centre for Evolutionary Biology and Biodiversity, School of Earth and Environmental Science, The University of Adelaide, Adelaide, SA 5005, Australia 3 Curator of Fishes, Museum and Art Gallery of the Northern Territory, PO Box 4646, Darwin, NT 0801, Australia 4 WIDB 401, Department of Biology, Brigham Young University, Provo, UT 84602, USA 5 Institute for Applied Ecology and Collaborative Research Network for Murray-Darling Basin Futures, University of Canberra, Canberra, ACT 2601, Australia 6 Aquatic Ecology Section, Arthur Rylah Institute for Environmental Research, Department of Environment and Primary Industries, 123 Brown Street, Heidelberg, VIC 3084, Australia 2 Received 6 September 2013; revised 22 October 2013; accepted for publication 22 October 2013 Freshwater biodiversity is under ever increasing threat from human activities, and its conservation and management require a sound knowledge of species-level taxonomy. Cryptic biodiversity is a common feature for aquatic systems, particularly in Australia, where recent genetic assessments suggest that the actual number of freshwater fish species may be considerably higher than currently listed. The freshwater blackfishes (genus Gadopsis) are an iconic group in south-eastern Australia and, in combination with their broad, naturally divided distribution and biological attributes that might limit dispersal, as well as ongoing taxonomic uncertainty, they comprise an ideal study group for assessing cryptic biodiversity. We used a multigene molecular assessment including both nuclear (51 allozyme loci; two S7 introns) and matrilineal markers (cytb) to assess species boundaries and broad genetic substructure within freshwater blackfishes. Range-wide examination demonstrates the presence of at least six candidate species across two nominal taxa, Gadopsis marmoratus and Gadopsis bispinosus. Phylogeographical patterns often aligned to purported biogeographical provinces but occasionally reflected more restricted and unexpected relationships. We highlight key issues with taxonomy, conservation, and management for a species group in a highly modified region. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540. ADDITIONAL KEYWORDS: conservation – cryptic species – drainage divides – ESU – phylogeography – sea level changes. INTRODUCTION Australia is regarded as one of the world’s top 17 megadiverse countries (Williams et al., 2001) and this *Corresponding author. E-mail: [email protected] is generally reflected in the species richness and levels of endemism displayed for many organismal groups (Chapman, 2009). Among non-marine vertebrates, however, there is one glaring exception. Australia’s freshwater fish fauna has long been regarded as depauperate compared to those found in other © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 521 522 M. P. HAMMER ET AL. regions of comparable size and climate (Lundberg et al., 2000; Allen, Midgley & Allen, 2003), such as continental USA, which has over three times as many described species (Page & Burr, 1991). Traditional explanations for this anomaly have invoked a range of climatic, topographic, and biogeographical factors (Allen et al., 2003). However, an alternate explanation, originally proposed by Lundberg et al. (2000), and more recently taken up by ourselves (Adams et al., 2013; Hammer, Adams & Hughes, 2013a; Unmack, 2013), is that Australia’s low number of species also reflects the level of detailed taxonomic effort devoted to this neglected group. In other words, we hypothesize that the Australian freshwater fish fauna is characterized by high levels of cryptic biodiversity (sensu Beheregaray & Caccone, 2007). Although most of the obviously-distinctive morphotypic forms among the Australian freshwater fishes have been described (Allen et al., 2003), there have been few detailed assessments for a considerable number of species whose geographical distributions span multiple river basins and even major biogeographical regions (Unmack, 2001, 2013). In comparison, freshwater fishes displaying a similar geographical pattern in the USA have typically been shown by detailed taxonomic investigation to contain cryptic species (Near & Benard, 2004; Berendzen, Olson & Barron, 2009; April et al., 2011). Supporting this expectation, baseline molecular genetic studies have detected likely candidate species in a large proportion of the geographically-widespread Australian freshwater fish ‘species’ surveyed to date (Hammer et al., 2013a). In the present study, we assess the generality of our prediction that current species counts for the Australian freshwater fish fauna under-estimate the actual numbers by at least 50% (Hammer et al., 2013a; Unmack, 2013) by exploring cryptic biodiversity in one of the more prominent groups, the freshwater blackfishes (genus Gadopsis). Two species are currently recognized; the river blackfish Gadopsis marmoratus Richardson, which occurs in the extensive Murray-Darling Basin, and coastal systems from eastern Victoria to South Australia and northern Tasmanian, and the two-spined blackfish Gadopsis bispinosus Sanger, which is limited to upland streams within the southern Murray-Darling Basin (Fig. 1). Gadopsis marmoratus reaches a moderate size (i.e. 350–625 mm and approximately 5 kg) and, accordingly, has greater recreational and cultural value for angling and eating (Lake, 1967; Jackson et al., 1996) than the smaller G. bispinosus (which grows to approximately 350 mm; Lintermans, 2007). Freshwater blackfishes are mysterious fish and have proved to be an enigmatic group with regard to both broader phylogentic affinities and taxonomy. Gadopsis species are largely nocturnal and display cryptobenthic behaviour. An unusual combination of anatomical features led to early hypotheses of close relationships to various marine suborders such as Blennioidei, Gadoidei, Ophidioidei, and Trachinoidei (Ovenden, White & Sanger, 1988), until Johnson (1984) placed them in the family Percichthyidae. Subsequent molecular work supported this hypothesis (Jerry, Eliphinstone & Baverstock, 2001; Near et al., 2012), with Gadopsis placed as the first branching member of the family (Near et al., 2012). Their phylogenetic relationships within Percichthyidae imply an ancient history (early Tertiary or older). The species-level taxonomic history of freshwater blackfishes has also been confounded, in this instance by limited, variable morphological characters for consistent discrimination. Gadopsis marmoratus was first described with the ambiguous type locality of ‘rivers in the southern parts of Australia’ (Richardson, 1848: 123). de Castelnau (1872) speculated that there may be additional species in Victoria and subsequently Gadopsis gracilis McCoy and Gadopsis gibbosus McCoy were recognized from the Yarra and Bunyip rivers, respectively, followed by Gadopsis fuscus Steindachner from the ‘freshwaters of South Australia’. Thereafter, Ogilby (1913) synonymized all these species under G. marmoratus. Parrish (1966) informally described Gadopsis tasmanica from Tasmania, which Sanger (1984) subsequently determined to be conspecific with mainland populations of G. marmoratus. Finally, Sanger (1984) described G. bispinosus on the basis of differences in dorsal spine counts and distinctive colour patterns, resulting in the two-species taxonomic framework currently accepted (Allen et al., 2003; Hoese et al., 2007). Nevertheless, taxonomic speculation continues to plague Gadopsis on two fronts. First, other undescribed forms have been suggested in the literature. It has long been known that G. marmoratus from southern Victoria and Tasmania grow much larger than those from the Murray-Darling Basin (Ogilby, 1913; Lintermans, 2007), and Parrish (1966) and Sanger (1986) have recognized fish from western Victoria as having lower dorsal spine counts. Second, several studies have examined genetic variation using allozymes (Sanger, 1986; Ryan, Miller & Austin, 2004) and mitochondrial (mt)DNA variation (Ovenden et al., 1988; Waters, Lintermans & White, 1994; Miller et al., 2004), all of which supported the distinctiveness of northern populations and, to some extent, those in western Victoria. However, combined morphological and molecular studies have been limited by a lack of comprehensive geographical coverage, which has hindered interpretations and left unresolved the status of all putative undescribed forms. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES 523 Figure 1. Locality data for all Gadopsis samples examined. For the corresponding site details, see Table 1. The shaded area identifies the known distribution of the two species currently recognized within Gadopsis. The symbol shape refers to the three deeper lineages identified, which represent ‘northern’ G. marmoratus (squares), ‘southern’ G. marmoratus (circles), and G. bispinosus (triangles), with the different shadings representing each candidate species. As a prominent inhabitant of the streams and rivers of south-eastern Australia, Gadopsis has been relatively well studied ecologically, with several lifehistory attributes being revealed that could contribute to their decline after environmental change. First, freshwater blackfishes are regarded as habitat specialists, typically occurring in areas with perennial flow, relatively low salinity, and high levels of physical cover (Koehn, 1987; Lloyd, 1987; Bond & Lake, 2003). Second, they have slow growth, low fecundity, large demersal larvae, specific spawning sites (i.e. confined spaces such as hollow logs and boulder crevices), and exhibit male parental care (Jackson, 1978; Khan, Wilson & Khan, 2003; O’Connor & Zampatti, 2006). Third, as high-order predators, consuming small fishes, crustaceans, and invertebrates, they are affected by the overall health of aquatic ecosystems (Koehn & O’Connor, 1990; Harris & Silveira, 1999). Finally, individual blackfish are reported to have restricted movement patterns, including small home ranges (< 30 m) and site fidelity (Lintermans, 1998; Khan, Khan & Wilson, 2004), with radio-tracking documenting some localized latitudinal and longitudinal movement of up to 200 m (Koster & Crook, 2008). This combination of intrinsic characteristics combined with perturbations caused by hydrological changes, drought, clearance of surrounding lands and riparian zones (e.g. leading to siltation), in-stream modifications (e.g. snag removal, weir construction), fish introductions, and overfishing has seen declines © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 524 M. P. HAMMER ET AL. in freshwater blackfish distribution and abundance, especially in the Murray-Darling Basin (Morris et al., 2001; Hammer, Wedderburn & van Weenan, 2009). Although neither species is considered threatened nationally, G. bispinosus is protected in the Australian Capital Territory, and G. marmoratus is protected in Queensland and South Australia and has two regional populations conservation listed; the upper Wannon River in Victoria as Critically Endangered (DSE, 2013) and the Snowy River in New South Wales as Endangered (FSC, 2008). Gadopsis species in Victoria and Tasmania have size and bag limits, with a closed season in southern Victoria. There are obvious conservation implications should either of the currently described freshwater blackfishes be found to harbour additional cryptic species (Bickford et al., 2007; Vanhove et al., 2012). The present study presents a spatially comprehensive, multigene assessment of Gadopsis, with the aim of answering three key questions. First, is there any compelling genetic evidence for the existence of additional candidate species, as predicted by our ‘cryptic biodiversity’ hypothesis for the Australian freshwater fish fauna? Our primary aim here is to identify all likely and potential species-level differences among populations to facilitate future taxonomic and molecular appraisals. Second, does the genetic diversity present within each Gadopsis candidate species have a strong phylogeographical component? Based on aspects of their biology (e.g. sedentary, habitat specificity, low fecundity, demersal larvae), we predict that all Gadopsis species should display phylogeographical substructuring. Third, are there any major conservation management issues emanating from our conclusions? MATERIAL AND METHODS TAXON SAMPLING Our sampling strategy was designed to cover the entire range of both Gadopsis species, with the goal of sampling as many separate drainages as possible. A total of 151 individuals from 74 sites (Fig. 1) were sampled for variation in the mitochondrial cytochrome b (cytb) gene, usually with two individuals per population (Table 1). For allozymes, 200 individuals from 69 sites were examined (some sites lacked frozen material or were geographically close to other included populations), with a mean sample size of 2.9 (Table 1). A subset of 72 individuals representing all key genetic lineages was sequenced for two introns of the nuclear S7 gene (Table 1), thus providing an additional independent assessment of the phylogenetic affinities among lineages. Fish were euthanized with clove oil, then sampled (whole or via a small lateral tissue section), with tissues snap frozen in liquid nitrogen and subsequently stored at −70 °C. Fish providing tissue samples or individuals from the same localities were retained as voucher specimens, and have been deposited at the Australian, South Australian and Victorian museums. These samples can be identified based on their station code (Table 1). For two sites with locally threatened populations (sites 47 and 48), fin clips were taken from live fish, which were then returned to the point of capture. ALLOZYME ELECTROPHORESIS Muscle homogenates were subjected to allozyme electrophoresis on cellulose acetate gels (Cellogel™), as described previously (Richardson, Baverstock & Adams, 1986). Thirty-six enzymes or non-enzymatic proteins produced zymograms of sufficient intensity and resolution to permit allozymic interpretation: ACON, ACP, ACYC, ADA, ADH, AK, ALD, AP, CK, ENOL, EST, FDP, FUM, GAPD, GLO, GOT, GP, GPD, GPI, GSR, IDH, LDH, MDH, ME, MPI, NDPK, PEPA, PEPB, PEPD, PGAM, 6PGD, PGK, PGM, PK, TPI, and UGPP. Details of enzyme and locus abbreviations, enzyme commission numbers, electrophoretic conditions, and stain recipes are provided in Richardson et al. (1986) and Hammer et al. (2007). Alphabetic and numerical designations were assigned to allozymes and multiple loci respectively, both in order of increasing electrophoretic mobility (i.e. Acona, Aconb, Adh1, Adh2). DNA ISOLATION, AMPLIFICATION, AND SEQUENCING Total DNA was obtained from approximately 0.25 cm3 of caudal fin or muscle via DNeasy Tissue Extraction Kits (Qiagen Inc.). We amplified the cytb gene and the first 28 bp of the Thr tRNA via polymerase chain reaction (PCR) using two primers that flanked the region. Most samples were amplified using two sets of overlapping primers because sequencing of the full gene was problematic. The primers GadF TTCAACTATAAGAACTAGAAT and HD.Gad.628 TTT GTRTTTGAGTTTAAKCC and Gad.505 TCAGTAG ATAATGCTACTCT and PP.Thr.41 AGGATTTTAACC TCTGGCGTCCG amplified all but the first four bases of cytb plus 28 bp of the Thr tRNA (collectively hereafter simply referred to as cytb). We also amplified the first two introns of the nuclear gene S7 via nested PCR using the primers: 1F TGGCCTCTTCC TTGGCCGTC and 3R GCCTTCAGGTCAGAGTTCAT in the first reaction followed by 1F.2 CTCTTCCTTG GCCGTCGTTG and 2R.67 TACCTGGGARATTCC AGACTC and 2F.2 GCCATGTTCAGTACCAGTGC and 3R GCCTTCAGGTCAGAGTTCAT in the second reactions. Primers 1F and 3R are from Chow & © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 Station code TR02-24 PU02-63 PU02-64 PU02-66 n/a PU02-81 PU02-97 PU02-101 PU02-95 PU02-72 TAS05-07 TAS03-09 TAS03-10 TAS03-11 TAS03-17 TAS03-20 TAS03-03 TAS03-04 TAS03-27 TAS03-23 PU02-78 PU02-105 PU02-82 PU03-05 TR02-268 TR02-16 TR02-210 PU02-85,PU02-108 TR02-373 PU02-110 PU02-109 FISH93:Gell PU03-08/09 PU02-112, PU03-07 FISH93:Darl PU09-122 PU00-19 FISH93:Mud PU00-12,02–117/118 Site 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 Back Creek, Noorinbee North, VIC Delegate River, Delgate, NSW Brodribb River, VIC Haunted Stream, VIC Thompson River, VIC Latrobe River, Noojee, VIC Greig Creek, Yarrum, VIC Tin Mine Creek, VIC Deep Creek, Forster, VIC Blackfish Creek, Wilsons Prom, VIC Styx River, Bushy Park, TAS Wye River, Swansea, TAS North George River, TAS Ansons River, Ansons Bay, TAS Boobyalla River, Winnaleah, TAS Great Forester River, Scottsdale, TAS Minnow River, Beulah, TAS Leven River, Gunns Plains, TAS Black River, Mawbanna, TAS Relapse Creek, TAS Turtons Creek, VIC Minnieburn Creek, VIC Tarago River, VIC Diamond Creek, Tonimbuk, VIC Donnellys Creek, Healesville, VIC Running Creek, Kinglake, VIC Lerderderg River, VIC Barwon River, Winchelsea, VIC Kuruc-A-Ruc Creek, Dereel, VIC Ford River, VIC Loves Creek, Gellibrand, VIC Gellibrand River, Gellibrand, VIC Brucknells Creek, Naringal East, VIC Mount Emu Creek, Panmure, VIC Darlots Creek, Haywood, VIC Bridgewater Lakes, VIC Stokes River, Digby, VIC Muddy Creek, Hamilton, VIC Wannon River, Grampians NP, VIC Locality Table 1. Locality data for all Gadopsis populations examined East Gippsland Snowy Snowy Tambo Thompson Latrobe South Gippsland South Gippsland South Gippsland South Gippsland Derwent East East East Piper Piper Mersey Smithton Smithton Arthur South Gippsland Bunyip Bunyip Bunyip Yarra Yarra Werribee Barwon Corangamite Otway Otway Otway Hopkins Hopkins Portland Portland Glenelg Glenelg Glenelg Drainage 2 2 1 2 2 2 2 1 2 2 2 2 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 3 2 3 2 2 2 4 Cytochrome b 1 2 1 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 2 1 1 1 1 2 S7 4 4 1 5 0 5 5 1 5 2 2 2 0 1 1 1 2 2 2 2 5 2 5 2 2 2 6 2 2 0 6 4 10 2 4 0 2 2 4 Allozymes −37.42944 −37.04750 −37.61278 −37.45278 −37.86722 −37.88194 −38.44917 −38.62500 −38.60806 −39.02194 −42.70139 −41.95028 −41.28333 −41.18028 −41.03167 −41.20750 −41.42861 −41.26944 −40.99139 −41.16056 −38.53444 −38.23556 −37.95694 −38.00472 −37.63639 −37.49444 −37.61722 −38.27917 −37.83750 −38.73528 −38.50472 −38.52000 −38.39222 −38.33667 −38.14694 −38.32075 −37.80028 −37.77167 −37.37361 Latitude 149.20750 148.81167 148.67472 147.76556 146.40806 145.89250 146.68361 146.32333 146.20861 146.41778 146.90750 147.95361 148.00000 148.14500 147.82167 147.55833 146.43167 146.03028 145.37445 145.43583 146.24611 145.83445 145.91417 145.73222 145.53333 145.24972 144.42222 143.97584 143.79861 143.42056 143.55056 143.53833 142.80917 142.72889 141.77083 141.40469 141.52945 141.95917 142.49861 Longitude SEV SEV SEV SEV SEV SEV SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SEV SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SBA SWV#1 SWV#1 SWV#2 NGW NGW NGW NGW Cand spp CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 525 FISH93:Gr SE03-11 TR02-194 SE02-103 FISHADD6:C43 TR02-172 FISHY4:Ang1+ ML05-03 ML05-01 SE03-05 PU00-08 PU00-06 TR02-245 TR02-236 TR02-230 PU00-05 PU99-80 TR02-312 TR02-295 PU99-82 TR02-455 ACT03-02 PU99-37 PU99-38 TR01-305 PU99-46 TR12-110 PU02-05 TR12-109 FISHLAB:MA-1131 + FISHLAB:MA-1124 + n/a FISHY4:2BF-1 + FISHY4:Cud1+ ACT03-01 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 Glenelg River, Harrow, VIC Ewens Ponds, Mt Gambier, SA Mosquito Creek, Langkoop, VIC Henry Creek, Kingston, SA Tookayerta Creek, Mt Compass, SA Nangkita Creek, Mt Compass, SA Angas River, Strathalbyn, SA Rodwell Creek, Mt Barker, SA Marne River, Black Hill, SA McKenzie River, Zumsteins, VIC Fyans Creek diversion, VIC Mount Cole Creek, Warrak, VIC Nowhere Creek, Elmhurst, VIC Avoca River, Mt Lonarch, VIC Birch Creek, Clunes, VIC Seven Creek, Strathbogie, VIC King River, Moyhu, VIC Scrubby River, Carboor East, VIC Kiewa River, Kergunyah, VIC Coppabella Creek, Coppabella, NSW Stony Creek, Carabost, NSW Catherines Creek, Jerrawa, NSW Shawns Creek, Coonabarabran, NSW McDonald River, Bendemeer, NSW Molong Creek, Uralla, NSW Browns Creek, Killarney, QLD Criss Cross Creek, VIC Taggerty River, Marysville, VIC Goulburn River, Knockwood, VIC Hollands Creek, Dodds Bridge, VIC Hollands Creek, Fords Bridge, VIC Stony Creek, VIC Ovens River, Bright, VIC Cudgewa Creek, Cudgewa, VIC Cotter River, Vanaties Crossing, ACT Locality Glenelg Millicent Millicent Millicent Lower Murray Lower Murray Lower Murray Lower Murray Lower Murray Wimmera Wimmera Wimmera Wimmera Avoca Loddon Goulburn Ovens Ovens Kiewa Upper Murray Murrumbidgee Lachlan Castlereagh Namoi Gwydir Condamine Goulburn Goulburn Goulburn Broken Broken Ovens Ovens Upper Murray Murrumbidgee Drainage 1 2 2 1 1 4 2 3 4 4 1 1 2 2 2 2 2 3 1 1 2 2 2 2 2 2 2 3 2 2 2 2 3 3 4 Cytochrome b 0 1 1 1 1 1 1 0 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 2 2 1 1 1 1 S7 1 2 2 1 1 4 2 4 4 4 1 1 2 2 2 2 2 3 1 1 2 2 2 2 2 2 5 3 4 7 7 0 4 3 4 Allozymes −37.15722 −38.02528 −37.10306 −36.46500 −35.35694 −35.34472 −35.25639 −35.18583 −34.70222 −37.08056 −37.10889 −37.26306 −37.14278 −37.25111 −37.28861 −36.88194 −36.58222 −36.66389 −36.32889 −35.74028 −35.62889 −34.78750 −31.26528 −30.88194 −30.66750 −28.35389 −37.44470 −37.50583 −37.42215 −36.79448 −36.82566 −36.66667 −36.71417 −36.12306 −35.34528 Latitude 141.59722 140.77945 141.03444 139.89278 138.72556 138.66250 138.89722 138.91139 139.49889 142.37028 142.55889 143.13917 143.28889 143.37778 143.81056 145.69056 146.39083 146.55667 147.03028 147.72445 147.71917 149.09389 149.13556 151.15583 151.28139 152.33972 145.47387 145.77250 146.23926 146.13565 146.13882 146.48333 146.93306 147.83250 148.89111 Longitude NGW NGW NGW NGW NMD#2 NMD#2 NMD#1 NMD#1 NMD#1 NGW NGW NGW NGW NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 NMD#1 BG BG BG BE#1 BE#1 BE#1 BE#1 BE#1 BE#2 Cand spp Site refers to the localities shown in Fig. 1. Station codes can be used to track references to genetic material deposited in the South Australian Museum and morphological samples deposited in the Australian, South Australian, and Victorian museum collections. State abbreviations: NSW, New South Wales; QLD, Queensland; SA, South Australia; TAS, Tasmania; VIC, Victoria; sample sizes are shown for each component in the present study: cytochrome b, S7, and allozymes. Latitude and longitude are provided in decimal degrees. Candidate species (cand spp) represents the name given to each taxon identified in the present study. Numbers after a name indicate which populations contained different genetic substructure for allozyme variation. Station code Site Table 1. Continued 526 M. P. HAMMER ET AL. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES Takeyama (1998). Final concentrations for PCR components per 25-μL reaction were: 25 ng of template DNA, 0.25 μM of each primer, 0.625 units of Taq DNA polymerase, 0.1 mM of each dNTP, 2.5 μL of 10 × reaction buffer and 2.5 mM MgCl2. Amplification parameters were: 94 °C for 2 min followed by 35 cycles of 94 °C for 30 s, 48 °C for 30 s, and 72 °C for 90 s, and a final elongation step at 72 °C for 7 min. For the nested PCR, our first reaction was 10 μL with the same PCR conditions listed above. This first PCR reaction was then diluted to 1 : 99 and 1 μL was used in the second 25-μL reaction. All sequences obtained in the present study were deposited in GenBank, accession numbers KF894492–KF894638 and the sequence alignments were deposited in Dryad, doi:10.5061/dryad.jn7b2. ALLOZYME DATA ANALYSIS To escape the influence of a priori expectations based on geography or morphotype, our initial analysis of the allozyme data used individuals as the unit of analysis. The genetic affinities among individuals were explored using stepwise principal coordinates analysis (PCA). Here, we followed the rationale initially employed by Horner & Adams (2007) to define 23 valid Australian species across seven previouslyrecognized ‘morpho-species’ within the skink genus Cryptoblepharus, and subsequently used in other allozyme studies of species boundaries (Hammer et al., 2007; Adams et al., 2011; Unmack et al., 2012). Briefly, discrete PCA clusters of individuals were flagged as putative candidate species when found to be diagnosable from all other PCA clusters by fixed differences (no alleles in common) and/or near-fixed differences (the cumulative frequency of all shared alleles at that locus being ≤ 10%) at multiple allozyme loci (herein a minimum of three for a combined sample size per cluster of N > 20, and a minimum of four for a combined N ≤ 20). All discrete PCA clusters were subsequently subjected to further rounds of PCA to determine whether additional heterogeneity was hidden in deeper dimensions and, if so, whether this heterogeneity was consistent with the presence of additional candidate species (under the criteria described above). Each PCA was implemented on a pairwise matrix of Rogers’ genetic distance among the relevant individuals, sensu Horner & Adams (2007). Neither stepwise PCA, nor within-site statistical comparisons of observed genotype frequencies with those expected under Hardy–Weinberg equilibrium (HWE) indicated any departures from panmixia at individual sites. Consequently, we also conducted a site-based assessment of genetic affinities by generating a Neighbour-joining (NJ) tree from a pairwise matrix of unbiased Nei’s distance among all sites. All 527 protocols used for HWE testing and NJ tree construction are described in Hammer et al. (2007). ANALYSIS OF DNA SEQUENCE DATA Sequences were edited using CHROMAS LITE, version 2.0 (Technelysium) and imported into BIOEDIT, version 7.0.5.2 (Hall, 1999). Sequences coding for amino acids were aligned by eye and checked via amino acid coding in MEGA, version 5.05 (Tamura et al., 2011) to test for unexpected frame shift errors or stop codons. S7 sequences were aligned with the online version of MAFFT, version 7.046 (Katoh & Standley, 2013) using the very slow G-INS-i algorithm with the scoring matrix for nucleotide sequences set to 1PAM/K = 2, a gap opening penalty of 1.53, and an offset value of 0.5. Seven individuals had heterogametic cytb sequences (an unusually high frequency in our experience). All were sequenced again to confirm the result. In all cases, the odd heterogametic base pair was either unique within all Gadopsis, or unique within either nominal species from our dataset. We removed the odd heterogametic base pair from all analyses. For S7, 14 individuals were heterozygous; thus, we manually phased the S7 data. Ten individuals only had a single heterozygous position, and the alleles from the remaining four individuals were easily phased. Phylogenetic analyses were performed with maximum likelihood (ML) using GARLI, version 2.0 (Zwickl, 2006). We identified the best-fitting model of molecular evolution using the Akaike information criterion in MODELTEST, version 3.7 (Posada & Crandall, 1998) using PAUP* 4.0b10 (Swofford, 2002). For the cytb and S7 data sets, MODELTEST identified TVM+I+G and TIM+I as the best models respectively for our ML analyses. To obtain the ML topology, we ran GARLI with ten search replicates with the default setting values changed: streefname = random; attachmentspertaxon = 160 (cytb), 88 (S7); genthreshfortopoterm = 100 000; scorethreshforterm = 0.05; significanttopochange = 0.00001. For bootstrapping, we ran 1000 replicates with the previous settings with the changes: genthreshfortopoterm = 10 000; significanttopochange = 0.01; treerejectionthreshold = 20, as suggested in the GARLI manual to speed up bootstrapping. Trees were rooted using Murray cod Maccullochella peelii (Mitchell), carmelite concepción Percilia irwini Eigenmann, golden perch Macquaria ambigua (Richardson), and southern pygmy perch Nannoperca australis Günther because these are the closest sister groups to Gadopsis (Near et al., 2012). For the combined cytb/S7 analysis, we put the same individuals sequenced for S7 with the corresponding cytb sequence. When an individual had two different S7 alleles, we excluded the second allele. The ML trees were deposited in TreeBASE, © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 528 M. P. HAMMER ET AL. accession number 15174 (http://purl.org/phylo/ treebase/phylows/study/TB2:S15174). We calculated mean within- and among-lineage variation based on cytb using p-distances in MEGA. To estimate molecular divergence times, we used our multigene dataset in *BEAST (Heled & Drummond, 2010) as incorporated into BEAST 2.0.2 (Drummond et al., 2012). Input files were generated using BEAUti 2.0.2. The analysis used an uncorrelated log-normal relaxed molecular clock with rate variation following a tree prior using the calibrated Yule model. All outgroup taxa were removed to comply with the assumption of the Yule model of complete taxon sampling. For our *BEAST analysis, we grouped our sequences into six taxa based on the cytb results, with the exception that we removed anomalous cytb sequences for an odd lineage because they clearly represent a gene tree artefact (see Results). We used separate partitions for the cytb and S7 genes. For the model of sequence evolution, we used the RB BEAST add-on, which automatically adjusts the analysis to choose the best model of sequence evolution for each partition. Divergences based on pairwise comparisons in the present study were assumed to occur at approximately 1.0% per million years. Other studies on teleost fishes have calibrated molecular clocks for protein coding mtDNA genes and obtained values of between 0.68 and 1.66% pairwise divergence per million years (Burridge et al., 2008). Studies on Australian freshwater fishes using the cytb gene have used a rate of 1.0% (Unmack et al., 2013) or have obtained similar rates (0.84%, 0.78%) based on biogeographical calibrations (Unmack et al., 2011, 2012), respectively. Although molecular clock estimates vary and, in most cases, only provide crude estimates of divergence times (Donoghue & Benton, 2007; Pulquerio & Nichols, 2007), they can provide important insights into the approximate timing of divergences. Hence, we interpret our molecular clock findings in the present study with an appropriate level of caution. Multiple shorter runs were conducted to check for stationarity and also that independent runs were converging on a similar result. Final results from the *BEAST analyses were based on the combination of four separate runs for 100 million generations each, with parameters logged every 10 000 generations. Tree and logfile outputs were combined in LOGCOMBINER, version 2.0.2 with a burn-in of 10%. The combined logfile was examined in TRACER, version 1.5 (Rambaut & Drummond, 2007), whereas the age estimates were summarized using TREEANNOTATOR, version 1.7.5 (version 2.0.2 was giving false values; Drummond et al., 2012) with the mean age values placed on the maximum clade credibility tree found in the sample of trees generated from BEAST. RESULTS ALLOZYME ANALYSIS Fifty-one putative loci were interpretable, 12 of which were invariant amongst the 200 individuals screened (Table 2). The initial PCA (Fig. 2A) assigned every individual to one of three well separated and discrete clusters (i.e. no intermediate or hybrid forms were detected) in the first two PCA dimensions. One of these PCA clusters corresponded to G. bispinosus, whereas the other two are referred to as the ‘northern’ and ‘southern’ G. marmoratus clusters. All three clusters were diagnosable from one another at multiple allozyme loci (range 3–16) (Table 2), thus satisfying our criteria for recognition as candidate species. Importantly, each of these three clusters also harboured additional, species-level heterogeneity when subjected to follow-up PCAs (Fig. 2B, C, D). Accordingly G. bispinosus comprised the allopatric candidate species, ‘BG’ from the Goulburn Basin and ‘BE’ from all other eastern basins, diagnosable at three allozyme loci; ‘northern’ G. marmoratus included the two parapatric taxa, ‘NMD’ from the MurrayDarling Basin (excluding Wimmera Basin) and ‘NGW’ from the Millicent Coast, Glenelg and Wimmera basins, also diagnosable at three allozyme loci; and ‘southern’ G. marmoratus was split into the three geographically-proximate taxa ‘SEV’ from costal eastern Victoria and New South Wales (Snowy Basin), ‘SBA’ from Victorian and Tasmanian basins draining to Bass Strait, and ‘SWV’ from western Victoria (east of Glenelg Basin), each diagnosable from the others at three or four loci (Fig. 1, Tables 2, 3). Further rounds of PCA on the individuals representing each of these seven candidate species (not presented) confirmed the presence of more PCA subclusters within three of the candidate species (NMD, SWV, and BE) (Fig. 2). However, none of the three phylogeographical dichotomies detected (sites 44/45 versus the rest in NMD, site 35 versus sites 33/34 in SWV, site 74 versus the rest in BE) (Fig. 1) reached the threshold levels of divergence required to infer the presence of additional species. Nevertheless, the within-species break within SWV (three fixed differences) (Table 2) was close to our threshold level (four loci for a combined N ≤ 20) for the sample sizes involved. Summaries of the allelic profiles of all ten lineages identified by stepwise PCA within Gadopsis (seven candidate species plus three, within-species phylogeographical breaks) are presented in Table 2, whereas pairwise estimates of genetic divergence among candidate species are provided in Table 3. Importantly, all seven candidate species and two of the three within-species phylogeographical breaks are clearly supported by the site-based NJ analysis © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 b b94,a b d b b a b c c a b b b a e a b b55,c43,d b b60,c b b b b b e a c d90,c b98,c b b b c b d a e88,d Acon1 Acon2 Acp Acyc Ada Adh Ap Ck Enol2 Est1 Est2 Fdp1 Fdp2 Fum Got1 Got2 Gp Gpd Gpi1 Gpi2 Gsr Idh Ldh Mdh1 Mdh2 Mdh3 Mpi PepA1 PepA2 PepB PepD1 Pgam 6Pgd Pgk Pgm1 Pgm2 Pk2 Tpi2 Ugpp BG (12) b b a d b a a b96,a c c a b b b a e a b87,a c b b b b b b b e a c d b b c96,d b c b d c92,a d BE#2 (4) a63,b a63,b25,d b d b b a b c50,d c a b b b a e a b c b b a63,b b b b87,a b e a c d87,c b b b b b50,c b d c d c e a b58,c b b a c c c b98,a b b98,a b b d c b c90,b b a b b a b a d a a98,b b58,a b b b b c b b c98,b b SEV (24) c84,b e a c c99,b b a c c99,b c99,a a b b b b d c b c b99,c a b b97,c2,a c b b80,a18,c1,d d a85,b a99,b b52,a b98,c b99,a b99,a b c92,d6,a a65,b b c81,d10,a8,b b96,c SBA (61) c f a c c b b c c c a b b b b b c b c b a b b a96,c b a d a a a b b b b c b b c a SWV#1 (12) ‘Southern’ Gadopsis marmoratus c e a c c b b c c c62,b a b b b b d c b c a a b b a b a50,b d a a a b b b b c b b c a SWV#2 (4) c90,d e c b95,c3,a b c a98,d c c c a95,c b96,a b b b d98,f c b c b a b b c b b c98,b a a b95,c5 b b c b c b b88,a7,c c a NGW (22) c f c b b c a c c c a b b a b d c b a b b70,a b b c b b a a a b b b c a50,b c b b c a NMD#1 (35) ‘Northern’ Gadopsis marmoratus C f92,c6,g C B b94,a C C C c97,a c a b b a b d98,a1,c c98,b b c66,a19,b9,d b a b b c b b97,d a56,d a a b96,c b97,a b90,a c b94,a c b b c a98,b NMD#2 (5) For polymorphic loci, the frequencies of all but the rarer/rarest alleles are expressed as percentages and shown as superscripts (allowing the frequency of each rare allele to be calculated by subtraction from 100%). All individuals were invariant at 12 loci: Ak, Ald, Enol1, Gapd1, Gapd2, Glo, Me, Ndpk1, Ndpk2, PepD2, Pk1, and Tpi1. Sample sizes are shown in parentheses. BE#1 (21) Locus Gadopsis bispinosus Table 2. Allele frequencies at all variable allozyme loci for the ten lineages identified in Gadopsis CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES 529 530 M. P. HAMMER ET AL. A) B) NGW #1 Second Dimension #2 ‘northern’ G. marmoratus ‘southern’ NMD C) SEV G. bispinosus SBA SWV #2 #1 G. marmoratus D) #2 BE #1 BG First Dimension Figure 2. Stepwise principal coordinates analysis (PCA) of the 200 Gadopsis included in the allozyme study based on 51 loci. Relative PCA scores have been plotted for the first and second dimensions for the initial PCA on all individuals (A) and for each of the three primary follow-up PCAs (B, C, D). Individuals are labelled by final taxon name, using the symbols presented in Fig. 1. Envelope legend: thick solid line = subsequently found to include multiple taxa in the follow-up PCA; thin solid line = no evidence of multiple taxa in follow-up PCA; dashed line = phylogeographical lineages within candidate taxa (lineage codes as in Table 1). Percentage variation explained by the first and second dimensions, respectively: A, 36% and 27%; B, 51% and 13%; C, 36% and 25%; D, 56% and 18%; whereas the number of diagnostic differences between candidate species is given in Table 3. Table 3. Pairwise allozyme measures of genetic divergence among seven candidate species in Gadopsis Taxon BE BG SEV SBA SWV NGW NMD BE BG SEV SBA SWV NGW NMD (0.05) 3 17 15 17 17 18 0.09 (0.00) 18 16 18 16 17 0.41 0.43 (0.02) 3 4 8 9 0.40 0.42 0.09 (0.03) 3 7 8 0.45 0.46 0.13 0.12 (0.05) 9 9 0.42 0.37 0.18 0.17 0.24 (0.01) 3 0.47 0.43 0.24 0.23 0.24 0.07 (0.03) Lower left triangle = number of diagnostic differences (out of 51 loci surveyed); upper right triangle = unbiased Nei’s distance. Zero diagonal = mean between-site Nei’s distances for each candidate species. (Fig. 3). This analysis also infers the presence of another east–west phylogeographical break in one candidate species (SEV; sites 1–3 versus sites 4, 6, 21) (Fig. 1) and obvious population structure in another (SBA; several distinctive sites in the eastern end of its range) (Fig. 1). However, our small sample sizes preclude any quantitative assessment of this withinspecies heterogeneity. SEQUENCE ANALYSIS The cytb dataset consisted of 1165 bp for 151 individual fish (Table 1). Removal of fish with identical haplotypes within populations reduced the dataset to 75 individuals, plus four outgroup taxa. Four individuals from two populations (populations 6 and 21) had a 2-bp deletion in the fifth last codon of cytb, which resulted in a premature stop codon. We confirmed this result by amplifying the approximately 3700-bp region between the Leu and Pro tRNA genes and sequencing from a nested PCR reaction using our cytb primers to reduce the chance of this being the result of a nuclear pseudogene copy. The individual number of each fish from a population with a given haplotype is provided after the population name in Figure 4. For the cytb data, excluding outgroups, 885 © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES 1 2 3 SEV 4 6 21 35 SWV ‘southern’ 20 10 12 16 SBA Nei D = 0.05 11 19 18 26 22 31 32 24 15 25 28 27 29 NGW ‘northern’ SWV#1 17 7 G. marmoratus SWV#2 33 34 50 40 37 39 49 51 52 38 41 42 43 NMD 9 14 8 23 61 48 53 58 65 59 62 64 63 57 56 55 54 NMD#1 60 46 47 45 G. bispinosus BG BE 66 67 68 74 73 70 69 72 44 NMD#2 BE#2 BE#1 Figure 3. Neighbour-joining tree depicting the genetic affinities, based on unbiased Nei’s distance, among the 69 Gadopsis sites surveyed in the allozyme study. The major lineage branches are labelled according to both currentlydiagnosed and candidate species. Brackets indicate the phylogeographical breaks for principal coordinates analysis-defined groups within three of the candidate species (in accordance with the terminology of Table 1). of the 1165 bp were constant, 26 variable characters were parsimony uninformative, and 254 characters were parsimony informative. ML recovered one tree with a –ln score of −6650.037403 (Fig. 4). The S7 dataset consisted of 1547 aligned base pairs for 72 individual fish (Table 1). Phasing, followed by removal of fish with identical haplotypes within popu- 531 lations, reduced the dataset to 40 ingroup haplotypes plus four outgroup taxa. The individual number of each fish from a population with a given haplotype is provided after the population name in Figure 5. S7 alleles from heterozygous individuals were designated as A and B. The S7 nuclear data, excluding outgroups, consisted of 1547 aligned base pairs. Of those, 1467 were constant, 25 variable characters were parsimony uninformative, and 55 characters were parsimony informative. ML recovered one tree with a –ln score of −5566.193389 (Fig. 5). We explored concatenation of cytb and S7 data via ML for 72 individual Gadopsis plus four outgroup taxa. The combined dataset consisted of 2712 aligned base pairs. Of those, 1623 were constant, 427 variable characters were parsimony uninformative, and 622 characters were parsimony informative (excluding outgroups). ML recovered one tree with a –ln score of −12418.424149 (Fig. 6). The largest divergences (mean p-distances > 8.0%) (Table 4) within cytb and S7 corresponded to three deeper lineages, which represent ‘northern’ and ‘southern’ G. marmoratus and G. bispinosus (Figs 4, 5). The key difference in the deeper relationships between the tree topologies was that the combined and S7 DNA datasets supported the allozyme dataset in designating G. bispinosus as the first branching lineage (Fig. 3), whereas, for cytb, G. bispinosus was sister to ‘northern’ G. marmoratus. Bootstrap support for deeper relationships were strongest in the combined dataset, moderate for S7, and weak at one node in cytb (Figs 4, 5, 6). All three deeper lineages contained large divergences for cytb (3.1–7.1% mean p-distance) (Table 4), as well as moderate to strong bootstrap support (Fig. 4) that corresponded to SEV, SBA, NMD, NGW, BG, and BE. This corresponds to most of the candidate species identified by the allozyme results, with two exceptions. The first was that we found little independent support for taxon SWV because it was only shallowly nested within SBA (mean p-distance 1.2%) (Fig. 4, Table 4). The second was the presence of a distinctive lineage, herein called ‘SEVodd’ from sites 6 and 21 (Figs 1, 4). This lineage was sister to SEV and SBA and was deeply divergent (mean p-distance 5.7–6.4%) (Table 4). No support for this lineage was present in any of the nuclear datasets. The S7 dataset only provided limited resolution between candidate species within the deeper lineages, although this gene usually provides little resolution at equivalent nodes with lower divergences at cytb (Lavoue, Sullivan & Hopkins, 2003). For cytb, all mean pairwise and within group p-distances are presented in Table 4. Molecular clock estimates (in millions of years) obtained from the *BEAST analysis based on cytb and S7 sequences are presented based on their mean © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 532 M. P. HAMMER ET AL. Figure 4. Maximum likelihood tree for 72 Gadopsis sites based on analysis of mitochondrial cytochrome b and Thr tRNA haplotypes (1165 bp). The major lineage branches are labelled according to candidate species, and sites are bracketed according to membership of the principal coordinates analysis-defined phylogeographical breaks evident within three candidate species (in accordance with the terminology of Table 1). Each operational taxonomic unit (OTU) code is based on the population number and locality (for common haplotypes, the locality name is not included because of space limitations) described in Table 1 and Fig. 1. The geographical distribution of clades is shown in Fig. 1. The value at the end of the OTU code indicates the specific individuals with the same haplotype from each population. Bootstrap values shown are based on 1000 replicates. ▶ Table 4. Mean between-lineage cytochrome b p-distances as percentages (lower left) for Gadopsis BE BG SEVodd SEV SBA SWV NGW NMD BE BG SEVodd SEV SBA SWV NGW NMD (0.15) 4.6 9.5 9.7 9.4 9.7 8.1 8.4 (0.23) 9.8 9.9 9.6 9.9 8.0 8.9 (0.26) 5.7 5.8 6.4 11.6 11.9 (0.58) 3.1 3.7 11.4 10.9 (0.55) 1.2 11.4 11.0 (0.54) 11.5 11.3 (0.35) 7.1 (0.45) Mean within-lineage p-distance is shown on the diagonal. values and 95% highest posterior density (HPD) with the cytb tree (Fig. 4) because they had the same topology; all estimates had effective sample sizes > 850. Mean divergences between candidate sister species varied from 3.7–4.7 Mya, with the earliest branching lineage estimated to have diverged 12.0 Mya (Fig. 4). DISCUSSION Range-wide examination of freshwater blackfishes using multigene datasets demonstrates the presence of six or seven candidate species within Gadopsis, with cryptic biodiversity evident in both G. marmoratus and G. bispinosus. Here, we discuss aspects of their taxonomy, biogeography and conservation implications. BETWEEN CANDIDATE SPECIES PATTERNS It is clear that the current taxonomy of Gadopsis misrepresents actual species richness. Even past suggestions of divergent taxa within G. marmoratus such as ‘northern’ and ‘southern’ forms, plus western Victoria (i.e. NGW), underestimate diversity, partly as a result of limited sampling across the range of the genus (Sanger, 1986; Ovenden et al., 1988; Miller et al., 2004; Ryan et al., 2004). Based on the number of fixed allozyme differences, plus congruence for mtDNA and nuclear DNA markers, we propose that Gadopsis consists of at least six and possibly seven species. The Goulburn Basin populations of G. bispinosus (taxon BG) are substantially differentiated from populations originating from the rest of its range (taxon BE; three fixed differences, mean cytb p-distance of 4.6%; Tables 3, 4; mean age 4.7 Mya, 95% HPD = 1.6–7.7 Mya). The distinctiveness of BG is somewhat unexpected based on a lack of similarity to biogeographical patterns in other fishes (i.e. there are few narrow range endemic fishes in the MurrayDarling Basin). However, another upland habitat specialist is endemic to the Goulburn Basin, the barred galaxias Galaxias fuscus Mack (Raadik, 2011), and populations of N. australis are also divergent at a within taxa level (Unmack et al., 2013). The Goulburn Basin represents the western most populations of G. bispinosus, and potential long-term isolation appears to have been aided by the large reach of unsuitable habitat that would need to be crossed between river basins (Fig. 1). By contrast, the northernmost populations occurring in the upper Murrumbidgee Basin are also highly isolated, and showed moderate differences for allozymes (Figs 2, 3), although only slight DNA sequence divergence (Figs 4, 5). As suggested by Waters et al. (1994), the Murrumbidgee populations may have been established via headwater exchanges rather than via lowland connections. We can now clarify the distribution and distinctiveness of the informally recognized ‘northern’ G. marmoratus (taxon NMD). It is found throughout the Murray-Darling Basin with the exclusion of the currently endoreic Wimmera Basin. This distribution © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES ‘southern’ G. marmoratus 78 95 B 100 A node mean upper lower A 12.0 16.7 7.6 B 3.7 6.5 1.3 C 9.5 12.8 6.1 D 4.4 7.1 1.5 E 4.7 7.7 1.6 ‘northern’ G. marmoratus 69 D 56 C E 81 G. bispinosus 0.02 100 6 LaTrobe (1,2) SEVodd 21 Turtons (1,2) 79 1 Back (1,2) 99 2 Delegate (1,2) 3 Brodbribb (1) 4 Haunted (1) SEV 4 Haunted (2) 97 5 Thompson (1,2) 12 Wye (1,2) 9 Deep (1,2) 20 Relapse (1,2) 7 Greig (1,2) 31 Loves (1,2) 32 Gellibrand (2) 61 8 Tin Mine (1) 66 18 Leven (1) 91 10 Blackfish (1,2) 17 Minnow (1) 16 (1) 26 (1,2) 11 (1,2) 66 27 Lerderderg (1,2) 23 Tarago (1) SBA 22 Minnieburn (1,2) 23 Tarago (1) 24 Diamond (2) 56 24 Diamond (1) 28 Barwon (1,2) 66 29 Kuruc A Ruc (1,2) 52 25 Donnellys (1,2) 30 Ford (1) 99 31 Loves (2) 13 N George (1) 14 Ansons (1) 18 Leven (2) 70 15 Boobyalla (1) 19 Black (1,2) 99 34 Mt Emu (1,2) 33 Brucknells (1,2,3) SWV 99 35 Darlots (1,2) 86 35 Darlots (3) 39 Wannon (1,2,3,4) 51 Mount Cole (1) 52 Nowhere (1,2) 37 Stokes (1,2) 52 38 Muddy (2) 42 Mosquito (1) 36 Bridgewater (1,2) 91 41 Ewens (1,2) NGW 38 Muddy (1) 64 40 U Glenelg (1) 43 Henry (1) 99 42 Mosquito (2) 49 McKenzie (4) 50 Fyans (1) 49 McKenzie (1,3) 49 McKenzie (2) 62 44 Tookayerta (1) 45 Nangkita (2,3,4) 45 Nangkita (1) 61 Catherines (1,2) 83 57 Scrubby (1,3) 56 King (1) 55 Seven (1,2) 55 56 King (2) 57 Scrubby (2) 58 Kiewa (1) NMD 48 Marne (1,2,3,4) 58 47 Rodwell (1,2,3) 59 (1) 62 (1,2) 63 (1,2) 64 (1,2) 65 (1,2) 94 46 Angas (1,2) 86 53 Avoca (1,2) 86 54 Birch (1,2) 60 Stony (1,2) 84 66 Criss Cross (1) 66 Criss Cross (2) 52 67 Taggerty (1) 67 Taggerty (2) BG 99 67 Taggerty (3) 68 Goulburn 2 68 Goulburn (1) 69 Hollands (1,2) 70 Hollands (1,2) 74 Cotter (1) 100 74 Cotter (2,3) 71 Stony (2) BE 73 Cudgewa (1,2,3) 72 Ovens (1) 72 Ovens (2) 74 Cotter (4) © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 533 534 M. P. HAMMER ET AL. 100 2 Delegate (1,2) 3 Brodbribb (1) 5 Thompson (1) SEV 6 LaTrobe (1) 21 Turtons (1) 61 1 Back (1) 5 Thompson (1A) ‘southern’ 8 (1) 9 (1) 15 (1) 17 (1A) 22 (1) 23 (1) 24 (1) 25 (1) 27 (1A) 19 Black (1A) 20 Relapse (1) 11 Styx (2B) 17 Minnow (1) 19 Black (1B) 28 (1) 33 (1) 34 (1,2) 35 (2) SWV 91 10 Blackfish (1B) 11 Styx (2A) 12 (1A) 16 (1) SBA G. marmoratus 93 12 Wye (1B) 31 Loves (1A) 51 31 Loves (1B) 30 Ford (1) 7 Greig (1B) 63 18 Leven (1) 7 (1A) 27 (1B) 29 (1) 38 Muddy (1) 92 41 Ewens (1) NGW 42 Mosquito (2) ‘northern’ 36 (1) 37 (1A) 39 (1,2) 43 (1) 49 (1A) 51 (1) 52 (2) 96 49 McKenzie (1B) 50 Fyans (1) 37 Stokes (1B) 100 52 46 Angas (1) 69 62 Shawns (1) 44 Tookayerta (1) 45 Nangkita (1) 52 55 Seven (1) 54 Birch (1) 56 King (1) NMD 63 McDonald (1) 64 Molong (1) 51 65 Browns (1) 59 Coppabella (1) 57 Scrubby (1) 58 (1) 60 (1) 61 (1) 67 Taggerty (1) 68 Goulburn (2B) BG 80 66 Criss Cross (1,2) 67 (1B) 68 (1B,2A) 69 Hollands (1,2) 70 Hollands (1,2) G. bispinosus 73 Cudgewa (2) 56 71 Stony (1A) BE 71 Stony (1B) 86 72 Ovens (2) 87 0.002 74 Cotter (1) 81 Figure 5. Maximum likelihood tree for Gadopsis based on analysis of 40 haplotypes from the nuclear S7 gene (1547 aligned bp). The major lineage branches are labelled according to candidate species, and sites are bracketed according to membership of the principal coordinates analysis-defined phylogeographical breaks evident within three candidate species (in accordance with the terminology of Table 1). Each operational taxonomic unit (OTU) code is based on the population number and locality (for common haplotypes, the locality name is not included because of space limitations) described in Table 1 and Fig. 1. The value at the end of the OTU code indicates the specific individuals with the same haplotype from each population. Alleles from heterozygous individuals were designated as A and B. Bootstrap values shown are based on 1000 replicates. pattern demonstrates (at least for this species, which penetrates well upland) that there have been no headwater exchanges with surrounding coastal basins for considerable time (as is also the case for taxa BG and BE). The sister lineage to NMD is NGW, which is limited to south-eastern South Australia, and the Glenelg and Wimmera basins (Miller et al., 2004) (Fig. 1). These two ‘northern’ candidate species are diagnosable by three fixed allozyme differences, a mean cytb p-distance of 7.1% (Tables 3, 4), with a mean age of 4.4 Mya (95% HPD = 1.5–7.1 Mya). Observations from previous studies of lower spine counts for populations from the Glenelg Basin provide some preliminary morphological support for the diagnosis of candidate species NGW (Parrish, 1966; Sanger, 1984; Sanger, 1986), with the population from upper Wannon River having unusually low counts (Ryan et al., 2004). The Wimmera Basin appears to be a longisolated, former tributary to the Murray River, with previous connections perhaps severed as Lake Bungunnia drained approximately 700 000 years ago (Stephenson, 1986; McLaren et al., 2009). Indeed, McLaren et al. (2011) suggested the ancestral Murray River may have drained south to the ocean via the Glenelg Basin up until approximately 2.4 Mya prior to the damming and subsequent formation of Lake Bungunnia. This age estimate is within our 95% HPD between NGW and NMD (1.5–7.1 Mya), although clearly on the lower end of our range of estimates. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES 98 90 96 100 96 0.005 99 78 97 SEV SBA NGW 98 NMD 96 BG 100 BE Figure 6. Maximum likelihood tree for 72 combined S7 and mitochondrial cytochrome b and Thr tRNA Gadopsis sequences (2712 bp). The major lineages are labelled according to candidate species. The geographical distribution of clades is shown in Fig. 1. Bootstrap values shown are based on 1000 replicates. The Wimmera Basin has a very limited native fish fauna, probably consisting of only five species. Of those, three species (Gadopsis, N. australis, and a cryptic species in mountain galaxias Galaxias olidus Günther) have essentially identical, or very similar, genetic signatures with the same lineage in the Glenelg Basin rather than other adjacent basins in the Murray-Darling Basin (Raadik, 2011; Unmack et al., 2013). The separation between the candidate species NGW and SWV occurs just west of Portland, which is congruent with the deepest phylogenetic separation within Yarra pygmy perch Nannoperca obscura (Klunzinger) (Hammer et al., 2010). Similarly, both N. australis and N. obscura have deeper withinspecies divergences between Henry Creek and the Murray River mouth (Hammer et al., 2010; Unmack et al., 2013), which is spatially congruent with separation between NGW and NMD (Fig. 1). The other two (potentially three) candidate species are found in remaining coastal river basins (Fig. 1). Within this region, the divergence between SEV and SBA consists of three fixed differences, a mean cytb p-distance of 3.1% (Tables 3, 4), and a mean age of 3.7 Mya (95% HPD = 1.3–6.5 Mya). This divergence was broadly congruent with the boundary between Eastern and Bass provinces (Unmack, 2001, 2013) and is similar to the separation of cryptic species within N. australis (Unmack et al., 2011, 2013), Australian smelt Retropinna semoni (Weber) (Hammer et al., 2007), and Galaxias olidus (Raadik, 2011). This separation is likely a result of the existence of a drainage divide that forms during low sea levels (Unmack, 2001). Turtons Creek (site 20) represents a geographical outlier to this pattern because the Tarwin Basin is west of this boundary between the lineages (Fig. 1). Tarwin Basin is adjacent to the La 535 Trobe Basin (site 6) and the presence of SEV Gadopsis there is probably a result of natural movement of fish across the drainage divide between the La Trobe and Tarwin basins, based on evidence of similar faunal relationships in N. australis (Unmack et al., 2013). Interestingly, the La Trobe and Tarwin basins both shared the ‘SEVodd’ cytb lineage, which was sister to all other ‘southern’ G. marmoratus samples (Fig. 4), whereas nuclear data placed both populations in SEV (Figs 3, 5). Apparently, this unusual mitochondrial lineage managed to persist in part of the range of SEV (perhaps isolated in the upper Tarwin Basin above Turtons Falls). Further sampling is needed to determine whether SEV Gadopsis occur throughout the Tarwin Basin. The last candidate species, SWV, is the only one lacking congruent support across all three datasets. We detected three fixed differences to SBA (Table 3); however, p-distances for cytb were shallow (1.2%) (Table 4) and they did not form a reciprocally monophyletic lineage with SBA (Fig. 4). The S7 gene had less resolution between candidate species, although SWV populations were all identical to haplotypes from two SBA populations (Fig. 5). There are a range of possible explanations for this lack of congruence. The mtDNA results could reflect sporadic introgression between candidate species and the S7 results could be the result of a lack of lineage sorting and/or ancestral polymorphism. Alternately, our three fixed allozyme differences might not remain truly diagnostic if we had been able to sample more intensively in southwestern Victoria. Only further characterization of additional individuals and populations for a broader range of genetic markers and morphological characters will resolve the status of SWV. Clarification of morphological differences between candidate species is required to determine whether they can be diagnosed to provide additional lines of evidence to support their recognition as valid species, and to provide non-molecular characters for distinguishing them. Molecular data expedite the search for cryptic species and, after rigorous study, many of these prove to be morphologically distinctive, albeit sometimes in subtle ways (Beheregaray & Caccone, 2007). A major advantage of complementary taxonomic approaches using molecular and morphological techniques is that genetic evaluations can generate testable hypotheses for methods based on traditional morphology (i.e. merisitcs, morphometrics, osteology) and geometric morphometric approaches (Berendzen et al., 2009; McDowall, 2010). WITHIN CANDIDATE SPECIES PATTERNS Based on aspects of their biology (e.g. small home range, habitat specificity, low fecundity, demersal © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 536 M. P. HAMMER ET AL. larvae), we predicted that Gadopsis species should display phylogeographical substructuring. In addition, G. bispinosus (taxa BG, BE) and G. marmoratus (NMD) in the Murray-Darling Basin have more fragmented upstream distributions that would promote a higher level of substructure. Coastal basins should also display a high degree of differentiation (Unmack et al., 2009), being mostly isolated from one another except at times of low sea levels. Unmack et al. (2013) provided hypotheses for how genetic variation may be structured across south-eastern Australia based on continental shelf width. Essentially, drainages in central Victoria and northern Tasmania were more recently connected hydrologically via Bass Lake, which formed whenever sea levels were substantially lowered (Unmack, 2001, 2013). By contrast, potential connectivity between basins should be lower in regions to the east and west because drainages there are less inter-connected during low sea levels (Unmack et al., 2013). It should be noted that, although our assessment of within candidate species patterns is constrained by our small sample sizes, our results are still indicative of the broader patterns. Southern coastal drainages had the highest mean within-lineage p-distances for cytb (0.54–0.58% for SWV, SBA, and SEV) (Table 4), with intermediate values for Nei’s D for allozymes (0.02–0.05) (Table 3). In each case, populations within SBA were just as differentiated as populations within the regions to the east (SEV) and west (SWV), despite the potential connectivity provided by Lake Bass. Clearly, recent sea level changes have not provided gene flow between Victorian and Tasmanian drainages, perhaps because Lake Bass was not sufficiently fresh or the shallow lake represented unsuitable habitat for Gadopsis. However, even sites that would have had low sea level riverine connections (i.e. not via Lake Bass) contained different haplotypes (e.g. sites 4–6, 7–10, 22–24, 25–28) (Figs 1, 4) (Unmack et al., 2013), as well as some sites within the same river basin (e.g. sites 2–3, 5–6, 25–26) (Figs 1, 4). Geographical structuring within the MurrayDarling Basin was moderate, with a mean within group p-distance of 0.45% for cytb (Table 4), and an intermediate value of 0.03 for Nei’s D for allozymes (Table 3). Although genetic divergences overall conformed to expectations, the relative pattern was a result of genetic variation between southern sites because all northern sites contained the same single cytb haplotype that was also shared with a single site (59) in the upper Murray. This may be indicative of more recent range expansion (e.g. Late Pleistocene) upstream within the Murray-Darling Basin rather than relictual persistence. Out of all the cytb comparisons, G. bispinosus (BG, BE) had the lowest levels of variation, with mean within-lineage p-distances for cytb of 0.23% and 0.15%, respectively) (Table 4). However, in contrasting taxa, BE conformed to our prediction of strong substructure and had the equal highest value of Nei’s D calculated from allozymes (0.05) relative to other candidate species, whereas the range-restricted BG had very low variation (0.00) (Table 3). Our data highlight several populations as genetically divergent (more so for the nuclear markers) (Figs 2, 3), namely the Murrumbidgee Basin (site 74; allozyme lineage BE#2) as a disjunct subpopulation of BE; the two separate drainages, Mt Emu Creek (sites 33, 34; allozyme lineage SWV#1) and Darlots Creek (site 35; allozyme lineage SWV#2), within SWV; and Tookayerta catchment (site 44, 45; allozyme lineage NMD#2) in the Mount Lofty Ranges, which was the most distinctive drainage across the expansive distribution of NMD (and displayed numerous unique alleles) (Table 2). This latter result is surprising, given this is a small system (approximately 100 km2) that abuts other stream catchments with which it shares freshwater connections, albeit separated by a section of dense swamp in its lower reaches. Divergent populations represent obvious broad withinspecies units for conservation and management (sensu Moritz, 1994). CONSERVATION Freshwater fishes are a group under significant threat globally (Bruton, 1995; Collares-Pereira & Cowx, 2004). The strong focus of human activity on their specific habitat, coupled with changes to catchments, through, for example, water use, development, and species introductions, together place a heavy burden on naturally-restricted aquatic ecosystems and species. The biological attributes of Gadopsis species put them among the types of freshwater fishes most vulnerable to extinction as a result of humaninduced environmental change (Angermeier, 1995), with declines being realized across the range. The discovery of multiple cryptic species of Gadopsis, and generally strong within-species genetic structure indicative of low dispersal ability, implies that declines will have significant consequences for biodiversity conservation (Piggott, Chao & Beheregaray, 2011). In the present study, the broad ranges of G. marmoratus and G. bispinosus have been shown to encompass a number of narrow range cryptic species (BG, NGW, SWV), all with smaller ranges and with ongoing threats increasing the likelihood of extinction. Furthermore, extirpations of divergent isolated populations will not be reversed (e.g. Murrumbidge Basin BE#2, Tookayerta Creek MBD#2). Despite numerous known and reported translocations including © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES southern Tasmania from northern Tasmania (Ogilby, 1913; Jackson & Llewellyn, 1980), Snowy River (Craigie River) from Victoria in 1883 (Ogilby, 1913; Whitley, 1962), and southern Victorian fish into northern Victoria (Cadwallader & Backhouse, 1983), we did not detect any geographically mismatched populations within their native range (other than Turtons Creek, site 21, as discussed earlier). Nevertheless, stocking or translocation are a potential threat through competition, hybridization, and gene swamping (Utter, 2004). The likely inability of most blackfishes to move between disjunct habitats with unsuitable intervening connections has implications for their longterm survivorship. The lower and upper Murray, once considered to be connected by a continuous G. marmoratus population in the Murray River lowland channel post-European settlement, are now fragmented as a result of an altered, static, and turbid riverine environment (Hammer et al., 2009). Relictual populations in mid-elevation stream catchments have contracted significantly, often to refuge habitats, and face ongoing threats to their viability from land-use change, water abstraction, and introduced species, including salmonids. Documented population declines or extirpation have occurred for NMD in the Mount Lofty Ranges, for NGW in the Wimmera Basin, Mosquito Creek, and Henry Creek (site 43), and for SEV in the Snowy River (FSC, 2008; Hammer et al., 2009). Upland populations, primarily candidate species BG, BE, and SEV, face additional threats of wildfire- or forestry-related sediment slugs and reservoir expansion (Lyon & O’Connor, 2008; Lintermans, 2012). Record drought has provided cumulative impact, with urgent conservation measures including fish rescue and in situ watering undertaken in response to habitat drying for several populations in South Australia (Hammer et al., 2009, 2013b). Broadly, climate change appears set to exacerbate hydrological change from significant abstraction (e.g. increases in water temperature, reduced freshwater flows), especially for small remnant populations (Morrongiello et al., 2011). CONCLUSIONS Our hypothesis that the Australian freshwater fish fauna is characterized by high levels of cryptic biodiversity is further supported by a comprehensive examination of a prominent group, the freshwater blackfishes. Using defined operational criteria, incorporating range-wide and intensive taxon sampling, and considering multiple nuclear and matrilineal markers, we provide genetic evidence for tripling the number of species in this iconic genus. This occurs in a well studied (and for the most part intensively 537 developed and managed) part of Australia, and hence there are major implications for aquatic conservation and management (Koehn & O’Connor, 1990; Lintermans, 2007). Ongoing molecular assessments and concurrent morphological descriptions for all groups of freshwater fishes remain a priority for fully documenting the regional fauna. There is some urgency to this task, given the significant current threats to fishes and aquatic habitats, and a high potential for future development and climate change, leading to the distinct possibility that cryptic species could become extinct before they are even described. This has already happened both in Australia within Macquarie perch Macquaria australasica Cuvier (Faulks, Gilligan & Beheregaray, 2010; Hammer et al., 2013a) and elsewhere such as the eastern USA within southern cavefish Typhlichthys subterraneus Girard (Niemiller, Near & Fitzpatrick, 2012). ACKNOWLEDGEMENTS We thank the many people who helped with field work and other aspects relative to obtaining specimens, especially M. Baltzly, L. Farrington, C. Kemp, G. Knowles, Maree Hammer, M. Lintermans, J. Lyon, R. Remington, S. Ryan, and S. Westergaard. J. Jackson and M. Lintermans supplied collection records and local advice for their jurisdictions. We also thank K. Walker for his support of the present study and M. Lintermans, D. Gilligan and two anonymous reviewers for their helpful comments. All collections were obtained under permit from various state fisheries agencies and research conducted under approved institutional animal ethics guidelines. Work was supported in part by grants including an Australian Postgraduate Award from the University of Adelaide and the Cooperative Research Centre for Freshwater Ecology to MH and to PJU via the Murray-Darling Basin futures research supported through the Australian Government’s Collaborative Research Networks (CRN) Program. REFERENCES Adams M, Page TJ, Hurwood DA, Hughes JM. 2013. A molecular assessment of species boundaries and phylogenetic affinities in Mogurnda (Eleotridae): a case study of cryptic biodiversity in the Australian freshwater fishes. Marine and Freshwater Research 64: 920–931. Adams M, Wedderburn SD, Unmack PJ, Hammer MP, Johnson JB. 2011. Use of congeneric assessment to understand the linked genetic histories of two threatened fishes in the Murray-Darling Basin, Australia. Conservation Biology 25: 767–767. Allen GR, Midgley SH, Allen M. 2003. Field guide to the freshwater fishes of Australia. Perth: Western Australian Museum. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 538 M. P. HAMMER ET AL. Angermeier PL. 1995. Ecological attributes of extinctionprone species: loss of freshwater fishes of Virginia. Conservation Biology 9: 143–158. April J, Mayden RL, Hanner RH, Bernatchez L. 2011. Genetic calibration of species diversity among North America’s freshwater fishes. Proceedings of the National Academy of Sciences of the United States of America 108: 10602– 10607. Beheregaray LB, Caccone A. 2007. Cryptic biodiversity in a changing world. Journal of Biology 6: 9. Berendzen PB, Olson WM, Barron SM. 2009. The utility of molecular hypotheses for uncovering morphological diversity in the Notropis rubellus species complex (Cypriniformes: Cyprinidae). Copeia 2009: 661–673. Bickford D, Lohman DJ, Sodhi NS, Ng PKL, Meier R, Winker K, Ingram KK, Das I. 2007. Cryptic species as a window on diversity and conservation. Trends in Ecology and Evolution 22: 148–155. Bond NR, Lake PS. 2003. Characterizing fish-habitat associations in streams as the first step in ecological restoration. Austral Ecology 28: 611–621. Bruton MN. 1995. Have fishes had their chips? The dilemma of threatened fishes. Environmental Biology of Fishes 43: 1–27. Burridge CP, Craw D, Fletcher D, Waters JM. 2008. Geological dates and molecular rates: fish DNA sheds light on time dependency. Molecular Biology and Evolution 25: 624–633. Cadwallader PL, Backhouse GN. 1983. A guide to the freshwater fish of Victoria. Melbourne: Atkinson Government Printer. de Castelnau F. 1872. Contribution to the ichthyology of Australia. I. The Melbourne fish market. Proceedings of the Zoological and Acclimatisation Society of Victoria 1: 29–242. Chapman AD. 2009. Numbers of living species in Australia and the world, 2nd edn. Canberra: Australian Biological Resources Study (ABRS). Chow S, Takeyama H. 1998. Intron length variation observed in the creatine kinase and ribosomal protein genes of the swordfish Xiphias gladius. Fisheries Science 64: 397– 402. Collares-Pereira M, Cowx I. 2004. The role of catchment scale environmental management in freshwater fish conservation. Fisheries Management and Ecology 11: 303–312. Donoghue PC, Benton MJ. 2007. Rocks and clocks: calibrating the tree of life using fossils and molecules. Trends in Ecology & Evolution 22: 424–431. Drummond AJ, Suchard MA, Xie D, Rambaut A. 2012. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution 29: 1969–1973. DSE. 2013. Advisory list of threatened vertebrate fauna in Victoria 2013. Melbourne: Department of Sustainability and Environment, 18. Faulks LF, Gilligan DM, Beheregaray LB. 2010. Evolution and maintenance of divergent lineages in an endangered freshwater fish, Macquaria australasica. Conservation Genetics 11: 921–934. FSC. 2008. Fisheries Scientific Committee: final determination Gadopsis marmoratus – river blackfish population in the Snowy River catchment. Nelson Bay: New South Wales Fisheries, 4. Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/ NT. Nucleic Acids Symposium Series 41: 95–98. Hammer M, Wedderburn S, van Weenan J. 2009. Action plan for South Australian freshwater fishes. Adelaide: Native Fish Australia (SA) Inc, 206. Hammer MP, Adams M, Hughes JM. 2013a. Evolutionary processes and biodiversity. In: Humphries P, Walker KF, eds. Ecology of Australian freshwater fishes. Collingwood: CSIRO Publishing, 49–79. Hammer MP, Adams M, Unmack PJ, Walker KF. 2007. A rethink on Retropinna: conservation implications of new taxa and significant genetic sub-structure in Australian smelts (Pisces: Retropinnidae). Marine and Freshwater Research 58: 327–341. Hammer MP, Bice CM, Hall A, Frears A, Watt A, Whiterod NS, Beheregaray LB, Harris JO, Zampatti BP. 2013b. Freshwater fish conservation in the face of critical water shortages in the southern Murray-Darling Basin, Australia. Marine and Freshwater Research 64: 807– 821. Hammer MP, Unmack PJ, Adams M, Johnson JB, Walker KF. 2010. Phylogeographic structure in the threatened Yarra pygmy perch Nannoperca obscura (Teleostei: Percichthyidae) has major implications for declining populations. Conservation Genetics 11: 213–223. Hammer MP, Unmack PJ, Adams M, Raadik TA, Johnson JB. 2014. Data from: A multigene molecular assessment of cryptic biodiversity in the iconic freshwater blackfishes (Teleostei: Percichthyidae: Gadopsis) of southeastern Australia. Dryad Digital Repository. doi: 10.5061/ dryad.jn7b2. Harris JH, Silveira R. 1999. Large-scale assessments of river health using an Index of Biotic Integrity with lowdiversity fish communities. Freshwater Biology 41: 235–252. Heled J, Drummond AJ. 2010. Bayesian inference of species trees from multilocus data. Molecular Biology and Evolution 27: 570–580. Hoese DF, Bray DJ, Paxton JR, Allen GR, eds. 2007. Zoological catalogue of Australia, Volume 35, parts 1–3. Fishes. Collingwood: ABRS & CSIRO Publishing. Horner P, Adams M. 2007. A molecular-systematic assessment of species boundaries in Australian Cryptoblepharus (Reptilia: Squamata: Scincidae) – a case study for the combined use of allozymes and morphology to explore cryptic biodiversity. The Beagle, Records of the Museums and Art Galleries of the Northern Territory Supplement 3: 1–19. Jackson PD. 1978. Spawning and early development of the river blackfish, Gadopsis marmoratus Richardson (Gadopsiformes: Gadopsidae), in the McKenzie River, Victoria. Australian Journal of Marine and Freshwater Research 29: 293–298. Jackson PD, Koehn JD, Lintermans M, Sanger AC. 1996. Family Gadopsidae: freshwater blackfishes. In: McDowall © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 CRYPTIC BIODIVERSITY IN FRESHWATER BLACKFISHES RM, ed. Freshwater fishes of south-eastern Australia, 2nd edn. Chatswood: Reed Books, 186–190. Jackson PD, Llewellyn LC. 1980. Family Gadopsidae: river blackfish. In: McDowall RM, ed. Freshwater fishes of south-eastern Australia, 2nd edn. Sydney: AH & AW Reed, 160–161. Jerry DR, Eliphinstone MS, Baverstock PR. 2001. Phylogenetic relationships of Australian members of the family Percichthyidae inferred from mitochondrial 12S rRNA sequence data. Molecular Phylogenetics and Evolution 18: 335–347. Johnson GD. 1984. Percoidei: development and relationships. In: Moser GH, ed. Ontogeny and systematics of fishes. Gainesville, FL: American Society of Ichthyologists and Herpetologists, Special Publication, Volume 1. 464–498. Katoh K, Standley DM. 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Molecular Biology and Evolution 30: 772–780. Khan MT, Khan TA, Wilson ME. 2004. Habitat use and movement of river blackfish (Gadopsis marmoratus R.) in a highly modified Victorian stream, Australia. Ecology of Freshwater Fish 13: 285–293. Khan MT, Wilson ME, Khan TA. 2003. Is spawning by river blackfish hollow-limited in an agricultural catchment? Victorian Naturalist 120: 179–190. Koehn JD. 1987. Artificial habitat increases abundance of two-spined blackfish (Gadopsis bispinosus) in Ovens River, Victoria. Technical report series No. 56. Melbourne: Conservation, Forests and Lands: Fisheries Division Victoria, 20. Koehn JD, O’Connor WG. 1990. Biological information for the management of native freshwater fish in Victoria. Melbourne: Department of Conservation and Environment. Koster WM, Crook DA. 2008. Diurnal and nocturnal movements of river blackfish (Gadopsis marmoratus) in a southeastern Australian upland stream. Ecology of Freshwater Fish 17: 146–154. Lake JS. 1967. Freshwater fish of the Murray-Darling River system: the native and introduced species. Sydney: VCN Blight, Government Printer. Lavoue S, Sullivan JP, Hopkins CD. 2003. Phylogenetic utility of the first two introns of the S7 ribosomal protein gene in African electric fishes (Mormyroidea: Teleostei) and congruence with other molecular markers. Biological Journal of the Linnean Society 78: 273–292. Lintermans M. 1998. The ecology of the two-spined blackfish Gadopsis bispinosus (Pisces: Gadopsidae). MSc Thesis, Australian National University. Lintermans M. 2007. Fishes of the Murray-Darling Basin: an introductory guide. Canberra: Murray-Darling Basin Commission. Lintermans M. 2012. Managing potential impacts of reservoir enlargement on threatened Macquaria australasica and Gadopsis bispinosus in southeastern Australia. Endangered Species Research 16: 1–16. Lloyd LN. 1987. Ecology and distribution of the small native fish of the lower River Murray, South Australia and their 539 interactions with the exotic mosquitofish, Gambusia affinis holbrooki. MSc Thesis, University of Adelaide. Lundberg JG, Kottelat M, Smith GR, Stiassny MLJ, Gill AC. 2000. So many fishes, so little time: an overview of recent ichthyological discovery in continental waters. Annals of the Missouri Botanical Garden 87: 26–62. Lyon JP, O’Connor JP. 2008. Smoke on the water: can riverine fish populations recover following a catastrophic fire-related sediment slug? Austral Ecology 33: 794–806. McDowall RM. 2010. New Zealand freshwater fishes: an historical and ecological biogeography. New York, NY: Springer. McLaren S, Wallace M, Pillans B, Gallagher S, Miranda J, Warne M. 2009. Revised stratigraphy of the Blanchetown Clay, Murray Basin: age constraints on the evolution of paleo Lake Bungunnia. Australian Journal of Earth Sciences 56: 259–270. McLaren S, Wallace MW, Gallagher SJ, Miranda JA, Holdgate GR, Gow LJ, Snowball I, Sandgren P. 2011. Palaeogeographic, climatic and tectonic change in southeastern Australia: the Late Neogene evolution of the Murray Basin. Quaternary Science Reviews 30: 1086–1111. Miller A, Waggy G, Ryan S, Austin C. 2004. Mitochondrial 12S rRNA sequences support the existence of a third species of freshwater blackfish (Percicthyidae: Gadopsis) from south-eastern Australia. Memoirs of Museum of Victoria 61: 121–127. Moritz C. 1994. Defining ‘evolutionary significant units’ for conservation. Trends in Ecology and Evolution 9: 373–375. Morris SA, Pollard DA, Gehrke PC, Pogonoski JJ. 2001. Threatened and potentially threatened freshwater fishes of coastal New South Wales and the Murray-Darling Basin. Sydney: NSW Fisheries, 166. Morrongiello JR, Beatty SJ, Bennett JC, Crook DA, Ikedife DN, Kennard MJ, Kerezsy A, Lintermans M, McNeil DG, Pusey BJ. 2011. Climate change and its implications for Australia’s freshwater fish. Marine and Freshwater Research 62: 1082–1098. Near TJ, Benard MF. 2004. Rapid allopatric speciation in logperch darters (Percidae: Percina). Evolution 58: 2798– 2808. Near TJ, Sandel M, Kuhn KL, Unmack PJ, Wainwright PC, Leo Smith W. 2012. Nuclear gene-inferred phylogenies resolve the relationships of the enigmatic pygmy sunfishes, Elassoma (Teleostei: Percomorpha). Molecular Phylogenetics and Evolution 63: 388–395. Niemiller ML, Near TJ, Fitzpatrick BM. 2012. Delimiting species using multilocus data: diagnosing cryptic diversity in the southern cavefish, Typhlichthys subterraneus (Teleostei: Amblyopsidae). Evolution 66: 846–866. O’Connor JP, Zampatti BP. 2006. Spawning season and site location of Gadopsis bispinosus Sanger (Pisces: Gadopsidae) in a montane stream of southeastern Australia. Transactions of the Royal Society of South Australia 130: 227–232. Ogilby JD. 1913. Edible fishes of Queensland. Part II. The gadopseiform percoids. Memoirs of the Queensland Museum 2: 67–80. © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540 540 M. P. HAMMER ET AL. Ovenden JR, White RWG, Sanger AC. 1988. Evolutionary relationships of Gadopsis spp. inferred from restriction enzyme analysis of their mitochondrial DNA. Journal of Fish Biology 32: 137–148. Page LM, Burr BM. 1991. A field guide to freshwater fishes: North America north of Mexico. Boston, MA: Houghton Mifflin Harcourt. Parrish RH. 1966. A review of the Gadopsidae, with a description of a new species from Tasmania. MSc Thesis, Oregon State University. Piggott MP, Chao NL, Beheregaray LB. 2011. Three fishes in one: cryptic species in an Amazonian floodplain forest specialist. Biological Journal of the Linnean Society 102: 391–403. Posada D, Crandall KA. 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817– 818. Pulquerio MJ, Nichols RA. 2007. Dates from the molecular clock: how wrong can we be? Trends in Ecology & Evolution 22: 180–184. Raadik TA. 2011. Systematic revision of the mountain galaxias Galaxias olidus Günther species complex (Teleostei: Galaxiidae) in eastern Australia. PhD Thesis, University of Canberra. Rambaut A, Drummond AJ. 2007. Tracer, Version 1.5. Available at: http://beast.bio.ed.ac.uk/Tracer Richardson BJ, Baverstock PR, Adams M. 1986. Allozyme electrophoresis: a handbook for animal systematics and population studies. Sydney: Academic Press. Richardson J. 1848. Ichthyology of the voyage of HMS. Erebus & Terror. In: Richardson J, Gray JE, eds. The zoology of the voyage of HMS Erebus & Terror, under the command of Captain Sir James Clark Ross during the years 1839 to 1843. London: EW Janson, 1–139. Ryan SG, Miller AD, Austin CM. 2004. Allozyme variation and taxonomy of the river blackfish, Gadopsis marmoratus Richardson, in western Victoria. Proceedings of the Royal Society of Victoria 116: 191–199. Sanger A. 1984. Description of a new species of Gadopsis (Pisces: Gadopsidae) from Victoria. Proceedings of the Royal Society of Victoria 96: 93–97. Sanger AC. 1986. The evolution and ecology of the Gadopsis marmoratus complex. PhD Thesis, University of Melbourne. Stephenson A. 1986. Lake Bungunnia – a Plio-Pleistocene megalake in southern Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 57: 137–156. Swofford DL. 2002. PAUP*. Phylogenetic analysis using parsimony (*and other methods), Version 4. Sunderland, MA: Sinauer Associates. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28: 2731–2739. Unmack PJ. 2001. Biogeography of Australian freshwater fishes. Journal of Biogeography 28: 1053–1089. Unmack PJ. 2013. Biogeography. In: Humphries P, Walker KF, eds. Ecology of Australian freshwater fishes. Collingwood: CSIRO Publishing, 25–48. Unmack PJ, Bagley JC, Adams M, Hammer MP, Johnson JB. 2012. Molecular phylogeny and phylogeography of the Australian freshwater fish genus Galaxiella, with an emphasis on dwarf galaxias (G. pusilla). PLoS ONE 7: e38433. Unmack PJ, Bennin AP, Habit EM, Victoriano PF, Johnson JB. 2009. Impact of ocean barriers, topography, and glaciation on the phylogeography of the catfish Trichomycterus areolatus (Teleostei: Trichomycteridae) in Chile. Biological Journal of the Linnean Society 97: 876– 892. Unmack PJ, Hammer MP, Adams M, Dowling TE. 2011. A phylogenetic analysis of pygmy perches (Teleostei: Percichthyidae) with an assessment of the major historical influences on aquatic biogeography in southern Australia. Systematic biology 60: 797–812. Unmack PJ, Hammer MP, Adams M, Johnson JB, Dowling TE. 2013. The role of continental shelf width in determining freshwater phylogeographic patterns in south-eastern Australian pygmy perches (Teleostei: Percichthyidae). Molecular Ecology 22: 1683–1699. Utter F. 2004. Population genetics, conservation and evolution in salmonids and other widely cultured fishes: some perspectives over six decades. Reviews in Fish Biology and Fisheries 14: 125–144. Vanhove MPM, Economou AN, Zogaris S, Larmuseau MHD, Giakoumi S, Kalogianni E, Volckaert FAM, Huyse T. 2012. Phylogenetics and biogeography of the Balkan ‘sand gobies’ (Teleostei: Gobiidae): vulnerable species in need of taxonomic revision. Biological Journal of the Linnean Society 105: 73–91. Waters JM, Lintermans M, White RWG. 1994. Mitochondrial DNA variation suggests river capture as a source of vicariance in Gadopsis bispinosus (Pisces: Gadopsidae). Journal of Fish Biology 44: 549–551. Whitley GP. 1962. The slippery or freshwater blackfish. Education Gazette 56: 95–96. Williams J, Read C, Norton A, Dovers S, Burgman M, Proctor W, Anderson H. 2001. Biodiversity, Australia state of the environment report 2001 (theme report). Collingwood: CSIRO Publishing. Zwickl D. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. PhD Thesis, The University of Texas. ARCHIVED DATA Data deposited in the Dryad digital repository (Hammer et al., 2014). © 2014 The Linnean Society of London, Biological Journal of the Linnean Society, 2014, 111, 521–540
© Copyright 2024 Paperzz