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