A revision of the gudgeon genus Hypseleotris (Gobiiformes: Gobioidei: Eleotridae) of northwest Australia, describing three new species and synonymizing the genus Kimberleyeleotris

Abstract

Species within the northwest Australian clade of Hypseleotris (six species) and the genus Kimberleyeleotris (two species) are reviewed following the recording of new populations in the region and a molecular study of the group that identified three undescribed candidate species. Based on the analysis of extensive morphological and nuclear and mitochondrial molecular datasets, Kimberleyeleotris is here formally synonymised with Hypseleotris. Furthermore, three species from the Kimberley region, Western Australia, are described to science: Hypseleotris maranda sp. nov., Hypseleotris wunduwala sp. nov., and Hypseleotris garawudjirri sp. nov. The presence of, or number of scales across the head and body, the pattern of sensory papillae on the head, fin ray counts, dorsal and anal fin colouration (particularly in breeding males), and body depth, can be used to distinguish the members of the northwest Australia lineage. Furthermore, the newly described species were genetically separated from all northwest Australian congeners by K2P distances ranging from 7.8–11.3% based on the CO1 gene, and 7.7–16.3 % based on the entire mitochondrial genome. Two of the new species, H. maranda sp. nov. and H. wunduwala sp. nov., have extremely narrow ranges being found in single sub-catchments of the Roe and King Edward Rivers respectively. On the other hand, H. garawudjirri sp. nov. is moderately widespread, being found across the Charnley, Calder, and Sale rivers. While the conservation risk to H. maranda sp. nov. and H. wunduwala sp. nov. is inherently high due to their small range, there are currently no obvious local threatening processes to either of these species given their remote locations that are little impacted by human activities.

Full text

https://doi.org/10.11646/zootaxa.5311.3.2
http://zoobank.org/urn:lsid:zoobank.org:pub:FBADEB71-1057-4E84-8B70-A30E99F27041
340 Accepted by R. Pethiyagoda: 8 May 2023; published: 30 Jun. 2023
Article ZOOTAXA
ISSN 1175-5326 (print edition)
ISSN 1175-5334 (online edition)
Zootaxa 5311 (3): 340–374
https://www.mapress.com/zt/
Copyright © 2023 Magnolia Press
A revision of the gudgeon genus Hypseleotris (Gobiiformes: Gobioidei:
Eleotridae) of northwest Australia, describing three new species and
synonymizing the genus Kimberleyeleotris
JAMES J. SHELLEY1,2*, AURÉLIEN DELAVAL1,3 & MATTHEW C. LE FEUVRE1
1School of BioSciences, University of Melbourne, Victoria 3010, Australia
MATTHEW C. LE FEUVRE: https://orcid.org/0000-0001-9592-5927
2Ichthyology, Sciences Department, Museums Victoria, Victoria 3001, Australia
3Present address: Institute of Marine Research, Bergen 5817, Norway
AURÉLIEN DELAVAL: https://orcid.org/0000-0002-2223-2932
*Correspondence:
�
jamesshelley85@hotmail.com; https://orcid.org/0000-0002-2181-5888
Abstract
Species within the northwest Australian clade of Hypseleotris (six species) and the genus Kimberleyeleotris (two species)
are reviewed following the recording of new populations in the region and a molecular study of the group that identified
three undescribed candidate species. Based on the analysis of extensive morphological and nuclear and mitochondrial
molecular datasets, Kimberleyeleotris is here formally synonymised with Hypseleotris. Furthermore, three species from
the Kimberley region, Western Australia, are described to science: Hypseleotris maranda sp. nov., Hypseleotris wunduwala
sp. nov., and Hypseleotris garawudjirri sp. nov. The presence of, or number of scales across the head and body, the pattern
of sensory papillae on the head, fin ray counts, dorsal and anal fin colouration (particularly in breeding males), and body
depth, can be used to distinguish the members of the northwest Australia lineage. Furthermore, the newly described
species were genetically separated from all northwest Australian congeners by K2P distances ranging from 7.8–11.3%
based on the CO1 gene, and 7.7–16.3 % based on the entire mitochondrial genome. Two of the new species, H. maranda
sp. nov. and H. wunduwala sp. nov., have extremely narrow ranges being found in single sub-catchments of the Roe and
King Edward Rivers respectively. On the other hand, H. garawudjirri sp. nov. is moderately widespread, being found
across the Charnley, Calder, and Sale rivers. While the conservation risk to H. maranda sp. nov. and H. wunduwala sp.
nov. is inherently high due to their small range, there are currently no obvious local threatening processes to either of these
species given their remote locations that are little impacted by human activities.
Key words: Eleotridae, range-restricted, freshwater, biodiversity, taxonomy, systematics
Introduction
The gudgeon genus Hypseleotris is the most speciose eleotrid radiation in Australia with 13 described species
(Thacker et al. 2022a; b). All but one of these species are restricted to Australian fresh waters, while H. compressa
is found in estuaries and lower parts of rivers around much of the coastline of Australia and the southern coastline of
the nearby island of New Guinea. These two islands shared a land bridge as recently as the Last Glacial Maxima (de
Groeve et al. 2022). A further six species are found across the Indo-Pacific region, mainly on islands. The ecology
of many of these species is poorly known, but like H. compressa most occupy estuaries and the lower parts of rivers
and are known or assumed to be diadromous (Keith & Mennesson 2023).
The Australian endemic species form two geographically separated clades that occur in the northwest and
southeast corners of the country (Thacker et al. 2022b). By virtue of the remoteness of northwest Australia, the
diversity and distribution of the freshwater fish fauna is considered poorly understood to western science (Shelley et
al. 2019). Research effort is limited compared to heavily populated southern regions of Australia and advancements
in knowledge of the region’s biodiversity tend to be sporadic, but substantial (Shelley et al. 2018a). Regardless
of the lesser research attention, the Kimberley region of northwest Australia has been shown to be a nationally

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 341
significant hotspot of freshwater fish endemism, with more endemic species than anywhere else in the country
(Shelley et al. 2018a; b).
For the northwest Australian Hypseleotris endemics, Shipway (1950) first described H. aurea from the Pilbara
region, Hoese & Allen (1983) then described H. ejuncida, H. kimberleyensis and H. regalis from the Kimberley
region, and Larson (2007) described H. barrawayi from the Arnhem Plateau. Further to this, Hoese & Allen (1987)
described a morphologically similar genus endemic to the Kimberley, Kimberleyeleotris, containing two species
K. hutchinsi and K. notata. Kimberleyeleotris was distinguished from Hypseleotris based on modest differences in
head papilla patterns, having reduced scales (although similar scale counts), the presence of a few small, loosely
attached vomerine teeth, and a slightly larger mouth. However, a recent phylogenetic analysis based on nuclear ultra-
conserved element sequences and complete mitochondrial genomes, which include all known Hypseleotris species
and Kimberleyeleotris hutchinsi samples, suggested the two genera were synonymous (Thacker et al. 2022b).
Since these initial scientific discoveries, surveys in the Kimberley region have revealed additional, geographically
separate populations of Hypseleotris in the Calder River catchment (Allen & Leggett 1990), the King Edward River
catchment (Morgan et al. 2006), the Charnley and Roe river catchments (Shelley et al. 2018a), and the Sale River
(Hammer & Moore, in prep). While the populations in the Calder and King Edward rivers were initially attributed to
H. kimberleyensis and H. ejuncida respectively, further collections and preliminary analysis indicated that they were
morphologically distinct from these species (Shelley et al. 2018a). The population found in the Roe River catchment
was also distinct in appearance from other described species (Shelley et al. 2018a).
Based on these developments, these three putative species were included in the analysis of Thacker et al.
(2022b) which indicated that they represent distinct evolutionary lineages that are as divergent, or more, than other
species in the genus. These recent collections and genetic analysis provide a sound basis for a formal taxonomic
examination of the new populations and revision of the northwest Australian Hypseleotris and Kimberleyeleotris.
Here, we summarise the nuclear and mitochondrial genomic evidence, and present colour photographs, distribution
data, and morphological data from all northwest Australian Hypseleotris species collected across their entire range to
demonstrate that the newly recorded populations represent three distinct species of Hypseleotris, which we describe
herein. Furthermore, we place Kimberleyeleotris morphological data in the context of all northwest Australian
Hypseleotris species to show that the morphological differences that were used to described Kimberleyeleotris fall
within the range of natural variation within the Hypseleotris, thus providing formal morphological support for their
synonymy. Documenting these species new to science contributes to our growing understanding of the Kimberley
biodiversity hotspot and provides taxonomic clarification among a species-rich complex of narrow-range freshwater
fishes.
Methods
Study material and location data
All type material used in this study, belonging the putative new species, was collected by authors JJS and MCL
during a comprehensive survey of the Kimberley region between 2012–2016. In total, 70 sites were sampled
across 19 major river catchments. This included the sampling of river catchments flowing off the western and
north-western Kimberley coast that are primarily only accessible by helicopter and have thus received little, or no,
research collection effort. Hypseleotris were collected from 15 sites across six major river catchments using dip,
fyke, and seine nets, and backpack electrofishing. Only 10 and 12 specimens could be collected from the Roe and
King Edward River catchments due to the difficulty in finding and catching them.
Following collection, selected specimens were euthanized (AQUI-S® solution), and species/morphotype
identification was conducted. Whole fish were either fixed in a 10 % solution of neutral buffered formalin for
morphological analysis or in pure ethanol for genetic analysis and stored in plastic bottles. After two weeks, formalin
preserved samples were soaked in freshwater and transferred into 70 % ethanol for long-term storage. All specimens
of Hypseleotris compared with imagery of the holotypes and some paratypes or typotypic specimens of each of
the northwest Australian Hypseleotris species (See Additional material section). Holotypes of each of the existing
species could not be physically examined by us due to changes in loan regulations and collection access limitations
during the COVID pandemic. However, a digital image of holotype and the corresponding data presented in Hoese
& Allen (1983) were available for comparison. A summary of substantiated Hypseleotris locality records in the

SHELLEY ET AL.
342 · Zootaxa 5311 (3) © 2023 Magnolia Press
Kimberley region is presented in Shelley et al. (2018a). The species records presented here utilise this dataset as
well as records from a recent survey of the western Kimberley catchments by Hammer & Moore (in prep). Together
there were 10 location records for the putative new species.
Additional material
Material from all other northwest Australia Hypseleotris species, and Kimberleyeleotris hutchinsi, were inspected
to check previous descriptions and to act as a point for comparison with the putative new species. The detail of this
additional material is as follows.
Hypseleotris regalis: WAM P.25028-007 (Paratypes), 5 (23–40 mm SL), collected with holotype, Australia,
Western Australia, Kimberley region, Wyulda Creek, about 2 km above junction with Roe River, approximately
15° 26’ S, 125° 37’ E, collected by G. Allen, 17 August 1974; NMV A.31770-001, 33 (23–40 mm SL), Australia,
Western Australia, Kimberley region, Prince Regent River, mid-section of Prince Regent River, 15°42’28.8”S,
125°26’20.4”E.
Hypseleotris aurea: WAM P.25122-006, 13 (33–44 mm SL), Australia, Western Australia, Pilbara region,
Murchison River, approximately 27° 48’ S, 114° 42’ E, collected by G. Allen, 23 November 1974.
Hypseleotris barrawayi: NTM S.15612-001 (Paratypes), 9 (18–44), Australia, Northern Territory, Arnhem Land
region, site 4, Nibuldakya Aquatic Survey, gorge tributary on upper reaches Katherine River, 13° 18.7’ S, 133° 6.5’
E, collected by D. Wedd and N. Smit, 7 July 1997.
Hypseleotris compressa: WAM P.25036-003, 35 (13–37 mm SL), Australia, Western Australia, Kimberley
region, Prince Regent River, King’s Cascades, 15° 37’ 19” S, 125° 18’ 18” E, collected by G. Allen, 27 August
1974.
Hypseleotris ejuncida: WAM P.25032-008 (Paratypes), 4 (30–40 mm SL), collected with the holotype,
Australia, Western Australia, Kimberley region, Gundarara Creek, about 2 km above junction with Prince Regent
River, approximately 15° 49’ S, 125° 37’ E, collected by G. Al1en, 21 August 1974. We note that this registration
number does not match that in Hoese & Allen (1983).
Hypseleotris kimberleyensis: WAM P.25454-007 (Paratypes), 3 (29.0–34.8 mm SL), collected with the holotype,
Australia, Western Australia, Kimberley region, Barnett River near Barnett Gorge, approximately 16° 32’ S, 126°
00’ E, Central Kimberley, Western Australia, collected by B. Hutchins and A. Chapman, 30 July 1975.
Kimberleyeleotris hutchinsi: WAM P.25684-007 (Paratypes), 14 (25–30 mm SL), collected with the holotype,
Australia, Western Australia, Kimberley region, tributary of Mitchell River, approximately 14° 49’ S, 125° 42’ E,
collected by B. Hutchins and T. Dryker, 31 October 1976.
Taxonomic approach
In this study we adopt the Unified Species Concept (USC), proposed by de Queiroz (2007), for determining discrete
species. The USC integrates a range of criteria drawn from other established species concepts that can be selectively
applied in accordance with the circumstances of the target group. In the case of allopatric groups, such as discussed
here, the diagnosis of evolutionary species is based on finding multiple, independent, genetic characters that
concordantly diagnose the presence of evolutionarily-independent taxa. Morphological and geographic information
must then be used to delineate biological or evolutionary species consistent with the USC.
In surveys conducted since the initial description of northwest Australian Hypseleotris species and their
distributions (Hoese & Allen 1983), new, geographically separate populations of Hypseleotris were discovered
in the neighbouring Charnley, Calder, and Sale River catchments (Allen & Leggett 1990; Shelley et al. 2018a;
Hammer & Moore in prep), the King Edward River catchment (Morgan et al. 2006), and the Roe River catchment
(Shelley et al. 2018a). As mentioned above, preliminary analysis indicated that there were three morphologically
distinct groups within these populations (Shelley et al. 2018a).
Based on these conclusions, samples of these three putative species were included in a phylogenetic analysis
of the entire Hypseleotris genus by Thacker et al. (2022b), based on nuclear ultra-conserved element sequences
containing 251 loci ranging from 802 to 2854 base pairs in length (95% CI ± 31) and mitochondrial genomes
composed of 16,729 base pairs. While the analyses presented in Thacker et al. (2022b) weren’t focused on species

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 343
distinction, the hypothesised phylogenies provide an indication of: (1) how divergent the three putative species
are relative to others in the genus, (2) concordance in relationships across hundreds of nuclear loci and complete
mitochondrial genomes, and (3) the evolutionary context for the species based on their phylogenetic placement
and estimated divergence times. Here we present the mitochondrial genome Maximum Likelihood tree and time
calibrated UCE Bayesian tree from Thacker et al. (2022b) to support the morphological evidence presented here and
aid with interpretation of the results. Furthermore, we calculate and present the proportion of genetic differences (p-
distance) in the Cytochrome c oxidase I (CO1) gene and the complete mitochondrial genome between species and
putative species found in northwest Australia. Such analysis of the mitochondrial genomes allows for two approaches
commonly used in assessing species-level identification, namely (1) a relative yardstick approach, comparing levels
of observed differentiation to that witnessed in broader Hypseleotris phylogeny; and (2) an indicative barcoding
K2P distance measure threshold of 1–3% for the CO1 gene (April et al. 2011; Ward 2009).
Molecular analysis
The complete methods and results for the analysis underlying the phylogenies presented here can be found in
Thacker et al. (2022b). In this section we present some details, adapted from Thacker et al. (2022b), which are key
to understanding and interpreting these results.
Thacker et al. (2022b) assembled a dataset of 40 individuals that encompassed the nine described Hypseleotris
species, six now described southeastern species, all known hemiclone hybrid crosses between selected southeastern
species, and the three putative species from the Kimberley region that are the focus of this study. Further to the target
group, they included Kimberleyeleotris hutchinsi and six outgroup taxa from the families Eleotridae, Butidae, and
Odontobutidae. Specimens of H. cyprinoides, the only Hypseleotris species that is not present in Australia but is
otherwise widespread throughout the Indo-Pacific, were also included from New Guinea (Indonesia), Papua New
Guinea, and South Africa.
Library preparation, enrichment, and UCE sequencing was conducted by Arbor Biosciences (Ann Arbor, MI,
USA). The UCE sequencing was performed in two batches using different targeted enrichment probe sets (0.5Kv1
and 1Kv1). The UCE sequences were then aligned in MAFFT version 7.4, using the FFT-NS-i algorithm. Finally,
alignments were screened for minimum taxonomic coverage requirements of 75 and 95%, and alignments were
concatenated into sequential format, ready for analysis. The mitochondrial genomes were assembled from UCE
off target reads by aligning the Trinity-assembled contigs for each species to the mitochondrial genome of Eleotris
acanthopoma (GenBank accession number AP004455) using the Map to Reference assembly tool in Geneious
Prime 2020.0.5 (www.geneious.com).
To estimate the phylogeny, partitioned maximum likelihood (ML) phylogenetic analyses were performed on
both the UCE and mitochondrial datasets using RAxML 8.2.12 (Stamatakis 2014) at the CIPRES science gateway
at phylo.org, using the XSEDE supercomputer (Miller et al. 2010). For Bayesian phylogenetic analysis, also on
both the UCE (with and without hemiclones) and mitochondrial datasets, a multi-threaded, MPI hybrid version of
ExaBayes was used to sample the Bayesian posterior distribution of phylogenetic trees with MCMC simulations,
also implemented at the CIPRES science gateway and using the XSEDE supercomputer (Miller et al. 2010).
The UCE phylogeny, which didn’t include hemiclonal hybrids, was used for time-calibration with MCMCTree,
implemented in PAML 4.8 (Yang 2007; Yang & Rannala 2006) and using the two-step approximate likelihood
calibration procedure of (Yang & Rannala 2006). The independent rates clock model, GTR + gamma substitution
model, and a root age estimate of 53.9 Ma were applied in line with the results of Thacker (2017). Three calibrations
were applied: root of Eleotridae at 45.5 Ma (95% highest posterior density interval: 34.0–57.6 Ma), divergence
of Giurus margaritacea and Mogurnda adspersa at 19.3 Ma (10.0–29.1 Ma), and root of Australian Hypseleotris
(excluding H. cyprinoides) at 9.0 Ma (4.3–15.8 Ma). These dates are legacy calibrations derived from fossil-based
calibrations in Thacker (2014) and Thacker (2017), and they were applied as ranges corresponding to the 95%
confidence intervals of the estimated ages.
Finally, to assess genetic similarity between populations of our target species, we calculated the proportion of
genetic differences (p-distance) between the mitochondrial genomes and the CO1 gene specifically using MEGA
6.06. This analysis was only run on the mitochondrial dataset as it is comparable with the species threshold criteria
presented in Ward et al. (2009) and April et al. (2011).

SHELLEY ET AL.
344 · Zootaxa 5311 (3) © 2023 Magnolia Press
Morphological data collection
Counts and measurements presented include those of Hoese & Allen (1983), Hoese & Allen (1987), and Larson
(2007). To help eliminate operator bias, counts and measurements were calibrated by re-examining the paratypes of
described species (see Additional material section). As in these papers, the methods here generally follow Hubbs &
Lagler (1970), with the following exceptions. The pterygiophore formula follows Birdsong et al. (1988). Transverse
scale counts backward are taken by counting the number of scale rows from the anal fin origin diagonally upward
and back toward the second dorsal fin base. Head length is taken to the upper attachment of the opercular membrane.
Only the right pectoral ray numbers were recorded. Finally, the segmented or branched caudal ray pattern (e.g.
9/8 or 9/7) is the number of segmented caudal rays attaching to the upper and lower hypural plates respectively.
Vertebral counts and branched and unbranched caudal ray counts were obtained by X-ray.
Measurements were taken with a Vernier calliper and dissecting microscope, recorded to the nearest 0.1 mm,
and converted to a percentage of the standard length (SL) or head length (HL).
Type material was deposited at Museums Victoria, Melbourne (NMV) and Western Australian Museum, Perth
(WAM), with additional material loaned from Museum and Art Gallery of the Northern Territory (previously
Northern Territory Museum), Darwin.
Results
Molecular analysis
After alignment and trimming, the 251 loci in the UCE dataset ranged between 802 and 2854 base pairs in length
(95% CI ± 31). The matrix of mitochondrial genomes was composed of 16,729 base pairs and included 34 taxa,
representing all the Hypseleotris species, Kimberleyeleotris hutchinsi, and the hemiclones. Among the outgroup
taxa only Mogurnda adspersa yielded a complete mitochondrial genome from the off target reads, so only that
species was used as the outgroup in phylogenetic analyses.
The phylogenetic hypothesis of Hypseleotris and outgroups based on analysis of UCE loci with both ML and
Bayesian inference is given in Fig. 1. The phylogeny received excellent support across most nodes, with 97–100%
bootstrap support in the ML analysis and 100% posterior probability in the Bayesian analysis. The only node
that did not receive strong support in both analyses was that separating H. barrawayi and the lineage containing
Kimberleyeleotris hutchinsi, H. kimberleyensis, and H. sp. 3 in the northwest clade (Bayesian = 100% posterior
probability; ML = 64% bootstrap support).
The crown age of Hypseleotris is estimated at 17.3 Ma (95% highest posterior density interval: 12.6–22.8 Ma)
and of Australian Hypseleotris at 10.1 Ma (7.4–13.4 Ma). At the crown of the Australian radiation, two main clades
form that encompass species from both southeast and northwest Australia, as expected from previous phylogenetic
analysis of the genus (Thacker & Unmack 2005). Within the Australian radiation, the northwest clade, plus H.
compressa, is dated at 8.9 Ma (6.5–11.7 Ma). Next, H. aurea from the Pilbara region of Western Australia diverges
from the rest of the northwest Australia clade at 7.2 Ma (5.3–9.6 Ma), followed by eight species distributed in the
Kimberley and neighboring Arnhem Land region. These include the named species H. kimberleyensis, H. barrawayi,
H. ejuncida, and H. regalis, Kimberleyeleotris hutchinsi, and the three putative Hypseleotris species dealt with here.
The Kimberley/Arnhem Land species comprise two clades that diverged at 6.2 Ma (4.5–8.2 Ma): one including H.
barrawayi, K. hutchinsi, H. kimberleyensis and H. sp. 3 (Carson River, King Edward River catchment), and a second
including H. sp. 1 (Garimbu Creek, Roe River catchment), H. sp. 2 (Bachsten River, Calder River catchment), H.
ejuncida, and H. regalis. Of the three putative species, H. sp. 3 (Carson River, King Edward River catchment)
diverged at 2.7 Ma (1.7–4.0 Ma), H. sp. 2 (Bachsten River, Calder River catchment) diverged at 5.1 Ma (3.6–7.0
Ma), and H. sp. 1 (Garimbu Creek, Roe River catchment) diverged at 5.8 Ma (4.2–7.7 Ma).
The phylogenetic hypothesis based on complete mitochondrial genomes is given in Fig. 2 and shows a similar
topology within the northwest Australian clade and the broader phylogeny.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 345
TABLE 1. Genetic distances (%) between northwest Australian species based on the Cytochrome c oxidase I (CO1) gene (values outside parentheses) and the entire mitochondrial
genome (values inside parentheses). Note that no data exists for H. notata.
H. aurea H. barrawayi H. compressa H. ejuncida H. garawudjirri
sp. nov.
H. hutchinsi H. kimberleyensis H. maranda
sp. nov.
H. regalis H. wunduwala
sp. nov.
H. aurea -
H. barrawayi 9.3 (8.8) -
H. compressa 5.2 (6.7) 8.2 (7.8) -
H. ejuncida 7.2 (7.8) 9.4 (8.5) 5.2 (6.9) -
H. garawudjirri sp. nov. 7.8 (13.7) 8.9 (15.0) 5.8 (13.7) 8.3 (14.5) -
H. hutchinsi 6.0 (6.4) 8.8 (8.3) 5.6 (6.3) 7.4 (7.3) 7.0 (13.1) -
H. kimberleyensis 8.7 (11.6) 9.7 (12.6) 7.7 (11.8) 9.0 (12.1) 8.0 (16.3) 7.8 (10.7) -
H. maranda sp. nov. 8.1 (8.9) 8.7 (9.0) 8.1 (7.7) 7.6 (8.4) 7.8 (15.1) 7.5 (8.4) 8.9 (12.8) -
H. regalis 7.1 (7.8) 9.3 (8.4) 5.8 (6.8) 1.9 (1.6) 8.4 (14.4) 7.3 (7.2) 9.0 (12.0) 7.9 (8.4) -
H. wunduwala sp. nov. 8.9 (10.4) 11.3 (11.4) 8.4 (10.4) 10.1 (10.8) 9.9 (16.0) 9.3 (9.7) 8.5 (10.7) 10.1 (11.7) 10.2 (10.2) -

SHELLEY ET AL.
346 · Zootaxa 5311 (3) © 2023 Magnolia Press
FIGURE 1. UCE phylogeny hypothesis, dated with MCMC tree. All nodes are supported at 1.0 posterior probability in the
Bayesian hypothesis, and 97–100% bootstrap support in the ML hypothesis, except where indicated; at those nodes the posterior
probability/bootstrap support are shown. Bars on nodes are 95% highest posterior density intervals of age estimates shown in the
lower scale. Hypseleotris individuals are labeled in parentheses with the locality where they were collected, not the entire range
of the taxon. The figure is modified from the phylogeny depicted in Thacker et al. 2022b.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 347
FIGURE 2. Hypothesized phylogeny based on analysis of complete mitochondrial genomes. All nodes are supported at 1.0
posterior probability in the Bayesian hypothesis, and 100% bootstrap support in the ML hypothesis, except where indicated;
at those nodes the posterior probability/bootstrap support are shown. Hypseleotris individuals are labeled in parentheses with
the locality where they were collected, not the entire range of the taxon. The figure is modified from the phylogeny depicted in
Thacker et al. 2022b.

SHELLEY ET AL.
348 · Zootaxa 5311 (3) © 2023 Magnolia Press
The results of the analysis of genetic distance between the northwest Hypseleotris species, based on the
mitochondrial dataset, are presented in Table 1. Across the northwest Australia clade, described and putative species
were separated from each other by a p-distance of 1.9 to 11.3% (median 8.2%) for the CO1 gene and 1.6 to 16.3%
(median 10.2%) across the mitochondrial genome. In the context of the northwest clade each of the putative species
exhibit genetic differentiation that is consistent with species level differences observed between described species,
often being greater than those. For instance, H. sp. 1 (Garimbu Creek, Roe River catchment) is separated from all
other species in the northwest Australia clade by a p-distance of 7.5 to 10.1% (median 8.1%) in the CO1 gene and
7.7 to 15.1% (median 8.9%) in the mitochondrial genome. H. sp. 2 (Bachsten River, Calder River catchment) was
separated from all other species in the clade by a p-distance 5.8 to 9.9% (median 8.0%) for the CO1 gene and 13.1
to 16.3% (median 14.5%). Finally, H. sp. 3 (Carson River, King Edward River catchment) was separated from all
other species in the clade by a p-distance 8.5 to 11.3% (median 9.9%) in the CO1 gene and 9.7 to 16.0% (median
10.7%) across the mitochondrial genome.
Finally, all species within the northwest clade exceeded the proposed K2P distance measure threshold for the
CO1 gene of 1–3% for species distinction, with the three putative species greatly exceeding this threshold (April
et al. 2011; Ward 2009). Taking each of these results into account, we consider that each of the putative species are
genetically distinct.
Morphological analysis
We collated a dataset consisting of 179 individuals for morphology: 13 H. sp. 1 (Garimbu Creek, Roe River
catchment), 12 H. sp. 2 (Bachsten River, Calder River catchment), 10 H. sp. 3 (Carson River, King Edward River
catchment), 9 H. aurea, 30 H. barrawayi, 30 H. compressa, 13 H. ejuncida, 25 H. hutchinsi, 14 H. kimberleyensis,
23 H. regalis. We summarize the characters and meristic counts in Table 2 to Table 9. Ultimately, we found a range
of characters that, in varying combinations, can reliably distinguish between each of the described and putative
species. We also summarize the morphological evidence, showing that the characters exhibited by Hypseleotris
hutchinsi (previously Kimberleyeleotris hutchinsi) fall within the range observed within the Hypseleotris genus. As
such we include it in the description of the Hypseleotris genus, specifically the presence of vomerine teeth.
We make these findings clear in a key to the Hypseleotris species of the northwest Australia. Together, the
congruent morphological evidence, nuclear and mitochondrial genetic data, and geographic information provide a
sufficient weight of evidence to justify the recognition of three new species, replacing the three candidate species
codes (H. sp. 1, H. sp. 2, and H. sp. 3). These species are formally described below, but for clarity, we use the new
scientific names throughout the rest of the paper and in all figures and tables: H. maranda sp. nov. for H. sp. 1; H.
garawudjirri sp. nov. for H. sp. 2 (formerly H. kimberleyensis); H. wunduwala sp. nov. for the H. sp. 3 (formerly
H. ejuncida in part).
TABLE 2. Frequency distribution of dorsal and anal fin soft ray counts in northwest Australian Hypseleotris species; data
on other species from Larson (2007) and Hoese and Allen (1983).
Species Second dorsal fin rays Anal fin rays
n8 9 10 11 12 13 9 10 11 12 13
H. aurea 9- 1 8 - - - - 1 7 1 -
H. barrawayi 31,30 - 4 26 1 - - 2 15 13 - -
H. compressa 31 - 24 7 - - - 1 16 14 - -
H. ejuncida 13 - 3 10 - - - 1 10 2 - -
H. garawudjirri sp. nov. 12 - - 5 6 - 1 - 2 9 1 -
H. hutchinsi 25 18 7 - - - - 12 13 - - -
H. kimberleyensis 14 - 1 12 1 - - 1 9 4 - -
H. maranda sp. nov. 13 - 1 6 5 1 - - 5 6 1 1
H. notata 2- 2 - - - - 1 1 - - -
H. regalis 23,6 1 4 18 - - - 2 4 - - -
H. wunduwala sp. nov. 10 - 2 8 - - - - 8 2 - -

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 349
TABLE 3. Frequency distribution of pectoral fin ray counts (right fin) in northwest Australian Hypseleotris species; data
on other species from Larson (2007) and Hoese and Allen (1983).
Species Pectoral fin rays
n13 14 15 16 17
H. aurea 10 - 1 6 3 -
H. barrawayi 31 - 7 22 2 -
H. compressa 29 - 7 20 1 1
H. ejuncida 12 - 6 6 - -
H. garawudjirri sp. nov. 12 - 5 7 - -
H. hutchinsi 25 - 4 20 1 -
H. kimberleyensis 13 - 8 5 - -
H. maranda sp. nov. 13 - - 13 - -
H. notata 2- 2 - - -
H. regalis 21 7 12 2 - -
H. wunduwala sp. nov. 10 3 6 1 - -
TABLE 4. Frequency distribution of reverse transverse scale counts in northwest Australian Hypseleotris species; data
on other species from Larson (2007) and Hoese and Allen (1983).
Species Reverse transverse scale counts
n5 6 7 8 9 10 11 12 13
H. aurea 9- - - - 1 1 5 1 1
H. barrawayi 30 - - - - - 4 9 14 3
H. compressa 22 - - - 1 9 12 - - -
H. ejuncida 11 - - 4 6 - 1 - - -
H. garawudjirri sp. nov. 12 - - - - 7 5 - - -
H. hutchinsi 12 1 3 2 - 3 3 - - -
H. kimberleyensis 11 - - - 1 9 1 - - -
H. maranda sp. nov. 13 - - - 1 4 5 2 1 -
H. notata 2- - - 2 - - - - -
H. regalis 20 - - 1 15 4 - - - -
H. wunduwala sp. nov. 10 - - - 2 3 5 - - -
TABLE 5. Frequency distribution of postdorsal scale counts in northwest Australian Hypseleotris species; data on other
species from Larson (2007) and Hoese and Allen (1983).
Species Postdorsal scales
n7 8 9 10 11 12 13
H. aurea 9 - - - 2 1 3 3
H. barrawayi 29 - 7 20 2 - - -
H. compressa 34 - 14 20 - - - -
H. ejuncida 12 - 11 1 - - - -
H. garawudjirri sp. nov. 12 - 1 9 1 1 - -
H. hutchinsi -No measurements taken
H. kimberleyensis 11 2 7 2 - - - -
H. maranda sp. nov. 13 - - 7 3 2 1 -
H. notata -No measurements taken
H. regalis 21 2 19 - - - - -
H. wunduwala sp. nov. 10 - 6 3 1 - - -

SHELLEY ET AL.
350 · Zootaxa 5311 (3) © 2023 Magnolia Press
TABLE 6. Frequency distribution of longitudinal scale counts in northwest Australian Hypseleotris species; data on other species from Larson (2007) and Hoese and Allen
(1983).
Species Longitudinal scale row count
n24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 43
H. aurea 9- - - - - - - - - - 1 - - 1 1 3 2 1
H. barrawayi 30 - - - - - - 7 5 11 6 1 - - - - - - -
H. compressa 24 - 1 3 7 8 5 - - - - - - - - - - - -
H. ejuncida 11 - - - - 1 3 4 3 - - - - - - - - - -
H. garawudjirri sp. nov. 12 - - - - - - 6 2 4 - - - - - - - - -
H. hutchinsi 20 - 2 - - 3 1 1 2 4 3 3 1 - - - - -
H. kimberleyensis 14 - - - - - 8 1 3 2 - - - - - - - - -
H. maranda sp. nov. 13 - - - - - 1 3 5 4 - - - - - - - - -
H. notata 2- 1 - 1 - - - - - - - - - - - - - -
H. regalis 22 5 8 9 - - - - - - - - - - - - - - -
H. wunduwala sp. nov. 10 - - - - 1 2 4 2 - 1 - - - - - - - -

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 351
TABLE 7. Frequency distribution of predorsal midline scale counts in northwest Australian Hypseleotris species; data on other species from Larson (2007) and Hoese and Allen
(1983).
Species Predorsal midline scale count
n0 2 3 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 23
H. aurea 9- - - - - - - - - - - - - 1 1 1 3 2 1 -
H. barrawayi 31 7 - 1 1 - 1 1 1 2 2 1 1 3 4 3 1 - 1 - 1
H. compressa 24 - - - - - - - - - - - 3 5 10 5 1 - - - -
H. ejuncida 11 - 1 1 - 1 - - 1 - 2 - 1 2 - - - 1 1 - -
H. garawudjirri sp. nov. 12 10 - - - - - - - - - - 1 1 - - - - - - -
H. hutchinsi 25 25 - - - - - - - - - - - - - - - - - - -
H. kimberleyensis 14 14 - - - - - - - - - - - - - - - - - - -
H. maranda sp. nov. 13 - - - - - - - - - - - 3 - 1 1 - 3 3 2 -
H. notata 22 - - - - - - - - - - - - - - - - - - -
H. regalis 21 - - - - - - - - - - - 10 9 2 - - - - - -
H. wunduwala sp. nov. 10 10 - - - - - - - - - - - - - - - - - - -

SHELLEY ET AL.
352 · Zootaxa 5311 (3) © 2023 Magnolia Press
TABLE 8. Body depth at anal origin and pelvic origin, expressed as percentages of standard length (rounded to the nearest whole number), of northwest Australian species of
Hypseleotris.
Species Body Depth at Pelvic Origin Body Depth at Anal Origin
n13 14 15 16 17 18 19 20 21 22 23 24 11 12 13 14 15 16 17 18 19 20 21 22 23 24
H. aurea 8- - - - - - - 1 1 3 3 - - - - - - - - - - 1 2 2 3 -
H. barrawayi 30 - 1 - 1 2 10 8 6 1 - 1 - - - - - - 4 4 11 7 2 1 1 - -
H. compressa 18,19 - - - - 1 1 3 3 3 4 1 2 - - - - - - - - 4 2 2 5 3 3
H. ejuncida 12 - - - 1 4 7 - - - - - - - - - - - - 2 10 - - - - - -
H. garawudjirri sp. nov. 12 - 1 1 3 1 2 3 1 - - - - - - - 1 1 2 2 3 2 - 1 - - -
H. hutchinsi 10 1 5 4 - - - - - - - - - 2 1 1 3 3 - - - - - - - - -
H. kimberleyensis 12 - 3 3 5 1 - - - - - - - - 2 1 5 2 2 - - - - - - -
H. maranda sp. nov. 12 - - 5 4 2 1 - - - - - - - - - - 1 2 5 3 1 - - - - -
H. notata 21 1 - - - - - - - - - - - 1 1 - - - - - - - - - - -
H. regalis 20 - - 2 5 4 7 2 - - - - - - - - - 3 3 10 3 1 - - - -
H. wunduwala sp. nov. 10 - - 3 3 2 2 - - - - - - - - - - 1 7 1 1 - - - - - -

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 353
TABLE 9. Descriptive morphological characters for northwest Australian Hypseleotris species examined in this study.
Species Cheek scales Opercle scales Prepelvic
scales
Nape Scales
H. aurea Covered in scales Covered in scales Absent Present, covering entire nape
H. barrawayi Scales present posterior to
eye, dorsal corner
Covered in scales Absent Often present, at least along
midline, but may cover broader
area. May also be absent.
H. compressa Covered in scales Covered in scales Present Present, covering entire nape,
strongly embedded under skin
H. ejuncida Often absent, if present they
are posterior to eye, towards
dorsal edge
Covered in scales Absent Present, covering most of
nape, but bald patch exists
immediately posterior to eyes
H. garawudjirri
sp. nov.
Fringed with scales Naked or sparsely
scaled
Absent Often absent, when present
scattered in no particular pattern
H. hutchinsi Absent Absent Absent Absent
H. kimberleyensis Absent Absent Absent Absent
H. maranda sp.
nov.
Covered in scales Covered in scales Absent Often present covering entire
nape, or are absent
H. notata Absent Absent Absent Absent
H. regalis Often present, either in small
patches or covering cheek, but
may be absent
Covered in scales Present Covered in small scales
H. wunduwala
sp. nov.
Absent Covered except
dorsal edge
Absent Absent
Systematics
GENUS: HYPSELEOTRIS
Hypseleotris Gill, 1863
Type species. Hypseleotris Gill, 1863: 270 (type species: Eleotris cyprinoides Valenciennes 1837, by original
designation)
Etymology. The generic name, Hypseleotris is from the Greek hypsos, meaning ‘height’, and the generic name
Eleotris to which the type species was originally designated.
Description. Genus of Eleotridae with compressed head, body, mouth small, oblique, not extending beyond
middle of eye. Anterior nostril a simple pore above middle of upper lip, posterior nostril a simple pore above
anterodorsal margin of eye. Gill opening extending to below posterior end of preoperculum. Pectoral base narrow,
with dark pigment blotch on dorsal margin. Pelvic fins separate, each pelvic fin with one spine and five segmented
rays. Body scales large, ctenoid; predorsal area (nape) and cheek bare or variably scaled with cycloid or a mix of
cycloid and ctenoid scales. Head pores present or absent. Sensory papillae in transverse rows on side and/or top of
head. Teeth small, monomorphic, in several rows in both jaws, few may also be on vomer. First dorsal VI–IX; dorsal
fin insertion pattern 3-1221 or 3-12210; second dorsal I, 8–13; anal I, 9–13, pectoral rays 13–17. Anal fin originating
directly below origin of second dorsal fin. Anal fin elements 6–11 preceding the first haemal spine. Vertebrae 25–33.
Species in the northwest Australian clade not strongly sexually dimorphic, but males exhibit brightly coloured
dorsal anal fins that are most vibrant during spawning periods. Most species from the southeast Australian clade are
strongly sexually dimorphic, with mature males exhibiting heads enlarged in the dorsal/nape region and brightly
coloured dorsal anal fins during the breeding season.
This description combines our observations with those presented in Thacker et al. (2022a) and Keith and
Mennesson (2023).

SHELLEY ET AL.
354 · Zootaxa 5311 (3) © 2023 Magnolia Press
Key to the species of Hypseleotris from northwest Australia
Multiple characters are provided to distinguish between species and should be considered in combination. Based on
field-based experience with the group, the key is tentatively considered valid for both juveniles and adults, although
all specimens formally examined were adults. All characters can be measured or observed in the field. This key is
most accurate for freshly caught individuals or those fixed in 10 % formalin or diluted ethanol (e.g. 70–80%), as
specimens fixed in pure ethanol are typically too brittle to make accurate counts related to fin rays and spines.
Key to Western Australian Species of Hypseleotris
1a Preoperculum with 2–4 pores. Predorsal scales reach forward to above middle of eyes. Scales from under first dorsal fin to
upper attachment of opercular membrane ctenoid; nape scales often ctenoid. Second dorsal fin-rays modally 1,9. Found across
the Kimberley region and beyond .................................................. H. compressa (Krefft, 1864)
1b No preopercular pores. Predorsal scales either absent or reach forward to or behind middle of eye. Scales from under first dorsal
fin to upper attachment of opercular membrane cycloid; nape scales absent or, if present, cycloid. Second dorsal fin-rays usually
1 spine and 10 or more rays ............................................................................. 2
2a Longitudinal scale count 34–43, rarely < 37. Body relatively deep, with depth at anal fin origin > 20 % of SL. Found in the
Murchison and Gascoyne rivers of the Pilbara region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. aurea (Shipway, 1950)
2b Longitudinal scale count 24–34, rarely 35–36. Body more slender, depth at anal fin origin typically < 20 % of SL . . . . . . . . 3
3a Head with few or no sensory papilla on lateral surfaces and dorsal surfaces (see Figs. 3 and 10) . . . . . . . . . . . . . . . . . . . . . . . 4
3b Head with extensive network of sensory papilla on lateral and dorsal surfaces (see Fig. 7 for example) . . . . . . . . . . . . . . . . . . 5
4a Head with a few rows of sensory papilla on lateral and dorsal surface of head (see Fig. 3). Pectoral fin rays 15. Predorsal
extensively scaled, midline scales 14–21. Top of eye does not extend beyond dorsal surface of head. In life, basal one-third of
dorsal and anal fins tan to brown (sometimes with white spots on second dorsal fin) followed distally by a thin white stripe,
followed distally by a broader black stripe, then thin white band on fin margin (colours most vibrant in adult males). Found in
Garimbu Creek, Roe River, northwestern Kimberley Plateau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. maranda sp. nov.
4b Head with no sensory papilla on lateral surface of head (see Fig. 10). Pectoral fin rays typically 13–14. Predorsal naked. Top of
eye extends notably above dorsal surface of head. In life, dorsal and anal fins tan to brown for basal third, followed distally by
moderately narrow black band, then thin white band at fin margin (colours most vibrant in adult males). Found in Carson River,
King Edward River, northern Kimberley Plateau. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. wunduwala sp. nov.
5a Predorsal midline scales present in most or all cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
5b Predorsal midline scales absent in most or all cases .......................................................... 7
6a Longitudinal scale count 24–26. Body moderately slender, body depth at anal fin origin 16–20% of SL. Predorsal extensively
scaled over entire nape, midline scales 14–16. Prepelvic region scaled. Body without vertical bars, but with 15–20 chevron
marks on side. Pelvic fin origin under posterior opercular margin. In life, basal one-third of dorsal and anal fins dark brown to
black (sometimes with white spots on second dorsal fin) followed distally by a thin white stripe, followed distally by a broader
black stripe, then thin white band at fin margin (colours most vibrant in adult males). Found in the Prince Regent and Roe rivers,
northwest Kimberley Plateau .................................................. H. regalis Hoese & Allen, 1983
6b Longitudinal scale count 28–34. Predorsal naked or partly scaled, with naked patches on nape. Pre-pelvic region unscaled.
Body with faint vertical bars anteriorly. Pelvic fin origin under or behind pectoral fin insertion . . . . . . . . . . . . . . . . . . . . . . . 8
7a Adpressed first dorsal fin reaches and most often surpasses second dorsal fin origin. Lateral line absent. Second dorsal fin rays
modally 8. Found in the Mitchell River, northern Kimberley Plateau . . . . . . . . . . . . . . . . H. hutchinsi (Hoese & Allen, 1983)
7b Adpressed first dorsal fin reaches or just falls short of second dorsal fin origin. Lateral line present. Second dorsal fin rays
modally 9 or more .................................................................................... 9
8a Reverse transverse scales usually 8 or less. Dark spot on pectoral fin base covers whole base. Body depth at anal fin origin
17–18 % of SL. In life, second dorsal and anal fins tan to orange-brown with numerous white spots surrounded by dark pigment
forming wavy bands, and with narrow light clear to white fin margin (colours most vibrant in adult males). Found in the Prince
Regent River, northwest Kimberley Plateau . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . H. ejuncida Hoese & Allen, 1983
8b Reverse transverse scales typically 11 or more. Body depth at anal fin origin 16–22 % of SL. In life, dorsal and anal fins with
broad yellowish-grey band through center of fin or slightly below center, with narrow bluish white band at fin margin (colours
most vibrant in adult males). Found in the Katherine River, Daly River, Arnhem Land region . . . . H. barrawayi Larson, 2007
9a Longitudinal scale count 25–27. Body depth at anal origin 12–13 % SL. Vomerine teeth present. Found in the Drysdale River,
northern Kimberley Plateau ................................................... H. notata (Hoese & Allen, 1983)

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 355
9b Longitudinal scale count 29–32. Body depth at anal origin 13–18 % SL, rarely 13% SL. Vomerine teeth absent . . . . . . . . . . 10
10a Postdorsal scales 7–9 (modally 8). Vertebrae 25. Preoperculum naked. In life, second dorsal and anal fins transparent to dusky
brown with narrow white mid-lateral band and fin margin (colours most vibrant in adult males). Found in the upper Fitzroy
River, southern Kimberley Plateau ....................................... H. kimberleyensis Hoese & Allen, 1983
10b Postdorsal scales 8–11 (modally 9). Vertebrae 26. Preoperculum fringed with scales around outer 1/2, sometime naked. In life,
dorsal and anal fins with broad bright orange submarginal band (sometimes split into two bands in first dorsal), and narrow
dusky-white band at fin margin (colours most vibrant in adult males). Found in the Charnley, Calder, and Sale rivers, western
Kimberley Plateau ................................................................. H. garawudjirri sp. nov.
Species treatments
Hypseleotris maranda, new species
Figs. 1–6; Tables 1–10
urn:lsid:zoobank.org:act:059B1C5D-9260-4D91-8AE3-6E53BBE99C9F
Recommended standard name: Roe River Gudgeon
Name in Language: Maranda (Ngarinyin)
Hypseleotris sp. 1 ‘Garimbu Gudgeon’—Shelley et al. 2018a:186−187; Thacker et al. 2022a: 3, 4, 6−10, 15, 18, 20
Holotype (measured): WAM P.35282.001 (29.1 mm SL), Australia, Western Australia, the Kimberley region,
Garimbu Creek around 2 km upstream of the large waterfall near the confluence with the Roe River, Roe River
catchment, 15° 20’ 30.01’’ S, 125° 33’ 22.88’’ E. The holotype was obtained using a dip net while snorkeling by J.
Shelley and M. Le Feuvre, July 7, 2013.
FIGURE 3. Papillae pattern of Hypseleotris maranda sp. nov. from the lateral profile. No pores are present on the dorsal
surface of the head. The drawing is a composite based on several specimens. The patterns are considered accurate within reason,
but the number of papillae is not.

SHELLEY ET AL.
356 · Zootaxa 5311 (3) © 2023 Magnolia Press
Paratypes (measured): 8 specimens, 26.7–37.0 mm SL. Paratypes were collected in line with the holotype.
NMV A.31917-013, 014, 020, 022, 023, 024, 025 (7), NMV A.31985-001 (1), Australia, Western Australia, the
Kimberley region, Garimbu Creek around 2 km upstream of the large waterfall near the confluence with the Roe
River, Roe River catchment.
Non-type material (measured): 4 specimens, 32.1–34.4 mm SL, Non-type material were collected in line
with the holotype. NMV A.31917-001 (32.1 mm SL), Australia, Western Australia, the Kimberley region, Garimbu
Creek around 2 km upstream of the large waterfall near the confluence with the Roe River, Roe River catchment.
The other three specimens were lost and are thus not registered. Details are provided in the Comments section in
this species treatment.
Diagnosis: Hypseleotris maranda sp. nov. differs from its congeners by the following combination of characters:
First dorsal fin dusky grey to transparent, with dusky grey band in basal third, followed distally by light dusky to
transparent band extending to margin, second dorsal fin with broad dusky grey basal band and light dusky to
transparent band extending to margin, anal fin dusky grey, with narrow light dusky margin, head with distinctively
sparse pattern of sensory papilla on lateral and dorsal surface (see Fig. 3). Pectoral fin rays 15, vertebrae 26,
preopercle covered with small scales, post-dorsal scale count greater than 9, usually 9, lateral scale count 30–34,
usually 32, and predorsal extensively scaled with 14–21 midline scales, number inconsistent, and top of eye does
not extend beyond dorsal surface of head.
FIGURE 4. Hypseleotris maranda sp. nov. (A) the preserved holotype (WAM P.35282.001, 29.1 mm SL male), from Garimbu
Creek, Roe River catchment, Western Australia (photo credit: James Shelley); and (B) a freshly caught specimen from Garimbu
Creek with natural coloration (photo credit: James Shelley and Matthew Le Feuvre).
Description: Based on 13 specimens, 26.7–37.0 mm SL. Counts for holotype (Fig. 4a) indicated by asterisk,
and number of specimens with a given count in parentheses. First dorsal spines VI* (13); second dorsal rays I,9 (1),
I,10* (6), I,11 (5), I,12 (1); anal rays I,10 (5), I,11* (6), I,12 (1), I,13 (1); pectoral rays 15* (13); segmented caudal

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 357
rays 15* (6), in 8/7 pattern; branched caudal rays 11* (6) in 6/5 pattern; longitudinal scale count 29* (1), 30 (3), 31
(5), 32 (4); transverse scales backward 8* (1), 9 (4), 10 (5), 11 (2), 12 (1); predorsal scale count 14*–21; postdorsal
scale count 9* (7), 10 (3), 11 (2), 12 (1); gill rakers on outer face of first arch 3+10 (1), 3+12 (2), 4+9 (1), 4+10 (5),
4+12 (1), 4+13 (1) (Tables 2–7).
See Table 10 for summary of morphometric variation. Body slender and compressed, more so posteriorly;
body depth at anal origin 16.0 (15.0–17.5) % of SL (Table 8); body depth at pelvic fin origin 17.0 (15.3–18.8) % of
SL. Caudal peduncle long, length 26.0 (21.5–28.3) % of SL. Caudal peduncle depth 9.5 (8.8–10.8) % of SL. Head
somewhat compressed, forming rough triangle (apex dorsally) in cross-section, usually deeper than wide at posterior
preopercular margin, head length 30.1 (28.9–31.0) % of SL, depth at posterior preopercular margin 59.0 (51.5–64.3)
% of HL, width at posterior preopercular margin 51.3 (35.5–58.0) % of HL. Large males with somewhat convex
nape but none with ‘hump’. Eyes do not protrude above head profile. Mouth short and oblique, forming an angle of
about 45° with body axis; jaws ending anterior to eye, below posterior nostril. Upper jaw length 27.5 (23.8–30.8)
% of HL; lips narrow; lower lip fused to chin anteriorly, side of lip free; no mental frenum present. Anterior naris
at end of very short tube just above upper lip; posterior naris oval, close to anterodorsal margin of eye. Eye width
25.3 (21.1–31.5) % of HL. Interorbital broad, its width approximately equal to eye width, 16.4 (12.5–21.0) % of HL.
Snout short, rounded to almost square in dorsal vie
w, gently rounded to slightly pointed in side view, its length about
equal to eye width, 27.6 (24.3–29.8) % of HL. Gill opening moderate, extending forward to under posterior margin
of eye. Opercle squat, slightly taller than it is long. Gill rakers slender near angle of arch and becoming progressively
shorter (but still slender) and more widely spaced anteriorly, longest raker (below angle of arch) nearly as long as
gill filaments (which are short); rakers on inner face of first and other arches short and stubby. Tongue tip blunt to
slightly concave. Teeth small, sharp and curved; teeth in both jaws in three to five rows anteriorly, closely packed,
curving inward, teeth in innermost tooth row slightly larger than others; rows narrowing toward side of each jaw,
so that posterior half of jaw with only one or two rows of teeth present. No vomerine teeth. No head pores. Sensory
papillae in reduced transverse pattern (Fig. 3).
TABLE 10. Morphometric variation in Hypseleotris maranda sp. nov. (values are percentages of denominators in ratios,
except for SL).
Character Holotype Paratypes (n=12)
Mean Minimum Maximum S.D.
Standard length 28.2 32.4 26.7 37.0
Head length 8.4 9.8 8.1 11.3 1.2
Head depth 4.5 5.8 4.5 7.3 0.9
Head width 3.8 5.0 2.9 6.4 1.0
Body depth at anus 4.6 5.2 4.0 6.4 0.8
Body depth at pelvic origin 4.3 5.5 4.3 6.9 0.8
Body width 3.1 3.9 3.1 5.0 0.6
Caudal peduncle length 7.9 8.4 7.4 10.0 0.8
Caudal peduncle depth 3.1 3.1 2.5 3.8 0.4
Snout length 2.2 2.7 2.2 3.2 0.3
Eye width 2.7 2.5 2.0 2.9 0.3
Upper jaw length 2.0 2.7 2.0 3.3 0.4
Interorbital width 1.7 1.6 1.1 2.2 0.4
Pectoral length 5.1 6.4 5.1 8.6 1.1
Pelvic length 5.9 5.8 4.2 7.4 0.8
Caudal length 5.9 6.7 5.7 8.4 0.9
Adpressed first dorsal length 5.3 5.0 3.8 6.0 0.8
Scales on body reaching forward to above pectoral fin base or further forward on to side of nape to above
preopercle; most body scales ctenoid, with cycloid scales under first dorsal fin so that ctenoid scales form a wedge
along side of body to behind pectoral fin; scales posterior to below first dorsal fin larger than those anteriorly.

SHELLEY ET AL.
358 · Zootaxa 5311 (3) © 2023 Magnolia Press
Prepelvic region naked. Pectoral fin base with few small cycloid scales. Predorsal scales small, cycloid; scales
embedded or partly so, nearly always non-imbricate, scattered in groups in rough line along midline of nape and
above opercle and preopercle in no particular pattern. Scales on side of head cycloid, firmly embedded; opercle and
preopercle with many small cycloid scales nearly covering each; two to four of opercle scales larger. Belly scales
cover most of surface with embedded cycloid scales.
First dorsal fin low, rounded to somewhat rectangular in form, no spines elongate, fin falling short of or just
touching second dorsal fin spine when adpressed; adpressed first dorsal length 15.0 (14.0–18.8) % of SL, not
differing between males and females. Anterior-most second dorsal and anal rays longest, declining notably in length
towards posterior creating a convex margin and asymmetric rounded fin shape, fin bases short, posterior rays falling
well short of caudal fin base; anterior-most second dorsal and anal rays unbranched. Pectoral fin pointed, slender,
central rays longest, 19.7 (18.1–23.2) % of SL; upper and lowermost two rays unbranched. Pelvic fin length 18.0
(15.6–21.0) % of SL; pelvic fins slender, pointed, fifth rays longest, fin rays with one branch point; fins falling short
of anus. Caudal fin truncate to slightly rounded; caudal fin length 20.7 (18.8–22.6) % of SL.
Colouration in alcohol: Head plain brown, paler ventrally; dorsal margin of opercle same as head. Body
plain brown, lighter ventrally, with 10 very faint narrow brown bars crossing dorsum, first bar crossing nape above
opercle, bars reaching down to top of midlateral scales. Scales on body narrowly margined with dark brown, usually
giving finely reticulate appearance. On mid-side of body, row of short dark brown vertical bars running from
anterior of caudal peduncle to caudal fin base. Distinct blackish, dark brown spot on lower half of caudal fin base.
First dorsal fin transparent to dusky grey, with darker base and light dusky to transparent band extending to
margin. Second dorsal fin with broad dusky grey basal band and light dusky to transparent band extending to
margin. Anal fin plain dusky grey, with narrow light dusky margin. Caudal fin usually plain light dusky with 4–5
slightly oblique narrow diffuse bars crossing basal half of fin, bars may break up into rows of dusky spots. Pectoral
fin transparent, grey-black blotch at base. Pelvic fins translucent to faintly dusky (Fig 4a).
FIGURE 5. Distribution of Hypseleotris maranda sp. nov. (white circle), H. garawudjirri sp. nov. (blue squares), and H.
wunduwala sp. nov., based on all known records. The collection location of each holotype is marked with an asterisk. Rivers in
which Hypseleotris are found are named.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 359
FIGURE 6. Garimbu Creek showing (A) typical habitat occupied by Hypseleotris maranda sp. nov.; and (B) a significant
waterfall in the lower catchment, just upstream of the junction with the Roe River, above which the species was found (never
below). The species was most commonly found along the tussock grass lined edges of the stream in slow flowing sections.

SHELLEY ET AL.
360 · Zootaxa 5311 (3) © 2023 Magnolia Press
Live colouration: Head and body tan to brown, dusky-white on ventral surface of head and abdomen, scale
margins narrowly outlined with dark brown. Dorsal saddles and vertical bars along side of body dusky to dark
brown. Internal narrow pale golden-brown stripe running from eye nearly to midcaudal fin base, following top of
vertebral column. Short black vertical blotch on lower base of caudal fin. Grey-black bar on pectoral base. Head
under eye pale to translucent. Scattered spots or blotches of iridescent gold pigment may show through body wall
of upper abdomen and side of head. Lips dusky grey to light brown. Pupil surrounded by rim of pale gold, most of
eye brownish-gold.
First dorsal fin dusky grey to transparent, with dusky grey band in basal third, followed distally by light dusky
to transparent band extending to margin. Second dorsal fin with broad dusky grey basal band and light dusky to
transparent band extending to margin. Anal fin dusky grey, with narrow light dusky margin. Caudal fin usually light
dusky with 4–5 slightly oblique, narrow, brown, diffuse bars crossing basal 3/4 of fin. Pectoral fin transparent, grey-
black bar at base. Pelvic fins translucent to faintly dusky (Fig. 4b).
Distribution: A narrow range Kimberley endemic only known from Garimbu Creek, a major tributary of
the Roe River, that drains northward off the Kimberley Plateau (Fig. 5). The species was observed over a 2 km
reach directly upstream of a substantial waterfall near the junction of Garimbu Creek and the Roe River (Fig. 6b).
However, it was not found directly below the falls. The species may be restricted to Garimbu Creek, but the precise
extent of the species’ range is unclear due to limited sampling in the area.
Ecology: Occurs in shallow to moderate depths, over rocky substrates, often near crevices and grass lined
stream edges (Fig. 6a). Little is known about its ecology, but the species is expected to be similar to others in the
genus that are oviparous benthic spawners with a diet consisting mainly of insect larvae and microcrustaceans.
Etymology: The specific name maranda, to be treated as a noun in apposition, is the Ngarinyin language word
for the catchment where the species is found (English name—Roe River). It is pronounced: Mar-un-da. Currently,
this is the only catchment in which the species is found. The name was provided by Ngarinyin elder Matthew Martin
with the assistance of Lloyd Nulgit. The Wilinggin Aboriginal Corporation represents all the Ngarinyin Traditional
Owners and their interests.
Comments: Three of the specimens used in data collected were left out to dry and were beyond recognition.
They were subsequently disposed of and were not registered. One specimen, NMV A.31917-001, had a tissue
sample taken from its side and was only used for meristic counts. Due to the poor condition of this specimen, it was
not included with the type material.
Hypseleotris garawudjirri, new species
Figs. 1–2, 5, 7–9; Tables 1–9, 11
urn:lsid:zoobank.org:act:0FAE9CB8-604C-462E-996C-3A752A8AACD7
Recommended standard name: Bachsten Gudgeon
Name in Language: Garawud jirri (Ngarinyin)
Hypseleotris kimberleyensis non Hoese & Allen 1983: 252; Allen & Leggett 1990:540 (in part); Allen et al. 2002:296 (in part);
Thacker & Unmack 2005: 2 (in part); Morgan et al. 2011:18, Fig. 43 (in part); Pusey et al. 2011:82 (in part); Morgan et al.
2014:18 (in part); Hoese 2018: online (in part)
Hypseleotris sp. 2 ‘Bachsten Gudgeon’− Shelley et al. 2018a:186−187; Thacker et al. 2022a: 3, 4, 6−10, 15, 18, 20
Holotype (measured): WAM P.35283.001 (35.9 mm SL), Australia, Western Australia, the Kimberley region,
Bachsten Creek at Bachsten Track crossing, Calder River catchment, 16° 03’ 17.73’’ S, 125° 12’ 49.58’’ E. The
holotype was obtained using a backpack electrofisher J. J. Shelley and M. C. Le Feuvre, August 1, 2014.
P
aratypes (measured): 11 specimens, 29.1–39.6 mm SL. NMV A.31740-018 to 20 (3) and NMV A.31740-
002 (1), 30.4–39.6 mm SL, Australia, Western Australia, the Kimberley region, Bachsten Creek at Bachsten Track
crossing, Calder River catchment, 16° 03’ 17.73’’ S, 125° 12’ 49.58’’ E, obtained with backpack electrofisher by J. J.
Shelley and M. C. Le Feuvre, August 1, 2014; NMV A.31766-007 and -008 (2), 29.1–36.8 mm SL, Australia, Western
Australia, the Kimberley region, Bachsten Creek at Bachsten Creek Bush Camp, Calder River catchment, 15° 59’
22.95’’ S, 125° 19’ 47.85’’ E, obtained with dip nets from the bank by J. J. Shelley and M. C. Le Feuvre, August 1,
2014; NMV A.31453-005 to -009 (5), 31.3–36.3 mm SL, Australia, Western Australia, the Kimberley region, Charnley
River around 2 km upstream of the confluence with Pearson Creek, Charnley River catchment, 16° 12’ 50.62” S, 125°
11’ 18.32” E obtained with backpack electrofisher by J. J. Shelley and M. C. Le Feuvre, August 4, 2014.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 361
FIGURE 7. Papillae pattern of Hypseleotris garawudjirri sp. nov. from (A) the lateral profile, and (B) the dorsal profile. The
drawing is a composite based on several specimens. The patterns are considered accurate within reason, but the number of
papillae is not.

SHELLEY ET AL.
362 · Zootaxa 5311 (3) © 2023 Magnolia Press
Non-type material (photographs examined): 2 specimens, 29.1–31.2 mm SL. At the time of publication these
new collections were not registered but are held at the WAM. Australia, Western Australia, the Kimberley region,
Upper Sale River gorge and swamp, Sale River catchment, 15° 55’ 37.88’’ S, 125° 7’ 16.39’’ E. Obtained using a
backpack electrofisher M. P. Hammer and G. I. Moore, July 25, 2022.
Diagnosis: Hypseleotris garawudjirri sp. nov. differs from its congeners by the following combination of
characters: dorsal and anal fins with broad bright orange submarginal band (sometimes split into two bands in first
dorsal), and narrow dusky-white band at fin margin, head with distinctively sparse pattern of sensory papilla on
lateral and dorsal surface (see Fig. 7). Pectoral fin rays 14–15 (usually 15), vertebrae 26, preopercle fringed with
scales, sometimes naked, postdorsal scale count 8–11, usually 9, longitudinal scale count 30–32, usually 30, and
predorsal typically naked, rarely scaled, and top of eye does not extend beyond dorsal surface of head.
Description: Based on 12 specimens, 29.1–39.6 mm SL. Counts for holotype (Fig. 8a) indicated by asterisk,
and number of specimens with a given count in parentheses. First dorsal spines VI* (9), VII (3); second dorsal rays
I,10* (5), I,11 (6), I,13 (1); anal rays I,10 (3), I,11* (8), I,12 (1); pectoral rays 14 (5), 15* (7); segmented caudal rays
15* (6), in 8/7 pattern; branched caudal rays 6/5* (7); longitudinal scale count 30 (6), 31 (2), 32* (4); transverse
scales backward 9* (7), 10 (5); predorsal scale count 0* (10), 14 (1), 15 (1); Postdorsal scale count 8 (1), 9* (9), 10
(1), 11 (1); gill rakers on outer face of first arch 2+9 (1), 2+10 (5), 3+8 (1), 3+10 (2), 3+11 (1) (Tables 2–7).
FIGURE 8. Hypseleotris garawudjirri sp. nov. (A) the preserved holotype (WAM P.35283.001, 35.9 mm SL male), from
Bachsten Creek, Calder River catchment, Western Australia (photo credit: James Shelley); and (B) a freshly caught specimen
from Bachsten Creek with natural coloration (photo credit: James Shelley and Matthew Le Feuvre).
See Table 11 for summary of morphometric variation. Body slender and compressed, more so posteriorly; body
depth at anal origin 17.5 (14.1–21.2) % of SL (Table 8); body depth at pelvic fin origin 17.0 (14.3–19.3) % of SL.
Caudal peduncle long, length 27.9 (24.7–30.3) % of SL. Caudal peduncle depth 10.1 (7.6–11.8) % of SL. Head
somewhat compressed, forming rough triangle (apex dorsally) in cross-section, usually deeper than wide at posterior
preopercular margin, head length 27.0 (24.6–29.3) % of SL, depth at posterior preopercular margin 61.2 (54.8–69.1)
% of HL, width at posterior preopercular margin 51.6 (43.2–60.8) % of HL. Large males with somewhat convex

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 363
nape but none with ‘hump’. Eyes do not protrude above head profile. Mouth short and oblique, forming an angle of
about 45° with body axis; jaws ending anterior to eye, below posterior nostril. Upper jaw length 26.4 (22.0–31.9) %
of HL; lips narrow; lower lip fused to chin anteriorly, side of lip free; no mental frenum present. Anterior naris at end
of very short tube just above upper lip; posterior naris almost triangular, with rounded corners, close to anterodorsal
margin of eye. Eye width 27.3 (25.8–29.6% of HL. Interorbital broad, its width approximately equal to eye width,
21.6 (16.3–26.9) % of HL. Snout short, rounded to almost square in dorsal view, gently rounded to slightly pointed in
side view, its length about equal to eye width, 23.7 (17.7–28.0) % of HL. Gill opening moderate, extending forward
to under posterior margin of preopercle or slightly further forward (but not reaching eye). Opercle squat, slightly
taller than it is long. Gill rakers slender near angle of arch and becoming progressively shorter (but still slender)
and more widely spaced anteriorly, longest raker (below angle of arch) nearly as long as gill filaments (which are
short); rakers on inner face of first and other arches short and stubby. Tongue tip blunt to slightly concave. Teeth
small, sharp and curved; teeth in both jaws in three to four rows anteriorly, closely packed, curving inward, teeth
in innermost tooth row slightly larger than others; rows narrowing toward side of each jaw. No vomerine teeth. No
head pores. Sensory papillae in reduced transverse pattern (Fig. 7).
TABLE 11. Morphometric variation in Hypseleotris garawudjirri sp. nov. (values are percentages of denominators in
ratios, except for SL).
Character Holotype Paratypes (n=12)
Mean Minimum Maximum S.D.
Standard length 35.9 33.8 29.1 39.6
Head length 10.0 9.2 8.0 10.9 1.1
Head depth 6.0 5.6 4.6 7.3 1.0
Head width 6.1 4.8 3.5 6.1 0.9
Body depth at anus 6.7 5.9 4.5 7.7 1.1
Body depth at pelvic origin 6.3 5.8 4.6 7.4 0.9
Body width 4.7 4.4 3.2 6.0 0.9
Caudal peduncle length 10.0 9.4 7.9 10.2 0.7
Caudal peduncle depth 4.1 3.4 2.4 4.7 0.7
Snout length 2.4 2.2 1.4 2.8 0.4
Eye width 2.7 2.5 2.1 3.1 0.3
Upper jaw length 3.2 2.4 1.8 3.2 0.6
Interorbital width 2.5 2.0 1.3 2.9 0.5
Pectoral length 6.6 6.3 4.5 8.1 1.0
Pelvic length 5.7 6.2 4.9 7.8 1.1
Caudal length 7.9 7.3 6.3 8.4 0.8
Adpressed first dorsal length 6.7 5.7 4.2 7.2 1.1
Scales on body reaching forward to above pectoral fin base or further forward on to side of nape to above
preopercle; most body scales ctenoid, with cycloid scales under first dorsal fin so that ctenoid scales form a wedge
along side of body to behind pectoral fin; scales posterior to below first dorsal fin larger than those anteriorly.
Prepelvic region naked. Pectoral fin base with few small cycloid scales. Predorsal scales small, cycloid, variably
present (absent in eight); scales embedded or partly so, nearly always non-imbricate, cover entire nape and above
opercle and preopercle when present. Scales on side of head cycloid, firmly embedded; opercle and preopercle with
patch of cycloid scales nearly covering each. Two to four of scales on opercle larger. Belly scales either completely
absent (in seven) or with few embedded cycloid scales near anus (in four) (Table 9).

SHELLEY ET AL.
364 · Zootaxa 5311 (3) © 2023 Magnolia Press
FIGURE 9. A stretch of (A) the Calder River, and (B) the Charnley River where the Hypseleotris garawudjirri sp. nov. was
found. The species showed an affinity for edge habitat in slow flowing stream sections such as these.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 365
First dorsal fin low, rounded to somewhat rectangular in form, no spines elongate, fin falling short of or just
touching second dorsal fin spine when adpressed; adpressed first dorsal length 16.7 (14.4–19.3) % of SL, not
differing between males and females. Anterior-most second dorsal and anal rays longest but not greatly so, steadily
declining in length towards posterior creating a posterior leaning truncate fin shape, fin bases short, posterior rays
falling well short of caudal fin base; anterior-most second dorsal and anal rays unbranched. Pectoral fin pointed,
slender, central rays longest, 18.5 (14.3–22.0) % of SL; upper and lowermost two rays unbranched. Pelvic fin length
18.2 (15.1–21.1) % of SL; pelvic fins slender, pointed, fifth rays longest, fin rays with one branch point; fins falling
short of anus. Caudal fin truncate to slightly rounded; caudal fin length 21.6 (19.6–23.1) % of SL.
Colouration in alcohol: Head dark to plain brown, paler ventrally; dorsal margin of opercle dark. Body plain
brown, lighter ventrally, with 10−11 very faint narrow brown bars crossing dorsum, becoming fainter posteriorly,
first bar crossing nape above pectoral fin base, bars reaching down to midlateral line. Scales on body narrowly
margined with dark brown, usually giving finely reticulate appearance. On mid-side of body, covering and extending
just below midlateral line, row of short dark brown vertical bars running from abdomen to caudal fin base, bars may
coalesce with brown bars crossing dorsum or be indiscernible. Distinct blackish, dark brown spot on lower half of
caudal fin base.
First dorsal fin transparent to dusky, with dusky base, a transparent to faintly dusky band above this, a dusky
submarginal band of about same width as transparent band above this, and a light dusky fin margin. Second dorsal
fin with broad dusky grey basal band, a transparent submarginal band, and a light, dusky margin. Anal fin plain
dusky grey, with narrow light dusky margin. Caudal fin usually light dusky, several rows of dusky spots crossing
basal half of fin. Pectoral fin transparent, dark blotch at base. Pelvic fins translucent to faintly dusky (Fig. 8a).
Live colouration: Head and body tan to brown, slightly lighter tan-brown on ventral surface of head and
abdomen, scale margins narrowly outlined with dark brown. Dorsal saddles and vertical bars alongside of body
dusky to dark brown. Series of small dark blotches/bars below midline black. Internal narrow pale golden-brown
stripe running from eye nearly to midcaudal fin base, following top of vertebral column. Short black vertical blotch
on lower base of caudal fin. Grey-black bar on pectoral base. Head above eye brown to dark brown. Head under
eye light brown to translucent. Scattered spots or blotches of iridescent gold pigment may show through body wall
of upper abdomen and side of head. Lips dusky grey to light brown. Pupil surrounded by rim of pale gold, most of
eye brownish-gold.
Dorsal and anal fins dusky grey to transparent basally, sometimes with whitish spots, followed distally by broad
bright orange submarginal band (sometimes split into two bands in first dorsal), and narrow dusky-white band at fin
margin. Caudal fin with light orange to dusky in middle rays, bright orange outer rays for ¾ of ray length, the light
dusky fin margin. Covered in white spots that may form 3–4 wavy, slightly convex bands (Fig. 8b).
Distribution: A relatively widespread Kimberley endemic, known from the neighboring Charnley, Calder and
Sale rivers that drain westward off the Kimberley Plateau. Sampling is fairly limited in these catchments, but the
species appears to be widespread within them (Fig. 5). Interestingly, the species appears to be absent or at least rare
in the Isdell River, which would likely connect with the Charnley and Calder rivers via Walcott Inlet during large
floods that lower the salinity of the inlet. These catchments otherwise share a very similar fish community.
Ecology: Prefers slow flowing deep-water habitat over sandy to rocky substrates, often near crevices, around
aquatic plants or near woody debris (Fig. 9a and 9b). Little is known about its ecology, but the species is expected
to be similar to others in the genus that are oviparous benthic spawners with a diet consisting mainly of insect larvae
and microcrustaceans. Breeding was observed during the dry season, but the range of size classes observed indicates
that they likely breed several times throughout the year.
Etymology: The specific name garawudjirri is a combination of the words Garawud Jirri, that mean ‘light/little
fish–floating around’ in the Ngarinyin language, to be treated as a noun in apposition. It is pronounced: Gaarr—Arrd
Jiddy. The name was provided by Ngarinyin elder Matthew Martin in consultation with the Wilinggin Aboriginal
Corporation. The Wilinggin Aboriginal Corporation represents all the Ngarinyin Traditional Owners and their
interests.
Comments: A total of 15 specimens of Hypseleotris were recently recorded in the Sale River catchment as part
of the Wilinggin-West Kimberley Bush Blitz expedition. The collectors, fish taxonomists Michael P. Hammer (NTM)
and Glen I. Moore (WAM), identified the specimens as H. sp. 2 in the field and provided photographs to the authors
who agreed with the identification bases on observable characters and the proximity to previously known populations.
As such, these specimens are included as a population of H. garawudjirri but were not formally examined.

SHELLEY ET AL.
366 · Zootaxa 5311 (3) © 2023 Magnolia Press
Hypseleotris wunduwala, new species
Figs. 1–2, 5, 10–12; Tables 1–9, 12
urn:lsid:zoobank.org:act:485C81EF-2E0E-47F6-B46F-0C7FED511ABB
Recommended standard name: Carson River Gudgeon
Name in language: Wunduwala (Belaa)
Hypseleotris ejuncida non Hoese and Allen 1983:255; Morgan et al. 2011:18, Fig. 42 (in part); Morgan et al. 2014:18 (in part);
Morgan et al. 2006:51; Morgan et al. 2010: 355−356, 362; Pusey et al. 2011:82 (in part).
Hypseleotris sp. 3 ‘King Edward Gudgeon’ − Shelley et al. 2018a:184−185; Thacker et al. 2022a: 3, 4, 6−10, 15, 18, 20; Hoese
2018: online (in part).
Holotype (measured): WAM P.35284.001 (20.3 mm SL), Australia, Western Australia, the Kimberley region, Mool
Mool (or Mur Mur) Lagoon near the junction of the Kalumburu Road and Carson River Road, 14° 27’ 14.15’’ S,
126° 40’ 59.09’’ E. The holotype was obtained using a seine net by J. Shelley and M. Le Feuvre, July 9, 2014.
Paratypes (measured): 9 specimens, 18.5–22.2 mm SL. Paratypes were collected in line with the holotype.
NMV A.31777-027 to 035 (9), Australia, Western Australia, the Kimberley region, Mool Mool (or Mur Mur) Lagoon
near the junction of the Kalumburu Road and Carson River Road.
Diagnosis: Hypseleotris wunduwala sp. nov. differs from its congeners by the following combination of
characters: Dorsal and anal fins tan to brown for basal third, followed distally by moderately narrow black band,
then thin white band at fin margin, head with no sensory papilla on lateral surface (see Fig. 10). Pectoral fin rays
13–15 (usually 14), vertebrae 26, preopercle naked, postdorsal scale count 8–10, usually 8, longitudinal scale count
28–33, usually 30, predorsal naked, and top of eye extends notably beyond dorsal surface of head.
Description: Based on 10 specimens, 18.5–22.2 mm SL. Counts for holotype (Fig. 11a) indicated by asterisk,
and number of specimens with a given count in parentheses. First dorsal spines VI* (10); second dorsal rays I,9 (2),
I,10* (8); anal rays I,10* (8); pectoral rays 13* (3), 14 (6), 15 (1); segmented caudal rays 15* (6), in 8/7 pattern;
branched caudal rays 11 (6) in 6/4 pattern; longitudinal scale count 28 (1), 29 (2), 30* (4), 31 (2), 33 (1); transverse
scales backward 8 (2), 9 (3), 10* (5); no predorsal scales; Postdorsal scale count 8 (6), 9* (3), 10 (1); gill rakers on
outer face of first arch 2+8 (2), 2+9 (2), 2+10 (1), 2+11 (1), 3+9 (2), 3+10 (1), 4+10 (1) (Tables 2–7).
See Table 12 for summary of morphometric variation. Body slender and compressed, more so posteriorly;
body depth at anal origin 16.2 (14.7–17.7) % of SL (Table 8); body depth at pelvic fin origin 16.2 (15.2–17.9) %
of SL. Caudal peduncle long, length 26.0 (23.6–29.6) % of SL. Caudal peduncle depth 9.7 (7.6–10.9) % of SL.
Head somewhat compressed, forming rough triangle (apex dorsally) in cross-section, usually deeper than wide
at posterior preopercular margin, head length 28.5 (27.1–30.2) % of SL, depth at posterior preopercular margin
55.1 (51.5–57.7) % of HL, width at posterior preopercular margin 48.1 (45.1–53.2) % of HL. Males and females
with essentially strait head profile, only very slightly convex. Eyes protrude above head profile. Mouth short and
oblique, forming an angle of about 45° with body axis; jaws ending anterior to eye, below posterior nostril. Upper
jaw length 25.7 (21.7–28.2) % of HL; lips narrow; lower lip fused to chin anteriorly, side of lip free; no mental
frenum present. Anterior naris at end of very short tube just above upper lip; posterior naris almost triangular with
rounded corners, close to anterodorsal margin of eye. Eye width 33.1 (30.5–36.0) % of HL. Interorbital broad,
its width approximately equal to eye width, 14.9 (12.0–18.2) % of HL. Snout short, rounded to almost square in
dorsal view, quite pointed with rounded tip in side view, its length about equal to eye width, 20.6 (16.1–23.7) % of
HL. Gill opening moderate, extending forward to under posterior margin of preopercle or slightly further forward
(but not reaching eye). Opercle slender, longer than it is tall. Gill rakers slender near angle of arch and becoming
progressively shorter (but still slender) and more widely spaced anteriorly, longest raker (below angle of arch) nearly
as long as gill filaments (which are short); rakers on inner face of first and other arches short and stubby. Tongue tip
blunt to slightly concave, may be slightly folded, giving tridentate appearance. Teeth small, sharp and curved; teeth
in both jaws in three to five rows anteriorly, closely packed, curving inward, teeth in innermost tooth row slightly
larger than others; rows narrowing toward side of each jaw, so that posterior half of jaw with only one or two rows
of teeth present. No vomerine teeth. No head pores. Sensory papillae in reduced transverse pattern (Fig. 10).

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 367
FIGURE 10. Papillae pattern of Hypseleotris wunduwala sp. nov. from (A) the lateral profile, and (B) the dorsal profile. The
drawing is a composite based on several specimens. The patterns are considered accurate within reason, but the number of
papillae is not.

SHELLEY ET AL.
368 · Zootaxa 5311 (3) © 2023 Magnolia Press
TABLE 12. Morphometric variation in Hypseleotris wunduwala sp. nov. (values are percentages of denominators in
ratios, except for SL).
Character Holotype Paratypes (n=10)
Mean Minimum Maximum S.D.
Standard Length 20.3 20.6 18.5 22.2
Head length 5.6 5.9 5.2 6.3 0.4
Head depth 3.2 3.2 3.0 3.5 0.1
Head width 2.6 2.8 2.5 3.1 0.2
Body depth at anus 3.1 3.3 2.9 3.7 0.3
Body depth at pelvic origin 3.3 3.3 3.0 3.7 0.2
Body width 2.3 2.4 2.3 2.6 0.1
Caudal peduncle length 6.0 5.4 5.0 6.0 0.3
Caudal peduncle depth 2.2 2.0 1.7 2.3 0.2
Snout length 1.1 1.2 1.0 1.4 0.2
Eye width 1.9 1.9 1.8 2.1 0.1
Upper jaw length 1.5 1.5 1.4 1.7 0.1
Interorbital width 1.0 0.9 0.7 1.1 0.1
Pectoral length 3.6 4.2 3.4 4.9 0.6
Pelvic length 4.0 4.1 3.7 4.8 0.4
Caudal length 4.8 5.1 4.7 5.6 0.3
Adpressed first dorsal length 3.6 3.8 3.5 4.1 0.2
Scales on body reaching forward to above pectoral fin base or slightly further forward to posterior edge of
operculum; most body scales ctenoid, with cycloid scales under first dorsal fin so that ctenoid scales form a wedge
along side of body to behind pectoral fin; scales posterior to below first dorsal fin larger than those anteriorly.
Prepelvic region naked. Pectoral fin base with few small cycloid scales. Predorsal and sides of head scaleless.
Opercle with patch of cycloid scales nearly covering, two to four of them larger; preopercle naked. Belly naked and
unpigmented, transparent (Table 9).
First dorsal fin low, rounded to somewhat rectangular in form, no spines elongate, fin falling short of or just
touching second dorsal fin spine when adpressed; adpressed first dorsal length 18.3 (17.1–19.7) % of SL, not
differing between males and females. Anterior-most second dorsal and anal rays longest, declining notably in length
towards posterior creating a convex margin and asymmetric, rounded fin shape, fin bases short, posterior rays
falling well short of caudal fin base; anterior-most second dorsal and anal rays unbranched. Pectoral fin pointed,
slender, central rays longest, 20.3 (17.4–24.3) % of SL; upper and lowermost two rays unbranched. Pelvic fin length
20.1 (16.5–23.0) % of SL; pelvic fins slender, pointed, fifth rays longest, fin rays with one branch point; fins falling
short of anus. Caudal fin truncate to slightly rounded; caudal fin length 24.9 (22.7–26.2) % of SL.
Colouration in alcohol: Head light brown to tan, paler ventrally; opercle light dusky to transparent. Body light
brown, slightly lighter ventrally, with 9−10 very faint narrow brown bars crossing dorsum, first bar crossing nape
above opercle, bars reaching down on to side of abdomen anteriorly, bars progressively becoming much shorter
posteriorly. Scales on body narrowly margined with dark brown, usually giving finely reticulate appearance. On
mid-side of body, row of short dark brown X-shaped blotches running from anterior of first dorsal fin to caudal fin
base, blotches may overlap with brown bars crossing dorsum. Distinct blackish, dark brown spot on lower half of
caudal fin base.
First and second dorsal fins and anal fin light dusky to transparent. Caudal fin light dusky to transparent with 4
slightly oblique narrow diffuse bars crossing basal half of fin, bars may break up into rows of dusky spots. Pectoral
fin transparent. Pelvic fins transparent to faintly dusky (Fig. 11a).

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 369
FIGURE 11. Hypseleotris wunduwala sp. nov. (A) the preserved holotype (WAM P.35284.001, 20.3 mm SL male), from Mool
Mool (or Mur Mur) Lagoon, King Edward River catchment, Western Australia (photo credit: James Shelley); and (B) a freshly
caught specimen from Mool Mool (or Mur Mur) Lagoon with natural coloration (photo credit: James Shelley and Matthew Le
Feuvre).
Live colouration: Head and body tan to brown, dusky-white on ventral surface of head and abdomen, scale
margins narrowly outlined with dark brown. Dorsal saddles and vertical bars along side of body dark brown to
black. Series of black X-shaped markings crossing midline. Internal narrow pale golden-green to golden-brown
stripe running from eye nearly to midcaudal fin base, following top of vertebral column. Short black vertical blotch
on lower base of caudal fin. Grey-black bar on pectoral base. Head above eye brown to dark brown with black
speckling. Head under eye dusky white to translucent. Scattered spots or blotches of iridescent gold pigment may
show through body wall of upper abdomen and side of head. Lips dusky grey to brown to dark brown. Pupil
surrounded by rim of pale gold, most of eye brownish-gold.
Dorsal and anal fins tan to brown for basal third, followed distally by moderately narrow black band, then thin
white band at fin margin. Caudal fin usually dusky brown for basal ¾, distal ¼ dusky-white to transparent, 3–4
slightly oblique, narrow, brown, diffuse bars crossing basal 3/4 of fin (Fig. 11b).
Distribution. A narrow range Kimberley endemic know only from three sites in the Carson River, a tributary
of the King Edward River that flows northward off the Kimberley Plateau. One of these sites was a small (~6,250
m2) floodplain lagoon (billabong), while the other two sites were closely situated in the Carson River mainstream.
The King Edward River and Carson River catchment have been the focus of considerable sampling effort, although
much of the lower King Edward River mainstream, between King Edward Falls and Kalumburu, has received little
sampling effort due to difficulty of access. It appears that the species is genuinely restricted to the lower Carson
River but may still be shown to occupy the lower King Edward River (Fig. 5).

SHELLEY ET AL.
370 · Zootaxa 5311 (3) © 2023 Magnolia Press
FIGURE 12. Carson River showing (A) Mool Mool (or Mur Mur) Lagoon, off Carson River, King Edward River catchment,
the type location of Hypseleotris wunduwala sp. nov.; and (B) typical riverine habitat in the lower Carson river where the species
is also found. The species was most common amongst pandanus roots and over vegetation at the edge of the waterbody.

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 371
Ecology. Prefers slow-flowing deep-water habitat over sandstone and mixed rocky substrates, often near
crevices, around aquatic plants, or near woody debris (Fig. 12a and 12b). The species was found in highest
abundance in a floodplain lagoon (Fig. 12a). Individuals orientate themselves almost vertically in the water column
close to cover. Little is known of its ecology, but the species is expected to be similar to others in the genus that
are oviparous benthic spawners with a diet consisting mainly of insect larvae and microcrustaceans. Breeding was
observed during the dry season in individuals as little as 20 mm SL. Maturation at this small size indicates that
they are truly diminutive relative to their congeners, rather than their small size simply being an artifact of limited
sampling or sampling bias.
Etymology: The specific name wunduwala, to be treated as a noun in apposition, is the ‘bush name’ of Dolores
Cheinmora (deceased) who, alongside her daughter Agnes Charles, was a part of the fish survey during which the
species was first collected (Morgan et al. 2009). It is pronounced: Wundu—wala. Dolores is from the Kwini people
and spoke the Belaa language. She also authored a book on the ethno-ecology of Kwini people (Cheinmora et al.
2017). Her country is the King George River area. Her family who chose the specific epithet, also suggested the
common name be the Carson River Gudgeon, which reflects the current day name for the catchment in which the
species is found.
Discussion
Here we describe three morphologically and genetically distinct taxa from the remote Kimberley region in northwest
Australia. Furthermore, we conclude that Kimberleyeleotris is a synonym of Hypseleotris, based on a comprehensive
analysis of new genetic and morphological evidence. Together, the new data extends the description for the genus
Hypseleotris and provides important taxonomic resolution for the northwest clade. The number of species within
the northwest Australian clade of Hypseleotris is now 11, with eight of these being endemic to the Kimberley
region, one being endemic to each of the Pilbara and Arnhem Plateau (Jawoyn Country) regions, and H. compressa
being widespread across Australia. The genus as a whole contains 17 valid species. The presence and number of
scales across the head and body, the pattern of sensory papillae on the head, fin ray counts, dorsal and anal fin
colouration (particularly in breeding males), and body depth, can be used to distinguish between the species within
the northwest Australia lineage.
Two of the newly described species are narrow-range endemics, with H. maranda and H. wunduwala having
extremely small distributions within sub-catchments of the Roe and King Edward rivers respectfully. This is not
uncommon in the region where at least a further 25 fish species are restricted to one or two catchments or sub-
catchments (Shelley et al. 2018a). The three species described here are considered well protected from most regional
anthropogenic disturbances due to their remoteness and given that there is low-impact land use in their respective
ranges. Hypseleotris maranda occurs entirely within the Prince Regent National Park and H. garawudjirri occurs
predominantly within the Prince Regent National Park and Charnley River-Artesian Range Wildlife Sanctuary.
Hypseleotris wunduwala occurs within the Carson River postural lease, managed by the Kalumburu Aboriginal
Corporation. However, the station runs relatively few head of free-range cattle to supply the local community
in the town of Kalumburu and the threat of these activities to freshwater environments is considered low. That
said, the species seems to be most concentrated in the type locality, Mool Mool Lagoon, a very small (~6,250 m2)
floodplain lagoon. The lagoon is not fenced, and cattle access could impact on the quality of this small, isolated
habitat especially during dry periods. The potential for accelerated climate change to impact on each of these species
is difficult to determine in the absence of detailed ecological information, although their restricted ranges would
suggest narrow environmental tolerances and, consequently, high vulnerability to environmental change (Le Feuvre
et al. 2021).
The members of the northwest Hypseleotris clade predominantly diverged in the late Miocene and Pleistocene
(Fig. 1), a time when Australia was becoming much cooler and drier and freshwater ecosystems were experiencing
dramatic environmental change (Martin 2006; McGowan et al. 2004). Phylogenetic evidence suggests that the
Pilbara region, the Kimberley Plateau and Arnhem Plateau have provided mesic refuges from overall aridity, likely
as high elevation and topographic complexity acted to generate orographical rainfall, provide shade, and maintain
suitable surface water habitat (Pepper et al. 2011b; a; Shelley et al. 2019). However, it is expected that several fish
species would have experienced population contraction and extinction in response to the changing environment

SHELLEY ET AL.
372 · Zootaxa 5311 (3) © 2023 Magnolia Press
(Shelley et al. 2018b). Further to this point, dispersal between and even within catchments in the Kimberley region
has been shown to be extremely rare in a range of fish species due to the rugged nature of the landscape which
physically restricts fish movement, except during extreme climatic events (Shelley et al. 2020, 2022). This would
have greatly limited opportunities for population expansion following a contraction. It also seems likely that
biological differences between species, such as dispersal syndromes and ecological niches, have dictated the ability
for a species to recolonise following population extinction or contraction to some degree (Shelley et al. 2022). It
may be that the more range restricted Hypseleotris species are among those that underwent significant population
contraction during the aridification of Australia. Furthermore, the restrictive nature of their home catchments, and
their respective biological traits, have limited their ability to subsequently expand their ranges.
The findings of this study are particularly important for future conservation strategies in the region as they
highlight the need for planning to occur at small spatial scales that reflect the small distribution of many of the
region’s fish fauna. Finally, despite broadscale sampling across the Kimberley Plateau, entire river systems are
only accessible via helicopter and remain unsampled. The narrow ranges of these species further suggest that future
sampling in these areas will uncover further important information for biodiversity conservation.
Acknowledgements
We acknowledge the Traditional Owners of the Country where the study species occur, the Wunambal Gaambera
and Wilinggin people, and recognise their continuing connection to land and water, paying our respects to Elders
past and present.
This work was supported by the Herman Slade Foundation, Holsworth Wildlife Research Endowment, Australian
Biological Resource Study, Bush Blitz, The Nature Conservancy, and the Thomas foundation. We thank Steve
Swearer and Tim Dempster (University of Melbourne) for their supervision and support during the broader project
that this was a part of, and Martin Gomon (NMV) and Michael Hammer (NTM) for their support in the preparation
of this manuscript. In addition, Di Bray (NMV), Glen Moore (WAM), Mark Allen (WAM) and Michael Hammer
provided access to specimens and information on collections when needed and each lodged specimens.
We are very grateful for the assistance of Tom Vigilante (Wunambal Gaambera Aboriginal Corporation) and
Rachel Treacy (Wilinggin Aboriginal Corporation) who helped coordinate engagement with the family of Dolores
Cheinmora and the Ngarinyin people, in particular Matthew Martin, that resulted in the naming of H. garawudjirri,
H.maranda, and H. wunduwala respectively. Finally, we thank the family of Dolores and all those involved in the
naming that we were not directly in contact with.
References
Allen, G.R., Midgley, S.H. & Allen, M. (2002) s.n. In: Field Guide to the Freshwater Fishes of Australia. Western Australian
Museum, Perth, Western Australia, pp. 236–239.
Allen, G.R. & Leggett, R. (1990) A collection of freshwater fishes from the Kimberley region of Western Australia. Records of
the Western Australian Museum, 14, 527–545.
April, J., Mayden, R.L., Hanner, R.H. & Bernatchez, L. (2011) Genetic calibration of species diversity among North America’s
freshwater fishes. Proceedings of the National Academy of Sciences, 108, 10602–10607.
https://doi.org/10.1073/pnas.1016437108
Birdsong, R.S., Murdy, E.O. & Pezold, F.L. (1988) A study of the vertebral column and median fin osteology in gobioid fishes
with comments on gobioid relationships. Bulletin of Marine Science, 42, 174–214.
Cheinmora, D., Charles, A., Karadada, T., Bernadette, W., Nyalerin, F., Waina, L., Punchi, M., Chalarimeri, A., Unghango, D.,
Saunders, T., Sefton, M., Vigilante, T. & Wightman, G. (2017) Belaa Plants and Animals: Biocultural Knowledge of the
Kwini People of the Far North Kimberley, Australia. Batchelor Institute Press, Batchelor, 283 pp.
de Queiroz, K. (2007) Species concepts and species delimitation. Systematic Biology 56, 879–886.
https://doi.org/10.1080/10635150701701083
de Groeve, J., Kusumoto, B., Koene, E., Kissling, W.D., Seijmonsbergen, A.C., Hoeksema, B.W., Yasuhara, M., Norder, S.J.,
Cahyarini, S.Y., van der Geer, A., Meijer, H.J.M., Kubota, Y. & Rijsdijk, K.F. (2022) Global raster dataset on historical
coastline positions and shelf sea extents since the Last Glacial Maximum. Global Ecology and Biogeography 31(11),
2162–2171.
https://doi.org/10.1111/geb.13573

A REVISION OF THE GENUS HYPSELEOTRIS OF NORTHWEST AUSTRALIA
Zootaxa 5311 (3) © 2023 Magnolia Press · 373
Hammer, M.P. & Moore, G.I. (2023) Wilinggin-West Kimberley Bush Blitz expedition: freshwater fishes. Report to the Director
of National Parks, Canberra. Western Australian Museum, Perth. [in prep]
Hoese, D.F. & Allen, G.R. (1983) A review of the gudgeon genus Hypseleotris (Pisces: Eleotridae) of Western Australia, with
descriptions of three new species. Records of the Western Australian Museum, 10, 243–261.
Hoese, D.F. & Allen, G.R. (1987) New Australian fishes Part 10. A new genus and two new species of freshwater Eleotrid fishes
(Gobioidei) from the Kimberley region of Western Australia. Memoirs of the Museum of Victoria, 48, 35–42.
https://doi.org/10.24199/j.mmv.1987.48.10
Hoese, D.F. (2018) Hypseleotris kimberleyensis Hoese & Allen 1982. In: Australian Faunal Directory. Australian Biological
Resources Study, Canberra. [viewed 19 April 2022]
Hoese, D.F. (2018) Hypseleotris ejuncida Hoese & Allen 1982. In: Australian Faunal Directory. Australian Biological Resources
Study, Canberra. [viewed 19 April 2022]
Hubbs, C.L. & Lagler, K.F. (1947) Fishes of the Great Lakes Region. Cranbrook Institute of Science, Bloomfield Hills, Michigan,
pp. 19–26.
Krefft, G. (1864) Notes on Australian freshwater fishes and descriptions of four new species. Proceedings of the Zoological
Society of London, 1864, 182–184.
Keith, P. & Mennesson, M.I. (2023) Revision of Hypseleotris (Teleostei: Eleotridae) from Indo-Pacific Islands using molecular
and morphometric approaches, with description of one new species. Zoological Journal of the Linnean Society, zlad003.
Larson, H.K. (2007) A new species of carp gudgeon, ‘Hypseleotris’ (Pisces: Gobioidei: Eleotridae), from the Katherine River
system, Northern Territory. Beagle: Records of the Museums and Art Galleries of the Northern Territory, 23, 111–117.
https://doi.org/10.5962/p.287430
Le Feuvre, M.C., Dempster, T., Shelley, J.J., Davis, A.M. & Swearer, S.E. (2021). Range restriction leads to narrower ecological
niches and greater extinction risk in Australian freshwater fish. Biodiversity and Conservation, 30 (11), 2955–2976.
Miller, M.A., Pfeiffer, W. & Schwartz, T. (2010) Creating the CIPRES Science Gateway for inference of large phylogenetic
trees. Proceedings of the Gateway Computing Environments Workshop (GCE), New Orleans, 2010, 1–8.
https://doi.org/10.1109/GCE.2010.5676129
Martin, H.A. (2006) Cenozoic climatic change and the development of the arid vegetation in Australia. Journal of Arid
Environments, 66, 533–563.
https://doi.org/10.1016/j.jaridenv.2006.01.009
McGowran, B., Holdgate, G.R., Li, Q. & Gallagher, S.J. (2004) Cenozoic stratigraphic succession in southeastern Australia.
Australian Journal of Earth Sciences, 51, 459–496.
https://doi.org/10.1111/j.1400-0952.2004.01078.x
Morgan, D., Charles, A., Nulgit, P. & Cheinmora, D. (2009) Fishes of the King Edward and Carson Rivers with their Belaa
and Ngarinyin names. Report to Land & Water Australia. Available from: https://researchportal.murdoch.edu.au/esploro/
outputs/report/Fishes-of-the-King-Edward-and/991005544318507891#file-0 (accessed 17 May 2023)
Pepper, M., Fujita, M.K., Moritz, C. & Keogh, J.S. (2011a) Palaeoclimate change drove diversification among isolated mountain
refugia in the Australian arid zone. Molecular Ecology, 20, 1529–1545.
https://doi.org/10.1111/j.1365-294X.2011.05036.x
Pepper, M., Ho, S., Fujita, M.K. & Keogh, J.S. (2011b) The genetic legacy of aridification: Miocene refugia fostered diversification
while Pleistocene climate cycles erased diversity in desert lizards. Molecular Phylogenetics and Evolution, 61, 750–759.
https://doi.org/10.1016/j.ympev.2011.08.009
Pusey, B.J., Kennard, M.J., Burrows, D., Perna, C., Kyne, P., Cook, B. & Hughes, J. (2011) Freshwater fish. In: Pusey, B.J.
(Ed.), Aquatic Biodiversity in Northern Australia: Patterns, Threats and Future. Charles Darwin University Press, Darwin,
pp. 82.
Shelley, J.J., Holland, O.J., Swearer, S.E., Dempster, T., le Feuvre, M.C., Sherman, C.D.H. & Miller, A.D. (2022) Landscape
context and dispersal ability as determinants of population genetic structure in freshwater fishes. Freshwater Biology, 67,
338–352.
https://doi.org/10.1111/fwb.13844
Shelley, J.J., Morgan, D.L., Hammer, M.P., Le Feuvre, M.C, Moore, G.I., Gomon, M.F., Allen, M.G. & Saunders, T. (2018a) A
Field Guide to the Freshwater Fishes of the Kimberley. Murdoch University Print Production Team, Perth, 262 pp.
Shelley, J.J., Swearer, S.E., Adams, M., Dempster, T., Le Feuvre, M.C., Hammer, M.P. & Unmack, P.J. (2018b) Cryptic
biodiversity in the freshwater fishes of the Kimberley endemism hotspot, northwestern Australia. Molecular Phylogenetics
and Evolution, 127, 843–858.
https://doi.org/10.1016/J.YMPEV.2018.06.032
Shelley, J.J., Dempster, T., Le Feuvre, M.C., Unmack, P.J., Laffan, S.W. & Swearer, S.E. (2019) A revision of the bioregionalisation
of freshwater fish communities in the Australian Monsoonal Tropics. Ecology and Evolution, 9 (8), 4568–4588.
https://doi.org/10.1002/ece3.5059
Shelley, J.J., Swearer, S.E., Dempster, T., Adams, M., Le Feuvre, M.C., Hammer, M.P. & Unmack, P.J. (2020) Plio‐Pleistocene
sea‐level changes drive speciation of freshwater fishes in north‐western Australia. Journal of Biogeography, 47, 1727–
1738.
https://doi.org/10.1111/jbi.13856
Shelley, J.J., Unmack, P.J., Dempster, T., Le Feuvre, M.C. & Swearer, S.E. (2019) The Kimberley, north-western Australia, as a

SHELLEY ET AL.
374 · Zootaxa 5311 (3) © 2023 Magnolia Press
cradle of evolution and endemic biodiversity: an example using grunters (Terapontidae). Journal of Biogeography 46 (11),
2420–2432.
https://doi.org/10.1111/jbi.13682
Shipway, B. (1950) Notes on the aquatic natural history of the lower Murchison River. Western Australian Naturalist, 2, 73–
77.
Stamatakis, A. (2014) RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics,
30, 1312–1313.
https://doi.org/10.1093/bioinformatics/btu033
Thacker, C. & Unmack, P.J. (2005) Phylogeny and biogeography of the eleotrid genus Hypseleotris (Teleostei: Gobioidei:
Eleotridae), with redescription of H. cyprinoides. Records of the Australian Museum, 57, 1–13.
https://doi.org/10.3853/j.0067-1975.57.2005.1436
Thacker, C.E. (2014) Species and shape diversification are inversely correlated among gobies and cardinalfishes (Teleostei:
Gobiiformes). Organisms Diversity & Evolution, 14, 419–436.
https://doi.org/10.1007/s13127-014-0175-5
Thacker, C.E. (2017) Patterns of divergence in fish species separated by the Isthmus of Panama. BMC Evolutionary Biology,
17, 1–14.
https://doi.org/10.1186/s12862-017-0957-4
Thacker, C.E., Geiger, D.L. & Unmack, P.J. (2022a) Species delineation and systematics of a hemiclonal hybrid complex in
Australian freshwaters (Gobiiformes: Gobioidei: Eleotridae: Hypseleotris). Royal Society Open Science, 9, 220201.
https://doi.org/10.1098/rsos.220201
Thacker, C.E., Shelley, J.J., McCraney, W.T., Adams, M., Hammer, M.P. & Unmack, P.J. (2022b) Phylogeny, diversification, and
biogeography of a hemiclonal hybrid system of native Australian freshwater fishes (Gobiiformes: Gobioidei: Eleotridae:
Hypseleotris). BMC Ecology and Evolution, 22, 1–23.
https://doi.org/10.1186/s12862-022-01981-3
Ward, R.D. (2009) DNA barcode divergence among species and genera of birds and fishes. Molecular Ecology Resources, 9,
1077–1085.
https://doi.org/10.1111/j.1755-0998.2009.02541.x
Yang, Z. (2007) PAML 4: phylogenetic analysis by maximum likelihood. Molecular Biology and Evolution, 24, 1586–1591.
https://doi.org/10.1093/molbev/msm088
Yang, Z. & Rannala, B. (2006) Bayesian estimation of species divergence times under a molecular clock using multiple fossil
calibrations with soft bounds. Molecular Biology and Evolution, 23, 212–226.
https://doi.org/10.1093/molbev/msj024