Range restriction leads to narrower ecological niches and greater extinction risk in Australian freshwater fish

Abstract

Human-induced environmental changes are accelerating biodiversity loss. Identifying which life-history traits increase extinction risk is important to inform proactive conservation. While geographically or numerically rare species are typically more vulnerable, ecological specialization may also increase extinction risk particularly when associated with rarity. We investigate whether regionally endemic freshwater fishes have more specialized diets and habitat requirements than more widely distributed, closely related species. We then use this information to assess extinction risk. Using closely-related widespread and endemic congeneric pairings from the Kimberley region of north-western Australia, we investigate whether there are ontogenetic diet shifts in 13 species and if some of these ontogenetic trophic units (OTUs) have narrow dietary niches. Using qualitative measures of habitat and presence/absence data, we also assess habitat specialization in 32 species. Overall, range-restricted species had narrower ecological niches. Ontogenetic diet shifts existed in 12 of 13 species and range-restricted species were more specialized for some or all of their OTUs compared to their widespread congenerics. Endemic species had a higher degree of variance in habitat use compared to their widespread congenerics, showing they had more specialized habitat requirements. As specialization is linked to extinction risk, the narrow niche breadth of small-ranged endemic fishes makes them more vulnerable to extinction than more cosmopolitan species. As many endemics from the Kimberley region have small ranges and/or low abundances, they may have an increased risk of extinction. By identifying which endemic species have narrow ecological niches, our study provides essential information for targeting proactive conservation efforts.

Full text

Vol.:(0123456789)
Biodiversity and Conservation
https://doi.org/10.1007/s10531-021-02229-0
1 3
ORIGINAL PAPER
Range restriction leads tonarrower ecological niches
andgreater extinction risk inAustralian freshwater fish
MatthewC.LeFeuvre1 · TimDempster1 · JamesJ.Shelley1 · AaronM.Davis2 ·
StephenE.Swearer1
Received: 14 April 2020 / Revised: 8 June 2021 / Accepted: 26 June 2021
© The Author(s), under exclusive licence to Springer Nature B.V. 2021
Abstract
Human-induced environmental changes are accelerating biodiversity loss. Identifying
which life-history traits increase extinction risk is important to inform proactive conserva-
tion. While geographically or numerically rare species are typically more vulnerable, eco-
logical specialization may also increase extinction risk particularly when associated with
rarity. We investigate whether regionally endemic freshwater fishes have more special-
ized diets and habitat requirements than more widely distributed, closely related species.
We then use this information to assess extinction risk. Using closely-related widespread
and endemic congeneric pairings from the Kimberley region of north-western Australia,
we investigate whether there are ontogenetic diet shifts in 13 species and if some of these
ontogenetic trophic units (OTUs) have narrow dietary niches. Using qualitative measures
of habitat and presence/absence data, we also assess habitat specialization in 32 species.
Overall, range-restricted species had narrower ecological niches. Ontogenetic diet shifts
existed in 12 of 13 species and range-restricted species were more specialized for some or
all of their OTUs compared to their widespread congenerics. Endemic species had a higher
degree of variance in habitat use compared to their widespread congenerics, showing they
had more specialized habitat requirements. As specialization is linked to extinction risk,
the narrow niche breadth of small-ranged endemic fishes makes them more vulnerable to
extinction than more cosmopolitan species. As many endemics from the Kimberley region
have small ranges and/or low abundances, they may have an increased risk of extinction.
By identifying which endemic species have narrow ecological niches, our study provides
essential information for targeting proactive conservation efforts.
Keywords Dietary diversity· Freshwater conservation· Habitat association· Intestinal
length· Melanotaeniidae· Terapontidae
Communicated by Mike Kevin Joy.
* Stephen E. Swearer
s.swearer@unimelb.edu.au
1 School ofBioSciences, University ofMelbourne, MelbourneVIC3010, Australia
2 Centre forTropical Water andAquatic Ecosystem Research, James Cook University, Townsville,
QLD, Australia

Biodiversity and Conservation
1 3
Introduction
Globally, human activities are imperiling species at an alarming rate (Ceballos et al.
2015). To address this extinction crisis, a growing body of literature aims to identify
which traits make particular species more vulnerable to extinction before declines occur
(Yu and Dobson 2000; Harcourt etal. 2002; Pritt and Frimpong 2010). Using such
insights to identify at-risk species can inform proactive conservation, which is generally
more successful and cheaper than reactive conservation (Kolar and Lodge 2002; Reed
and Shine 2002; Cardillo etal. 2006). Although species with small geographic ranges
(Gaston and Blackburn 1996; Pyron 1999; Le Feuvre etal. 2016) and/or low abun-
dances (McKinney 1997; Pritt and Frimpong 2010) are more vulnerable to extinction,
there is increasing awareness that ecological specialization can also increase extinction
risk (Munday 2004; Clavel etal. 2011).
Specialists are often well adapted to their local environment and perform well under
specific conditions. In contrast, generalists often utilize a wide variety of resources,
meaning they tend to perform consistently across a variety of habitats, but worse than
specialists in their preferred habitats (Hourigan and Reese 1987; Futuyma and Moreno
1988; Caley and Munday 2003; Reif etal. 2006). However, specialists are more likely to
have restricted ranges, potentially due to limits on resource availability and/or distribu-
tion, which may negate any potential benefits of specialization in reducing extinction
risk (Brown 1984; Slatyer etal. 2013). Furthermore, if the resources utilized by special-
ists decline in abundance or quality, specialists may be lost at a far greater rate than gen-
eralists (Munday 2004) because they are either unable to adapt or are outcompeted by
generalists in modified environments. This can result in a loss of redundancy and com-
plementarity within functional groups, increasing the risk of less resilient communities
and a reduction in key ecosystem services (Clavel etal. 2011).
Although specialization can take many forms, for instance temperature tolerance
(Calosi etal. 2010; Luna etal. 2012) and reproductive mode (Brändle etal. 2003), the
two most commonly investigated types of specialization linked to increased extinction
risk are dietary and habitat specialization. Although there are some exceptions (Heino
and Soininen 2006), the links between habitat specialization and small range size and/
or increased extinction risk are consistent across taxa including plants (Boulangeat
etal. 2012), invertebrates (Hughes 2000), parasites (Krasnov etal. 2008), fish (Pritt and
Frimpong 2010), birds (Symonds and Johnson 2006), marsupials (Fisher et al. 2003)
and primates (Harcourt etal. 2002). Dietary specialization is similarly linked to small
range and increased vulnerability (Harcourt etal. 2002; Beck and Kitching 2007; Boyles
and Storm 2007), however, this relationship is less robust, with several studies finding
no relationship between diet and either range size or extinction risk (Brashares 2003;
Symonds and Johnson 2006; Hobbs etal. 2010). Exceptions are generally a result of
species specializing on widespread resources (i.e., niche position, Heino and Soininen
2006; Lappalainen and Soininen 2006; Berkström etal. 2012) or high dispersal capac-
ity overriding the effects of specialization in determining range size (Bonte etal. 2004;
Youssef etal. 2011).
In addition, specialists often develop distinct morphologies, such as body and
tooth shape and intestinal length, that enable and enhance their specialization (Davis
and Betancur-R 2017). For example, intestinal length may indicate dietary specializa-
tion, with trophic level inversely related to intestinal length, as longer intestines are
required to effectively extract nutrients from plant matter and detritus. However, as long

Biodiversity and Conservation
1 3
digestive tracts require more energy to maintain, a trade-off exist between maximizing
energy extraction and minimizing energetic costs, so only specialist carnivores and her-
bivores have short and long intestines, respectively (Wagner etal. 2009; Davis etal.
2010, 2013).
In fishes, habitat specialists tend to be more vulnerable to extinction (Munday 2004;
Reynolds etal. 2005; Pritt and Frimpong 2010). Evidence suggests dietary specializa-
tion and extinction risk in fish are correlated, though few studies exist (Brooker etal.
2014). Verifying such a relationship is complicated by the prevalence and importance of
ontogenetic shifts in diet. Ontogenetic shifts in diet are driven by intrinsic (e.g., mouth
gape, gut length and body size) and extrinsic factors (e.g., preferred habitat type, preda-
tion risk and food availability) that change as individuals grow. As a result, rather than
looking at the diet of a species as a whole, it may be more informative to split species
into “ontogenetic trophic units” (OTUs). These OTUs correspond to distinctly different
trophic stages that occur as individuals grow (Stoner and Livingston 1984; Davis etal.
2011). Indeed, failure to account for OTUs may mask differences in dietary specializa-
tion between species and/or obscure developmental stages when dietary niches are nar-
row (Muñoz and Ojeda 1998).
The fishes of northern Australia present a model system for using specialization to iden-
tify species of conservation concern and inform proactive conservation for several reasons.
First, northern Australian rivers have mostly avoided large-scale development and remain
in good condition (Stein etal. 2002). Second, recent analyses indicate that northern Aus-
tralia is home to many fishes that have disproportionately low abundances and/or small
geographic ranges making them potentially vulnerable to extinction (Le Feuvre etal. 2016;
Le Feuvre 2017). If these species are also specialized, they will have a much greater risk of
extinction. Third, the region is earmarked for major development, particularly agriculture
and resource extraction, which will entail large-scale water capture, diversion and extrac-
tion and probable impacts on water quality. As a result, some rivers may be profoundly
altered in the coming decades (Petheram etal. 2014; Australian Government 2015).
The evidence to date, however, suggests that habitat and dietary specialization is typi-
cally rare in endemic northern Australian freshwater fishes. Omnivory, generalized habitat
requirements and ecological opportunism are common characteristics of northern Aus-
tralian fish and are thought to be adaptations to the seasonally variable climate across the
region (Pusey etal. 1995, 2010; Davis etal. 2011; Blanchette etal. 2013; Koehn and Ken-
nard 2013; Thorburn etal. 2014). Specialization, when it occurs, tends to be restricted to
areas where the environment is stable (Pusey etal. 2010; Davis etal. 2012). The predomi-
nant dietary specialization exhibited across northern Australia is piscivory. Feeding guilds
that are common elsewhere, such as detritivory and herbivory, are rare in Australia (Pusey
etal. 2010) but see Davis etal. (2011) (Terapontidae) and Ebner etal. (2011) (Sicydinnae).
Although the diets and habitat requirements of fish are reasonably well known for some
regions (e.g., north-eastern Queensland, Pusey etal. 2004) and taxonomic groups (Tera-
pontids, e.g., Davis etal. 2012), for many northern Australian fishes, particularly range
restricted species, their ecological requirements are poorly known.
Using congeneric pairs from multiple genera, we aimed to compare the dietary and hab-
itat specializations of endemic fauna with their widespread counterparts. Specifically, we
(1) investigate the relationship between body size and intestinal length, as intestinal length
may be indicative of specialization. We then (2) determine OTUs for each species and (3)
assess whether any OTUs are specialized. Next, we (4) analyze the degree of habitat spe-
cialization of each species. Finally, we (5) discuss whether endemics are more vulnerable
to extinction due to their ecological specialization.

Biodiversity and Conservation
1 3
Methods
Study species
We used exclusively freshwater teleost fish species that are endemic to the Kimberley
region of north-western Australia and widespread congenerics to investigate the relation-
ship between range size and dietary and habitat specialization. Species pairings were deter-
mined based on the co-occurrence of widespread and range restricted congeneric species
in the Kimberley region, for which adequate samples could be collected for dietary and/or
habitat analyses. For dietary analyses, we investigated diet breadth in 13 species from four
genera and habitat specialization in 32 species from 20 genera (Table1). This includes a
recently described species and genetically distinct candidate species (Shelley etal. 2017,
2018a, b, 2020a). For the Melanotaenia and terapontid species where both diet and habitat
information were available, the species pairs are closely related (Unmack etal. 2013; Shel-
ley etal. 2018b), so phylogenetic effects should be minimal. In species pairings with only
habitat data, the Neosilurus are similarly closely related (Huey etal. 2014), whereas the
Craterocephalus are more distantly related (Unmack and Dowling 2010).
Dietary specialization
We collected individuals from 11 species of terapontids and two species of melanotaeniids
from 12 catchments in 2012 − 2014 (Fig.1). We used a combination of backpack electro-
fisher, hook and line with spinner lures, and seine, fyke, and gill nets to collect fish for the
dietary analyses. Fyke nets were set overnight (set/cleared within approximately one hour
of sunset/sunrise respectively) and gill nets were set for approximately two hours before
sunset and regularly cleared to reduce impacts to bycatch. Within each genera most indi-
viduals were caught using the same suite of collection methods: Hephaestus were mostly
collected with gill nets, electrofishing and, to a lesser extent, hook and line with lures;
Syncomistes and Leiopotherapon were mostly collected with gill nets and electrofish-
ing; and Melanotaenia entirely with seine and fyke nets. While there is potential for net
feeding or collection methodology to confound the results (e.g., feeding while caught in
fyke nets), any effect on the overall conclusions of this study are likely to be negligible
as: (a) all capture methods except fyke nets largely preclude net feeding and (b) any biases
would be consistent within genera due to consistent collection methods. To control for
seasonal differences in diet (Thorburn etal. 2014), prey availability (Marchant 1982) and
habitat use (Jardine etal. 2012), we only used individuals collected during the dry season
(April-November).
We anaesthetized all fish in the field. For melanotaeniids, we preserved whole fish in
10% buffered formalin. For terapontids, we measured standard length (Ls), opened the
body cavity, removed the digestive tract, from the esophagus to the anus, and preserved it
in 10% buffered formalin.
In the laboratory, we measured Ls in Melanotaenia and then removed the entire diges-
tive tract. In both families, we measured intestinal length (LI) from the pyloric caeca to the
anus. In total, we assessed the lengths of 1056 digestive tracts. We then assessed stomach
fullness. If a qualitative assessment judged the stomach to be ≥ 20% full, we transferred the
gut contents into a petri dish and sorted prey items to the lowest practical taxonomic level
using field guides (Entwisle etal. 1997; Hawking and Smith 1997; Gooderham and Tsyrlin

Biodiversity and Conservation
1 3
Table 1 The species examined in the habitat and dietary analyses, whether they are range restricted or widespread, their distribution, the number of landscape and structural
habitats surveyed for each species, and their habitat preferences
Species Range Habitat Diet Distribution Landscape Structural Habitat
Ambassis macleayi W X King Edward River WA to Normanby River
QLD
13 33 Only found in permanent sites, mostly in shal-
low slow flow areas with rocky substrate and
edges with sediment substrate
Ambassis sp.a W X Fitzroy River WA to Murray River QLD and
Central Australia
47 111 Most common in shallow slow-flowing areas
Amiataba percoidesaW X Ashburton River WA to Brisbane River
QLD, and Central Australia
38 89
Craterocephalus helenae R X Drysdale River 8 22 Only found in permanent waterbodies with
sediment substrate and sandy edges. Most
commonly associated with shallow, mela-
leuca lined, slow-flow areas in the down-
stream reaches of the main channel
Craterocephalus lentiginosusaW X Fitzroy River WA, to Goyder River NT 20 52 Only found in rocky bottomed permanent
areas, mostly associated with shallow slow
flow areas
Glossamia aprion W X Fitzroy River WA to Clarence River NSW 23 59
Glossogobius giurus W X Ashburton River WA to Burnett River QLD 40 97
Hannia sp wintoni R X Prince Regent WA and Roe Rivers WA 9 22 Only found over rocky substrate
Hephaestus epirrhinosbR X X Drysdale River 8 22 Only in permanent waterbodies, and mostly
found in the main channel of the river
Hephaestus jenkinsi W X X Fitzroy River WA to Victoria River NT 48 114
Hypseleotris kimberleyensis R X Upper Fitzroy River 3 5 Found at permanent waterbodies in rocky,
pandanus lined gorges in tributaries of the
upper Fitzroy catchment. Generally associ-
ated with vertical edges
Kimberleyeleotris hutchensi R X Lower Mitchell River 7 14 Only found over rocky substrate at permanent
sites in the lower catchment and strongly
associated with vertical, pandanus lined
edges in the main river channel
Leiopotherapon macrolepis R X X Prince Regent WA and Roe Rivers WA 9 22 Rock substrate only

Biodiversity and Conservation
1 3
Table 1 (continued)
Species Range Habitat Diet Distribution Landscape Structural Habitat
Leiopotherapon unicolor W X X Murchison River WA to Murray River
QLD, Murray Darling Basin and Central
Australia
45 103
Melanotaenia australis W X X Ashburton River WA to Darwin Region NT 50 119
Melanotaenia gracillisbR X X Drysdale River only 8 22 Only found at permanent sites, most com-
monly over sediment based substrates
Mogurnda oligolepis W X Fitzroy River WA to Pentecost River WA 39 96
Nematolosa erebiaW X Ashburton River WA to South Coast QLD,
Murray Darling and Central Australia
40 97 Mostly in permanent sites
Neoarius graeffei W X Ashburton River WA to Hunter River NSW 24 56 Found permanent sites, mostly restricted to
deep waterholes
Neoarius midgleyi W X Fitzroy River WA to East Alligator River NT 23 57 Found in deep waterholes
Neosilurus ater W X Fitzroy River WA to Pioneer River QLD 32 75 Most common in deeper waterholes, and
generally more common downstream
Neosilurus hyrtliiaW X Fitzroy River WA to South Coast QLD,
Murray Darling and Central Australia
45 109 More commonly in shallower slow flow areas,
rarely downstream
Neosilurus pseudospinosusaR X Fitzroy River WA to Finnis River NT 32 78 Most common in riffles
Oxyeleotris selheimi W X Fitzroy River WA to Barron River QLD 21 50 Mostly in riffles
Strongylura krefftii W X Fitzroy River WA to Mary River QLD 25 62 Only in permanent, deep waterbodies
Syncomistes bonapartensis W X X King Edward WA to Finnis Rivers NT 28 70 Only found in rocky bottomed sites, mostly in
deep waterholes, generally associated with
pandanus
Syncomistes holsworthi R X X Durack River WA to Victoria River NT 9 18 Only found in rocky bottomed sites, mostly
associated with rock slab edges
Syncomistes kimberleyensis R X X Durack to Ord Rivers WA 9 18 Only found in rocky bottomed sites, mostly
in riffles

Biodiversity and Conservation
1 3
Table 1 (continued)
Species Range Habitat Diet Distribution Landscape Structural Habitat
Syncomistes rastellus R X X Drysdale River 8 22 Mostly found in vertical edged, rocky bot-
tomed deep waterholes in the upper catch-
ment
Syncomistes trigonicuscW X X Moran to Drysdale Rivers 23 60 Permanent sites only
Syncomistes versicolor R X Prince Regent River 5 14 N/A
Syncomistes wunambal R X X Mitchell River 7 14 Only found at permanent, rocky bottomed
sites, mostly in the lower catchment
Toxotoes chatareus W X Charnley River WA to Pascoe River QLD 28 67 Mostly deep waterholes
For range, W refers to widespread species and R to range restricted species
a Includes potential cryptic species
b A small population of these species exists in the King Edward River, but they are very poorly known and their taxonomy is unresolved so may or may not be the same species
c While this species has a relatively small range, it is spread across historically poorly connected catchments

Biodiversity and Conservation
1 3
2002). We then determined the wet weight of each prey type to the nearest 0.001g follow-
ing the methodology of Hyslop (1980). We used the mass method as it has been shown to
be most accurate (Ahlbeck etal. 2012). Due to a high level of among-individual variation
in diet, we grouped prey items into 23 broad categories (Appendix1). These categories
were chosen as they emphasized variation in food size, type and origin (i.e., benthos vs
water column, Davis etal. 2011).
The terapontid diet and intestinal length datasets were supplemented with data for 824
additional specimens (606 of which were Leiopotherapon unicolor) that were collected
from an additional 25 catchments across northern Australia (Fig.1; B.J. Pusey, D.L. Mor-
gan, unpubl. data). Collection and processing procedures were similar, however stomach
contents were determined using the volumetric approach of Hyslop (1980) (see Davis etal.
(2011) for further details). Comparative analyses indicated that the two approaches gave
similar results. In total, we analysed and compared the contents of 1799 stomachs.
Habitat specialization
During field surveys in the Kimberley region in June to August 2012 and 2013, we col-
lected data on habitat and fish catches (for 32 species) from 50 sites over 11 catchments
(Fig.1). As we only included species where we surveyed a variety of habitats across their
Fig. 1 Catchments where fish were caught for dietary analysis (shaded and numbered) and locations for fish
and habitat surveys (dots) across the Kimberley (top) and northern Australia (bottom). The black box on the
lower map shows the extent of the upper map

Biodiversity and Conservation
1 3
range (see above), 25 species and candidate species known from the Kimberley (Shel-
ley etal. 2018a, b) were excluded from analyses due to insufficient data on their habitat
requirements.
Habitat was qualitatively assessed using landscape (n = 5) characteristics at the site level
and structural (n = 6) traits at the microhabitat level (i.e., riffles, shallow pools and deep
pools, Appendix1). We used a backpack electrofisher to survey riffles, seine and fyke nets
to survey in shallow (generally < 1m) pools, and gill nets, hook and line and dip nets in
deep (generally > 1m) pools. This was due to the different efficiencies of each methodol-
ogy in each habitat and the inability of a single sampling method to survey every habitat
type and species effectively. Because abundance data is strongly influenced by collection
method (Jackson and Harvey 1997), we used presence/absence to remove this bias.
Data analysis
Diet
Differences in diet among species were explored in two ways. First, we investigated the
relationship between Ls and Li for each species to estimate trophic level, as shorter intes-
tines indicate a higher proportion of animal matter in the diet and longer intestines more
plant matter and/or detritus. Second, species were split into ontogenetic trophic units
(OTUs) based on size, as diet often varies with body size, due to both intrinsic and extrin-
sic factors changing with body size. We then investigated how diet varies among OTUs and
determined dietary niche breadth for each OTU.
To determine the relationship between Ls and Li for each genus, we ran ANCOVAs with
Ls as the dependent variable, species as the independent variable, and Li as the covari-
ate, using log transformed data. Where significant interactions occurred, we ran separate
regressions for each species. We also ran the same analyses using mass (g) instead of Ls,
but as log-transformed length and weight were highly correlated (r2 = 0.98), the analyses
gave identical results. Therefore, we only present analyses using Ls. As intestinal length is
correlated with body size, we would have preferred to generate a Ls independent measure
of Li sensu Wagner etal. (2009). However, as there was a significant interaction between
species and Ls, we could not use a common slope to standardize Li. As a result, non-over-
lapping 95% confidence intervals on linear regressions determined which values of Ls and
Li were significantly different between species within genera.
To determine OTUs and their differences, we transformed prey weights into propor-
tion of total diet for each prey category (weight of prey category/weight of entire gut con-
tents, excluding the unidentified fraction). To identify OTUs, we grouped species into size
classes. We split species that achieve a maximum Ls of < 200mm into 10mm size classes,
and split species growing to > 200mm Ls into 20mm size classes. We then averaged die-
tary data across each size class. For each species, we ran Similarity Percentages (SIM-
PER) analyses to determine the dissimilarity between different size classes, and the factors
influencing the dissimilarity. We also generated Principal component ordinations (PCOs)
to visualize shifts in diet (i.e., increases in diet dissimilarity between size classes), which
we used to delineate breaks between OTUs.
To determine how diet varied between OTUs within each genera, we used a two factor
nested Permutational Multivariate Analysis of Variance (PERMANOVA) based on 99,999
permutations of a Bray–Curtis dissimilarity matrix on dietary data, with species and OTU
nested within species as fixed factors. However, for Syncomistes, as there were seven rather

Biodiversity and Conservation
1 3
than two species in the comparisons, we investigated the effects of range restriction on diet
as well as species and OTU. We ran a PERMANOVA with distribution (widespread and
range restricted), species nested by distribution and OTU nested by species as fixed factors.
Significant terms were further investigated with a posteriori pairwise comparison using the
PERMANOVA t statistic, and the OTUs and food items contributing most to dissimilarity
in diet were revealed using SIMPER analyses. In addition, we investigated the effect of
catchment on each individual OTU using PERMANOVA, with catchment as a fixed factor.
As there was a strong effect of catchment on diet, the full PERMANOVAs outlined above
were performed at two levels, first in all sampled catchments where either species occurred,
and second, only in catchments where the species pairs were sympatric. This explicitly
tested whether generalists remain more generalized where they co-occur with specialists,
or whether specialist diets are merely constrained by catchment based resource differences.
However, as we only caught four individual Leiopotherapon unicolor co-occurring with L.
macrolepis, we expanded the co-occurring catchments to include the neighboring Mitchell,
King Edward and Drysdale rivers to get individuals from all OTUs. Similarly, as only four
Syncomistes bonapartensis were caught in the Drysdale River, the S. bonapartensis data
was expanded to include the neighboring King Edward catchment.
Finally, we averaged the diet data for each OTU and used PCOs for visualization.
In addition, we calculated standardized Levin’s niche breadth for each OTU using the
formula:
where

B
is Levin’s measure of niche breadth and

p
j
is the fraction of items in the diet
that are of food category j (Levins 1968). Niche breadth was then standardized using the
formula:
where

BA
is Levin’s standardized niche and n is the number of possible dietary items (Hurl-
bert 1978). To determine whether there was a consistent positive relationship between die-
tary niche breadth and range size, we then compared the niche breadth of widespread to
range-restricted OTUs using a t-test.
To improve normality, all dietary data was arcsine square-root transformed.
Habitat
We calculated the proportion of sites/microhabitats where each species was present (num-
ber of sites present/total number of sites) for each habitat value (Appendix1), restrict-
ing analyses to the catchment(s) where each species is found. For each of the 11 habitat
variables, we calculated the variance in the proportional data for each habitat value, with
higher variance indicating a higher degree of specialization. We used variance as a means
of determining habitat specialization over other approaches (Devictor etal. 2010), as it
distinguishes between species that are more common in one habitat category compared
to another (likely due to habitat preference; comparatively low variance) and species that
are very common in one habitat category and very rare or absent in another (likely due to
specialization; comparatively high variance). Similarly, coefficient of variation in habitat

B
=
1
∑p
2
j

B
A=

B−
1
n−1

Biodiversity and Conservation
1 3
associations can determine a species’ degree of habitat specialization (Julliard etal. 2006).
We then used this variance data for each habitat variable to run two-factor PERMANOVAs
based on 99,999 permutations of a Bray–Curtis dissimilarity matrix, with genus and dis-
tribution (widespread versus range restricted) as fixed factors. Significant terms were fur-
ther investigated with a posteriori pairwise comparison using the PERMANOVA t statistic,
and the habitat variables contributing most to dissimilarity of habitat types revealed using
SIMPER analyses. Comparison of habitat breadth was then visualized using PCOs. These
analyses were done for all genera and then for the co-occurring congeneric species alone.
We ran the analyses for co-occurring species at two levels, first in all sampled catchments
where either species occurred, and second, only in catchments where the species pairs were
sympatric. To improve normality, habitat variances were fourth-root transformed.
Results
Overall we found that range-restricted species tend to have more specialized diets and habi-
tat requirements than co-occurring widespread species from the same genus.
Diet
For each genus, we found a positive effect of Ls on Li accounting for the effects of spe-
cies and for species accounting for the effects of Ls (P < 0.0001 in all instances). We also
recorded a significant interaction between species and Ls, as slopes differed in steepness
between species (Appendixes 2 and 3) although all slopes remained positive (P < 0.0001
in all instances, Appendix4). Using separation in 95% confidence intervals, Hephaestus
jenkinsi Li was significantly longer than H. epirrhinos for a given Ls for individuals larger
than 48mm. All Leiopotherapon macrolepis have longer Li for a given Ls than L. unicolor.
Melanotaenia gracilis had a significantly shorter Li for a given Ls than M. australis for
individuals less than 55mm. Relationships were more complex for Syncomistes, with three
distinct groupings of intestinal lengths. Syncomistes bonapartensis has the shortest Li for a
given Ls, S. kimberleyensis, S rastellus and S. holsworthi have intermediate Li, and S. versi-
color, S. trigonicus and S. wunambal have the longest Li (Appendix5).
We identified 34 ontogenetic trophic units across the 13 species analysed, with each spe-
cies exhibiting 1–4 OTUs (Fig.2; Table2; Appendix9). In general, species exhibited one
of two developmental trajectories, either (a) starting out as an invertivore, then becoming
more omnivorous with prey size increasing with body size and, in some cases, ending up as
macrophagous carnivores, or (b) starting out as a meiophagous omnivore and becoming an
algivore–detritivore (Fig.2; Table2).
For each genus, nested PERMANOVAs revealed a significant effect of species in all
genera except Hephaestus. OTUs nested by species were different in all genera (P < 0.0001
in all instances). As we looked at the diet of seven species in Syncomistes rather than just
two in every other genus, we ran a different analysis that incorporated the role of distribu-
tion (range restricted versus widespread) and found that distribution, species nested by dis-
tribution and OTU nested by species were all significant (Table3; Appendixes 6 and 10).
There was a strong effect of catchment on prey type (e.g., proportion of algae to detritus,
the types of macroinvertebrates consumed and/or the proportion of macrocrustacea to fish)
but not the overall dietary guild (Appendix11). As a result, we re-ran the PERMANOVAs
but restricted the analyses to the catchments where species were sympatric. In Hephaestus

Biodiversity and Conservation
1 3
(P = 0.0004), Melanotaenia (P = 0.01) and East Kimberley Syncomistes (S. bonapartensis,
S. holsworthi, S. kimberleyensis, P < 0.0001) species were different, but in Leiopothera-
pon (P = 0.06) and Drysdale Syncomistes (S. bonapartensis, S. rastellus, S. trigonicus,
P = 0.2) they were not. In all genera, OTU nested within species differed (P < 0.002 in all
instances). Diet differed between distribution types among the East Kimberley Syncomistes
(P < 0.0001) but marginally not so among those from the Drysdale (P = 0.053, Appendixes
12 and 13). Individual variation was high, within and between catchments, especially
among the smallest OTUs (Appendixes 11 and 13).
Overall, widespread species had more generalized diets than range-restricted species. At
their smallest and largest OTUs, range restricted and widespread species were most similar.
However, at intermediate OTUs widespread species had far broader dietary niches (Fig.2;
Appendixes 6 and 10). Within Hephaestus, the widespread H. jenkinsi starts as an aquatic
invertivore, spends most of its life as an omnivore with prey size increasing with body size,
before becoming a macrophagous carnivore. In contrast, Hephaestus epirrhinos is entirely
carnivorous, with prey size increasing with body size. For Leiopotherapon, the widespread
L. unicolor follows the same trajectory as H. jenkinsi, whereas the endemic L. macrolepis
is a meiophagous omnivore at all body sizes, with prey size increasing slightly with body
size. Melanotaenia australis is omnivorous at all OTUs, with an increasing proportion of
algae in its diet, while the widespread M. gracilis is invertivorous at its smallest and larg-
est OTUs, with a period of omnivory at intermediate body sizes. For Syncomistes, species
Fig. 2 Principal component analysis of diet for the average values for all ontogenetic trophic units (OTUs).
Widespread species are black solid shapes and range restricted species are shown as outlined grey shapes,
with the numbers 1 − 4 representing the OTUs from smallest to largest body size. Pearson correlations
between OTUs and prey items > 0.5 are shown on the graphs. See Table 2 for description of each OTU
abbreviation and their size classes

Biodiversity and Conservation
1 3
tend to exhibit a higher degree of omnivory when small, and become algivore-detritivores
as they grow. The widespread Syncomistes bonapartensis and S. trigonicus, and to a lesser
extent the Drysdale endemic S. rastellus, spend longer as omnivores with S. trigonicus
maintaining a degree of omnivory in all OTUs.
Table 2 Ontogenetic trophic groups (OTUs) identified for each species, their size classes, the number of
individuals from each tropic group, each OTU’s standardized Levin’s niche breadth and the broad dietary
group assigned to each OTU
Asterisks indicate range-restricted species
Species OTU Size (mm) N Niche Breadth Dietary group
Hephaestus epirrhinos* HE1 ≤ 120 9 0.278 Aquatic invertivore
HE2 > 120 18 0.172 Macrophagous carnivore
Hephaestus jenkinsi HJ1 ≤ 40 11 0.272 Aquatic invertivore
HJ2 > 40, ≤ 120 95 0.426 Meiophagous omnivore
HJ3 > 120, ≤ 280 111 0.556 Macrophagous omnivore
HJ4 > 280 1 0.0003 Macrophagous carnivore
Leiopotherapon macrolepis* LM1 All 97 0.315 Meiophagous omnivore
Leiopotherapon unicolor LU1 ≤ 40 62 0.194 Aquatic invertivore
LU2 > 40, ≤ 80 553 0.465 Meiophagous omnivore
LU3 > 80, ≤ 120 187 0.653 Macrophagous omnivore
LU4 > 120 54 0.597 Macrophagous carnivore
Melanotaenia australis MA1 ≤ 40 87 0.638 Meiophagous omnivore
MA2 > 40 60 0.737 Meiophagous omnivore
Melanotaenia gracilis* MG1 ≤ 20 6 0.195 Invertivore
MG2 > 20, ≤ 50 60 0.457 Meiophagous omnivore
MG3 > 50 4 0.300 Invertivore
Syncomistes sp bonapartensis SB1 ≤ 80 40 0.133 Meiophagous omnivore/
detritivore
SB2 > 80, ≤ 160 57 0.090 Algivore detritivore/omnivore
SB3 > 160 37 0.076 Algivore detritivore
Syncomistes sp holsworthi* SH1 ≤ 60 2 0.074 Algivore detritivore/omnivore
SH2 > 60, ≤ 160 17 0.063 Algivore detritivore
SH3 > 160 10 0.045 Algivore detritivore
Syncomistes kimberleyensis* SK1 ≤ 50 11 0.116 Meiophagous omnivore
SK2 > 50 28 0.033 Algivore detritivore
Syncomistes rastellus* SR1 ≤ 40 1 0.161 Meiophagous omnivore/
detritivore
SR2 > 40, ≤ 90 9 0.074 Algivore detritivore/omnivore
SR3 > 90 17 0.070 Algivore detritivore
Syncomistes trigonicus ST1 ≤ 40 9 0.160 Meiophagous omnivore
ST2 > 40, ≤ 70 39 0.150 Algivore detritivore/omnivore
ST3 > 70 43 0.050 Algivore detritivore/omnivore
Syncomistes sp versicolor* SV1 ≤ 180 39 0.059 Algivore detritivore
SV2 > 180 2 0.001 Algivore
Syncomistes sp wunambal* SW1 ≤ 50 2 0.415 detritivore/omnivore
SW2 > 50 21 0.039 Algivore detritivore

Biodiversity and Conservation
1 3
Table 3 PERMANOVA results
testing for differences in dietary
and habitat preferences between
species
Bold indicates significant effects based on an alpha level of 0.05
Differences in dietary composition are between species and ontoge-
netic trophic units (OTUs) nested by species for four different genera
across the Kimberley region. For Syncomistes a second analysis was
run investigating the effects of distribution (widespread versus range
restricted), species nested by distribution and OTUs nested by species.
Differences in habitat preference are between genera and widespread
versus range-restricted species (distribution) for all species analysed
and for those just in co-occurring congeneric pairs
Source df Pseudo-F P Unique perms.
Diet
Hephaestus
Species 1 1.3 0.26 93828
OTU (Species) 4 6.5 < 0.0001 88437
Residual 239
Leiopotherapon
Species 1 15.6 < 0.0001 94136
OTU (Species) 3 26.9 < 0.0001 91099
Residual 948
Melanotaenia
Species 1 3.2 0.01 95223
OTU (Species) 2 5.9 < 0.0001 93920
Residual 213
Syncomistes — by species
Species 6 6.0 < 0.0001 92488
OTU (Species) 10 9.1 < 0.0001 90358
Residual 367
Syncomistes — by distribution
Distribution 1 9.1 < 0.0001 95629
Species (Dist.) 5 5.5 < 0.0001 92861
OTU (Species) 11 8.5 < 0.0001 90097
Residual 366
Habitat
All species
Genus 18 2.1607 0.0279 94047
Distribution 1 4.6699 0.0159 95532
Genus x distribution 5 1.6007 0.1804 94510
Residual 7
Total 31
Co-occurring species only
Genus 5 1.4955 0.1654 93064
Distribution 1 4.1186 0.03 80067
Genus × distribution 5 1.4917 0.1954 93902
Residual 4
Total 15

Biodiversity and Conservation
1 3
Standardized Levin’s niche breadth was twice as narrow for range restricted OTUs
(mean ± S.E = 0.16 ± 0.03) compared to widespread ones (0.33 ± 0.06; P = 0.03).
Habitat
Broadly, we found that range restricted species had more specialized habitat requirements
than widespread species with differences between genera. When comparing across all
genera and locations, significant and consistent effects of distribution (i.e., widespread v
range restricted) and genus on habitat breadth emerged (PERMANOVA results: Table3;
specific habitat preference: Table1; dissimilarities between co-occurring species within
each genera: Appendix14). For co-occurring species pairs across all locations, we only
recorded a significant effect of distribution (Table3). PCOs indicate a separation between
widespread and range restricted species across all genera except Neosilurus (Fig.3), indi-
cating a greater degree of habitat specialisation by range restricted species. However, when
analyses were restricted to just the catchments where species co-occurred, there was no
effect of distribution, a significant difference between genera (P = 0.028) and no interaction
(Appendix7).
Discussion
Our study is the first to use specialization to identify freshwater fish of conservation con-
cern in northern Australia. Range restricted freshwater fish are generally more ecologically
specialized than their widespread congenerics. Our findings add to the growing evidence
that the Kimberley, along with several other northern Australian regions, is a hotspot
of freshwater fish conservation concern, with risks likely to increase if the environment
changes.
Generally, our results are consistent with the hypothesis that range-restricted spe-
cies have specialized dietary and habitat requirements (Figs.2 and 3, Tables 1 and 2,
Appendix6). For four of the species pairs, evidence in addition to our results indicates
further specialization in range-restricted species (Appendix 8). As specialization is rare
in endemic, northern Australian freshwater fishes (Pusey etal. 2010; Davis etal. 2011),
Kimberley endemics are unique in their high degree of habitat and dietary specialization.
Indeed for terapontids, which made up the majority of our pairings, Davis etal. (2011)
notes their dietary versatility. However, this is not the case in the Kimberley, and may help
explain the familial diversity in the region (Shelley etal. 2018a). The high degree of spe-
cialization may be driven by the relative stability of the Kimberley’s rivers and the provi-
sion of numerous dry season refuges in the gorges typical of the region (Pepper and Keogh
2014; Shelley etal. 2019a). Generally, environmental stability is thought to be a driver of
specialization (Clavel etal. 2011), and the perennial rivers in northern Australia, such as
the Daly River and rivers of the Kimberley Plateau, are thought to support more specialist
species. (Pusey etal. 2010; Davis etal. 2012; Pepper and Keogh 2014).
Specialization could also be explained by the processes that have driven regional spe-
ciation; speciation in the Kimberley appears to be caused by sea level rise and fall due
to glacial cycles during the Pliocene and Pleistocene which connected and disconnected
many catchments every 40,000 to 100,000years, leading to allopatric speciation (Shelley
etal. 2019b, 2020b). When isolated, individuals may have adapted to the specific envi-
ronment of their disconnected catchment, and these specialist traits became fixed during

Biodiversity and Conservation
1 3
Fig. 3 Principal component ordination of habitat specialization in Kimberley freshwater fishes for all gen-
era (A) and just the widespread-range restricted congeneric pairs (B). For A, each genera is represented
numerically, whereas in B each genera is represented by a different symbol. In both graphs, black symbols
represent range-restricted species and grey symbols widespread ones. Pearson correlations between species
and habitat variables > 0.2 are shown on the graphs

Biodiversity and Conservation
1 3
the speciation process (Futuyma and Moreno 1988). When species rejoined their closely
related congener during lowered sea levels, specialization was possibly reinforced by niche
partitioning which allowed two similar species to coexist (McCormack etal. 2010).
In addition to having a broad diet, widespread species varied their diets across catch-
ments, more so than endemic species, indicating that they have an inherent capacity to shift
their diets in different environments. Dietary differences were generally in the proportion
of algae to detritus, the type of aquatic invertebrate eaten or the ratio of macroinvertebrates
to fish, rather than the overall dietary guild. Although these differences are likely driven
by variation in prey availability among catchments, it still indicates that widespread spe-
cies are adaptable and more resilient to change. For range restricted endemics found across
multiple catchments, only the Syncomistes holsworthi 3 OTU had different diets among
catchments, with 90% of the variation explained by a change in the proportion of detritus
to algae, rather than their feeding mode. Whether single catchment endemics could modify
their diets if moved into new catchments is unknown. For the species for which seasonal
data exists, some OTUs become increasingly omnivorous (Hephaestus jenkinsi 2, Melano-
taenia australis 1 and M. gracilis 2), others become more specialized (more algae, Synco-
mistes trigonicus 1 and 2), and some (H. epirrhinos 1) change their primary invertebrate
prey species (unpub. data), possibly in relation to seasonal prey availability (Marchant
1982). This suggests the diet of most specialists remains largely unchanged. Overall, our
results indicate that generalists modify their diets spatially and temporally, but, based on
the available evidence, specialists are less flexible.
As specialization is linked to extinction risk (Clavel etal. 2011), it appears that most
range restricted species in the Kimberley are more vulnerable to environmental change
than their widespread counterparts. However, specialization per se may not be the most
important factor in determining extinction risk; the resource a species is specialized on may
be more important. If a species has a high niche position (i.e., prefers atypical resources),
it may have a smaller range (Lappalainen and Soininen 2006) or be more vulnerable to
extinction (Johnson etal. 2002; Bonin 2012). In addition, the sensitivity of the required
resource to environmental change may also be important (Berkström etal. 2012; Brooker
etal. 2014). Many specialists are very abundant where found, as their required resource
may also be abundant (Reif etal. 2006). If the niche position hypothesis is correct, the
extinction risk to Kimberley specialists may be reduced as they are all specialized on rela-
tively abundant prey and habitat types.
In contrast to the niche position hypothesis, specialization alone, whether on abundant
or rare resources, may increase extinction risk (Reynolds etal. 2005; Passy 2012). So while
Kimberley specialists may be specialized on relatively common prey (e.g., fauna, and
algae and detritus), they may still be more vulnerable. For example, if fine sediment loads
increase due to changed land use or increased resource extraction, it could reduce the avail-
ability of algae, detritus and macroinvertebrates, and the feeding efficiency of both herbi-
vores and carnivores. While generalists may also experience declines in feeding efficiency
and reductions in the availability of some prey species, they should have greater capac-
ity to switch prey or increase dietary breadth in the face of environmental change (Wood
and Armitage 1997; Tebbett etal. 2017). Other factors, such as trophic level, may also be
important in determining extinction risk; high trophic level species tend to be more vulner-
able to extinction (McKinney 1997; Purvis etal. 2000) which may add further pressure on
the specialist carnivore H. epirrhinos.
An aspect of habitat specialization that we did not collect data on was ontogenetic
shifts in habitat use. Many fish species require specific nursery habitats (King 2004; Pusey
etal. 2004); for example Hephaestus fuliginosus (a close relative of H. jenkinsi) has an

Biodiversity and Conservation
1 3
ontogenetic shift from riffles that provide ample invertebrate prey and protection from
predators for smaller individuals to deeper pools as adults (Pusey etal. 2004). Our obser-
vations indicate that this is the case for the majority of grunters across the Kimberley, as
most juvenile terapontids were caught in riffles (Le Feuvre etal. unpubl. data). As a result,
juvenile grunters may be vulnerable to impacts that reduce the extent or duration of flow
in riffles (e.g., water extraction, storage and/or climate change). Unsurprisingly, impacts on
juvenile habitat has flow-on effects for adult populations; for example, the habitat breadth
of juvenile pomacentrids explained 74% of variation in response to habitat loss between
species on coral reefs (Wilson etal. 2008). Other riffle dwellers, such as the falsespine cat-
fish, Neosilurus pseudospinosus, are similarly vulnerable.
In addition, we did not quantify the availability of habitat and dietary resources, so we
could not distinguish between preference and species using resources based on resource
availability in the environment (Devictor etal. 2010). Despite these concerns, where spe-
cies occurred in sympatry, most species pairings exhibited different diets during at least
one OTU.
Our approach can be applied in other systems and taxa to identify vulnerable species
before populations decline. As quantifying specialization is intensive, it is particularly use-
ful as a second step once broad scale analyses have identified potentially vulnerable species
(e.g., Yu and Dobson 2000; Schipper etal. 2008) or areas of conservation concern (e.g.,
Brooks etal. 2006). Specialists can then be listed under Criterion E (Unfavorable Quantita-
tive Analysis) under the IUCN (Mace etal. 2008) before species are impacted. In addition,
knowing the ecological requirements of species provides specific information for conserva-
tion management.
Supplementary Information The online version contains supplementary material available at https:// doi.
org/ 10. 1007/ s10531- 021- 02229-0.
Acknowledgements We acknowledge the funding contributions for this project from the Hermon Slade
Foundation (HSF 11/4) and the Holsworth Wildlife Research Endowment. We thank the traditional own-
ers, landholders and the Department of Conservation for kindly giving us permission to work on their land.
Research was conducted under the Western Australian Department of Fisheries Instrument of Exemption
No. 2072 and the University of Melbourne Animal Ethics Permit No. 1212470.
Author contributions M.C.L., T.D., J.J.S. and S.E.S. designed the study; M.C.L., J.J.S. and A.M.D col-
lected all data; M.C.L. and S.E.S. conducted statistical analyses. All authors contributed critically to draft-
ing the manuscript.
Funding Hermon Slade Foundation (HSF 11/4) and the Holsworth Wildlife Research Endowment.
Data availability All data will be deposited in the Dryad Digital Repository.
Declarations
Ethical approval Research was conducted under the Western Australian Department of Fisheries Instrument
of Exemption no. 2072 and the University of Melbourne Animal Ethics Permit No. 1212470.
Consent for publication All authors gave final approval for publication.
References
Ahlbeck I, Hansson S, Hjerne O (2012) Evaluating fish diet analysis methods by individual-based model-
ling. Can J Fish Aquat Sci 69:1184–1201

Biodiversity and Conservation
1 3
Australian Government (2015) Our north, our future: white paper on developing northern Australia. Dep
Prim Minist and Cabinet Canberra
Beck J, Kitching IJ (2007) Correlates of range size and dispersal ability: a comparative analysis of sphingid
moths from the Indo-Australian tropics. Glob Ecol Biogeogr 16:341–349
Berkström C, Jones G, McCormick M, Srinivasan M (2012) Ecological versatility and its importance for the
distribution and abundance of coral reef wrasses. Mar Ecol Prog Ser 461:151–163
Blanchette ML, Davis AM, Jardine TD, Pearson RG (2013) Omnivory and opportunism characterize food
webs in a large dry-tropics river system. Freshwater Sci 33:142–158
Bonin MC (2012) Specializing on vulnerable habitat: acropora selectivity among damselfish recruits and the
risk of bleaching-induced habitat loss. Coral Reefs 31:287–297
Bonte D, Baert L, Lens L, Maelfait J-P (2004) Effects of aerial dispersal, habitat specialisation, and land-
scape structure on spider distribution across fragmented grey dunes. Ecography 27:343–349
Boulangeat I, Lavergne S, Van Es J, Garraud L, Thuiller W (2012) Niche breadth, rarity and ecological
characteristics within a regional flora spanning large environmental gradients: plant niche breadth and
ecological characteristics. J Biogeogr 39:204–214
Boyles JG, Storm JJ (2007) The perils of picky eating: dietary breadth is related to extinction risk in insec-
tivorous bats. PLoS ONE 2:e672
Brändle M, Stadler J, Klotz S, Brandl R (2003) Distributional range size of weedy plant species is correlated
to germination patterns. Ecology 84:136–144
Brashares JS (2003) Ecological, behavioral, and life-history correlates of mammal extinctions in West
Africa. Conserv Biol 17:733–743
Brooker RM, Munday PL, Brandl SJ, Jones GP (2014) Local extinction of a coral reef fish explained by
inflexible prey choice. Coral Reefs 33:891–896
Brooks TM, Mittermeier RA, da Fonseca GAB, Gerlach J, Hoffmann M, Lamoreux JF, Mittermeier CG,
Pilgrim JD, Rodrigues ASL (2006) Global biodiversity conservation priorities. Science 313:58–61
Brown JH (1984) On the relationship between abundance and distribution of species. Am Nat 124:255–279
Caley MJ, Munday PL (2003) Growth trades off with habitat specialization. Proc R Soc Lond Series B Biol
Sci 270:S175–S177
Calosi P, Bilton DT, Spicer JI, Votier SC, Atfield A (2010) What determines a species’ geographical range?
Thermal biology and latitudinal range size relationships in European diving beetles (Coleoptera:
Dytiscidae). J Anim Ecol 79:194–204
Cardillo M, Mace GM, Gittleman JL, Purvis A (2006) Latent extinction risk and the future battlegrounds of
mammal conservation. Proc Natl Acad Sci USA 103:4157–4161
Ceballos G, Ehrlich PR, Barnosky AD, García A, Pringle RM, Palmer TM (2015) Accelerated modern
human–induced species losses: entering the sixth mass extinction. Sci Adv 1:e1400253
Clavel J, Julliard R, Devictor V (2011) Worldwide decline of specialist species: toward a global functional
homogenization? Front Ecol Environ 9:222–228
Davis AM, Betancur-R R (2017) Widespread ecomorphological convergence in multiple fish families span-
ning the marine–freshwater interface. Proc R Soc B Biol Sci 284:20170565
Davis AM, Pearson RG, Pusey BJ, Perna C, Morgan DL, Burrows D (2011) Trophic ecology of northern
Australia’s terapontids: ontogenetic dietary shifts and feeding classification. J Fish Biol 78:265–286
Davis AM, Pusey BJ, Pearson RG (2012) Contrasting intraspecific dietary shifts in two terapontid assem-
blages from Australia’s wet-dry tropics: contrasting size-related diet habits in terapontids. Ecol Freshw
Fish 21:42–56
Davis AM, Pusey BJ, Thorburn DC, Dowe JL, Morgan DL, Burrows D (2010) Riparian contributions to the
diet of terapontid grunters (Pisces: Terapontidae) in wet-dry tropical rivers. J Fish Biol 76:862–879
Davis AM, Unmack PJ, Pusey BJ, Pearson RG, Morgan DL (2013) Ontogenetic development of intestinal
length and relationships to diet in an Australasian fish family (Terapontidae). BMC Evol Biol 13:53
Devictor V, Clavel J, Julliard R, Lavergne S, Mouillot D, Thuiller W, Venail P, Villéger S, Mouquet N
(2010) Defining and measuring ecological specialization. J Appl Ecol 47:15–25
Ebner BC, Thuesen PA, Larson H, Keith P (2011) A review of distribution, field observations and precau-
tionary conservation requirements for sicydiine gobies in Australia. Cybium 35:397–414
Entwisle TJ, Sonneman JA, Lewis SH (1997) Freshwater algae in Australia : a guide to conspicuous gener
St and Assoc Syd
Fisher DO, Blomberg SP, Owens IPF (2003) Extrinsic versus intrinsic factors in the decline and extinction
of Australian marsupials. Proc R Soc B Biol Sci 270:1801–1808
Futuyma DJ, Moreno G (1988) The evolution of ecological specialization. Annu Rev Ecol Syst 19:207–233
Gaston KJ, Blackburn TM (1996) Conservation implications of geographic range size-body size relation-
ships. Conserv Biol 10:638–646

Biodiversity and Conservation
1 3
Gooderham J, Tsyrlin E (2002) The Waterbug book: a guide to the freshwater macroinvertebrates of temper-
ate Australia. CSIRO Publishing, Collingwood
Harcourt AH, Coppeto SA, Parks SA (2002) Rarity, specialization and extinction in primates. J Biogeogr
29:445–456
Hawking JH, Smith FJ (1997) Colour guide to invertebrates of Australian inland waters. Co-op Res Cent for
Freshw Ecol Albury
Heino J, Soininen J (2006) Regional occupancy in unicellular eukaryotes: a reflection of niche breadth,
habitat availability or size-related dispersal capacity? Freshw Biol 51:672–685
Hobbs JA, Jones GP, Munday PL (2010) Rarity and extinction risk in coral reef angelfishes on isolated
islands: interrelationships among abundance, geographic range size and specialisation. Coral Reefs
29:1–11
Hourigan TF, Reese ES (1987) Mid-ocean isolation and the evolution of Hawaiian reef fishes. Trends Ecol
Evol 2:187–191
Huey JA, Cook BD, Unmack PJ, Hughes JM (2014) Broadscale phylogeographic structure of five freshwater
fishes across the Australian monsoonal tropics. Freshwater Sci 33:273–287
Hughes JB (2000) The scale of resource specialization and the distribution and abundance of lycaenid but-
terflies. Oecologia 123:375–383
Hurlbert SH (1978) The measurement of niche overlap and some relatives. Ecology 59:67–77
Hyslop EJ (1980) Stomach contents analysis—a review of methods and their application. J Fish Biol
17:411–429
Jackson DA, Harvey HH (1997) Qualitative and quantitative sampling of lake fish communities. Can J Fish
Aquat Sci 54:2807–2813
Jardine TD, Pusey BJ, Hamilton SK, Pettit NE, Davies PM, Douglas MM, Sinnamon V, Halliday IA, Bunn
SE (2012) Fish mediate high food web connectivity in the lower reaches of a tropical floodplain river.
Oecologia 168:829–838
Johnson CN, Delean S, Balmford A (2002) Phylogeny and the selectivity of extinction in Australian marsu-
pials. Anim Conserv 5:135–142
Julliard R, Clavel J, Devictor V, Jiguet F, Couvet D (2006) Spatial segregation of specialists and generalists
in bird communities. Ecol Lett 9:1237–1244
King AJ (2004) Ontogenetic patterns of habitat use by fishes within the main channel of an Australian flood-
plain river. J Fish Biol 65:1582–1603
Koehn JD, Kennard MJ (2013) Habitats. In: P. Humphries, K. Walker (eds) Ecology of Australian freshwa-
ter fishes, CSIRO Publishing, Collingwood, pp 25–48
Kolar CS, Lodge DM (2002) Ecological predictions and risk assessment for alien fishes in North America.
Science 298:1233–1236
Krasnov BR, Mouillot D, Khokhlova IS, Shenbrot GI, Poulin R (2008) Scale-invariance of niche breadth in
fleas parasitic on small mammals. Ecography 31:630–635
Lappalainen J, Soininen J (2006) Latitudinal gradients in niche breadth and position—regional patterns in
freshwater fish. Naturwissenschaften 93:246–250
Le Feuvre MC, Dempster T, Shelley JJ, Swearer SE (2016) Macroecological relationships reveal conser-
vation hotspots and extinction-prone species in Australia’s freshwater fishes. Glob Ecol Biogeogr
25:176–186
Le Feuvre MC (2017) Triple jeopardy in the tropics: assessing extinction risk in Australia’s freshwater bio-
diversity hotspot. Diss Univ Melb
Levins R (1968) Evolution in changing environments: some theoretical explorations. Princeton University
Press, Princeton
Luna B, Pérez B, Torres I, Moreno JM (2012) Effects of incubation temperature on seed germination of
mediterranean plants with different geographical distribution ranges. Folia Geobot 47:17–27
Mace GM, Collar NJ, Gaston KJ, Hilton-Taylor C, Akçakaya HR, Leader-Williams N, Milner-Gulland E, j.
& Stuart, S.N. (2008) Quantification of extinction risk: IUCN’s system for classifying threatened spe-
cies. Conserv Biol 22:1424–1442
Marchant R (1982) Seasonal variations in the macroinvertebrate fauna of billabongs along Magela Creek,
Northern Territory. Mar Freshw Res 33:329–342
McCormack JE, Zellmer AJ, Knowles LL (2010) Does niche divergence accompany allopatric divergence
in aphelocoma jays as predicted under ecological speciation?: insights from tests with niche models.
Evolution 64:1231–1244
McKinney ML (1997) Extinction vulnerability and selectivity: combining ecological and paleontological
views. Annu Rev Ecol Syst 1:495–516
Munday PL (2004) Habitat loss, resource specialization, and extinction on coral reefs. Glob Change Biol
10:1642–1647

Biodiversity and Conservation
1 3
Muñoz AA, Ojeda FP (1998) Guild structure of carnivorous intertidal fishes of the Chilean coast: implica-
tions of ontogenetic dietary shifts. Oecologia 114:563–573
Passy SI (2012) A hierarchical theory of macroecology. Ecol Lett 15:923–934
Pepper M, Keogh JS (2014) Biogeography of the Kimberley, Western Australia: a review of landscape evo-
lution and biotic response in an ancient refugium. J Biogeogr 41:1443–1455
Petheram C, Gallant J, Wilson P, Stone P, Eades G, Rogers L, Read A, Tickell S, Commander P, Moon
A, McFarlane D, Marvanek S (2014) Northern rivers and dams: a preliminary assessment of surface
water storage potential for northern Australia. CSIRO, Australia
Pritt JJ, Frimpong EA (2010) Quantitative determination of rarity of freshwater fishes and implications for
imperiled-species designations: determining rarity for imperiled species. Conserv Biol 24:1249–1258
Purvis A, Gittleman JL, Cowlishaw G, Mace GM (2000) Predicting extinction risk in declining species.
Proc R Soc b: Biol Sci 267:1947–1952
Pusey B, Kennard M, Arthington A (2004) Freshwater fishes of North-Eastern Australia. CSIRO Publish-
ing, Collingwood
Pusey BJ, Arthington AH, Read MG (1995) Species richness and spatial variation in fish assemblage struc-
ture in two rivers of the wet tropics of northern Queensland, Australia. Environ Biol Fishes 42:181–199
Pusey BJ, Arthington AH, Stewart-Koster B, Kennard MJ, Read MG (2010) Widespread omnivory and low
temporal and spatial variation in the diet of fishes in a hydrologically variable northern Australian
river. J Fish Biol 77:731–753
Pyron M (1999) Relationships between geographical range size, body size, local abundance, and habitat
breadth in North American suckers and sunfishes. J Biogeogr 26:549–558
Reed RN, Shine R (2002) Lying in wait for extinction: ecological correlates of conservation status among
australian elapid snakes. Conserv Biol 16:451–461
Reif J, Hořák D, Sedláček O, Riegert J, Pešata M, Hrázský Z, Janeček Š, Storch D (2006) Unusual abun-
dance–range size relationship in an Afromontane bird community: the effect of geographical isolation?
J Biogeogr 33:1959–1968
Reynolds JD, Webb TJ, Hawkins LA (2005) Life history and ecological correlates of extinction risk in
European freshwater fishes. Can J Fish Aquat Sci 62:854–862
Schipper J, Chanson JS, Chiozza F, Cox NA, Hoffmann M, Katariya V, Lamoreux J, Rodrigues ASL, Stuart
SN, Temple HJ, Baillie J, Boitani L, Lacher TE, Mittermeier RA, Smith AT, Absolon D, Aguiar JM,
Amori G, Bakkour N, Baldi R, Berridge RJ, Bielby J, Black PA, Blanc JJ, Brooks TM, Burton JA,
Butynski TM, Catullo G, Chapman R, Cokeliss Z, Collen B, Conroy J, Cooke JG, da Fonseca GAB,
Derocher AE, Dublin HT, Duckworth JW, Emmons L, Emslie RH, Festa-Bianchet M, Foster M, Foster
S, Garshelis DL, Gates C, Gimenez-Dixon M, Gonzalez S, Gonzalez-Maya JF, Good TC, Hammerson
G, Hammond PS, Happold D, Happold M, Hare J, Harris RB, Hawkins CE, Haywood M, Heaney LR,
Hedges S, Helgen KM, Hilton-Taylor C, Hussain SA, Ishii N, Jefferson TA, Jenkins RKB, Johnston
CH, Keith M, Kingdon J, Knox DH, Kovacs KM, Langhammer P, Leus K, Lewison R, Lichtenstein G,
Lowry LF, Macavoy Z, Mace GM, Mallon DP, Masi M, McKnight MW, Medellín RA, Medici P, Mills
G, Moehlman PD, Molur S, Mora A, Nowell K, Oates JF, Olech W, Oliver WRL, Oprea M, Patterson
BD, Perrin WF, Polidoro BA, Pollock C, Powel A, Protas Y, Racey P, Ragle J, Ramani P, Rathbun G,
Reeves RR, Reilly SB, Reynolds JE, Rondinini C, Rosell-Ambal RG, Rulli M, Rylands AB, Savini S,
Schank CJ, Sechrest W, Self-Sullivan C, Shoemaker A, Sillero-Zubiri C, Silva ND, Smith DE, Srini-
vasulu C, Stephenson PJ, van Strien N, Talukdar BK, Taylor BL, Timmins R, Tirira DG, Tognelli MF,
Tsytsulina K, Veiga LM, Vié J-C, Williamson EA, Wyatt SA, Xie Y, Young BE (2008) The status of
the world’s land and marine mammals: diversity, threat, and knowledge. Science 322:225–230
Shelley JJ, Delaval A, Le Feuvre MC (2017) A revision of the grunter genus Syncomistes (Teleostei, Tera-
pontidae, Syncomistes) with descriptions of seven new species from the Kimberley region, northwest-
ern Australia. Zootaxa 4367:1–103
Shelley JJ, Delaval A, Le Feuvre MC, Dempster T, Raadik TA, Swearer SE (2020a) Revision of the genus
Hannia (Teleostei, Terapontidae), with description of a new species, Hannia wintoni, from the Kim-
berley, Western Australia. Zootaxa 4869:562–586
Shelley JJ, Dempster T, Le Feuvre MC, Unmack PJ, Laffan S, Swearer SE (2019a) A revision of the biore-
gionalisation of freshwater fish communities in the Australian monsoonal tropics. Ecol Evol 9:4568–
4588. https:// doi. org/ 10. 1002/ ece3. 5059
Shelley JJ, Morgan DL, Hammer MP, Le Feuvre MC, Moore GI, Gomon MF, Allen MG, Saunders T
(2018a) A field guide to the freshwater fishes of the Kimberley. Murdoch University Publishing Team,
Perth
Shelley JJ, Swearer SE, Adams M, Dempster T, Le Feuvre MC, Hammer MP, Unmack PJ (2018b) Cryptic
biodiversity in the freshwater fishes of the Kimberley endemism hotspot, northwestern Australia. Mol
Phylogenet Evol 127:843–858

Biodiversity and Conservation
1 3
Shelley JJ, Swearer SE, Dempster T, Adams M, Le Feuvre MC, Hammer MP, Unmack PJ (2020b) Plio-
Pleistocene sea-level changes drive speciation of freshwater fishes in north-western Australia. J Bioge-
ogr 47:1727–1738. https:// doi. org/ 10. 1111/ jbi. 13856
Shelley JJ, Unmack PJ, Dempster T, Le Feuvre MC, Swearer SE (2019b) The Kimberley, north-western
Australia, as a cradle of evolution and endemic biodiversity: an example using grunters (Terapontidae).
J Biogeogr 46(11):2420–2432. https:// doi. org/ 10. 1111/ jbi. 13682
Slatyer RA, Hirst M, Sexton JP (2013) Niche breadth predicts geographical range size: a general ecological
pattern. Ecol Lett 16:1104–1114
Stein JL, Stein JA, Nix HA (2002) Spatial analysis of anthropogenic river disturbance at regional and conti-
nental scales: identifying the wild rivers of Australia. Landsc Urban Plan 60:1–25
Stoner AW, Livingston RJ (1984) Ontogenetic patterns in diet and feeding morphology in sympatric sparid
fishes from seagrass meadows. Copeia 1984:174–187
Symonds MRE, Johnson CN (2006) Range size-abundance relationships in Australian passerines. Glob
Ecol Biogeogr 15:143–152
Tebbett SB, Goatley CHR, Bellwood DR (2017) Fine sediments suppress detritivory on coral reefs. Mar
Pollut Bull 114:934–940
Thorburn DC, Gill H, Morgan DL (2014) Predator and prey interactions of fishes of a tropical Western Aus-
tralia river revealed by dietary and stable isotope analyses. J R Soc West Aust 97:363–387
Unmack PJ, Allen GR, Johnson JB (2013) Phylogeny and biogeography of rainbowfishes (Melanotaeniidae)
from Australia and New Guinea. Mol Phylogenet Evol 67:15–27
Unmack PJ, Dowling TE (2010) Biogeography of the genus Craterocephalus (Teleostei: Atherinidae) in
Australia. Mol Phylogenet Evol 55:968–984
Wagner CE, McIntyre PB, Buels KS, Gilbert DM, Michel E (2009) Diet predicts intestine length in Lake
Tanganyika’s cichlid fishes. Funct Ecol 23:1122–1131
Wilson SK, Burgess SC, Cheal AJ, Emslie M, Fisher R, Miller I, Polunin NVC, Sweatman HPA (2008)
Habitat utilization by coral reef fish: implications for specialists vs. generalists in a changing environ-
ment. J Anim Ecol 77:220–228
Wood PJ, Armitage PD (1997) Biological effects of fine sediment in the lotic environment. Environ Manage
21:203–217
Youssef S, Baumel A, Véla E, Juin M, Dumas E, Affre L, Tatoni T (2011) Factors underlying the narrow
distribution of the Mediterranean annual plant Arenaria provincialis (Caryophyllaceae). Folia Geobot
46:327–350
Yu J, Dobson FS (2000) Seven forms of rarity in mammals. J Biogeogr 27:131–139
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.