See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/263491033

Gut content and stable isotope analyses provide complementary

understanding of ontogenetic dietary shifts and trophic relationships among
fishes in a tropical river

Article  in  Freshwater Biology · October 2012

DOI: 10.1111/j.1365-2427.2012.02858.x

CITATIONS READS

123 1,646

5 authors, including:

Aaron M Davis Melanie Blanchette

James Cook University Edith Cowan University
54 PUBLICATIONS   2,154 CITATIONS    32 PUBLICATIONS   639 CITATIONS   

SEE PROFILE SEE PROFILE

Bradley James Pusey Tim Jardine

University of Western Australia University of Saskatchewan
152 PUBLICATIONS   7,825 CITATIONS    129 PUBLICATIONS   4,896 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

1000 IRES Project View project

Catchment to Reef Project View project

All content following this page was uploaded by Tim Jardine on 30 October 2017.

The user has requested enhancement of the downloaded file.

Freshwater Biology (2012) 57, 2156–2172 doi:10.1111/j.1365-2427.2012.02858.x

Gut content and stable isotope analyses provide

complementary understanding of ontogenetic dietary shifts
and trophic relationships among fishes in a tropical river
A A R O N M . D A V I S * , †, M E L A N I E L . B L A N C H E T T E †, B R A D L E Y J . P U S E Y ‡, T I M D . J A R D I N E § A N D
RICHARD G. PEARSON†
*
Centre for Tropical Water and Aquatic Ecosystem Research, James Cook University, Townsville, QLD 4811, Australia
†
School of Marine and Tropical Biology, James Cook University, Townsville, QLD 4811, Australia
‡
Centre of Excellence in Natural Resource Management, University of Western Australia, Albany, WA 6330, Australia
§
Tropical Rivers and Coastal Knowledge, Australian Rivers Institute, Griffith University, Nathan, QLD 4111, Australia

SUMMARY
1. Despite widespread recognition of the role of body size in fish trophic ecology, little attention
has been focused on this issue in isotopic studies, particularly in tropical systems.
2. We used analyses of stomach contents and stable isotopes to examine size-related shifts in diet
in a terapontid fish assemblage in the Australian wet–dry tropics. Stomach content analysis
identified substantial ontogenetic dietary shifts in all species, corresponding to changes in body
size–isotope trajectories for two species. Shifts away from relatively specialised diets of heavily
13
C-depleted insect larvae to consumption of a range of items across multiple basal carbon sources
appeared to be the proximate cause of observed isotopic changes.
3. Allochthonous organic matter in the form of C3 riparian vegetation was particularly important
to smaller terapontids before larger fish shifted to a broad range of dietary items and similarly
broad range of basal carbon sources.
4. While there was general agreement between d13C and stomach content analysis, there was
minimal concurrence between the latter and d15N isotopic derivation of estimates of trophic
position. Due to factors such as omnivory, isotopically overlapping basal sources and uncertainties
about rates of isotopic fractionation in both predator and prey species, stomach content analysis
provides an essential complement to isotopic methodologies in tropical systems.
5. Given that basal sources supporting any individual species can change markedly with
ontogeny, consideration of intraspecific, size-related variation is necessary in isotopic studies of
food web structure.

Keywords: body size, feeding ecology, mixing model, omnivory, organic matter

to identify carbon sources or food web structure (Hyslop,

Introduction
1980). SCA has several inherent limitations such as
Body size is a fundamental determinant of fish tropho- providing only a short-term (hours to days) dietary
dynamics, with size-related dietary shifts prevalent in the ‘snapshot’ of recently ingested items (Hyslop, 1980), the
life histories of many fish species (Mittelbach & Persson, requirement of high sampling frequency to obtain a
1998; Jennings et al., 2002; Davis et al., 2011a). The result- reliable time-integrated overview of dietary habits and
ing ‘ontogenetic niches’ (sensu Werner & Gilliam, 1984) minimal indication of the degree of assimilation of dietary
can greatly complicate our understanding of feeding items (Parkyn, Collier & Hicks, 2001). These limitations
interactions in aquatic food webs. Dietary studies have have led to increasing emphasis on stable isotope analysis
traditionally relied upon stomach content analysis (SCA) (SIA) as a tool to assess aquatic food web structure and

Correspondence: Aaron M. Davis, Centre for Tropical Water and Aquatic Ecosystem Research, James Cook University, Townsville, QLD 4811,
Australia. E-mail: [email protected]

2156  2012 Blackwell Publishing Ltd

 Isotopic ecology of some tropical fishes 2157
function (Vander Zanden, Cabana & Rasmussen, 1997; terapontid grunters inhabiting the Burdekin River, one of
Post, 2002). A major advantage of isotopic approaches is Australia’s largest wet–dry tropical catchments. The aims
that they provide temporally integrated information of the study were to (i) compare the correspondence
(weeks to months) on dietary habits, reflecting foods that between SCA and SIA in determining diet, (ii) examine
are actually assimilated by the consumer. Stomach content changes in trophic position with ontogeny and (iii) assess
data can, however, provide information on taxonomic and whether proportions of food exploited by terapontids
size composition of diets and predator–prey interactions vary in relation to body size.
in complex systems where species consume a diversity of
items that may be difficult to identify from stable isotope
Methods
ratios alone (Layman, Winemiller & Arrington, 2005a). A
combination of SIA and SCA is increasingly being used to Study sites
improve interpretation of feeding studies and aquatic
Fish were collected from the upper Burdekin catchment in
food webs (Parkyn et al., 2001; Mantel, Salas & Dudgeon,
the wet–dry tropics of north-eastern Australia (Fig. 1). The
2004; Layman et al., 2005a).
Burdekin catchment is the fifth largest in Australia
Stable isotope analysis is also being increasingly used to
(130 000 km2). Study sites were located upstream of the
clarify ontogenetic dietary shifts in fishes. The use of d15N
Burdekin Falls Dam in the Basalt, Cape ⁄ Campaspe and
as an indicator of trophic position suggests that size-
Burdekin rivers and Keelbottom Creek. The region expe-
related dietary transitions are common, although the
magnitude of effects can vary (Jennings et al., 2002;
Galván, Sweeting & Reid, 2010). The majority of demon- 145° 146° 147° 148°

strated isotopic shifts have been from simple, often

18°
plankton-driven marine or freshwater food chains where Burdekin catchment
Sample location
size-structured feeding is expected to be pronounced Locality

(Post, 2003; Galván et al., 2010). Food webs in tropical Lower Burdekin river
Burd Study river
ek
rivers, however, tend to be characterised by omnivory and in
19°
Townsville
by being short, diffuse and highly interconnected (Jepsen

Ck.
Riv
& Winemiller, 2002; Winemiller, 2004; Douglas, Bunn & er Ayr

Davies, 2005; Layman et al., 2005a,b; Pusey et al., 2010;

Basalt R.
Rayner et al., 2010). Trophic enrichment of nitrogen
20°
isotopes in tropical rivers is often less than the predicted Charters Towers
Bo
3–4& (Kilham et al., 2009), with many large predatory fish Ri

w
ve

en
r
in these rivers occupying similar trophic positions to e R. Lake Bro
ke
ap Dalrymple
C

smaller-bodied fishes (Layman et al., 2005b), suggesting

nR
iver
21°
that many species feed across multiple trophic levels
(Jepsen & Winemiller, 2002; Douglas et al., 2005). In these Sutt
or
more trophically complex systems, with a greater diversity
of production sources and weaker size structuring, isoto-
r e
Belyando Riv

22°
pic evidence of size-related diet shifts may be difficult to
document. Simultaneous studies of SCA and SIA of
different size classes of species from tropical fresh waters
are rare, with many studies limiting isotopic assessments
23°
to adult fish to avoid the confounding effects of ontogeny.
Northern Australia’s terapontid grunters are an ideal
Alpha
group to examine the utility of SIA in discerning
ontogenetic dietary shifts in tropical freshwater fishes.
24°
The Terapontidae is one of the most trophically diverse of
Australia’s freshwater fish families, with pronounced
ontogenetic dietary shifts (determined by SCA) a prom- 0 25 50 100 150 200
km

inent aspect of species’ dietary ecology (Pusey, Kennard &

Arthington, 2004; Davis et al., 2010; Davis, Pearson & Fig. 1 Study site locations in the Upper Burdekin River catchment.
Pusey, 2011b). This study examines the diets of the Data  Commonwealth of Australia (Geoscience Australia) 1990.

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

2158 A. M. Davis et al.
riences a sub-humid tropical (monsoonal) climate, char- samples were collected at each site via 20 vertical hauls of
acterised by pronounced seasonality in rainfall and a 63-lm net through the water column.
discharge. Highest annual river flows typically occur Potential basal food sources collected at each site (when
from December to April (the ‘wet season’), with lowest present) were aquatic and terrestrial vascular plants,
flows between August and October (the ‘dry season’). The submerged leaves, filamentous algae, seston (a surrogate
flow regime of the Burdekin River is amongst the most for phytoplankton, also including traces of suspended
variable in the world for rivers of comparable size particulate matter) and biofilm (epilithon). Multiple leaves
(Puckridge et al., 1998), and flows in the upper Burdekin of the dominant terrestrial and aquatic vascular plants
study area have been classified as ‘unpredictable inter- occurring at each site were clipped directly from the plant.
mittent’ (Kennard et al., 2010). This flow regime class is Submerged leaf packs were collected from the substratum
notable for extreme intra- and interannual flow variability at each site and stored in plastic bags for transport.
and variable timing of maximum flows. Regional vegeta- Samples of filamentous algae were collected with forceps
tion is dominated by sclerophylous Eucalyptus and Acacia from the substratum and stored in the same manner.
woodlands, and the riparian zone typically composed of Seston was collected using 20 vertical hauls of a 20-lm
Melaleuca and occasional open vine thickets (Pearson, plankton net at each site. For biofilm collection, three
1991). The upper catchment is sparsely populated, the stones at each site were scrubbed with a toothbrush, and
predominant land use being low-intensity cattle grazing. the collected material was washed through 800- and 250-
The upper Burdekin River system is of low gradient, with lm sieves to remove coarse detritus and macroinverte-
an under-fit channel, steep banks and minimal off- brates, respectively, with the resultant slurry from each
channel, floodplain habitat. Despite its size, the Burdekin stone being collected in an individual tube. With
River has low diversity of instream habitats. The river is the exception of plankton and biofilm samples (see
largely characterised by long shallow reaches dominated below), all samples were stored on ice in the field and
by a sand and fine gravel substratum (Pearson, 1991). immediately frozen upon return to the laboratory prior to
processing.
Fish specimens were measured (standard length in mm)
Study species
and weighed prior to excision of the stomach and viscera
The four species of Terapontidae (grunters) that occur in from the body cavity. Stomachs estimated to be more than
the Burdekin River are collectively the most abundant 20% full were transferred to a glass dish, and the
fishes, by numbers and biomass, in this system (Pusey, contribution of each food item was determined by the
Arthington & Read, 1998). They comprise three of Austra- indirect volumetric method of Hyslop (1980). Food items
lia’s more widespread fish species – spangled perch were identified to the lowest practical taxonomic level
Leiopotherapon unicolor (Günther), barred grunter Amniataba within 46 categories, including species level for fish,
percoides (Günther) and sooty grunter Hephaestus fuliginosus family level for macroinvertebrates and functional groups
(Macleay), which are common elements of the fish assem- for items such as filamentous algae and detritus
blage, and the small-headed grunter Scortum parviceps (Table S1). For fish longer than 40 mm SL, a fillet of
(Macleay), which is endemic to the Burdekin catchment. dorsal white muscle was collected for SIA, as this tissue
typically has lower lipid and inorganic carbonate content
than other tissues and yields lower variability in d15N and
Sampling and sample processing
d13C values (Pinnegar & Polunin, 1999). For fish <40 mm
Fish were collected at twelve sites across major tributaries SL, whole fish excluding head, tail and viscera were used
of the upper Burdekin River from September to Novem- for analysis. Different tissues tend to follow similar
ber 2009 (late dry season), using electro-fishing (boat patterns of isotopic change, so should not have unduly
mounted and backpack) and gill netting, with efforts influenced resulting trends (Doucett et al., 1999). Samples
made to sample as wide a range of size classes for each of abdominal muscle were excised from larger decapods
species as possible at all sites. Macroinvertebrates were crustaceans. All aquatic macroinvertebrates were first
collected using a 2-m hand-held 250-lm mesh dip net pooled into functional feeding groups according to
from all major habitat types (edge, bed, macrophytes, taxonomy (e.g. family) and, if a greater sample size was
runs) at each site. Samples were pooled to represent site- needed, combined according to feeding mode based on
level abundance and diversity. Animals were sorted live Hawking & Smith (1997), but always at least within the
on-site and placed into distilled water for a minimum of same taxonomic order, and processed whole as a com-
3 h to facilitate purging of gut contents. Zooplankton posite sample. All animal tissue samples were oven-dried
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
 Isotopic ecology of some tropical fishes 2159
at 55 C to a constant mass before being ground to a fine and the pellet re-suspended in the solvent mixture. These
homogeneous powder using a pestle and mortar. steps were repeated at least three times or until the
Leaves of all aquatic and terrestrial vascular plants and solvent ran clear. Finally, the pellet was dried (60 C) and
filamentous algae were washed with distilled water, ground. About 1000 lg of the sample was analysed for
brushed and visually inspected to ensure removal of d13C and d15N using the isotopic analysis techniques
contaminants prior to oven-drying. Zooplankton and outlined above.
phytoplankton samples underwent initial processing d13C values of all terapontids were corrected for the
prior to refrigerated storage (0 C). Each zooplankton potential effects of lipid content prior to any analyses.
sample was poured through 250- and 60-lm sieves and Based on the subset of methanol–chloroform extracted
rinsed with distilled water, with the 60-lm-size fraction samples, the relationship between bulk tissue d13C, lipid-
retained for storage. Each phytoplankton sample was extracted d13C and bulk tissue C : N was used to develop
poured through 60- and 20-lm sieves to remove contam- the following lipid correction equation for terapontids:
inants larger than 60 lm, rinsed thoroughly with distilled
water and the 20-lm-size fraction was retained. Biofilm d13 Ccorrected ¼ d13 Cuntreated þ 3:33 logðC : Nbulk Þ
samples were centrifuged in an Eppendorf 5702 centrifuge  3:62ðr2 ¼ 0:64Þ:
at approximately 112 g for 10 min to concentrate material.
The chlorophyll-rich top fraction was collected from each This equation fell between the correction predictions of
60-mL tube, then filtered through pre-combusted What- Post et al. (2007) and Logan et al. (2008).
man GF ⁄ C glass fibre filter papers (0.7 lm) and stored
frozen in aluminium foil prior to oven-drying. Data analysis
Dietary shifts. Due to the questionable applicability of the
Stable isotope analysis taxonomic species as a valid functional ecological unit in
Isotopic analysis for d15N, d13C, %C, %N and C ⁄ N was fish ecology, we employed the concept of the ‘ontogenetic
conducted by the Colorado Plateau Stable Isotope trophic unit’ (OTU; sensu Stoner & Livingston, 1984),
Laboratory at Northern Arizona University. Samples were where an individual terapontid species was subdivided
analysed in continuous-flow mode using a Thermo- into size classes (standard length: <40, 40–80, 80–160 and
Finnigan Delta plus Advantage gas isotope-ratio mass >160 mm) with different known ecological (trophic) roles
spectrometer interfaced with a Costech Analytical ECS4010 (Davis et al., 2011a). Dietary data were square-root trans-
elemental analyszer. Data were normalised using four formed, and the level of dietary overlap between largest
internationally accepted isotope reference standards and smallest OTU’s was compared using the Bray–Curtis
(International Atomic Energy Agency standards CH6, similarity index, which provides values ranging from 0%
CH7, N1 and N2). External precision on these standards (no shared species) to 100% (all prey species shared and
was ±0.10& or better for d13C and ±0.20& or better for consumed in the same proportion), and can be as a
d15N. d13C and d15N data are expressed relative to standard measure of dietary overlap (Marshall & Elliott, 1997).
reference material: Vienna Pee Dee Belemnite for carbon Although, in reality, there are no defined critical levels
and atmospheric N for nitrogen. with which similarity values can be compared, the
intensity of overlap was assessed as high (>60), interme-
diate (40–60) or low (<40), following Ross (1986).
Lipid correction
Following Logan et al. (2008), a subset of samples from Size-related shifts in trophic position (SCA). Stomach con-
each fish species was analysed in duplicate (one lipid- tent data were used to estimate a unique trophic position
extracted, one non-extracted) to develop an appropriate for individual fish following the formula:
lipid correction equation. Lipid extraction was done using X
TPSCA ¼ ðVi Ti Þ þ 1
the modified Folch, Lees & Stanley (1957) technique at the
Colorado Plateau Stable Isotope Laboratory. Approxi- where TPSCA is the trophic position of a species size class
mately 10–20 mg of ground tissue was soaked in a 2 : 1 weighted by the volumetric contribution of the ith prey
chloroform : methanol (by volume) solvent mixture and item (Vi), and Ti is trophic position of the ith prey item
the material suspended by stirring. After 15 min, the (sensu Winemiller, 1990; Vander Zanden et al., 1997).
sample was centrifuged (300 g for 5 min), the supernatant Trophic positions of prey items were allocated by major
discarded (i.e. the analysis was not quantitative for lipids) taxonomic groups ranging from primary producers
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
2160 A. M. Davis et al.
(algae, aquatic macrophytes, terrestrial vegetation) and d15N fractionation rates of primary invertebrate con-
detritus-inorganic material (trophic position 1.0) to sumers can be substantially lower than those of secondary
predaceous invertebrates and fish (trophic position 3.0). consumers such as fish (Vander Zanden & Rasmussen,
Estimated trophic positions of prey items were assigned 2001; Vanderklift & Ponsard, 2003). Therefore, to provide
according to the concept of the dominant functional feeding context to changes in d15N values of terapontids during
group (sensu Merritt & Cummins, 1996) of the majority of ontogeny, mean d15N values for different prey items were
members of the group (Table S1). Trophic categories for also plotted to compare d15N values and fractionation
macroinvertebrate groups were derived from Hawking & rates of invertebrate and fish prey items occupying the
Smith (1997). Feeding designations and trophic positions same nominal trophic positions (Table S1). Differences in
for the common prey fish species were based on the trophic the mean d15N values between fish and invertebrate
habits outlined by Pusey et al. (2010). Trends in trophic consumers in each trophic category (i.e. herbivores–
position according to size (standard length) at a species detritivores, omnivores and carnivores) were assessed
level, as well as for all individuals of each species at each using one-way A N O V A .
site, were analysed using linear regression.
Estimation of size-related assimilated basal source material from
SIA. Changes in d13C according to standard length for
Stable isotope analyses
individuals of each species were analysed using linear
Size-related shifts in trophic position (SIA). It is widely regression. The relative contribution of different basal
recognised that estimation of trophic position from SIA is carbon sources to diet of different OTUs was assessed
sensitive to the assumed nitrogen fractionation value used using Stable Isotope Analysis in R (SIAR; Parnell et al.,
in calculations (DeNiro & Epstein, 1978; Post, 2002; 2010) a Bayesian mixing model that runs on the R
Vanderklift & Ponsard, 2003). Following the method platform (R Development Core Team, 2009). Bayesian
described in Winemiller, Akin & Zeug (2007), trophic mixing models have the advantage of allowing the
position based on isotopic data was calculated using the variation and uncertainties associated with isotopic esti-
formula: mates and trophic enrichment to be propagated through
the model, with outputs being more reflective of the
TPSIA ¼ ½ðd15 Nconsumer  d15 Nsource Þ=3:3 þ 1
natural variability within a system. The SIAR model is fit
where d15Nsource was the mean of all basal sources via Markov Chain Monte Carlo methods producing
occurring at each site, and the denominator 3.3 was an simulations of values of dietary proportions of sources
estimated mean trophic enrichment (fractionation) be- consistent with the data using a Dirichlet prior distribu-
tween consumers and their diet. The 3.3& estimate was tion (Parnell et al., 2010). Prior to SIAR analyses, related
calculated using the isotopic data from specimens of all basal sources with similar isotopic compositions were
terapontid species at <40 mm SL. All terapontids in this grouped to minimise the number of sources and simplify
size range have similar invertivorous diets (Davis et al., the range of possible solutions (Phillips, Newsome &
2011a), dominated by larvae of three orders of aquatic Gregg, 2005). Biofilm typically has a large component of
insect (Ephemeroptera, Diptera and Trichoptera) that are attached filamentous algae (Rasmussen, 2010), and explor-
lower-level consumers. The mean d15Nsource was sub- atory analyses identified d13C values for filamentous algae
tracted from the d15N of all terapontids <40 mm SL at each as having a significant correlation with biofilm across sites
site and divided by 2 (to account for the secondary (r2 = 0.51, P < 0.05), so the two sources were combined at
consumer status of fish in this size range). This metric is site level and termed ‘benthic algae’. Five food sources
analogous to the calculation of average fractionation rate were used in the mixture modelling: terrestrial C4 grasses,
throughout the food web proposed by Kilham et al. (2009). terrestrial C3 vegetation (all other terrestrial vegetation
The resultant 3.3& fractionation estimate was slightly and submerged leaf packs), seston, aquatic macrophytes
above the average fractionation rate of 3.0& derived for (submerged) and benthic algae.
white muscle tissue from the meta-analysis of Vanderklift The SIAR model was run at both the catchment level
& Ponsard (2003) and aligned closely with average (using the overall mean and standard deviation for each
fractionation rates (3.4&) derived from the meta-analysis of the five basal sources and the overall mean for each
of Post (2002). Trends in trophic position according to OTU) and the individual site level (using the values for
standard length for individuals of each species at each available basal sources at each site and the site mean for
site, as well as at a broader species level, were analysed each OTU) to assess broad- and fine-scale trends in size-
using linear regression. related feeding of the species. Each terapontid species was
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
 Isotopic ecology of some tropical fishes 2161
coded into its size class groups (OTUs). Trophic enrich- tures of all four species’ diets (Fig. 2). All species were
ment factors (TEFs) for nitrogen were corrected using the invertivorous in their smaller size classes, with diets
value of 3.3 ± 0.5& (mean ± 1 SD as above). A d13C dominated by Diptera larvae (particularly Chironomidae
fractionation rate of 0.4 ± 1.3& was assumed following and Simuliidae), as well as Trichoptera and Ephemerop-
Post (2002). The SIAR mixing model was run for 500,000 tera. As they grew, fish shifted away from this reliance on
iterations, discarding the first 50 000 samples. The result- invertebrates, although the nature of these shifts varied
ing distributions of probability–density functions of the among species. Amniataba percoides continued to consume
feasible foraging solutions produced by SIAR allowed large numbers of invertebrates in all size classes, but
direct identification of the most probable solution for filamentous algae and aquatic macrophytes became
basal carbon sources supporting the OTU of each species increasingly important with size. Hephaestus fuliginosus
(Parnell et al., 2010). Upper and lower 95% credibility displayed pronounced size-related dietary change, with
intervals were used to describe the contribution for each smaller invertebrates diminishing in importance and
diet item (Phillips & Gregg, 2003). larger animals such as fish (primarily Melanotaenia splen-
dida), shrimp (Palaemonidae) and plants (aquatic macro-
phytes, filamentous algae) dominating the diet of larger
Results
individuals. With increased size, L. unicolor shifted from
Stomach content analysis invertivory to generalised carnivory, consuming an
increasingly broad array of predominantly animal prey
A total of 337 stomachs from the four terapontid species
(aquatic insects, terrestrial insects, prawns and fish).
were examined. Ontogenetic shifts were prominent fea-

Fig. 2 Volumetric proportions (%) of major prey items across Burdekin terapontid size classes. Prey categories are based on those presented in
Table S1. The number of individuals (n) examined in each size class is indicated.

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

2162 A. M. Davis et al.
Scortum parviceps changed from a largely invertivorous strated a significant negative trend in trophic position as it
diet to an almost exclusive diet of aquatic and terrestrial shifted from a diet dominated by invertebrates to herbiv-
plant material and detritus in larger specimens. These ory in larger size classes. The remaining terapontids fed
diets and ontogenetic shifts are largely consistent with across trophic positions through much of their life
previous reports of the trophic ecology of these species histories, often ranging between primary consumers
(Pusey et al., 2010; Davis et al., 2011a,b). Bray–Curtis (trophic position 2) through to tertiary predators (trophic
analysis demonstrated high dietary similarity within position about 4) within a similar size range.
species’ OTUs across sites (A.M. Davis, unpublished
data), concurring with previous reports (Pusey et al.,
Stable isotope analyses
2010; Davis et al., 2011a,b).
Within-species Bray–Curtis dietary overlap values Size-related shifts in trophic position. At the species level,
based on SCA between the smallest and largest OTUs of there was a small positive relationship between trophic
A. percoides, H. fuliginosus, L. unicolor and S. parviceps level, indicated by d15N, and standard length for H. fulig-
were 45.6, 2.27, 11.53 and 28.83, respectively. With the inosus and L. unicolor, but not for A. percoides or S. parviceps
exception of A. percoides, where overlap between size (Fig. 4). The weak positive relationships (<0.5 trophic
classes was moderate (i.e. Bray–Curtis dietary overlap position) occurred between individuals <40 mm SL and
between 40 and 60), size-related diet shifts in all other those >160 mm SL in both species. These relationships
species resulted in low (>40) overlap. The extent of the between trophic position and standard length were also
ontogenetic dietary shift in H. fuliginosus was particularly largely reflected at a site level, with significant positive
pronounced, with essentially no similarity in diet between relationships between d15N and standard length evident at
the < 40 and > 160 mm SL size classes. several sites for both H. fuliginosus and L. unicolor, with
Linear regression of trophic position estimates calcu- positive but non-significant trends evident at most remain-
lated from SCA showed that, with the exception of ing sites (Table S2). There were no significant relationships
S. parviceps, significant size-related shifts in trophic posi- between diet-based and isotope-based estimates of trophic
tion were not apparent (Fig. 3). Scortum parviceps demon- position for individual fishes in any of the species (Fig. 5),

Fig. 3 Size-related trophic position estimates for Burdekin River terapontids based on stomach content analysis (SCA). The regression line
represents a significant relationship between standard length and trophic position for Scortum parviceps (r2 = 0.796, P < 0.001).

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

 Isotopic ecology of some tropical fishes 2163

Fig. 4 Size-related trophic position estimates for Burdekin River terapontids based on stable isotope analysis (SIA). Solid lines represent
significant regression relationships between standard length and trophic position (for Hephaestus fuliginosus r2 = 0.164, P < 0.01; for Leiopo-
therapon unicolor r2 = 0.06, P < 0.05).

indicating low correspondence between the two tech- length were largely reflected at the site level, with
niques. Isotope-based trophic positions for S. parviceps significant positive relationships evident at several sites
tended to be higher than those derived from SCA. for both H. fuliginosus and particularly L. unicolor (Ta-
Comparison of d15N values for fish and invertebrate ble S2), and similar but non-significant trends in regres-
prey in terapontid diets indicated that fish from all trophic sions at most remaining sites.
positions (herbivore–detritivore, omnivore and carnivore)
consistently had elevated d15N values compared with Estimation of size-related assimilated basal source material from
invertebrate consumers ostensibly occupying similar tro- isotopic data. Mean d15N and d13C values of basal sources
phic niches (Fig. 6). There were significant differences across the 12 sites revealed low variability within catego-
between mean d15N values of fish and invertebrates at all ries, with source isotopic ratios generally consistently
positions of the food web (herbivores: F1 = 60.8, P < 0.001; positioned relative to each other (Fig. S1; Table S3):
omnivores: F1 = 45.6, P < 0.001; carnivores: F1 = 163.8, terrestrial C3 vegetation was the most 13C-depleted source
P < 0.001). at all sites, aquatic macrophytes were typically the most
13
C-enriched aquatic plant at most sites, while C4 grasses
Size-related shifts in d13C. Regression analysis indicated were the most distinct basal sources, exhibiting the most
no significant rate of increase in d13C with standard length enriched d13C values. Mean d13C values of the autochtho-
in A. percoides or S. parviceps at the species level (Fig. 7). A nous basal sources (benthic algae, aquatic macrophytes
significant positive relationship between d13C and stan- and seston) were not well differentiated, although there
dard length was evident in L. unicolor with individuals was some divergence in d15N across these sources. The
becoming progressively enriched in d13C as length dominant animal prey, including fish, in terapontid diets
increased (r2 = 0.413, P < 0.001). A positive size-related exhibited intermediate d15N between terapontids and
shift in d13C enrichment was also evident in H. fuliginosus basal sources.
(r2 = 0.392, P < 0.001) with increased size. These relation- There was considerable overlap in the d13C and d15N
ships between overall species’ d13C values and standard ratios of many terapontid OTUs, particularly among
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
2164 A. M. Davis et al.

Fig. 5 Comparison of trophic position values derived from stomach content analysis (SCA) and stable isotope analysis (SIA) for four Burdekin
River terapontid species. The diagonal line represents total correspondence between the two methodologies. For Amniataba percoides r2 = 0.004,
P > 0.05; Hephaestus fuliginosus r2 = 0.001, P > 0.05; Leiopotherapon unicolor r2 = 0.04, P > 0.05; and Scortum parviceps r2 = 0.20, P > 0.05.

smaller (<80 mm SL) sizes (Fig. S1). The ontogenetic shifts

demonstrated in the regression analyses of individual
species’ isotopic values are also evident in the d13C–d15N
biplots. Progressive enrichment in both d13C and d15N was
particularly evident in the sequential OTUs of H. fuligi-
nosus and L. unicolor as their isotopic values diverged
from their high overlap as juveniles. There was some
variation in isotopic values of several of the significant
prey items in terapontid diet. The invertebrate larvae
(Diptera, Ephemeroptera, Lepidoptera and Trichoptera)
that predominated in the diets of smaller terapontids
(<80 mm SL) were relatively depleted in d13C. Prey items
that increased in importance as species such as H. fulig-
inosus and L. unicolor grew (e.g. palaemonid shrimps, the
fish M. splendida and aquatic macrophytes) were relatively
enriched in d13C.
The ability of the SIAR mixing model to resolve
proportions of different food sources in terapontid diets
Fig. 6 Mean (±SE) d15N values for basal sources, with invertebrates
and fish grouped according to trophic position classifications out- varied between size classes at both site and catchment
lined in Table S1. levels. Results were similar at both scales so are presented
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
 Isotopic ecology of some tropical fishes 2165

Fig. 7 Trends in size-related d13C for Burdekin River terapontids. Solid lines represent significant regression relationships between standard
length and d13C.

here for the catchment scale only. The carbon sources basal sources (benthic algae, seston and aquatic macro-
supporting the diet of H. fuliginosus and L. unicolor phytes) made the dominant contribution to diet, except
became less well resolved as fish increased in size where the model could not discriminate any dominant
(Fig. 8). Terrestrial C3 vegetation made the dominant source. Instances where an individual carbon source was
contribution in both species for size classes >80 mm SL; clearly dominant for a species’ OTU at a site (i.e. minimum
however, the C3 pathway progressively diminished in contribution to model outputs >0.4 for the lower 95%
importance for fish over 80 mm in both species, with a probability value) were rare but, when evident, generally
range of basal carbon sources making similar feasible took the form of C3 vegetation as the major source for fish
contributions to the diet. A similar (though less pro- OTUs smaller than 80 mm SL. Multiple carbon sources
nounced) trend in relative importance of C sources was making likely contributions to OTUs (>0.2 for the lower
evident in A. percoides. C3 vegetation was the major 95% probability value) were commonplace across species
contributor to diet in specimens smaller than 40 mm SL (particularly in larger size classes) and sites.
(95% credibility interval, 45–62%), remaining the domi- These SIA results align well with dietary shifts docu-
nant C source in all size classes, although as fish grew its mented from SCA. SIAR modelling of source contribu-
importance diminished. No obvious size-related patterns tions to diet of the key invertebrates in juvenile stomach
in relative contributions of different carbon sources were contents showed a major reliance on C3 vegetation (A. M.
evident for S. parviceps, with seston, C3 vegetation, Davis, unpubl. data), highlighting their role in transfer of
aquatic macrophytes and biofilm making similar contri- energy to these secondary consumers. The combined
butions to diet across all size classes. volumes of these 13C-depleted invertebrates decreased
These catchment-scale trends were also largely reflected markedly as standard length increased for L. unicolor,
in SIAR results conducted at a site scale. C3 vegetation was H. fuliginosus and S. parviceps (Fig. S2). In contrast, volu-
typically the primary carbon source supporting A. perco- metric contribution of these insect taxa to the diet of A.
ides, L. unicolor and H. fuliginosus in OTUs smaller than percoides remained relatively constant as fish size in-
80 mm SL at most sites (Table S4). C3 vegetation dimin- creased, with this species also remaining heavily reliant
ished markedly in significance for the larger OTUs of on C3 vegetation across all size classes. Other prey items
L. unicolor and H. fuliginosus at most sites, where different contributing to the diet of larger size classes of H. fulig-

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

2166 A. M. Davis et al.

Fig. 8 Boxplots derived from the Stable Isotope Analysis in R mixing model showing the contribution of different primary carbon sources to the
diets of Burdekin terapontid size classes using d13C and d15N isotopes. The proportions show credibility intervals plotted at 95, 75 and 50%
credibility intervals. Carbon sources are labelled: SE, seston; AM, aquatic macrophytes; C3V, C3V terrestrial vegetation; BA, benthic algae and
C4, C4 terrestrial grasses.

inosus and L. unicolor (i.e. Palaemonidae, M. splendida) two terapontids exhibited a corresponding enrichment in
demonstrated reliance on a broad range of relatively 13C- 13
C with increased size and reliance on a diversity of 13C-
enriched basal sources (A. M. Davis, unpubl. data). These enriched basal sources.

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

 Isotopic ecology of some tropical fishes 2167
demonstrate a significant decrease in trophic position
Discussion
through SCA, exhibited no correlation between body size
Stable isotope analysis and SCA both indicated that and trophic position according to SIA.
feeding across several trophic positions with several basal Caut, Angulo & Courchamp (2009) noted that dispar-
sources (i.e. omnivory) is important in the diets of ity between stomach content and isotopic estimates of
Burdekin River terapontids. This finding supports the trophic position can be due to several factors such as
view that omnivory is an adaptive response to seasonal insufficient sample sizes, biases in the contributions of
variations in water-level and trophic resources that basal sources used to derive trophic position estimates
characterise hydrologically variable tropical river systems from d15N data and errors in the assumed d15N
(Lowe-McConnell, 1975; Jepsen & Winemiller, 2002). In a fractionation rates. The sample of only 18 individuals
system as variable as the Burdekin River, species traits are of S. parviceps in this study may have been insufficient
expected to revolve around resource generalism and for accurate appraisal of diet, but it can be noted that
opportunistic foraging (Townsend & Hildrew, 1994; Poff SCA for this species was similar to that in much larger
& Allan, 1995; Pusey et al., 2010). However, the contribu- studies in the same catchment (Pusey et al., 2010; Davis
tions of individual sources varied considerably over the et al., 2011b). Sample sizes for the other species were
life history of fish. Both SIA and SCA approaches relatively high and also concurred with previous
suggested that the early life-history stages of all terapon- research, indicating that SCA-derived estimates of tro-
tid species were heavily reliant on a limited set of prey phic position were robust with regard to sample size.
(insect larvae) and C3 terrestrially derived plants. The The issue of assumed fractionation rates is particularly
diets of two of species in particular (H. fulginosus and pertinent to trophic position estimates in isotopic studies.
L. unicolor) became more generalised with size and as a Scortum parviceps demonstrated the greatest disparity
result were less well resolved with SIAR. Therefore, while between the analytical methods, at least in relation to
large terapontids may be generalists, smaller individuals trophic position, with SIA significantly overestimating
are more specialised. Identifying these stages is critical to trophic position in larger size classes compared with SCA.
understanding the trophic ecology of these species, and The dietary shift from carnivory (invertivory) in juveniles
ultimately, the broader food web. to herbivory in the largest size classes should have been
accompanied by a relative reduction in d15N as fish grew.
Apparent overestimation of the trophic position of nom-
Correspondence and mismatches between SCA and SIA
inally herbivorous species using SIA is not uncommon
The two methods for assessing size-related dietary shifts (Carseldine & Tibbetts, 2005; Layman et al., 2005b; Mill,
produced both congruent and conflicting results. There Pinnegar & Polunin, 2007; Winemiller et al., 2011). Differ-
was broad agreement between SCA and carbon mixing ences in food quality and feeding and excretion rates can
model data. Shifts away from insect prey evident from cause herbivorous fishes to deviate significantly from the
SCA were reflected in increasingly enriched 13C isotope frequently cited 3–4% trophic-step enrichment applicable
data as both L. unicolor and H. fuliginosus increased in to more carnivorous species (Mill et al., 2007), an effect
size, with corresponding mixing model outputs demon- that could account for the anomalies in the trophic
strating progressively reduced importance of C3 riparian position estimates for S. parviceps. Scortum parviceps differs
vegetation (with low d13C). The ontogenetic dietary shifts from other sympatric terapontids not only in diet (Davis
revealed by both SIA and SCA were so profound in et al., 2011b; this study) but also in exhibiting many of the
H. fuliginosus and L. unicolor that different life stages ecomorphological features of a specialised herbivore, such
essentially acted as different trophic species. In contrast, as flattened dentition and a long intestine (Vari, 1978;
A. percoides, which exhibited small size–related shifts from Davis, Pusey & Pearson, 2012). Changes in trophic
its insect diet, exhibited relatively minor shifts in its enrichment in relation to ontogeny constitutes an infor-
organic carbon sources through its life history. However, mation gap in isotopic ecology and represents an issue of
there was little agreement between the SCA and SIA considerable significance in herbivores, many of which,
estimates of trophic position. Hephaestus fuliginosus and like S. parviceps, switch from carnivory to herbivory
L. unicolor, which exhibited no significant increases in (Carseldine & Tibbetts, 2005).
trophic position on the basis of SCA, demonstrated The variable allocation of assimilated elements to differ-
significant, although minor, increases in trophic position ent physiological processes via ‘isotopic routing’ has been
according to SIA. Scortum parviceps, the only species to suggested as another explanation for discrepancies

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

2168 A. M. Davis et al.
between ingested foods and isotopic values of different
Shifts in basal production sources and the effects of
tissues (Gannes, O’Brien & Martinez del Rio, 1997; Perga &
omnivory
Gerdeaux, 2005). While many studies have documented
ontogenetic relationships between body size and d13C and There is long-standing debate as to the dominant basal
d15N isotopes, these effects could result from changing carbon sources supporting tropical aquatic food webs
physiological allocation of isotopes or changes in tissue (Thorp & Delong, 2002; Douglas et al., 2005; Zeug &
turnover rates during ontogeny (Overman & Parrish, 2001), Winemiller, 2008; Lau, Leung & Dudgeon, 2009). This
although these phenomena are poorly described (Fry & study identified a range of autochthonous and alloch-
Arnold, 1982; Doucett et al., 1999). The fact that stomach thonous carbon sources supporting terapontid diets over
content data aligned closely with d13C isotopic changes for different stages of their life history, although C3
three of the four species in this study provides more vegetation was particularly important in the smaller
credence to isotopic changes in d13C being directly attrib- size classes. Species shifted away from these relatively
utable to dietary shifts. Isotopic change in ectotherms (such specialised juvenile diets to feeding across multiple
as fish) is dominated by the effects of growth, which slows carbon sources as they grew. It is increasingly recogni-
as fish age (Hesslein, Hallard & Ramlal, 1993; Suzuki et al., sed that single conceptual models of carbon dynamics
2005). Therefore, old slow-growing fish such as the large are not broadly applicable to different river systems,
size classes reported here may not respond isotopically habitat units or individual species (Hoeinghaus, Winem-
despite a recent change in diet, especially considering that iller & Agostinho, 2007; Zeug & Winemiller, 2008).
elimination of 15N-enriched nitrogen may occur slowly in Similarly, with ontogenetic dietary shifts being such a
muscle tissue of fish (MacNeil, Drouillard & Fisk, 2006). fundamental component of fish dietary ecology, a single
This latter point may be particularly relevant for species conceptual entity is unlikely to adequately represent a
such as S. parviceps that become more herbivorous as they given species at all times.
age, switching from a diet with a high N content and high Many of the size-related dietary shifts observed here
d15N (animal protein) to one with low N content and low occurred with only minor changes in trophic position.
d15N (plant protein). Even in species such as L. unicolor and H. fuliginosus
Another potential source of error in isotope-derived where isotope-based increases in tropic position were
trophic position estimates is the change in the relative evident, these changes were small (<0.5 trophic position)
fractionation rates of key prey species that drive size- and may well have been due to changes in fractionation
related diet shifts. Several studies have indicated that rates of prey rather than changes in trophic position per se.
d15N fractionation rates of vertebrates can be higher Highly variable tropical systems characterised by short
than those for invertebrates (Vander Zanden & Ras- food chains, omnivory and little size-based feeding
mussen, 2001; Vanderklift & Ponsard, 2003). A similar (Douglas et al., 2005; Layman et al., 2005b) may be
outcome was apparent in this study, with invertebrate conducive to major size-related trophic shifts coupled
prey across all trophic guilds demonstrating much with minimal apparent change in trophic position. This
lower d15N values than corresponding fish species in study demonstrates the use of SIA and d15N-derived
the same trophic guilds. While it could be argued that estimates of trophic position may yield limited, and
these results underline the value of SIA in revealing the potentially erroneous, insights into the nature and extent
true trophic position of consumers, it is also possible of ontogenetic diet shifts in many species. It is important
that the ontogenetic increases in d15N for species such to note that due to resourcing constraints, this study was
as H. fuliginosus and L. unicolor simply reflect higher based on samples from 1 year’s late dry season. Addi-
relative d15N fractionation rates of larger-bodied and tional research is required to make robust inferences
longer-lived prey (fish, macrocrustacea) that underpin about basal production sources supporting terapontid
ontogenetic changes in diet. Using d15N as a proxy for diets, as there may be substantial hydrology-mediated
trophic position may be valid for specialised predators shifts in the primary production sources (Zeug & Winem-
in relatively simple, linear food chains with significant iller, 2008; Jardine et al., 2012).
size-based feeding (see Post, 2003), but the habits of
more generalised carnivores and omnivores that feed
Limitations of mixing model approaches
among multiple trophic levels and include taxonomic
groups with varying fractionation rates may be expected While isotopic mixing models have emerged as a
to introduce considerable uncertainty into the assign- popular tool in isotopic studies, some caution is war-
ment of trophic levels.
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
 Isotopic ecology of some tropical fishes 2169
ranted in the interpretation of outputs. Mixing model fish. Additionally, the meaningful resolution of SIA is
outputs do not provide a definitive solution for the contingent upon primary food sources that exhibit distinct
organic matter sources supporting particular species, isotopic values. Because of the diversity of potential basal
instead outlining a distribution of mass-balanced solu- sources in riverine ecosystems, often with considerable
tions from a nominated set of possible contributions isotopic overlap, diets and functional roles of consumers
(Phillips & Gregg, 2003). The ultimate resolution of can be difficult to resolve solely on the basis of SIA. A
mixing model outputs is therefore dependent on the range of associated uncertainties regarding fractionation
isotopic distinctiveness of basal sources. When multiple rates through trophic levels, and the confounding role of
carbon sources demonstrate considerable overlap in ontogeny, further challenge meaningful interpretation of
isotopic value, as in this study, resolution of model SIA mixing model outputs. Due to the cosmopolitan
inputs can be limited. The apparent role of particular feeding habits of tropical species and wide-ranging
carbon sources may simply reflect their similarity in solutions emerging from mixing model predictions (par-
isotopic value to other contributing sources, rather than ticularly in larger size classes of several species), concur-
their real contributions to higher-level consumers. There rent use of SCA was valuable in disentangling the trophic
is also increasing theoretical and empirical evidence that complexity and specific structure of predator–prey inter-
generalist populations, which broadly utilise a wide actions. Results of this study also indicate that for certain
diversity of resources, are often composed of a heter- fish species, particularly those that exhibit pronounced
ogeneous assemblage of relatively specialised individu- size-related dietary shifts, the basal sources may change
als of the same species (Bolnick et al., 2007). While markedly with size. Ignoring the possibility of size-related
beyond the scope of this study, populations composed shifts in organic matter supporting fish species is a
of a collection of specialised individuals reliant on a simplistic and potentially flawed approach to definition
range of different carbon sources will certainly pose of aquatic food web function.
challenges for mixing models applied at a broader
site or population level (as in this study). Another of the
most cited caveats for mixing model analyses is their Acknowledgments
susceptibility to assumed TEFs that have not been
Fish were collected under a Queensland General Fisher-
validated at species or tissue levels (Caut et al., 2009;
ies Permit (No. 80302) and a James Cook University
Bond & Diamond, 2011). However, one of the major
Ethics Permit (A1391). The map of the Burdekin catch-
strengths of the SIAR platform is the capacity to
ment was constructed by Adella Edwards, James Cook
incorporate variation in TEFs into the model; the
University School of Earth and Environmental Science,
current study used large standard deviations of 0.5
with data provided by Geoscience Australia. We are
and 1.4& for nitrogen and carbon, respectively, likely
grateful to landowners for access to sites on their
leading to larger error estimates around source propor-
properties, and we thank the volunteers who assisted
tions.
in the field. Two anonymous reviewers provided
comments that greatly improved an earlier version of
Advantages of the dual approach the study manuscript.

Stable isotope analysis and SCA both demonstrated the

important role of ontogenetic dietary shifts in the trophic References
ecology of Burdekin River terapontid species. However,
Bolnick D.I., Svanback R., Araujo M.S. & Persson L. (2007)
each method had inherent limitations in defining trophic Comparative support for the niche variation hypothesis
shifts and elucidating food web structure. It has been long that more generalized populations also are more hetero-
appreciated that stable isotope ratios are most informative geneous. Proceedings of the National Academy of Sciences of the
when used in conjunction with stomach content analyses United States of America, 104, 10075–10079.
(Mantel et al., 2004; Layman et al., 2005a,b), but this dual Bond A. & Diamond A. (2011) Recent Bayesian stable-isotope
approach may be particularly relevant to tropical aquatic mixing models are highly sensitive to variation in discrim-
ecosystems. With the widespread omnivory and minor ination factors. Ecological Applications, 21, 1017–1023.
increases in trophic position through life history, even in Carseldine L. & Tibbetts I.R. (2005) Dietary analysis of the
species with marked ontogenetic dietary shifts, the results herbivorous hemiramphid Hyporhamphus regularis ardelio:
of SIA considered in isolation may provide minimal an isotopic approach. Journal of Fish Biology, 66, 1589–
1600.
insight into food web structure and the functional role of
 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172
2170 A. M. Davis et al.
Caut S., Angulo E. & Courchamp F. (2009) Variation in Hoeinghaus D.J., Winemiller K.O. & Agostinho A.A. (2007)
discrimination factors (d15N and d13C): the effect of diet Landscape-scale hydrologic characteristics differentiate
isotopic values and applications for diet reconstruction. patterns of carbon flow in large river food webs. Ecosys-
Journal of Applied Ecology, 46, 443–453. tems, 10, 1019–1033.
Davis A.M., Pearson R.G. & Pusey B.J. (2011b) Contrasting Hyslop E.J. (1980) Stomach content analysis – a review of
intraspecific dietary shifts in two terapontid assemblages methods and their application. Journal of Fish Biology, 17,
from Australia’s wet-dry tropics. Ecology of Freshwater Fish, 411–429.
21, 42–56. Jardine T.D., Pusey B.J., Hamilton S.K., Pettit N.E., Davies
Davis A.M., Pearson R.G., Pusey B.J., Perna C., Morgan D.L. P.M., Douglas M.M. et al. (2012) Fish mediate high food
& Burrows D. (2011a) Trophic ecology of northern Aus- web connectivity in the lower reaches of a wet-dry tropical
tralia’s terapontids: ontogenetic dietary shifts and feeding floodplain river. Oecologia, 168, 829–838.
classification. Journal of Fish Biology, 78, 265–286. Jennings S., Greenstreet S.P.R., Hill L., Piet G.J., Pinnegar J.K.
Davis A.M., Pusey B.J. & Pearson R.G. (2012) Trophic ecology & Warr K.J. (2002) Long term trends in the trophic
of terapontid fishes (Pisces: Terapontidae): the role of structure of the North Sea fish community: evidence from
morphology and ontogeny. Marine and Freshwater Research, stable isotope analysis, size-spectra and community met-
63, 128–141. rics. Marine Biology, 141, 1085–1097.
Davis A.M., Pusey B.J., Thorburn D.C., Dowe J.L., Morgan Jepsen D.B. & Winemiller K.O. (2002) Structure of tropical
D.L. & Burrows D. (2010) Riparian contributions to the diet river food webs revealed by stable isotope ratios. Oikos, 96,
of terapontid grunters in wet-dry tropical rivers. Journal of 46–55.
Fish Biology, 76, 862–879. Kennard M., Pusey B., Olden J., Mackay S., Stein J. & Marsh N.
DeNiro M.J. & Epstein S. (1978) Influence of diet on the (2010) Ecohydrological classification of natural flow regimes
distribution of carbon isotopes in animals. Geochimica to support environmental flow assessments: an Australian
Cosmochimica Acta, 42, 495–506. case study. Freshwater Biology, 55, 171–193.
Doucett R.R., Booth R.K., Power G. & McKinley R.S. (1999) Kilham S.S., Hunte-Brown M., Verburg P., Pringle C.M.,
Effects of spawning migration on the nutritional status of Whiles M.R., Lips K.R. et al. (2009) Challenges for inter-
anadromous Atlantic salmon (Salmo salar): insights from preting stable isotope fractionation of carbon and nitrogen
stable isotope analysis. Canadian Journal of Fisheries and in tropical aquatic ecosystems. Verhandlungen Internationale
Aquatic Sciences, 56, 2172–2180. Vereinigung für Theoretische und Angewandte Limnologie, 30,
Douglas M.M., Bunn S.E. & Davies P.M. (2005) River and 749–753.
wetland food webs in Australia’s wet-dry tropics: general Lau D.C.P., Leung K.M.Y. & Dudgeon D. (2009) What
principles and implications for management. Marine and does stable isotope analysis reveal about trophic rela-
Freshwater Research, 56, 329–342. tionships and the relative importance of allochthonous
Folch J., Lees M. & Stanley G.H.S. (1957) A simple method for and autochthonous resources in tropical streams? A
the isolation and purification of total lipids from animal synthetic study from Hong Kong. Freshwater Biology, 54,
tissues. Journal of Biological Chemistry, 226, 497–509. 127–141.
Fry B. & Arnold C.R.E. (1982) Rapid 13C ⁄ 12C turnover during Layman C.A., Winemiller K.O., Arrington D.A. & Jepsen D.B.
growth of brown shrimp (Penaeus aztecus). Oecologia, 54, (2005b) Body size and trophic position in a diverse tropical
200–204. food web. Ecology, 86, 2530–2535.
Galván D.E., Sweeting C.J. & Reid W.D.K. (2010) Power of Layman C.A., Winemiller K.O. & Arrington D.A. (2005a)
stable isotope techniques to detect size-based feeding in Describing the structure and function of a Neotropical
marine fishes. Marine Ecology Progress Series, 407, 271– river food web using stable isotopes, stomach contents, and
278. functional experiments. In: Dynamic Food Webs: Multispecies
Gannes L.Z., O’Brien D.M. & Martinez del Rio C. (1997) Assemblages, Ecosystem Development and Environmental
Stable isotopes in animal ecology: assumptions, caveats, Change (Eds P.C. de Ruiter, V. Wolters & J.C. Moore), pp.
and a call for more laboratory experiments. Ecology, 78, 395–406. Elsevier, Amsterdam.
1271–1276. Logan J.M., Jardine T.D., Miller T.J., Bunn S.E., Cunjak R.A. &
Hawking J.H. & Smith F.J. (1997) Colour Guide to the Lutcavage M.E. (2008) Lipid corrections in carbon and
Invertebrates of Australian Inland Waters, 2460 pp. Cooper- nitrogen stable isotope analyses: comparison of chemical
ative Research centre for Freshwater Ecology, Albury, extraction and modeling methods. Journal of Animal Ecol-
NSW. ogy, 77, 838–846.
Hesslein R.H., Hallard K.A. & Ramlal P. (1993) Replacement Lowe-McConnell R.H. (1975) Fish Communities in Tropical
of sulfur, carbon, and nitrogen in tissue of growing broad Freshwaters. Longman Press, London.
whitefish (Coregonus nasus) in response to a change in diet MacNeil M.A., Drouillard K.G. & Fisk A.T. (2006) Variable
traced by d34S, d13C, and d15N. Canadian Journal of Fisheries uptake and elimination of stable nitrogen isotopes between
and Aquatic Sciences, 50, 2071–2076.

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

 Isotopic ecology of some tropical fishes 2171
tissues in fish. Canadian Journal of Fisheries and Aquatic Puckridge J.T., Sheldon F., Walker K.F. & Boulton A.J. (1998)
Sciences, 63, 345–353. Flow variability and the ecology of large rivers. Marine and
Mantel S.K., Salas M. & Dudgeon D. (2004) Food web Freshwater Research, 49, 55–72.
structure in a tropical forest stream. Journal of the North Pusey B.J., Arthington A.H. & Read M.G. (1998) Freshwater
American Benthological Society, 23, 728–755. fishes of the Burdekin River, Australia: biogeography,
Marshall S. & Elliott M. (1997) A comparison of univariate history and spatial variation in community structure.
and multivariate numerical and graphical techniques for Environmental Biology of Fishes, 53, 303–318.
determining inter- and intraspecific feeding relationships Pusey B.J., Arthington A.H., Stewart-Koster B. & Kennard
in estuarine fish. Journal of Fish Biology, 51, 526–545. M.J. (2010) Trophic generalism in freshwater fish assem-
Merritt R.W. & Cummins K.W. (1996) An Introduction to the blages of a hydrologically variable northern Australian
Aquatic lnsects of North America, 3rd edn. Kendall Hunt river. Journal of Fish Biology, 77, 731–753.
Publishing, Iowa, U.S.A. Pusey B.J., Kennard M.J. & Arthington A.H. (2004) Freshwater
Mill A.C., Pinnegar J.K. & Polunin N.V.C. (2007) Explaining Fishes of North-Eastern Australia. CSIRO Publishing, Col-
isotope trophic-step fractionation: why herbivorous fish lingwood, Australia.
are different. Functional Ecology, 21, 1137–1145. R Development Core Team. (2009) R: A Language and
Mittelbach G.G. & Persson L. (1998) The ontogeny of Environment for Statistical Computing. R Foundation for
piscivory and its ecological consequences. Canadian Journal Statistical Computing, Vienna. Accessed online 20 Septem-
of Fisheries and Aquatic Sciences, 55, 1454–1465. ber 2009 at: http://www.r-project.org/.
Overman N.C. & Parrish D.L. (2001) Stable isotope compo- Rasmussen J.B. (2010) Estimating terrestrial contribution to
sition of walleye: 15N accumulation with age and area- stream invertebrates and periphyton using a gradient-
specific differences in d13C. Canadian Journal of Fisheries and based mixing model for delta 13C. Journal of Animal Ecology,
Aquatic Sciences, 58, 1253–1260. 79, 393–402.
Parkyn S.M., Collier K.J. & Hicks B.J. (2001) New Zealand Rayner T.S., Pusey B.J., Pearson R.G. & Godfrey P.C. (2010)
stream crayfish: functional omnivores but trophic preda- Food web dynamics in an Australian Wet Tropics River.
tors? Freshwater Biology, 46, 641–652. Marine and Freshwater Research, 61, 909–917.
Parnell A., Inger R., Bearhop S. & Jackson A.L. (2010) Source Ross S.T. (1986) Resource partitioning in fish assemblages: a
partitioning using stable isotopes: coping with too much review of field studies. Copeia, 1986, 352–388.
variation. PLoS ONE, 5, e9672. Stoner A.W. & Livingston R.D. (1984) Ontogenetic patterns
Pearson R.G. (1991) Ecology of the Burdekin River. Australian in diet and feeding morphology in sympatric sparid
Centre for Tropical Freshwater Research Report No. 91 ⁄ 01. fishes from seagrass meadows. Copeia, 1984, 174–187.
James Cook University, Townsville, Queensland, 36 pp. Suzuki K.W., Kasai A., Nakayama K. & Tanaka M. (2005)
Perga M.E. & Gerdeaux D. (2005) ‘Are fish what they eat’ all Differential isotopic enrichment and half life among tissues
year round? Oecologia, 144, 598–606. in Japanese temperate bass (Lateolabrax japonicus) juveniles:
Phillips D.L. & Gregg J.W. (2003) Source partitioning using implications for analyzing migration. Canadian Journal of
stable isotopes: coping with too many sources. Oecologia, Fisheries and Aquatic Sciences, 62, 671–678.
136, 261–269. Thorp J.H. & Delong A.D. (2002) Dominance of autochtho-
Phillips D.L., Newsome S.D. & Gregg J.W. (2005) Combining nous autotrophic carbon in food webs of heterotrophic
sources in stable isotope mixing models: alternative meth- rivers. Oikos, 96, 543–550.
ods. Oecologia, 144, 520–527. Townsend C.R. & Hildrew A.G. (1994) Species traits in
Pinnegar J.K. & Polunin N.V.C. (1999) Differential fraction- relation to a habitat templet for river systems. Freshwater
ation of d13C and d15N among fish tissues: implications for Biology, 31, 265–275.
the study of trophic interactions. Functional Ecology, 13, Vander Zanden M.J., Cabana G. & Rasmussen J.B. (1997)
225–231. Comparing the trophic position of freshwater fish calcu-
Poff N.L. & Allan J.D. (1995) Functional organisation of lated using stable nitrogen isotopes (dl5N) and literature
stream fish assemblages in relation to hydrological vari- dietary data. Canadian Journal of Fisheries and Aquatic
ability. Ecology, 76, 606–627. Sciences, 54, 1142–1158.
Post D.M. (2002) Using stable isotopes to estimate trophic Vander Zanden M.J. & Rasmussen J.B. (2001) Variation in d15N
position: models, methods, and assumptions. Ecology, 83, and d13C trophic fractionation: implications for aquatic food
703–718. web studies. Limnology and Oceanography, 46, 2061–2066.
Post D.M. (2003) Individual variation in the timing of ontoge- Vanderklift M.A. & Ponsard S. (2003) Sources of isotopic
netic niche shifts of largemouth bass. Ecology, 84, 1298–1310. variation in consumer-diet d15N: a meta-analysis. Oecologia,
Post D.M., Layman C.A., Arrington D.A., Takimoto G., 74, 231–235.
Quattrochi J. & Montaña C.G. (2007) Getting to the fat of Vari R.P. (1978) The terapon perches (Percoidei: Teraponi-
the matter: models, methods and assumptions for dealing dae): a cladistic analysis and taxonomic revision. Bulletin of
with lipids in stable isotope analyses. Oecologia, 152, 179–189. the American Museum of Natural History, 159, 175–340.

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

2172 A. M. Davis et al.
Werner E.E. & Gilliam J.F. (1984) The ontogenetic niche and prey items (grey circles) and terapontid fishes (open
species interactions in size structure populations. Annual circles) collected from Upper Burdekin aquatic habitats.
Review of Ecology and Systematics, 15, 393–415. Figure S2. Percentage of stomach contents contributed
Winemiller K.O. (1990) Spatial and temporal variation in by highly 13C depleted invertebrate families (Diptera,
tropical fish trophic networks. Ecological Monographs, 60, Ephemeroptera, Trichoptera and Lepidoptera larvae)
331–367.
plotted against standard length for four Burdekin River
Winemiller K.O. (2004) Floodplain river food webs: general-
terapontids.
izations and implications for fisheries management. In:
Proceedings of the Second International Symposium on the
Table S1. Estimated trophic position values for prey
Management of Large Rivers for Fisheries Volume II (Eds R. categories used in calculation of trophic position for
Welcomme & T. Petr), pp. 285–309. Regional Office for Asia Burdekin River terapontids.
and the Pacific, Bangkok, Thailand. Table S2. Numbers (and size range, mm SL) of each
Winemiller K.O., Akin S. & Zeug S.C. (2007) Production terapontid species collected at each site in the Burdekin
sources and food web structure of a temperate tidal catchment.
estuary: integration of dietary and stable isotope data. Table S3. Average d13C and d15N isotope values for
Marine Ecology Progress Series, 343, 63–76. basal sources at 12 upper Burdekin River sites.
Winemiller K.O., Zeug S.C., Robertson C.R., Winemiller B.K. Table S4. Stable Isotope Analysis in R (SIAR) modeling
& Honeycutt R.L. (2011) Food web structure of coastal summaries for terapontid ontogenetic trophic units
streams in Costa Rica revealed by dietary and stable
(OTUs) according to site.
isotope analyses. Journal of Tropical Ecology, 27, 463–476.
As a service to our authors and readers, this journal
Zeug S.C. & Winemiller K.O. (2008) Evidence supporting the
importance of terrestrial carbon in a large-river food web.
provides supporting information supplied by the authors.
Ecology, 89, 1733–1743. Such materials are peer-reviewed and may be re-orga-
nized for online delivery, but are not copy-edited or
typeset. Technical support issues arising from supporting
Supporting Information information (other than missing files) should be ad-
Additional Supporting Information may be found in the dressed to the authors.
online version of this article:
Figure S1. Stable isotope values (mean ± 1 SD for d13C (Manuscript accepted 27 June 2012)
versus d15N) of potential primary sources (solid circles),

 2012 Blackwell Publishing Ltd, Freshwater Biology, 57, 2156–2172

View publication stats