Trophic ecology of euryhaline and coastal elasmobranchs in northern Australia

Sharon Louise Every

A thesis submitted for the degree of Doctor of Philosophy at Charles Darwin University and The Australian National University

12th of May 2017

© Copyright by Sharon Louise Every

All Rights Reserved

Plate 1 South Alligator River 2014

ii

Plate 2 Speartooth Shark Glyphis glyphis

Declaration by Author

This work contains no material which has been accepted for the award of any other degree or diploma in any university or other tertiary institution and, to the best of my knowledge and belief, contains no material previously published or written by another person, except where due reference has been made in the text.

I give consent to this copy of my thesis, when deposited in the Libraries of my two Universities, being made available for loan and photocopying online via the

Universities open access repositories.

I have received copyright permissions from the Journal of Experimental Marine

Biology and Ecology and Marine Ecology Progress Series to use my publications in this thesis.

Sharon Every

12th of May 2017

iii Acknowledgements

Thank you to the Bininj and Mungguy people for allowing me to conduct this research on their traditional country.

I would like to thank my principal supervisor, Associate Professor David Crook who guided the formation of key ecological questions, provided expertise in aquatic and fish ecology, kept me on track and provided critical support and guidance throughout my candidature. Also to my co-supervisors; Associate

Professor Christopher Fulton who helped direct me to consider key areas of ecological theory, and what general philosophy I was trying to address in this thesis as well as statistical assistance and ensuring that my writing was succinct,

Dr. Heidi Pethybridge for her guidance within the laboratory and introducing me to biotracer techniques, Dr. Peter Kyne for teaching me fieldwork practice, and the ecology and biology of elasmobranchs. I thank all my supervisors for the timely feedback, encouragement and assistance. Thanks must also go to

Professor Karen Edyvane and Dr. Emile Ens for assisting at the beginning of my

PhD.

Thank you to Dr. David Parry who accepted my scholarship to North Australia

Marine Research Association, then to Dr. Edward Butler, the current director of

NAMRA who has assisted and mentored me throughout my PhD candidature. By being a part of NAMRA I was fortunate to be made welcome as a part of the

Australian Institute of Marine Science (AIMS) and the Arafura Timor Research

Facility. I thank all my colleagues, particularly Dr. Claire Streten for providing invaluable advice and feedback.

iv

To Dr. Thor Saunders and Northern Territory Fisheries for their expertise and equipment, particularly to Grant Johnson for assisting in the field and his guidance in the collection of elasmobranchs. I also appreciate all the people that assisted in the field including Mark Grubert, Duncan Buckle, Roy Tipagola,

Kirsty McAllister and Crew of Solander RV, Dominic Valdez and Francisco

Zillamarin and volunteers.

I thank the Commonwealth Scientific Industrial Research Organization (CSIRO) for allowing me the use of Dr. Peter Nichols lipid laboratories. Dr. Nichols also provided invaluable assistance with lipid analysis along with Dr. Peter Mansuer.

I also wish to thank my relatives for allowing me to stay in their homes in

Hobart whilst working at CSIRO.

I would like to thank my late father Capt. Mark Every and mother Lesley who sparked and encouraged my interested in marine life by taking my brother and I snorkelling and diving and allowing us free rein at the beach to explore rock pools. I also thank my mother for her continuous support throughout my PhD candidature.

v Table of Contents Declaration by author ...... iii Acknowledgements ...... iv Table of Contents ...... vi Abbreviations ...... viii List of Figures ...... xi List of Tables ...... xiii Abstract ...... 1 1 General Introduction ...... 3 1.1 Trophic ecology ...... 3 1.1.1 Food webs...... 3 1.2 Trophic cascades and keystone species ...... 5 1.3 Ecological Niche...... 6 1.4 Trophic ecology of cartilaginous fishes ...... 8 1.5 Trophic specialisation ...... 10 1.6 Measuring trophic niche and resource overlap ...... 12 1.6.1 Stable carbon and nitrogen isotopes as trophic indicators ...... 13 1.6.2 Stable isotope analysis of elasmobranchs ...... 15 1.6.3 Fatty acids as food web tracers ...... 17 1.6.4 Fatty acids in elasmobranchs ...... 19 1.7 Study Site: South Alligator River ...... 20 1.8 Thesis Outline & Aims ...... 28 2 Comparison of fin and muscle tissues for analysis of signature fatty acids in tropical euryhaline sharks ...... 32 2.1 Abstract ...... 33 2.2 Introduction ...... 34 2.3 Methods ...... 36 2.3.1 Ethics statement ...... 36 2.3.2 Tissue sampling and preparation ...... 37 2.1.1 Lipid and fatty acid extraction ...... 38 2.3.3 Statistical analyses ...... 39 2.4 Results ...... 40 2.4.1 Intraspecific tissue differences ...... 42 2.4.2 Interspecific differences ...... 48 2.5 Discussion ...... 50 2.6 Conclusion ...... 55 2.7 Acknowledgments ...... 57 3 Niche metrics suggest euryhaline and coastal elasmobranchs provide trophic connections among marine and freshwater biomes in northern Australia ...... 58 3.1 Abstract ...... 59 3.2 Introduction ...... 60 3.3 Methods ...... 64 3.3.1 Site description, sample collection and preparation ...... 64 3.3.2 Stable isotopes ...... 69 3.3.3 Fatty acids ...... 71 3.3.4 Statistical Analysis ...... 73 3.3.5 Niche area calculations ...... 74

vi 3.4 Results ...... 76 3.4.1 Stable isotope values, trophic position estimates, and niche metrics ...... 76 3.4.2 Fatty acid biomarkers and niche metrics ...... 84 3.4.3 Trophic niche metric comparisons ...... 88 3.5 Discussion ...... 89 3.6 Acknowledgments ...... 97 4 A seasonally dynamic estuarine ecosystem provides a diverse prey base for elasmobranchs ...... 97 4.1 Abstract ...... 99 4.2 Introduction ...... 100 4.3 Methods ...... 102 4.3.1 Elasmobranch and potential prey collection...... 102 4.3.2 Tissue sampling & preparation ...... 104 4.3.3 Stable Isotope Analysis ...... 105 4.3.4 Fatty Acid Analysis ...... 106 4.3.5 Assessment of food web structure ...... 107 4.3.6 Isotope mixing models of sharks and putative prey taxa ...... 109 4.3.7 Fatty acid prey-predator linkages ...... 111 4.3.8 Individual specialisation of fatty acids ...... 112 4.4 Results ...... 112 4.4.1 Food web structure and connectivity among putative prey taxa ...... 112 4.4.2 Trophic linkages between sharks and putative prey taxa...... 119 4.4.3 Shark intraspecific differences ...... 123 4.5 Discussion ...... 123 4.5.1 Seasonal and spatial patterns of trophic range and dietary diversity among putative prey taxa ...... 124 4.5.2 Links between putative prey to elasmobranchs ...... 125 4.6 Conclusions ...... 129 4.7 Acknowledgements ...... 130 5 General Discussion ...... 131 5.1 Assessment of biomarkers ...... 133 5.1.1 Source tissues for trophic biomarkers in elasmobranchs ...... 133 5.1.2 Trophic biomarkers within coastal and euryhaline elasmobranchs and their putative prey species ...... 134 5.2 Trophic niche of coastal and euryhaline elasmobranchs ...... 137 5.2.1 Trophic position ...... 138 5.2.2 Resource partitioning and linking freshwaters with coastal biomes ...... 139 5.2.3 Intraspecific differences within the coastal and euryhaline elasmobranchs .... 142 5.3 Ecological role of coastal and euryhaline elasmobranchs ...... 144 5.4 Potential impacts to euryhaline and coastal elasmobranchs food webs due to changes in the conditions of tropical riverine environments ...... 144 5.5 Future Directions ...... 147 6 References ...... 150 7 Appendix ...... 187 7.1 Elasmobranch fatty acid data ...... 187 7.2 Stable Isotope niche ...... 189 7.3 Fatty acid niche overlap ...... 190 7.4 Indicator FAs from Legendre indicator species analysis ...... 191 7.5 Putative prey fatty acid data ...... 193 7.6 Principal coordinate ordination of fatty acids within putative prey ...... 194

vii Abbreviations °C Degrees Celsius

G Delta

µg Microgram

µl Microlitre

Z Omega

ANCOVA Analysis of covariance

ANOSIM Analysis of similarity

ANOVA Analysis of variance

ARA Arachidonic acid

CD Centroid distance

CR G13C range cf Compare with cm Centimetre

DCM Dichloromethane

DDF Diet discrimination factors

DHA Docosahexaenoic acid

DPA Docosapentaenoic acid

DSQ GC-MS Dual stage quadruple gas chromatography-mass spectrometry

EFA Essential fatty acid

EPA Eicosapentaenoic acid

FA Fatty acid

FAA Fatty acid analysis

FAME Fatty acid methyl esters

FALD Fatty aldehyde analysed as dimethyl acetal g Gram

GC Gas chromatograph

GC-MS Gas chromatography-mass spectrometry

HCl Hydrochloric acid

viii Inc Incorporation

IUCN International union for conservation of nature km Kilometre

LA Linoleic acid labdsv Ordination and Multivariate Analysis for Ecology

LSC Lyo screen control

Max Maximum

MCMC Markov chain Monte Carlo mg Milligram mL Millilitre

MNND Mean nearest neighbour distance mm Millimetre

Min Minimum min Minutes

MeOH Methanol nMDS Non-metric multidimensional scaling

MUFA Monounsaturated fatty acids

N Nitrogen n Number

NaCl Sodium Chloride

NR G15N range

PCA Principal coordinates analysis

PCO Principal component analysis

PERMANOVA Permutational multivariate analysis of variance

PERMDISP Homogeneity of dispersion test

PNG Papua New Guinea ppm Parts per million

PUFA Polyunsaturated fatty acid

QFASA Quantitative fatty acid signature analysis

ix QGIS Quantum Geographic Information System

NT Northern Territory

RInSP R individual specialisation package rpm Revolutions per minute

SI Stable isotopes

SIA Stable isotopes analysis

SIAR Stable Isotopes Analysis in R

SIBER Stable Isotope Bayesian Ellipses in R

SDNND the standard deviation of mean nearest neighbour distance

SIMPER Similarities of variance

SFA Saturated fatty acids

SD Standard deviation

SE Standard error

SEAC Ellipse area of stable isotopes

TA Total Area of stable isotope convex hulls

TFA Total fatty acids

TL Total length

TLE Total lipid extract

TNW Total niche width

TP Trophic position

UK United Kingdom

USA United States of America v/v Volume/volume percent

WIC Within individual variation

x List of Figures Figure 1.1 Global map of total marine chondrichthyan species richness (reproduced from: Dulvy et al. 2014) 9 Figure 1.2 Representation of the differences between feeding types in consumer populations. (Bearhop et al. 2004) 11 Figure 1.3 Map of Northern Territory, Australia, showing Kakadu National Park (White line indicates the approximate park boundaries), and the rivers within, including the study site, the South Alligator River. Inset shows this region in relation to the rest of Australia. 22 Figure 2.1 Comparison of the relative means (± standard deviation) of (A) saturated, (B) monounsaturated, and (C) polyunsaturated fatty acid profiles based on fin and muscle tissues taken from three shark species (Carcharhinus leucas, Glyphis garricki and G. glyphis) from the South Alligator River, Kakadu National Park, Australia. 43 Figure 2.2 Comparison of the fatty acid (a) 20:5Z3, (b) 20:4Z6, (c) 22:4Z6 and (d) 20:3Z6 (%) relative means (± standard deviation) within fin and muscle tissues taken from three shark species (Carcharhinus leucas, Glyphis garricki and G. glyphis) from the South Alligator River, Kakadu National Park, Australia 45 Figure 2.3 Ordination (nMDS) of fatty acid profiles from the fin and muscle tissues of the three shark species (A) Carcharhinus leucas, (B) Glyphis garricki, (C) G. glyphis from the South Alligator River, Kakadu National Park, Australia. 47 Figure 2.4 % Contribution of fatty acids that caused the main differences between fin and muscle profiles from SIMPER analysis in (a) Carcharhinus leucas (b) Glyphis garricki and (c) G. glyphis from the South Alligator River, Kakadu National Park, Australia. 49 Figure 3.1 Map of the South Alligator River, Northern Territory, Australia, showing capture locations of elasmobranchs. Dotted lines represent the approximate boundary of each sampling site. Inset shows where the river is in relation to northern Australia. (Background map from Stamen Toner, and formed using QGIS Development Team, ȋʹͲͳͷȌȌǤ‡†ȋ„Ž—‡Ȍ–‡š–‹†‹ ƒ–‡•†”›ȋ™‡–Ȍ•‡ƒ•‘•ƒŽ‹‹–›ȋΩȌ˜ƒŽ—‡•Ǥ 68 Figure 3.2 Biplot of mean (±SD) G13C and G15N values for seven species of euryhaline and coastal elasmobranchs from South Alligator River, Kakadu National Park, Australia. Primary producer values along the x Ȃ axis are G13C ranges for seston, epiphytes and seagrass from the combined wet and dry season, Embley River Estuary, Gulf of Carpentaria (Loneragan et al., 1997). 77 Figure 3.3 (a) Bivariate plot of isotopic space depicting wet vs dry niche areas within standard ellipse (SEAC) of G13C and G15N values of Carcharhinus leucas (blue) Glyphis garricki (black), and Rhizoprionodon taylori (red) from Kakadu National Park, Australia (b) Bayesian confidence intervals of isotopic niche area; black dots represent their mode, shaded boxes represent the 50, 75 and 95% credible intervals from dark to light grey chance 83 Figure 3.4 Principal coordinates analysis of fatty acids in muscle tissue of Carcharhinus leucas, Glyphis garricki and Rhizoprionodon taylori, Carcharhinus amboinensis and Urogymnus dalyensis in both the wet and dry seasons from the South Alligator River, Kakadu National Park, Australia. 85 Figure 3.5 Comparisons of the posterior distributions of the probabilistic niche overlap metrics of 5 major essential fatty acids (EFAs) between 3 shark species: Carcharhinus leucas (blue), Glyphis garricki (black) and Rhizoprionodon taylori (red) from the South Alligator River, Kakadu National Park, Australia. Also shown on the bottom row is seasonal overlap for G. garricki (wet season = black, dry season = grey). Niche overlap is characterized by the probability that an individual from one shark species (or season) is found within the niche region of another shark species (or season). Posterior means and 95% credible intervals are displayed in light blue 86 Figure 3.6 Schematic of the essential fatty niche space and extent of overlap among three shark species, Carcharhinus leucas (blue), Glyphis garricki (black) and Rhizoprionodon taylori (red) in the South Alligator River, Kakadu National Park, Australia. Niche spaces are scaled relative to each other; central axis omitted. 88 Figure 3.7 (a) Bivariate plot of isotopic space depicting niche areas within standard ellipse (SEAC) of G13C and G15N of Carcharhinus leucas (blue) Glyphis garricki (black), and

xi Rhizoprionodon taylori (red) from Kakadu National Park, Australia (b) Bayesian confidence intervals of isotopic niche area. 189 Figure 3.8 (a) Ten random elliptical projections of trophic niche region for each elasmobranch and pair of major essential fatty acids (FA) and (b) G. garricki wet vs dry season major (elliptical plots). Also displayed are one-dimensional density plots (lines) and two- dimensional scatterplots to demonstrate normality. 190 Figure 4.1 Plot of mid-trophic species (black dots) (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Nemapteryx armiger, Polydactylus macrochir and Rhinomugil nasutus) G13C and G15N showing standard deviation error bars overlayed over the wet (coloured circles) and dry (coloured triangles) season shark (Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori) isotope values adjusted for trophic discrimination 115 Figure 4.2 Principal coordinate ordination of essential fatty acids (EFA)>0.05% in percentage abundance in prey and shark species (black circles surrounding symbol), with vector overlays of tŠ‡‘•–‹ˆŽ—‡–‹ƒŽ •ȋ‡ƒ”•‘ǯ• ‘””‡Žƒ–‹‘ƒ„‘˜‡ͲǤͳȌǤ 118 Figure 4.3 Box and whisker plots from MixSIAR of a) Seasonal difference of shark diet. b) Sharks and the proportion of the source (Prey) that makes up their diet. Prey grouped based on G13C and source of C estimated from Loneragan et al. (1997) estuarine = Rhinomugil nasutus, mid-river = Polydactylus macrochir + Lates calcarifer + N. armiger, freshwater = Macrobrachium equidens. 121 Figure 4.4 Principal coordinate ordination of essential fatty acids (EFA)>0.05% in percentage abundance in prey species, ™‹–Š˜‡ –‘”‘˜‡”Žƒ›•‘ˆ–Š‡‘•–‹ˆŽ—‡–‹ƒŽ •ȋ‡ƒ”•‘ǯ• correlation above 0.1). 194

xii List of Tables

Table 1.1 Species of elasmobranchs recorded in the South Alligator River, Northern Territory, —•–”ƒŽ‹ƒǡ™‹–Š•—ƒ”›‘ˆ”‡’‘”–‡††‹‡–ƒ”›‹–‡•ǤDz˜ƒŽ—‡•dz”‡ˆ‡”•–‘–Š‡—„‡”‘ˆ individuals examined in each of the cited studies. 21 Table 2.1 Number and total length (TL) of specimens from which samples of fin and muscle tissue were taken for fatty acid analysis in three shark species from the South Alligator River, Australia (Size range +/- SD). 37 Table 2.2 Comparisons of the relative abundance of fatty acids (FA) (mean % r standard deviation) between fin and muscle tissue in Carcharhinus leucas, Glyphis garricki and G. glyphis, from the South Alligator River, Australia. 41 Table 2.3 Paired t-tests comparing the concentrations of three major fatty acid (FA) classes detected within the fin and muscle tissues from each of three euryhaline shark species, Carcharhinus leucas, Glyphis garricki, and G. glyphis, from the South Alligator River, Australia. Significant (P<0.05) result shown in bold. 44 Table 2.4 Paired t-tests comparing the concentrations of four essential fatty acids detected within the fin and muscle tissues from each of three euryhaline shark species, Carcharhinus leucas, Glyphis garricki, and G. glyphis, from the South Alligator River, Australia. Significant (P<0.05) result shown in bold. 46 Table 3.1 Number (n) and total length (TL) of seven euryhaline and coastal elasmobranch species from the South Alligator River, Kakadu National Park, Australia from which muscle tissue samples were taken for stable isotope (SI) and fatty acid (FA) analysis. Included is the division of species between the wet and dry seasons and the sex ratio male to female. 66 Table 3.2 Mean (±SD) values of fatty acids (FA) (% mean of the relative abundance of FA> 0.5%), G13Ωƒ†G15Ω‹—• Ž‡–‹••—‡ˆ”‘•‡˜‡‡Žƒ•‘„”ƒ Š•’‡ ‹‡• ‘ŽŽ‡ –‡† from the South Alligator River, Kakadu National Park, Australia. Included are calculations ˆ‘”–”‘’Š‹ ’‘•‹–‹‘ȋȌǡ–‘–ƒŽƒ”‡ƒ‘ˆ•–ƒ„Ž‡‹•‘–‘’‡ ‘˜‡šŠ—ŽŽ•ȋΩ2), ellipse area of stable isotopes (SEAC Ω2) and essential fatty acids (EFA) niche space (shaded grey). SFA: saturated fatty acids; MUFA: monounsaturated fatty acids; PUFA: polyunsaturated fatty acids; wet:dry: wet and dry seasons 78 Table 3.3 ANOVA and ANCOVA (shaded) of stable isotope values, the ratio of Z3/Z6 and the fatty acid 18:2Z6 from the muscle of Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori from the South Alligator River, Kakadu National Park, Australia ‘’ƒ”‡†–‘–Š‡Dz™‡–dzƒ†Dz†”›dz•‡ƒ•‘ǡ™‹–Š–‘–ƒŽŽ‡‰–ŠȋȌ–Š‡ ‘˜ƒ”‹ƒ–‡Ǥ‹‰‹ˆ‹ ƒ– values are in bold. 79 Table 3.4 Isotopic overlap (%) of stable isotopes (SI) ellipses (SEAC) and convex hull (TA) of Carcharhinus leucas, Glyphis garricki and Rhizoprionodon taylori from the South Alligator River, Kakadu National Park, Australia. Also included are the probabilities of sharks being in the major essential fatty acids (EFA) niche space of each other with a set confidence interval (CI) of 95%. Ȃ no test completed 82 Table 3.5 Mean values for all fatty acids (FA) (% mean of the relative abundance of FA> 0.5%), and their standard deviation in muscle tissue from six elasmobranchs collected from the South Alligator River, Kakadu National Park, Australia. 187 Table 4.1 Number (n) and total length of 4 sharks (Every et al., 2017) and 6 putative prey species caught from the South Alligator River, Kakadu National Park, Australia from which muscle tissue samples were taken for stable isotope (SIA) and fatty acid analysis (FAA). Wet and dry species number, Total length (TL) (± standard deviation (SD)), sex ratio and habitat are also included. 103 Table 4.2 ƒ›‡ǯ•‡–”‹ •‘ˆ–Š‡‘—–ŠŽŽ‹‰ƒ–‘”‹˜‡”‹†-trophic taxa and shark species. Numbers of species (n) at each site and season are included, those with n values < 5 were omitted from these analysis (highlighted grey). TA = Total Area, NR = range of G15N, CR = range of G15N, CD= centroid distance, MNND =mean nearest neighbour distance, SDNND = standard deviation of nearest neighbour distance. 114 Table 4.3 Mean values for essential fatty acids (EFA> 0.5%), G13C and G14N and their standard deviation in muscle (SD) tissue from six potential prey species of shark collected from the

xiii South Alligator River, Kakadu National Park, Australia. Included are indicator FAs for each species with a p<0 .05. 116 Table 4.4 Comparison of species, season (wet and dry) and site locations (1 - 4) of essential fatty acids from the mid-taxa species (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Nemapteryx armiger, Polydactylus macrochir and Rhinomugil nasutus) in the South Alligator River, Kakadu, Australia using PERMANOVA. DF = degrees of freedom. 119 Table 4.5 Stable Isotope mixing model results from four species of sharks, Carcharhinus leucas, Glyphis garricki, G. glyphis and R. taylori in the South Alligator River, Australia. Results are the percentage mean proportion of the shark consuming from each prey group and the combined results over all prey groups and the difference between seasons ± standard deviation (SD). 122 Table 4.6 Fatty acids (FA) within prey (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Nemapteryx armiger, Polydactylus macrochir and Rhinomugil nasutus) and sharks species (Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori) found in the South Alligator River, Kakadu, Australia that can be used as indictors for a specific species according to a Dufrêne- Legendre indicator species analysis (Dufrêne & Legendre 1997). Note: The analysis was unable to find indicator FAs for P. macrochir and L. calcarifer. 191 Table 4.7 Mean values for fatty acids (FA> 0.5%) and their standard deviation in muscle (SD) tissue from six potential prey species of shark collected from the South Alligator River, Kakadu National Park, Australia. Included are indicator FAs for each species with a p <0 .05 193

xiv Abstract

‘™Ž‡†‰‡‘ˆƒ‘”‰ƒ‹•ǯ•–”‘’Š‹ ‡ ‘Ž‘‰›‹•‹’‘”–ƒ–so that the structure, function and predications of their ecosystems can be determined. Temporal and spatial variations in the trophic niches of organisms can be quantitatively examined to determine the essential resources for survival of species and how changes in the availability of resources affect ecosystem processes.

In this study I utilised biochemical tracers within muscle and fin tissue to examine the trophic ecology of coastal and euryhaline elasmobranchs (sharks and rays) within the South Alligator River, Northern Territory, Australia. Firstly,

I examined the suitability of different tissue types for profiling fatty acids (FA).

Both fin and muscle tissue contained important dietary FAs, although some discrepancies between FAs in tissue types suggested muscle tissue as the preferred option for use in this study.

I then explored elasmobranch dietary connections using FAs and stable isotopes

(SI) and through a comparison of putative prey species. Coastal and euryhaline elasmobranchs were distinguished by their trophic sources Ȃ river/estuary and marine Ȃ with varying degrees of trophic niche overlap among species. Putative prey species were generally found to exhibit omnivory and showed seasonal differences in their biochemical tracers. While most euryhaline elasmobranchs were generalist consumers, the different species tended to consume prey from either marine or freshwater sources.

1 Collectively, the analysis of biomarkers suggests that coastal and euryhaline elasmobranchs support large-scale cross-biome connectivity in tropical aquatic ecosystems. Consequently, changes in the movement and abundance of euryhaline elasmobranchs could have major implications for tropical ecosystem function and connectivity. Whilst this assemblage of elasmobranchs is relatively well protected in the World Heritage Kakadu National Park, similar assemblages elsewhere face multiple threats. It is therefore important that integrated management approaches, which work across marine-freshwater boundaries, are used to protect these high mobile and ecologically important components of tropical aquatic food webs.

2 1 General Introduction

1.1 Trophic ecology

Ecology is the study of the interactions between organisms and their environment that ultimately shape their distribution, abundance and biology

(Krebs 2009, Loreau 2009). The subfield of trophic ecology considers the dietary aspects of organismal interactions via predation and competition.

Understanding trophic ecology is necessary to determine how communities are organised, and how biomass, nutrients and energy are transferred within and among ecosystems (Winemiller & Polis 2013). Studying dietary aspects of populations within an assemblage (a group of two or more species that share geographic and temporal space) enables conservation and natural resource managers to detect changes in trophic dynamics, and plan ecosystem-based strategies to protect key functions, and assist the restoration of disturbed biomes (Polis 1994, Hooper et al. 2005, Stachowicz et al. 2007).

1.1.1 Food webs

The basic premise of food webs was first introduced by Elton (1927), who stated that species are arranged in food chains, combining to form food cycles limited by primary production by plants that is linked to solar energy and nutrient availability. Food webs are conceptual or quantitative constructs that attempt to describe the strength and direction of a network of feeding interactions between producers and consumers (Polis & Strong 1996, Dunne

2012). As such, food webs represent the transference of energy throughout a community via linear connections (food chains) between prey and consumers

(Polis & Strong 1996, Dunne 2012). Most food webs consist of four to five

3 trophic levels with rare, large-bodied apex predators at the highest trophic level, whilst primary producers and detrital pools are at the lowest level (Cohen et al. 1993, Montoya et al. 2006). However, in order to develop accurate models of energy flow through food webs, it is necessary to also consider indirect effects within food webs (e.g. predation effects of species other than their direct prey, top-down and bottom-up effects), multiple species interactions, and temporal dynamics in the relative abundance of producers and consumers

(Polis 1994, Montoya et al. 2006, Dunne 2012, Polis & Winemiller 2013).

The emergence of new biomarkers and other information on the trophic ecology of organisms has allowed for detailed understanding of complex food webs, such as host parasitoids and mutualistic food webs (Ings et al. 2009), while more Ǯclassicǯfood webs that focus on direct trophic linkages have been advanced by filling gaps in dietary information. At the core of these advances is the critical examination of species interactions (Polis & Strong 1996), and measures of turnover that include the metabolic rates and body sizes of species

(Bascompte 2009). These models must also account for species that may occupy several trophic levels, which can set up multiple interactions among and between trophic levels (Polis & Strong 1996). Classic food webs are valuable when species-specific data is required at a community level, when considering the links between herbivores and consumers and quantifying energy flows

(Polis 1994, Ings et al. 2009). Characterising food webs is important for ecosystem based management practice, as accurate interpretation of trophic linkages combined with other ecosystem data (including human usage) is a requirement for these food web models (Curtin & Prellezo 2010).

4 1.2 Trophic cascades and keystone species

The removal of a specific species from a community can have significant and often cascading effects on food webs, such as dramatic shifts in relative abundance and species richness (Paine 1969, Estes et al. 2011). Early work on cascading effects focussed on changes caused from the reduction of producers resulting in the subsequent reduction of consumers at each trophic level Ȃ known as bottom-up effects (Elton 1927, Lindeman 1942). However, it was suggested that herbivores would rarely be food limited, and so their biomass must be largely controlled by consumers via top-down effects (Hairston et al.

1960). Therefore, consumers could enable producers to flourish by limiting herbivore populations via top down effects (Hairston et al. 1960). More recently these fundamental questions have highlighted the complexity of food webs as both top-down and bottom-up effects have been demonstrated to work concurrently. For example, a four decade study in the North Sea discovered that bottom-up processes occurred in this ecosystem resulting from sea temperature increases causing changes in planktonic groups, whilst top-down effects were occurring due to shifts in higher-order consumers via commercial fishing causing a trophic cascade (Lynam et al. 2017).

Top-down effects are created within food webs when certain keystone species

(often apex predators) strongly influence connections among the entire food web. Keystone species were first identified by Paine (1969), who experimentally removed the predatory seastar Pisaster sp. along sections of the rocky intertidal community on the Pacific Northwest coast of the USA (Paine

1966). His landmark work revealed how these predatory species shaped and

5 stabilised the distribution and abundance of their primary prey (mussels), which in turn, affected competitive interactions for space and broader patterns of diversity in this ecological community (Paine 1966). Variation in keystone species can also change due to shifts in prey availability of another influential predator. For example, an increase in abundance of the mesopredator Sea otter

Enhydra lutris indirectly caused the reduction of kelp growth through predation of sea urchins. However, during tŠ‡ͳͻͻͲǯs an increase in predation on E. lutris from Killer Whales Orcinus orcas caused a severe decline in E. lutris abundance, releasing predation pressure of sea urchins, increasing sea urchin abundance and therefore an increase in grazing on the primary producer, kelp (Estes

1998).

Numerous examples of such trophic cascades have been found across multiple ecosystems as a result of the loss of keystone species, including wolves Canis lupus, grizzly bears Urtsus arctos (Berger et al. 2001, Winnie 2012), and dingoes

Canis lupus dingo (Letnic et al. 2012).

1.3 Ecological niche

The niche concept describes the essential abiotic and biotic conditions an organism requires for their survival (Hutchinson 1957, Chase & Leibold 2003), and how biotic interactions can affect access to their necessary resources. When multiple axes of environmental conditions and resources are considered, a theoretical multivariate hyper-volume can be constructed for species niches

(Hutchinson 1957, Holt 2009). This niche concept can provide a framework for

6 examining how the trophic ecology of species can shape their role in community food webs (Layman et al. 2007). Species that have similar niches may compete with each other for resources, and as a result, behavioural or other phenotypic adaptations may evolve to reduce niche overlap and competition (Hutchinson

1957, Chase & Leibold 2003, Hirzel & Le Lay 2008, Krebs 2009, McInerny &

Etienne 2012). These adaptations lead to interspecific differences such as spatial or temporal differences in feeding times (Platell & Potter 2001). Species may also evolve intraspecific differences as a result of density-dependent competition amongst individuals, which can lead to differences in feeding preferences and ontogenetic differences (Bolnick et al. 2011).

In terms of ecological theory, niches are often considered as either

Ǯˆ—†ƒ‡–ƒŽǯ‘”Ǯ”‡ƒŽ‹•‡†ǯ (Hutchinson 1957, Krebs 2009). The fundamental nich‡†‡• ”‹„‡•–Š‡–‘–ƒŽ’‘••‹„Ž‡Ǯecological spaceǯ that a species can inhabit given resource availability (Hutchinson 1957, Pulliam 2000, Araújo & Guisan

ʹͲͲ͸ǡ ‹”œ‡ŽƬ‡ƒ›ʹͲͲͺȌǤ ”‡ƒŽ‹–›ǡŠ‘™‡˜‡”ǡƒ•’‡ ‹‡•ǯ†‹•–”‹„—–‹‘‹•ƒŽ•‘ shaped by competition for limited resources with other species. This imposes limitations on the availability of resources, resul–‹‰‹Ǯ”‡ƒŽ‹œ‡†‹ Š‡•ǯthat are narrower than the fundamental niche.

Ǯ‹ Š‡™‹†–Šǯ is used to describe –Š‡˜ƒ”‹‡–›‘ˆ”‡•‘—” ‡•‹ƒǮ‡ ‘Ž‘‰‹ ƒŽ•’ƒ ‡ǯ that a species exploits within a given ecosystem. Niche width can be constrained by intraspecific competition and resources (Roughgarden 1972) or increased by facilitation (where both species benefit from their interaction) (Bruno et al.

2003). Niche width can also be affected by feeding preferences of organisms and

7 interactions between species and individuals. Ultimately, such factors influence the habitat preferences, morphology, movement patterns and distribution of species (Van Valen 1965, Bearhop et al. 2004). Although sympatric species may overlap in their niches to varying degrees, individuals may alter their resource use to reduce competition via temporal, morphological, prey switching or ontogenetic shifts.

1.4 Trophic ecology of cartilaginous fishes

Fishes of the Class Chondrichthyes (cartilaginous fishes) are widespread in aquatic environments, with around 1200 species identified (Weigmann 2016), many of which are apex or meso predators in marine waters. Approximately 90 chondrichthyans are either obligate freshwater or euryhaline (moving between freshwater and marine areas) species (Dulvy et al. 2014), and about one third of chondrichthyans are of conservation concern (Dulvy et al. 2014) due to overfishing, unmanaged bycatch and habitat destruction (Stevens et al. 2000,

Dulvy et al. 2014). Life history traits such as late maturation and low fecundity render many chondrichthyans at high risk of population depletion, especially when faced with multiple, often synergistic, threats (Stevens et al. 2000, Dulvy et al. 2014). The ecological implications of this vulnerability are likely to be profound, as many chondrichthyans are thought to be high to middle order predators that may play critical roles in shaping community interactions, food web structure, and the flow of energy and biomass among biomes (Myers &

Worm 2003, Young et al. 2015).

8 Elasmobranchs (sharks and rays) are the main representative of the Class

Chondrichthyes in tropical waters. Of the approximately 465 species that occur in tropical waters, the Indo-West Pacific region has the highest diversity (301 species), of which over 80% are endemic; other tropical regions have less than 30% of this level of species richness (e.g., Eastern Pacific = 81, Western

Atlantic = 95, Eastern Atlantic = 70; White & Sommerville 2010, Dulvy et al.

2014; Fig. 1.1).

Figure 1.1 Global map of total marine chondrichthyan species richness (reproduced from: Dulvy et al. 2014)

Increasing our understanding of the trophic ecology of elasmobranchs is crucial to their conservation as it can assist in identifying critical habitat and their relationships with other species (Heithaus et al. 2008, Hammerschlag et al.

2010). As higher-order predators, elasmobranchs often influence ecosystems by exerting top-down control on prey species. Reductions in the abundance of these predators can impact biomes by shortening food chains (Estes et al. 2011) and initiating trophic cascades where abundances of lower-order predators

9 and/or herbivores increase as predation pressure is removed (Heithaus et al.

2008, Ruppert et al. 2013). For example, over-fishing of large predators such as sharks and herbivorous fish in coral reefs in the Caribbean lead to an increase in algal cover of coral that ultimately resulted in large-scale mortality of coral

(Hughes 1994, Bascompte et al. 2005).

1.5 Trophic specialisation in elasmobranchs

Adding to the complexity of trophic ecology are the differences in trophic niches among species and among individuals of the same species. For example, some species appear to be generalists at a population level, with a wide range of food items consumed. However, when examined at an individual level, there may be strong specialisation in diet among individuals (Bearhop et al. 2004). According to Bearhop et al. (2004), consumers can be classified as either Dz•’‡ ‹ƒŽ‹•–•dzor one of –™‘–›’‡•‘ˆDz‰‡‡”ƒŽ‹•–dz (Fig. 1.2). Specialists are species that consume a narrow range of prey types whereas generalists consume a much wider range of prey items. Within the generalists, Type A species are those in which all individuals consume a wide range of prey, whilst Type B generalists are species in which individuals within a population specialize on different prey from the same available range of prey items. Intraspecific differences in diet can result from ontogenetic differences and sexual dimorphism (Schoener 1983), but genetic and environmental influences on ecological traits may also play roles in niche variation within a species (Price 1987, Fry et al. 1999, Bolnick et al. 2002,

Hammerschlag-Peyer et al. 2011, Araújo et al. 2011).

10

Figure 1.2 Representation of the differences between feeding types in consumer populations. (Bearhop et al. 2004)

Determining the degree of specialisation among and within species is important, as generalist species tend to be relatively resistant to environmental change because they are able to consume a variety of food sources, depending on their availability, whilst specialists with restricted dietary preferences tend to be more vulnerable (Bolnick et al. 2003).

Early investigations of elasmobranch diet indicated that many species

(particularly those from the Order Carcharhiniformes) are generalists

(Wetherbee & Cortés 2004). However, more recent research has shown that intra-specific variation in diet may be widespread among the apex predators

(Araújo et al. 2011), whilst some species have been shown to have highly specialised diets (e.g. Australian Weasel Shark Hemigaleus australiensis consume 97% benthic octopus; Taylor & Bennett 2008). This research shows that understanding of the trophic ecology of cartilaginous fishes, and their roles in food webs, requires consideration of both inter- and intra-specific variation in diet.

11

1.6 Measuring trophic niche and resource overlap

To compare and measure trophic niche and resource overlap, the variables that define the trophic niche of species need to be identified and quantified across the spectrum of dietary items they may consume over space and time (Bearhop et al. 2004). Measurements of trophic niches for investigating the diet of aquatic species has largely involved methods such as stomach content analysis or gastric lavage that involve major or fatal interventions with study individuals

(e.g. Cortés 1997, Wetherbee & Cortés 2004, Barnett et al. 2010). Early elasmobranch dietary studies were limited to prey lists of their stomach contents (e.g. Sadowsky 1971), which was later expanded to include the relative importance of prey items (Hyslop 1980, Cortés 1997). However, these studies were often lethal and required large numbers of individuals to meet statistical requirements (Hyslop 1980). Moreover, stomach content samples presented only a small temporal snapshot of patterns of consumption (Shiffman et al.

2012). Elasmobranchs are also prone to regurgitating their stomach contents when caught (Arrington et al. 2002), which can affect the efficacy of these traditional contents-based techniques. Similarly, different rates of digestion can make the taxonomic identification of prey species difficult (Hyslop 1980). Since dietary studies are temporally constrained and may require the death of elasmobranchs, such approaches are not generally appropriate or ethical for species of conservation concern (Hammerschlag & Sulikowski 2011).

12 The development of biochemical tracers (e.g. stable isotopes (SI) and fatty acids

(FA)) has allowed for some recent advances in how we estimate and understand trophic resource use and connections in ecological communities (Watt &

Ferguson 2015, Young et al. 2015, Rude et al. 2016). A key advantage of these methods is being able to quantify resource use across a period of days to years, depending on the rates of biomarker turnover in the tissue(s) being tested.

Further, given that only small amounts of tissue are often required for analysis, biochemical tracers require non-fatal and relatively less intervention with study individuals than traditional approaches to dietary analysis. For the effective use of biochemical tracers in trophic dynamics, biochemical tracers must be

‹†‡–‹ˆ‹‡†™‹–Š‹–Š‡ ‘•—‡”ǯ•ˆ‘‘†•‘—” ‡ƒ†„‡†‹‰‡•–‡†ƒ†ƒ„•‘”„‡†‹–‘ tissue with limited modification in a qualitative or preferably quantitative way

(Dalsgaard et al. 2003).

1.6.1 Stable carbon and nitrogen isotopes as trophic indicators

Stable isotopes (SI) have been increasingly used as biochemical tracers in trophic studies in both terrestrial and aquatic species (Post 2002, Fry 2006,

Hussey et al. 2012). In trophic studies, stable isotopes are used to measure differences in the isotopic forms of an element to make inferences about food web structure. The majority of carbon on earth has 6 neutrons and protons in the nucleus (12C), however, heavier isotopes have an extra neutron (13C) (Fry

2006). Both isotopic forms behave the same way (e.g. processes in organisms), but 13C bonds and its attractive forces are stronger, causing 13C to react more slowly (Fry 2006). This difference causes a separation (fractionation) between

13 the heavy and light isotope, which means more of the light isotope goes into the product of reactions compared to the heavy isotope. The ratio of 13C to 12C

(G13C) can be compared among species or individuals to determine the source of primary production upon which their food relies (Smith & Epstein 1971, Fry

2006). As plant species have differences in physiological and metabolic pathways (C3/C4), carbon is processed in slightly different ways among species, which causes the amount of 12C and13C in each plant source to vary (Smith &

Epstein 1971). Another aspect to consider is that 12C and13C can be lost through excretion and respiration and should be accounted for when determining sources of primary production (Fry 2006). As 12C is the light isotope it is usually utilised first, causing changes in G13C (Fry 2006). ƒ˜‡”ƒ‰‡‘ˆͳΩ‘ˆG13C is depleted at each trophic level (DeNiro & Epstein 1978).

The fractionation of 14N/15N (G15N) occurs in the same manner as G13C, however, atmospheric nitrogen is converted by microbes and absorbed into plants to form proteins (Fry 2006). Nitrogen then moves through the food chain in a relatively constant, measurable rate, as an animal consumes plants and prey

(Cabana & Rasmussen 1994). Therefore, apex consumers will have the highest

G15N, whilst producers will have the lowest. However, like carbon, nitrogen is excreted, mostly in the form of urea. This causes a difference in G15N in each

–”‘’Š‹ Ž‡˜‡Ž‘ˆ̱͵ǤͶΩǡalthough this difference can range from between 0.6 Ȃ

ͷǤͻΩȋ—‡–ƒŽǤʹͲͳ͵ȌǤ”ƒ†‹–‹‘ƒŽ‘†‡Ž• ƒŽ —Žƒ–‡trophic position (TP)

—•‹‰–Š‡ ‘•–ƒ–͵ǤͶΩ‡”‹ Š‡–ˆƒ –‘”•–ƒ”–‹‰ˆ”‘–Š‡‘™G15N value of a primary producer or consumer. However, these models do not account for differences in in the enrichment of G15N at each trophic level and can

14 underestimate the TP of apex predators (Hussey et al. 2014a,b). Therefore, a narrow discrimination model can be more appropriate to calculate TP, which accounts for changes in the enrichment of G15N at each trophic level and increasing the TP of apex predators (Hussey et al. 2014a,b).

Together, both G13C and G15N represent most of the diet of a consumer and can be beneficial when used in combination: for example, as a proxy for trophic niche (Newsomes et al. 2007, Jackson et al. 2011), which can be compared to establish resource partitioning (Vaudo & Heithaus 2011). Both isotopes can be used to examine intraspecific differences in diet and link prey to consumers via isotope mixing models (Phillips et al. 2014).

1.6.2 Stable isotope analysis of elasmobranchs

Stable isotope analysis is being increasingly used to study the trophic ecology of elasmobranchs and has been particularly successful at estimating trophic levels and quantifying food web structure (Kerr et al. 2006, Carlise et al. 2012,

Pethybridge et al. 2012, Frisch et al. 2016). For example, organochlorine contaminants and G13C and G14N have been used to describe food web linkages and trophic level in the Near Threatened Greenland Shark Somniosus microcephalus in Cumberland Sound and Davis Strait, Canada (Fisk et al. 2002).

This study found that S. microcephalus is a high-order predator that consumes benthic species, although the stomach contents indicated a higher trophic level than what G14N revealed (Fisk et al. 2002). G14N levels were similar to Turbot

Scophthalmus maximus and Ringed Pusa hispida and Harp seals Pagophphilus

15 groenlandicus were found in the stomach content of S. microcephalus, however the organochlorine contaminants were much higher within S. microcephalus tissue than prey tissue, indicating that G14N levels were not sufficient to determine their trophic position. Therefore the use of multiple biochemical tracers was beneficial to avoid discrepancies in the interpretation of biochemical tracers (Fisk et al. 2002).

G13C, G14N and mercury have also been analysed in deep-sea chondrichthyans to examine the bioaccumulation of contaminants within these species

(Pethybridge et al. 2012). Other notable examples of SI applications include the investigation of ontogenetic changes in the trophic position of Blue Shark

Prionace glauca, Mako Shark Isurus oxyricnchus, Common Thresher Shark

Alopias vulpinus (Estrada et al. 2003), intraspecific specialisations in Tiger

Sharks Galeocerdo cuvier compared to Bull Sharks Carcharhinus leucas (Matich et al. 2011) and the dietary habitats of a number of elasmobranch species (e.g.

Yick et al. 2011, Polo-Silva et al. 2012, Tilley et al. 2013, Thorburn 2014 et al.).

The accuracy of SI methods when using elasmobranch tissue has also been investigated to account for the unique physiology of elasmobranchs, including potentially confounding factors such as the high lipid content of elasmobranch livers and high concentrations of urea in the muscle tissue (Logan et al. 2008,

Logan & Lutcavage 2010, Hussey et al. 2012, Kim & Koch 2012), as well as maternal effects (lipid reserves in the neonates liver from the mother) (Olin et al. 2011). Methods developed to overcome these issues include the removal of urea and lipids prior to analyses (Kim & Koch 2012) and lipid normalization

16 models using the ratio of bulk C: bulk N as a proxy for lipid content (Logan et al.

2008).

1.6.3 Fatty acids as food web tracers

Fatty acids are the basic building blocks of lipids and can be used in a similar manner to SIs because they can be passed up through the food chain in a predictable manner. Fatty acids are synthesised or ingested by organisms and play important roles within cell membranes and as cell signals and hormone precursors (Sargent et al. 1999a). How FAs behave in a species is important to understand because it can affect interpretation of food webs; depending on the physiological processes within a species, FAs may increase or decrease in tissues depending on their energy or biochemical cellular requirements

(Dalsgaard et al. 2003).

There are three main groups of FAs classified based on the number of double bonds within the carbon chain; saturated fatty acids (SFA) (no double bonds within the carbon chain), monounsaturated FA (MUFA) (one double bond), and polyunsaturated FAs (PUFA) (multiple double bonds). Most consumers are unable to synthesize all the required PUFA, however some of these FAs are essential to certain species and are termed essential FAs (EFA) such as many of those in the omega 3/omega 6 (Z3/Z6) groups (Sargent et al. 1995). Because of this, and the fact that EFAs are produced in algae and bacteria, they can be traced throughout the food web and used in trophic analysis (Lee et al. 1971,

Graeve et al. 1994, Hall et al. 2006). Some FAs can also be helpful to identify as

17 markers of a certain species or habitat (Dalsgaard et al. 2003, Kelly & Scheibling

2012), although caution must be applied because such biochemical tracers can be found in similar proportions in more than one organism. Therefore, it is often beneficial to use multiple biotracers when attempting to elucidate food web structure (Dalsgaard et al. 2003, Kelly & Scheibling 2012). For example, 18:1Z7 has been linked to freshwater species, whilst a low ratio of Z3/Z6 can suggest abiotic differences such as marine and/or tropical waters (Speers-Roesch et al.

2008). Thus the ratio of Z3/Z6 (e.g. a low ratio) may indicate two possible abiotic factors, having another supporting biochemical tracer such as (high)

18:1Z7 supports the species consuming prey in freshwater.

In trophic studies, FA profiles or signatures are often used to describe the significant FAs identified in the tissues of prey and predators. Once identified these profiles can be used to distinguish the trophic characteristics of species, and can be qualitatively compared to determine dietary links between the species (e.g. Iverson et al. 1993, Käkelä et al. 1993). Iverson et al. (2004) extended this concept and developed a mixing model (quantitative fatty acid signature analysis (QFASA)) that can quantitatively analyse the proportion of

FAs within consumers. To account for the metabolism of the predator and differences in the way FAs are assimilated, calibration coefficients are calculated and used within the model. Although an important method for inferring the diet of consumers, some species are difficult to maintain or rare, therefore calibration coefficients may not be able to be calculated through feeding experiments (Budge et al. 2004, 2011, 2012, Thiemann et al. 2011, Bromaghin et al. 2015).

18 1.6.4 Fatty acids in elasmobranchs

Lipids and FA research using elasmobranch tissue was at first descriptive, reporting nutritional compositions for human consumption (e.g., Nichols et al.

2001, Stoknes et al. 2004, Økland et al. 2005). At this time dietary studies utilising FAs were largely in teleost fishes particularly in relation to aquaculture.

For example, Turbot Scophthalmus maximus was found to utilise 20:4Z6

(arachidonic acid) from fish feed in the promotion of growth (Castell et al.

1994), and similar relationships were also found with other FAs in juvenile fish species (Scott & Middleton 1979, Sargent, McEvoy, et al. 1999, Tocher 2010).

The suitability of FAs to examine aspects of the feeding ecology of elasmobranches was initially examined in a suite of deep-sea species

(Pethybridge et al. 2010). Further studies followed using two methods, stomach content and FA analysis, to determine the feeding ecology of deep-sea chondrichthyans (Pethybridge et al. 2011a). The advantage of using both methods was that stomach content data was able to validate FA profiles

(Pethybridge et al. 2011a). Subsequent experimental studies examined the uptake of FAs in Port Jackson sharks Heterodontus portusjacksoni (Beckmann et al. 2013, 2014). These studies were among the first controlled dietary experiments on elasmobranchs, which provided important information on which tissues (liver and muscle) are useful for dietary studies and the integration times of FAs into elasmobranch tissues. Fatty acids have now been used to infer the trophic ecology in various species of elasmobranchs from

Whale Sharks Rhincodon typus (Couturier et al. 2013, Rohner et al. 2013), to C.

19 leucas (Belicka et al. 2012), S. microcephalus and White Sharks Carcharodon carcharias (Davidson & Cliff 2014) among others.

1.7 Study site: South Alligator River

The research outlined in this thesis was conducted in the South Alligator River in the Northern Territory (NT), Australia (Fig 1.3). The South Alligator River is a relatively pristine tidal river in northern Australia situated in the World

Heritage listed Kakadu National Park. It supports important populations of coastal and euryhaline elasmobranchs, some of which are of conservation concern. Species that utilise the river include euryhaline species C. leucas,

Northern River Shark Glyphis garricki, Speartooth Shark G. glyphis, Freshwater

Whipray, Urogymnus dalyensis and Largetooth Sawfish Pristis pristis and the coastal species, Australian Sharpnose shark Rhizoprionodon taylori and Pigeye

Shark C. amboinensis, both of which can be found in the river estuary (Table

1.1).

20 Table 1.1 Species of elasmobranchs recorded in the South Alligator River, Northern Territory, Australia, with summary of ”‡’‘”–‡††‹‡–ƒ”›‹–‡•ǤDzn ˜ƒŽ—‡•dz”‡ˆ‡”•–‘–Š‡—„‡”‘ˆ‹†‹˜‹†—ƒŽ•‡šƒ‹‡†‹‡ƒ Š‘ˆ–Š‡ ‹–‡†•–—†‹‡•Ǥ

Elasmobranch consumer Published dietary information

Scientific Common Habitat Species n values Method and references

Carcharhinus leucas Bull Shark Euryhaline Teleost fish (e.g. Ariidae 48 (NT, Aus) Stomach content analysis & SI (Tillett et al. 2014) spp., Polydactylus macrochir), 76 (WA, NT, Aus) Stomach content analysis (Thorburn & Rowland 2008) elasmobranchs (Dasyatis spp., small 64 (Florida, USA) Stomach content analysis (Snelson et al. 1984) Carcharhinidae spp.), cephalopods, 309 Natal Coast, South Stomach content analysis (Cliff & Dudley 1991) Cetaceans Africa Carcharhinus Pigeye Shark Coastal Teleost fish, 106 (NT, Aus) Stomach content & SI (Tillett et al. 2014) amboinensis cephalopods

Glyphis garricki Northern River Euryhaline Teleost fish (e.g. Ariidae 5 (WA, Aus) x-ray (Thorburn & Morgan 2004) Shark spp., P. macrochir )

Glyphis glyphis Speartooth Shark Euryhaline Teleost fish 7 (QLD, Aus) Stomach content (Peverell et al. 2006) (e.g. Nematalosa spp., Ariidae spp.) 1 (PNG) Spines and bones found in mouth (White et al. 2015) Crustaceans (e.g. Macrobrachium spinipes*) Pristis pristis Largetooth Euryhaline Teleost fishes (e.g. N. 9 (WA, Aus) Stomach content (Thorburn et al. 2007) Sawfish graeffei) Crustaceans (M. spinipes) Insect parts Rhizopriodon taylori Australian Coastal Teleost fish (e.g. 168 (Qld, Aus) Stomach content (Simpfendorfer 1998) Sharpnose Shark Leiognathidae, Clupeidae) Crustaceans 116 (Qld, Aus) Stable isotopes (e.g. Penaeidae) (Munroe, Heupel, et al. 2014) Cephalopods Urogymnus dalyensis Freshwater Euryhaline - - - Whipray

21

Figure 1.3 Map of Northern Territory, Australia, showing Kakadu National Park (White line indicates the approximate park boundaries), and the rivers within, light blue lines indicate approximate flood boundaries of the Alligator River region, including the study site, the South Alligator River. Inset shows this region in relation to the rest of Australia.

Most of the euryhaline elasmobranchs that inhabit the lower reaches of the

South Alligator River are juveniles to sub-adult and remain in the estuary for approximately three to five years, suggesting that the riverine areas are a nursery for these species (Heupel et al. 2007). The South Alligator has a broad diversity of potential prey (Bishop, Allen, Pollard, et al. 1980; Pusey et al. 2016)

22 and is linked hydrologically to an extensive floodplain that provides annual pulses of energy and nutrients (Pettit et al. 2016a). Despite the physiological challenges associated with living in a macrotidal and highly seasonal habitat

(e.g., strong current, highly variable salinity), such systems may offer protection from larger sharks and other apex predators due to the high turbidity (Wolanski

& Chappell 1996) and the fact that variable salinity may prevent larger sharks from entering the river since very few elasmobranchs inhabit fresh or brackish water in the adult phase (Compagno & Cook 1995, Heupel & Simpfendorfer

2008).

1.7.1 Study Species

Although a number of Carcharhinidae and Batoidea species occasionally utilise the estuaries in northern Australia, only five elasmobranch species have been recorded in both estuarine and freshwater habitats; Northern River Shark

Glyphis garricki, Speartooth Shark G. glyphis, Bull Shark Carcharhinus leucas,

Freshwater Whipray Urogymnus dalyensis and the Largetooth Sawfish Pristis pristis. All of these species are found in the South Alligator River along with the coastal species, the Australian Sharpnose Shark Rhizoprionodon taylori that are commonly found in estuaries in northern Australia. Northern River Shark

Glyphis garricki is considered Critically Endangered (Pogonoski & Pollard 2003) and Speartooth Shark G. glyphis is Endangered (Compagno et al. 2009) according to the IUCN Red List of Threatened Species. Both species occur across northern Australia (Thorburn & Morgan 2004, Peverell et al. 2006, Pillans et al.

2009, Morgan et al. 2011) and G. glyphis is believed to be extirpated along the

23 east coast of Australia, as no individuals have been caught since the first recorded specimen in 1983 in the Bizant River (Pillans et al. 2009). This decline is believed to be caused by mortality associated with bycatch from extensive fishing activity in this region (Pillans et al. 2009).

Whilst Glyphis species appear morphologically similar to each other, a range of key characteristics can be used to distinguish each species. For example, the position of the watermark boundary (the dark and light and colouration on the body) is over an eye diameter below the watermark in G. garricki where as in G. glyphis the watermark is just below the eye (Compagno et al. 2008). Also G. garricki lacks a black blotch on the ventral tip of pectoral fins, which is found on

G. glyphis (Compagno et al. 2008). Neonates of both species are approximately

50 Ȃ 65 cm (Stevens et al. 2005, White et al. 2015) at birth and adult G. glyphis have been estimated to reach 260 cm total length (White et al. 2015) similar to

G. glyphis (Pillans et al. 2009).

Parturition in G. glyphis may occur just before the wet season, enabling juveniles to access a wide range of prey as the water levels rise. Glyphis glyphis have been found in a range of salinities of 0.8 Ȃ 28.0 ppm, with juveniles being found in the lower salinities and adults in marine regions (Pillans et al. 2009, White et al.

2015) whilst G. garricki were found in 2.0 Ȃ 36 ppm (Pillans et al. 2009). Both species move upstream with the flooding tide and downstream with the ebbing tide (Pillans et al. 2009). Both Glyphis spp occur more commonly in rivers of high turbidity (Pillans et al. 2009), avoid very low salinities and migrate to the mouth of the river in response to freshwater inflows (Kyne 2016). The trophic

24 ecology of both is poorly understood, with dietary data limited to opportunistic stomach dissections and x-rays (Thorburn & Morgan 2004, Peverell et al. 2006,

White et al. 2015). The limited data collected to date suggest that the diet is limited to teleost fish and crustaceans (Macrobrachium spp.) (Table 1.1).

The Bull Shark Carcharhinus leucas is a common species that occurs across northern Australia from Bunbry, Western Australia to Sydney, New South

Wales. It can be separated from Glyphis spp. by their broader snout, larger eye, fin and lower watermark boundary (Last & Stevens 2009). Adults return to estuaries to give birth to live young that are from 69 Ȃ 75 cm total length. After being born neonates may swim upstream (Thorburn & Rowland 2008). Despite its prevalence in subtropical to tropical rivers and coastal biomes they are considered Near Threatened (IUCN Red List) (Simpfendorfer & Burgess 2009).

Aspects of the biology of C. leucas have been well studied across its range.

Within northern Australian river systems, C. leucas utilises freshwater habitats and ranges downstream for over 200 km (Thorburn & Rowland 2008). Stomach content analyses in the USA, South Africa, South America and Australia have identified a range of prey species mainly teleost spp., elasmobranchs and crustaceans (Sadowsky 1971, Snelson et al. 1984, Cliff & Dudley 1991, Thorburn

& Rowland 2008, Tillett et al. 2014, Thorburn et al. 2014). Chemical biotracers in juvenile C. leucas have suggested maternal dependency and utilisation of a range of marine, estuarine to freshwater basal sources (Belicka et al. 2012a,

Matich & Heithaus 2015). There was also some evidence of dietary intraspecific

25 difference in a population of C. leucas in the Everglades, Florida (Matich &

Heithaus 2015) and adults were found to be apex predators (Daly et al. 2013).

Other euryhaline species recorded in the South Alligator are rare. For example, the Freshwater Whipray U. dalyensis is a recently described species that is considered of Least Concern (Last & Manjaji-matsumoto 2006) that occurs in estuarine and freshwater habitats from north Western Australia to Queensland.

A study of the movements of this species in the Wenlock River, Queensland showed that U. dalyensis correlated their movement with tidal and lunar cycle during the dry season and the diel cycle in the wet (Campbell et al. 2012).

However, their movement was limited to an 8 kilometre section of the river, although during the wet season males moved up to 60 kilometres down river

(Campbell et al. 2012). Information on the trophic ecology of this species is lacking.

Largetooth Sawfish Pristis pristis occurs from northern Western Australia to

Queensland and is Critically Endangered on the IUCN Red List (Kyne et al.

2013). They are largely found in shallow water and juveniles inhabit the fresh water reaches of rivers and floodplains for approximately four to five years where they move tidally and migrate to marine and estuarine waters (Thorburn et al. 2007, Whitty et al. 2008, 2009). Parturition occurs during the wet season, possibly to ensure the success of neonates as they may be reliant on wet season flows to avoid potential predators such as C. leucas (Morgan et al. 2011,

Thorburn et al. 2014). Limited stomach dissections suggest a diet primarily composed of teleost fish and some crustaceans (Thorburn et al. 2007) (Table

26 1.1). Little is known in regards to their trophic ecology, however when compared alongside an assemblage of freshwater taxa in West Australia, P. pristis was found to be reliant on teleost species and may also be an important prey source for C. leucas (Thorburn et al. 2014).

Of the coastal species in the South Alligator region the Australian Sharpnose

Shark Rhizoprionodon taylori is a small (22 Ȃ 78 cm) carcharhinid, endemic to tropical Australian waters (Compagno 1984) and is considered a species of

Least Concern (IUCN Red List). It is relatively fecund compared to other carcharhinids and embryos are held in diapause until the water is at its warmest (Simpfendorfer 1992). This is believed to allow embryos to be born when the most diverse and abundant prey is present (Simpfendorfer 1992).

North Queensland populations of R. taylori were found to be resident in areas of approximately 100 kilometres2, commonly moving between embayments whilst foraging (Munroe et al. 2015). The diet of R. taylori from stomach content analysis consists of a wide range of teleost and penaeids (Simpfendorfer 1998).

G13C and G14N (including trophic position) has been found to vary amongst individuals and a large niche range was calculated, indicating a range of prey sources and a generalist feeding strategy (Munroe et al. 2014a, b). Although well studied in northern Queensland, dietary data is lacking from populations in the

‡–”ƒŽƒ†™‡•–‡”’ƒ”–•‘ˆ–Š‡•’‡ ‹‡•ǯ”ƒ‰‡where the species occurs surrounding large riverine systems.

Carcharhinus amboinensis is found in the nearshore regions of Indo-West

Pacific and down to the subtropical regions of Australia and is currently

27 considered Data Deficient (IUCN Red List). They are similar to C. leucas in appearance but have slightly larger dorsal fin and the angle of the notch in the anal fin is acute compared to a right angle in C. leucas (Compagno et al. 2005).

These sharks utilise estuarine areas of rivers but are intolerant to freshwater outflows (Knip, et al. 2011a). This habitat is particularly important for juvenile

C. amboinensis (Knip, et al. 2011a) and it is likely that this species has some dietary overlap with C. leucas.

1.8 Thesis Outline & Aims

As outlined above, it is critical to understand the trophic interactions of species in order to understand how ecosystems respond to changes in factors such as harvest pressure and climate. This is especially the case for species, such as apex and mesopredators that have the potential to shape the structure and function of biotic assemblages and food webs. Compared to other tropical rivers outside of Australia where euryhaline elasmobranchs reside, the South Alligator

River is a relatively undisturbed river system. This means that dietary baseline data can be collected from elasmobranchs and their food webs in this system and potentially applied to other more disturbed systems. Identification of prey and basal sources has the potential to provide managers and policy makers with key information to develop strategies to protect estuary/river ecosystems, and thus, assist in the recovery of euryhaline elasmobranchs and other threatened species.

28 In this thesis, I aim to utilise biomarkers to examine the trophic biology and ecology of coastal and euryhaline sharks within the South Alligator River. In doing so, I explore three key questions in the main data chapters. In Chapter 2, I explore the utility of different shark tissues for profiling the fatty acids consumed by three species of elasmobranch: C. leucas, G. garricki and G. glyphis.

Specifically, I address two questions: (i) How similar are the FA profiles between muscle and fin tissue of C. leucas, G. garricki and G. glyphis? and (ii) Are fin clippings suitable for future FA dietary studies? The FA profiles of muscle and fin tissues are quantitatively compared to determine the similarities and differences between the tissues amongst species. I attempt to explain the differences between the FAs in fin and muscle tissue and suggest why fins could be used for future dietary studies on the condition that further research is carried out on fin FAs. Based on these findings, I elected to use muscle tissue both SI and FA analysis to inform dietary data in the remainder of the thesis.

In Chapter 3, I apply two sets of biomarkers to examine aspects of euryhaline elasmobranchs food webs including trophic position, basal source, niche and niche overlap. To determine this, I asked the following questions (i) What are the trophic positions of euryhaline and coastal sharks of the South Alligator

River and how do they compare as an assemblage? (ii) Which basal sources

(marine, estuarine, freshwater) contribute to the elasmobranches diet? and (iii)

What is the extent of dietary overlap within and among species? Using G15N, the trophic positions of elasmobranchs were estimated. Comparisons of G13C from other studies in northern Australia (e.g. Loneragan et al. 1997) were then used to estimate the basal sources of primary production supporting elasmobranchs.

29 Niche sizes were determined for each species where sample size was sufficient and used to calculate the SI overlap of elasmobranch assemblage. Fatty acid signatures were then used to calculate niche size and overlap using a recently developed Bayesian method (Swanson et al. 2015). The difference between data from SI and FA analysis are discussed and combined to describe the food web supporting elasmobranchs in the system. Elasmobranchs were divided into two separate groups, one estuarine/marine and the other freshwater, suggesting that they are feeding over a range of biomes with limited seasonal differences.

To further explore differences between and amongst these species, I then explore the biomarkers of elasmobranch putative prey species in the next chapter.

In Chapter 4, biochemical tracers from a range of putative prey are compared against those of the elasmobranchs examined in Chapter 3 to describe the structure and function of food webs. The following two questions were addressed: (i) What can biochemical tracers of putative prey and the euryhaline species C. leucas, G. garricki and one coastal species, R. taylori reveal about the structure and function of food webs within the South Alligator River? (ii) Are there trophic connections between putative prey and elasmobranch species and can this be used to assess their diet? To describe the structure and function of

ˆ‘‘†™‡„•™‹–Š‹–Š‡‘—–ŠŽŽ‹‰ƒ–‘”‹˜‡”ǡ ˆ‹”•–†‡–‡”‹‡ƒ›‡ǯ• community wide metrics using G13C and G15N and then explore the potential links between putative prey and the three elasmobranch species using both sets of biochemical tracers. An overall comparison of prey and shark SI and FAs is first described to examine the similarities between prey and sharks. Stable

30 isotope Bayesian mixing models are then used to find the most likely contribution of prey SI within shark SI. The differences amongst shark FAs are finally compared to determine intraspecific differences. From here, I discuss the variation found amongst all species, sites and seasonal differences of putative prey species.

Finally, I synthesise the new insights gained by this research in Chapter 5, with the broader body of work on the trophic role and connections provided by elasmobranchs in marine and freshwater ecosystems. In this way, I reveal what is known of tropical estuarine food webs and the role of predators within those food webs. Management options are then considered for these ecosystems and future directions are discussed.

31 2 Comparison of fin and muscle tissues for analysis of signature fatty acids in tropical euryhaline sharks

This chapter has been published in the Journal of Experimental Marine

Ecology and Biology.

Every SL, Pethybridge HR, Crook DA, Kyne PM, Fulton CJ (2016) Comparison of

fin and muscle tissues for analysis of signature fatty acids in tropical

euryhaline sharks. J Exp Mar Bio Ecol 479:46Ȃ53

Below is a list of the co-authors and their contribution. All authors contributed to concept design and writing.

Heidi Pethybridge Ȃ Contributed to the experimental design, knowledge of multivariate and lipid analysis, fish biology and ecology and assited with planning and editing manuscript.

Christopher Fulton provided advice and assistance on the statistical analyses, suggested key literature, and assisted with editing of the published manuscript.

Peter Kyne Ȃ Contributed knowledge of elasmobranch biology and ecology, and assisted with fieldwork and tissue collection.

David Crook Ȃ Contributed knowledge of fish biology and ecology and statistical analysis of fish and riverine ecology and editing manuscript.

32 2.1 Abstract

Fatty acid (FA) analysis can provide an effective, non-lethal method of elucidating the trophic ecology of fish. One method utilized in the field is to collect biopsied muscle tissue, but this can be problematic in live sharks due to a thick dermal layer with extensive connective tissue. The aim of this research was to determine whether fin and muscle tissue yield similar FA profiles in three species of tropical euryhaline sharks: Carcharhinus leucas, Glyphis garricki and Glyphis glyphis. Fatty acid profiles were detectable in fin clips as small as 20 mg (~5 mm x 6 mm) and muscle biopsies >10 mg mass. Overall profiles in relative (%) FA composition varied significantly between fin and muscle tissues for C. leucas and G. garricki (global R-values = 0.204 and 0.195, P < 0.01), but not

G. glyphis (global R-value = 0.063, P = 0.257). The main FAs that contributed to these differences were largely 18:0 for C. leucas, 20:4Z6 for G. garricki and

20:5Z3 for G. glyphis, which reflect the different physiological functions and turnover rates of the two tissues. Notably, no significant differences were detected between tissue types for the major classes of FAs and abundant dietary essential FAs. It was concluded that FA profiles from either fin clips or muscle tissue may be used to examine the trophic ecology of these tropical euryhaline sharks when focusing on dietary essential FAs. Given that some non-essential

FAs were different, caution should be applied when comparing FA profiles across different tissue types.

33 2.2 Introduction

Many shark, ray and chimaera species (Class Chondrichthyes) are susceptible to severe population reductions as a result of negative anthropogenic influences such as over-exploitation and habitat destruction, with an estimated 24% of chondrichthyan species considered to be threatened (Dulvy et al., 2014).

Reductions in the abundance of apex or meso-predators such as sharks can cause changes in ecosystems through competitive release, resulting in the alteration of fish population dynamics (Stevens et al., 2000). It is important, therefore, to understand the trophic ecology of sharks to evaluate the consequences of reductions in their abundance. Given the rarity and/or threatened status of many shark species, non-lethal and minimally intrusive methods for determining diet are often required.

Prey consumption analyses in sharks have traditionally involved stomach content analyses, which require major intervention (e.g., gastric lavage) or lethal dissection (Barnett et al., 2010; Cortés, 1999). In recent times, less invasive, but still highly informative techniques have been used, such as stable isotopes (e.g., Hussey et al., 2011a; Speed et al., 2011) and lipid and fatty acid

(FA) profiling (e.g., Couturier et al., 2013a; Rohner et al., 2013). Fatty acids have been validated in determining the dietary sources of sharks through comparisons with stomach content analysis (Pethybridge et al., 2011a) and in vivo (Beckmann et al., 2013). This concept works due to the inability of most high-‘”†‡”’”‡†ƒ–‘”•–‘•›–Š‡•‹œ‡•’‡ ‹ˆ‹  •ǡ•— Šƒ•ʹʹǣͷɘ͵ƒ†ʹʹǣ͸ɘ͵

(Iverson, 2009) that are only found in primary producers or lower order

34 consumers. The detection of such FAs within the tissues of a consumer suggests direct or secondary consumption of specific taxa such as autotrophic algae, diatoms and bacteria (Dalsgaard et al., 2003; Parrish et al., 2013). In addition to dietary information, FA analysis has been used to acquire information on elasmobranch (shark and ray) bioenergetics, life-history and physiology

(Beckmann et al., 2014a; Pethybridge et al., 2014, 2011b).

Fatty acids are vital for cell and organelle function in living organisms, especially essential FAs (EFA) that are involved in critical physiological functions (Tocher, 2003). While many FAs can only be assimilated by consumers through their diet, some FAs necessary for physiological and structural functions are produced de novo (Tocher, 2003). Given the variety of tissue structure and functionality within multicellular animals, FA profiles can vary among tissue types. For instance, different shark tissues have been found to preferentially store higher saturated fats (SFA) and polyunsaturated fats

(PUFA) in structural tissues (e.g., muscle), while higher monounsaturated fats

(MUFA) are often found in tissues used for energy storage (e.g., liver,

(Pethybridge et al., 2010)). While liver tissue can provide the most temporally sensitive indicator of dietary change in sharks (Beckmann et al., 2014b), it requires lethal sampling. Muscle tissue provides dietary information integrated over longer time periods, but can be problematic to collect in live sharks due to a thick dermal layer with extensive connective tissue (Tilley et al., 2013).

Although fin clips are used extensively in shark genetic studies (e.g., Lewallen et al., 2007), and are recognised as a viable tissue for stable isotope analysis (e.g.,

35 Hussey et al., 2011b; Olin et al., 2014), their utility for FA analysis has not yet been determined.

Shark fins consist of cartilage and some connective tissue, muscle and vascularisation, with an outer dermal layer covered with denticles. This composition of various tissue types has the potential to influence the FA profiles of fins versus muscle tissue, given the tissue-based differences reported for stable isotope analysis of G13C (Hussey et al., 2010). Here, FA profiles obtained from fin tissue and non-lethal muscle biopsies are examined to determine whether they differ from the same three species of tropical euryhaline elasmobranchs: Bull Shark Carcharhinus leucas, Northern River Shark Glyphis garricki, and Speartooth Shark Glyphis glyphis. River sharks (Glyphis species) are globally threatened and rare species (Pillans et al., 2009) with little information available on their biology, including trophic ecology. In doing so, the utility of fin tissue was explored as a non-lethal method for examining FA profiles in future dietary analyses of potentially important apex predators in tropical river ecosystems.

2.3 Methods

2.3.1 Ethics statement

This study was conducted with the approval of the Charles Darwin University animal ethics committee (Approval A12016 and A11041) in conjunction with permits from NT Fisheries and Kakadu National Park (Permit RK805).

36 2.3.2 Tissue sampling and preparation

Sharks from each of the three target species (Table 2.1) were captured from the

South Alligator River, Kakadu National Park, Australia, between March 2013

and July 2014 using 4 or 6 in. gill nets, or hook and line. Tissues were collected

from each temporarily restrained (<5 min) individual before they were released

back into the water. All sharks were juveniles or sub-adults (Table 1). Muscle

tissue biopsies (mean wet weight 0.025 g) were collected from the caudal

peduncle using a 3Ȃ5 mm biopsy punch (Stiefel, USA), along with a fin clip

sample (~15 mm2 and 0.03 g) from the rear tip of a pectoral fin (Lewallen et al.,

2007). Tissue samples were immediately placed in liquid nitrogen (Ȃ196°C) for

up to 1 week during fieldwork, then transferred to a Ȃ20°C freezer. To avoid

degradation of the sample from defrosting and refreezing, all frozen muscle

samples were dissected in the freezer to remove dermal layers and as much

connective tissue as possible to ensure only muscle tissue was sampled. While

initial samples were extracted from wet tissue, these samples were freeze-dried

for analysis.

Table 2.1 Number and total length (TL) of specimens from which samples of fin and muscle tissue were taken for fatty acid analysis in three shark species from the South Alligator River, Australia (Size range +/- SD).

Species n Min TL (cm) Max TL Mean TL Sex ratio

(cm) (cm) M:F

Carcharhinus leucas 17 74.5 82.5 78.49r3.48 8:9

Glyphis garricki 11 75.5 140.5 96.45r19.60 7:4

Glyphis glyphis 4 71.0 85.0 76.80r6.25 1:3

37 2.1.1 Lipid and fatty acid extraction

Total lipid content was extracted using the modified Bligh and Dyer (1959) method using a one-phase dichloromethane (DCM):Methanol (MeOH):milliQ

H2O solvent mixture (10:20:7.5 mL) which was left overnight. After approximately 12 hours, the solution was broken into two phases by adding 10 mL of DCM and 10 mL of saline milliQ H2O (9 g sodium chloride (NaCl) L-1) to give a final solvent ratio of 1:1:0.9. The lower layer was drained into a 50 mL round bottom flask and concentrated using a rotary evaporator. The extract was transferred in DCM to a pre-weighed 2 mL glass vial. The solvent was blown down under a constant stream of nitrogen gas, and the round bottom flask rinsed three times with DCM into the vial. The total lipid extract (TLE) was

†”‹‡†‹–Š‡˜‹ƒŽ–‘ ‘•–ƒ–™‡‹‰Š–ƒ†ʹͲͲɊŽ‘ˆ™ƒ•ƒ††‡†Ǥ‘”‡Ž‡ƒ•‡ fatty acids from the lipid backbone, 10mg of TLE was added per 1.5 mL of DCM and transmethylated in MeOH:DCM:hydrochloric acid (HCl) (10:1:1 v/v) for 2 hours at 100°C. After cooling, 1.5 mL Milli-Q water was added and FA were extracted three times with 1.8 mL of hexane:DCM (4:1 v/v), after which individual tubes were vortexed and centrifuged at 2000 rpm for 5 mins. After each extraction, the upper organic layer was removed under a nitrogen gas stream. A known concentration of internal injection standard (19:0 FAME or

23:0 FAME) preserved in DCM was added before 0.2 Pl of this solution was injected into an Agilent Technologies 7890B gas chromatograph (GC) (Palo Alto,

California USA) equipped with an EquityTM-1 fused silica capillary column (15 m x 0.1 mm internal diameter and 0.1 Pm film thickness), a flame ionization detector, a splitless injector and an Agilent Technologies 7683B Series auto- sampler. At an oven temperature of 120°C, samples were injected in splitless

38 mode and carried by helium gas. Oven temperature was raised to 270°C at 10°C min-1, and then to 310°C at 5°C min-1. Peaks were quantified using Agilent

Technologies ChemStation software (Palo Alto, California USA). Confirmation of peak identifications was by GC-mass spectrometry (GC-MS), using an on-column of similar polarity to that described above and a Finnigan Thermoquest DSQ GC-

MS system. Only fin and muscle tissue samples that were above 0.02 g and 0.01 g in mass, respectively, were used in these analyses, as lower sample masses compromised analytical detection.

Total FAs were determined in mg/g and calculated based on the total area of peaks of all FAs divided by the internal standard, times, the mass and volume of internal standard, the mass of the tissue and dilution factors.

2.3.3 Statistical analyses

Fatty acids were expressed as a percentage of total FAs in the sample, and FAs that accounted for less than 0.5% were excluded from statistical analyses.

Paired t-tests were used to detect significant differences in the means of the major classes of total FAs (SFA, PUFA, MUFA) and four abundant EFAs within matched pairs of fin and muscle tissues from each individual for each shark species. t-Tests were carried out on these EFAs to determine the extent of their influence in causing the differences between the tissues. Analysis of similarity

(ANOSIM) was then applied to the multivariate FA profiles (31 FAs) obtained from each tissue type in a single factorial design to examine differences in overall FA profiles from the two tissue types. As fin and muscle tissues were

39 extracted from the same individual, a dissimilarity matrix was used based on binomial deviance to accommodate the non-independence of samples (Clarke and Warwick, 2001). Where differences were detected by ANOSIM, similarities of variance (SIMPER) were used to determine the dietary FAs that contributed most to these differences, by indicating the percentage contribution of each FA based on the Euclidian dissimilarity of each pair. All multivariate analyses were performed using PRIMER (v6), while univariate analyses were performed using the base package of R (R Core Development Team, 2014).

2.4 Results

A total of 65 FAs were identified across the three shark species, with 31 FAs having relative mean values greater than 0.5% (Table 2.2). These 31 FAs made up 68Ȃ97% of total FAs, whereas the mean sum of the remaining 34 minor FAs ranged from 4Ȃ8%. Total FA was higher in muscle than fin in all sharks with large standard deviations in C. leucas and G. garricki whilst G. glyphis had less variation (Table 2.2).

40 Table 2.2 Comparisons of the relative abundance of fatty acids (FA) (mean % r standard deviation) between fin and muscle tissue in Carcharhinus leucas, Glyphis garricki and G. glyphis, from the South Alligator River, Australia.

C. leucas G. garricki G. glyphis

Muscle Fin Muscle Fin Muscle Fin

16:0 10.63 ±5.12 10.00 ±4.92 11.13 ±2.68 10.24 ±2.99 9.54 ±1.80 11.29 ±3.63 17:0 0.51 ±0.44 0.91 ±0.35 0.71 ±0.12 1.13 ±0.39 0.80 ±0.17 1.14 ±0.31 18:0 17.94 ±5.54 19.85 ±5.69 17.51 ±4.04 17.18 ±2.63 17.64 ±1.93 17.01 ±2.55 20:0 0.63 ±0.64 0.59 ±0.26 1.30 ±2.81 1.01 ±2.02 0.32 ±0.03 0.32 ±0.09 22:0 0.51 ±0.37 1.43 ±2.14 2.08 ±3.74 0.81 ±0.63 0.59 ±0.16 0.67 ±0.21 24:0 0.42 ±0.28 1.17 ±0.63 0.30 ±0.08 0.74 ±0.29 0.54 ±0.19 0.78 ±0.3 15:1 1.35 ±1.33 0.96 ±0.81 2.30 ±1.53 0.94 ±0.61 1.42 ±0.56 0.57 ±0.3 16:1Z7 1.89 ±1.44 1.53 ±1.18 0.94 ±0.40 1.03 ±0.51 0.90 ±0.28 0.93 ±0.36 17:1 1.12 ±0.77 2.59 ±1.48 1.10 ±0.32 3.04 ±1.86 2.66 ±1.95 2.64 ±1.26 18:1Z9 16.50 ±6.35 14.52 ±3.43 12.19 ±5.51 10.97 ±2.49 10.35 ±1.7 10.34 ±0.92 18:1Z7 5.36 ±2.49 3.71 ±1.37 5.53 ±1.72 3.64 ±1.54 5.47 ±0.84 3.79 ±0.53 17:1Z6 0.51 ±0.19 1.00 ±0.46 0.64 ±0.25 0.65 ±0.18 0.83 ±0.40 0.74 ±0.20 19:1 0.41 ±0.19 0.67 ±0.25 0.35 ±0.12 0.34 ±0.08 0.30 ±0.23 0.41 ± 07 20:1Z9 1.21 ±0.66 0.83 ±0.35 0.86 ±0.60 0.78 ±0.66 0.57 ±0.26 0.68 ±0.20 20:1Z5 0.54 ±1.08 0.48 ±0.37 0.18 ±0.14 0.17 ±0.07 0.16 ±0.03 0.24 ±0.19 22:1Z11 2.13 ±5.76 0.25 ±0.51 0.19 ±0.24 0.18 ±0.16 0.12 ±0.40 0.07 ±0.02 24:1Z9 0.83 ±0.43 0.64 ±0.32 0.57 ±0.23 0.65 ±0.25 0.96 ±0.66 1.02 ±0.25 18:2b 0.62 ±0.33 2.60 ±0.98 0.28 ±0.21 1.38 ±0.53 0.49 ±0.13 1.32 ±0.69 18:2c 0.97 ±0.74 0.51 ±0.58 0.13 ±0.32 0.31 ±0.41 0.12 ±0.09 0.14 ±0.02 18:2Z6 0.55 ±0.91 0.56 ±0.55 2.34 ±1.14 1.83 ±0.68 2.36 ±0.97 1.58 ±0.38 20:2 3.02 ±2.23 1.00 ±1.29 0.55 ±1.02 0.56 ±1.12 0.29 ±0.15 0.21 ±0.07 20:2Z6 0.59 ±0.87 0.27 ±0.5 0.76 ±0.28 0.42 ±0.16 0.83 ±0.23 0.51 ±0.22 20:3Z9# 8.36 ±6.41 7.91 ±4.43 1.80 ±4.70 1.52 ±3.66 0.35 ±0.18 0.29 ±0.10 20:3Z6 0.32 ±6.41 0.28 ±0.25 0.67 ±0.33 0.56 ±0.20 0.55 ±0.26 0.65 ±0.37 22:3 0.94 ±0.79 0.82 ±0.71 1.33 ±1.13 2.29 ±1.17 2.57 ±1.37 1.45 ±0.85 20:4Z6 3.18 ±4.38 5.66 ±3.37 10.47 ±4.68 15.46 ±5.46 14.76 ±3.96 12.43 ±4.74 22:4Z6 1.51 ±2.09 2.50 ±1.26 4.44 ±1.85 6.07 ±2.48 3.91 ±3.00 6.96 ±2.28 20:5Z3 0.52 ±0.20 1.30 ±3.27 0.89 ±0.71 0.97 ±0.75 0.94 ±0.32 4.74 ±7.22 22:5Z3 1.91 ±1.40 3.01 ±2.12 0.80 ±1.27 0.34 ±0.72 1.47 ±1.92 0.24 ±0.42 22:5Z6 0.89 ±0.61 1.01 ±0.46 1.87 ±0.67 1.52 ±0.55 1.76 ±0.15 1.97 ±1.11 22:6Z3 4.25 ±2.39 2.22 ±1.37 7.38 ±3.82 4.10 ±1.90 7.97 ±2.99 4.32 ±2.28

ȭ •δͲǤͷΨȗ 8.38 ±8.74 5.04 ±1.67 4.13 ±1.84 5.31 ±1.22 5.03 ±1.25 4.99 ±1.37 ȭSFA 30.66 ±8.75 33.97 ±17.81 33.05 ±7.54 31.14 ±6.18 29.46 ±3.16 31.23 ±7.56 ȭMUFA 32.01 ±6.71 27.26 ±5.15 24.55 ±7.99 22.09 ±4.31 23.18 ±4.44 20.74 ±1.17 ȭPUFA 27.60 ±12.20 29.7 ±7.89 33.63 ±8.32 37.08 ±5.33 37.66 ±6.72 36.20 ±9.15 Z3/Z6 0.99 ±0.45 0.65 ±1.20 0.44 ±0.65 0.23 ±0.33 0.43 ±0.61 0.38 ±1.08 i17:0 0.71 ±0.34 1.95 ±1.08 1.78 ±4.06 0.75 ±0.73 0.87 ±0.31 0.97 ±0.44 16:0FALD 0.61 ±0.64 0.57 ±0.34 1.46 ±1.01 1.25 ±0.62 1.44 ±1.43 1.64 ±1.12 18:0FALD 0.88 ±0.40 1.49 ±1.02 0.86 ±0.46 1.74 ±1.80 0.98 ±0.58 2.82 ±0.99

TFAs (mg/g) 2.56 ±4.16 1.76 ±1.35 4.21 ±5.95 3.66 ±03.71 1.06 ±0.23 0.56 ±0.37

FAs <0.5% include ͳͶǣͲͳͷǣͲǡƒͳͷǣͲǡͳͷǣͲǡͳͶǣͳǡͳ͸ǣͳɘͳ͵ǡͳ͸ǣͳɘͻǡͳ͸ǣͳɘ͹ǡͳ͸ǣͳɘͷǡ ͳ͹ǣͳɘͺΪƒͳ͹ǣͲǡͳͺǣͳɘ͹ǡͳͺǣͳɘͷǡͳͺǣͳǡͳͻǣͳǡʹͲǣͳɘ͹ǡʹͲǣͳɘͳͳǡʹͲǣͳɘͷǡʹʹǣͳɘͻǡʹʹǣͳɘ͹ǡ ʹͶǣͳɘͳͳǡʹͶǣͳɘ͹ǡ 16:4+16:3, 18:2avǡͳͺǣͶɘ͵ǡͳͺǣ͵ɘ͸ǡͳͺǣ͵ɘ͵ǡʹͲǣͶɘ͵ȀʹͲǣʹǡʹͳǣͷɘ͵ǡʹͳǣ͵ǡ 22:2av, 22:2bv, i16:0, 18:1FALD # ʹͲǣ͵ɘͻ‹†‡–‹ˆ‹‡†„ƒ•‡†‘ ‘’ƒ”‹•‘™‹–Š‘–Š‡”C. leucas fatty acid literature; a standard was not available at the time of analyses. v = unable to identify bonds as standard was not available at the time of analyses. FA - Fatty acids, TFA Ȃ total fatty acids, SFA- saturated fatty acids, MUFA - monounsaturated fatty acids, PUFA - polyunsaturated fatty acids. FALD Ȃ fatty aldehyde analysed as dimethyl acetal.

41 2.4.1 Intraspecific tissue differences

No significant differences in the proportions of the main FA classes were detected between fin and muscle for these three species, with the exception of

MUFA in C. leucas where higher amounts were found in muscle (Table 2.3; Fig.

2.1). For all species, large intraspecific variability (standard deviations [SD]) in the major FA classes was observed in both fin and muscle tissues (Table 2.2; Fig

2.1). Standard deviations for most FAs were similar for both muscle and fin for a given species. There were, however, substantial differences in the degree of intraspecific variability in several FAs between muscle and fin. In C. leucas, for

‡šƒ’Ž‡ǡͳ͸ǣͲ ǡͳ͹ǣͲǡͳͺǣʹɘ͸ǡʹͲǣͶɘ͸ǡʹͲǣͳɘͷƒ†ʹʹǣͶɘ͸were more variable in muscle than fins, whilst the opposite was the case for ʹͲǣͷɘ͵ƒ†

20:2. In G. garricki, iͳ͹ǣͲǡͳͺǣʹ„ǡͳͺǣͳɘͻ™‡”‡‘”‡˜ƒ”‹ƒ„Ž‡‹—• Ž‡ǡ while16:0FALD, 17:1 and 18:0FALD were more variable in fins. In G. glyphis,

16:0FALD, 17:1, and ʹʹǣͶɘ͸™‡”‡‘”‡˜ƒ”‹ƒ„Ž‡‹—• Ž‡ǡ™Š‹Ž‡ͳͺǣʹ„ǡʹͲǣͷɘ͵

ƒ†ʹͶǣͳɘͻ™‡”‡‘”‡˜ƒ”‹ƒ„Ž‡‹ˆ‹•Ǥ

42

Figure 2.1 Comparison of the relative means (± standard deviation) of (A) saturated, (B) monounsaturated, and (C) polyunsaturated fatty acid profiles based on fin and muscle tissues taken from three shark species (Carcharhinus leucas, Glyphis garricki and G. glyphis) from the South Alligator River, Kakadu National Park, Australia.

43

Table 2.3 Paired t-tests comparing the concentrations of three major fatty acid (FA) classes detected within the fin and muscle tissues from each of three euryhaline shark species, Carcharhinus leucas, Glyphis garricki, and G. glyphis, from the South Alligator River, Australia. Significant (P<0.05) result shown in bold.

Major FA Class Species t Score df p-Value C. leucas -1.595 16 0.130

G. garricki 0.649 10 0.531 Saturated G. glyphis -0.775 3 0.494 C. leucas 2.279 16 0.037 Monounsaturated G. garricki 1.237 10 0.244 G. glyphis 1.429 3 0.249 C. leucas -0.785 16 0.444

G. garricki -1.541 10 0.154 Polyunsaturated G. glyphis -0.990 3 0.395

In both the muscle and fin clips of C. leucas, the FAs with highest relative

ƒ‘—–•™‡”‡ͳͺǣͲǡͳͺǣͳɘͻǡͳ͸ǣͲǡƒ†ʹͲǣ͵ɘͻǡ‹‘”†‡”‘ˆ†‡ ”‡ƒ•‹‰”‡Žƒ–‹˜‡ importance (Table 2.2). In G. garricki muscle, the 4 dominant FAs were 18.0,

ͳͺǣͳɘͻǡͳ͸ǣͲƒ†ʹͲǣͶɘ͸ǡƒ†ƒŽ–Š‘—‰Š–Š‡•ƒ‡ •™‡”‡†‘‹ƒ–‹–Š‡ˆ‹•ǡ

–Š‡‘”†‡”‘ˆ‹’‘”–ƒ ‡™ƒ•†‹ˆˆ‡”‡–ȋͳͺǣͲǡʹͲǣͶɘ͸ǡͳͺǣͳɘͻƒ†ͳ͸ǤͲǢƒ„Ž‡

2.2). For G. glyphis muscle and fin, the two dominant FAs were consistently 18:0

ˆ‘ŽŽ‘™‡†„›ʹͲǣͶɘ͸ǡŠ‘™‡˜‡”ǡ–Š‡—• Ž‡Šƒ†Š‹‰Š‡”Ž‡˜‡Ž•‘ˆͳͺǣͳɘͻ–Šƒͳ͸ǣͲǢ the opposite was true for fins for this species. t-Tests of the major EFAs

(20:4Z6, 22:6Z3, 20:5Z3, 20:3Z9) found in the fins and muscles indicated no significant difference among tissue types, except for 20:4Z6 in G. garricki (Table

2.4, Fig 2.2).

44 Figure 2.2 Comparison of the fatty acid (a) 20:5Z3, (b) 20:4Z6, (c) 22:4Z6 and (d) 20:3Z6 (%) relative means (± standard deviation) within fin and muscle tissues taken from three shark species (Carcharhinus leucas, Glyphis garricki and G. glyphis) from the South Alligator River, Kakadu National Park, Australia

45

Table 2.4 Paired t-tests comparing the concentrations of four essential fatty acids detected within the fin and muscle tissues from each of three euryhaline shark species, Carcharhinus leucas, Glyphis garricki, and G. glyphis, from the South Alligator River, Australia. Significant (P<0.05) result shown in bold.

Abundant Species t score df p-Value EFA C. leucas 0.97 16 0.34 20:5Z3 G. garricki 0.34 10 0.74 G. glyphis 0.89 3 0.44 C. leucas 2.02 16 0.06 20:4Z6 G. garricki 2.29 10 0.04 G. glyphis -0.58 3 0.59 C. leucas 1.58 16 0.13 22:4Z6 G. garricki 1.74 10 0.11 G. glyphis 1.20 3 0.31 C. leucas -0.31 16 0.76 20:3Z9 G. garricki -0.14 10 0.89 G. glyphis -0.37 3 0.74

Multivariate analysis revealed a large amount of overlap in the overall FA profiles obtained from the fins and muscles of each species (Fig. 2.3). The overall FA profile, however, had significant but weak differences that were detected between fins and muscles for C. leucas (global R-value = 0.204, P <

0.01) and G. garricki (global R-value = 0.195, P < 0.01), but not in G. glyphis

(global R-value = 0.063, P = 0.257).

46

Figure 2.3 Ordination (nMDS) of fatty acid profiles from the fin and muscle tissues of the three shark species (A) Carcharhinus leucas, (B) Glyphis garricki, (C) G. glyphis from the South Alligator River, Kakadu National Park, Australia.

47 2.4.2 Interspecific differences

Similar relative amounts of SFA were observed in all three species (range 29.46 to 33.97%), while C. leucas had higher amounts of MUFA and lower amounts of

PUFA than G. garricki and G. glyphis. Both Glyphis species had less variation in the SD of FAs between fin and muscle tissues than C. leucas. There were 11 EFAs that were detected in all species that were >0.5% and 10 EFAs that had minor contributions (<0.5%) for C. leucas and G. glyphis, and 8 in G. garricki (Table

2.2). Notably, the muscle of C. leucas consistently had higher relative amounts of all four EFAs, while in G. garricki and G. glyphis the relative amounts varied according to the specific EFA (Fig 2.2).

The FAs contributing to these significant but weak differences in the multivariate analysis varied among species (Table 2.2; Fig. 2.4). In C. leucas,

18:0, 20:3Z9, 18:1Z9 and 16:0 contributed to 58% of the differences between fin and muscle, whereas in G. garricki, 56% of the differences were due to

20:4Z6, 18:1Z9, 20:3Z9 and 22:6Z9. The FAs contributing 60% of the difference between tissue types in G. glyphis were all EFAs, as well as 20:5Z3,

20:4Z6 and 22:4Z6. Fatty acids that appeared to be in similar amounts among tissue types were 16:0, 18:0, 20:0, 19:1, 20:1ɘͻǡ 20:1ɘͷǡʹͲǣ͵ɘ͸and 24:1ɘ9.

There was considerable variation among individuals as shown by the large standard deviations for 20:5Z3 in G. glyphis (fin and muscle) and C. leucas

(muscle), 20:4Z6 and 22:4Z6 in the fin of C. leucas, and 20:3Z9 in both tissue types in G. garricki. The ‡ƒ”ƒ–‹‘‘ˆɘ͵Ȁɘ͸ •™ƒ•Š‹‰Š‡”‹–Š‡—• Ž‡ compared to the fins of all species.

48

Figure 2.4 % Contribution of fatty acids that caused the main differences between fin and muscle profiles from SIMPER analysis in (a) Carcharhinus leucas (b) Glyphis garricki and (c) G. glyphis from the South Alligator River, Kakadu National Park, Australia.

49 2.5 Discussion

Overall FA profiles did appear to differ according to tissue type within the two shark species C. leucas and G. garricki, but not G. glyphis, which suggests caution must be applied when selecting which tissue type to use for future dietary studies in these and other chondrichthyan species. Sample size for G. glyphis was low which may partially account for the differences between the species, however this species was included due to its rarity (Pillans et al., 2009).

Differences in the overall FA profiles among tissue types were expected and are likely due to functional and dietary differences of certain FAs and their affiliation with different structural tissue types, which can be difficult to separate. Most of these differences in fin and muscle tissue were due to non- essential FAs and there were some important similarities that were apparent among the two tissue types in terms of key FAs. This included important EFAs, which suggests that the potentially less intrusive use of fin tissues may be effective for future studies wishing to explore dominant trophic patterns in these tropical euryhaline sharks.

Similarity in the proportions of major classes of FAs among tissues types and species suggest they are most likely involved with structural or physiological functions common to tropical sharks. Conversely, FAs in higher quantities in

‡‹–Š‡”–Š‡—• Ž‡‘”ˆ‹ȋ‡Ǥ‰Ǥǡͳ͹ǣͲǡʹʹǣͶɘ͸ƒ†ʹͲǣ͵ɘͻȌ ‘—Ž† be linked to specific structures, physiology or functions (e.g., locomotion) of those tissues

(Pethybridge et al., 2010) or indicate temporal differences in diet (discussed

„‡Ž‘™ȌǤ‘–ƒ„Ž›ǡ‘—”•–—†›•’‡ ‹‡•ǯ—• Ž‡–‹••—‡•™‡”‡†‘‹ƒ–‡†„› ǡƒ• has been found in the Port Jackson Shark Heterodontus portusjacksoni

50 (Beckmann et al., 2014b) and deep water shark species (Pethybridge et al.,

2010). Polyunsaturated FAs also dominates in the sub-dermal tissue of the Reef

Manta Ray Manta alfredi and the Whale Shark Rhincodon typus (Couturier et al.,

2013b) and, typically in the muscle tissues of teleost fish (Belling et al., 1997;

Økland et al., 2005). In contrast, shark liver tissue, which has been shown to be more representative of diet (Beckmann et al., 2014b), is typically dominated by energy-rich MUFA.

•‹‰•‹‰ƒ–—”‡ ƒƒŽ›•‹•–‘„‡––‡”—†‡”•–ƒ†ƒ•’‡ ‹‡•ǯ–”‘’Š‹ ‡ ‘Ž‘‰› should take into account known trophic markers and EFA, particularly if they show highly variable patterns among tissues types. Commonly used estuarine- based trophic markers, detected in this study that were variable between fin clips and muscle, included those produced by bacteria (17:0, i17:0), diatoms, algae, mangroves and terrestrial plants (18:2Z6, 20:4Z3, 20:4Z6, and 20:5Z3), and dinoflagellates (22:6Z3; (Alfaro et al., 2006; Kelly and Scheibling, 2012;

Sargent et al., 1989)). Many other FAs are considered to be trophic markers for particular taxon or trophic groups and were also variable between the fin clips and muscles. For example, 18:1Z7 is characteristic of bacteria (Kelly and

Scheibling, 2012), 20:1Z9, 20:1Z11 and 22:1Z11 of copepods (Falk-Petersen et al., 2002; Kelly and Scheibling, 2012), 16:1Z7 of diatoms and mangrove (Kelly and Scheibling, 2012; St. John and Lund, 1996), and 22:0 and 24:0 of mangrove and terrestrial plants (Joseph et al., 2012; Rossi et al., 2008). That these particular FAs were variable between the tissue types indicates tissue differences, however the fact that these known markers were found in the fins supports their utility for dietary studies.

51

Determining the importance of FA profile differences between fins and muscle for dietary analysis requires differentiation between FAs that are assimilated

ˆ”‘ƒ‹†‹˜‹†—ƒŽǯ•†‹‡–ȋ•— Šƒ• •Ȍˆ”‘–Š‘•‡’”‘†— ‡†de novo (Tocher,

2003). Essential FA profiles found in muscle and fin tissue of these tropical shark species were dominated by the Z6 FAs, which are formed through the linoleic pathway. In this pathway, 20:4Z6 is elongated to 22:4Z6 (Tocher, 2010) and as there are only small amounts of precursors to 20:4Z6 it is likely that it has been accumulated by diet. Importantly, the differences between tissues in

20:4Z6 and 22:4Z6 were proportional across tissues within species, suggesting similar processes are occurring in the fin and muscle. These processes may be occurring at different rates since 20:4Z6 in G. garricki was the only significantly different EFA in univariate analysis. As only one EFA differed the combination of non-essential FAs may be more important in influencing differences than individual EFAs. Therefore the lack of significant differences between most fin and muscle EFAs, the low r values in the ANOSIM and that similar processes are likely occurring in fin and muscle suggests that both tissue types are appropriate for trophic studies.

Variation in a range of FAs among tissue types can indicate variable uptake of particular tissues over time. For example, the EFA 20:3Z9 was a major contributor to differences between fin and muscle in both C. leucas and G. garricki. This unusual FA has also been detected in some C. leucas in the Florida

Everglades and, along with other Z6 and Z3 PUFA were linked to deficiency in

EFA in these sharks (Belicka et al., 2012a). It was also found that 18:1Z9

52 contributed to the dissimilarity of fin and muscle FA profiles in C. leucas and G. garricki. Present in high relative levels in a range of organisms, this FA can often be an indication of carnivory (Falk-Petersen et al., 2002; Kelly and Scheibling,

2012).

The fins in all species did accumulate FAs that are linked to diet and many of the

FAs, particularly the EFAs, varied between the fin clips and muscle in similar ways. This suggests that the same processes are occurring in both tissues.

Differences in the FA profiles of various elasmobranch tissues is now becoming well established (Beckmann et al., 2013; Pethybridge et al., 2010), with the first controlled experiments indicating the uptake of FA can vary considerably across shark muscle, liver and blood serum (Beckmann et al., 2014b).

Saturated FAs (SFA), such as 16:0 and 18:0, also contributed to differences between fins and muscle in C. leucas (and to some extent G. garricki), which is interesting because these SFA are ubiquitous in animals and variations are expected among tissue types according to rates of cellular metabolism (Tocher,

2003). Most fin tissue is cartilage, and so would be expected to have slower metabolism and tissue turnover rates than muscle (Malpica-cruz et al., 2012).

Certainly, studies measuring stable isotopes have found that cartilage and fin have a slower turnover rate than muscle and blood (MacNeil et al., 2006;

Malpica-cruz et al., 2012). It is therefore likely that the FA profiles of fins are representing another time period in the diet and habitat usage of these sharks.

Such variances in FA profiles among fins and muscle could be particularly useful in providing scientists with key insights into the trophic ecology of species

53 occupying dynamic tropical river environments that experience a monsoonal wetȂdry cycle (Warfe et al., 2011).

This study found highly variable amounts of total FA in the muscle and fin both within and between species emphasising the importance of adequate sample sizes. Researchers could maximise the utility of such tissue samples in rare/threatened species, especially when sampling adults with larger shark fins, as some of the muscular tissue layers could be dissected and used to obtain stable isotope evidence (Hussey et al., 2011a). Moreover, comparisons could be made between muscle tissue profiles and connective tissue/cartilage profiles to explore temporal differences.

Apart from intraspecific differences across FA profiles there were also interspecific differences such as the variation in 20:4Z6 across species. These differences may be indicative of dietary and perhaps environmental change as

Z6 have been identified as environmental indicators of temperature and increases in the relative amounts of the FA, 20:4Z6, and dominance of Z6 pathways have been linked to tropical waters (Couturier et al., 2013b; Sinclair et al., 1986). Furthermore, experimental work with seals and salmon found

18:1Z9 was assimilated into muscle and adipose fins directly from their diet

(Budge et al., 2004; Skonberg et al., 1994). Therefore the differences in the amount of 18:1Z9 in these shark species may suggest separation between their trophic levels. Since more 18:1Z9 was found in the muscle than the fin, this could indicate an increase in consumption of higher order consumers with age.

54 It could, however, also be due to de novo synthesised 14:0 and 16:0 (Dalsgaard et al., 2003).

Ontogeny, sex-based physiology and different movement patterns can all be reflected in FA profiles of different tissues (Belicka et al., 2012a; Parrish et al.,

2013). All the sharks studied here were juvenile to sub-adult individuals and as such were not sexually mature, with some individuals showing open umbilical scars indicating they were neonates, which implies a short period of active feeding. Consequently, it is highly likely that the fins of some small individuals

(e.g. <100 cm total length) may be reflecting a stronger maternal signature than muscle tissue, due to differences in metabolism and structural turnover among the two tissues types (Belicka et al., 2012a). Such effects may also explain some of the high degree of variation found within species, as these sharks were not only sampled from different stages of ontogeny, but also across a range of seasons (Sargent et al., 1999b; Tocher, 2010). While it is difficult to obtain a fully replicated stratified sample of tissues among a range of developmental stages and body sizes in rare and/or difficult to sample animals, the potential for ontogenetic and sex-based influences upon FA profiles should be considered in future studies, where possible.

2.6 Conclusion

An understanding of differences in FA profiles obtained from different tissue types is important when utilizing FAs to elucidate the trophic ecology of higher order consumers such as sharks. Fatty acid profiles in the fins and muscles

55 reflected FAs, which have previously been used as biomarkers in trophic studies of marine predators (Dalsgaard et al., 2003; Kelly and Scheibling, 2012). Similar proportions of dominant FAs, particularly EFAs, were found to occur among the muscle and fin tissues from these tropical euryhaline shark species, along with some strong similarities between the two Glyphis species (which potentially could be explained by their genetic similarity (Wynen et al., 2009). Collectively, this suggests comparable assimilation and usage processes may be occurring in both tissue types for these major FAs. Whilst muscle and fins are not directly interchangeable in dietary analyses, both tissue types have measurable quantities of dietary EFAs in the FA profiles of both tissues, suggesting that diet is being reflected and should have utility in future shark trophic studies.

Slight differences in the proportion of some EFAs within the different tissue types can provide key opportunities (e.g., temporal hindcasting of seasonal prey consumption), but also signal caution in applying these analyses to understanding patterns of diet. As fins consist of multiple tissues, each tissue type may have slightly different proportions of FAs dependent on the physiological needs of that tissue as compared to muscle where only one tissue type is present. Temporal variations in habitat usage and ontogeny will be reflected at different time scales of tissues due to turnover rates of FA that are not yet well understood. A priority for future research should be exploring links between FA profiles in these tissues and rates of assimilation in the various chondrichthyan tissues, to provide opportunities for temporal exploration of diet. Where possible, this should also include investigation of potential prey sources in controlled settings to validate the dietary links and examine FA

56 synthesis pathways. What is clear is the need for further work on elucidating fine scale differences between tissues in order to determine the suitability of tissue FA analysis for dietary studies.

2.7 Chapter Acknowledgments

This research was conducted on the traditional country of the Bininj and

Mungguy people. We thank Peter Nichols, Peter Mansour, Grant Johnson, Mark

Grubert, Duncan Buckle, Roy Tipiloura, Dominic Valdez, Francisco Zillamarin, field volunteers, Edward Butler, colleagues in the ATRF and the crew of R.V.

Solander for fieldwork and laboratory assistance. Charles Darwin University,

CSIRO and the North Australia Marine Research Alliance (NAMRA) provided funding, alongside collaborative partnerships in the Marine Biodiversity and

‘”–Š‡”—•–”ƒŽ‹ƒ —„•‘ˆ—•–”ƒŽ‹ƒ ‘˜‡”‡–ǯ•ƒ–‹‘ƒŽ˜‹”‘mental

Research Program (NERP)

57 3 Niche metrics suggest euryhaline and coastal elasmobranchs provide trophic connections among marine and freshwater biomes in northern Australia

This chapter has been published in Marine Ecology Progress Series

Every SL, Pethybridge HR, Fulton CJ, Kyne PM, Crook DA (2017) Niche metrics

suggest euryhaline and coastal elasmobranchs provide trophic connections

among marine and freshwater biomes in northern Australia. Mar Ecol Prog

Ser 565:181Ȃ196

Below is a list of the co-authors and their contribution. All authors contributed to concept design and writing.

Heidi Pethybridge Ȃ Contributed knowledge of multivariate and lipid and stable isotope analysis, fish biology and ecology and developing and editing manuscript.

Christopher Fulton provided advice and assistance with statistical analyses and data presentation, as well as assistance with manuscript planning and editorial comments on the manuscript text.

Peter Kyne Ȃ Contributed knowledge of elasmobranch biology and ecology, and assisted with fieldwork and tissue collection.

David Crook Ȃ Contributed knowledge of fish biology and ecology and statistical analysis of fish and riverine ecology and editing manuscript.

58

3.1 Abstract

Tropical elasmobranchs could play significant roles in connecting coastal and river ecosystems, yet few studies have explored the trophic ecology of elasmobranch species that may link these biomes. We investigated the trophic niches of seven such species in northern Australia during the tropical monsoonal wet and dry seasons, using stable carbon (G13C) and nitrogen (G15N) isotopes (SI), and fatty acid (FA) biomarkers taken from muscle tissue. Both SI and FA metrics suggested significant niche partitioning between species, with two distinct guilds: a marine food web based on epiphytes and seagrass (low

G13C), and an estuarine/freshwater food web with a seston base (higher G13C). A large overlap in SI niche areas and higher mean trophic positions (4.1 Ȃ 4.8) were evident in species accessing marine diets (Carcharhinus leucas,

Rhizoprionodon taylori) when compared to species predominantly feeding in estuaries (3.2 Ȃ 3.6; Glyphis garricki, G. glyphis). Across all seasons, G. garricki had the greatest FA niche space, and variable overlap with two other species (R. taylori, C. leucas). Although limited seasonal effects were apparent for individual

FA biomarkers, SI niche metrics revealed greater niche areas and inter-specific partitioning during the dry season for three species. Subtle differences in niche metrics derived from SI and FAs were likely due to disparate turnover times, and the statistical approach of each metric (two-dimensional versus multi- dimensional). Collectively, our analyses suggest these tropical coastal and euryhaline elasmobranchs consume prey from a range of sources to provide trophic connections across marine, estuarine and freshwater biomes.

59

3.2 Introduction

Resource partitioning, whereby species vary their use of habitat and dietary items over space and time (Ross 1986), is fundamental to our understanding of coastal ecosystem structure and function (Kitchell et al. 2002). As elasmobranchs (sharks and rays) have key roles in aquatic food webs (Stevens et al. 2000) and are demographically vulnerable predators (Dulvy et al. 2014) it is important to understand how they utilize these resources. Elasmobranchs are typically mid to upper level predators, and can provide ecosystem stability by influencing the abundance and health of prey species at multiple trophic levels, and connect otherwise distinct food webs (Rooney et al. 2006, Heithaus et al.

2013). To understand these trophic interactions, it is important to define elasmobranch niches, connections to different biomes, seasonal shifts, and the potential for dietary overlap between species.

Niche theory suggests species use a set of resources that form their unique space, an Dzn Ȃ †‹‡•‹‘ƒŽ‹ Š‡Š›’‡”˜‘Ž—‡dz(Hutchinson 1957), which may be measured and compared among locations, seasons and species. In some cases, species can share resources to produce niche overlap. Whether such overlap leads to competition is dependent on the spatial and temporal extent of shared resource use. In the case of trophic overlap, sympatric species may target different prey items (Yick et al. 2011), and/or have temporal or spatial differences in consumption that negate competitive interactions (Ross 1986).

Indeed, the extent of trophic overlap has been found to be highly variable in a

60 number of elasmobranch assemblages (Papastamatiou et al. 2006, Yick et al.

2011, Tilley et al. 2013, Heithaus et al. 2013). Such variation demonstrates the differences between the trophic resource use of elasmobranchs and highlights the importance of determining diet for each species, particularly in assembelages where there are limited data, such as in species that are rare, threatened and/or data deficient, for instance, euryahline elasmobranchs

(Lucifora et al. 2015).

Euryhaline elasmobranchs are capable of tolerating a wide range of salinities and may complete different life history stages in marine, estuarine or freshwater habitats. Tropical regions are particularly important in this regard, given the high diversity of euryhaline elasmobranchs in these biomes (Lucifora et al. 2015). Due to their reliance on rivers and estuaries, euryhaline elasmobranchs must adjust to fluctuations in salinity and turbidity over tidal and seasonal cycles. These fluctuations affect not only their physiology, but also the abundance of their potential prey. Environmental changes are particularly pronounced in tropical rivers due to the variation between high-flow

‘•‘‘ƒŽǮ™‡–ǯ•‡ƒ•‘• and low-”ƒ‹ˆƒŽŽǮ†”›ǯ•‡ƒ•‘•(Douglas et al. 2005,

Warfe et al. 2011). Such changes may affect the utilization of trophic resources and partitioning within tropical elasmobranch assemblages that are poorly understood at present. However, recent studies have indicated the existence of dietary overlap among juvenile Pigeye Shark Carcharhinus amboinensis and Bull

Shark C. leucas in northern Australia (Tillett et al. 2014) and among C. leucas and Largetooth Sawfish Pristis pristis in Western Australia (Thorburn et al.

2014).

61

Traditional trophic niche metrics reflecting diet breath or richness are based on prey composition data from stomach content analysis, which can be invasive or fatal, often requires large sample numbers, and provides only a brief temporal snapshot of diet (Hussey et al. 2012). By analysing small tissue samples for biochemical tracers, a more time-integrated assessment of trophic resource use can be obtained in a relatively non-invasive manner. Stable isotopes (SI) and fatty acids (FAs), are biotracers found in animal tissues that can be analysed to estimate time-integrated trophic niches (Hussey et al. 2011, Jackson et al. 2011,

Sardenne et al. 2016). An understanding of the fractionation of stable isotopes

(e.g., G13C, G15N) through food webs, allows isotopes to be used for a variety of applications, including estimation of niche area, resource partitioning and dietary overlap. Specifically, values of G13C may be used to identify species habitat associations through the comparison of G13C values with primary producers. The base of a food web may consist of C3 or C4 plants, bacteria or detrital matter, each having different biochemical pathways (Peterson & Fry

1987). These differences may result in G13C values that can be used to determine which species utilize those food webs. Values of G15N can estimate trophic position, because G15N is passed up food chains via the consumption of proteins so that higher order predators tend to be more enriched in G15N (Hussey et al.

2012).

Fatty acids can also be traced through food webs, and are particularly useful for detecting basal sources and trophic interactions via essential FAs (EFAs) that can only be obtained through dietary sources (Iverson 2009). These EFAs are

62 synthesized by different primary and secondary producers such as dinoflagellates, diatoms, algae or during microbial and bacterial processes

(Dalsgaard et al. 2003, Parrish 2013). Fatty acids often have faster turnover rates than isotopes (e.g. 14 weeks in shark muscle tissue (Beckmann et al.

2014b), cf. 6 Ȃ 12 months for G13C and G15N, (Malpica-cruz et al. 2012, Hussey et al. 2012), and have been used to reveal physiological and environmental changes over relatively fine scales, such as the EFA trophic biomarker omega 3

[Z3] / omega 6 [Z6] which can be used to find seasonal differences within species. For example, Albacore Tuna Thunnus alalunga caught in waters off eastern Australia had an increased ratio of Z3/Z6 with decreasing water temperatures (Pethybridge et al. 2015). The ratio of Z3/Z6 has also been shown to be influenced by changes in salinity, either increasing (e.g. Sturgeon

Acipenser naccarii, Martínez-Álvareza et al. 2005) or decreasing (e.g. White-edge

Freshwater Whipray Himantura signifer, Speers-Roesch et al. 2008) with increasing salinity. Fatty acid 18:2Z6 (LA) is another biomarker that may be useful in estuarine environments, as it can indicate terrestrial sources when compared to marine phytoplankton (Napolitano et al. 1997, Budge & Parrish

1998).

In this study, we use SI and FA biochemical tracers to evaluate trophic resource partitioning and seasonal variation in trophic resource utilization among seven species of euryhaline and coastal elasmobranchs in the wet-dry tropical region of northern Australia. Northern Australia provides an excellent setting for studying elasmobranch trophic ecology due to the relatively pristine state of the estuarine and riverine ecosystems (Warfe et al. 2011) and the high diversity of

63 sympatric elasmobranch species (Last 2002). The objectives of our study were to use SI and FA tracers to: (i) estimate and compare trophic position estimates among species, (ii) identify basal source contributions (marine, estuarine, freshwater), and (iii) apply biochemical niche metrics (measurement of the distance between biochemical tracers) (Jackson et al. 2011, Swanson et al.

2015) to determine the extent of dietary overlap within and among species and across seasons.

3.3 Methods

3.3.1 Site description, sample collection and preparation

Seven species of euryhaline and coastal elasmobranch were collected in the

South Alligator River, Northern Territory, Australia from March 2013 to July

2014 (Table 3.1, Fig. 3.1). This is a relatively pristine, macro-tidal river system with a ~ 5 km wide mouth (Wolanski & Chappell 1996), and tidal influence to

100 km upstream (Wolanski & Chappell 1996), where it narrows to ~ 20 m wide. The mean channel depth is ~7.5 m with a largely muddy bed (Wolanski &

Chappell 1996), and mangroves dominate the estuary up to the limit of tidal influence (Mitchell et al. 2007). Strong seasonal fluctuations in rainfall, which can be broadly classified as wet season (November to April; 2013 Ȃ 2014 mean total (± standard error, SE) rainfall 241.3 ± 36.2 mm/season and dry season

(May to October; 2013 Ȃ 2014 mean total rainfall 18.2 ± 12.7, mm/season

(Australian Government Bureau of Meteorology, 2016)), contribute to large changes in river hydrology, turbidity and salinity. The mouth of the river had mean salinity values (± standard deviation, SD) of 34.5 ± ͲǤʹΩȋ†”›Ȍƒ†ͳ͹Ǥͳ±

64 4.3Ωȋ™‡–Ȍǡ™Š‡”‡ƒ•‹–Š‡‹†-lower region of the river it was 21.9 ± 5.3Ω

(dry) and 0.4 ± ͳǤͷΩȋ™‡–ȌǤ

65 Table 3.1 Number (n) and total length (TL) of seven euryhaline and coastal elasmobranch species from the South Alligator River, Kakadu National Park, Australia from which muscle tissue samples were taken for stable isotope (SI) and fatty acid (FA) analysis. Included is the division of species between the wet and dry seasons and the sex ratio male to female.

Total n Range TL (cm) Sex ratio Habitat n M:F Species SIA FAA

Scientific Name Common Name Wet Dry Wet Dry SIA FA SIA FA Coastal / marine Carcharhinus Pigeye Shark 2 0 2 0 1 68 68 1:1 0:1 amboinensis Carcharhinus leucas Bull Shark 34 28 6 20 2 72-93 72-139 18:16 13:9 Euryhaline Glyphis garricki Northern River Shark 42 22 20 12 13 56-141 45-61 23:19 15:10 Euryhaline Glyphis glyphis Speartooth Shark 8 2 3 2 6 71-120 78-136 1:4 2:6 Euryhaline Urogymnus dalyensis Freshwater Whipray 2 0 2 0 2 56-110 56-110 0:2 0:2 Euryhaline Pristis pristis Largetooth Sawfish 2 2 0 0 0 87-96 - 1:1 0:0 Euryhaline Australian Sharpnose Coastal/marine Rhizoprionodon taylori 38 8 30 1 28 34-87 34-87 11:27 23:6 Shark

66

Four main sampling sites (Fig. 3.1) were chosen along the river to encompass a range of distances from the river mouth and where elasmobranchs were captured (limited by the rarer species). As such, these sites included a spectrum of salinities (see Fig. 3.1), and habitat characteristics (e.g. distance from mouth, river flow, width of river, abundance and species composition of riparian vegetation). A range of techniques was used at each site to capture species and obtain tissue samples during the live capture and release of individuals in the wild. Rhizoprionodon taylori and Carcharhinus amboinensis were captured with baited line in the lower estuary and coastal region of the South Alligator River,

Kakadu National Park (Fig. 3.1). Northern River Shark Glyphis garricki,

Speartooth Shark G. glyphis, C. leucas, Pristis pristis and Freshwater Whipray

Urogymnus dalyensis were caught in mid-lower estuarine reaches with a combination of 10.2 to 15.2 cm gill nets and baited lines.

67

Figure 3.1 Map of the South Alligator River, Northern Territory, Australia, showing capture locations of elasmobranchs. Dotted lines represent the approximate boundary of each sampling site. Inset shows where the river is in relation to northern Australia. (Background map from Stamen Toner, and formed using QGIS Development Team, (2015)). Red (blue) text indicates dry (wet) •‡ƒ•‘•ƒŽ‹‹–›ȋΩȌ˜ƒŽ—‡•Ǥ

68 A 5 mm biopsy punch (Stiefel, USA) was used to collect muscle tissue from between the second dorsal and the caudal fin, just anterior and lateral of the caudal peduncle. Total, standard and precaudal length and clasper size were also measured and umbilical scarring in neonates were noted. Within 5 minutes of capture, the tissue sample was immediately placed on liquid nitrogen (-

196°C) for preservation and initial storage in the field. Salinity and other water quality parameters were measured at each site using a hand held multi- parameter instrument (Sonde 6000, YSI Inc. Xylem, USA) (Fig. 1) during each sampling event.

Within one week, tissue samples (mean wet mass 30.0 ± 20.0 mg) were transferred to a -20°C freezer, and then later freeze-dried (Alpha 1-4 LSC, Christ,

Germany) at -20°C for 21 hours and then -30°C for three hours. To avoid degradation of the sample from defrosting and refreezing, frozen muscle samples were dissected in the freezer to remove dermal layers and as much connective tissue as possible. Where sample tissue masses were low, tissue was only used for SIA and not FA analysis (Table 3.1) except for G. glyphis where the opposite was true. Mean tissue (±SD) wet mass for SIA 2.0 ± 1.0 mg and FAs

36.0 ± 28.0 mg.

3.3.2 Stable isotopes

Muscle tissue was rinsed in milli-Q water and sonicated to remove excess urea as per Kim & Koch (2012). Tissues were then freeze-dried to a constant mass and pulverized using a combination of micro-scissors and micro-pestle or a

69 coarse pestle and ceramic mortar. A subset of material was weighed to 400 to

1000 Pg in tin cups for combustion in a Sercon Europa EA-GSL elemental analyser (Sercon Ltd, UK), which were then analysed with a Sercon Hydra 20-22 duel-inlet gas isotope ratio mass spectrometer (Sercon Ltd, UK) at the

Australian Rivers Institute, Griffith University, Queensland. The international standards used to determine the relative G13C and G15N were Peedee Belemnite

Carbonate and Atmospheric Nitrogen with a precision of (±1 SD) 0.03 and

ͲǤͲͻΩˆ‘”G15N and G13C, respectively. We did not use mathematical models to correct G13C values for lipids as lipid content in all samples was deemed to be low as inferred by the percent ratios of carbon to nitrogen (C:N; mass percent composition of tissue sample) being < 3.5 (Table 3.2, Post et al. 2007) and by the low lipid levels inferred from total FAs (Every et al. 2016).

Trophic positions (TPs) for each species were calculated using narrowing discrimination with increasing dietary G15N values (Hussey et al. 2014a, b). This works on the concept that as trophic position (TP) increases, the dietary discrimination factors (the increase of G15N at each TP) decreases (Hussey et al.

2014a, b). Popeye Mullet (Rhinomugil nasutus, G15N ͸Ǥ͸άͳǤʹΩ, TP 2.9; Froese

& Pauly 2015) was used as the baseline consumer collected at the same location and over the same time period as the elasmobranchs. Popeye Mullet were chosen because they have a wide distribution throughout the river, covering a range of salinities and habitat type (Carpenter & Niem 1999). Also their G15N values were similar regardless of where they were caught within the river, negating the need to account for differing baselines. Finally we suggest that the

TP of 2.9 for R. nasutus is a reasonable estimate as other Mugilidae grazersȄ

70 Mugil bananensis, M. curema and Liza falcipinnis Ȅall have TPs of 2.8, 2.7 and

3.0, respectively (Faye et al. 2011) and feed in a manner similar to R. nasutus, consuming algae and insects along the water surface and bank edges (Faye et al.

2011). The following equation was used with ߚ଴and ߚଵvalues from the meta- analysis of Hussey et al. (2014a, b):

Ž‘‰ሺߜଵହܰ െ ߜଵହܰ ሻ െ Ž‘‰ሺߜଵହܰ െ ߜଵହܰ ሻ ܶܲ ൌ ௟௜௠ ௕௔௦௘ ௟௜௠ ்௉ ൅ ܶܲ ݇ ௕௔௦௘

where TPbase ‹•–Š‡–”‘’Š‹ ’‘•‹–‹‘‘ˆDz„ƒ•‡Ž‹‡dz ‘•—‡”ȋR. nasutus) and

15 15 15 ଵହ G NTP is the consumer G N value. k (rate at which G NTP approachesߜ ܰ௟௜௠ is calculated via the formula:

ଵହ ሺߚ଴ െ ߜ ܰ௟௜௠ሻ ݇ ൌ െŽ‘‰ ଵହ െߜ ܰ௟௜௠ where term:

ଵହ െߚ଴ ߜ ܰ௟௜௠ ൌ ߚଵ

Here, ߚ଴is the slope, ߚଵ is the intercept and k is calculated from ߚ଴ = 5.9 [4.5,

7.3] and ߚଵ= -0.3 [-0.4, -0.1] (highest posterior density median [95% uncertainty intervals]) from Hussey et al. (2014a, b).

3.3.3 Fatty acids

Lipid was extracted quantitatively using the modified Bligh & Dyer (1959) method (detailed in Every et al. 2016), which utilizes an overnight one-phase extraction process of methanol: dichloromethane (DCM): milli-Q water (2:1:0.8 by volume). The following day DCM and saline milli-Q water were added so that

71 the final volume was 1:1:0.9. The lower phase was removed into a round bottom flask and solvents were evaporated using a rotary evaporator in a bath of 40°C.

Remaining lipid was transported with DCM into a pre-weighed vial, blown down with nitrogen gas and dried to a constant weight. All vials were made up to final concentration of 10 mg lipid to 1.5 ml DCM and were stored in a -20°C freezer until further analysis within a couple of days.

To liberate the FAs from the lipid backbone the total lipid extract was transmethylated based on the methods in Parrish et al. (2015) and Miller et al.

(2006). Briefly, 50 Pl of the extract was transferred to a test tube and blown down with nitrogen gas before adding 3 ml of a methylating solution (methanol, hydrochloric acid and chloroform at a ratio of 10:1:1 volume). After heating for

2 hours at 100°C and cooled, 1 ml of milli-Q water was added to the test tube, and then extracted with 1.8 ml of hexane and chloroform (4:1), vortexed and centrifuged for 5 min. The upper, organic layer containing the fatty acid methyl esters (FAME) was placed into a clean vial. This extraction process was repeated three times. After adding a known amount of internal standard (C23 or C19

FAME), 0.2 Pl was injected into an Agilent Technologies 7890B gas chromatograph (GC) (Palo Alto, California, USA) equipped with an Equity-1 fused silica capillary column (15 x 60.1 mm internal diameter and 0.1 Pm film thickness), a flame ionization detector, a splitless injector and an auto-sampler.

Peaks were quantified using Agilent Technologies ChemStation software (Palo

Alto, California, USA). Confirmation of peak identifications was by GC-mass spectrometry (GC-MS), using a column of similar polarity to the Equity-1 fused silica capillary column and a Finnigan Thermoquest DSQ GC-MS system.

72

3.3.4 Statistical Analysis

Analysis of variance (ANOVA) was used to determine differences in stable isotopes among the fixed factors of species and seasons. Due to a significant relationship between G13C and total length (TL) in C. leucas (which was not evident in the other taxa), an analysis of covariance (ANCOVA) was run with total length (TL) as a covariate to account for this species. Assumptions of normality, variance, homogeneity of slopes and colinearity were tested through visual inspection of boxplots, calculating Leverage, ‡•‹†—ƒŽ•ƒ†‘‘ǯ•D and graphed in a Residual vs Fitted, Normal Q-Q and Residual vs Fitted values plots prior to accepting the models. No transformations were necessary.

Fatty acids were calculated as a percentage of total FAs, and those with group means < 0.5% were excluded from further analyses. To evaluate interspecific differences in FA profiles among species, a permutational multivariate analysis of variance (PERMANOVA) was applied to a Euclidean distance matrix comprised of the FAs (variables) for all the elasmobranch samples, using Type

III sum of squares (partial) and a minimum of 9900 unique permutations. To determine which FAs contributed most to the observed interspecific differences, an analysis of similarity (SIMPER) was conducted. To explore FA profiles between the species and season, a principal coordinates analysis (PCA) was constructed and arranged into a difference matrix based on Euclidean distances.

PERMANOVA analysis was used to explore the significance of seasonal effects in

G. garricki (the only species with sufficient seasonal replication, Table 3.1).

73 ANOVA was then used to explore interspecific differences in Z3/Z6 between G. garricki, G. glyphis, C. leucas and R. taylori. As significant effects were found, we then used an ANCOVA ™‹–Šƒ•ƒ ‘˜ƒ”‹ƒ–‡ǡ„‡–™‡‡‡ƒ Š•Šƒ”ǯ•Z3/Z6 ratio to consider possible allometric effects. 18:2Z6 (LA) was also explored for interspecific differences to determine which species of shark may have the most terrestrial input, which is suggestive of floodplain resources (Napolitano et al.

1997).

Univariate analyses were performed in R using the R core packages (R

Development Core Team 2014), SIBER (Jackson et al. 2011) and one FA multivariate analysis used the package NicheROVER to determine overlap between sharks (Swanson et al. 2015). All other multivariate analyses were performed in PRIMER (v6) and PERMANOVA+ (v1) (Clarke & Gorley 2006). All p-values < 0.05 were considered significant.

3.3.5 Niche area calculations

Using the R package Stable Isotopes Analysis in R (SIAR) a SI Bayesian ellipse

‘†‡Ž™ƒ• ƒŽ —Žƒ–‡†ƒ†–Š‡ ‡ŽŽ‹’•‡ƒ”‡ƒȋΩ2) was corrected for sample size

(SEAC) (Jackson et al. 2011). SEAC was calculated from the covariance matrix determined via Bayesian inference. Another metric of isotopic niche areas, the

‘˜‡šŠ—ŽŽ–‘–ƒŽƒ”‡ƒȋΩ2) (TA) was also calculated (Layman, et al. 2007a,

Jackson et al. 2011) however this was not compared across seasons due to low sample numbers. Although TA in a G15N Ȃ G13C biplot is strongly affected by sample size (Jackson et al. 2011, Syväranta et al. 2013), it was calculated to

74 explore total niche width as a proxy for the extent of trophic diversity within a food web, presence of outliers, and a conservative estimate of maximum potential overlap among species (Layman et al. 2012a, Heithaus et al. 2013).

Drawing SEAC in isotopic space (G13C and G15N) that incorporates ~40 % of the sampled individuals is thought to provide a more accurate measure of niche width than TA, and is less biased by sample size (Jackson et al. 2011, Syväranta et al. 2013). Ellipse confidence intervals for the area of the ellipses were calculated from Bayesian likelihoods, whereby the probability of one ellipse being bigger than the other can be determined by the uncertainty in the probable values. Overlaps of SEAC were calculated by finding the area of species

Dzƒdz‡ŽŽ‹’•‡ƒ†–Š‡ƒ”‡ƒ‘ˆ–Š‡‘˜‡”Žƒ’‘ˆ•’‡ ‹‡•Dz„dz™‹–Š•’‡ ‹‡•Dzƒdzƒ†–Š‡ converted to a percentage. This could only be performed for the three species C. leucas, G. garricki and R. taylori, which total n values for samples over both seasons were >30, which is considered optimum for these analyses (Syväranta et al. 2013). It should be noted that samples were not even across seasons, with wet:dry sampling for C. leucas and R. taylori being 28:6 and 8:30, respectively.

As such, the C. leucas dry season ellipse may be slightly biased due to their small n value (8 is considered the minimum; Jackson et al. 2011), yet it was included to provide a descriptive comparison. Glyphis garricki seasonal samples were relatively even 22:20 wet:dry. Ellipses were also calculated for the combined sampling period to ensure that these results were not overly biased.

Fatty acid niche space and overlap was explored by calculating hyper-ellipses, using a multivariate extension on the bivariate approach of Jackson et al. (2011) described above. Here, Bayesian priors are used to determine the hyper-volume

75 and the probability of finding one species in another species niche space

(Swanson et al. 2015). To avoid degenerate covariance matrices in the absence of prior information, the number of niche dimensions fitted from the dependent variables should be lower than the number of sample observations (Swanson et al. 2015). Therefore, only the five most abundant EFAs were selected (20:4Z6,

20:3Z9, 22:5Z6, 22:6Z3 and 22:4Z6), given the sample numbers (~20) for each of the three shark species analysed (G. garricki, C. leucas, R. taylori). Each EFA was then represented as an axis, calculated in FA hyperspace and projected onto a two dimensional plot as a probabilistic projection to display FA niche space.

Niche space was calculated based on the volume of ellipses, under the assumption they were multivariate normal. This differs from niche area calculations for SIs, as they are not multidimensional and area is calculated from two dimensions. Percentage overlap was calculated based on the probability that one species would be found in the niche of another species.

Throughout this chapter we refer to niche metric as the measurement of both the two-dimensional bivariate aspect of niche, niche area (as for SI), and the multivariate niche space (as for FAs).

3.4 Results

3.4.1 Stable isotope values, trophic position estimates, and niche metrics

Muscle G13C and G15N values in the seven elasmobranch species revealed two distinct trophic guilds (Fig. 3.2). The highest values of both elements were in C. leucas, R. taylori, C. amboinensis and P. pristis (Table 3.2, Fig.3.2). The

76 elasmobranch guild with the lowest mean G13C and G15N values included G. glyphis, U. dalyensis and G. garricki (Table 3.2, Fig. 3.2). Taxa had a significant effect on G15N values (p < 0.01, Fdf = 16.61, 115, R2 = 0.3) and G13C values (p < 0.01,

Fdf = 50.71,115, R2 = 0.6). The post hoc test for G15N showed no significant differences between R. taylori and C. leucas, and between G. garricki and G. glyphis, whereas all others species differed significantly. For G13C values, only G. garricki and G. glyphis were not significantly different from each other (Table

3.3).

Figure 3.2 Biplot of mean (±SD) G13C and G15N values for seven species of euryhaline and coastal elasmobranchs from South Alligator River, Kakadu National Park, Australia. Primary producer values along the x Ȃ axis are G13C ranges for seston, epiphytes and seagrass from the combined wet and dry season, Embley River Estuary, Gulf of Carpentaria (Loneragan et al., 1997).

77 Table 3.2 Mean (±SD) values of fatty acids (FA) (% mean of the relative abundance of FA> 0.5%), G13Ωƒ†G15Ω‹—• Ž‡–‹••—‡ˆ”‘•‡˜‡ elasmobranch species collected from the South Alligator River, Kakadu National Park, Australia. Included are calculations for trophic position (TP), total area of •–ƒ„Ž‡‹•‘–‘’‡ ‘˜‡šŠ—ŽŽ•ȋΩ2), ellipse area of stable isotopes (SEAC Ω2) and essential fatty acids (EFA) niche space (shaded grey). SFA: saturated fatty acids; MUFA: monounsaturated fatty acids; PUFA: polyunsaturated fatty acids; wet:dry: wet and dry seasons

SI C. amboinensis C. leucas G. garricki G. glyphis U. dalyensis P. pristis R. taylori G13C -13.4 -14.9±2.3 -18.9±1.3 -18.9±1.7 -20.2 -13.4 -15.1±1.3 G15N 11.7 12.95±2.8 9.3±1.9 7.6±3.6 8.9 12.6 11.1±2.5 C:N 2.6 2.8±0.2 2.8±0.2 2.7±0.1 3.2 2.7 2.8±0.4 TP 4.2 4.8±0.9 3.6±0.4 3.2±0.8 3.4 4.5 4.1±0.6 SEAC 18.5 6.9 10.3 (Wet:Dry) (16.9:17.1) (9.8:7.0) (2.1:11.7) TA 79.1 28.9 39.1 FA ȭ  42.5 31.9±8.7 30.4±8.1 27.5±7.2 30.3 43.6±12.3 ȭ  24.7 29.7±5.0 22.2±3.6 18.8±2.6 26.1 20.9±3.5 ȭ  26.2 30.3±2.2 38.4±3.4 44.0±4.3 34.7 30.9±3.4 ͳͺǣͳɘͻ 14.1 16.2±6.3 11.2±4.9 8.2±3.6 13.8±0.5 9.5±2.7 18:2bv 0.6 0.6±0.4 0.2±0.2 0.5±0.2 0.4±0.3 0.1±0.1 18:2cv 0.1 0.9±0.8 0.1±0.3 0.1±0.1 0.1±0.1 0.13±0.02 ͳͺǣʹɘ͸ 0.6 0.8±0.9 1.8±1.1 1.8±1.1 3.8±3.7 0.9±1.0 20:2v 0.2 2.8±2.2 0.5±0.8 0.3±0.2 0.2±0.1 0.1±0.1 ʹͲǣʹɘ͸ 0.4 0.5±0.8 0.9±0.4 0.8±0.2 0.3±0.0 0.6±0.2 ʹͲǣ͵ɘͻ# 0.8 7.6±6.2 1.4±3.4 0.3±0.3 0.3±0.1 0.1±0.1 ʹͲǣ͵ɘ͸ 0.2 0.4±0.3 0.8±0.4 0.6±0.3 0.3±0.1 0.4±0.2 22:3v 1.7 1.01±0.7 1.7±1.8 2.6±2.4 2.9±3 1.1±1.2 ʹͲǣͶɘ͸ 9.6 5.3±5.5 11.3±4.7 14.1±4.4 14.4±2.4 8.1±2.1 ʹʹǣͶɘ͸ 3.5 2.1±2.1 6.4±4.2 8.0±6.5 0.2±0.2 3.0±1.4 ʹͲǣͷɘ͵ 0.9 0.5±0.2 0.9±0.8 1.1±0.6 5.2±6.8 1.6±0.8 ʹʹǣͷɘ͵ 0.9 1. 9±1.2 1.5±2.0 1.6±1.8 1.1±1.6 1.1±1.7 ʹʹǣͷɘ͸ 1.8 1.2±0.7 2.3±1.1 2.3±1.1 1.2±0.8 2.2±0.8 ʹʹǣ͸ɘ͵ 4.8 4.8±2.3 8.3±4.6 9.8±8.8 4.3±0.2 11.4±5.6 ȭZ3/ȭZ6 0.4 0.70±0.4 0.5±0.6 0.4±0.8 0.5±0.0 0.9±1.4 EFA niche 13413.7 114886.8 455.6 space 40732.8: (Wet:Dry) 64824.2

# 20:3Z9 identified based on comparison with other C. leucas fatty acid literature; a standard was not available at the time of analyses. v = unable to identify bonds as standard was not available at the time of analyses.

78 Table 3.3 ANOVA and ANCOVA (shaded) of stable isotope values, the ratio of Z3/Z6 and the fatty acid 18:2Z6 from the muscle of Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori from the South Alligator River, Kakadu National Park, Australia ‘’ƒ”‡†–‘–Š‡Dz™‡–dzƒ†Dz†”›dz•‡ƒ•‘ǡ™‹–Š–‘–ƒŽŽ‡‰–ŠȋȌ–Š‡ ‘˜ƒ”‹ƒ–‡Ǥ‹‰‹ˆ‹ ƒ–˜ƒŽ—‡• are in bold.

Species Test Variable( Tracer DF F- P -Value Residual R2 t- Significant Tukey post hoc s) ratio standard value comparisons (p<0.05) Error All ANO Species G15N 115 16.6 < 0.01 2.5 0.3 G. garricki Ȃ C. leucas VA G. glyphis Ȃ C. leucas R. taylori- G. garricki R. taylori Ȃ G. glyphis G13C 115 50.7 < 0.01 1.7 0.6 G. garricki Ȃ C. leucas G. glyphis Ȃ C. leucas R. taylori Ȃ G. garricki R. taylori Ȃ C. leucas R. taylori Ȃ G. glyphis Z3/Z6 3, 7.2 < 0.01 0.6 0.2 G. garricki Ȃ C. leucas 76 G. glyphis Ȃ C. leucas 18:2Z6 3, 6.4 <0.01 1.0 0.2 G. garricki Ȃ C. leucas 76 R. taylori Ȃ G. garricki

C. leucas Season Ɂ15N 30 0.8 0.4 2.6 0.0 Ɂ13C 30 5.4 0.03 1.9 0.2 ANC Season + Ɂ13C 29 5.1 0.01 1.8 0.3 OVA TL Season Ɂ13C 0.7 0.4 TL Ɂ13C 0.05 -2.1 G. garricki ANO Season Ɂ15N 40 0.0 0.9 1.9 0.0 VA Ɂ13C 40 1.3 0.2 1.3 0.0 R. taylori Season Ɂ15N 36 1.4 0.2 2.5 0.0 Ɂ13C 36 1.7 0.2 1.3 0.0

79 The relationship between isotope values and TL of sharks were compared and were not significant except for G13C values in C. leucas, which had a negative relationship with total length (p < 0.01, Fdf = 10.31, 30, R2 = 0.3). There was a significant relationship between the ratio of C:N and G13C (p < 0.01, Fdf =

14.8.71,117, R2 = 0.1). Seasonal effects for G13C values were not found in C. leucas as the inclusion of TL as a covariate negated the seasonal effect and revealed TL to be the only significant variable (Table 3.2). Based on G15N values, the calculated

TP ranged from 3.2 ± 0.8 Ȃ 4.8 ± 0.9 and was lowest in G. glyphis and highest in C. leucas (Table 3.2).

Of the three species for which convex hull total area (TA) was calculated, C. leucas was the largest followed by R. taylori and G. garricki. Carcharhinus leucas moderately overlapped with R. taylori, although some individuals almost fell into the isotopic niche area of G. garricki. Glyphis garricki had the least overlap with other species: only the outliers of R. taylori and C. leucas shared their TA.

Corrected standard ellipse areas (SEAC) of stable isotopes ranged from 6.86 Ȃ

ͳͺǤͶ͸Ω2 with confidence intervals ranging from 97 Ȃ 99%. Carcharhinus leucas had the largest SEAC followed by R. taylori and G. garricki (Table 3.4, Appendix

7.2). When both seasons were combined the SEAC of C. leucas and R. taylori overlapped and were clearly separated from G. garricki (Table 3.4, Appendix

7.2). The wet season SEAC of R. taylori was almost completely within the wet season ellipse of C. leucas and over the dry season niche space increased in both species, particularly in R. taylori (Table 3.4, Fig. 3.3 a, b). Seasonal differences in niche metrics (Table 3.2, Fig. 3.3 a, b) and overlap (Fig. 3.2, Table 3.4 a, b) of SEAC

80 were found in each of the three species, with the least change occurring in G. garricki. The SEAC of C. leucas in the dry season overlapped with that of G. garricki. However, there was no overlap between C. leucas and G. garricki when all samples were combined (Fig. 3.3 a, b, Table 3.4).

81 Table 3.4 Isotopic overlap (%) of stable isotopes (SI) ellipses (SEAC) and convex hull (TA) of Carcharhinus leucas, Glyphis garricki and Rhizoprionodon taylori from the South Alligator River, Kakadu National Park, Australia. Also included are the probabilities of sharks being in the major essential fatty acids (EFA) niche space of each other with a set confidence interval (CI) of 95%. Ȃ no test completed

Biochemical method: SI SI EFA

Niche metric: % of species a among % Overlap % Overlap

species b ellipse (SEAC) TA Probability*

Season: Both Dry Wet Both Both

Species comparisons

a x b

C. leucas x G. garricki 0 27.4 0 22.5 61.7

C. leucas x R. taylori 36.9 27.6 12.2 37.0 0.4

G. garricki x R. taylori 0 0 0 8.4 0.6

G. garricki x C. leucas 0 66.5 0 - 13.2

R. taylori x C. leucas 66.1 40.2 99.8 - 11.2

R. taylori x G. garricki 0 0 0 - 79.9

Season comparisons

G. garricki (Dry) x (Wet) 94.5 - 50.3

G. garricki (Wet) x (Dry) 67.8 - 35.4

C. leucas (Dry) x (Wet) 36.3 - -

C. leucas (Wet) x (Dry) 36.6 - -

R. taylori (Dry) x (Wet) 79.0 - -

R. taylori (Wet) x (Dry) 14.1 - -

ȗ‘ˆͳ•’‡ ‹‡•„‡‹‰‹ƒ‘–Š‡”ǯ•‹ Š‡•’ƒ ‡Ǥ

82

Figure 3.3 (a) Bivariate plot of isotopic space depicting wet vs dry niche areas within standard ellipse (SEAC) of G13C and G15N values of Carcharhinus leucas (blue) Glyphis garricki (black), and Rhizoprionodon taylori (red) from Kakadu National Park, Australia (b) Bayesian confidence intervals of isotopic niche area; black dots represent their mode, shaded boxes represent the 50, 75 and 95% credible intervals from dark to light grey chance

83 3.4.2 Fatty acid biomarkers and niche metrics

Sixty-five FAs were identified in the elasmobranch muscle tissue, of which 31 were detected with mean values above 0.5% for any one species (Table 3.2, and

Appendix 7.1). Relative abundance of saturated FAs (SFAs) were similar in C. leucas, G. garricki, G. glyphis and U. dalyensis ranging from 27.5 ± 7.2 to 32.0 ±

8.7% but were higher in R. taylori and C. amboinensis (43.6 ± 12.3% and 42.5 respectively). The monounsaturated FAs (MUFA) were relatively low in G. glyphis, G. garricki and R. taylori (means ranging from 18.8 to 22.2%) compared to the other elasmobranchs, whose mean relative abundance ranged from 24.8 to

29.7%. Relative amounts of polyunsaturated FAs (PUFA) were highly variable between species (means ranging from 26.2 to 44.0%) with it lowest in C. amboinensis and highest in the Glyphis species. Polyunsaturated FAs were similar between C. leucas and R. taylori (30.3 ± 2.2% and 30.9 ± 3.4%, respectively) and were dominated by three EFAs: 20:4Z6 (ARA), 22:6Z3 (DHA) and 22:4Z6

(adrenic acid).

There were significant differences in FA profiles among species (PERMANOVA:

Fdf = 9.23, p < 0.01) and there was no effect of season on the FA profile of G. garricki (Fdf = 0.71, p = 0.60). Elasmobranch FAs showed some grouping between and within species, such as in R. taylori and G. garricki (Fig. 3.4). The greatest overlap was between G. glyphis and G. garricki and a few individuals of C. leucas.

The least overlap was in R. taylori, separated by the SFAs: 17:0, 18:0 and 16:0.

However, there were also a few dry season R. taylori outliers near G. glyphis and

U. dalyensis. A large subgroup of C. leucas (and some individuals of G. garricki) were also separated from the other species by the FAs, 18:2c, 20:3Z9, 18:2b,

84 18:1Z9 and 20:2, and notably, these sharks did not appear to have any unique differences such as size or sex. Whereas the Glyphis spp. were largely separated by EFAs: 20:3Z6, 22:6Z3, 22:4Z6, 22:5Z6 (Z6 DPA) and 20:4Z6 (ARA). There was a shift in posterior means from the dry to the wet season from ~ 35 Ȃ 50%, that may suggest an increase in niche space during the wet season (Fig. 3.5 and

Appendix 7.3).

Figure 3.4 Principal coordinates analysis of fatty acids in muscle tissue of Carcharhinus leucas, Glyphis garricki and Rhizoprionodon taylori, Carcharhinus amboinensis and Urogymnus dalyensis in both the wet and dry seasons from the South Alligator River, Kakadu National Park, Australia.

85

Figure 3.5 Comparisons of the posterior distributions of the probabilistic niche overlap metrics of 5 major essential fatty acids (EFAs) between 3 shark species: Carcharhinus leucas (blue), Glyphis garricki (black) and Rhizoprionodon taylori (red) from the South Alligator River, Kakadu National Park, Australia. Also shown on the bottom row is seasonal overlap for G. garricki (wet season = black, dry season = grey). Niche overlap is characterized by the probability that an individual from one shark species (or season) is found within the niche region of another shark species (or season). Posterior means and 95% credible intervals are displayed in light blue

The highest mean ratio of Z3/Z6 was in R. taylori and the lowest was in G. glyphis ranging from (0.9 ± 1.4 to 0.4 ± 0.8) (Table 3.2). An ANOVA using sharks species (consisting of C. leucas, G. garricki, G. glyphis and R. taylori) and Z3/Z6 as factors revealed significant interspecific differences (p = 0.01, Fdf = 7.23,76, R2=

0.2), whilst post hoc tests found significant differences between G. garricki and C.

86 leucas, and G. glyphis and C. leucas but not between R. taylori and the other sharks; G. garricki, G. glyphis and C. leucas. Total length had a positive significant effect on Z3/Z6 ratios in C. leucas (p = 0.05, Fdf = 4.21, 21, R2 = 0.18), G. garricki (p

= 0.02, Fdf = 6.21, 23, R2 = 0.9) but not in R. taylori or G. glyphis. A significant difference was also found between the elasmobranches using the terrestrial marker 18:2Z6 (p < 0.01, Fdf = 6.43, 76, R2 = 0.2) and post hoc tests found that G. garricki and C. leucas, and G. garricki and R. taylori (Table 3.3) were the only pairs that had significantly different means.

The largest FA niche space occurred in G. garricki followed by C. leucas and R. taylori (Table 3.2, Fig. 3.5, 3.6) across all sites and seasons. The chance of G. garricki „‡‹‰‹ƒ‘–Š‡”•’‡ ‹‡•ǯ ‹ Še space was low (0.6% for R. taylori and

13.2% for C. leucas) yet C. leucas and R. taylori had larger probabilities for overlapping with G. garricki FA niche space (79.9% for R. taylori and 61.7% for C. leucas). Slight seasonal differences were evident in G. garricki FA niche space with a 15% probability of overlap between the wet and dry seasons (Fig. 3.5).

87

Figure 3.6 Schematic of the essential fatty niche space and extent of overlap among three shark species, Carcharhinus leucas (blue), Glyphis garricki (black) and Rhizoprionodon taylori (red) in the South Alligator River, Kakadu National Park, Australia. Niche spaces are scaled relative to each other; central axis omitted.

3.4.3 Trophic niche metric comparisons

Differences in niche metrics, including inter-specific overlap, were apparent between the two SI niche metrics (SEAC and TA) and between both these and

EFA niche spaces. Between the two SI niche metrics, TA consistently gave larger area values and showed greater percentage overlap between species compared to SEAC. Despite this, the ordering of species niche areas were similar being highest in C. leucas followed by R. taylori, and G. garricki (Table 3.2). In contrast,

FA niche spaces were highest in G. garricki followed by C. leucas and R. taylori

(Table 3.2). When comparing percent overlap in SEAC and EFA niche metrics, differences in model output was observed (Table 3.4). Greatest differences

88 between methods were observed between R. taylori and G. garricki where there was a large chance of overlap in EFA niche space but not in SEAC. Carcharhinus leucas and R. taylori overlap was also different with less chance of overlap for SI than EFA niche metrics (Table 3.4). Seasonal differences between G. garricki SI and FA niche metrics were also apparent with greater overlap in SI SEAC (67.8 to

94.5%) and FA niche space (35.4 to 50.3%).

3.5 Discussion

The complementary analyses of SI and FA niche metrics in the current study revealed significant differences in the trophic resource use of tropical elasmobranchs, with indications of resource partitioning among three species

(and limited evidence in another four species) occupying the South Alligator

River and adjacent coastal waters. Although SI niche metrics indicated clear separation between coastal/marine and euryhaline elasmobranchs, the difference was less pronounced based on FA niche metrics alone. This is likely due to the faster turnover of FAs (Kirsch et al. 1998, Beckmann et al. 2014b) compared to SI (Logan & Lutcavage 2010), differences in the statistical methods employed, and that both tracers provide different representations of an

‘”‰ƒ‹•ǯ•‹ Š‡Ǥ‘‰ƒ ‘—‹–›‘ˆ‘”‰ƒ‹••ǡ ‹ Š‡•’ƒ ‡”‡ˆŽ‡ –•–Š‡ trophic transfer of marker FAs from basal sources to predators, whilst, SI niche

ƒ”‡ƒ•”‡’”‡•‡–„‘–Šƒ‘”‰ƒ‹•ǯ•–”‘’Š‹ ’‘•‹–‹‘ƒ†‹–•„ƒ•ƒŽ ƒ”„‘•‘—” ‡Ǥ

Not surprisingly, differences between these biochemical niche metrics and more traditional niche metrics based on taxonomic descriptions of diet composition

89 have also been observed through recent efforts to compare SI, TA and the omnivory index derived from the Ecopath ecosystem modelling framework

(Layman, et al. 2007a, Navarro et al. 2011). Despite these differences, each method gives complementary understanding of ecological interactions and habitat use.

In this study, interspecific differences in the SI niche area of elasmobranchs appear to be largely driven by G13C, indicating that the feeding locations of each species are variable or that mobile prey species are being consumed by specific elasmobranchs. Several species (Rhizoprionodon taylori, Carcharhinus amboinensis, C. leucas and Pristis pristis) were characterized by values that are typically reported in marine environments (-14 to -ͳ͸ΩǢ —••‡›‡–ƒŽǤʹͲͳͳb,

Munroe et al. 2015). G13C values for P. pristis and C. amboinensis were similar to previous studies (Thorburn et al. 2014, Tillett et al. 2014), which may have been sourced from seagrass, whilst C. leucas values were between both marine epiphytes and seagrass, and R. taylori had values closer to epiphytes (Loneragan et al. 1997). In contrast, Glyphis garricki, G. glyphis and Urogymnus dalyensis had lower G13C values suggestive of estuarine/freshwater seston signatures (-18.8 to

-ʹ͵ǤʹΩǢ‘‡”ƒ‰ƒ‡–ƒŽǤͳͻͻ͹Ȍ.

In terms of trophic position (TP), the seven elasmobranch species were broadly similar, based on the spread of G15N values; with C. leucas having the highest values and mean TP (4.8 ± 0.9). This result conforms with C. leucas stomach contents analyzed in northern Australia which mainly consisted of teleost fish and occasional larger predators such as P. pristis and the freshwater crocodile

90 Crocodylus porosus (Thorburn & Rowland 2008). Although our analysis used juveniles, our TP estimates were higher than adult TP (4.6 ± 0.2) and sub-adult

TP (4.4 ± 0.3) C. leucas reported in Mozambique (Daly et al. 2013) and Western

Australia TP (~ 4.4) (Thorburn et al. 2014) and in the same range (~3.8 to 5.4) as those from South Africa (Hussey et al. 2014a). This variation is likely due to the different statistical methods of determining TP as well as dietary and environmental variation (Peterson & Fry 1987). There is currently no literature reporting TP estimates for Glyphis spp. and U. dalyensis and only the second for P. pristis whose TP was in the same range (3.6 to 4.7) as Thorburn et al. (2014). The low TP calculated in this study for G. glyphis (3.2 ± 0.8) suggests that they consume lower order prey and that their TP was more similar to rays than sharks of this body size (Hussey et al. 2014a). The TP in R. taylori were in the same range as those caught in the Cleveland Bay, Queensland, but were higher than those caught in nearby bays (Munroe et al. 2014a). This could be attributed to differences in prey TP, environmental conditions or baseline G15N values, which are all closely coupled.

Based on the analysis of elasmobranch FA profiles and biomarkers, clear interspecific differences were apparent which mostly supported the two guilds found from SIs. All species had marked variation in their individual FAs, which was similar to other northern Australian tropical sharks (Nichols et al. 2001) and

C. leucas (Belicka et al. 2012a). The Glyphis spp. grouped together and had high levels of Z6 FA biomarkers and 18:2Z6 (LA) which are all characteristic of terrestrial and freshwater sources (Napolitano et al. 1997, Budge & Parrish

1998). In contrast, R. taylori, had higher relative levels of SFA, and 20:5Z3 (EPA)

91 and 22:6Z3 (DHA) that are characteristic of marine food webs based on diatoms and flagellates, respectively (Parrish 2013). Higher relative levels of the FA biomarker, 18:1Z9, linked to piscivory (Dalsgaard et al. 2003, Kelly & Scheibling

2012), was reported for C. leucas, which supports the high TP estimates based on

G15N values. However, the complete FA profile of C. leucas, that consisted of high relative levels of 18 PUFA, 16:1Z7, and low 22:6Z3 (DHA) suggests a greater intake of estuarine than marine prey, contrary to the predominant marine based

SI signatures. This suggests that C. leucas are utilizing food webs across the marine to freshwater spectrum, which was previously found in C. leucas (Matich

& Heithaus 2014). Glyphis spp. also had links to marine sources indicated by the moderate proportion of 22:6Z3 (DHA) and the higher amount of 20:4Z6 (ARA) which implies links between benthic marine and estuarine/coastal prey (Hall et al. 2006, McMeans et al. 2013).

The FA biomarker Z3/Z͸ ƒŠ‡Ž’ˆ—”–Š‡”†‡ˆ‹‡ƒ•’‡ ‹‡•ǯ‹ Š‡ƒ†–Š‡‹”„ƒ•ƒŽ sources as higher ratios (> ~1.5) indicate a preference for marine resources whilst low ratios (< ~1.5) suggest a dependence on freshwater resources

(Martínez-Álvareza et al. 2005, Özogul et al. 2007). The elasmobranchs in the current study appeared to all have a freshwater influence based on the above range, however, there was a difference between species, in that R. taylori (a marine species) had the highest ratio (0.9), whilst G. garricki (0.5) and G. glyphis

ȋͲǤͶȌ ‘ŽŽ‡ –‡†—’•–”‡ƒ–‡†‡†–‘„‡Ž‘™‡”ǤŠ‡ɘ͵Ȁɘ͸”ƒ–‹‘‹C. leucas and G. garricki were significantly different, which may reflect a preference for feeding in differing salinities, especially since C. leucas had G13C values that were more

92 indicative of marine sources (although caught in similar areas). The Z3/Z6 ratio may also be related to condition, growth and maturation in elasmobranchs, as a significant (positive) relationship was found between TL and the ratio in C. leucas and G. garricki, which were mostly juveniles. This relationship may reflect physiological changes that occur during sexual maturation, such as that found in juvenile teleost fish where the Z3/Z6 ratios decreased during sexual maturation

(Uysal & Aksoylar 2005) or even changes in the way they use the river during ontogeny.

The niche metrics of elasmobranch species in this study suggest that there is a broad prey base in the South Alligator River. The largest SI niche area was observed for C. leucas - a species reported to consume a broad range of taxa including catfish (ariids), rays (batoids), and other carcharhinid sharks (Snelson et al. 1984, Tillett et al. 2014). However the FA niche space for C. leucas was intermediate between G. garricki (being very large) and R. taylori (very small), which suggests individual dietary specializations in C. leucas (Matich et al. 2011), prey switching according to the local abundance of prey (Matich & Heithaus

2014), and/or that there is a maternal influence (Olin et al. 2011, and see below).

In contrast, R. taylori exhibited an intermediate SI niche area, and a much smaller

FA niche space, which is somewhat at odds with the relatively broad diet (e.g. prawns and teleost fishes) identified in stomach contents (Simpfendorfer 1998).

Interestingly, some R. taylori in Queensland occupied areas near the river outflows during the wet season rather than moving to seagrass beds with the majority of the population (Munroe et al. 2014b). In light of these findings, it may be that the narrow niche of R. taylori in the South Alligator estuary results from

93 some individuals of this species targeting prey around riverine outflows.

Collectively, these trophic niches suggest some important considerations for conservation and management of these species. For example, the strong overlap between R. taylori and C. leucas, and the relatively narrower trophic niche of R. taylori suggests this species may be in competition with the more abundant C. leucas and vulnerable to disturbances (Layman et al. 2007a). Similarly, the greater dependence of G. garricki on riverine resources compared to the other elasmobranchs highlights the importance of this habitat to G. garricki in order for them to forage, grow and maintain healthy populations.

Temporal changes in trophic niche size or overlap among populations are likely in environments that are highly dynamic. Seasonal differences in biota are relatively common in tropical river systems due to the large outflow of freshwater during the wet season changing the sources of energy and nutrients, and increasing primary and secondary production (Winemiller & Jepsen 1998).

Our results may indicate seasonal changes in the niche metrics of all species tested, but not in the individual SI or FA biomarkers. This may be because the ellipses are highlighting subtle population differences as they compare the posterior mean values of the two isotopes and the covariance matrix of each species while PERMANOVA and ANOVA compare individual treatment means.

The largest seasonal differences in niche metrics were for SI areas for R. taylori, which were much smaller in the wet than the dry season. This is likely to reflect a preference for a particular food type that is more prevalent in the wet season whereas in the dry they may need to source a variety of food to meet their nutritional requirements. For both C. leucas and G. garricki larger SI areas were

94 observed in the wet than the dry, reflecting a greater breadth of available food and habitat, although the large overlap between niche areas suggest that they are consuming similar prey type over both seasons. This was further supported by G. garricki having the larger wet season EFA niche space.

Observed differences between SI and FA niche metrics, particularly within C. leucas may also be a result of maternal influences, driven by temporal differences in muscle tissue turnover rates between the two biochemical tracers. The significant correlation between TL and C. leucas G13C values, suggests that these sharks rely on different food sources as they grow. Maternal signatures have been found in C. leucas in both G13C and the FA 20:3Z9 which was used as a biomarker for EFA deficiency (Matich et al. 2010, Olin et al. 2011, Belicka et al.

2012a) and very weakly in G15N (Matich et al. 2010). The lack of maternal signature in isotopes within our sharks may be attributed to C. leucas feeding at a similar TP to their mothers, that their size range was too small to find a size- related difference or there were differences caused from the baseline consumer.

Tissues in C. leucas have shown a decline in isotopic values until they reach 110 cm TL (Matich et al. 2010) whereas our TL range was from 72 to 93 cm and open umbilical scars were found on ~50 % of individuals. The two juvenile P. pristis may also have had some maternal influence as they shared higher SI values.

Differences detected between the biochemical niche metrics are also associated with the statistical methods used. FA niche space is a probability projection rather than geometric (Swanson et al. 2015), and thus calculates the likelihood of

‘‡•’‡ ‹‡•„‡‹‰‹–Š‡‘–Š‡”•ǯ‹ Š‡•’ƒ ‡Ǥ  ‘–”ƒ•–ǡC is reliant on the

95 confidence interval set for the ellipse and only one measurement of overlap is calculated. For example R. taylori had a high probability of being in G. garricki FA niche space but there was no SEAC overlap between these species. In determining

FA niche space, only five EFAs could be used, which meant there was an underlying assumption that these FAs were the only FAs within the species that were important for differentiating niches. A more complete understanding of trophic niche based on FA profiles could be achieved with larger sample sizes.

Better understanding of FA trophic markers in tropical watersheds would increase our ability to distinguish food sources within and among rivers, estuaries and coastal ecosystems. Likewise, understanding how biomarkers respond to environmental cues, and profiling potential prey items, would greatly increase our ecological knowledge of species resource use and trophic interactions in coastal and estuarine ecosystems.

This study identified two separate feeding guilds and was the first to use SIA and

FA analysis in combination to measure trophic niche metrics and explore resource partitioning among an assemblage of elasmobranchs. Differences in FA and SI niche metrics highlighted the advantages of combining such analyses. Our approach has demonstrated that elasmobranchs within the South Alligator River display partitioning in trophic resource use, particularly across marine and freshwater food webs. In particular, some species (C. leucas and G. garricki) are utilizing food webs across the marine to freshwater spectrum, which suggests that these fish provide broad and important cross-biome trophic linkages in tropical coastal ecosystems. Whilst other species (R. taylori) have a predominantly marine based niche that is relatively narrow but still provides

96 limited connections through to the estuary. Given the potential for complexity in resource use, also highlighted by seasonal shifts in SI and FA niche metrics, it appears likely these elasmobranchs play important roles in food web connectivity in tropical aquatic ecosystems.

3.6 Chapter Acknowledgments

This research was conducted on the traditional country of the Bininj and

Mungguy people. We gratefully acknowledge the traditional custodians for allowing access to this country. We thank Peter Nichols, Peter Mansour, Grant

Johnson, Mark Grubert, Duncan Buckle, Roy Tipiloura, Dominic Valdez, Francisco

Zillamarin, Edward Butler, Claire Streten, Diane Purcell, Kirsty McAllister and the crew of R.V. Solander for fieldwork and laboratory assistance and to Martin Lysy

(nicheROVER) and Andrew Jackson (SIAR) for their advice and assistance.

Charles Darwin University, CSIRO and the North Australia Marine Research

Alliance (NAMRA) provided funding, alongside collaborative partnerships in the

ƒ”‹‡‹‘†‹˜‡”•‹–›ƒ†‘”–Š‡”—•–”ƒŽ‹ƒ —„•‘ˆ—•–”ƒŽ‹ƒ ‘˜‡”‡–ǯ•

National Environmental Research Program (NERP). This study was conducted with the approval of the Charles Darwin University Animal Ethics Committee

(A12016), in conjunction with permits from NT Fisheries and Kakadu National

Park (RK805). Finally, we thank four anonymous reviewers for their helpful comments on an earlier draft of this manuscript.

97

4 A seasonally dynamic estuarine ecosystem provides a diverse prey base for elasmobranchs

This chapter is currently submitted to Estuaries and Coasts on the 25th of April.

Below is a list of the co-authors and their contribution. All authors contributed to concept design and writing.

Heidi Pethybridge Ȃ Contributed knowledge of multivariate and lipid and stable isotope analysis, fish biology and ecology.

Christopher Fulton provided advice and assistance with the statistical analyses as well as editorial comments on the structure and content of the manuscript text and data presentations.

Peter Kyne Ȃ Contributed knowledge of elasmobranch biology and ecology, and assisted with fieldwork and tissue collection.

David Crook Ȃ Contributed knowledge of fish biology and ecology and statistical analysis of fish and riverine ecology and manuscript planning and development.

98 4.1 Abstract

Food web studies connecting species of high conservation concern with their prey are limited in tropical river systems, particularly for predators such as sharks. Stable carbon (G13C) and nitrogen (G15N) isotopes (SI) and fatty acid (FA) biochemical tracers taken from putative prey species in the estuary of the South

Alligator River, northern Australia, were compared with existing information on four tropical elasmobranchs (Carcharhinus leucas, Glyphis garricki, G. glyphis and

Rhizoprionodon taylori) to examine food web structure and infer dietary linkages over distinct wet and dry seasons along a longitudinal gradient of sites.

Upstream food webs had the greatest trophic diversity (G15N range and the

ƒ••‡„Žƒ‰‡•ǯ–‘–ƒŽ‹ Š‡ƒ”‡ƒȌǤŠ‡”‡™‡”‡•‹‰‹ˆ‹ ƒ–†‹ˆˆ‡”‡ ‡•‹ •ƒ† • among species across all sites and seasons, with G13C values indicating distinct foraging habitats. Mixing models of SI indicated that G. glyphis and G. garricki had similar fresh water and midȂriver diets, although outliers were found.

Multivariate analyses revealed some intraspecific differences, yet specialisation indices suggested these four sharks are trophic generalists that utilise a diverse prey base.

99 4.2 Introduction

Food webs in tropical floodplain rivers are highly connected, dominated by seasonal hydrological cycles and have short food chains and variable ecological communities (Douglas et al. 2005, Blanchette et al. 2014). Consumers such as euryhaline and coastal elasmobranchs (sharks and rays) provide potentially important connections across ecosystems due to their high mobility and high trophic position, and are crucial in the maintenance of community structure and ecosystem function in many estuaries (Last 2002, Every et al. 2017). Estuarine and coastal ecosystems may act as nurseries for sharks (Heupel et al. 2007), afford protection from predation and provide a diverse source of prey (Cyrus &

Blaber 1992, Heupel et al. 2007). However, many of these ecosystems have been affected by habitat disturbance and fishing pressure (Gallagher et al. 2012, Dulvy et al. 2014) that have contributed to the decline of many estuarine species, including elasmobranchs (Lucifora et al. 2015). In order to conserve and manage these species, there is a need to increase our knowledge of the dietary requirements and potential trophic specialisation of euryhaline elasmobranchs

(Montoya et al. 2006) to better understand functional differences, overlaps and their dependency on particular species and habitats (Young et al. 2015, Grubbs et al. 2016).

Previous work examining dietary composition in tropical euryhaline elasmobranchs has been largely limited to more ubiquitous species such as the bull shark Carcharhinus leucas. However, other species are also important components of the elasmobranch fauna of rivers and estuaries in the Indo-

Pacific, but are not well studied. In northern Australia there is a paucity of data

100 on the trophic ecology of coastal and euryhaline elasmobranchs, with previous studies focusing on adult to sub-adult (Tillett et al. 2014) and juvenile C. leucas and large tooth sawfish Pristis pristis (Thorburn & Rowland 2008, Thorburn et al.

2014). Although studies that have used stomach content analysis are beneficial, provide only a brief snapshot in time and often miss soft-bodied prey or over- represent certain groups (e.g. crustaceans) due to differential rates of digestion and/or complex temporal patterns in consumption. Advances in techniques such as biochemical analysis of stable isotopes (SI) and fatty acids (FA) have allowed for broader time scales of trophic biology to be explored based on tissues from a suite of sharks and rays (MacNeil et al. 2005, Hussey et al. 2011a, Pethybridge et al. 2011a, Couturier et al. 2013a, Rohner et al. 2013, Every et al. 2016).

–ƒ„Ž‡‹•‘–‘’‡•‘ˆ ƒ”„‘ȋɁ13Ȍƒ†‹–”‘‰‡ȋɁ15N) have been used to determine niche area and overlap (Vaudo & Heithaus 2011, Every et al. 2017,

Chapter 3) food web structure (Abrantes & Sheaves 2009, Tilley et al. 2013) and

‘—‹–›‡–”‹ •ȋƒ›ƒƬ‘•–ʹͲͲͷǡ”‹†ǯ‘—”Ƭ—„‘‹•ʹͲͳ͵Ȍ. Isotopic mixing models (Layman & Allgeier 2012, Parnell et al. 2013, Tilley et al. 2013) can be particularly useful to trace which prey or prey group (source) is likely to have been consumed by a predator (Peterson & Fry 1987). Complementary FA analyses have also been helpful in interpreting isotopic food web indices, as they provide greater specification of basal sources and can confirm trophic linkages

(Budge et al. 2002, Iverson 2009, Kelly & Scheibling 2012). Many fatty acids are passed from producers to consumers with limited fractionation or synthesis, as vertebrates are unable to synthesize essential FAs (EFA). The combination of both SI and FA analyses provides additional support to each method in the

101 interpretation of the trophic ecology of consumers (Belicka et al. 2012a,

McMeans et al. 2013).

The objective of the current study was to assess the structure of food webs within a tropical riverine system in northern Australia. This was explained by undertaking SI and FA analyses on a suite of mid-trophic taxa, and using published data on juvenile euryhaline (Carcharhinus leucas, Glyphis garricki, G. glyphis) and coastal (Rhizoprionodon taylori) elasmobranchs to explore possible

–”‘’Š‹  ‘‡ –‹‘•Ǥ‡—•‡†ƒ›ƒǯ•‡–”‹ •–‘ƒ••‡••–”‘’Š‹ •–”— –—”‡‘ˆ food webs, and mixing models and intraspecific differences to assess diet of these elasmobranch species, two of which are of high conservation concern.

4.3 Methods

4.3.1 Elasmobranch and potential prey collection

Three euryhaline elasmobranch species (Carcharhinus leucas, Glyphis garricki, G. glyphis) and one coastal species (Rhizoprionodon taylori) were collected in the

South Alligator River, Australia from March 2013 to July 2014 (Table 4.1) as part of previous studies (Every et al. 2016, 2017). Rhizoprionodon taylori were captured by baited line in the mouth of the river and G. garricki, G. glyphis and C. leucas were collected further upstream, with a combination of gill nets and baited lines. All sharks were measured and biopsied before being released at the site of capture.

102

Table 4.1 Number (n) and total length of 4 sharks (Every et al., 2017) and 6 putative prey species caught from the South Alligator River, Kakadu National Park, Australia from which muscle tissue samples were taken for stable isotope (SIA) and fatty acid analysis (FAA). Wet and dry species number, Total length (TL) (± standard deviation (SD)), sex ratio and habitat are also included.

Shark Species Sex TL±SD FAA SIA Habitat ratio (cm) M:F

Scientific Name Common Name Wet Dry Wet Dry Carcharhinus leucas Bull Shark 24:16 82.2±16.3 20 2 27 6 euryhaline Glyphis garricki Northern River Shark 22:19 94.5±24.6 12 13 22 20 euryhaline Glyphis glyphis Speartooth Shark 3:7 88.7±23.3 2 3 2 3 euryhaline Rhizoprionodon taylori Australian Sharpnose Shark 7:21 54.8±12.1 1 24 4 27 coastal Potential Prey Johnius novaeguineae Paperhead Croaker 7.9±3.5 8 14 11 8 estuarine Lates calcarifer Barramundi 39.9±13.8 6 17 8 20 estuarine Macrobrachium equidens Rough River Prawn 7.4±1.6 21 12 22 15 euryhaline Nemapteryx armiger Threadfin Catfish 27.9±5.0 10 12 12 16 estuarine /euryhaline? Polydactylus macrochir King Threadfin Salmon 44.8±18.8 8 7 8 8 euryhaline Rhinomugil nasutus Popeye Mullet 16.6±6.6 7 7 7 11 estuarine

103 Sampling for prey occurred in the same 4 sites where sharks were collected for an earlier study (Every et al. 2017) over the wet (monsoon) (November Ȃ April) and dry (May Ȃ October) seasons. Briefly, site 1 was the furthest upstream and had a mean •ƒŽ‹‹–›ȋΩ±SD) during the dry season of 21.9±5.3 and of 0.4±1.5 during the wet season, whilst at site 4 salinity was high in the dry (34.5±0.2) and lower in the wet (17.1±4.3) (Every et al. 2017). Prey were captured using a range of sampling methods: a ~5 m wide beam trawl, gill nets (mesh size ranging from

10 Ȃ 30 cm), a cast net, and custom made wire rectangular marine and opera crab pots. Prey species were also caught during gill net and line fishing for sharks. Six putative prey species (Table 4.1) were chosen for analysis as these:

(1) appeared in sufficient numbers to be considered a significant part of the food web; (2) represented a range of tropic levels; and, (3) had been reported previously in the stomachs of study elasmobranchs (Snelson et al. 1984,

Simpfendorfer 1998, Thorburn & Morgan 2004, Peverell et al. 2006). Prey species consisted of five teleost fishes and one crustacean (Table 4.1).

4.3.2 Tissue sampling & preparation

For teleost fishes, only muscle tissue of prey was used so that larger fish could be released, which involved using a scalpel to lift scales (where present) and cut a small square of tissue from the caudal peduncle region. Smaller fish of less than

25 cm total length were euthanized in 20 L of river water in AQUI-S® (20 mg/L)

(Lower Hutt, New Zealand; sensu Turchini et al. 2011; Matley et al. 2016), and then the right side of the body filleted. The invertebrate Macrobrachium equidens were also euthanized in the same way before muscle was dissected from within

104 the 2nd to 4th abdominal segments, taking particular care not to include other tissue (e.g. exoskeleton, gut).

Immediately after collection, all prey tissue was stored in liquid nitrogen at Ȃ

196ǏC, and within a week transferred to a -ʹͲǏˆ”‡‡œ‡”—–‹Ž‹–™ƒ•ˆ”‡‡œ‡-dried for analysis. Preparation of samples was undertaken in the freezer to avoid tissue degeneration. All tissue except muscle was removed and the muscle sample divided and weighed separately for SI and FA analyses. Mean (± standard deviation (SD)) dry sample weight was 1.96±0.16 mg across all prey types.

4.3.3 Stable Isotope Analysis

Prey muscle tissue was freeze-dried to a constant weight and then pulverised using a combination of micro-scissors and a small polyethene pestle, or a coarse pestle and ceramic mortar. Muscle tissue was weighed to between 400-2200 Pg.

To combust and analyze samples, a SerCon Europa EA-GSL elemental analyzer and Hydra 20-22 isotope ratio mass spectrometer (Sercon Ltd, UK) was used at the Australian Rivers Institute, Griffith University. Relative G13C and G15N were calculated using the Peedee Belemnite Carbonate international standards for

G13ƒ†–‘•’Š‡”‹ ‹–”‘‰‡™‹–Šƒ’”‡ ‹•‹‘‘ˆȋͳȌͲǤͲ͵ƒ†ͲǤͲͻΩˆ‘”

G15N and G13C, respectively. Due to the low lipid content in the muscle, lipid corrections were not necessary except for N. armiger which had a mean C:N ratio of 4.3±1.1, which was over the recommended level of 3.5, therefore the following formula was applied (Post et al. 2007):

105 G13Cnormalized = G13C untreated -3:32 + 0:99 X C:N

Since SI analysis required a smaller amount of animal tissue to analyse, more individuals (cf. to FA analysis) could be examined with this method.

4.3.4 Fatty Acid Analysis

Prey FAs were quantitatively extracted from muscle tissue via direct transmethylation (Parrish et al. 2015). Fatty acids were liberated from the lipids within the tissue sample via solvent extraction. Tissues were freeze dried, weighed, and 3 ml of MeOH: hydrochloric acid (HCl): DCM (10:1:1) was added, vortexed and placed in heating block at 85°C for 2 hours. After cooling, 1 ml of milli-Q H2O was added and the FA solution was extracted with 1.8 ml of 4:1 hexane:DCM solution and then vortexed for five minutes in a centrifuge to form the lipid bilayer. The upper layer was then transferred using DCM and blown down under a constant stream of N2. The extraction process was repeated two more times before a known concentration of internal standard was added. Final concentrations of 10 mg lipid to 1.5mL DCM were made and stored in a -20°C freezer until further analysis within 7 days of extraction.

Fatty acid composition was quantified by an Agilent Technologies 7890B gas chromatograph (GC) (Palo Alto, California USA) and an Agilent Technologies

7683B Series auto-sampler. Peaks were quantified using Agilent Technologies

ChemStation software (Palo Alto, California USA), and identifications confirmed by GC-mass spectrometry (GC-MS) using a column of similar polarity to that

106 described above and a Finnigan Thermoquest DSQ GC-MS. Fatty acids were converted to a percentage. FAs with values <0.5% were not included in statistical analysis.

4.3.5 Assessment of food web structure

–ƒ„Ž‡‹•‘–‘’‡†ƒ–ƒ™ƒ•—•‡†–‘ ƒŽ —Žƒ–‡ƒ›ƒǯ••‹š‡–”‹ •(Layman et al.

2007a) of seasonal and spatial trophic diversity in both putative prey and shark consumer species across each site and season. The first four metrics are measures of the assemblage trophic diversity, whilst the last two measure the relative space between each other (Layman et al. 2007a). Specifically, these are:

(i) G15N range (NR), (ii) G13C range (CR) (both of which are the isotopic distances between species), (iii) total area (TA) (the assemblage combined isotopic niche space occupied), (iv) centroid distance (CD) (a function of species spacing), (v) mean nearest neighbour distance (MNND), and (vi) the standard deviation of

MNND (SDNND) (both of which are measures of species density) (Layman et al.

2007a). Metrics were calculated using the mean from each species group in the R package SIAR 4.2.2, which uses Bayesian approaches to account for uncertainty in the derived means of convex hulls, removes potential errors and therefore increases the validity in the estimates of community metrics (Jackson et al.

2011). For this analysis there needs to be a minimum of 5 individuals for each species, thus in some sites and seasons, species were omitted (Table 2) to minimise sample size biases (Jackson et al. 2011). All sites and seasons were included in one analysis so that those species where n was < 5 in specific sites could be compared holistically.

107

To examine differences in the SI compositions of the putative prey taxa an analysis of variance (ANOVA) was used followed by pairwise Tukey tests with

Bonferroni adjustments for multiple comparisons. Evaluations of Q-Q plots and residual vs fitted graphs indicated that no data transformations were required to satisfy model assumptions.

Permutational analysis of variance (PERMANOVA) was used to explore significant differences between species EFAs in multivariate space. A homogeneity of dispersion test (PERMDISP) revealed an uneven distribution of multivariate variance (p < 0.01, Fdf = 5.40, 5, 120). However, PERMANOVA has been found to be relatively robust to such dispersion issues (Clarke & Gorley

2006) (e.g. in our case, site or season). In these analyses the PERMANOVA (with

9999 permutations) was used to test for a significant difference between prey, prey and season, prey and capture location (sites 1, 2, 3 and 4) as factors and finally as prey, season and capture location. A pairwise test was also carried out with species and season as the factor. To assist in the interpretation of the

PERMANOVA and to visualize these differences a principal component analysis

ȋȌ™‹–Šƒ‡ƒ”•‘ǯ• ‘””‡Žƒ–‹‘‘ˆͳΨ™ƒ• ƒŽ —Žƒ–‡†Ǥ

To determine which FAs may be unique to each prey species a Dufrêne-

Legendre indicator species analysis ((R package; labdsv (Roberts

2016))(Dufrêne & Legendre 1997) was applied. This was developed to determine which species could be used as indicators for various habitats, however we have used the same calculations to determine which FAs occur more

108 frequently in each species Ȃ –Š‡”‡ˆ‘”‡‘—”•’‡ ‹‡•ƒ”‡–Š‡ǮŠƒ„‹–ƒ–ǯƒ†–Š‡ •

ƒ”‡–Š‡Ǯ•’‡ ‹‡•ǯƒ ‘”†‹‰–‘Dufrêne & Legendre (1997). To calculate the indicator value for FAs based on the relative frequency and association of FAs within each species we need to determine the presence / absence (Pij) of FAs in a species and the abundance of FAs in the species (Xij):

Where:

FA = i

Species = j nc = number of samples in cluster c (for cluster c in set K) f = relative frequency a = abundance of FAs

݀௜ǡ௖ = Indictor value (IndVal)

௖ܲ௜ǡ௝אσ௝ ݂௜ǡ௖ ൌ  ݊௖

௖ݔ௜ǡ௝ሻȀ݊௖אሺσ௝ ܽ௜ǡ௖ ൌ ௄  ௖ݔ௜ǡ௝ሻȀ݊௞ሻאσ௞ୀଵሺሺσ௝

݀௜ǡ௖ ൌ ݂௜ǡ௖ ൈܽ௜ǡ௖

An indicator value and p Ȃ value are assigned to each FA for that particular species. The addition of the p - value, was an adaptation in the R package; labdsv

(Roberts 2016) from the oringnal calculations of Dufrêne & Legendre (1997).

4.3.6 Isotope mixing models of sharks and putative prey taxa

Mixing models for SI were created using the Bayesian models package MixSIAR

(Moore & Semmens 2008) in R (R Core Development Team, 2014). These

109 models use Markov chain Monte Carlo (MCMC) to calculate uninformed priors based on the data given (we used 10,000 iterations). They were designed to be robust, allow multiple sources to be used and enable priors and uncertainty measures to be included (Moore & Semmens 2008). As recent work has found that more than three sources can undervalue minor dietary items (Brett 2014), prey data was grouped based on the divisions created by their G13C values.

Similar G13C values such as what was found here have previously been linked to carbon sources in tropical riverine waters (including their estuaries and surrounding seagrasses) and so our putative prey species have been classified accordingly (Loneragan et al. 1997). Group 1 prey had G13C values closer to freshwater signatures and consisted of Lates calcarifer, Macrobrachium equidens and Johnius novaeguineae. Group 2 consisted of Polydactylus macrochir and

Neoarius armiger were higher in G14N than the other species and had G13C values that were in between estuarine and freshwater signatures whilst Group 3 consisted only of Rhinomugil nasutus, which had a G13C value closer to an estuarine signature. Residual errors were included in the model (Parnell et al.

2010) and uncertainties consisted of elemental concentrations based on the mass of each tissue (Parnell et al. 2010) and diet discrimination factors (DDF, the fractionation ‘ˆɁ15N ƒ†Ɂ13C when passed through a food chain). Since DDFs have not been determined for these prey species and the South Alligator River appears to have more than one food web the use of a consumer baseline from our caught prey could cause confounding results. For these reasons we used Ɂ15N

DDFs estimated from Bunn et al (2013) study. In this study Ɂ15N DDF were calculated from a range of species in lotic environments from northern Australia and Papua New Guinea using a regression analysis and comparisons of literature.

110 We then compared the feeding behaviours to our species and used those DDFs

(see table 4.2).

Table 4.2 Putative prey species caught in the South Alligator River, Kakadu National Park, Australia and their estimated diet discrimination factor (DDF) used in the mixing model, based on Bunn et al., (2013).

Species Feeding method Diet discrimination factor (DDF) Macrobrachium equidens Predatory invertebrate 1.8±1.7 (March et al. 2002 based on Macrobrachium spp.) Johnius novaeguineae Omnivorous fish (predatory 4.3±1.5 invertebrates/algae) (Sasaki 2001) Lates calcarifer Predatory fish (Davis 1985) 5.7±1.6 Polydactylus macrochir Predatory fish (Brewer et al. 5.7±1.6 1995) Neoarius armiger Predatory fish (Blaber 1993) 4.3±1.5 Rhinomugil nasutus Omnivorous fish (algae / 3.9±1.4 herbivores invertebrates) (Froese & Pauly 2015)

A carbon DDF of 1.3±0.3Ω was used for all species, following McCutchan et al.

(2003). Differences in isotopic integration amongst species were accounted for by the prey sampling regime (see above) (Post 2002, Moore & Semmens 2008), sufficient prey n values (Table 1), and by using the average size range ±20% for each prey species.

4.3.7 Fatty acid prey-predator linkages

Prey EFAs and shark EFAs (see Every et al. (2017) were compared with a main model PERMANOVA and a pairwise PERMANOVA the fixed factor was species and a Type III (partial) sum of squares was used for both analyses. To compliment this, similarity percentages (SIMPER) based on Bray Curtis distances

(Euclidean distance gives average squared distance not average similarity) was

111 used to calculate the average similarity between the FA profiles of individuals within a species.

4.3.8 Individual specialisation of fatty acids

To explore the degree of individual specialisation we used the FA data from

Every et al. (2017) to calculate indices based on Roughgarden (1972), which calculates the proportion of total niche width (TNW) and within individual variation (WIC) in fatty acids. These were determined using the R individual specialisation package (RInSP) (Zaccarelli et al. 2013). This test is useful when there are more than two variables; therefore the use of two SIs is not appropriate. Values of TNW/WIC closer to 1 indicate no intraspecific differences whilst 0 suggests a high degree of individual specialization. Monte Carlo tests are then followed to determine a p value. All FA > 0.5% were included as there is an increase in accuracy when there are more variables associated with each individual (Bolnick et al. 2002, Zaccarelli et al. 2013).

4.4 Results

4.4.1 Food web structure and connectivity among putative prey taxa

Across all sites ther‡™ƒ•ƒ‘˜‡”ƒŽŽ†‡ ”‡ƒ•‡‹ƒ›‡ǯ•‡–”‹ •‘ˆǡǡǡ

CD and SDNND from sites 1 (river mouth) Ȃ 4 (upstream), whilst MMND stayed relatively constant, apart from a slight increase of NR and MMND at site 2 during the wet season (Table 4.3). When all sites were pooled there were distinct differences in all metrics between the dry and wet seasons: TA = 22.0±2.2 and

33.7±2.7, NR = 5.0±0.4, 7.0±0.5 and CR = 7.8±0.38 and 9.6±0.5. Spatial

112 differences were also apparent among sites, with site 1 having higher CR, particularly during the wet season (9.1±0.4) compared to the dry season

(7.4±0.5). The number of trophic levels for this assemblage remained fairly constant across sites except for site 4, which was very low (1.6±0.3). The trophic ecologies (MNND metric) of the assemblage were largely similar, however this metric doubled from site 1 (3.3±0.3) to site 2 (6.0±0.7) during the wet season and was the lowest at site 1 during the dry season (2.3±0.2).

113

Table 4.3 ƒ›‡ǯ•‡–”‹ •‘ˆ–Š‡‘—–ŠŽŽ‹‰ƒ–‘”‹˜‡”‹†-trophic taxa and shark species (Site 1 river mouth Ȃ Site 4 upstream). Numbers of species (n) at each site and season are included, those with n values < 5 were omitted from these analysis (highlighted grey). TA = Total Area, NR = range of G15N, CR = range of G15N, CD= centroid distance, MNND =mean nearest neighbour distance, SDNND = standard deviation of nearest neighbour distance.

Site All 1 2 3 4 Season Dry Wet Dry Wet Dry Wet Dry Wet Dry Wet n C. leucas 6 28 6 25 0 3 0 0 0 0 G. garricki 20 22 10 6 8 14 2 2 0 0 G. glyphis 3 0 3 2 0 0 0 0 0 0 R. taylori 0 0 0 0 0 0 0 0 30 8 J. novaeguineae 14 8 9 0 2 2 3 6 0 0 L. calcarifer 20 8 11 8 5 0 4 0 0 0 M. equidens 15 22 3 8 7 0 5 11 0 3 N. armiger 16 12 7 2 3 6 6 4 0 0 P. macrochir 8 8 0 0 1 1 7 0 0 7 R. nasutus 11 7 5 0 6 0 0 7 0 0 TA 22.0±2.2 33.7±2.7 17.2±3.4 14.2±3.1 8.5±2.9 0.0±0.0 5.2±1.5 2.8±2.1 - 0.0±0.0 NR 5.0±0.4 7.0±0.5 4.7±0.7 3.9±0.4 1.9±0.6 4.3±0.6 3.5±0.7 3.8±0.5 - 2.5±0.6 CR 7.8±0.38 9.6±0.5 7.4±0.5 9.1±0.4 9.7±0.7 4.1±0.6 4.4±0.7 6.9±0.5 - 2.0±0.6 CD 2.7±0.1 3.2±0.1 2.4±0.2 3.5±0.2 3.5±0.2 3.0±0.3 2.3±0.3 2.9±0.2 - 1.6±0.3 MNND 4.6±0.1 2.0±0.2 2.3±0.2 3.3±0.3 3.1±0.3 6.0±0.7 3.1±0.4 3.8±0.3 - 3.2±0.6 SDNND 2.0±0.2 0.8±0.2 1.2±0.3 1.6±0.4 1.5±0.5 0.0±0.0 0.9±0.5 0.7±0.5 - 0.0±0.0

Mid-trophic level species that were putative prey of sharks differed significantly

for both G15N (p < 0.01, R2= 14.18, Fdf = 4.725, 143) and G13C (p < 0.01, R2=

72.70, Fdf = 76.175, 143). Pairwise comparisons were all significant (p < 0.01)

for G13C, with the exception of N. armiger and P. macrochir. In contrast, there was

a low range of mean G15N values and pairwise comparisons for G15N were non-

significant (Fig. 4.1). Polydactylus macrochir had the highest mean G15N value

followed by L. calcarifer and N. armiger both of which were highly variable,

indicated by their large standard deviations (SD) that extended past P. macrochir

(Fig. 4.1). Species with similar G13C consisted of L. calcarifer and M. equidens

114 having lower G13C mean values, J. novaeguineae and N. armiger low G13C and G15N, and R. nasutus had the highest G13C values (Table 4.4, Fig. 4.1).

Figure 4.1 Plot of mid-trophic species (black dots) (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Nemapteryx armiger, Polydactylus macrochir and Rhinomugil nasutus) G13C and G15N showing standard deviation error bars overlayed over the wet (coloured circles) and dry (coloured triangles) season shark (Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori) isotope values adjusted for trophic discrimination.

115

Table 4.4 Mean values for essential fatty acids (EFA> 0.5%), G13C and G14N and their standard deviation in muscle (SD) tissue from six potential prey species of shark collected from the South Alligator River, Kakadu National Park, Australia. Included are indicator FAs for each species with a p <0.05.

Polydactylus Johnius Lates Macrobrachium Nemapteryx Rhinomugil Species macrochir novaeguineae calcarifer equidens armiger nasutus SI (ppm)

G13C -17.4±1.3 -19.3±1.6 -22.9±2 -21.4±1 -21.2±2.1 -14.6±1.9 G15N 8.3±1.8 8.2±1.1 8.2±1.7 8.8±1.5 9.3±3.8 6.6±1.1 C/N 2.8±0.3 2.9±0.1 3.1±0.8 2.8±0.0 4.2±1.1 2.9±0.2

EFA (%) ͳͺǣʹɘ͸ 2.3±1.1 1.9±0.8 3.9±1.9 9.5±5.0 3.4±2.9 1.0±1.3 ͳͺǣ͵ɘ͵ 0.6±0.3 0.2±0.1 1.2±1.1 1.1±1.2 2.0±2.7É 0.5±0.5 ʹͲǣʹɘ͸ 0.3±0.0 0.3±0.1 0.3±0.1 0.4±0.1 0.7±0.2 0.1±0.2 ʹͲǣ͵ɘ͸ 0.2±0.2 0.5±0.6 0.6±0.2 0.2±0.1 0.4±0.2 0.4±0.2 ʹͲǣ͵ɘͻȗ 0.1±0.1 0.6±2.2 0.0±0.0 0.2±0.4 0.0±0.0 0.0±0.0 ʹͲǣͶɘ͵ȀʹͲǣʹ 0.3±0.1 0.4±0.2 0.4±0.3 1.6±0.8É 0.4±0.3 0.6±0.3 ʹͲǣͶɘ͸ 10.6±2.3 9.8±2.5 10.6±4.7 10.3±2.8 7.1±3.9 5.7±2.0 ʹͲǣͷɘ͵ 4.9±1.5 4.9±2.1 1.7±2.4 9.0±2.9 É 2.6±2.2 8.7±4.0 22:2a 0.5±0.1 0.6±0.3É 0.3±0.2 0.0±0.0 0.2±0.1 0.2±0.1 22:3 2.1±0.4 1.8±0.9 1.8±1.2 0.6±0.3 2.0±0.9 3.3±3.6 ʹʹǣͶɘ͸ 1.3±0.5 2.1±2.0 2±0.9 0.5±0.4 1.9±0.8 0.7±0.3 ʹʹǣͷɘ͵ 0.0±0.0 0.2±0.7 0.6±1.1 0.1±0.2 0.2±0.9 3.1±4.1 É ʹʹǣͷɘ͸ 1.9±0.6 3.0±1.2É 1.8±1.2 0.6±0.2 1.2±0.6 1.0±0.4 ʹʹǣ͸ɘ͵ 13.6±3.2 16.6±4.5É 5.6±3.1 4.7±1.4 7.1±5.4 10±4.9

EFA< 0.5% 0.0±0.0 0.2±0.1 0.2±0.1 02±0.2 0.1±0.1 0.3±0.1

SAT 4.5±0.3 4.2±0.2 4.8±0.9 4.6±1.0 5.4±1.1 4.6±0.6 MUFA 1.6±0.1 1.5±0.1 2.0±0.4 1.5±0.3 1.8±0.4 1.7±0.6 PUFA 2.3±0.2 2.6±0.2 1.8±0.5 2.3±0.4 1.7±0.6 2.1±0.7 ɘ͵Ȁɘ͸ 0.7±0.2 0.8±0.5 1.8±0.8 0.7±0.2 1.0±0.4 0.6±1.0

É an indicator fatty acid for the species with p values < 0.05 based on Dufrêne and Legendre (1997), 20:3Z9* identified based on comparison with other C. leucas fatty acid literature; a standard was not available at the time of analyses EFA (essential fatty acids) <0.5 include 18:3Z6, 18:4Z3, 18:2a, 18:2b, 18:2c, 21:5Z3, 21:3, 22:2b, SAT Ȃ saturated fatty acid, MUFA - monounsaturated fatty acids, PUFA Ȃ polyunsaturated fatty acid

116 Significant seasonal differences were found between the wet and dry G13C values of prey but not G15N. Capture location was not significant in G15N (p = 0.08, R2=

4.5, Fdf = 2.33, 145) but was in G13C (p = <0.01, R2= 17.5, Fdf = 10.33, 145).

Significantly different pairs were found between sites 1 and 3 (t = 0.5, p <0.01), sites 1 and 4 (t = 1.0, p <0.01), and sites 1 and 2 (t = 0.7, p < 0.01).

Fourteen EFAs with >0.5 % representation within tissues appeared to separate across three broad divisions within these potential prey taxa (Table 4.4, Fig. 4.2,

4.2 and see Appendix 7.4). One group consisted largely of M. equidens, the second, P. macrochir, R. nasutus and J. novaeguineae and the third N. armiger and

L. calcarifer. However, it should be noted that individual N. armiger were dispersed over all groups whilst individual R. nasutus were spread amongst groups of J. novaeguineae, P. macrochir and N. armiger. The main EFA that separated M. equidens from the other prey species was 18:2Z6, whilst L. calcarifer were divided into two subgroups by a number of EFA; however the most influential were 20:2Z6 and 22:4Z6. The larger group of N. armiger was separated principally by 20:2Z6, J. novaeguineae 20:5Z6 and P. macrochir and

22:5Z6 separated R. nasutus.

117

Figure 4.2 Principal coordinate ordination of essential fatty acids (EFA)>0.05% in percentage abundance in prey and shark species (black circles surrounding •›„‘ŽȌǡ™‹–Š˜‡ –‘”‘˜‡”Žƒ›•‘ˆ–Š‡‘•–‹ˆŽ—‡–‹ƒŽ •ȋ‡ƒ”•‘ǯ• ‘””‡Žƒ–‹‘ above 0.1).

Significant differences in EFA profiles were found amongst prey species and there were significant interactions between species x season and species x capture location; but not between season x capture location (Table 4.5). Lates calcarifer and P. macrochir were very similar to each other and could not be separated by their FA profile using the Dufrêne - Legendre indicator species analysis. Johnius novaeguineae had the most FAs (6) that resulted in their separation from the other species with p-values <0.05, N. armiger and R. nasutus

118 had four, whilst M. equidens had three (Table 4.4; Appendix 7.5, 7.6: for indicator values and specific FAs).

Table 4.5 Comparison of species, season (wet and dry) and site locations (1 - 4) of essential fatty acids from the mid-taxa species (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Nemapteryx armiger, Polydactylus macrochir and Rhinomugil nasutus) in the South Alligator River, Kakadu, Australia using PERMANOVA. DF = degrees of freedom.

Variable DF Pseudo-F P(perm) Unique

perms

Species 3 11.6 <0.01 9932

Capture location 1 6.2 <0.01 9952

Species x Season 3 2.3 <0.01 9925

Species x Capture location** 8 1.6 0.01 9892

Season x Capture location** 2 1.4 0.2 9947

Species x Season x Capture 2 4.4 <0.01 9944 location **

**Not all species were included in capture location

4.4.2 Trophic linkages between sharks and putative prey taxa

Stable isotope analysis indicated that the majority of C. leucas had G13C values that were higher than most prey species within the South Alligator River system with R. nasutus being the most notable exception (Fig. 4.1). However, some individuals of C. leucas were also isotopically similar to P. macrochir and N. armiger (Fig. 4.1). Rhizoprionodon taylori were similar in G13C values to C. leucas and were also similar to R. nasutus and N. armiger. Stable isotope signatures within Glyphis spp. were similar to many of the prey species, particularly J.

119 novaeguineae, and M. equidens. The majority of G. garricki isotopic values were close to M. equidens, whilst in another group of G. garricki, isotopic values were similar to L. calcarifer.

When G13C and G15N values were pooled across all shark species and compared to each prey group in the mixing model, there was ~ 23% difference (Table 4.6).

There was little difference between the consumption of prey during the wet and dry seasons, with Group 2 (consisting of signatures between estuarine and freshwater) having the most difference (2.9±6.0%) and Group 1 (consisting of freshwater signatures) having the least (0.6±1.4%) (Table 4.6, Fig. 4.3a).

Differences in prey consumption by shark species appeared to be more important (3.7± 2.7%) than seasonal variation (0.8±1.6%). Carcharhinus leucas consumed prey from Group 2 and 3 (consisting of estuarine signatures) whilst R. taylori showed the greater consumption of prey species from Group 3 (Table 4.6,

Fig. 4.3b). Glyphis garricki and G. glyphis had the highest mean consumption from

Group 1 and 3. The Glyphis spp consumed the most from Group 1 although G. garricki had the highest proportion (67.8±14.3%) compared to G. glyphis (Table

4.6, Fig 3b). Interestingly, the Glyphis spp. consumed the lowest amount from the

Group 2 yet both consumed almost one third of prey from the Group 3. However,

R. taylori consumed the most of the four sharks from within Group 3.

120 Figure 4.3 Box and whisker plots from MixSIAR of a) Seasonal difference of shark diet. b) Sharks and the proportion of the source (Prey) that makes up their diet. Prey grouped based on G13C and source of C estimated from Loneragan et al. (1997) Group 1 consists of Lates calcarifer, Macrobrachium equidens, Johnius novaeguineae, Group 2 Polydactylus macrochir, Neoarius armiger and Group 3 was only Rhinomugil nasutus.

121 Table 4.6 Stable Isotope mixing model results from four species of sharks, Carcharhinus leucas, Glyphis garricki, G. glyphis and R. taylori in the South Alligator River, Australia. Results are the percentage mean proportion of the shark consuming from each prey group and the combined results over all prey groups and the difference between seasons ± standard deviation (SD). Group 1 consists of Lates calcarifer, Macrobrachium equidens, Johnius novaeguineae, Group 2 Polydactylus macrochir, Neoarius armiger and Group 3 was only Rhinomugil nasutus.

Prey group: Group 1% ± SD Group 2 % ± SD Group 3 % ± SD

Shark species

All Sharks 33.5±20.6 21.6±18.4 44.9±22.0

C. leucas 3.3±6.4 50.6±19.4 46.1±18.4

G. garricki 67.8±14.3 1.6±4.8 30.6±14.0

G. glyphis 61.2±19.2 4.9±9.9 33.9±18.5

R. taylori 13.9±1.39 13.4±16.2 72.7±17.5

Wet 2.8±3.9 47.4±12.2 49.8±10.9

Dry 3.5±5.3 44.5±18.2 52.0±16.5

Significant differences in EFA profiles were found among all shark and prey species (p < 0.01, Fdf = 20.269). Pairwise tests of EFA profiles further confirmed this for all species pairs (all p < 0.05) except for G. garricki and G. glyphis (p =

0.4), which was not found to be significantly different. A PCO indicated that C. leucas, G. garricki and G. glyphis all had a diverse array of EFAs (Fig. 4.2) and shared FAs with P. macrochir, L. calcarifer, J. novaeguineae, M. equidens, N. armiger and R. nasutus. However, there were slight interspecific differences between the sharks. Glyphis garricki and G. glyphis had high relative levels of

18:2Z6, which was not present in C. leucas, whilst G. glyphis also had high

122 contributions of 20:5Z3. Each of these FAs were also present in P. macrochir, J. novaeguineae, L. calcarifer, M. equidens, R. nasutus and N. armiger.

4.4.3 Shark intraspecific differences

Most shark species had over 65.0% average similarity of FAs among individuals

(C. leucas (67.4 %), G. garricki (68.9%), G. glyphis (67.7%) and R. taylori

(77.7%)). The four sharks had similar FAs WIC/TIC indices and only C. leucas

(0.90 (p < 0.01)) and G. garricki (0.92 (p < 0.01)) had significant values, whereas in G. glyphis (0.94, p=0.28) and R. taylori (0.95 p=1) these values were not significant. Only G. garricki could be compared for seasonal differences, which indicated very little change in FAs between the wet (0.92, p < 0.01) and dry season (0.93, p < 0.93).

4.5 Discussion

Spatial and seasonal differences in stable isotopes and fatty acids were found in the trophic range and diversity of putative prey of four species of sharks that utilise the South Alligator River. Whilst there were significant differences between putative prey species, their biochemical tracer compositions overlapped. This suggests that these species are consuming a range of basal sources and that there is a high degree of omnivory or consumption of omnivores amongst species (Jepsen & Winemiller 2002). This was particularly the case within L. calcarifer and N. armiger as they had a large G15N variance among individuals, which may indicate feeding from a range of trophic levels.

This high degree of omnivory and trophic generalism support the general fifth

123 principal of river and wetland food webs in the wetȂdry topics as outlined by

Douglas et al. (2005). This principal suggests that food chains are short, that species often feed across a number of trophic levels, and that there is relatively low dietary specialisation in tropical rivers (Douglas et al. 2005). The comparison of both SI and FA biochemical tracers demonstrated likely dietary links between prey and elasmobranchs. All four shark species were found to be generalist consumers of near shore and estuarine prey species, with little change over the two seasons.

4.5.1 Seasonal and spatial patterns of trophic range and dietary diversity among

putative prey taxa

The influx of organic sources at certain points along the river may explain the spatial differences in prey in the South Alligator River (Pusey et al. 2016). For example, site 1 had the greatest range of basal sources based on the CR. This site was the furthest upstream and has a mixture of terrestrial, freshwater and some limited marine basal sources, as has been found in other estuaries (Atwood et al.

2012). During the wet season in the upper river (sites 1 and 2), the trophic ecology of species appears to overlap more than during the dry season. This is perhaps a function of the changes in abiotic factors such as salinity and changes in hydrological patterns (Jardine et al. 2015, Pusey et al. 2016), which could similarly explain a slight decrease in spatial trophic diversity from the upper to lower river reaches. This can arise because the majority of species are unable to inhabit the mid-reaches of the river (Pusey et al. 2016) due to fluctuating

124 conditions (e.g. salinity) caused by both seasonal and tidal influences (Warfe et al. 2011, Jardine et al. 2015).

Like many tropical rivers season influenced the isotopic and FA composition in putative prey species and thus the trophic structure of the river (Winemiller &

Jepsen 1998, Roach et al. 2009, Ward et al. 2016). However, seasonal shifts in individual FA and SI biotracers were not reported previously in these elasmobranchs (Every et al. 2016). This may indicate that sharks are moving to consume their preferred prey or that they are consuming a variety of prey from a range of sites which may make identifying seasonal change difficult. Other large predators such as the estuarine crocodile Crocodylus porous and L. calcarifer across northern Australia were also found to consume prey whose basal sources were from outside their immediate surroundings (Jardine et al. 2016).

4.5.2 Links between putative prey to elasmobranchs

Putative prey showed significantly different mean G13C values among species suggesting that they are utilising a range of basal sources, such has been found in other wetȂdry tropical rivers (Jardine et al. 2016). Large variance in G15N may be attributed not only to omnivory or consumption of omnivores but may also be related to ontogenetic change as L. calcarifer has been found to switch from the consumption of smaller teleosts and Macrobrachium spp at 40 cm total length, to

Ariidae and Polynemidae prey alongside an increase in consumption of Mugilid and Engraulid fishes (Davis 1985). Whilst specific dietary studies have not been conducted for N. armiger, dietary ontogenetic change has been reported in other

125 Neoarius species (Dantas et al. 2012, Thorburn et al. 2014). Alternatively, this may be attributed to the varied diet of N. armiger that is reported to include teleosts, polychaetes and crustacea (Blaber et al. 1994). Due to the similarities and differences of biochemical tracers amongst the prey assemblage it appears that P. macrochir may feed on J. novaeguineae, as their EFAs overlap (high in

22:5Z6) and P. macrochir had higher G15N than J. novaeguineae. Neoarius armiger was the only teleost species that showed similarities to the biochemical profile of the crustacean M. equidens, which was very different to other putative prey species in that it was high in 18:2Z6.

Based on both biochemical tracers it appeared that not all sharks consumed putative prey species where they appeared to be sympatric. Carcharhinus leucas was a prime example of this, with most individuals caught at site 1 not appearing to consume prey with freshwater G13C values that were caught at these same sites. Although C. leucas had the highest mean G13C value, the mixing model indicated that they are consuming the majority of prey species from Group 2

(50.6±19.4) and 3 (46.1±18.4), which consisted of species with higher G13C values (more estuarine signatures) such as N. armiger. Similarly, other Neoarius spp. have been commonly reported in the stomach content analysis of populations of C. leucas in other estuarine ecosystems (Snelson et al. 1984,

Thorburn & Rowland 2008). However, as C. leucas had a high G13C value, C. leucas are likely to be consuming other estuarine prey species such as larger L. calcarifer (Heithaus et al. 2013) from these or nearby coastal locations that were not caught in this study. Especially since the majority of C. leucas did not appear to have strong trophic links with any of the putative prey explored here and had

126 significant intraspecific variation. This discrepancy in SI signatures versus capture location suggests high levels of prey movement may be occurring, or that a maternal signature is present within the sharks (Every et al. 2017). Movement data from the Shark River Estuary, Florida, USA, similarly found that elasmobranch species opportunistically captured prey entering the river from the flood plains (Matich & Heithaus 2014). We consider it unlikely that an abundant prey source was missed in our study, as a range of fishing methods were used to collect prey.

Fatty acids indicated that there were dietary links between a small cluster of C. leucas and N. armiger, and L. calcarifer and R. nasutus. These individuals had a size range of 69.5 Ȃ 99. 5 cm, which is approximately the same size range of the entire cohort so ontogenetic change is unlikely to explain these differences.

Interestingly, when the other sharks were included amongst the prey, C. leucas were very similar to G. garricki and R. taylori FAs, which may suggest that C. leucas is consuming them or they are consuming similar prey. Although difficult to evaluate without investigating stomach contents, elasmobranchs (including other C. leucas) have been found in the gut of C. leucas along with a variety of teleosts fish (Snelson et al. 1984) and in Australia C. leucas, crocodiles, pigs and birds (Thorburn & Rowland 2008).

Stable isotope mixing models of R. taylori suggest that they are also consuming the majority of Group 3 prey (with more estuarine signatures) (55.0 ± 34.0%) with some from Group 2 (13.4 ± 16.2%) and only a very small proportion of prey from the Group 1 (with more riverine signatures) (13.9 ± 1.39). The EFA profiles

127 also support the isotope mixing model as they suggest that R. taylori are consuming J. novaeguineae, P. macrochir, N. armiger with some individuals being close to R. nasutus. Previous studies of R. taylori indicate that they consume marine species (Simpfendorfer 1998, Munroe et al. 2014a), however it appears that they do consume prey that have assimilated biotracers from freshwater habitats. This is interesting, as R. taylori was not found to enter the river in a movement study in Queensland (Munroe et al. 2015). Some of R. taylori Ǯ•  profiles did not appear to have dietary links to any of the putative prey species and so may be consuming other marine prey (largely teleost spp.) similar to what was found in stomach content analysis (Simpfendorfer 1998). Therefore there may be some degree of resource partitioning occurring amongst the population that was not observed here, perhaps because the sampling effort was concentrated in the estuary.

Other shark species that can tolerate the riverine conditions are likely to access more riverine prey. For example, our study indicated that G. garricki are primarily consuming species from the freshwater prey group (Group 1) and had a low degree of intraspecific differences. This was supported by G. garricki EFA profiles, which indicated links with the freshwater and estuarine prey L. calcarifer and N. armiger and possibly P. macrochir and J. novaeguineae.

Corroborating these findings were the stomach contents of 6 individual G. garricki where Neoarius spp. and P. macrochir were also found (Thorburn &

Morgan 2004). Although G. glyphis was very similar to G. garricki, they consumed slightly more estuarine prey (Group 3)(33.9 ± 18.5% compared to 30.6 ± 14.0%) and less freshwater prey (61.2± 19.2% compared to 67.9 ± 14.3%). Their EFA

128 were associated with only L. calcarifer and N. armiger and the stomach contents of seven individuals indicated Nematolsa erebi, the freshwater prawn M. spinipes and spines of catfish were also found (Peverell et al. 2006). Both Glyphis spp. show a reliance on riverine resources, particularly G. garricki due to their apparent preference for upriver putative prey species. In contrast, C. leucas and

R. taylori had strong links to the mid-river prey and very low proportion of freshwater prey according to SI mixing models. This suggests that all four shark species have important trophic connections to the riverine environment.

4.6 Conclusions

Seasonal and spatial differences were found in the South Alligator River with the most trophic diversity and biochemical tracer variance at site 1 in the upper river. This variation in dietary biochemical tracers indicates the complexity of food webs in this river and appears to be a common feature of tropical estuaries

(Magnone et al. 2015). All sharks appeared to be generalist species, which may be due to the large amount of putative prey species (Pusey et al. 2016). Further exploration is required to explain why individual shark biotracers do not show evidence of seasonal change, yet prey species did. Another finding was that C. leucas had predominantly marine-based signatures, yet they were captured 80 km upstream. Both these aspects require further investigation such as combining movement (e.g. acoustic telemetry) and biochemical tracer data. Maternal signatures in the shark species may also affect these data, however there were no significant maternal signatures found in these individuals (Every et al. 2017).

One potential way to further our knowledge of the trophic ecology of these

129 species using FAs would be to conduct feeding trials so that the differing physiological responses to individual FAs can be calculated in dietary mixing models similar to isotopes. The use of both FAs and SI indicate that elasmobranchs consume a variety of prey taxa, with slight differences between elasmobranch species prey choice. However, all the elasmobranchs had trophic links to the river suggesting that riverine biomes are a very important source of diverse prey for these species of conservation concern.

4.7 Chapter acknowledgements

This research was conducted on the traditional country of the Bininj and

Mungguy people. We gratefully acknowledge the traditional custodians for allowing access to this country. We thank Peter Nichols, Peter Mansour, Grant

Johnson, Mark Grubert, Duncan Buckle, Roy Tipiloura, Dominic Valdez, Francisco

Zillamarin, Edward Butler, Claire Streten, Kirsty McAllister, Mirjam Kaestli and the crew of R.V. Solander for fieldwork and laboratory assistance and to Martin

Lysy (nicheROVER) and Andrew Jackson (SIAR) for their advice and assistance.

Charles Darwin University, CSIRO and the North Australia Marine Research

Alliance (NAMRA) provided funding, alongside collaborative partnerships in the

Marine Biodiversity and No”–Š‡”—•–”ƒŽ‹ƒ —„•‘ˆ–Š‡—•–”ƒŽ‹ƒ ‘˜‡”‡–ǯ•

National Environmental Research Program (NERP). All procedures performed in this study were conducted with the approval of the Charles Darwin University

Animal Ethics Committee (A12016), in conjunction with permits from NT

Fisheries and Kakadu National Park (RK805).

130 5 General Discussion

Emerging evidence suggests that tropical elasmobranchs are important trophic links in aquatic ecosystems (e.g. Borrell et al. 2010, Blanco-Parra et al. 2011, Wai et al. 2012, Polo-Silva et al. 2012, Tilley et al. 2013). This thesis adds to this evidence by focusing on a group of tropical coastal and euryhaline elasmobranchs (sharks and rays) in the South Alligator River, Northern

Territory, Australia. Although these species utilise a variety of dietary items from fresh, estuarine and marine sources, each species showed preferences for a particular source (Chapter 3). As highly-mobile trophic generalists, these elasmobranchs appear to provide an important vector for transporting assimilated energy and nutrients among river, estuary and nearby coastal marine biomes (Chapter 3, 4). Through the application of multiple trophic chemical biotracers, it was revealed that coastal and euryhaline elasmobranchs provide diverse trophic linkages across a tropical aquatic environment.

Chemical biotracers were used to identify some of the ecological roles of these species in the South Alligator River with smaller sample sizes than would have been required for traditional stomach content analysis. Moreover, chemical biotracers provided an integrated understanding of the contributions of various prey species and basal resources to the food webs upon which elasmobranchs rely. This represents a significant advance on information from stomach content analyses alone. From a methodological viewpoint, I found that the use of shark fin samples is a viable, non-lethal method for fatty acid (FA) analysis (Chapter 2), thus increasing the options for trophic studies in species of conservation concern

131 (Heupel & Simpfendorfer 2010, Hussey et al. 2012). The validation of a non- lethal method is particularly significant given that chondrichthyans tend to have slow growth and reproduction rates, and more than 50% of chondrichthyan species are currently considered as threatened or Near Threatened under IUCN

Red List of Threatened Species criteria (Dulvy et al. 2014).

Combined stable isotope and fatty acid analyses (Chapter 3) provided evidence that many coastal and euryhaline elasmobranchs are generalist mesopredators, with perhaps only Carcharhinus leucas considered as a higher order predator.

Notably, all of these elasmobranchs utilised a diversity of aquatic prey (Chapter

4). Most of the putative prey appeared to be omnivores, which may explain the variability found in the signatures of the elasmobranch consumers (Chapter 4).

These findings (high levels of generalism and omnivory) are similar to other aquatic consumers in tropical rivers, such as teleosts and macroinvertebrates

(Jepsen & Winemiller 2002, Douglas et al. 2005, Blanchette et al. 2014), which suggest an overarching need to take advantage of temporally variable food sources. The inundation of floodwaters during the wet season results in large abiotic and biotic changes and is a feature of tropical riverine food webs

(Winemiller & Jepsen 1998, Jardine et al. 2015, Ward et al. 2016a). This seasonal change was evident amongst biochemical tracers within putative prey species in the South Alligator River. However, this potential seasonality was only evident within the niche metrics for these elasmobranchs (Chapter 3, 4). The reasons for low temporal variation in trophic niche among the elasmobranchs is unclear, but may reflect elasmobranchs having a greater capacity to move long distances in

132 order to seek their preferred prey, before being forced to switch dietary items due to low availability.

5.1 Assessment of biomarkers

5.1.1 Source tissues for trophic biomarkers in elasmobranchs

Essential fatty acids (EFA) are largely dietary related, as vertebrates are unable to synthesize EFA (Dalsgaard et al. 2003). Among-tissue differences in FA profiles were found in C. leucas and G. garricki but not in G. glyphis. However, when all FAs >0.5% were compared individually, most of the EFAs were not significantly different among tissue types. Ultimately, muscle tissue was found to be more suitable for this study for FA analysis as further analysis is required to address the issues raised in Chapter 2, such as the discrepancies in non-dietary

FAs between fin and muscle tissue. As mentioned, there is however the possibility of species differences within tissues of the same type, therefore some caution is required when interpreting tissues in species that previously have not been analysed. Controlled studies on the uptake of FAs in a range of shark species are required to further validate the use of fin and muscle tissue. This has occurred in Heterodontus portusjacksoni (Beckmann et al. 2014b), where muscle, liver and blood serum were all found to have differences in the proportion of FAs that were likely related to physiological function. However, as this ancient Order represents a small temperate water species compared to the larger, tropical species examined in this thesis (Compagno et al. 2005), there are likely physiological differences that need to be considered. Thus, whilst the use of muscle and fin tissue provides a non-lethal means for trophic analyses,

133 validation for different species and consideration of variation among tissue types remains necessary when interpreting data using this approach.

5.1.2 Trophic biomarkers within coastal and euryhaline elasmobranchs and their

putative prey species

Whilst using the whole FA profile from a species can elucidate inter- and intraspecific differences, comparing individual or small groups of FA biomarkers can refine and add additional information in food web analyses. For example, by using the ratio of Z3/Z6 a significant difference was identified between elasmobranchs with marine/estuarine G13C signatures and those with freshwater G13C signatures (Every et al. 2017, Chapter 3). However, there are caveats to utilising these biomarkers, as other factors may influence results. For example, maternal reserves in the liver of neonates may be reflected in the neonatal signatures, which may confound the interpretation of results (Olin et al.

2011, Every et al. 2017, Chapter 3). The South Alligator population of C. leucas had comparatively high ratios of Z3/Z6 and higher G13C values which suggests a marine signature (Martínez-Álvareza et al. 2005, Özogul et al. 2007), yet these species were caught in upstream sites with strong freshwater influence. Also, the bioaccumulation of some of the EFAs detected in South Alligator River elasmobranchs, such as 22:5Z6, has been associated with increased growth of cod and scallops during early ontogeny (Parrish 2013). This evidence suggests that the physiological functions of individual EFAs must be considered to ensure that trophic interactions are interpreted correctly.

134 Some FAs are especially useful for determining basal sources within ecosystems.

For example, FA biomarkers from both terrestrial and marine biomes were found in all species in the South Alligator River, including those that can indicate sources from the terrestrial environment (Napolitano et al. 1997, Budge &

Parrish 1998). This FA indicated significant differences between Glyphis spp. and

C. leucas and R. taylori, suggesting that Gylphis spp. consumed more prey from the upper estuary than C. leucas and R. taylori. This was supported by all putative prey species having the largest proportion of terrestrial FAs compared to elasmobranchs, particularly Macrobrachium equidens which was collected under overhanging macrophyte branches by the edge of the river, possibly suggesting close links to terrestrial sources. Stable isotopes indicated that the migratory

Macrobrachium spinipes transports nutrients along the river, upstream as adults and downstream as larvae, which could suggest that its congener (M. equidens) is similarly transporting material from floodplains (Jardine et al. 2012a, Novak et al. 2015). However, recently it was found that the amount of marine-derived material carried by M. spinipes upstream was negligible, possibly due to the fast turnover off their tissue (Novak et al. 2017). Other FAs were linked to marine food webs, such as EPA 20:5Z3 and DHA 22:6Z3, which were found in slightly higher proportions in C. leucas and R. taylori (Chapter 3) than the other species, further suggesting that they consume marine prey sources.

Bacterial markers were revealed within elasmobranch tissue that also appeared high in relative abundance in the putative prey species M. equidens (Chapter 4).

These bacterial FAs consist of odd number monosaturated FAs, such as 15.1

(George & Parrish 2015). As the riverbed consisted primarily of mud, which is

135 likely to have a high proportion of bacterial FAs, this strong bacterial signature was not surprising (Bachok et al. 2003, Torres-Ruiz et al. 2007); however, in elasmobranchs odd number FAs are uncommon. In tropical river systems, a large proportion of primary production occurs within the periphyton (Pettit, et al.

2016b, Ward et al. 2016b), which consists of cyanobacteria, algae and detritus

(Geesey et al. 1978). All of these are likely to produce odd number FAs (Belicka et al. 2012b). This provides a possible explanation for the higher proportion of odd number FAs passed up the food web and accumulated in elasmobranchs in the South Alligator River, whereas elasmobranchs relying on more marine food webs would be expected to have less odd number FAs (Pethybridge et al. 2010,

Couturier et al. 2013b). In order to test this hypothesis conclusively, FAs of periphyton would need to be analysed and compared to elasmobranchs and other members of the food web.

The FA 20:3Z9 was also found in high quantities in C. leucas, which was also reported in a population of C. leucas in the Florida Everglades (Belicka et al.

2012a). This FA is considered to be indicative of EFA deficiencies, as the lack of

Z3 and Z6 encourages the synthesis of 20:3Z9 through the Z9 pathway starting from oleic acid 18:1Z9 (Le et al. 2009). This suggests that either this population of C. leucas is nutrient deficient or there is another explanation, such as the FA having an important physiological role. Determining the relative influences of nutrient deficiency versus environmental or other factors on 20:3Z9 assimilation in C. leucas in northern Australia is an interesting area for future research.

136 Given the large number of FAs within elasmobranch muscle tissue, as well as the diverse ecological roles and range of habitats that elasmobranchs occupy, a cautious approach is required when making inferences about trophic links.

Diverse trophic relationships can be difficult to elucidate through remote metrics, such as muscle tissue compared to tissue from prey. The approach employed in this thesis was to combine multiple biomarkers, with different tissue processing/retention times, to examine both short- and long-term patterns of diet consumption (MacNeil et al. 2005, Pethybridge et al. 2010,

Beckmann et al. 2014b, Matich et al. 2015). Based on these results, it can be concluded that integrated approaches using multiple biotracers provides a powerful means of making food web inferences that would not be possible via the use of a single biotracer.

5.2 Trophic niche of coastal and euryhaline elasmobranchs

Collectively, evidence is provided to support coastal and euryhaline elasmobranch and putative prey species trophic position (TP), basal source, niche metrics, overlap and dietary composition of elasmobranchs via FA and SI biochemical tracers. Although there was some overlap in niche space among species, these results provide evidence of resource partitioning and overlap within this elasmobranch assemblage. This suggests that each species has some specific dietary requirements, however, the presence of niche overlap may indicate competition for trophic resources, particularly amongst the Glyphis spp.

137 5.2.1 Trophic position

Trophic position is important in the context of food web analysis, as it indicates the length of food chains and the number of trophic links from basal sources to apex predators. Because of this, the TP of elasmobranchs in the South Alligator

River were estimated order to understand their ecological role. Of the species studied, juvenile C. leucas had a TP of 4.8±0.9, which suggests that C. leucas is an apex predator in this region (Chapter 3). Previous studies had also concluded that C. leucas is an apex predator that consumes a wide range of prey including other elasmobranchs (Snelson et al. 1984, Thorburn & Rowland 2008). As such,

C. leucas appears to be a potential keystone species whose abundance could affect a range of ecological processes in tropical rivers and estuaries.

Based on the current evidence, the South Alligator River assemblage appears to have short food chains, which is indicative of dynamically stable environments.

For instance, estuaries can be under constant physical change, yet maintain a relatively stable assemblage of species through time (Vander Zanden & Fetzer

2007). Food chain length and TP also provide an indication of the functional groups that species are consuming. For example, juvenile G. glyphis had the lowest TP of the elasmobranchs (3.2±0.8) suggesting that they are consuming primary consumers. In contrast, tertiary consumers such as adult Carcharodon carcharias prey on mammals and have a mean TP of 5.2, whilst piscivorous sharks that consume planktivores have mean TP of 4.1 (Hussey et al. 2015).

Through estimating the TP of an assemblage of elasmobranchs in the South

Alligator River, I also provide baseline data that can be used to monitor shifts in trophic levels over time. For example, mean TP have been used in ecosystem

138 modelling to identify ecosystems shifts due to disturbances such as overfishing

(Branch et al. 2010, Pauly et al. 1998). However, TP can be difficult to estimate within estuarine environments due to the presence of multiple food webs, abundance of middle order prey, increased diversity of basal sources and incidence of omnivory (Layman et al. 2007, Abrantes & Sheaves 2009). This can result in difficulty finding a baseline consumer that can assimilate a range of basal sources. Another solution may be the use of multiple baseline consumers.

5.2.2 Resource partitioning and linking freshwaters with coastal biomes

Carbon isotopes (G13C) provide a means to understand the base of food webs. I explored G13C variation and showed that the basal sources supporting elasmobranchs in the South Alligator River could be partitioned into two broad groups: marine/estuarine and freshwater/estuarine based on the G13C values of species and compared to primary sources in a northern Queensland tropical system (Loneragan et al. 1997). Slight seasonal differences were found in the multivariate niche metrics of C. leucas and G. garricki, but not within individual biochemical markers. Much of the niche variation of species appeared to be driven by G13C. For example Rhizoprionodon taylori, Carcharhinus amboinensis, C. leucas and Pristis pristis appeared to have values that typified marine environments, whilst Glyphis garricki, G. glyphis and Urogymnus dalyensis were characterized by values representative of the estuarine/freshwater environment.

139 Comparisons between putative prey species and elasmobranch consumers G15N and G14C, using assemblageȂwide measures of trophic structure amongst the sampled species, revealed seasonal differences and potential spatial hotspots within the catchment (Chapter 3). This may be typical of tropical rivers in the wet-dry tropics. For example, Blanchette et al. (2014) found that the Burdekin

River catchment, Queensland had as much spatial variability in basal sources and food web structure, as occurred across several other rivers in the region. In the

South Alligator River, the most upstream site had the greatest range of basal carbon sources and the most spread of trophic levels, indicating a diverse and productive section of the river. During the wet season this site also had an increase in trophic diversity compared to the dry season SIs (Chapter 4). The upstream site was where the higher numbers of elasmobranchs were caught in this study; a result that may reflect the effects of regular flooding on primary productivity and habitat diversity at the site (Junk et al. 1989, Pusey et al. 2016).

High abundance of euryhaline elasmobranchs in the upper estuarine reaches of the South Alligator River may also potentially be a response to the distribution and abundance of prey species. Based on the capture data (Table, 3.1, 4.1, Fig.

3.1) C. leucas was the most commonly caught upstream at site 1 whilst G. garricki was primarily collected further upriver and downstream of this site. Glyphis garricki G13C signatures were also found to have more freshwater basal sources than C. leucas, suggesting resource partitioning is occurring within this assemblage of elasmobranchs. Whilst there are trophic differences among species, it is difficult to know without direct testing whether this represents a preference for a particular prey type, abiotic conditions or an outcome of

140 competition. The presence of C. leucas may also influence other individuals of C. leucas and other species within the assemblage through fear mediated behaviour change (Ritchie & Johnson 2009): i.e., when the fear of being eaten or attacked

ƒ—•‡•ƒ Šƒ‰‡‹ƒ‹†‹˜‹†—ƒŽǯ•„‡Šƒ˜‹‘—”•— Šƒ•‘˜‹‰–‘ƒŽ‡••’”‘†— –‹˜‡ habitat. This behavioural change may then influence what other species are consuming and their movements. Predator avoidance has been suggested in juvenile C. limbatus where they were not present in areas of high prey density

(Heupel & Hueter 2002). It is also possible that G. garricki move into these areas over different time scales, consuming slightly different food types which may reduce competition pressure, as was found in four sympatric species

(Rhizoprionodon terraenovae, C. limbatus, C. isodon, C. brevipinna) within

Apalachicola Bay, Florida (Bethea et al. 2004). Glyphis glyphis and particularly U. dalyensis and Pristis pristis were rarely caught in this river during this study and it is possible that interactions with predators contribute to this distributional pattern. In the Fitzroy River, Western Australia, 23 of P. pristis examined were found to have bites on their bodies attributed to either Crocodylus johnstoni or C. leucas, which may suggest their behaviour ȋ‘˜‡‡–—’•–”‡ƒ‹–‘Ǯ•ƒˆ‡”ǯ waters) is influenced by predator avoidance (Morgan et al. 2017).

Although niche overlap of biochemical tracers usually represents shared resources, cannibalistic behaviours or maternal signatures may also cause overlap (Heithaus et al. 2013; Olin et al. 2014). This may have occurred between

R. taylori and C. leucas as R. taylori were caught in the mouth of the river and were found to have a strong SI trophic overlap with C. leucas. Whether this was based on C. leucas consuming similar prey or whether C. leucas is consuming R.

141 taylori is not clear. Also, these species were not caught in the same location, therefore their prey must either be highly mobile or C. leucas is moving further down river (or R. taylori is moving upstream) than expected, considering their

G13C values and what is known of their movement patterns in other locations

(Chapter 3, Munroe et al. 2014b). Both adult and juvenile C. leucas are known to consume elasmobranchs (Snelson et al. 1984; Thorburn et al. 2008), therefore it is possible a maternal influence is present even though it was only found in G13C in this study (Chapter 3). Data on movements of this population is required, along with a greater understanding of maternal signatures in C. leucas neonates, to fully resolve our understanding of trophic niche overlap.

5.2.3 Intraspecific differences within the coastal and euryhaline elasmobranchs

Although each elasmobranch species appeared to consume a subset of available prey, niche widths were relatively large, particularly for C. leucas (SI-based) and

G. garricki (FA-based), which indicates that each have a generalist diet (Chapter

3). However, when the individual mean value of muscle tissue SI and FAs at a species level were considered, large standard deviations were revealed. This can sometimes indicate intraspecific differences within species Ȃ i.e. generalist type

B (Bearhop et al. 2004; see Chapter 1, 3), however, more refined equations using total niche width can be used to compare an assemblage of species (Bolnick et al.

2002). These intraspecific equations have been developed that utilise multiple variables such as stomach content analysis and stable isotopes to calculate significant indices of intraspecific difference within a species (Roughgarden

1972, Bolnick et al. 2002, Matthews & Mazumder 2004, Araújo et al. 2011). As

142 this study did not use stomach content analysis, only FAs could be used to calculate the elasmobranchs specialisation index (Chapter 4). Based on the FAs in the elasmobranchs of the South Alligator River there was very limited individual specialisation, contrary to what was indicated by the large standard deviations in both SIs. Generalist diets are common in many freshwater taxa in tropical riverine environments (Douglas et al. 2005, Pusey et al. 2010). This study demonstrates that this broad pattern also extends to coastal and euryhaline elasmobranchs in these systems.

An important caveat to note here is that the individuals sampled for all species

(except R. taylori) were juveniles. It is possible that dietary specialisation changes with ontogeny (e.g., Sandbar Shark Carcharhinus plumbeus (McElroy et al. 2006)). To further our knowledge of the trophic ecology of euryhaline elasmobranchs over their full life history, samples taken from adult individuals across tissue structures that assimilate material over longer periods (e.g. skeletal elements) may be beneficial. For example, Vander Zanden et al. (2010) analysed the scutes of Loggerhead Turtles Caretta caretta to examine of the SI composition of material deposited over multiple years (up to 12 years). This analysis indicated strong individual differences amongst individuals in contrast to the previously hypothesised generalist dietary mode for the species (Vander

Zanden et al. 2010). Similar approaches have also been used to examine ontogenetic shifts in diet of teleost fishes using otoliths (McMahon et al. 2011).

Whilst elasmobranchs do not possess otoliths or similar bony growths, several studies have analysed calcified cartilage such as vertebrae to examine

143 ontogenetic dietary change in elasmobranchs (e.g., C. carcharias Kerr et al. 2006,

Estrada et al. 2006).

5.3 Ecological role of coastal and euryhaline elasmobranchs

Apex predators, such as some elasmobranches, have been hypothesised to exert top-down control on aquatic communities via food web interactions (Estes 1998,

Williams et al. 2004). However, quantifying this trophic influence has been difficult due to the high degree of mobility and ontogenetic habitat shifts displayed by sharks and rays (Heupel et al. 2014, Grubbs et al. 2016). However, a mass balance model of trophic networks to determine feeding relationships was used by Libralato et al. (2006) to establish that elasmobranchs often function as keystone species that have a larger influence on an ecosystem than what their biomass might predict (Power et al. 1996). This thesis did not aim to specifically determine if the elasmobranchs within the South Alligator are keystone species.

However, considering the trophic connections they provide across biomes, it appears likely that euryhaline elasmobranchs at least perform an important role in cross-biome ecosystem links through trophic interactions.

5.4 Potential impacts of changes in tropical riverine environments

When considering the impacts of disturbance to coastal and euryhaline elasmobranch food webs in the South Alligator River, we must consider both elasmobranch life history and their trophic ecology. For instance, R. taylori is a relatively fecund shark, had a medium sized niche width relative to other elasmobranchs in the South Alligator and has the ability to utilise estuaries and

144 coastal areas. This species, therefore, is likely to be one of the least at-risk species of this assemblage, as long as the northern coastal ecosystem remains in good condition (Munroe et al. 2014b, 2015, Every et al. 2017, Chapter 3, 4).

However, most euryhaline elasmobranchs are of conservation concern because they require riverine environments as well as marine biomes, have low fecundity and are slow growing (Dulvy et al. 2014). Added to these general risk factors is the fact that female C. leucas, P. pristis, G. glyphis and G. garricki are philopatric

(returning to their natal habitat to give birth) (Phillips et al. 2011, Tillett et al.

2012, Feutry et al. 2014, 2017). As nursery areas for these species, tropical river systems are extremely important for the reproductive output of euryhaline elasmobranchs. Although there has been increased research on euryhaline elasmobranchs since concern was raised regarding their conservation status in

–Š‡ͳͻͻͲǯ•ȋ‘’ƒ‰‘Ƭ‘‘ͳͻͻͷȌǡŽ‹––Ž‡”‡ƒ‹•‘™ƒ„‘—––Š‡‹”ƒ†—Ž–Ž‹ˆ‡ stages (except for C. leucas), particularly the Glyphis spp. For example, the first adult G. glyphis ever recorded was found in 2014 in a market place in Papua New

Guinea (White et al. 2015). The length of time that different species remain in riverine/estuarine environments is also uncertain, although for G. glyphis it has been estimated that they remain in rivers for approximately 6 years (Feutry et al.

2017). Therefore there is a need for research into the adult life stages of Glyphis spp.

Perhaps the greatest risk for coastal and euryhaline elasmobranchs is physical changes to their habitat that would limit connectivity, and/or change prey availability (Atwood et al. 2012). Seasonal flow regime is critical to community structure in tropical wetȂdry rivers such as the South Alligator (Warfe et al.

145 2011, Jardine et al. 2015, Pettit et al. 2016a). As terrestrial floodplains become inundated, nutrients and energy from these highly productive areas are transported back to the river channel both as dissolved solutes and as assimilated material within the tissues of fishes and other mobile organisms

(Jardine et al. 2012a, Polis & Winemiller 2013). Thus, connectivity of floodplains to the main river is considered a major driver of biotic productivity in tropical rivers and their estuaries (e.g. Warfe et al. 2011; Jardine et al. 2012b; Pettit et al.

2016a).

As connectivity is critical to biota in tropical rivers, any changes to water flow that reduce connectivity could have serious ramifications for euryhaline elasmobranchs as it may interfere with their ability to consume prey or avoid larger predators. For example, a barrage in the Fitzroy River, Western Australia, prevents migration of fish and elasmobranchs except during large floods, causing species such as P. pristis, C. leucas and L. calcarifer to be trapped below the barrage where they are highly vulnerable to predation and harvest (Thorburn et al. 2007). These types of barriers may prevent the transport of nutrients up or down stream via large predators such as elasmobranchs. This may cause habitat fragmentation that potentially causes niche width collapse, potentially leading to population decline and even local extirpation (Layman et al. 2007a). Changes such as this are unlikely in the South Alligator River and nearby rivers because they are protected by a World Heritage National Park (Kakadu National Park).

However, few other rivers nationally and globally are afforded such protection and are therefore vulnerable to the effects of human disturbance on ecosystem connectivity.

146 5.5 Future Directions

Although this thesis has contributed to the growing body of work on the trophic ecology of elasmobranchs, there are key areas for further research. From a methodological standpoint, a key area for future work is an improved understanding of biotracer uptake into different elasmobranch tissues to assist using complex biotracer data to infer trophic linkages over time and space in the wild. Although controlled feeding studies are challenging in large elasmobranchs due to issues of their husbandry (Hussey et al. 2012), a greater number of experimental studies on a wide range of organisms would be one avenue for addressing this knowledge gap. As the metabolism of elasmobranch fins are slower compared to other tissues (Malpica-cruz et al. 2012), the rate FA and SI are taken up by muscle and fins should be compared concurrently. If tissues varied, then these tissues could be analysed in unison to determine temporal dietary differences. Ontogenetic differences in fin tissue may also be important as adult fin tissue may differ from that of smaller animals as cartilaginous sections will calcify with age (Compagno 1999). In order to determine the effect of this variability in fins should be sampled over different sections of the fin, and comparisons made between different age groups.

Combining evidence from independent sources is often an effective way to get a holistic picture of trophic connections across the landscape. In that sense, integration of spatial movement data with seasonal shifts in dietary signature would be a key area for future research on the trophic ecology of euryhaline elasmobranchs. Not only will this help confirm the trophic hypotheses developed here, but also indicate whether there are focal feeding areas or if elasmobranch

147 predators forage widely on dispersed prey. Biotelemetry (e.g. radio and acoustic telemetry) has been used previously to determine the movement patterns of a number of fish species in tropical riverine environment such as L. calcarifer

(Crook et al. 2016), euryhaline sharks (Pillans et al. 2009, Whitty et al. 2009,

Heupel et al. 2010, Matich & Heithaus 2015) and is currently being used in the

South Alligator River on some of the studied elasmobranch species. When compiled, this data could be combined with the results in this thesis to help explain the movement patterns of species and their influence on the trophic ecology of the system.

Explicit protection of coastal and euryhaline elasmobranchs in ecosystem-based management approaches is a unique challenge requiring cross-agency collaboration. Evidence in this study suggests that tropical coastal, estuarine and freshwater environments are linked via large-scale cross-biome trophic connections underpinned by mobile elasmobranchs and other mobile organisms.

Such connections suggest that large-scale management approaches, which cover multiple biomes, are required to protect these species and further our understanding of the impact of the loss of these predatory elasmobranchs from aquatic ecosystems. While this will mean otherwise disparate freshwater and marine-focused management agencies will need to closely collaborate on their approaches, the imperative to do so is there. It is now widely recognised that sustainable fisheries should be managed in estuaries and coastal waters globally to minimise the incidence of elasmobranchs of conservation concern being caught as by-catch (Wiley & Simpfendorfer 2007, Blaber et al. 2009, Speed et al.

2010). Moreover, these unique predatory species have a strong potential to be a

148 key force in shaping aquatic communities and ecosystem function. Coastal and euryhaline elasmobranchs could provide key focal groups for aquatic conservation and management agencies to develop adaptive management approaches that protect aquatic ecosystem integrity.

149 6 References

Abrams P (1987) On classifying interactions between populations. Oecologia

73:272Ȃ281

Abrantes K, Sheaves M (2009) Food web structure in a near-pristine mangrove

area of the Australian Wet Tropics. Estuar Coast Shelf Sci 82:597Ȃ607

Alfaro, AC, Thomas F, Sergent L, Duxbury M (2006) Identification of trophic

interactions within an estuarine food web (northern New Zealand) using

fatty acid biomarkers and stable isotopes. Estuar Coast Shelf Sci 70:271Ȃ

286.

Araújo MS, Bolnick DI, Layman CA (2011) The ecological causes of individual

specialisation. Ecol Lett 14:948Ȃ58

Araújo MB, Guisan A (2006) Five (or so) challenges for species distribution

modelling. J Biogeogr 33:1677Ȃ1688

””‹‰–‘ǡ‹‡‹ŽŽ‡”ǡ‘ˆ–—• ǡ‹ȋʹͲͲʹȌ ‘™‘ˆ–‡†‘ˆ‹•Š‡•Dz”—

‘‡’–›dzǫ ‘Ž‘‰›ͺ͵ǣʹͳͶͷȂ2151

Atwood TB, Wiegner TN, MacKenzie RA (2012) Effects of hydrological forcing on

the structure of a tropical estuarine food web. Oikos 121:277Ȃ289

Bachok Z, Mfilinge PL, Tsuchiya M (2003) The diet of the mud clam Geloina

coaxans (Mollusca, Bivalvia) as indicated by fatty acid markers in a

subtropical mangrove forest of Okinawa, Japan. J Exp Mar Bio Ecol 292:187Ȃ

197

Barnett A, Redd KS, Frusher SD, Stevens JD, Semmens JM (2010) Non-lethal

method to obtain stomach samples from a large marine predator and the

use of DNA analysis to improve dietary information. J Exp Mar Bio Ecol

393:188Ȃ192

150 Bascompte J (2009) Disentangling the web of life. Science 325:416Ȃ419

Bascompte J, Melián CJ, Sala E (2005) Interaction strength combinations and the

overfishing of a marine food web. Proc Natl Acad Sci U S A 102:5443Ȃ5447

Bearhop S, Adams CE, Waldron S, Fuller RA, Macleod H (2004) Determining

trophic niche width: a novel approach using stable isotope analysis. J Anim

Ecol 73:1007Ȃ1012

Beckmann CL, Mitchell JG, Seuront L, Stone DAJ, Huveneers C (2013)

Experimental evaluation of fatty acid profiles as a technique to determine

dietary composition in benthic elasmobranchs. Physiol Biochem Zool

86:266Ȃ278

Beckmann CL, Mitchell JG, Seuront L, Stone DAJ, Huveneers C (2014a) From egg

to hatchling: preferential retention of fatty acid biomarkers in young-of-the-

year Port Jackson sharks Heterodontus portusjacksoni. J Fish Biol 85:944Ȃ

952

Beckmann CL, Mitchell JG, Stone DAJ, Huveneers C (2014b) Inter-tissue

differences in fatty acid incorporation as a result of dietary oil manipulation

in Port Jackson sharks (Heterodontus portusjacksoni). Lipids 49:577Ȃ590

Belicka LL, Matich P, Jaffé R, Heithaus MR (2012a) Fatty acids and stable isotopes

as indicators of early-life feeding and potential maternal resource

dependency in the Bull Shark Carcharhinus leucas. Mar Ecol Prog Ser

455:245Ȃ256

Belicka LL, Sokol ER, Hoch JM, Jaffé R, Trexler JC (2012b) A molecular and stable

isotopic approach to investigate algal and detrital energy pathways in a

freshwater marsh. Wetlands 32:531Ȃ542

Belling GB, Abbey M., Campbell JH, Campbell GR, (1997) Lipid content and fatty

151 acid composition of 11 species of Queensland (Australia) fish. Lipids 32,

621Ȃ625.

Berger J, Stacey PB, Bellis L, Johnson MP (2001) A mammalian predator-prey

imbalance: grizzly bear and wolf extinction affect avian neotropical

migrants. Ecol Appl 11:947Ȃ960

Bethea DM, Buckel JA, Carlson JK (2004) Foraging ecology of the early life stages

of four sympatric shark species. Mar Ecol Prog Ser 268:245Ȃ264

Blaber SJM, Brewer DT, Salini JP (1994) Diet and dentition in tropical ariid

catfishes from Australia. Environ Biol Fishes 40:159Ȃ174

Blaber SJM, Dichmont CM, White WT, Buckworth RC, Sadiyah L, Iskandar B,

Nurhakim S, Pillans RD, Andamari R, Dharmamadi, Fahmi (2009)

Elasmobranchs in southern Indonesian fisheries: the fisheries, the status of

the stocks and management options. Rev Fish Biol Fish 19:367Ȃ391

Blanchette ML, Davis AM, Jardine TD, Pearson RG (2014) Omnivory and

opportunism characterize food webs in a large dry-tropics river system.

Freshw Sci 33:142Ȃ158

Blanco-Parra M-P, Galván-Magaña F, Márquez-Farías JF, Niño-Torres CA (2011)

Feeding ecology and trophic level of the banded guitarfish, Zapteryx

exasperata inferred from stable isotopes and stomach contents analysis.

Environ Biol Fishes 95:65Ȃ77

Bligh EG, Dyer WJ (1959) A rapid method of total lipid extraction and

purification. Can J Biochem Physiol 37:911Ȃ917

Bishsop KA, Allen SA, Pollard DA (1980) Ecological studies on the freshwater

fishes of the Alligator Rivers Region, Northern Territory; Volume II:

Synecology

152 Bishsop KA, Allen SA, Pollard DA (1980) Ecological studies on the freshwater

fishes of the Alligator Rivers Region, Northern Territory; Volume 1: Outline

of the study summary conclusions and recommendations.

Bolnick DI, Amarasekare P Araújo MS, Burger R, Levine, Jonathan M, Novak M,

Rudolf VHW, Schreiber SJ, Urban MC, Vasseur DA (2011) Why intraspecific

trait variation matters in community ecology. Trends Ecol Evol 26:183Ȃ192

Bolnick DI, Svanbäck R, Fordyce JA, Yang LH, Davis JM, Hulsey CD, Forister ML

(2003) The ecology of individuals: incidence and implications of individual

specialization. Am Nat 161:1Ȃ28

Bolnick DI, Yang LH, Fordyce JA, Davis JM, Svanbäck R (2002) Measuring

individual-level resource specialization. Ecology 83:2936Ȃ2941

Borrell A, Cardona L, Kumarran RP, Aguilar A (2010) Trophic ecology of

elasmobranchs caught off Gujarat, India, as inferred from stable isotopes.

ICES J Mar Sci 68:547Ȃ554

Branch TA, Watson R, Fulton EA, Jennings S, McGilliard CR, Pablico GT, Ricard D,

Tracey SR (2010) The trophic fingerprint of marine fisheries. Nature

468:431Ȃ435

Brett MT (2014) Resource polygon geometry predicts Bayesian stable isotope

mixing model bias. Mar Ecol Prog Ser 514:1Ȃ12

Brewer (2014) D.T., Blaber, S.J.M., Salini, J.P. and Farmer, M.J., 1995. Feeding

ecology of predatory fishes from Groote Eylandt in the Gulf of Carpentaria,

Australia, with special reference to predation on penaeid prawns. Estuar

Coast Shelf Sci, 40:577Ȃ 600.

”‹†ǯ‘—”ǡ—„‘‹• ȋʹͲͳ͵Ȍ •‘–‘’‹ †‹˜‡”•‹–›‹†‹ ‡•ǣ ‘™sensitive to

food web structure? (DW Pond, Ed.). PLoS One 8:e84198

153 Bromaghin JF, Rode KD, Budge SM, Thiemann GW (2015) Distance measures and

optimization spaces in quantitative fatty acid signature analysis. Ecol Evol

5:1249Ȃ62

Bruno JF, Stachowicz JJ, Bertness MD (2003) Inclusion of facilitation into

ecological theory. Trends Ecol Evol 18:119Ȃ125

Budge SM, Iverson SJ, Bowen WD, Ackman RG (2002) Among- and within-species

variability in fatty acid signatures of marine fish and invertebrates on the

Scotian Shelf, Georges Bank, and southern Gulf of St. Lawrence. Can J Fish

Aquat Sci 59:886Ȃ898

Budge SM, Cooper MH, Iverson SJ (2004) Demonstration of the deposition and

modification of dietary fatty acids in Pinniped blubber using radiolabelled

precursors. Physiol Biochem Zool 77:682Ȃ687

Budge SM, Parrish C (1998) Lipid biogeochemistry of plankton, settling matter

and sediments in Trinity Bay, Newfoundland. II. Fatty acids. Org Geochem

Budge SM, Penney SN, Lall SP (2011) Response of tissue lipids to diet variation in

Atlantic salmon (Salmo salar): Implications for estimating diets with fatty

acid analysis. J Exp Mar Bio Ecol 409:267Ȃ274

Budge SM, Penney SN, Lall SP, Trudel M (2012) Estimating diets of Atlantic

salmon (Salmo salar) using fatty acid signature analyses; validation with

controlled feeding studies. Can J Fish Aquat Sci 69:1033Ȃ1046

Bunn SE, Leigh C, Jardine TD (2013) Diet-tissue fractionation of G15N by

consumers from streams and rivers. Limnol Oceanogr 58:765Ȃ773

Cabana G, Rasmussen JB (1994) Modelling food chain structure and contaminant

bioaccumulation using stable nitrogen isotopes. Nature 372:255Ȃ257

Callaway RM (2009) Facilitation and the organization of plant communities. In:

154 The Princeton guide to ecology.p 478Ȃ490

Campbell HA, Hewitt M, Watts ME, Peverell SC, Franklin CE (2012) Short- and

long-term movement patterns in the freshwater whipray (Himantura

dalyensis) determined by the signal processing of passive acoustic telemetry

data. Mar Freshw Res 63:341Ȃ350

Carlisle AB, Kim SL, Semmens BX, Madigan DJ, Jorgensen SJ, Perle CR, Anderson

SD, Chapple Tk, Kanive PE, Block BA (2012) Using stable istope analysis to

understand the migration and trophic ecology of norhteastern Pacific white

sharks (Carcharodon carcharias). PLoS One 7:e30492

Carpenter KE, Niem VH (1999) The living marine resources of the Western

Central Pacific. Volume 4. Bony fishes part 2 (Mugilidae to Carangidae). FAO

Species Identif Guid Fish Purp 4:2069Ȃ2790

Castell JD, Bell JG, Tocher DR, Sargent JR (1994) Effects of purified diets

containing different combinations of arachidonic and docosahexaenoic acid

on survival, growth and fatty acid composition of juvenile turbot

(Scophthalmus maximus). Aquaculture 128:315Ȃ333

Chase JM, Leibold MA (2003) Ecological niches: linking classical and

contemporary approaches. University of Chicago Press

Clarke KR, Gorley R (2006) PRIMER v6 PRIMER-E Ltd. Plymouth, UK

Clarke K, Warwick R, (2001) Change in marine communities: An approach to

statistical analysis and interpretation, 2nd ed. PRIMER-E, Plymouth.

Cliff G, Dudley SFJ (1991) Sharks caught in the protective gill nets off Natal, South

Africa. 4. The bull shark Carcharhinus leucas Valenciennes. South African J

Mar Sci 10:253Ȃ270

Cohen JE, Pimm SL, Yodzis P, Saldana J (1993) Body sizes of animal predators

155 and animal prey in food webs. J Anim Ecol 62:67

Compagno LJV (1984) FAO Species catalogue. Vol. 4. Sharks of the world. An

annotated and illustrated cataloge of shark species known to date. Part 2.

Carcharhiniformes. FAO Fish Synopsis 4:251Ȃ655

Compagno LJV (1999) Endoskeleton. In: Hamlet WC (ed) Sharks, skates, and

rays: the biology of elasmobranch fishes, 1st edn. The Johns Hopkins

University Press, Balitmore, p 69Ȃ93

Compagno LJV, Cook SF (1995) The exploitation and conservation of freshwater

elasmobranchs: Status of taxa and prospects for the future. J Aquaric Aquat

Sci 7:62Ȃ90

Compagno LJV, Dando M, Fowler SL (2005) Sharks of the World. Princeton

University Press, Hong Kong

Compagno LJV, Pogonoski JJ, Pollard DA (2009) Glyphis glyphis. IUCN Red List

Threat Species 2009 eT39379A10221801

Compagno LJV, White WT, Last PR (2008) Glyphis garricki sp. nov., a new species

of river shark (Carcharhiniformes: Carcharhinidae) from northern Australia

and Papua New Guinea, with a redescription of Glyphis glyphis. Descr New

Aust Chondrichthyans: 213Ȃ225

Cortés E (1997) A critical review of methods of studying fish feeding based on

analysis of stomach contents: application to elasmobranch fishes. Can J Fish

Aquat Sci 54:726Ȃ738

Couturier LIE, Rohner CA, Richardson AJ, Pierce SJ, Marshall AD, Jaine FRA,

Townsend KA, Bennett MB, Weeks SJ, Nichols PD (2013b) Unusually high

levels of n-6 polyunsaturated fatty acids in whale sharks and reef manta

rays. Lipids 48:1029Ȃ1034

156 Couturier LIE, Rohner CA, Richardson AJ, Marshall AD, Jaine FRA, Bennett MB,

Townsend KA, Weeks SJ, Nichols PD (2013a) Stable isotope and signature

fatty acid analyses suggest reef manta rays feed on demersal zooplankton.

PLoS One 8:e77152

Cortés E, (1999) Standardized diet compositions and trophic levels of sharks.

ICES J Mar Sci 56:707Ȃ717.

Crook DA, Buckle DJ, Allsop Q, Baldwin W, Saunders TM, Kyne PM, Woodhead JD,

Maas R, Roberts B, Douglas MM (2016) Use of otolith chemistry and acoustic

telemetry to elucidate migratory contingents in barramundi Lates calcarifer.

Mar Freshw Res 68:1554Ȃ1566

Curtin R, Prellezo R (2010) Understanding marine ecosystem based

management: A literature review. Mar Policy 34:821Ȃ830

Cyrus DP, Blaber SJM (1992) Turbidity and salinity in a tropical northern

Australian estuary and their influence on fish distribution. Estuar Coast

Shelf Sci 35:545Ȃ563

Dalsgaard J, John MA St., Kattner G, Müller-Navarra DC, Hagen W (2003) Fatty

acid trophic markers in the pelagic marine environment. Adv Mar Biol

46:225Ȃ340

Daly R, Froneman PW, Smale MJ (2013) Comparative feeding ecology of Bull

Sharks (Carcharhinus leucas) in the coastal waters of the southwest Indian

Ocean inferred from stable isotope analysis. PLoS One 8:1Ȃ11

Dantas DV, Barletta M, Ramos J de AA, Lima ARA, Costa MF da (2012) Seasonal

diet shifts and overlap between two sympatric catfishes in an estuarine

nursery. Estuaries and Coasts 36:237Ȃ256

Davidson B, Cliff G (2014) Comparison of pinniped and cetacean prey tissue

157 lipids with lipids of their elasmobranch predator. In Vivo 28:223Ȃ8

Davis TLO (1985) The food of barramundi, Lates calcarifer (Bloch), in coastal and

inland waters of Van Diemen Gulf and the Gulf of Carpentaria, Australia. J

Fish Biol 26:669Ȃ682

DeNiro MJ, Epstein S (1978) Influence of diet on the distribution of carbon

isotopes in animals. Geochim Cosmochim Acta 42:495Ȃ506

Douglas MM, Bunn SE, Davies PM (2005) River and wetland food webs in

—•–”ƒŽ‹ƒǯ•™‡–Ȃdry tropics: general principles and implications for

management. Mar Freshw Res 56:329Ȃ342

Dufrêne M, Legendre P (1997) Species assemblages and indicator species: the

need for a flexible asymmetrical approach. Ecol Monogr 67:345Ȃ366

Dulvy NK, Fowler SL, Musick JA, Cavanagh RD, Kyne PM, Harrison LR, Carlson JK,

Davidson LNK, Fordham S V, Francis MP, Pollock CM, Simpfendorfer CA,

Burgess GH, Carpenter KE, Compagno LJ V, Ebert DA, Gibson C, Heupel MR,

Livingstone SR, Sanciangco JC, Stevens JD, Valenti S, White WT (2014)

š–‹ –‹‘”‹•ƒ† ‘•‡”˜ƒ–‹‘‘ˆ–Š‡™‘”Ž†ǯ••Šƒ”•ƒ†”ƒ›•ǤŽ‹ˆ‡͵ǣͳȂ34

Dunne JA (2012) Food Webs. In: Meyers RA (ed) Computational Complexity:

Theory, Techniques, and Applications. Springer New York, New York, NY, p

1155Ȃ1176

Elton CS (1927) Animal ecology. University of Chicago Press

Estes JA (1998) Killer whale predation on sea otters linking oceanic and

nearshore Ecosystems. Science 282:473Ȃ476

Estes JA, Terborgh J, Brashares JS, Power ME, Berger J, Bond WJ, Carpenter SR,

Essington TE, Holt RD, Jackson JBC, Marquis RJ, Oksanen L, Oksanen T

(2011) Trophic downgrading of planet Earth. Science 333:301Ȃ6

158 Estrada JA, Rice AN, Lutcavage ME, Skomal GB (2003) Predicting trophic position

in sharks of the north-west Atlantic Ocean using stable isotope analysis. J

Mar Biol Assoc UK 83:1347Ȃ1350

Estrada JA, Rice AN, Natanson LJ, Skomal GB (2006) Use of isotopic analysis of

vertebrae in reconstructing ontogenetic feeding ecology in white sharks.

Ecology 87:829Ȃ34

Every SL, Pethybridge HR, Crook DA, Kyne PM, Fulton CJ (2016) Comparison of

fin and muscle tissues for analysis of signature fatty acids in tropical

euryhaline sharks. J Exp Mar Bio Ecol 479:46Ȃ53

Every SL, Pethybridge HR, Fulton CJ, Kyne PM, Crook DA (2017) Niche metrics

suggest euryhaline and coastal elasmobranchs provide trophic connections

among marine and freshwater biomes in northern Australia. Mar Ecol Prog

Ser 565:181Ȃ196

Feutry P, Berry O, Kyne PM, Pillans RD, Hillary RM, Grewe PM, Marthick JR,

Johnson G, Gunasekera RM, Bax NJ, Bravington M (2017) Inferring

contemporary and historical genetic connectivity from juveniles. Mol Ecol

26:444Ȃ456

Feutry P, Kyne PM, Pillans RD, Chen X, Naylor GJ, Grewe PM (2014)

Mitogenomics of the Speartooth Shark challenges ten years of control region

sequencing. BMC Evol Biol 14:232

Fisk, AT, Tittlemier, SA, Pranschke, JL, Norstrom, RJ, (2002) Using anthropogenic

contaminants and stable isotopes to assess the feeding ecology of Greenland

sharks. Ecology 83:2162-2172

159 Frisch, A.J., Ireland, M., Rizzari, J.R., Lönnstedt, O.M., Magnenat, K.A., Mirbach, C.E.

and Hobbs, J.P.A., 2016. Reassessing the trophic role of reef sharks as apex

predators on coral reefs. Coral Reefs, 35:459Ȃ472.

Gallagher AJ, Kyne PM, Hammerschlag N (2012) Ecological risk assessment and

its application to elasmobranch conservation and management. J Fish Biol

80:1727Ȃ48

Geesey GG, Mutch R, Costerton JW, Green RB (1978) Sessile bacteria: An

important component of the microbial population in small mountain

streams. Limnol Oceanogr 23:1214Ȃ1223

George EM, Parrish CC (2015) Invertebrate uptake of lipids in the vicinity of

Atlantic salmon (Salmo salar) aquaculture sites in British Columbia. Aquac

Res 46:1044Ȃ1065

Grubbs RD, Carlson JK, Romine JG, Curtis TH, McElroy WD, McCandless CT,

Cotton CF, Musick JA (2016) Critical assessment and ramifications of a

purported marine trophic cascade. Sci Rep 6:20970

ƒ›‡ǡ‹–‘†‡‘”ƒ‹•ǡƒˆˆ”ƒ› ǡƒ†‹‘ǡŠ‹ƒ™ǡ‘ ǯŠ ‡ȋʹͲͳͳȌ

Structure and seasonal variability of fish food webs in an estuarine tropical

marine protected area (Senegal): Evidence from stable isotope analysis.

Estuar Coast Shelf Sci 92:607Ȃ617

Falk-Petersen S, Dahl, TM, Scott, CL, Sargent, JR, Gulliksen B, Kwasniewski S, Hop,

H, Millar, RM, 2002. Lipid biomarkers and trophic linkages between

ctenophores and copepods in Svalbard waters. Mar. Ecol. Prog. Ser. 227,

187Ȃ194. doi:10.3354/meps227187

Fisk AT, Tittlemier SA, Pranschke JL, Norstrom RJ (2002) Using anthropogenic

contaminants and stable isotopes to assess the feeding ecology of Greenland

160 Sharks. Ecology 83:2162Ȃ2172

Froese R, Pauly D (2015) FishBase. www.fishbase.org (accessed 15 Oct 2015)

Fry B (2006) Stable Isotope Ecology, 3rd edn. Springer Science + Buisness Media

LLC, New York, NY

Fry B, Mumford PL, Tam F, Fox DD, Warren GL, Havens KE, Steinman AD (1999)

Trophic position and individual feeding histories of fish from Lake

Okeechobee, Florida. Can J Fish Aquat Sci 56:590Ȃ600

Graeve M, Kattner G, Hagen W (1994) Diet-induced changes in the fatty acid

composition of Arctic herbivorous copepods: Experimental evidence of

trophic markers. J Exp Mar Bio Ecol 182:97Ȃ110

Hairston NG, Smith FE, Slobodkin LB (1960) Community structure, population

control and competition. Am Nat 94:421Ȃ425

Hall D, Lee SY, Meziane T (2006) Fatty acids as trophic tracers in an

experimental estuarine food chain: Tracer transfer. J Exp Mar Bio Ecol

336:42Ȃ53

Hammerschlag-Peyer CM, Yeager LA, Araújo MS, Layman CA (2011) A

hypothesis-testing framework for studies investigating ontogenetic niche

shifts using stable isotope ratios. PLoS One 6:e27104

Hammerschlag N, Ovando D, Serafy J (2010) Seasonal diet and feeding habits of

juvenile fishes foraging along a subtropical marine ecotone. Aquat Biol

Hammerschlag N, Sulikowski J (2011) Killing for conservation: the need for

alternatives to lethal sampling of apex predatory sharks. Endanger Species

Res 14:135Ȃ140

Heithaus MR, Frid A, Wirsing AJ, Worm B (2008) Predicting ecological

consequences of marine top predator declines. Trends Ecol Evol 23:202Ȃ

161 210

Heithaus MR, Vaudo JJ, Kreicker S, Layman CA, Krützen M, Burkholder DA,

Gastrich K, Bessey C, Sarabia R, Cameron K, Wirsing AJ, Thomson JA,

Dunphy-Daly MM (2013) Apparent resource partitioning and trophic

structure of large-bodied marine predators in a relatively pristine seagrass

ecosystem. Mar Ecol Prog Ser 481:225Ȃ237

Heupel MR, Carlson JK, Simpfendorfer CA (2007) Shark nursery areas: concepts,

definition, characterization and assumptions. Mar Ecol Prog Ser 337:287Ȃ

297

Heupel MR, Hueter RE (2002) Importance of prey density in relation to the

movement patterns of juvenile blacktip sharks (Carcharhinus limbatus)

within a coastal nursery area. In: Marine and Freshwater Research. CSIRO

Publishing, p 543Ȃ550

Heupel MR, Knip DM, Simpfendorfer CA, Dulvy NK (2014) Sizing up the

ecological role of sharks as predators. Mar Ecol Prog Ser 495:291Ȃ298

Heupel MR, Simpfendorfer CA (2010) Science or slaughter: need for lethal

sampling of sharks. Conserv Biol 24:1212Ȃ1218

Heupel MR, Simpfendorfer CA (2008) Movement and distribution of young bull

sharks, Carcharhinus leucas in a variable estuarine environment. Aquat Biol

1:277-289

Heupel MR, Yeiser BG, Collins AB, Ortega L, Simpfendorfer CA (2010) Long-term

presence and movement patterns of juvenile bull sharks, Carcharhinus

leucas, in an estuarine river system. Mar Freshw Res 61:1Ȃ10

Hirzel AH, Lay G Le (2008) Habitat suitability modelling and niche theory. J Appl

Ecol 45:1372Ȃ1381

162 Holt RD (2009) Bringing the Hutchinsonian niche into the 21st century:

Ecological and evolutionary perspectives. Proc Natl Acad Sci U S A

106:19659Ȃ19665

Hooper DU, Chapin III FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, Lawton JH,

Lodge DM, Loreau M, Naeem S, Schmid B, Setälä H, Symstad AJ, Vandermeer

J, Wardle DA (2005) Effects of biodiversity on ecosystem functioning: A

consensus of current knowledge. Ecol Monogr 75:3Ȃ35

Hughes TP (1994) Catastrophes, phase shifts, and large-scale degradation of a

Caribbean coral reef. Science 265:1547Ȃ1551

Hussey NE, Brush J, McCarthy ID, Fisk, AT, (2010) Ɂ15ƒ†Ɂ13C diet-tissue

discrimination factors for large sharks under semi-controlled conditions.

Comp Biochem Physiol A Mol Integr Physiol 155:445Ȃ453.

Hussey NE, Chapman DD, Donnelly E, Abercrombie DL, Fisk AT (2011a) Fin-icky

samples: an assessment of shark fin as a source material for stable isotope

analysis. Limnol Oceanogr Methods 9:524Ȃ532

Hussey NE, Dudley SFJ, Mccarthy ID, Cliff G, Fisk AT (2011b) Stable isotope

profiles of large marine predators: viable indicators of trophic position, diet,

and movement in sharks? Can J Fish Aquat Sci 68:2029Ȃ2045

Hussey NE, MacNeil MA, Fisk AT (2010) The requirement for accurate diet-tissue

discrimination factors for interpreting stable isotopes in sharks.

Hydrobiologia 654:1Ȃ5

Hussey NE, MacNeil MA, McMeans BC, Olin JA, Dudley SFJ, Cliff G, Wintner SP,

Fennessy ST, Fisk AT (2014a) Rescaling the trophic structure of marine food

webs. Ecol Lett 17:239Ȃ250

Hussey NE, MacNeil, A M, McMeans, BC, Olin, JA, Dudley, SF, Cliff, G, Wintner, SP,

163 Fennessy, ST. and Fisk, AT, (2014b) Corrigendum to Hussey et al.(). Ecology

letters, 17:768-768.

Hussey NE, MacNeil MA, Olin JA, McMeans BC, Kinney MJ, Chapman DD, Fisk AT

(2012) Stable isotopes and elasmobranchs: tissue types, methods,

applications and assumptions. J Fish Biol 80:1449Ȃ1484

Hussey NE, MacNeil MA, Siple MC, Popp BN, Dudley SFJ, Fisk AT (2015)

Expanded trophic complexity among large sharks. Food Webs 4:1Ȃ7

Hutchinson GE (1957) Concluding remarks. Cold Spring Harb Symp Quant Biol

22:415Ȃ427

Hyslop EJ (1980) Stomach contents analysis: a review of methods and their

application. J Fish Biol 17:411Ȃ429

Ings TC, Montoya JM, Bascompte J, Blüthgen N, Brown L, Dormann CF, Edwards

F, Figueroa D, Jacob U, Jones JI, Lauridsen RB, Ledger ME, Lewis HM, Olesen

JM, Veen FJF Van, Warren PH, Woodward G (2009) Ecological networks -

Beyond food webs. J Anim Ecol 78:253Ȃ269

Iverson SJ (2009) Tracing aquatic food webs using fatty acids: from qualitative

indicators to quantitative determination lipids in aquatic ecosystems. In:

Kainz M, Brett MT, Arts MT (eds) Lipids in Aquatic Ecosystems. Springer

New York, New York, NY, p 281Ȃ307

Iverson SJ, Bowen WD, Boness DJ, Oftedal OT (1993) The effects of maternal size

and milk anergy output on pup growth in grey seals (Halichoerus grypus).

Physiol Zool 66:61Ȃ88

Iverson SJ, Field C, Don Bowen W, Blanchard W, Don Bowen W, Blanchard W

(2004) Quantitative fatty acid signature analysis: a new method of

estimating predator diets. Ecological 74:211Ȃ235

164 Jackson AL, Inger R, Parnell AC, Bearhop S (2011) Comparing isotopic niche

widths among and within communities: SIBER Ȃ Stable Isotope Bayesian

Ellipses in R. J Anim Ecol 80:595Ȃ602

Jardine TD, Bond NR, Burford MA, Kennard MJ, Ward DP, Bayliss P, Davies PM,

Douglas MM, Hamilton SK, Melack JM, Naiman RJ, Pettit NE, Pusey BJ, Warfe

DM, Bunn SE (2015) Does flood rhythm drive ecosystem responses in

tropical riverscapes? Ecology 96:684Ȃ692

Jardine TD, Hunt RJ, Pusey BJ, Bunn SE (2011) A non-lethal sampling method for

stable carbon and nitrogen isotope studies of tropical fishes. Mar Freshw

Res 62:83Ȃ90

Jardine TD, Pettit NE, Warfe DM, Pusey BJ, Ward DP, Douglas MM, Davies PM,

Bunn SE (2012a) Consumer-resource coupling in wet-dry tropical rivers. J

Anim Ecol 81:310Ȃ322

Jardine TD, Pusey BJ, Hamilton SK, Pettit NE, Davies PM, Douglas MM, Sinnamon

V, Halliday IA, Bunn SE (2012b) Fish mediate high food web connectivity in

the lower reaches of a tropical floodplain river. Oecologia 168:829Ȃ838

Jardine TD, Rayner TS, Pettit NE, Valdez D, Ward DP, Lindner G, Douglas MM,

Bunn SE (2016) Body size drives allochthony in food webs of tropical rivers.

Oecologia:1Ȃ13

Jepsen DB, Winemiller KO (2002) Structure of tropical river food webs revealed

by stable isotope ratios. Oikos 96:46Ȃ55

Joseph MM, Renjith, KR, Ratheesh Kumar CS, Chandramohanakumar N, (2012)

Assessment of organic matter sources in the tropical mangrove ecosystems

of cochin, southwest India. Environ. Forensics 13:262Ȃ271.

doi:10.1080/15275922.2012.676600

165 Junk WJ, Bayley PB, Sparks RE (1989) The flood pulse concept in river-floodplain

systems. Can Spec Publ Fish Aquat Sci 106:110Ȃ127

Käkelä R, Hyvärinen H, Vainiotalo P (1993) Fatty acid composition in liver and

blubber of the saimaa ringed seal (Phoca hispida saimensis) compared with

that of the ringed seal (Phoca hispida botnica) and grey seal (Halichoerus

grypus) from the Baltic. Comp Biochem Physiol -- Part B Biochem 105:553Ȃ

565

Kelly J, Scheibling R (2012) Fatty acids as dietary tracers in benthic food webs.

Mar Ecol Prog Ser 446:1Ȃ22

Kerr LA, Andrews A, Cailliet GM, Brown TA, Coale KH (2006) Investigations of

Delta(14)C, delta(13)C, and delta(15)N in vertebrae of white shark

(Carcharodon carcharias) from the eastern North Pacific Ocean. Environ Biol

Fishes 77:337Ȃ353

Kim SL, Koch PL (2012) Methods to collect, preserve, and prepare elasmobranch

tissues for stable isotope analysis. Environ Biol Fishes 95:53Ȃ63

Kirsch PE, Iverson SJ, Bowen WD (1998) Dietary effects on the fatty acid

signature of whole Atlantic cod (Gadus morhua). Can J Fish Aquat Sci

55:1378Ȃ1386

Kitchell JF, Essington TE, Boggs CH, Schindler DE, Walters CJ (2002) The role of

sharks and longline fisheries in a pelagic ecosystem of the Central Pacific.

Ecosystems 5:202Ȃ216

Knip DM, Heupel MR, Simpfendorfer CA, Tobin AJ (2011a) Ontogenetic shifts in

movement and habitat use of juvenile pigeye sharks Carcharhinus

amboinensis in a tropical nearshore region. Mar Ecol Prog Ser 425:233Ȃ246

Knip DM, Heupel MR, Simpfendorfer CA, Tobin AJ, Moloney J (2011b) Wet-

166 season effects on the distribution of juvenile pigeye sharks, Carcharhinus

amboinensis, in tropical nearshore waters. Mar Freshw Res 62:658Ȃ667

Krebs CJ (2009) Ecology: the experimental analysis of distribution and

abundance (B Wilbur, Ed.), 6th edn. Benjamin Cummings, San Francisco

Kyne PM (2016) Sharks and rays of northern Australian rivers.

Kyne PM, Carlson JK, Smith K (2013) Pristis pristis. The IUCN Red List of

Threatened Species 2013: e.T18584848A18620395.

Last PR (2002) Freshwater and estuarine elasmobranchs of Australia. In: Fowler

SL, Reed TM, Dipper FA (eds) Elasmobranch biodiversity, conservation and

management. Proceedings of the International Seminar and Workshop.p

185Ȃ193

Last PR, Manjaji-matsumoto BM (2006) Himantura dalyensis sp. nov, a new

estuarine whipray (Myliobatoidei: Dasyatidae) from northern Australia.

Last PR, Stevens JD (2009) Sharks and Rays of Australia, Second. CSIRO

Publishing, Collingwood

Layman CA, Araujo MS, Boucek R, Hammerschlag-Peyer CM, Harrison E, Jud ZR,

Matich P, Rosenblatt AE, Vaudo JJ, Yeager LA, Post DM, Bearhop S (2012a)

Applying stable isotopes to examine food-web structure: an overview of

analytical tools. Biol Rev Camb Philos Soc 87:545Ȃ562

Layman CA, Allgeier JE (2012b) Characterizing trophic ecology of generalist

consumers: a case study of the invasive lionfish in The Bahamas. Mar Ecol

Prog Ser 448:131Ȃ141

Layman CA, Arrington DA, Montaña CG, Post DM (2007a) Can stable isotope

ratios provide for community-wide measures of trophic structure? Ecology

88:42Ȃ48

167 Layman CA, Quattrochi JP, Peyer CM, Allgeier JE (2007b) Niche width collapse in

a resilient top predator following ecosystem fragmentation. Ecol Lett

10:937Ȃ944

Le HD, Meisel JA, Meijer VE de, Gura KM, Puder M (2009) The essentiality of

arachidonic acid and docosahexaenoic acid. Prostaglandins Leukot Essent

Fat Acids 81:165Ȃ170

Lee RF, Nevenzel JC, Paffenhöfer GA (1971) Importance of wax esters and other

lipids in the marine food chain: Phytoplankton and copepods. Mar Biol

9:99Ȃ108

Lewallen EA, Anderson TW, Bohonak, AJ, (2007) Genetic structure of leopard

shark (Triakis semifasciata) populations in California waters. Mar. Biol. 152,

599Ȃ609.

Letnic M, Ritchie EG, Dickman CR (2012) Top predators as biodiversity

regulators: the dingo Canis lupus dingo as a case study. Biol Rev Camb Philos

Soc 87:390Ȃ413

Libralato S, Christensen V, Pauly D (2006) A method for identifying keystone

species in food web models. Ecol Modell 195:153Ȃ171

Lindeman RL (1942) The trophic-dynamic aspect of ecology. Ecology 23:399Ȃ

417

Logan JM, Jardine TD, Miller TJ, Bunn SE, Cunjak RA, Lutcavage ME (2008) Lipid

corrections in carbon and nitrogen stable isotope analyses: comparison of

chemical extraction and modelling methods. J Anim Ecol 77:838Ȃ846

Logan JM, Lutcavage ME (2010a) Stable isotope dynamics in elasmobranch

fishes. Hydrobiologia 644:231Ȃ244

Logan JM, Lutcavage ME (2010b) Reply to Hussey et al.: The requirement for

168 accurate diet-tissue discrimination factors for interpreting stable isotopes in

sharks. Hydrobiologia 654:7Ȃ12

Loneragan NR, Bunn SE, Kellaway DM (1997) Are mangrove and seagrasses

sources of organic carbon for penaeid prawns in a tropical estuary? A

multiple isotope study. Mar Biol 130:289Ȃ300

Loreau M (2009) Communities and ecosystems. In: The Princeton guide to

ecology.p 230

Lucifora LO, Carvalho MR de, Kyne PM, White WT (2015) Freshwater sharks and

rays. Curr Biol 25:R971ȂR973

Lynam CP, Llope M, Möllmann C, Helaouët P, Bayliss-Brown GA, Stenseth NC

(2017) Interaction between top-down and bottom-up control in marine

food webs. Proc Natl Acad Sci 114:201621037

MacNeil, MA, Drouillard, KG, Fisk, AT. 2006. Variable uptake and elimination of

stable nitrogen isotopes between tissues in fish. Can. J. Fish. Aquat. Sci. 63,

345Ȃ353

MacNeil MA, Skomal GB, Fisk AT (2005) Stable isotopes from multiple tissues

reveal diet switching in sharks. Mar Ecol Prog Ser 302:199Ȃ206

McElroy, W.D., Wetherbee, B.M., Mostello, C.S., Lowe, C.G., Crow, G.L. and Wass,

R.C., 2006. Food habits and ontogenetic changes in the diet of the sandbar

shark, Carcharhinus plumbeus, in Hawaii. Env. Biol. Fish76:81-92.

Magnone L, Bessonart M, Gadea J, Salhi M (2015) Trophic relationships in an

estuarine environment: A quantitative fatty acid analysis signature

approach. Estuar Coast Shelf Sci 166:24Ȃ33

Malpica-cruz L, Herzka SZ, Sosa-nishizaki O, Lazo JP (2012) Tissue-specific

isotope trophic discrimination factors and turnover rates in a marine

169 elasmobranch: empirical and modeling results. Can J Fish Aquat Sci 69:551Ȃ

564

Martínez-Álvareza RM, Sanza A, García-Gallegoa M, Domezain A, Domezain J,

Carmonac R, M. del Valle, Ostos-Garridoc AEM (2005) Adaptive branchial

mechanisms in the sturgeon Acipenser naccarii during acclimation to

saltwater. Comp Biochem Physiol Biochem Physiol Part A 141:183Ȃ190

Matich P, Heithaus MR (2015) Individual variation in ontogenetic niche shifts in

habitat use and movement patterns of a large estuarine predator

(Carcharhinus leucas). Oecologia 178:347Ȃ359

Matich P, Heithaus MR (2014) Multi-tissue stable isotope analysis and acoustic

telemetry reveal seasonal variability in the trophic interactions of juvenile

bull sharks in a coastal estuary. J Anim Ecol 83:199Ȃ213

Matich P, Kiszka JJ, Heithaus MR, Mourier J, Planes S (2015) Short-term shifts of

stable isotope (G13C, G15N) values in juvenile sharks within nursery areas

suggest rapid shifts in energy pathways. J Exp Mar Bio Ecol 465:83Ȃ91

Matich P, Heithaus MR, Layman CA (2010) Size-based variation in intertissue

comparisons of stable carbon and nitrogen isotopic signatures of bull sharks

(Carcharhinus leucas) and tiger sharks (Galeocerdo cuvier). Can J Fish Aquat

Sci 67:877Ȃ885

Matich P, Heithaus MR (2015) Individual variation in ontogenetic niche shifts in

habitat use and movement patterns of a large estuarine predator

(Carcharhinus leucas). Oecologia 178:347Ȃ359

Matich P, Heithaus MR, Layman CA (2011) Contrasting patterns of individual

specialization and trophic coupling in two marine apex predators. J Anim

Ecol 80:294Ȃ305

170 Matley JK, Fisk AT, Tobin AJ, Heupel MR, Simpfendorfer CA (2016) Diet-tissue

discrimination factors and turnover of carbon and nitrogen stable isotopes

in tissues of an adult predatory coral reef fish, Plectropomus leopardus.

Rapid Commun Mass Spectrom 30:29Ȃ44

Matthews B, Mazumder A (2004) A critical evaluation of intrapopulation

variation of delta13C and isotopic evidence of individual specialization.

Oecologia 140:361Ȃ371

McCutchan JHJ, Lewis WMJ, Kendall C, McGrath CC (2003) Variation in trophic

shift for stable isotope ratios of carbon, nitrogen, and sulfur. Oikos 102:378Ȃ

390

McMahon KW, Fogel ML, Johnson BJ, Houghton LA, Thorrold SR (2011) A new

method to reconstruct fish diet and movement patterns from delta C-13

values in otolith amino acids. Can J Fish Aquat Sci 68:1330Ȃ1340

McInerny GJ, Etienne RS (2012) Stitch the niche - a practical philosophy and

visual schematic for the niche concept (S Higgins, Ed.). J Biogeogr 39:2103Ȃ

2111

McMeans BC, Arts MT, Lydersen C, Kovacs KM, Hop H, Falk-Petersen S, Fisk AT

(2013) The role of Greenland sharks (Somniosus microcephalus) in an Arctic

ecosystem: assessed via stable isotopes and fatty acids. Mar Biol 160:1223Ȃ

1238

Miller MR, Nichols PD, Barnes J, Davies NW, Peacock EJ, Carter CG (2006)

Regiospecificity profiles of storage and membrane lipids from the gill and

muscle tissue of Atlantic salmon (Salmo salar L.) grown at elevated

temperature. Lipids 41:865Ȃ876

Mitchell AL, Lucas RM, Donnelly BE, Pfitzner K, Milne AK, Finlayson M (2007) A

171 new map of mangroves for Kakadu National Park, Northern Australia, based

on stereo aerial photography. Aquat Conserv Mar Freshw Ecosyst 17:446Ȃ

467

Montoya JM, Pimm SL, Solé R V (2006) Ecological networks and their fragility.

Nature 442:259Ȃ264

Moore JW, Semmens BX (2008) Incorporating uncertainty and prior information

into stable isotope mixing models. Ecol Lett 11:470Ȃ80

Morgan DL, Somaweera R, Gleiss AC, Beatty SJ, Whitty JM (2017) An upstream

migration fought with danger: freshwater sawfish fending off sharks and

crocodiles. Ecology 98:1465Ȃ1467

Morgan DL, Whitty JM, Phillips NM, Thorburn DC, Chaplin JA, R M (2011) North-

western Australia as a hotspot for endangered elasmobranchs with

particular reference to sawfishes and the Northern River Shark. J R Soc West

Aust 94:345Ȃ358

Munroe SEM, Heupel MR, Fisk AT, Logan JM, Simpfendorfer CA (2015) Regional

movement patterns of a small-bodied shark revealed by stable-isotope

analysis. J Fish Biol 86:1567Ȃ1586

Munroe SEM, Heupel MR, Fisk AT, Simpfendorfer CA (2014a) Geographic and

temporal variation in the trophic ecology of a small-bodied shark: evidence

of resilience to environmental change. Can J 72:343Ȃ351

Munroe SEM, Simpfendorfer CA, Heupel MR (2014b) Habitat and space use of an

abundant nearshore shark, Rhizoprionodon taylori. Mar Freshw Res 65:959Ȃ

968

Myers RA, Worm B (2003) Rapid worldwide depletion of predatory fish

communities. Nature 423:280-283

172 Napolitano GE, Pollero RJ, Gayoso AM, MacDonald BA, Thompson RJ (1997) Fatty

acids as trophic markers of phytoplankton blooms in the Bahia Blanca

estuary (Buenos Aires, Argentina) and in Trinity Bay (Newfoundland,

Canada). Biochem Syst Ecol 25:739Ȃ755

Navarro J, Coll M, Louzao M, Palomera I, Delgado A, Forero MG (2011)

Comparison of ecosystem modelling and isotopic approach as ecological

tools to investigate food webs in the NW Mediterranean Sea. J Exp Mar Bio

Ecol 401:97Ȃ104

Newsomes SD, Martinez C, Bearhop S, Phillips DL, Martinez del Rio C (2007) A

niche for isotopic ecology. Front Ecol Environ 5:429Ȃ436

Nichols PD, Rayner M, Stevens JD (2001) A pilot investigation of northern

Australian shark liver oils: characterization and value-adding a pilot

investigation of northern Australian shark liver oils: characterization and

value-adding. CSIRO Marine Research.

Novak PA, Bayliss P, Crook DA, Garcia EA, Pusey BJ, Douglas MM (2017) Do

upstream migrating, juvenile amphidromous shrimps, provide a marine

subsidy to river ecosystems? Freshw Bio l65:1Ȃ14

Novak PA, Douglas MM, Garcia EA, Bayliss P, Pusey BJ (2015) A life-history

account of Macrobrachium spinipes (Schenkel, 1902) (Cherabin) in a large

tropical Australian river. Freshw Sci 34:620Ȃ633

Økland HMW, Stoknes IS, Remme JF, Kjerstad M, Synnes M (2005) Proximate

composition, fatty acid and lipid class composition of the muscle from deep-

sea teleosts and elasmobranchs. Comp Biochem Physiol B Biochem Mol Biol

140:437Ȃ443

Olin JA, Hussey NE, Fritts M, Heupel MR, Simpfendorfer CA, Poulakis GR, Fisk AT

173 (2011) Maternal meddling in neonatal sharks: implications for interpreting

stable isotopes in young animals. Rapid Commun Mass Spectrom 25:1008Ȃ

1016

Olin JA, Poulakis GR, Stevens PW, DeAngelo JA, Fisk,AT, (2014) Preservation

effects on stable isotope values of archived elasmobranch fin tissue:

comparisons between frozen and ethanol-stored samples. Trans Am Fish

Soc 143, 1569Ȃ1576.

Özogul Y, Özogul F, Alagoz S (2007) Fatty acid profiles and fat contents of

commercially important seawater and freshwater fish species of Turkey: A

comparative study. Food Chem 103:217Ȃ223

Paine RT (1966) Food Web Complexity and Species Diversity. Am Nat 100:65Ȃ75

Paine RT (1969) A Note on Trophic Complexity and Community Stability. Am Nat

103:91Ȃ93

Papastamatiou YP, Wetherbee BM, Lowe CG, Crow LG (2006) Distribution and

diet of four species of carcharhinid shark in the Hawaiian Islands: evidence

for resource partitioning and competitive exclusion. Mar Ecol Prog Ser

320:239Ȃ251

Parnell AC, Inger R, Bearhop S, Jackson AL (2010) Source partitioning using

stable isotopes: Coping with too much variation. PLoS One 5:1Ȃ5

Parnell AC, Phillips DL, Bearhop S, Semmens BX, Ward EJ, Moore JW, Jackson AL,

Grey J, Kelly DJ, Inger R (2013) Bayesian stable isotope mixing models.

Environmetrics 24:387Ȃ399

Parrish CC (2013) Lipids in marine ecosystems. ISRN Oceanogr 2013:1Ȃ16

Parrish CC, Nichols PD, Pethybridge HR, Young JW (2015) Direct determination

of fatty acids in fish tissues: quantifying top predator trophic connections.

174 Oecologia 177:85Ȃ95

Parrish CC, Pethybridge HR, Young JW, Nichols, PD, (2013). Spatial variation in

fatty acid trophic markers in albacore tuna from the southwestern Pacific

OceanȄ’‘–‡–‹ƒŽDz–”‘’‹ ƒŽ‹œƒ–‹‘dz•‹‰ƒŽǤ‡‡’‡ƒ‡•Ǥƒ”– Top. Stud.

Oceanogr. 1Ȃ9.

Pauly D, Christensen V, Dalsgaard J, Froese, R, Torres, F, 1998. Fishing down

marine food webs. Science 27:860Ȃ863

Peterson BJ, Fry B (1987) Stable isotopes in ecosystem studies. Annu Rev Ecol

Syst 18:293Ȃ320

Pethybridge HR, Butler EC V, Cossa D, Daley RK, Boudou A (2012) Trophic

structure and biomagnification of mercury in an assemblage of deepwater

chondrichthyans from southeastern Australia. Mar Ecol Prog Ser 451:163Ȃ

174

Pethybridge HR, Daley RK, Nichols PD (2011) Diet of demersal sharks and

chimaeras inferred by fatty acid profiles and stomach content analysis. J Exp

Mar Bio Ecol 409:290Ȃ299

Pethybridge HR, Daley RK, Virtue P, Nichols PD (2010) Lipid composition and

partitioning of deepwater chondrichthyans: inferences of feeding ecology

and distribution. Mar Biol 157:1367Ȃ1384

Pethybridge HR, Daley RK, Nichols PD, (2011a) Diet of demersal sharks and

chimaeras inferred by fatty acid profiles and stomach content analysis. J.

Exp. Mar. Bio. Ecol. 409:290Ȃ299.

Pethybridge HR, Daley RK, Virtue P, Nichols PD, (2011b) Lipid (energy) reserves,

utilisation and provisioning during oocyte maturation and early embryonic

development of deepwater chondrichthyans. Mar. Biol. 158:2741Ȃ2754.

175 Pethybridge HR, Parrish CC, Bruce BD, Young, JW, Nichols PD, (2014) Lipid, fatty

acid and energy density profiles of White Sharks: insights into the feeding

ecology and ecophysiology of a complex top predator. PLoS One 9, e97877.

Pethybridge HR, Parrish CC, Morrongiello J, Young JW, Farley JH, Gunasekera RM,

Nichols PD (2015) Spatial patterns and temperature prediction of tuna fatty

acids: tracing essential nutrients and changes in primary producers. PLoS

One 10:e0131598

Pettit NE, Naiman RJ, Warfe DM, Jardine TD, Douglas MM, Bunn SE, Davies PM

(2016a) Productivity and Connectivity in Tropical Riverscapes of Northern

Australia: Ecological Insights for Management. Ecosystems:1Ȃ23

Pettit NE, Ward DP, Adame MF, Valdez D, Bunn SE (2016b) Influence of aquatic

plant architecture on epiphyte biomass on a tropical river floodplain. Aquat

Bot 129:35Ȃ43

Peverell SC, Mcpherson GR, Garrett RN, Gribble NA (2006) New records of the

River Shark Glyphis (Carcharhinidae) reported from Cape York Peninsula,

northern Australia. Zootaxa 68:53Ȃ68

Phillips NM, Chaplin JA, Morgan DL, Peverell SC (2011) Population genetic

structure and genetic diversity of three critically endangered Pristis

sawfishes in Australian waters. Mar Biol 158:903Ȃ915

Phillips DL, Inger R, Bearhop S, Jackson AL, Jonathan W, Parnell AC, Semmens BX,

Ward EJ (2014) Best practices for use of stable isotope mixing models in

food web studies. Can J Zool 92:823Ȃ835

Pillans RD, Stevens JD, Kyne PM, Salini JP (2009) Observations on the

distribution, biology, short-term movements and habitat requirements of

river sharks Glyphis spp. in northern Australia. Endanger Species Res

176 10:321Ȃ332

Platell ME, Potter IC (2001) Partitioning of food resources amongst 18 abundant

benthic carnivorous fish species in marine waters on the lower west coast of

Australia. J Exp Mar Bio Ecol 261:31Ȃ54

Pogonoski JJ, Pollard DA (2003) Glyphis garricki. SSG Aust Ocean Reg Work

March 2003 SSG Austra

Polis GA (1994) Food webs, trophic cascades and community structure. Aust J

Ecol 19:121Ȃ136

Polis GA, Strong DR (1996) Food web complexity and community dynamics. Am

Nat 147:813Ȃ846

Polis GA, Winemiller KO (2013) Food webs: integration of patterns and

dynamics. Springer Sci Bus Media

Polo-Silva C, Galván-Magaña F, Delgado-Huertas A (2012) Trophic inferences of

blue shark (Prionace glauca) in the Mexican Pacific from stable isotope

analysis in teeth. Rapid Commun Mass Spectrom 26:1631Ȃ8

Post DM (2002) Using stable isotopes to estimate trophic position: models,

methods, and assumptions. Ecology 83:703Ȃ718

Post DM, Layman CA, Arrington DA, Takimoto G, Quattrochi J, Montaña CG

(2007) Getting to the fat of the matter: models, methods and assumptions

for dealing with lipids in stable isotope analyses. Oecologia 152:179Ȃ189

Power ME, Tilman D, Estes JA, Menge BA, Bond WJ, Mills S, Daily G, Castilla JC,

Lubchenco J, Paine RT, Mills LS (1996) Challenges Quest for Keystones. Am

Inst Biol Sci 46:609Ȃ620

”‹ ‡ȋͳͻͺ͹Ȍ‹‡–ƒ”‹ƒ–‹‘‹ƒ’‘’—Žƒ–‹‘‘ˆƒ”™‹ǯ• ‹ Š‡•Ǥ ‘Ž‘‰›

Pulliam HR (2000) On the relationship between niche and distribution. Ecol Lett

177 3:349Ȃ361

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

77:731Ȃ753

Pusey BJ, Kennard MJ, Larson HK, Alsop Q, Hammer M, Buckle D, (2016)

Estuarine fishes of the South Alligator River, Kakadu National Park,

northern Australia. Mar Freshw Res 67:1797Ȃ1812

R Core Development Team, (2014) R: A Language and Environment for

Statistical Computing

Ritchie EG, Johnson CN (2009) Predator interactions, mesopredator release and

biodiversity conservation. Ecol Lett 12:982Ȃ998

Roach KA, Winemiller KO, Layman CA, Zeug SC (2009) Consistent trophic

patterns among fishes in lagoon and channel habitats of a tropical floodplain

river: Evidence from stable isotopes. Acta Oecologica 35:513Ȃ522

‘„‡”–•ȋʹͲͳ͸Ȍƒ ƒ‰‡DzŽƒ„†•˜Ǥdz”†‹ƒ–‹‘—Ž–‹˜ƒ”

Rohner CA, Couturier LIE, Richardson AJ, Pierce SJ, Prebble CEM, Gibbons MJ,

Nichols PD (2013) Diet of whale sharks Rhincodon typus inferred from

stomach content and signature fatty acid analyses. Mar Ecol Prog Ser

493:219Ȃ235

Rossi S, Youngbluth, MJ, Jacoby, CA, Pages F, Garrofe X, (2008) Fatty acid trophic

markers and trophic links among seston, crustacean zooplankton and the

siphonophore Nanomia cara in Georges Basin and Oceanographer Canyon

(NW Atlantic). Sci Mar 72:403Ȃ416.

Roughgarden J (1972) Evolution of Niche Width. Am Nat 106:683Ȃ718

178 Rooney N, McCann K, Gellner G, Moore JC (2006) Structural asymmetry and the

stability of diverse food webs. Nature 442:265Ȃ269

Ross ST (1986) Resource partitioning in fish assemblages: a review of field

studies. Copeia 2:352Ȃ388

Ruppert JLW, Travers MJ, Smith LL, Fortin M-J, Meekan MG (2013) Caught in the

middle: combined impacts of shark removal and coral loss on the fish

communities of coral reefs. PLoS One 8:e74648

Rude, N.P., Trushenski, J.T. and Whitledge, G.W., 2016. Fatty acid profiles are

biomarkers of fish habitat use in a river- floodplain ecosystem.

Hydrobiologia, 773:63-75.

Sadowsky V (1971) Notes on the bull shark Carcharhinus leucas in the lagoon

region of Cananéia, Brazil. Bol do Inst Ocean 20:71Ȃ78

Sardenne F, Bodin N, Chassot E, Amiel A, Fouché E, Degroote M, Pethybridge HR

(2016) Trophic niches of sympatric tropical tuna in the Western Indian

Ocean inferred by stable isotopes and neutral fatty acids. Prog Oceanogr

146:75Ȃ88

Sargent JR, Bell JG, Bellio MG, Henderson RJ, Tocher DR (1995) Requirement

criteria for essential fatty acids. J Appl Ichthyol 11:183Ȃ198

Sargent JR, Bell G, McEvoy L, Tocher DR, Estevez A (1999a) Recent developments

in the essential fatty acid nutrition of fish. Aquaculture 177:191Ȃ199

Sargent JR, Henderson, JR, Tocher, DR, (1989) The Lipids, in: Halver, J.E. (Ed.),

Fish Nutrition. Academic Press, San Diego, p. 153Ȃ218.

Sargent JR, McEvoy L, Estevez A, Bell G, Bell M, Henderson JR, Tocher DR (1999b)

Lipid nutrition of marine fish during early development: current status and

future directions. Aquaculture 179:217Ȃ229

179 Schoener TW (1983) Field Experiments on Interspecific Competition. Am Nat

122:240Ȃ285

Scott AP, Middleton C (1979) Unicellular algae as a food for turbot (Scophthalmus

maximus L.) larvae - The importance of dietary long-chain polyunsaturated

fatty acids. Aquaculture 18:227Ȃ240

Shiffman DS, Gallagher AJ, Boyle MD, Hammerschlag-Peyer CM, Hammerschlag N

(2012) Stable isotope analysis as a tool for elasmobranch conservation

research: a primer for non-specialists. Mar Freshw Res 63:635

Simpfendorfer CA (1992) Reproductive Strategy of the Australian Sharpnose

Shark, Rhizoprionodon taylori (Elasmbanchii:Carcharhinidae), from

Cleveland Bay, Noethern Queensland. Aust J Mar Freshw Res 43:67Ȃ75

Simpfendorfer CA (1998) Diet of the Australian Sharpnose Shark,

Rhizoprionodon taylori, from northern Queensland. Mar Freshw Res 49:757Ȃ

761

Simpfendorfer CA, Burgess GH (2009) Carcharhinus leucas. IUCN Red List Threat

Species 2009 eT39372A10187195

‹ Žƒ‹” ǡǯ‡ƒK., Naughton, JM, (1986) Polyunsaturated fatty acid types in

Australian fish. Prog. Lipid Res. 25:81Ȃ82.

Skonberg DI, Rasco BA, Dong FM, (1994) Fatty acid composition of salmonid

muscle changes in response to a high oleic acid diet. J. Nutr. 124:1628Ȃ1638.

Smith BN, Epstein S (1971) Two categories of 13C/12C ratios for higher plants.

Plant Physiol 47:380Ȃ384

Snelson FFJ, Mulligan TJ, Williams SE (1984) Food habits, occurrence, and

population structure of the bull shark, Carcharhinus leucas, in Florida

coastal lagoons. Bull Mar Sci 34:71Ȃ80

180 Speed CW, Field IC, Meekan MG, Bradshaw CJA (2010) Complexities of coastal

shark movements and their implications for management. Mar Ecol Prog Ser

408:275Ȃ293

Speed CW, Meekan MG, Field IC, McMahon CR, Abrantes K, Bradshaw CJA, (2011)

Trophic ecology of reef sharks determined using stable isotopes and

telemetry. Coral Reefs 31:357Ȃ367.

Speers-Roesch B, Ip YK, Ballantyne JS (2008) Plasma non-esterified fatty acids of

elasmobranchs: comparisons of temperate and tropical species and effects

of environmental salinity. Comp Biochem Physiol A Mol Integr Physiol

149:209Ȃ16

St. John MA, Lund T, (1996) Lipid biomarkers: linking the utilization of frontal

plankton biomass to enhanced condition of juvenile North Sea cod. Mar.

Ecol. Prog. Ser. 131:75Ȃ85.

Stachowicz JJ, Bruno JF, Duffy JE (2007) Understanding the effects of marine

biodiversity on communities and ecosystems. Annu Rev Ecol Evol Syst

38:739Ȃ766

Stevens JD, Bonfil R, Dulvy NK, Walker PA (2000) The effects of fishing on sharks,

rays, and chimaeras (chondrichthyans), and the implications for marine

ecosystems. ICES J Mar Sci 57:476Ȃ494

Stevens JD, Pillans R, Salini J (2005) Conservation assessment of Glyphis sp. A

(speartooth shark), Glyphis sp. C (northern river shark), Pristis microdon

(freshwater sawfish) and Pristis zijsron (green sawfish). Australian

Government, Canberra.

Stoknes IS, Økland HMW, Falch E, Synnes M (2004) Fatty acid and lipid class

composition in eyes and brain from teleosts and elasmobranchs. Comp

181 Biochem Physiol - B Biochem Mol Biol 138:183Ȃ191

Swanson HK, Lysy M, Power M, Stasko AD, Johnson JD, Reist JD (2015) A new

probabilistic method for quantifying n-dimensional ecological niches and

niche overlap. Ecology 96:318Ȃ324

Syväranta J, Lensu A, Marjomäki TJ, Oksanen S, Jones RI (2013) An empirical

evaluation of the utility of convex hull and standard ellipse areas for

assessing population niche widths from stable isotope data. PLoS One

8:e56094

Taylor SM, Bennett MB (2008) Cephalopod dietary specialization and

ontogenetic partitioning of the Australian weasel shark Hemigaleus

australiensis White, Last & Compagno. J Fish Biol 72:917Ȃ936

Thiemann GW, Iverson SJ, Stirling I, Obbard ME (2011) Individual patterns of

prey selection and dietary specialization in an Arctic marine carnivore.

Oikos 120:1469Ȃ1478

Thorburn DC, Gill HS, Morgan DL (2014) Predator and prey interactions of fishes

of a tropical Western Australia river revealed by dietary and stable isotope

analyses. J R Soc West Aust 97:363Ȃ387

Thorburn DC, Morgan DL (2004) The northern river shark Glyphis sp. C

(Carcharhinidae) discovered in Western Australia. Zootaxa 685:1Ȃ8

Thorburn DC, Morgan DL, Rowland AJ, Gill HS (2007) Freshwater Sawfish Pristis

microdon Latham, 1794 (Chondrichthyes: Pristidae) in the Kimberley region

of Western Australia. Zootaxa 41:27Ȃ41

Thorburn DC, Rowland AJ (2008) Juvenile bull sharks Carcharhinus leucas

(Valenciennes, 1839) in northern Australian rivers. Beagle Rec Museums Art

Gall North Territ 24:79Ȃ86

182 Tillett BJ, Meekan MG, Field IC (2014) Dietary overlap and partitioning among

three sympatric carcharhinid sharks. Endanger Species Res 25:283Ȃ293

Tillett BJ, Meekan MG, Field IC, Thorburn DC, Ovenden JR (2012) Evidence for

reproductive philopatry in the bull shark Carcharhinus leucas. J Fish Biol

80:2140Ȃ58

Tilley A, López-Angarita J, Turner JR (2013) Diet reconstruction and resource

partitioning of a Caribbean marine mesopredator using stable isotope

Bayesian modelling. PLoS One 8:e79560

Tocher DR (2010) Fatty acid requirements in ontogeny of marine and freshwater

fish. Aquac Res 41:717Ȃ732

Tocher DR, (2003) Metabolism and functions of lipids and fatty acids in teleost

fish. Rev. Fish. Sci. 11:107Ȃ184.

Torres-Ruiz M, Wehr JD, Perrone AA (2007) Trophic relations in a stream food

web: importance of fatty acids for macroinvertebrate consumers. J North

Am Benthol Soc 26:509Ȃ522

Turchini GM, Francis DS, Senadheera SPSD, Thanuthong T, Silva SS De (2011)

Fish oil replacement with different vegetable oils in Murray cod: Evidence of

ƒDz‘‡‰ƒ-͵•’ƒ”‹‰‡ˆˆ‡ –dz by other dietary fatty acids. Aquaculture

315:250Ȃ259

Valen L Van (1965) Morphological variation and width of ecological niche. Am

Nat 99:377Ȃ390

Vaudo JJ, Heithaus MR (2011) Dietary niche overlap in a nearshore

elasmobranch mesopredator community. Mar Ecol Prog Ser 425:247Ȃ260

Uysal K, Aksoylar MY (2005) Seasonal variations in fatty acid composition and

the n-6/n-3 fatty acid ratio of pikeperch (Sander lucioperca) muscle lipids.

183 Ecol Food Nutr 44:23Ȃ35

Wai T-C, Yeung JWY, Lam VYY, Leung KMY, Dudgeon D, Williams GA (2012)

Monsoons and habitat influence trophic pathways and the importance of

terrestrial-marine linkages for estuary sharks. Ecosphere 3:art8

Ward DP, Pettit NE, Adame MF, Douglas MM, Setterfield SA, Bunn SE (2016a)

Seasonal spatial dynamics of floodplain macrophyte and periphyton

abundance in the Alligator Rivers region (Kakadu) of northern Australia.

Ecohydrology 9:1675Ȃ1686

Ward DP, Pettit NE, Adame MF, Douglas MM, Setterfield SA, Bunn SE (2016b)

Seasonal spatial dynamics of floodplain macrophyte and periphyton

abundance in the Alligator Rivers region (Kakadu) of northern Australia.

Ecohydrology 9:1675Ȃ1686

Warfe DM, Pettit NE, Davies PM, Pusey BJ, Hamilton SK, Kennard MJ, Townsend

SA, Bayliss P, Ward DP, Douglas MM, Burford MA, Finn M, Bunn SE, Halliday

ȋʹͲͳͳȌŠ‡Dz™‡–-†”›dz‹–Š‡™‡–-dry tropics drives river ecosystem

structure and processes in northern Australia. Freshw Biol 56:2169Ȃ2195

ƒ––ǡǤǤƒ† ‡”‰—•‘ǡǤ ǤǡʹͲͳͷǤ ƒ––›ƒ ‹†•ƒ†•–ƒ„Ž‡‹•‘–‘’‡•ȋɁ13C and

Ɂ15N) reveal temporal changes in narwhal (Monodon monoceros) diet linked

to migration patterns. Mar Mammal Sci 31:21-44.

Weigmann S (2016) Annotated checklist of the living sharks, batoids and

chimaeras (Chondrichthyes) of the world, with a focus on biogeographical

diversity. J Fish Biol 88:837Ȃ1037

Wetherbee BM, Cortés E (2004) Food consumption and feeding habits. In:

Carrier JC, Musick JA, Heithaus MR (eds) Biology of Sharks and their

Relatives. CRC Press LLC, p 225Ȃ246

184 White WT, Appleyard SA, Sabub B, Kyne PM, Harris M, Lis R, Baje L, Usu T, Smart

JJ, Corrigan S, Yang L, Naylor GJ (2015) Rediscovery of the threatened river

sharks, Glyphis garricki and G. glyphis, in Papua New Guinea. PLoS One

10:e0140075

White WT, Sommerville E (2010) Elasobrnachs of Tropical Marine Ecosystems.

In: Sharks and Their Relatives II: Biodiversity, Adaptive Physiology, and

Conservation.p 159Ȃ191

Whitty JM, Morgan DL, Peverell SC, Thorburn DC, Beatty SJ (2009) Ontogenetic

depth partitioning by juvenile freshwater sawfish (Pristis microdon:

Pristidae) in a riverine environment. Mar Freshw Res 60:306

Whitty JM, Phillips NM, Morgan DL, Chaplin JA, Thorburn DC, Peverell SC (2008)

Habitat associations of Freshwater Sawfish (Pristis microdon) and Northern

River Sharks (Glyphis sp. C): including genetic analysis of P. microdon across

northern Australia.

Wiley TR, Simpfendorfer CA (2007) The ecology of elasmobranchs occurring in

the Everglades National Park, Florida: implications for conservation and

management. Bull Mar Sci 80:171Ȃ189

Williams TM, Estes JA, Doak, Daniel F, Springer, Alan M (2004) Killer appetites:

Assessing the role of Predators in Ecological communities. Ecology

85:3373Ȃ3384

Winemiller KO, Jepsen DB (1998) Effects of seasonality and fish movement on

tropical river food webs. J Fish Biol 53:267Ȃ296

Winemiller KO, Polis GA (2013) Food webs: what can they us about the world?

In: Polis GA, Winemiller KO (eds) Food webs: integration of patterns and

dynamics, 2nd edn. Springer-Science+Buisness Media, BV, USA, p 1Ȃ22

185 Winnie JA (2012) Predation risk, elk, and aspen: tests of a behaviorally mediated

trophic cascade in the Greater Yellowstone Ecosystem. Ecology 93:2600Ȃ14

Wolanski E, Chappell J (1996) The response of tropical Australian estuaries to a

sea level rise. J Mar Syst 7:267Ȃ279

Wynen L, Larson HK, Thorburn DC., Peverell SC, Morgan DL, Field IC, Gibb K,

(2009) Mitochondrial DNA supports the identification of two endangered

river sharks (Glyphis glyphis and Glyphis garricki) across northern Australia.

Mar. Freshw. Res. 60:554Ȃ562.

Yick JL, Tracey SR, White RWG (2011) Niche overlap and trophic resource

partitioning of two sympatric batoids co-inhabiting an estuarine system in

southeast Australia. Appl Icthyology 27:1272Ȃ1277

Young JW, Hunt BP V, Cook TR, Llopiz JK, Hazen EL, Pethybridge HR, Ceccarelli D,

Lorrain A, Olson RJ, Allain V, Menkes C, Patterson TA, Nicol SJ, Lehodey P,

Kloser RJ, Arrizabalaga H, Anela Choy C (2015) The trophodynamics of

marine top predators: Current knowledge, recent advances and challenges.

Deep Sea Res Part II Top Stud Oceanogr 113:170Ȃ187

Zaccarelli N, Bolnick DI, Mancinelli G (2013) RInSp: an R package for the analysis

of individual specialization in resource use. Methods Ecol Evol 4:1018Ȃ1023

Zanden HB Vander, Bjorndal KA, Reich KJ, Bolten AB (2010) Individual

specialists in a generalist population: results from a long-term stable

isotope series. Biol Lett 6:711Ȃ714

Zanden MJ Vander, Fetzer WW (2007) Global patterns of aquatic food chain

length. Oikos 116:1378Ȃ1388

186 7 Appendix 7.1 Elasmobranch fatty acid data

Table A7.1 Mean values for all fatty acids (FA) (% mean of the relative abundance of FA> 0.5%), and their standard deviation in muscle tissue from six elasmobranchs collected from the South Alligator River, Kakadu National Park, Australia.

C. FA amboinensis C. leucas G. garricki G. glyphis U. dalyensis R. taylori 16:0 8.6 10.6±4.8 8.0±5.6 8.0±5.5 13.7±3.5 13.5±5.1 17:0 0.6 0.4±0.3 0.6±0.2 0.68±0.4 1.1±0.0 0.9±0.2 18:0 32.5 19.8±6.9 19.5±5.2 17.1±4.5 13.8±5.8 28.3±7.4 22:0 0.3 0.6±0.4 1.7±2.9 0.8±0.6 0.8±0.4 0.5±0.6 24:0 0.5 0.4±0.3 0.5±0.4 1.0±0.9 0.7±0.7 0.4±0.3 15:1 0.2 1.3±1.3 1.5±1.5 0.8±0.8 1.1±0.3 0.7±0.4 ͳ͸ǣͳɘ͹ 0.5 1.9±1.4 0.7±0.6 0.7±0.5 1.6±0.2 0.7±0.5 17:1 1.3 1.3±0.9 1.0±0.3 2.0 ±1.7 3.8±0.7 0.5±0.2 ͳͺǣͳɘͻ 14.1 16.2±6.3 11.2±4.9 8.2±3.6 13.8±0.5 9.5±2.7 ͳͺǣͳɘ͹ 6.2 4.9±2.5 5.2±1.8 4.3±2.0 4.0±1.3 7.2±2.0 ͳ͹ǣͳɘ͸ 0.5 0.5±0.2 0.5±0.3 0.6±0.5 0.8±0.4 0.6±0.3 ʹͲǣͳɘͻ 0.8 1.0±0.4 1.0±0.5 0.8±0.7 0.3±0.3 0.7±0.3 ʹʹǣͳɘͳͳ 0.1 1.7±5.0 0.2±0.2 0.1±0.0 0.1±0.1 0.5±2.1 ʹͶǣͳɘͻ 0.9 0.8±0.4 0.9±0.7 1.2±0.9 0.6±0.9 0.4±0.1 18:2bv 0.6 0.6±0.4 0.2±0.2 0.5±0.2 0.4±0.3 0.1±0.1 18:2cv 0.1 0.9±0.8 0.1±0.3 0.1±0.1 0.1±0.1 0.13±0.0 ͳͺǣʹɘ͸ 0.6 0.8±0.9 1.8±1.1 1.8±1.1 3.8±3.7 0.9±1.0 20.20 0.2 2.8±2.2 0.5±0.8 0.3±0.2 0.2±0.1 0.1±0.1 ʹͲǣʹɘ͸ 0.4 0.5±0.8 0.9±0.4 0.8±0.2 0.3±0.0 0.6±0.2 ʹͲǣ͵ɘͻ# 0.8 7.6±6.2 1.4±3.4 0.3±0.3 0.3±0.1 0.1±0.1 ʹͲǣ͵ɘ͸ 0.2 0.4±0.3 0.8±0.4 0.6±0.3 0.3±0.1 0.4±0.2 22:3 1.7 1.01±0.7 1.7±1.8 2.6±2.4 2.9±3.0 1.1±1.2 ʹͲǣͶɘ͸ 9.6 5.3±5.5 11.3±4.7 14.1±4.4 14.4±2.4 8.1±2.1 ʹʹǣͶɘ͸ 3.5 2.1±2.1 6.4±4.2 8.0±6.5 0.2±0.2 3.0±1.4 ʹͲǣͷɘ͵ 0.9 0.5±0.2 0.9±0.8 1.1±0.6 5.2±6.8 1.6±0.8 ʹʹǣͷɘ͵ 0.9 1. 9±1.2 1.5±2.0 1.6±1.8 1.1±1.6 1.1±1.7 ʹʹǣͷɘ͸ 1.8 1.2±0.7 2.3±1.1 2.3±1.1 1.2±0.8 2.2±0.8 ʹʹǣ͸ɘ͵ 4.8 4.8±2.3 8.3±4.6 9.8±8.8 4.3±0.2 11.4±5.6

ȭδͷΨ • 0.1 5.5±0.9 5.6±0.9 6.1±1.0 5.1±0.8 3.03±0.5

i17:0 0.6 0.8±0.5 1.0±2.8 0.6±0.4 0.7±0.6 0.3±0.1 16:0FALD 0.6 0.8±0.7 1.1±1.0 1.1±1.3 0.9±0.6 0.76±0.6 18:0FALD 1.6 0.9±0.6 1.3±1.1 1.7±1.3 2.2±1.4 0.5±0.2 FAs <0.5% include ͳͶǣͲͳͷǣͲǡƒͳͷǣͲǡͳͷǣͲǡͳͶǣͳǡͳ͸ǣͳɘͳ͵ǡͳ͸ǣͳɘͻǡͳ͸ǣͳɘ͹ǡͳ͸ǣͳɘͷǡ ͳ͹ǣͳɘͺΪƒͳ͹ǣͲǡͳͺǣͳɘ͹ǡͳͺǣͳZ5, 18:1, 19:1, 20:1Z7, 20:1ZͳͳǡʹͲǣͳɘͷǡʹʹǣͳɘͻǡʹʹǣͳɘ͹ǡʹͶǣͳɘͳͳǡ

187 ʹͶǣͳɘ͹ǡͳ͸ǣͶΪͳ͸ǣ͵ǡͳͺǣʹƒvǡͳͺǣͶɘ͵ǡͳͺǣ͵ɘ͸ǡͳͺǣ͵ɘ͵ǡʹͲǣͶɘ͵ȀʹͲǣʹǡʹͳǣͷɘ͵ǡʹͳǣ͵ǡʹʹǣʹƒv, 22:2bv, i16:0, 18:1FALD # 20:3Z9 identified based on comparison with other C. leucas fatty acid literature; a standard was not available at the time of analyses. v = unable to identify bonds as standard was not available at the time of analyses. FA - Fatty acids, SFA- saturated fatty acids, MUFA - monounsaturated fatty acids, PUFA - polyunsaturated fatty acids. FALD - fatty aldehyde analysed as dimethyl acetal.

188 7.2 Stable Isotope niche

Figure A7.1 (a) Bivariate plot of isotopic space depicting niche areas within standard ellipse (SEAC) of G13C and G15N of Carcharhinus leucas (blue) Glyphis garricki (black), and Rhizoprionodon taylori (red) from Kakadu National Park, Australia (b) Bayesian confidence intervals of isotopic niche area.

189 7.3 Fatty acid niche overlap

Figure A7.2 (a) Ten random elliptical projections of trophic niche region for each elasmobranch and pair of major essential fatty acids (FA) and (b) G. garricki wet vs dry season major (elliptical plots). Also displayed are one-dimensional density plots (lines) and two-dimensional scatterplots to demonstrate normality.

190 7.4 Indicator FAs from Legendre indicator species analysis

Table 7.2 Fatty acids (FA) within prey (Johnius novaeguineae, Lates calcarifer, Macrobrachium equidens, Neoarius armiger, Polydactylus macrochir and Rhinomugil nasutus) and sharks species (Carcharhinus leucas, Glyphis garricki, G. glyphis and Rhizoprionodon taylori) found in the South Alligator River, Kakadu, Australia that can be used as indictors for a specific species according to a Dufrêne- Legendre indicator species analysis (Dufrêne & Legendre 1997). Note: The analysis was unable to find indicator FAs for P. macrochir and L. calcarifer.

Indicator Indicator FAs Species Value P - value 20.3Z9 C. leucas 0.7 0.00 20.2 0.7 0.00 18.2c 0.3 0.00 18.1Z9 0.2 0.00 18.2b 0.2 0.00 20.1Z9 0.2 0.02 22.1Z11 0.4 0.05 15.1 G. garricki 0.2 0.01 20.2Z6 0.2 0.02 20.3Z6 0.2 0.04 22.0 0.2 0.06 16.0FALD 0.2 0.08 i17.0 0.2 0.60 20.4Z6 G. glyphis 0.2 0.00 18.0FALD 0.2 0.00 22.4Z6 0.2 0.00 24.1Z11 0.2 0.00 18.0 R. taylori 0.2 0.00 18.1Z7 0.2 0.00 24.0 J. novaeguineae 0.3 0.00 22.6Z3 0.2 0.00 22.5Z6 0.2 0.00 20.1Z7 0.3 0.00 24.1Z9 0.2 0.00 22.2a 0.2 0.03 18.2Z6 M. equidens 0.3 0.00 20.4Z3/20.2 0.3 0.00 20.5Z3 0.3 0.00 17.0 0.2 0.00 17.1 0.2 0.01 20.0 0.2 0.14 14.0 N. armiger 0.5 0.00 16.0 0.2 0.00 20.1Z11 0.2 0.00 18.3Z3 0.3 0.01

191 16.1Z7 R. nasutus 0.3 0.00 15.0 0.3 0.00 22.5Z3 0.3 0.00 17.1Z6 0.2 0.01 22.3 0.1 0.64

192 7.5 Putative prey fatty acid data

Table A7.3 Mean values for fatty acids (FA> 0.5%) and their standard deviation in muscle (SD) tissue from six potential prey species of shark collected from the South Alligator River, Kakadu National Park, Australia. Included are indicator FAs for each species with a p <0 .05

Polydactylus Johnius Lates Macrobrachium Nemapteryx Rhinomugil Species macrochir novaeguineae calcarifer equidens armiger nasutus 14:0 0.3±0.4 0.0±0.0 0.2±0.2 0.2±0 2.0±2.6 0.8±1.0 15:0 0.5±0.3 0.3±0.1 0.7±0.5 0.5±0.1 0.6±0.3 É 1.3±1.1 É 16:0 20.5±2.6 16.4±2.4 19.6±4.3 20.5±7.3 25.9±7.4 É 22.9±5.7 17:0 1.2±0.1 1.2±0.5 1.5±0.6 1.8±0.7 É 1.5±0.4 0.9±0.5 18:0 12.6±1.2 12.5±1.3 14.1±5.9 12±2.8 12±1.6 9.9±2.1 20:0 0.3±0.0 0.4±0.2 0.4±0.1 1.0±0.3 0.4±0.1 0.3±0.1 22:0 0.5±0.1 0.9±0.2 1±0.8 1.0±0.4 0.4±0.1 0.3±0.1 24:0 0.4±0.0 1.6±0.5É 0.7±0.4 0.3±0.1 0.2±0.1 0.6±0.4 15:1 0.5±0.4 0.3±0.1 0.7±0.7 0.7±0.3 0.4±0.4 0.3±0.2 ͳ͸ǣͳɘ͹ 3.1±1.7 1.8±0.9 2.8±2.3 2.0±1.2 3.0±1.7 5.8±4.1É 17:1 1.1±0.3 0.9±0.3 1.8±1.5 2.2±0.8 É 0.6±0.6 0.5±0.3 ͳ͹ǣͳɘ͸ 0.8±0.2 0.7±0.2 1.0±0.4 0.5±0.2 0.6±0.1 1.2±1.0 É ͳͺǣͳɘ͹ 3.4±0.5 3.0±1.4 3.2±1.1 2.6±0.8 3.2±0.6 3.4±0.7 ͳͺǣͳɘͻ 9.7±1.2 8.5±1.9 12.8±4.5 9.8±3.2 11.8±4.2 8.2±4.5 ʹͲǣͳɘ͹ 0.2±0.1 0.6±0.7É 0.1±0.0 0.0±0.0 0.2±0.1 0.3±0.2 ʹͲǣͳɘͻ 0.5±0.3 0.3±0.1 0.3±0.2 0.1±0.0 0.7±0.3 0.2±0.3 ʹͲǣͳɘͳͳ 0.1±0.1 0.1±0.1 0.3±0.2 0.2±0.1 0.5±0.3É 0.1±0.0 ʹͶǣͳɘͻ 0.5±0.4 1.4±0.5É 0.5±0.3 0±0.0 0.3±0.2 0.7±0.8 ʹͶǣͳɘͳͳ 0.0±0.0 0.1±0.0 0.1±0.1 0±0.0 0.0±0.0 0.1±0.2

< 0.5% 0.0±0.0 0.0±0.0 0.1±0.0 0.1±0.0 0.1±0.0 0.1±0.0

16:0FALD 0.5±0.4 0.5±0.2 0.3±0.3 0.1±0.2 0.4±0.4 0.4±0.4 18:0FALD 0.4±0.1 0.4±0.2 0.4±0.4 0.4±0.6 0.3±0.2 0.4±0.5 i17:0 0.3±0.1 0.3±0.0 0.6±0.2 0.3±0.1 0.5±0.1 0.0±0.0

É an indicator fatty acid for the species with p values < 0.05 based on Dufrêne and Legendre (1997), FA <0.5 includes 14:1, 15:0, 16:1Z5, 16:1Z7, 16:1Z9, 16:1Z13, 16:4+16:3, , 17:1Z8+17:0, 18:1, 18:1Z5, 18:1Z7, 18:2, 18:3Z6, 18:4Z3, 19:1, 20:1Z5, 21:5Z3, 21:3, 22:2b, 22:1Z7, 22:1Z9, , 24:1Z7, 18:1FALD, i15:0, i16:0 FALD - fatty aldehyde analysed as dimethyl acetal

193 7.6 Principal coordinate ordination of fatty acids within putative

prey

Figure A7.3 Principal coordinate ordination of essential fatty acids (EFA)>0.05% in percentage abundance in prey species, with vector overlays of –Š‡‘•–‹ˆŽ—‡–‹ƒŽ •ȋ‡ƒ”•‘ǯ• ‘””‡Žƒ–‹‘ƒ„‘˜‡ͲǤͳȌǤ

194